Psychosocial interventions for men with prostate cancer

  • Review
  • Intervention

Authors


Abstract

Background

As the incidence and prevalence of prostate cancer continue to rise, the number of men needing help and support to assist them in coping with disease and treatment-related symptoms and their psychosocial effects is likely to increase.

Objectives

To evaluate the effectiveness of psychosocial interventions for men with prostate cancer in improving quality of life (QoL), self-efficacy and knowledge and in reducing distress, uncertainty and depression.

Search methods

We searched for trials using a range of electronic databases including the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE and PsycINFO to October 2013, together with handsearching of journals and reference lists.

Selection criteria

Randomised controlled trials of psychosocial interventions for men at any stage of prostate cancer. We included psychosocial interventions that explicitly used one or a combination of the following approaches: cognitive behavioural, psychoeducational, supportive and counselling. Interventions had to be delivered or facilitated by trained or lay personnel.

Data collection and analysis

Pairs of review authors independently extracted data and assessed risk of bias. We analysed data using standardised mean differences (SMDs), random-effects models and 95% confidence intervals (CIs).

Main results

Nineteen studies comparing psychosocial interventions versus usual care in a total of 3204 men with prostate cancer were included in this review. All but three of these studies were conducted in the United States.

Men in the psychosocial intervention group had a small, statistically significant improvement in the physical component of general health–related quality of life (GHQoL) at end of intervention (1414 participants, SMD 0.12, 95% CI 0.01 to 0.22) based on low-quality evidence. A small improvement in favour of psychosocial interventions (SMD 0.24, 95% CI 0.02 to 0.47) was also seen in the physical component of GHQoL at end of intervention for group-based interventions. No clear evidence of benefit was found for GHQoL scores at end of intervention with individual-based interventions compared with controls. Also, no clear evidence suggested that psychosocial interventions were beneficial in improving the physical component of GHQoL at four to six and at eight to 12 months post-intervention. In addition, no clear evidence showed benefit associated with psychosocial interventions for the mental component of GHQoL at end of intervention (1416 participants, SMD -0.04, 95% CI -0.15 to 0.06) based on moderate-quality evidence. Results for the mental component of GHQoL at four to six and at eight to 12 months post-intervention were compatible with benefit and harm. At end of intervention, cancer-related QoL showed a small improvement following psychosocial interventions (SMD 0.21, 95% CI 0.04 to 0.39), but at eight and 12 months, the effect was compatible with benefit and harm. For prostate cancer–specific and symptom-related QoL, the differences between groups were not significant.

No clear evidence indicated that psychosocial interventions were beneficial in improving self-efficacy at end of intervention (337 participants, SMD 0.16, 95% CI -0.05 to 0.38) based on very low-quality evidence in three studies that assessed individual-based interventions. The results for self-efficacy at six to eight and at 12 months post-intervention were compatible with benefit and harm. Men in the psychosocial intervention group had a moderate increase in prostate cancer knowledge at end of intervention (506 participants, SMD 0.51, 95% CI 0.32 to 0.71) based on very low-quality evidence in two studies; this increase was also observed in the subgroups of group-based and individual-based interventions. A small increase in knowledge with psychosocial interventions was noted at three months post-intervention (SMD 0.31, 95% CI 0.04 to 0.58).

The results for uncertainty (916 participants, SMD -0.05, 95% CI -0.35 to 0.26) and distress (916 participants, SMD 0.02, 95% CI -0.11 to 0.15) at end of intervention were compatible with both benefit and harm based on very low-quality evidence. No clear evidence suggests that psychosocial interventions were beneficial in reducing uncertainty and distress between groups at six to eight and at 12 months post-intervention. Finally, no clear evidence of benefit is associated with psychosocial interventions for depression at end of intervention (434 participants, SMD -0.18, 95% CI -0.51 to 0.15) based on very low-quality evidence. Individual-based interventions significantly reduced depression when compared with usual care groups. The results for depression at six and at 12 months post-intervention were compatible with benefit and harm.

The overall risk of bias in the included studies was unclear or high, primarily as the result of performance bias.

No data regarding stage of disease or treatment with androgen deprivation therapy (ADT) were extractable for subgroup analysis. Only one study addressed adverse effects. High attrition could indicate that some participants may not have been comfortable with the interventions.

Authors' conclusions

Overall, this review shows that psychosocial interventions may have small, short-term beneficial effects on certain domains of well-being, as measured by the physical component of GHQoL and cancer-related QoL when compared with usual care. Prostate cancer knowledge was also increased. However, this review failed to demonstrate a statistically significant effect on other domains such as symptom-related QoL, self-efficacy, uncertainty, distress or depression. Moreover, when beneficial effects were observed, it remained uncertain whether the magnitude of effect was large enough to be considered clinically important. The quality of evidence for most outcomes was rated as very low according to GRADE, reflecting study limitations, loss to follow-up, study heterogeneity and small sample sizes. We were unable to perform meaningful subgroup analyses based on disease stage or treatment modality. Although some findings of this review are encouraging, they do not provide sufficiently strong evidence to permit meaningful conclusions about the effects of these interventions in men with prostate cancer. Additional well-done and transparently reported research studies are necessary to establish the role of psychosocial interventions in men with prostate cancer.

Résumé scientifique

Interventions psychosociales chez les hommes atteints d'un cancer de la prostate

Contexte

Comme l'incidence et la prévalence du cancer de la prostate continuent d’augmenter, le nombre d'hommes nécessitant une aide et un soutien pour faire face à la maladie et aux symptômes liés au traitement et à leurs effets psychosociaux est susceptible d'augmenter.

Objectifs

Évaluer l'efficacité des interventions psychosociales chez les hommes atteints d'un cancer de la prostate dans l'amélioration de la qualité de vie (QdV), de l'auto-efficacité et des connaissances, ainsi quedans la diminution de la détresse, de l’incertitude et de la dépression.

Stratégie de recherche documentaire

Nous avons recherché des essais au moyen d'un éventail de bases de données électroniques, y compris le registre Cochrane des essais contrôlés (CENTRAL), MEDLINE, EMBASE et PsycINFO jusqu' à octobre 2013, ainsi que dans une recherche manuelle de journaux et de références bibliographiques.

Critères de sélection

Essais contrôlés randomisés d'interventions psychosociales chez les hommes à n'importe quel stade du cancer de la prostate. Nous avons inclus les interventions psychosociales qui utilisaient explicitement une ou une combinaison des méthodes suivantes: comportement cognitif, psychoéducation, soutien et conseils. Les interventions devaient être administrées ou facilitées par un personnel profane ou formé.

Recueil et analyse des données

Deux auteurs de la revue ont indépendamment extrait les données et évalué le risque de biais. Nous avons analysé les données en utilisant les différences moyennes standardisées (DMS), les modèles à effets aléatoires et les intervalles confiance (IC) à 95%.

Résultats principaux

Dix-neuf études comparant les interventions psychosociales versus les soins habituels chez un total de 3204 hommes atteints d'un cancer de la prostate ont été inclues dans cette revue. Toutes, sauf trois de ces études ont été réalisées aux États-Unis.

Les hommes dans le groupe d'intervention psychosociale présentaient une petite amélioration statistiquement significative dans la composante physique de la qualité de vie liée à la santé générale (QdVSG) à la fin de l'intervention (1414 participants, DMS de 0,12, IC à 95% de 0,01 à 0,22) sur la base de preuves de faible qualité. Une petite amélioration en faveur des interventions psychosociales (DMS 0,24, IC à 95% de 0,02 à 0,47) était également observée dans la composante physique de QdVSG à la fin de l'intervention pour les interventions de groupe. Aucune preuve de bénéfice n'a été observée pour les scores de QdVSG à la fin de l'intervention avec des interventions individuelles par rapport aux groupes témoins. De même, aucune preuve claire ne suggérait que les interventions psychosociales étaient bénéfiques pour améliorer la composante physique de QdVSGde4 à 6 et de 8 à 12 mois après l’intervention. En outre, aucune preuve claire n'a montré un bénéfice associé à des interventions psychosociales pour la composante mentale de QdVSG à la fin de l'intervention (1416 participants, DMS – de 0,04, IC à 95% - de 0,15 à 0,06), preuves de qualité modérée. Les résultats pour la composante mentale de QdVSGde4 à 6 et de 8 à 12 mois post-intervention étaient compatibles avec les effets bénéfiques et délétères. À la fin de l'intervention, la QdV liée au cancer a montré une petite amélioration après les interventions psychosociales (DMS 0,21, IC à 95% de 0,04 à 0,39), mais à 8 et 12 mois, l'effet était compatible avec des effets bénéfiques et délétères. Pour les symptômes de la QdV du cancer de la prostate, les différences entre les groupes n'étaient pas significatives.

Aucune preuve claire ne montrait que les interventions psychosociales étaient bénéfiques pour améliorer l'efficacité personnelle à la fin de l'intervention (337 participants, DMS de 0,16, IC à 95% -de 0,05 à 0,38), preuves de très faible qualité dans trois études ayant évalué des interventions individuelles. Les résultats pour l'auto-efficacité au bout de 6 à 8 et à 12 mois post-intervention étaient compatibles avec les effets bénéfiques et délétères. Les hommes dans le groupe d'intervention psychosociale présentaient une augmentation modérée des connaissances sur le cancer de la prostate à la fin de l'intervention (506 participants, DMS de 0,51, IC à 95% de 0,32 à 0,71), preuves de très faible qualité provenant de deux études; cette augmentation était également observée dans les interventions en sous-groupe et les interventions en individuel. Les connaissances augmentaient légèrement à trois mois post-intervention (DMS 0,31, IC à 95% de 0,04 à 0,58).

Les résultats pour l'incertitude (916 participants, DMS de -0,05, IC à 95% -de 0,35 à 0,26) et pour la détresse (916 participants, DMS de 0,02, IC à 95% - de 0,11 à 0,15) à la fin de l'intervention étaient compatibles avec les bénéfices et les effets délétères, preuves de très faible qualité. Aucune preuve claire ne suggère que les interventions psychosociales étaient bénéfiques en termes de réduction de l'incertitude et de la détresse entre les groupes au bout de 6 à 8 et à 12 mois post-intervention. Enfin, aucune preuve de bénéfice n’est associée à des interventions psychosociales pour la dépression à la fin de l'intervention (434 participants, DMS de -0,18, IC à 95% - de 0,51 à 0,15), preuves de très faible qualité. Les interventions individuelles réduisaient significativement la dépression par rapport aux soins habituels en groupe. Les résultats pour la dépression à 6 et à 12 mois post-intervention étaient compatibles avec les effets bénéfiques et délétères.

Le risque de biais global dans les études incluses était incertain ou élevé, principalement en raison de la performance du biais.

Aucune donnée concernant la phase de la maladie ou le traitement par l’hormonothérapie anti-androgénique (ADT) étaient extractibles pour l'analyse en sous-groupes. Une seule étude examinait les effets indésirables. Le taux d'attrition élevé pourrait indiquer que certains participants n’étaient pas à l'aise lors de ces interventions.

Conclusions des auteurs

Dans l'ensemble, cette revue montre que les interventions psychosociales peuvent avoir des effets bénéfiques de petite taille et de courte durée dans certains domaines concernant le bien-être, tels que mesurés par la composante physique de la QdVSG liée au cancer en comparaison aux soins habituels. Les connaissances sur le cancer de la prostate augmentaient également. Cependant, cette revue n'a pas permis de démontrer d’effet statistiquement significatif dans d'autres domaines, tels que les symptômes relatifs à la qualité de vie, l'efficacité personnelle, l’incertitude, la détresse ou la dépression. De plus, lorsque les effets bénéfiques étaient observés, il n’était pas certain que l'ampleur de l'effet soit suffisamment vaste pour être considérée comme cliniquement importante. La qualité des preuves pour la plupart des résultats a été considérée comme très faible conformément à la méthode GRADE, ce qui reflète les limites des études, la perte de suivi,l'hétérogénéité des études et la taille réduite des échantillons. Nous n’avons pas été en mesure d'effectuer des analyses en sous-groupes basées sur le stade de la maladie ou la modalité de traitement. Bien que certains résultats de cette revue soient encourageants, ils ne fournissent pas suffisamment de preuves solides permettant d’apporter des conclusions significatives concernant les effets de ces interventions chez les hommes atteints d'un cancer de la prostate. Des études supplémentaires effectuées correctement et rapportées de manière transparente sont nécessaires afin d'établir le rôle des interventions psychosociales chez les hommes atteints d'un cancer de la prostate.

摘要

攝護腺癌男性患者的心理社會介入

背景

由於攝護腺癌的發生率和盛行率持續攀升,需要幫助和支持以協助他們因應疾病和治療相關症狀及其心理社會影響的男性人數,也可能增加。

目的

針對改善生活品質 (quality of life, QoL)、自我效能 (self-efficacy),和知識,以及減輕苦惱、不確定感和沮喪感,評估心理社會介入對攝護腺癌男性患者的療效。

搜尋策略

我們利用許多電子資料庫搜尋試驗,包括考科藍對照試驗中央註冊 (Cochrane Central Register of Controlled Trials, CENTRAL)、MEDLINE、EMBASE和PsycINFO (截至2013年10月為止),並以人工的方式搜尋期刊和參考文獻清單。

選擇標準

本次文獻回顧收錄任一期數攝護腺癌男性患者的心理社會介入隨機對照試驗。我們納入明確使用1種或併用數種下列方法的心裡社會介入:認知行為介入、心理教育介入、支持性介入和諮商。必須由受過訓練或外行人負責進行或協助執行介入措施。

資料收集與分析

由2位文獻回顧作者組成一對,獨立萃取資料並評估偏差風險。我們利用標準化平均差 (SMD)、隨機效果模式 (random-effect model) 和95%信賴區間 (CI) 進行資料分析。

主要結果

本次文獻回顧納入19篇比較心理社會介入和一般照護的試驗,共計包含3204名攝護腺癌男性患者。除了3篇試驗以外,所有其他試驗皆於美國進行。

心理社會介入組的男性患者,於介入結束時的一般健康相關生活品質 (general health–related quality of life, GHQoL) 生理層面,具有顯著的小幅改善 (1414名受試者,SMD為0.12,95% CI為0.01至0.22),證據品質偏低。至於接受團體介入的患者,於介入結束時,接受心理社會介入患者的GHQoL生理層面,也有些許改善 (SMD為0.24,95%CI為0.02至0.47)。並無明確證據顯示,相較於對照組,接受個別介入的患者,於介入結束時GHQoL分數有所改善。同樣亦無明確證據顯示,於介入後4至6個月和8至12個月時,心理社會介入對改善GHQoL的生理層面具有益處。此外,並無明確證據證明心理社會介入於介入結束時,對GHQoL的心理層面具有益處 (1416名受試者,SMD為 -0.04,95% CI為 -0.15至0.06),證據品質中等。於介入後4至6個月和8至12個月的GHQoL心理層面結果,正面和負面影響不相上下。在介入結束時,心理社會介入可略微改善癌症相關生活品質 (SMD為0.21,95% CI為0.04至0.39),但在8和12個月時,正面和負面的影響不相上下。對於攝護腺癌專屬生活品質和症狀相關生活品質,治療組間並無顯著差異。

並無明確證據顯示在介入結束時,心理社會介入有助於改善自我效能 (337名受試者,SMD為0.16,95% CI為 -0.05至0.38),但此項資料得自3篇證據品質極低的個別介入試驗。介入後6至8個月和12個月時,對自我效能的正面影響和負面影響相當。心理社會介入組的男性患者,於介入結束時對攝護腺癌的知識增加程度中等 (506名受試者,SMD為0.51,95% CI為0.32至0.71),此項資料來自2篇證據品質極低的試驗;至於團體介入和個別介入的子群體,知識的增加情況與上述相似。在介入後3個月,可見心理社會介入組患者的知識微幅增加 (SMD為0.31,95% CI為0.04至0.58)。

介入結束時對不確定感 (916名受試者,SMD為 -0.05,95% CI為 -0.35至0.26) 和苦惱 (916名受試者,SMD為 0.02,95% CI為 -0.11至0.15) 的結果,正面和負面影響相似,證據品質極低。並無明確證據顯示,在介入後6至8個月和12個月時,心理社會介入有助於減輕患者的不確定感和苦惱。最後,並無明確證據證明在介入結束時,心理社會介入對沮喪的情緒有所助益 (434名受試者,SMD為 -0.18,95% CI為 -0.51至0.15),證據品質極低。相較於一般照護組,個別介入可顯著減輕患者的沮喪感。在介入後第6個月和第12個月時,對沮喪結果的正面和負面影響相仿。

納入試驗的整體偏差風險不明或偏高,主要是表現性偏差所致。

並無關於疾病分期或雄性素阻斷療法 (androgen deprivation therapy, ADT) 的資料可供萃取,以進行子群體分析。只有1項試驗探究不良作用。高損耗表示接受介入可能使某些受試者感覺不自在。

作者結論

整體而言,此次文獻回顧顯示,相較於一般照護,心理社會介入對某些健康領域 (採用GHQoL生理層面和癌症相關生活品質進行測量),可能具有些微的短期效果。也可能可以增加患者對攝護腺癌的知識。不過本次文獻回顧未能證實對其他領域具統計顯著影響,例如症狀相關生活品質、自我效能、不確定感、苦惱或沮喪感。此外,即使觀察到正面效果,卻仍不確定介入的影響幅度是否足以視為具臨床重要性。依據GRADE,大部分結果的證據品質皆屬極低,反映這些試驗有其限制、無法追蹤受試者、試驗具異質性和收錄的樣本數少。我們無法依據疾病分期或治療模組,進行有意義的子群體分析。儘管本次文獻回顧的部分結果令人心生鼓舞,但並未提供充分證據,以致無法針對這些介入對攝護腺癌男性患者的療效,提出有意義的結論。必須藉助其他公開通報的優良研究試驗,確立心理社會介入對攝護腺癌男性患者的角色。

譯註


翻譯者:臺北醫學大學實證醫學研究中心
本翻譯計畫由衛生福利部補助經費,臺北醫學大學實證醫學研究中心、台灣實證醫學學會及東亞考科藍聯盟(EACA)統籌執行。

Zusammenfassung

Psychosoziale Interventionen für Männer mit Prostatakarzinom

Hintergrund

Da die Inzidenz und Prävalenz des Prostatakarzinoms weiter steigen, ist es wahrscheinlich, dass mehr Männer Hilfe und Unterstützung bei der Bewältigung der Krankheit und behandlungsbedingter Symptome sowie den psychosozialen Auswirkungen benötigen.

Ziele

Bewertung der Alltagswirksamkeit von psychosozialen Interventionen für Männer mit Prostatakarzinom zur Verbesserung der Lebensqualität, der Selbstwirksamkeit und des Krankheitswissens sowie zur Verringerung von Sorgen, Unsicherheit und Depression.

Literatursuche

Die Studiensuche umfasste eine Reihe von elektronischen Datenbanken, einschließlich ‚Cochrane Central Register of Controlled Trials‘ (CENTRAL), MEDLINE, EMBASE und PsycINFO bis Oktober 2013, kombiniert mit der manuellen Suche in Fachzeitschriften und Referenzlisten.

Auswahlkriterien

Randomisierte kontrollierte Studien über psychosoziale Interventionen für Männer in jedem Prostatakrebsstadium. Wir schlossen Studien mit psychosozialen Interventionen ein, die explizit eine einzelne oder eine Kombination der folgenden Methoden verwendeten: kognitive Verhaltenstherapie, Psychoedukation, Unterstützung und Beratung. Die Interventionen mussten durch geschultes Personal oder Laien erbracht oder ermöglicht werden.

Datenerhebung und -analyse

Zwei Review-Autoren extrahierten unabhängig voneinander die Daten und bewerteten das Risiko für Bias. Zur Datenanalyse benutzen wir die standardisierte mittlere Differenz (SMD), random-effects Modelle und 95% Konfidenzintervalle (KI).

Wesentliche Ergebnisse

Es wurden 19 Studien mit insgesamt 3.204 Männern in dieses Review eingeschlossen, die psychosoziale Interventionen mit der üblichen Versorgung verglichen. Alle bis auf drei dieser Studien wurden in den Vereinigten Staaten durchgeführt.

Männer der psychosozialen Interventionsgruppe wiesen eine kleine, statistisch signifikante Verbesserung der physischen Komponente der allgemeinen gesundheitsbezogenen Lebensqualität am Ende der Intervention auf (1.414 Teilnehmer, SMD 0,12, 95% KI 0,01 bis 0,22). Diese Aussage basiert auf niedriger Qualität der Evidenz. Eine kleine Verbesserung zu Gunsten der psychosozialen Interventionen (SMD 0,24, 95% KI 0,02 bis 0,47) war auch in der physischen Komponente der allgemeinen gesundheitsbezogenen Lebensqualität am Ende der gruppenbasierten Interventionen zu sehen. Im Vergleich von Einzelinterventionen mit Kontrollen wurde kein eindeutiger Nutzenbeweis für das Ergebnis der allgemeinen gesundheitsbezogenen Lebensqualität gefunden. Ebenfalls wurde keine eindeutige Evidenz gefunden, dass psychosoziale Interventionen zur Verbesserung der physischen Komponente der allgemeinen gesundheitsbezogenen Lebensqualität nach vier bis sechs und nach acht bis 12 Monaten postinterventionell nützlich waren. Darüber hinaus fanden wir postinterventionell keine eindeutige Evidenz für den Nutzen von psychosozialen Interventionen für die psychische Komponente der allgemeinen gesundheitsbezogenen Lebensqualität (1.416 Teilnehmer, SMD -0,04, 95% KI -0,15 bis 0,06). Diese Aussage basiert auf moderater Qualität der Evidenz. Die Ergebnisse der psychischen Komponente der allgemeinen gesundheitsbezogenen Lebensqualität wiesen postinterventionell nach vier bis sechs und nach acht bis 12 Monaten Nutzen und Schaden auf. Am Ende der psychosozialen Intervention zeigte sich eine kleine Verbesserung der krebsspezifischen Lebensqualität (SMD 0,21, 95% KI 0,04 bis 0,39), aber nach acht und 12 Monaten war die Wirkung für Nutzen und Schaden vergleichbar. Für die Prostatakrebs-spezifische und symptombezogene Lebensqualität waren die Unterschiede zwischen den Gruppen nicht signifikant.

Wir fanden keine eindeutige Evidenz, dass psychosoziale Interventionen am Ende nützlich bei der Verbesserung der Selbstwirksamkeit waren (337 Teilnehmer, SMD 0,16, 95% KI -0,05 bis 0,38). Dieses Ergebnis basiert auf einer sehr niedrigen Qualität der Evidenz aus drei Studien, die Einzelinterventionen beurteilten. Die Ergebnisse der Selbstwirksamkeit nach sechs bis acht und nach 12 Monaten postinterventionell waren bezüglich Nutzen und Schaden vergleichbar. Männer in der psychosozialen Interventionsgruppe hatten einen moderaten Wissenszuwachs über Prostatakrebs am Ende der Intervention (506 Teilnehmer, SMD 0,51, 95% KI 0,32 bis 0,71), basierend auf sehr niedriger Qualität der Evidenz aus zwei Studien. Dieser Zuwachs wurde auch in den Untergruppen der gruppenbasierten Interventionen und Einzelinterventionen beobachtet. Drei Monate nach der Intervention wurde eine kleine Wissenszunahme durch psychosoziale Interventionen festgestellt (SMD 0,31, 95% KI 0,04 bis 0,58).

Die Evidenz für Unsicherheit (916 Teilnehmer, SMD -0,05, 95% KI -0,35 bis 0,26) und Sorgen (916 Teilnehmer, SMD 0,02, 95% KI -0,11 bis 0,15) war am Ende der Intervention bezüglich Nutzen und Schaden vergleichbar, basierend auf sehr niedriger Qualität der Evidenz. Wir fanden keine klare Evidenz dafür, dass psychosoziale Interventionen in Bezug auf die Verringerung der Unsicherheit und Sorgen nach sechs bis acht und nach 12 Monaten postinterventionell nützlich waren. Schließlich zeigte sich am Ende von psychosozialen Interventionen keine eindeutige Evidenz für den Nutzen gegen Depressionen (434 Teilnehmer, SMD -0,18, 95% KI -0,51 bis 0,15). Die Aussagen basieren auf sehr niedriger Qualität der Evidenz. Im Vergleich zu den Kontrollgruppen mit üblicher Versorgung reduzierten Einzelinterventionen Depressionen deutlich. Die Ergebnisse für Depressionen nach sechs und 12 Monaten postinterventionell waren bezüglich Nutzen und Schaden vergleichbar.

Insgesamt war das Risiko für Bias der eingeschlossenen Studien unklar oder hoch, vor allem als Folge des Performance-Bias.

Für die Subgruppenanalyse waren keine Daten zum Krankheitsstadium oder zur Behandlung mit Androgendeprivationstherapie verfügbar. Nur eine Studie untersuchte unerwünschte Wirkungen. Der hohe Verlust bei der Nachbeobachtung könnte bedeuten, dass einige Teilnehmer sich mit den Interventionen möglicherweise unwohl gefühlt haben.

Schlussfolgerungen der Autoren

Zusammenfassend zeigt dieses Review, dass psychosoziale Interventionen im Vergleich zur üblichen Versorgung kleine und kurzfristige positive Auswirkungen auf bestimmte Bereiche des Wohlbefindens haben können, wie die Messungen der physischen Komponente der allgemeinen gesundheitsbezogenen und der krebsspezifischen Lebensqualität gezeigt haben. Das Krankheitswissen über Prostatakrebs nahm auch zu. Dennoch konnte dieses Review keine statistisch signifikante Wirkung in anderen Domänen, wie der symptombezogenen Lebensqualität, der Selbstwirksamkeit, Unsicherheit, Sorgen oder Depression zeigen. Wenn darüber hinaus eine vorteilhafte Wirkung beobachtet wurde, blieb es ungewiss, ob die Größe des Effekts groß genug war, um als klinisch relevant angesehen zu werden. Die Qualität der Evidenz für die meisten Endpunkte wurde nach GRADE als sehr gering bewertet, was Studienlimitationen, Verluste in der Nachbeobachtung, Studienheterogenität und kleine Stichproben wiederspiegelt. Wir waren nicht in der Lage, sinnvolle Subgruppenanalysen basierend auf dem Krankheitsstadium oder der Behandlungsmethode durchzuführen. Obwohl einige Ergebnisse dieses Reviews ermutigend sind, basieren sie auf nicht ausreichend starker Evidenz, um aussagekräftige Schlussfolgerungen über die Wirkung dieser Interventionen bei Männern mit Prostatakarzinom zu ermöglichen. Zusätzliche, gut geplante und transparent berichtete Studien sind notwendig, um die weitere Rolle von psychosozialen Interventionen bei Männern mit Prostatakarzinom zu ermitteln.

Anmerkungen zur Übersetzung

S. Schmidt, Koordination durch Cochrane Schweiz.

Plain language summary

Psychosocial interventions for men with prostate cancer

Approximately one in six men will be diagnosed with prostate cancer in their lifetime, and the number of men living with, or after, the diagnosis will continue to rise. Some of them will experience problems such as difficulty in passing urine, pain and sexual dysfunction, and these may cause distress, anxiety, uncertainty or depression. Their quality of life may therefore be affected. Health professionals often provide interventions to address some of these prostate cancer–related symptoms and the psychosocial problems that they may cause. However as yet, we do not know how effective, collectively, these interventions are.

In this review, we assessed studies that compared the effectiveness of psychosocial interventions versus the usual care that participants received. The evidence is current to October 2013. Nineteen studies, with a total of 3204 men at different stages of prostate cancer, were included. All but three of these studies were carried out in the United States. The review focused on the following outcomes: quality of life, self-efficacy, knowledge, uncertainty, distress and depression.

The results show that psychosocial interventions resulted in small improvements in quality of life at end of intervention; however, it is not clear whether these small improvements actually made a difference to participants. The interventions were also effective in increasing knowledge up to three months after the intervention. No evidence suggests that these interventions improved self-efficacy or reduced uncertainty, distress and depression in men with prostate cancer. No clear evidence shows that group-based interventions were, on the whole, more effective than individual-based ones for most of the outcomes measured in this review. All of these findings should be treated with caution, as the quality of the evidence was largely rated as very low. Some of these studies were small, and the number of participants who dropped out of some studies was high. The men were at various stages of prostate cancer. Interventions also varied in terms of content, delivery and how long they lasted. More and better research is needed to determine which types of psychosocial interventions are effective for meeting the needs of men at different stages of prostate cancer.

Résumé simplifié

Interventions psychosociales chez les hommes atteints d'un cancer de la prostate

Environ un homme sur six est diagnostiquéavec un cancer de la prostate et le nombre d'hommes vivants avec, ou ayant souffert de ce cancer, continue d’augmenter. Certains d'entre eux vont ressentir des problèmes, tels que des difficultés à uriner, de la douleur et un dysfonctionnement sexuel, et ceux-ci peuvent provoquer de la détresse, de l'anxiété, de l'incertitude ou de la dépression. Leur qualité de vie peut donc être affectée. Les professionnels de la santé fournissent souvent des interventions visant à corriger certains de ces symptômes liés au cancer de la prostate pouvant entraîner des problèmes psychosociaux. Cependant, nous ne connaissons pas encore l’efficacité de ces interventions collectives.

Dans cette revue, nous avons évalué les études qui comparaient l'efficacité des interventions psychosociales par rapport aux soins habituels reçus par les participants. Les preuves sont à jour en octobre 2013. Dix-neuf études, avec un total de 3204 hommes à différents stades de cancer de la prostate, ont été inclues. Toutes, sauf trois de ces études ont été réalisées aux États-Unis. La revue s'est concentrée sur les critères de jugement suivants : la qualité de vie, l'efficacité personnelle, les connaissances, l’incertitude, l'angoisse et la dépression.

Les résultats montrent que les interventions psychosociales entraînaient de petites améliorations sur la qualité de vie à la fin de l'intervention; cependant, il n'est pas clair si ces petites améliorations apportaientun changement significatif chez les participants. Les interventions étaient également efficaces pour augmenter les connaissances jusqu' à trois mois après l'intervention. Aucune preuve ne suggère que ces interventions amélioraient l'efficacité personnelle ou réduisaient l'incertitude, l'angoisse et la dépression chez les hommes atteints d'un cancer de la prostate. Dans l'ensemble, aucune preuve claire ne montre que les interventions de groupes étaient plus efficaces que les interventions individuelles pour la plupart des critères de jugement mesurés dans cette revue. Tous ces résultats doivent être considérés avec prudence, car la qualité des preuves a été largement considérée comme étant très faible. Certaines de ces études étaient de petite taille et le nombre de participants ayant abandonné était élevé. Les hommes étaient à différents stades de cancer de la prostate. Les interventions variaient également en termes de contenu, de délivrance et de leur duration. D'autres recherches de meilleure qualité sont nécessaires pour déterminer quels types d'interventions psychosociales sont efficaces pour répondre aux besoins des hommes à différents stades de cancer de la prostate.

Notes de traduction

Traduit par: French Cochrane Centre 14th January, 2014
Traduction financée par: Financeurs pour le Canada : Instituts de Recherche en Santé du Canada, Ministère de la Santé et des Services Sociaux du Québec, Fonds de recherche du Québec-Santé et Institut National d'Excellence en Santé et en Services Sociaux; pour la France : Ministère en charge de la Santé

淺顯易懂的口語結論

攝護腺癌男性患者的心理社會介入

約有六分之一的男性在其一生中可能被診斷罹患攝護腺癌,且罹患攝護腺癌的男性患者人數仍持續增加。部分患者可能出現諸如解尿困難、疼痛和性功能障礙等問題,這些問題可能讓他們感到苦惱、焦慮、產生不確定感或感覺沮喪。患者的生活品質可能因此受到影響。健康專業人員經常提供介入,協助患者處理一些攝護腺癌相關症狀,以及導因於症狀的心理社會問題。無論如何,迄今為止我們仍不清楚這些介入的整體效果。

在本次文獻回顧中,我們針對比較心理社會介入和一般照護療效的試驗進行評估。最新資料可追溯至2013年10月。本次文獻回顧納入19篇試驗,共計收錄3204名男性患者。除了3篇試驗以外,所有其他試驗皆於美國進行。本次文獻回顧著重下列結果:生活品質、自我效能、知識、不確定感、苦惱和沮喪感。

分析結果顯示在介入結束時,心理社會介入可微幅改善患者的生活品質;不過並不清楚這些微小的改善,是否確實能使患者變得不同。心理社會介入也可有效增加患者的知識,直至介入結束後3個月。並無證據顯示這些介入可改善攝護腺癌男性患者的自我效能,或減輕他們的不確定感、苦惱和沮喪感。就本次文獻回顧測量的大部分結果,並無明確證據顯示團體介入 (就整體而論) 的效果,優於個別介入。所有上述結果皆須謹慎解讀,因為大部分試驗的證據品質極低。部分試驗規模很小,有些試驗退出的患者人數很多。試驗收錄的受試者為各種期別攝護腺癌的男性患者。介入的內容、進行方式和持續時間也不盡相同。必須進行更多且更優良的研究,才能判斷哪些心理社會介入,可有效滿足各種期別攝護腺癌男性患者的需求。

譯註


翻譯者:臺北醫學大學實證醫學研究中心
本翻譯計畫由衛生福利部補助經費,臺北醫學大學實證醫學研究中心、台灣實證醫學學會及東亞考科藍聯盟(EACA)統籌執行。

Laienverständliche Zusammenfassung

Psychosoziale Interventionen für Männer mit Prostatakarzinom

Bei etwa einem von sechs Männern wird im Laufe ihres Lebens Prostatakrebs diagnostiziert, und die Anzahl der Männer, die mit oder nach dieser Diagnose leben wird weiter steigen. Bei einigen von ihnen werden Probleme wie Schwierigkeiten beim Wasserlassen, Schmerzen und sexuelle Funktionsstörungen auftreten, die wiederum Kummer, Angstgefühle, Unsicherheit oder Depression verursachen können. Ihre Lebensqualität kann daher beeinträchtigt werden. Gesundheitsfachleute bieten oft Interventionen an, um einige dieser Symptome von Prostatakrebs und die psychosozialen Probleme, die sie verursachen können anzugehen. Aber noch wissen wir nicht, wie effektiv diese Interventionen insgesamt sind.

In diesem Review untersuchten wir Studien, die die Alltagswirksamkeit von psychosozialen Interventionen mit der üblichen Pflege, die die Teilnehmer erhielten verglichen. Die Evidenz ist auf dem Stand von Oktober 2013. Neunzehn Studien mit insgesamt 3.204 Männern mit verschiedenen Stadien von Prostatakrebs, wurden eingeschlossen. Alle bis auf drei dieser Studien wurden in den Vereinigten Staaten durchgeführt. Dieser Review fokussiert die folgenden Endpunkte: Lebensqualität, Selbstwirksamkeit, Unsicherheit, Kummer und Depression.

Die Ergebnisse zeigen, dass psychosoziale Interventionen zu kleinen Verbesserungen in der Lebensqualität am Ende der Intervention führen; es ist jedoch nicht klar, ob diese kleinen Verbesserungen tatsächlich einen Unterschied für die Teilnehmer bewirken. Die Interventionen waren auch bei der Erhöhung des Wissens bis zu drei Monate nach der Intervention wirksam. Keine Evidenz deutet darauf hin, dass diese Interventionen die Selbstwirksamkeit verbessern oder Unsicherheit, Kummer und Depressionen bei Männern mit Prostatakrebs reduzieren. Keine klare Evidenz zeigt, dass gruppenbasierte Interventionen insgesamt wirksamer auf die Endpunkte waren, die in diesem Review gemessen wurden, als Individuelle Interventionen. Alle Ergebnisse sollten mit Vorsicht interpretiert werden, da die Qualität der Evidenz weitgehend als sehr niedrig bewertet wurde. Manche Studien waren klein und die Anzahl der Teilnehmer, die die Studien abbrachen war hoch. Die Männer hatten verschiedene Stadien des Prostatakrebs. Die Interventionen unterschieden sich auch hinsichtlich ihres Inhalts, der Vorgehensweise und Dauer. Mehr und bessere Forschung ist notwendig um festzustellen, welche Art von psychosozialen Interventionen wirksam für die Bedürfnisse von Männern mit Prostatakrebs in verschiedenen Stadien ist.

Anmerkungen zur Übersetzung

S. Schmidt, Koordination durch Cochrane Schweiz.

Summary of findings(Explanation)

Summary of findings for the main comparison. Psychosocial interventions compared with usual care for men with prostate cancer
  1. 1Lack of blinding.
    2Lack of allocation concealment.
    3Selective reporting.
    495% CI crosses the minimal important difference.
    5I2 > 50%.
    6Unclear blinding.
    7Unclear allocation concealment.
    8Significant loss to follow-up.
    9Instruments used were not validated.
    10Findings consistent with both clinically significant benefit and harm.

Psychosocial interventions compared with usual care for men with prostate cancer
Patient or population: patients with prostate cancer
Settings: hospital and community
Intervention: psychosocial interventions
Comparison: usual care
OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
Usual care Psychosocial interventions
General healthrelated quality of life: physical component at end of interventionSee commentMean general health–related quality of life: physical component at end of intervention in the intervention groups was
0.12 standard deviations higher
(0.01 to 0.22 higher)
-1414
(six studies)
⊕⊕⊝⊝
low 1,2,3,4
SMD 0.12 (95% CI 0.01 to 0.22)
General healthrelated quality of life: mental component at end of interventionSee commentMean general health–related quality of life: mental component at end of intervention in the intervention groups was
0.04 standard deviations lower
(0.15 lower to 0.06 higher)
-1416
(six studies)
⊕⊕⊕⊝
moderate 1,2,3
SMD -0.04 (95% CI -0.15 to 0.06)
Self-efficacy at end of interventionSee commentMean self-efficacy at end of intervention in the intervention groups was
0.16 standard deviations higher
(0.05 lower to 0.38 higher)
-337
(three studies)
⊕⊝⊝⊝
very low 1,2,4,5
SMD 0.16 (95% CI -0.05 to 0.38)
Prostate cancer knowledge at end of interventionSee commentMean prostate cancer knowledge at end of intervention in the intervention groups was
0.51 standard deviations higher
(0.32 to 0.71 higher)
-506
(two studies)
⊕⊝⊝⊝
very low 6,7,8,9
SMD 0.51 (95% CI 0.32 to 0.71)
Uncertainty at end of interventionSee commentMean uncertainty at end of intervention in the intervention groups was
0.05 standard deviations lower
(0.35 lower to 0.26 higher)
-916
(two studies)
⊕⊝⊝⊝
very low 1,2,5,10
SMD -0.05 (95% CI -0.35 to 0.26)
Distress at end of interventionSee commentMean distress at end of intervention in the intervention groups was
0.02 standard deviations higher
(0.11 lower to 0.15 higher)
-916
(two studies)
⊕⊝⊝⊝
very low 1,2,10
SMD 0.02 (95% CI -0.11 to 0.15)
Depression at end of interventionSee commentMean depression at end of intervention in the intervention groups was
0.18 standard deviations lower
(0.51 lower to 0.15 higher)
-434
(three studies)
⊕⊝⊝⊝
very low 3,5,6,7,10
SMD -0.18 (95% CI -0.51 to 0.15)
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval; SMD: Standardised mean difference.
GRADE Working Group grades of evidence.
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Background

Description of the condition

Prostate cancer is the fifth most common cancer in the world among individuals of both sexes combined, and it is the second most common cancer in men (Cancer Research UK 2013). In the United Kindom (UK), prostate cancer is the most common cancer in men, with around 41,000 new cases reported in 2010 (Cancer Research UK 2013). The incidence figures vary according to race and country of residence. About one in six men will be diagnosed with prostate cancer during their lifetime (American Cancer Society 2013). The highest incidence of the disease is seen in Australia and New Zealand, followed by North America and Europe. Incidence rates in the Western developed countries continue to rise. In the United States of America (USA) alone, it was estimated that 238,590 new cases of prostate cancer would be diagnosed in 2013 (American Cancer Society 2013). Over the past 25 years, the five-year relative survival rate (adjusted for normal life expectancy) for all stages of prostate cancer combined increased from 68% to 100% (American Cancer Society 2013), and more than 2.5 million men who are currently living in the USA have been diagnosed with the disease (Prostate Cancer Foundation 2013). This had led to increasing focus on survivorship issues, which researchers have started to address over the past two decades. A large body of evidence suggests that a diagnosis of cancer and the treatment that follows give rise to significant psychosocial problems, including distress, anxiety and depression. Quality of life in men with localised prostate cancer, especially in the first six months post-diagnosis, may be reduced. Reeve 2012 found significant decrements in physical, mental and social aspects of cancer patients’ lives when compared with men without cancer.

The cancer journey does not end with treatment. Several studies have reported problems related to sexual function and urinary incontinence, relationships (between partners and within the family) and financial strain. Anxiety and uncertainty about disease recurrence and spread of the cancer are constant worries for some. Most men with prostate cancer adapt and cope with the disease and with its treatment, but a significant minority (almost a third) have moderate and severe unmet needs for psychosocial support (Ames 2009; Ernstmann 2009; White 2012). Health professionals and researchers are rising to the challenges presented by prostate cancer by developing and testing interventions designed for those men with cancer and their partners who need help and support. To date, no systematic review of randomised controlled trials has evaluated the effectiveness of psychosocial interventions for men with prostate cancer in terms of quality of life, self-efficacy, knowledge, uncertainty, distress and depression.

Description of the intervention

Different types of psychosocial interventions are available, such as psychoeducational therapy, cognitive behavioural therapy (group or individual), group supportive therapy and individual supportive therapy, including counselling (Newell 2002). In this review, we have used the term 'psychosocial intervention' to describe all of the above, as well as those interventions designed to provide information, physical exercise, relaxation, or art and music therapies. Psychosocial interventions can be delivered face-to-face, online, by telephone or through a combination of these approaches.

How the intervention might work

Studies on the effects of a cancer diagnosis and of the ensuing treatment have shown that the overall quality of life of some sufferers can be reduced (McCaughan 2007). Psychosocial interventions have been developed to help men cope with the negative effects of a cancer diagnosis. Most psychosocial interventions are based on theories that emphasise the importance of self-management and positive coping strategies in maintaining or enhancing the person’s self-efficacy and competence to deal with the effects of the illness. Two of the most common theories that underpin self-management programmes and behaviour change are social cognitive theory (Bandura 1997) and stress and coping (Lazarus 1984). Self-efficacy, which is a central construct of the social cognitive theory, is the belief in one’s capabilities to organise and execute the courses of action required to manage prospective situations (Bandura 1997).

The transactional model of stress and coping (Lazarus 1984) serves as a framework for evaluating the processes of coping with stressful events. When people are faced with a threat, they appraise the threat in terms of how challenging or controllable it is (primary appraisal). They also appraise their own and external resources at their disposal to help them face the threat (secondary appraisal). How they eventually cope depends on these appraisals. If the disease is seen as overwhelming and their own and external resources inadequate, individuals may feel incapable of coping (Mishel 2002). The aim of psychosocial interventions, in this case, is to help the patient to cognitively reframe anxiety- and uncertainty-provoking events as something more manageable and less threatening (Mishel 2002).

Cognitive behavioural change interventions commonly include goal setting, motivational interviewing, problem-solving and coping skills training, environmental change (barrier reduction), behavioral contracting, self-monitoring and use of incentives/rewards and social support, in addition to information and homework/skill rehearsal (Peyrot 2007).

Why it is important to do this review

This review is timely because both the incidence and the prevalence of prostate cancer continue to rise. A substantial number of survivors and those close to them need help and support to cope with the physical and psychosocial problems that may arise. “A cultural shift in the approach to care and support [is needed] for people affected by cancer to a greater focus on recovery, health and well-being after treatment" (DoH 2010).

Policymakers and practitioners require evidence to draw upon when dealing with psychosocial problems in this population. Organisations such as the National Institute for Health and Care Excellence need evidence on which to base their guidelines (Wilt 2001).

A systematic review of the effectiveness of these interventions will inform us of the nature of the evidence so far. Equally important is the need to review the quality of these studies, the types of interventions, the outcome measures and the tools or instruments used to collect and analyse data. Recommendations for future research may help others to address the limitations of previous studies and to continue to add to the body of knowledge in this area. This may lead to better use of research funding and evidence of better quality provided to practitioners.

Objectives

To evaluate the effectiveness of psychosocial interventions for men with prostate cancer in improving quality of life (QoL), self-efficacy and knowledge and in reducing distress, uncertainty and depression.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised controlled trials only.

Types of participants

Studies including men diagnosed with prostate cancer (any stage) or mixed cancers were eligible if separate data for men with prostate cancer were available. Studies were eligible regardless of whether they included partners or couples.

Types of interventions

We included psychosocial interventions explicitly using one or a combination of the following approaches.

  • Cognitive behavioural therapy.

  • Psychoeducational therapy.

  • Supportive therapy.

  • Counselling.

We included individual-based and group-based interventions, whether delivered via telephone, online, at home or during clinic visits (or a combination of these) by lay or trained personnel (or a combination of both).

Eligible control intervention was the usual care received by participants as part of their medical treatment.

Studies were excluded from the review if they:

  • involved surgical or pharmacological interventions;

  • involved alternative therapy interventions such as acupuncture, massage or herbal remedies;

  • involved people with different types of cancer, whereby data specifically related to prostate cancer were not available; or

  • involved peer support groups or one-to-one peer support without facilitation/input from a professional or a trained helper.

Types of outcome measures

As consistent with psychosocial intervention research, a large number of outcomes, using a range of standardised and validated tools, were measured in the included studies. Researchers in these studies did not always use the same scales to measure the same outcomes. When possible and appropriate, we compared the results of different tools for the same outcomes. The outcomes and tools for each of the included studies are listed in the Characteristics of included studies. They are also summarised in the Main results section.

Primary outcomes

Primary outcomes were:

  • quality of life;

  • self-efficacy;

  • knowledge;

  • uncertainty;

  • distress; and

  • depression.

Self-efficacy in this review refers to belief in one's ability to deal with problems related to cancer (Weber 2007).

Uncertainty, in this context, is the inability to determine the meaning of illness-related events (Mishel 2002).

Secondary outcomes

None.

Search methods for identification of studies

See Appendices.

Electronic searches

We followed the principles for searching relevant studies as suggested by Egger 2001. The following electronic databases were searched.

  • The Cochrane Central Register of Controlled Trials (CENTRAL; The Cochrane Library, 2013, Issue 9).

  • MEDLINE (1946 to October Week 1 2013).

  • EMBASE (1974 to 21 October 2013).

  • PsycINFO (1806 to October Week 2 2013).

Searches were carried out on Ovid, and the search strategies are provided in Appendix 1, Appendix 2, Appendix 3 and Appendix 4, respectively.

Two published papers were in Mandarin and one was in French. They were translated into English for this review by academics fluent in the relevant languages.

Searching other resources

Online facilities (on journal websites) such as ‘similar articles’, ‘related articles’ and ‘articles from the same author’ were used to locate papers on psychosocial interventions for prostate cancer survivors, and to find other papers that authors of included studies had published on the same trial. We handsearched reference lists of relevant studies and systematic reviews. Additionally, we handsearched the following journals from 2000 to September 2013.

  • European Journal of Cancer Care.

  • Journal of Advanced Nursing.

  • Journal of Clinical Nursing.

Data collection and analysis

Selection of studies

After removing duplicates, we assigned an identification code to all records retrieved through electronic searches. Study selection at this stage consisted of pairs of review authors reading titles and abstracts to decide whether to include, exclude or wait for the full paper to make a decision. The first review author (KP) formed a pair with each of four other review authors (SM, EM, CS and EJH), and each review pair independently assessed studies for inclusion. Another review author (JN), who was not included in the pairings, was consulted when a third opinion was required.

The three inclusion criteria, at this stage, were 'randomised controlled trial', 'psychosocial intervention' and 'men with prostate cancer'. Full papers of all selected titles and abstracts were obtained and read by each pair, and their decisions on whether the study met these three criteria were justified (on a template). As new records were uncovered through other sources, such as by handsearching of journals or reference lists of other papers, the same process was followed again.

Data extraction and management

We developed a form that could be used to standardise and record the data extracted from selected studies. The main information was recorded under the following headings: aims or hypotheses, methods (e.g. type of design, number of groups, number in each group at each follow-up time point), participants (e.g. stage of cancer, type of treatment, age, inclusion/exclusion criteria, ethnicity, baseline characteristics), intervention and control groups (e.g. type, content, duration and location of intervention, who delivered the intervention and information about the control group), outcomes and outcome measures (e.g. outcomes relevant to this review, measurement tools, validity and reliability of tools) and results (e.g. whether all outcomes were reported, tests carried out, loss to follow-up, intention-to-treat analysis, main findings). The same pairing process described under Selection of studies was followed to ensure objective and independent data extraction. To ensure rigour, review authors met to extract data from two papers to ‘pilot test’ the form, discuss any ambiguities and resolve interpretation differences.

Assessment of risk of bias in included studies

Each pair of review authors used the risk of bias tool recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011a) to independently assess the risk of bias in the selected studies. To ensure rigour and consistency and to avoid disparities in judgement, all those involved in this process met and ‘pilot tested’ this exercise by assessing the risk of bias in two studies and comparing the results. Each study was assessed for risk of bias in the following domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting and other possible sources of bias. The descriptors ‘high risk’, ‘low risk’ and ‘unclear risk’ were used to rate the level of bias. Decisions of each review author had to be supported by evidence (or lack thereof) from the published material. Differences in judgement within pairs of review authors were resolved through discussion or by seeking a third opinion from another review author (JN).

Measures of treatment effect

All reported data in the included studies for the outcomes of interest were continuous. Means and standard deviations (SDs) were extracted, and, when appropriate, standard errors (SEs) were converted into SDs for meta-analysis by using the standardised mean difference (SMD) and 95% confidence intervals (CIs). All data were analysed using Review Manager 5 (RevMan 2012).

Unit of analysis issues

All studies used a simple parallel-group clinical trial design. For analysis of data from studies with more than one psychosocial intervention group, we compared each psychosocial intervention group versus the control group separately. The population of the control group was divided, but the standard deviation was left unchanged, as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011b).

Dealing with missing data

In some cases, more than one paper was published on the same study. If two reports presented the same data, they were counted only once. Therefore each report of an included study was scanned for data. Authors were contacted for missing data required for meta-analysis and for assessment of the risk of bias. Published protocols and other papers published on the included studies were assessed for information missing from the main paper or report. Replacement values were not inputted. Missing data in individual studies are reported in a narrative form in the Main results section.

Assessment of heterogeneity

Heterogeneity was assessed by using the I² statistic. I² greater than 50% was investigated, as this may have represented substantial heterogeneity (Higgins 2002).

Assessment of reporting biases

Study authors were contacted and most responded, although not all missing information was available. We planned to assess the presence of publication bias by using a funnel plot; however, too few studies were included in the meta-analysis carried out in this review (Sterne 2011).

Data synthesis

Some outcomes were measured by using a range of scales. Standardised mean differences and random-effects models were used to account for heterogeneity in the studies. However, the fixed-effect model was selected for settings in which the random-effects model would be expected to distort the weighting considerably in favour of the smaller studies (Higgins 2002). The Cohen 'd' guide to interpretation of effect size was used, and the width of the confidence interval was taken into account (Higgins 2002). When quantitative data were missing, a narrative summary of findings was reported.

Subgroup analysis and investigation of heterogeneity

We proposed to carry out predefined subgroup analyses of different types of interventions, ethnic groups, age groups and stages of disease. When possible, subgroup analyses were performed to compare group-based and individual-based interventions. Data on ethnicity and on different age groups were not presented separately in the included studies. Only one study included an all-black American cohort, and another an all-Hispanic cohort. Data were insufficient for subgroup analysis related to stage of disease, and available data in the included studies were reported using different disease stage classification systems.

We attempted to carry out post hoc subgroup analysis related to treatment with androgen deprivation therapy—a potential confounding variable; however, the data were insufficient.

Sensitivity analysis

We planned to perform sensitivity analyses to assess the impact of risk of bias, missing data and heterogeneity indicated by an I2 statistic greater than 50%. However, too few trials were identified for this to be done in a meaningful way.

Summary of findings table

We used the GRADE framework to rate the quality of evidence for each main outcome and reported the results in a 'Summary of findings' table (Guyatt 2011). A 'high’ quality rating means that further research is very unlikely to change our confidence in the estimate of effect; a 'moderate’ quality rating means that further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. On the other hand, an evidence rating of 'low’ quality means that further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. A 'very low’ quality evidence rating means that we are very uncertain about the estimate.

Results

Description of studies

See Characteristics of included studies and Characteristics of excluded studies.

Results of the search

Figure 1 provides details on the numbers of records identified, accessed, assessed and selected. Nineteen randomised controlled trials were included in this review.

Figure 1.

Study flow diagram.

Included studies

Participants

A total of 3204 men with prostate cancer (mean 168.6; SD 158.6; range 29 to 740) were randomly assigned to psychosocial intervention or control groups in the 19 included studies. Of these, 2746 remained in the study until the final follow-up time point, yielding a completion rate of 85.7%. Attrition ranged between 1.7% (Davison 1997) and 36.5% (Parker 2009). A total of 1555 participants (mean 57.5; SD 31.6; range 14 to 372) were included in the 26 psychosocial intervention groups (some studies had more than one psychosocial intervention group). The number of participants in the 19 control groups was 1222 (mean 67.9; SD 78.6; range 15 to 368).

Five studies (Ames 2011; Bailey 2004; Campbell 2007; Manne 2011; Zhang 2006) were described by the authors as 'pilot studies’. All but three studies were carried out in the USA; one study (Berglund 2007) was performed in Sweden, one in Australia (Chambers 2013) and the other in Canada (Davison 1997). Seven studies (Campbell 2007; Giesler 2005; Manne 2011; Mishel 2002; Mishel 2009; Northouse 2007; Thornton 2004) included partners or primary support persons.

The ethnic composition of the study populations was not provided in two studies (Berglund 2007; Davison 1997). Most members of the sample (85%) in the Australian study Chambers 2013 were of "British/Scottish/Welsh/Irish descent; the remainder were of European, Asian, or other ethnic background" (personal correspondence). In the other 16 studies, 1444 participants (mean 103.4; SD 59.8; range 0 to 226) were described as 'white' or 'Caucasian'. The number of participants described as ‘African American’ or ‘black’ was 393 (mean 24.8; SD 27.1; range 5 to 105). In one small study (Campbell 2007), all of the 40 participants were African Americans. In another study (Mishel 2002), in which a deliberate attempt was made to recruit more African Americans, the proportion of this group was 44% (n = 105). The number of ‘Hispanic’ or ‘Latino’ participants was 193 in the seven studies that provided this information. In some studies, these individuals were probably included in the ‘other’ category. One study (Penedo 2006) had a large Hispanic population (78 of 191; 40.8%). In another study (Penedo 2007), the total population was Hispanic (n = 93).

Four studies (Chambers 2013; Davison 1997; Mishel 2009; Thornton 2004) recruited participants before they had received treatment. In 12 studies (Ames 2011; Berglund 2007; Campbell 2007; Giesler 2005; Lepore 2003; Manne 2011; Mishel 2002; Northouse 2007; Penedo 2006; Penedo 2007; Weber 2007; Zhang 2006), most or all of the participants had completed one or more forms of treatment (surgery, radiotherapy, chemotherapy or hormone therapy). One study (Parker 2009) recruited men who were undergoing radical prostatectomy, another (Carmack Taylor 2006) enrolled only those receiving hormone therapy and another (Bailey 2004) recruited only those who had been in 'watchful waiting'.

Stage of disease

Detailed figures on disease stage of participants were provided in eight studies (Bailey 2004; Chambers 2013; Giesler 2005; Lepore 2003; Manne 2011; Mishel 2002; Parker 2009; Thornton 2004). However, because of the diversity of stage classification systems in the selected studies (‘ABCD’ or 'Jewett', ‘TNM’ and ‘AJCC’), no overall accurate figures of the stage of disease of these participants could be calculated. The other 11 studies provided incomplete or vague information on the composition of their sample with regards to disease status. Participants in Ames 2011 were described as being in the biochemical recurrent stage. Berglund 2007 reported that 42 participants (20%) had metastasis but gave no precise figures on the stage of disease of the other participants. In Carmack Taylor 2006, the number of participants at the different stages of the disease was not given, although eligibility criteria included ‘those receiving continuous androgen–ablation therapy, regardless of disease stage'. Campbell 2007 did not report detailed figures on disease stage but described those eligible for inclusion as beyond the acute diagnosis and treatment phase for prostate cancer. Participants in Davison 1997 were described as ‘newly diagnosed’. In Mishel 2009, recruitment inclusion criteria included 'participants at T1 or T2 stage', and the final sample was described as “with localized or locally advanced prostate cancer”; no detailed figures were presented. Participants in Northouse 2007 were described as being in the newly diagnosed stage (65%), in the advanced stage (21%) and in the biochemical recurrence stage (14%). Eligibility criteria in Penedo 2007 included men with stage I or II localised prostate cancer within three months of diagnosis, but no figures were available on the disease stage of participants. Similarly in Zhang 2006, one of the inclusion criteria specified participants at stage I to III prostate cancer, although the numbers of participants at different stages of disease were not given.

Androgen deprivation therapy (ADT) can impact quality of life and depression, among other variables. In studies of psychosocial interventions for prostate cancer patients, ADT can be a confounding variable if it is not controlled. Of the 19 studies included in this review, five reported that some but not all participants received ADT or other forms of hormone therapy (Berglund 2007; Campbell 2007; Chambers 2013; Manne 2011; Northouse 2007). In one study (Carmack Taylor 2006), all participants received continuous ablation therapy, but in another (Weber 2007), no participants received ADT. In the remaining 12 studies, no information indicated whether participants received ADT, although the participants' stages of disease suggest that some may have received the treatment.

Interventions

A total of 26 psychosocial intervention groups and 19 control groups were included in the 19 included studies. Thirteen studies (Ames 2011; Bailey 2004; Campbell 2007; Chambers 2013; Davison 1997; Giesler 2005; Manne 2011; Northouse 2007; Penedo 2006; Penedo 2007; Thornton 2004; Weber 2007; Zhang 2006) each had one psychosocial intervention group and one control group. Five studies (Carmack Taylor 2006; Lepore 2003; Mishel 2002; Mishel 2009; Parker 2009) each had two psychosocial intervention groups and one control group. Berglund 2007 had three psychosocial intervention groups and one control group.

Interventions varied in terms of aims, types, delivery and dose (see Characteristics of included studies). Thirteen (Ames 2011; Berglund 2007; Campbell 2007; Carmack Taylor 2006; Giesler 2005; Lepore 2003; Manne 2011; Mishel 2002; Northouse 2007; Penedo 2006; Penedo 2007; Weber 2007; Zhang 2006) aimed to help men with prostate cancer cope with the effects of the disease and the treatment they received. The other six (Bailey 2004; Chambers 2013; Davison 1997; Mishel 2009; Parker 2009; Thornton 2004) targeted participants before they received treatment. Bailey 2004 focused on helping men to cognitively reframe and manage the uncertainty of watchful waiting. The intervention in Chambers 2013 was designed to reduce decision and cancer-specific distress and to improve quality of life. The intervention in Davison 1997 was intended to reduce stress and anxiety in relation to treatment decision making. Mishel 2009 targeted participants' decision-making consultation with their physicians. The intervention in Parker 2009 aimed to reduce presurgical stress. The aim in Thornton 2004 was to help participants learn basic skills to facilitate communication with the medical team during the presurgical consultation and during the recovery period. Bailey 2004 focused on helping men to cognitively reframe and manage the uncertainty of watchful waiting.

Different categorisations of psychosocial interventions have been assigned. Fawzy 1995 lists four: behavioural training, individual psychotherapy, group interventions and education. Raingruber 2011 also offers four categorisations: cognitive behavioural interventions, supportive interventions, group interventions and telephone-assisted interventions. In practice, not all psychosocial interventions fit neatly into one of these ideal types. Often interventions comprise elements from two or more of the intervention types listed above. An education intervention can also be supportive and can be delivered to participants in groups or by telephone. It is the theoretical perspective upon which an intervention is based and its main focus (i.e. behavioural change, education or support) that justify its description as cognitive behavioural, educational or supportive.

A review of psychological interventions for women with metastatic cancer (Mustafa 2013) categorised their interventions into two types: cognitive behavioural and supportive/expressive group therapy. In this Cochrane review, the interventions seem to fit two types: 'cognitive behavioural' and 'supportive/educational'. This categorisation is based on the explicit reference in some of these studies to the theoretical underpinnings of interventional approaches and to their main components. Raingruber 2011 describes cognitive behavioural interventions as those based on the belief that how individuals view situations influences their emotional responses or problem-solving abilities. The interventions in seven studies included in this review can be described as cognitive behavioural: the structured behavioral intervention (Ames 2011); the cognitive skills training intervention based on cognitive behavioural theoretical objectives (Campbell 2007); the lifestyle programme, based on the cognitive behavioural curriculum (Carmack Taylor 2006); the stress management intervention based on cognitive behavioural principles (Parker 2009); the cognitive-behavioural stress management intervention (Penedo 2006; Penedo 2007) and intimacy-enhancing therapy that uses techniques drawn from cognitive behavioural and behavioural marital therapy (Manne 2011).

Raingruber 2011 describes supportive interventions as those focused on supportive psychosocial interventions that include listening, validation, stress management, problem solving and education related to diagnosis. The interventions reported in the 12 other included studies can be described as supportive/educational as they comprised a number of these components and placed a strong focus on education. However, it must be emphasised that considerable overlap has been noted between these categories; therefore, it is not wise to carry out subgroup analyses to compare them.

The interventions in this review can be differentiated according to the format in which they were delivered, that is, in groups or on an individual basis (face-to-face or by telephone). The interventions in eight studies were group-based (Ames 2011; Berglund 2007; Carmack Taylor 2006; Lepore 2003; Manne 2011; Penedo 2006; Penedo 2007; Zhang 2006). The 11 remaining studies administered individual-based interventions by telephone (Bailey 2004; Campbell 2007; Chambers 2013; Mishel 2002; Mishel 2009), face-to-face (Davison 1997; Parker 2009; Thornton 2004; Weber 2007), face-to-face and by telephone (Giesler 2005) or by home visits and by telephone (Northouse 2007). As these two formats (group-based and individual-based) are distinct, subgroup analysis to compare their effects was carried out.

Frequency and duration of interventions

In accordance with the recommendations of Voils 2012, the dosage of interventions is reported here in terms of frequency, duration and amount. In ten studies (Ames 2011; Bailey 2004; Berglund 2007; Campbell 2007; Lepore 2003; Manne 2011; Mishel 2002; Penedo 2006; Penedo 2007; Weber 2007), the interventions were delivered once a week. This ranged between five and 10 weekly sessions. In two studies (Northouse 2007; Zhang 2006), the interventions were delivered bi-weekly, and in another (Giesler 2005), they were provided monthly. In Carmack Taylor 2006, interventions were delivered weekly and bi-weekly. In Parker 2009, the two sessions took place approximately one to two weeks before surgery. In the remaining three studies, telephone calls were delivered over a period of seven to 10 days (Mishel 2009) and in a one-off face-to-face session (Davison 1997; Thornton 2004). The intervention in Chambers 2013 was delivered over five telephone sessions—two before and three after treatment.

The duration of interventions ranged from a one-off session (Davison 1997; Thornton 2004) to 20 weeks (Carmack Taylor 2006) with a mean of 6.7 weeks. In the 14 studies that provided data, the duration of contact (amount) ranged from 13 minutes (Bailey 2004) to 135 minutes (Berglund 2007); the mean duration of contact was 74.8 minutes.

The background of those delivering the interventions varied. In eight studies (Bailey 2004; Berglund 2007; Chambers 2013; Davison 1997; Giesler 2005; Mishel 2002; Mishel 2009; Northouse 2007), nurses delivered or were involved in delivering the intervention. In seven studies (Campbell 2007; Lepore 2003; Parker 2009; Penedo 2006; Penedo 2007; Thornton 2004; Zhang 2006), a psychologist (trained in health or clinical psychology) administered the intervention. A physiotherapist delivered the exercise intervention in Berglund 2007. In Carmack Taylor 2006, the facilitator who delivered the intervention was supervised by a clinical psychologist, and in Manne 2011, the intervention was delivered by ‘trained therapists’. Trained former patients, who had undergone prostatectomy, were paired with participants in Weber 2007. In Ames 2011, a multidisciplinary team delivered the intervention.

In 12 studies (Ames 2011; Bailey 2004; Berglund 2007; Carmack Taylor 2006; Chambers 2013; Davison 1997; Lepore 2003; Parker 2009; Penedo 2006; Penedo 2007; Weber 2007; Zhang 2006), the participants were men only, and in the other seven, participants' partners or primary support persons were also involved. In Lepore 2003, female family members convened in a separate room for their own discussion while the men received the intervention.

Excluded studies

Thirty studies were excluded after full texts were assessed by two review authors. The main reasons for exclusion included the following: 14 had a mixed cancer sample but no separate data for prostate cancer were available; 11 did not provide psychosocial interventions; two did not measure outcomes relevant to this review; one had a very small sample and reported combined data for the intervention and control groups; two had two interventions each and no usual care group. See Characteristics of excluded studies.

Risk of bias in included studies

Details of risk of bias assessment in the 19 studies are given in Characteristics of included studies and are summarised in Figure 2 and Figure 3.

Figure 2.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Figure 3.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Allocation

Seven studies (Ames 2011; Berglund 2007; Davison 1997; Lepore 2003; Mishel 2002; Mishel 2009; Thornton 2004) did not provide a description of the actual randomisation method, although some information, for example, on stratification or ‘blocking’ was given. Random sequence generation information was available for the other 12 studies from published papers or from personal correspondence with the study authors. Methods used included ‘coin toss’, a table of random numbers and minimisation.

In three studies (Campbell 2007; Lepore 2003; Weber 2007), opaque envelopes were used as the method to conceal allocation. In Manne 2011, randomisation was concealed from staff, but information about the actual method of concealment was not available from the study author (personal correspondence). Concealment was not considered an issue in three studies (Giesler 2005; Northouse 2007; Zhang 2006) because "the group that participants had been assigned to was obvious to all" (Giesler 2005, personal communication) or “participants were told of group allocations as per IRB [Institutional Review Board] requirements” (Northouse 2007, personal communication). In the other 11 studies, allocation concealment was not mentioned or described.

Blinding

Psychosocial interventions are normally tested against usual care (control) or other psychosocial interventions. As no placebo is involved, it is not possible to conceal group assignment from participants or from personnel delivering the intervention. Sometimes the same personnel can be involved in more than one intervention in the same study (e.g. Lepore 2003; Parker 2009). Although lack of blinding of participants and personnel would increase the likelihood of bias in these studies, researchers in these studies were constrained by the nature of the psychosocial interventions, as it is impossible to blind recipients and interventionists to treatment.

In three studies (Berglund 2007; Northouse 2007; Thornton 2004), participants were told of group assignment. In three other studies, they were not informed of group assignment (Lepore 2003; Penedo 2006; Penedo 2007). In the remaining studies, blinding of participants was not mentioned. Blinding of personnel was not mentioned or described in all included studies, except in Lepore 2003. In Manne 2011, allocation to groups was “concealed from staff before assignment but not afterwards” (personal communication).

Blinding of assessors (data collectors) was mentioned in 11 studies (Carmack Taylor 2006; Chambers 2013; Giesler 2005; Lepore 2003; Manne 2011; Northouse 2007; Parker 2009; Penedo 2006; Penedo 2007; Weber 2007; Zhang 2006). One study (Bailey 2004) stated that “data collectors did not deliver the intervention but were aware of group assignment at the time of enrollment” (p. 342). In two studies (Mishel 2002; Mishel 2009), it was not clear whether data collectors were blinded to group assignment. In the remaining five studies, this issue was not mentioned.

Incomplete outcome data

Only two studies (Manne 2011; Northouse 2007) carried out intention-to-treat analysis in relation to missing data. In six studies (Ames 2011; Giesler 2005; Lepore 2003; Mishel 2002; Mishel 2009; Thornton 2004), missing data were related to the number of participants in each group at each of the follow-up time points. Most of the other included studies explained the reasons for attrition. Only eight of 19 studies (Berglund 2007; Carmack Taylor 2006; Chambers 2013; Manne 2011; Northouse 2007; Parker 2009; Penedo 2006; Penedo 2007) provided a CONSORT (Consolidated Standards of Reporting Trials) or CONSORT-type flow diagram.

Selective reporting

Eight studies (Berglund 2007; Davison 1997; Giesler 2005; Lepore 2003; Manne 2011; Thornton 2004; Weber 2007; Zhang 2006) did not report data on one or more prespecified outcomes or at one or more prespecified time points. In the other 11 studies, prespecified outcome data were reported.

Other potential sources of bias

Other potential sources of bias could be the smaller number of participants than was required to detect effect size in the included studies: 41 in Bailey 2004; 40 in Campbell 2007; and 29 in Zhang 2006. In nine studies, the attrition rate was greater than 15%, and in five (Berglund 2007; Campbell 2007; Parker 2009; Penedo 2007; Thornton 2004), the rate was greater than 20% (range 23.7% to 36.5%).

Effects of interventions

See: Summary of findings for the main comparison Psychosocial interventions compared with usual care for men with prostate cancer

The follow-up time points in the included studies can be classified into three periods: end of intervention (up to three months post-intervention), medium term (four to six months post-intervention) and long term (eight to 12 months post-intervention). The timing of 'end of intervention' varied from two weeks to six months, depending on the duration of the intervention. Therefore, six months post-baseline was in fact 'end of intervention' in a study in which the intervention lasted six months. Only one study (Chambers 2013) collected data at two years post-treatment.

Quality of life

Quality of life was measured in 16 of 19 included studies.

General health-related quality of life

Nine studies (Ames 2011; Carmack Taylor 2006; Chambers 2013; Giesler 2005; Lepore 2003; Northouse 2007; Parker 2009; Thornton 2004; Weber 2007) used the extensively validated Short Form-36 (SF-36) to measure general health–related quality of life. Two studies (Campbell 2007; Zhang 2006) used the Medical Outcomes Study Physical Functioning Measure (MOS-PF). However, extractable data for meta-analysis were not available in five studies (Ames 2011; Giesler 2005; Thornton 2004; Weber 2007; Zhang 2006). Quality of life in the SF-36 is normally presented in the form of two composite scores: physical and mental. Weber 2007 provided scores for each domain but not summary scores for the physical and mental health components.

Analysis of data from six studies with extractable data (Campbell 2007; Carmack Taylor 2006; Chambers 2013; Lepore 2003; Northouse 2007; Parker 2009) showed a small, statistically significant improvement in the physical component of general health–related quality of life at the end of intervention in favour of the psychosocial intervention when compared with usual care (SMD 0.12, 95% CI 0.01 to 0.22, P = 0.03) (Analysis 1.1; Figure 4). The total population involved in these studies numbered 1414. The quality of evidence was rated as low (Summary of findings for the main comparison), and the 95% CI crosses the minimal important difference. Five studies reported data at four to six months in a total of 1357 participants (Carmack Taylor 2006; Chambers 2013; Lepore 2003; Northouse 2007; Parker 2009), and four studies reported data at eight to 12 months in a total of 1215 participants for the physical component of general health–related quality of life (Chambers 2013; Lepore 2003; Northouse 2007; Parker 2009). No significant differences in effects were found for the physical component of general health–related quality of life at four to six months (SMD 0.02, 95% CI -0.09 to 0.13) (Analysis 1.2) and at 8 to 12 months (SMD 0.14, 95% CI -0.03 to 0.32) (Analysis 1.3). The only significant difference identified from subgroup analysis of group-based and individual-based interventions for the physical component of general health–related quality of life was a small, statistically significant improvement in favour of psychosocial interventions when compared with usual care in group-based interventions at end of intervention (SMD 0.24, 95% CI 0.02 to 0.47, P = 0.03) (Analysis 1.1; Figure 4).

Figure 4.

Forest plot of comparison: 1 Psychosocial intervention versus usual care, outcome: 1.1 General health–related quality of life: physical component at end of intervention.

Data from 1416 participants in six studies (Campbell 2007; Carmack Taylor 2006; Chambers 2013; Lepore 2003; Northouse 2007; Parker 2009) showed no significant difference in effects on the mental component of general health–related quality of life at end of intervention when psychosocial intervention was compared against usual care (SMD -0.04, 95% CI -0.15 to 0.06) (Analysis 1.4). The quality of evidence was rated as moderate (Summary of findings for the main comparison). Five studies reported data at four to six months in a total of 1357 participants (Carmack Taylor 2006; Chambers 2013; Lepore 2003; Northouse 2007; Parker 2009), and four studies reported data at eight to 12 months in a total of 1215 participants for the mental component of general health–related quality of life (Chambers 2013; Lepore 2003; Northouse 2007; Parker 2009). No significant differences in the mental component of general health–related quality of life were found at four to six months (SMD -0.08, 95% CI -0.19 to 0.03) (Analysis 1.5), and at eight to 12 months, post-intervention (SMD -0.04, 95% CI -0.15 to 0.08) (Analysis 1.6). Only one study (Chambers 2013) measured the effects of a psychoeducational intervention on health-related quality of life at 24 months post-treatment. It reported no change in the physical and mental components at this time point. Subgroup analysis also showed no difference in the mental component of general health–related quality of life between psychosocial interventions and usual care for group-based and individual-based interventions at these time points.

Of the six studies for which extractable data for meta-analysis (in relation to quality of life) were available, two described participants as being at stages I to III of prostate cancer (Northouse 2007; Parker 2009). In Carmack Taylor 2006, participants were recruited, regardless of stage, and no other detail was provided. Participants in Lepore 2003 were at T1, T2 and T3 stages; no patient at the metastatic stage was recruited. Participants in Chambers 2013 were newly diagnosed with localised prostate cancer. In the remaining study (Campbell 2007), participants were described as ‘beyond acute treatment phase’. It seems that most participants in these studies were in the localised or locally advanced stage of prostate cancer. However, no adequate extractable data were available for subgroup comparison.

Cancer-related quality of life

Cancer-related quality of life was measured in three studies (Northouse 2007; Penedo 2006; Penedo 2007) using the Functional Assessment of Cancer Therapy-General (FACT-G). A small, statistically significant improvement in cancer-related quality of life was seen at end of intervention in favour of the psychosocial intervention when compared with usual care (497 participants, SMD 0.21, 95% CI 0.04 to 0.39, P = 0.02) (Analysis 1.7). Subgroup analysis showed no difference in effect at end of intervention between the subgroups of group-based and individual-based interventions (262 participants, SMD 0.30, 95% CI -0.04 to 0.64; and 235 participants, SMD 0.16, 95% CI -0.09 to 0.42, respectively). Only Northouse 2007, which included an individual-based intervention, reported data at eight months and at 12 months post-intervention; no significant differences in cancer-related quality of life were noted between groups at these two time points (228 participants, SMD 0.01, 95% CI -0.25 to 0.27; and 218 participants, SMD 0.03, 95% CI -0.24 to 0.29, respectively) (Analysis 1.8). In the three studies that provided data on stage of cancer, participants were described as being in prostate cancer stages I to III (Northouse 2007) and in stages I and II (Penedo 2006; Penedo 2007). No separate data were provided for each disease stage to allow subgroup analysis to be carried out.

Prostate cancer–specific quality of life

Three studies (Ames 2011; Northouse 2007; Thornton 2004) used the Functional Assessment of Cancer Therapy–Prostate (FACT-P) to measure prostate cancer–specific quality of life. No data could be extracted from Ames 2011 and Thornton 2004 for a meta-analysis. Analysis of data from Northouse 2007, with a total of 219 participants, showed no significant difference in prostate cancer–specific quality of life between psychosocial intervention and usual care groups at end of intervention (SMD 0.13, 95% CI -0.14 to 0.39) and at eight months (SMD -0.07, 95% CI -0.34 to 0.19) and at 12 months (SMD 0.14, 95% CI -0.13 to 0.40) post-intervention (Analysis 1.9). In Northouse 2007, the intervention was individual-based. No data on stage of disease could be extracted for subgroup analysis.

Symptom-related quality of life

The Expanded Prostate Cancer Index Composite (EPIC) was used in three studies (Campbell 2007; Chambers 2013; Northouse 2007) to measure symptom-related quality of life. It consists of four domains: urinary, bowel, sexual and hormonal. Composite scores are given for each of these domains, but no overall score is given for the scale. Data were available at end of intervention from three studies for the urinary, bowel and sexual domains (Campbell 2007; Chambers 2013; Northouse 2007); and from two studies for the hormonal domain (Campbell 2007; Northouse 2007). Analysis 1.10 shows no significant differences in quality of life between psychosocial intervention and usual care groups at end of intervention in the urinary domain (946 participants, SMD 0.18, 95% CI -0.16 to 0.52), the bowel domain (946 participants, SMD -0.02, 95% CI -0.15 to 0.10), the sexual domain (946 participants, SMD 0.03, 95% CI -0.10 to 0.16) and the hormonal domain (265 participants, SMD 0.17, 95% CI -0.36 to 0.70). At six to eight months post-intervention, two studies reported data from a total of 887 participants for the urinary, bowel and sexual domains (Chambers 2013; Northouse 2007). No significant differences were identified between psychosocial intervention and usual care at six to eight months for these domains (Analysis 1.11). Two studies reported data at 12 months post-intervention from a total of 864 participants for the urinary, bowel and sexual domains (Chambers 2013; Northouse 2007). At 12 months post-intervention (Analysis 1.12), no significant differences in effects on quality of life were evident in these domains. All three studies contributing to these analyses were individual-based. Only Northouse 2007 and Chambers 2013 provided data on quality of life and disease stage, but they were not comparable. No further analysis was possible.

The UCLA Prostate Cancer Index (PCI) was used in three studies (Lepore 2003; Parker 2009; Weber 2007), which included a total of five psychosocial intervention groups to measure six symptom-related quality of life domains: urinary function, urinary bother, sexual function, sexual bother, bowel function and bowel bother. No composite score is given for the scale, only for each domain. All three studies reported data at end of intervention for the six symptom-related quality of life domains (Lepore 2003; Parker 2009; Weber 2007). No significant differences were found between psychosocial intervention and usual care for the following domains at end of intervention: urinary function (428 participants, SMD 0.14, 95% CI -0.06 to 0.34), urinary bother (428 participants, SMD 0.01, 95% CI -0.19 to 0.21), sexual function (428 participants, SMD 0.00, 95% CI -0.20 to 0.20), sexual bother (428 participants, SMD 0.17, 95% CI -0.03 to 0.37), bowel function (430 participants, SMD -0.07, 95% CI -0.27 to 0.13) and bowel bother (428 participants, SMD 0.02, 95% CI -0.18 to 0.22) (Analysis 1.13). Two studies with a total of four psychosocial intervention groups reported data from 357 participants at six months and from 351 participants at 12 months for each of the six symptom-related quality of life domains (Lepore 2003; Parker 2009). No significant differences in these six domains of symptom-related quality of life were found between groups at six months and at 12 months (Analysis 1.14; Analysis 1.15). As no summative score was assigned for the UCLA PCI tool, it was not possible to undertake subgroup analysis related to group-based versus individual-based interventions.

The Prostate Cancer Quality of Life Instrument (PCQoL) was used to measure symptom-related quality of life in Giesler 2005. However, no data could be extracted for analysis. No comparable data on disease stage were available for further analysis.

Quality of life in the remaining studies

Five studies measured and reported quality of life but had insufficient extractable data for meta-analysis. Ames 2011 used the SF-36 and reported that general health–related quality of life, in the psychosocial intervention group, decreased at end of intervention, compared with the control group. At six months, the psychosocial intervention group had a higher quality of life than the control group. Prostate cancer‒specific quality of life (using the FACT-P) was higher for the psychosocial intervention group than for the control group at end of intervention, but it decreased at six months post-intervention. Giesler 2005 used the Prostate Cancer Quality of Life Instrument to measure the urinary, sexual, bowel and cancer worry domains of quality of life. Although no extractable data were available for comparison, it was reported that “the intervention had significant beneficial effects on several of the disease-specific outcomes, with the most consistent effects accruing to the sexual outcomes” (p. 757). The psychosocial intervention was focused on problems related to sexual, urinary and bowel dysfunction, cancer worry, depression, fatigue and pain. Berglund 2007 measured quality of life by using the validated European Organization for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC QLQ–C30). Although no data were available for meta-analysis, it was reported that quality of life did not change following the three psychosocial interventions in the study. In Bailey 2004, participants were asked to rate on a scale of 0 to 10 (higher score indicating higher quality of life) their quality of life at the time of study and the quality of life they expected in the six months following the intervention. It was reported that at baseline, quality of life scores ranged from three to 10, but the post-intervention range was four to 10 (Bailey 2004). Participants' expectations of quality of life at six months were reported to have increased significantly more in the psychosocial intervention group than in the control group. Similarly, Mishel 2009 asked participants to rate their overall quality of life by asking them how positive they felt about their quality of life. No significant differences in results were reported post-intervention. Lack of validated measures makes these findings questionable.

Self-efficacy

Self-efficacy, in this review, refers to the belief and confidence in one's ability to manage problems related to chronic illness, including cancer (Weber 2007). Self-efficacy was reported in three studies (Campbell 2007; Northouse 2007; Weber 2007). It appears that self-efficacy was measured in Lepore 2003 (this is mentioned in an accompanying paper by Helgeson 2006), but no data were available. The validated instruments used to measure the effects of interventions in the other three studies were the self-efficacy Symptom Control Inventory (Campbell 2007), the Lewis Cancer Self-efficacy Scale (Northouse 2007), and the Stanford Inventory of Cancer Patient Adjustment (Weber 2007). The total number of participants available for analysis was 337. The analysis of these three studies showed no significant difference in self-efficacy when psychosocial intervention was compared with usual care at end of intervention (SMD 0.16, 95% CI -0.05 to 0.38) (Analysis 1.16; Figure 5). The quality of evidence was rated as very low (Summary of findings for the main comparison). The high heterogeneity (I² = 65%) across these studies raises questions about bias in each of the individual studies. In Campbell 2007, the sample was small (n = 40) and attrition was high overall (25%) and was higher in the psychosocial intervention group (40%). In Weber 2007, standard deviations were high in the usual care and psychosocial intervention groups (43.8 and 40.6, respectively) in a sample of 72 participants. In Northouse 2007, in which the sample was much larger (n = 235) and the SDs were 19 (psychosocial intervention group) and 20.2 (usual care group), no difference in self-efficacy scores was seen between psychosocial intervention and usual care groups at end of intervention (SMD 0.01, 95% CI -0.25 to 0.26) (Analysis 1.16; Figure 5). A fixed-effect model was selected for this analysis, as a random-effects model distorted the weighting considerably in favour of the smaller studies (Higgins 2002).

Figure 5.

Forest plot of comparison: 1 Psychosocial intervention versus usual care, outcome: 1.16 Self-efficacy at end of intervention.

Only Northouse 2007 had extractable data at eight and 12 months. No significant differences in self-efficacy were evident between groups at eight months (228 participants, SMD -0.05, 95% CI -0.31 to 0.21) and at 12 months (218 participants, SMD -0.07, 95% CI -0.34 to 0.20) (Analysis 1.17).

All interventions included in the analyses for self-efficacy were individual-based. Subgroup analysis of disease stage and self-efficacy was considered. However, only Northouse 2007 provided details of the disease stage of participants but no separate data on the scores according to disease stage.

Knowledge

Three studies (Lepore 2003; Mishel 2002; Mishel 2009) assessed the effects of psychosocial interventions on level of knowledge among men with prostate cancer. Mishel 2002 did not provide information on the numbers of participants in the psychosocial intervention and control groups at baseline or at the follow-up time points and therefore was excluded from the analysis. These investigators found "no significant differences between the two treatment groups and the control group [...] for cancer knowledge" (Mishel 2002, p. 1861).

In Lepore 2003, two psychosocial interventions (‘education only’ and 'education plus discussion’) and one control group were included. Data on knowledge were collected at baseline and at two weeks post-baseline. In Mishel 2009, two psychosocial interventions were provided: ‘treatment direct’ (i.e. intervention administered directly to men with prostate cancer) and ‘treatment supplemented’ (i.e. intervention administered to men with prostate cancer and their partners). Data were collected at four weeks and at three months post-baseline. Both Lepore 2003 and Mishel 2009 gave the number of participants in each group at baseline, but no such information was provided at follow-up time points. Therefore baseline data on group allocation were used for analysis of the other time points.

Lepore 2003 developed their own Prostate Cancer Knowledge ‘13-item quiz’. No information on the reliability and validity of the instrument was reported, except that “a urologist verified the accuracy of the items” (p. 445). Mishel 2009 used the Prostate Cancer 20–Item Knowledge Scale, which they developed and used in a previous study. The Kuder–Richardson Formula 20 (KR-20) reliability of the tool was reported to range from 0.67 to 0.73. No information on the validity of the tool was given. Figure 6 shows a moderate and statistically significant increase in prostate cancer knowledge in the psychosocial intervention groups when compared with usual care at end of intervention (SMD 0.51, 95% CI 0.32 to 0.71, P < 0.00001) (Analysis 1.18). The total number of participants in these two studies was 506. This evidence should be put in the context of the lack of tried and tested tools for measuring knowledge, the unavailability of data on the number of participants in each group at each of the follow-up time points and the subgroup differences (I² = 65.2%) between these studies. The quality of evidence was rated as very low (Summary of findings for the main comparison). Statistically significant improvements in prostate cancer knowledge were noted in the psychosocial intervention group when compared with usual care at end of intervention in the subgroups for both group-based (250 participants, SMD 0.68, 95% CI 0.41 to 0.95, P < 0.00001) and individual-based interventions (256 participants, SMD 0.35, 95% CI 0.08 to 0.62, P = 0.01) (Analysis 1.18).

Figure 6.

Forest plot of comparison: 1 Psychosocial intervention versus usual care, outcome: 1.18 Prostate cancer knowledge at end of intervention.

Only Mishel 2009 measured the effects of two interventions (both individual-based) on this outcome at three months post-baseline. A small, statistically significant increase in prostate cancer knowledge was noted at this time point (256 participants, SMD 0.31, 95% CI 0.04 to 0.58, P = 0.02) (Analysis 1.19). Separate scores for disease stage were not provided.

Uncertainty

Uncertainty, in this review, is defined as the inability to determine the meaning of illness-related events (Mishel 2002). Only three studies (Chambers 2013; Mishel 2002; Northouse 2007) measured the effects of psychosocial intervention on uncertainty. All three studies were individual-based. No extractable data from Mishel 2002 were available for comparison, but the authors reported "a significant decline over time in uncertainty, but this change did not differ by treatment group" (p. 1861). In Chambers 2013, decisional uncertainty was measured by a subscale of the Decisional Conflict Scale-Revised. Northouse 2007 used the 28-Item Mishel Uncertainty in Illness Scale (a validated tool) to measure uncertainty at end of intervention and at eight months and 12 months post-baseline. No significant differences in effects were observed between psychosocial intervention and control groups at end of intervention (SMD -0.05, 95% CI -0.35 to 0.26) (Analysis 1.20). The quality of evidence was rated as very low (Summary of findings for the main comparison). No significant differences in uncertainty were observed between groups at six to eight months (SMD 0.04. 95% CI -0.09 to 0.17) (Analysis 1.21) and at 12 months (SMD 0.07, 95% CI -0.06 to 0.21) (Analysis 1.22). These two studies had a total population of 916. No data were available for disease stage subgroup analysis.

Distress

Five studies (Chambers 2013; Manne 2011; Mishel 2002; Northouse 2007; Zhang 2006) measured the effects of psychosocial interventions on distress among men with prostate cancer. In Chambers 2013, cancer-specific psychological distress was measured by the Revised Impact of Events Scale. Manne 2011 used the Impact of Events Scale to measure cancer-specific distress and the Psychological Distress Scale of the Mental Health Inventory and reported “no significant treatment main effects for general distress, cancer-specific distress" (p. 1202). Zhang 2006 reported using the McCorkle and Young Distress Scale to measure symptom distress in a small sample of 29 men with prostate cancer. Data on the effects of the psychosocial intervention on distress were not reported in the study and were not available from the main author (personal communication). Mishel 2002 used the 15-Item Symptom Distress Scale to “index the degree of problems the patient had with treatment side effects” (p. 1860) but provided no overall score for symptom distress. Northouse 2007 used the 16-Item Symptom Scale of the Omega Screening Questionnaire to measure General Symptom Distress among men with prostate cancer at baseline and at four, eight and 12 months post-baseline. They reported that a supportive-educative family intervention had no significant effects on men’s distress at the three post-intervention time points. No data could be extracted from three studies (Manne 2011; Mishel 2002; Zhang 2006). When the findings from Chambers 2013 and Northouse 2007 were combined, they showed no significant difference in distress between psychosocial intervention and usual care groups at end of intervention (SMD 0.02, 95% CI -0.11 to 0.15) (Analysis 1.23). The quality of evidence was rated as very low (Summary of findings for the main comparison). Also no difference in distress between psychosocial intervention and control groups was noted at six to eight months (SMD 0.06, 95% CI -0.07 to 0.19) and at 12 months thereafter (SMD 0.02, 95% CI -0.11 to 0.15) (Analysis 1.23). These two studies (Chambers 2013; Northouse 2007) were individual-based with a total population of 916. No data could be extracted for disease stage subgroup analysis.

Depression

Effects on depression were measured in six studies (Berglund 2007; Carmack Taylor 2006; Davison 1997; Giesler 2005; Lepore 2003; Weber 2007). No extractable data were available for Davison 1997 for inclusion in a meta-analysis. However, it was reported that no significant differences in mean scores of depression were seen between psychosocial intervention and control groups at end of intervention. Similarly, Giesler 2005 did not provide relevant data for comparison of the effects of their intervention on depression, and this study was excluded from the meta-analysis. Whether the interventions had any effect on depression was not mentioned, although data on the moderating effects of depression on other variables were presented.

In Berglund 2007, depression was measured with the Hospital Anxiety and Depression Scale at six and 12 months post-intervention. This study included three psychosocial intervention groups and one control group. Data at six months post-intervention were not reported, and extractable data for meta-analysis were not available. However, Berglund 2007 concluded that “it could not be assumed that, the interventions had any effect on [...] depressive symptoms” (p. 88). The small sample size in each group and 20% attrition “resulted in decreasing power at each assessment period" (Berglund 2007, p. 88).

In Weber 2007, the effects of a dyadic social support intervention on depression were measured by the Geriatric Depression Scale (a validated instrument). Data were collected at baseline and at four and eight weeks post-intervention, but the scores at four months were not presented. Weber 2007 reported that, compared with the control group, the dyadic social support group had significantly less depression (mean = 2.53 and 0.92, respectively) at eight weeks post-intervention. However, some of the coefficients and standard deviations in this study were extreme (Chambers 2011).

Depression at end of intervention

Three studies (Carmack Taylor 2006; Lepore 2003; Weber 2007) used validated instruments to measure depression at end of intervention in a total of 434 participants. Both Carmack Taylor 2006 and Lepore 2003 used the Center for Epidemiological Studies–Depression Scale; each study had two psychosocial intervention arms and one control group. Weber 2007 used the Geriatric Depression Scale. Analysis 1.24 shows no significant difference in depression between groups at end of intervention (SMD -0.18, 95% CI -0.51 to 0.15). The total population in these studies was 434. The quality of evidence was rated as very low (Summary of findings for the main comparison). Analysis 1.24 showed a statistically significant reduction in depression in favour of the psychosocial intervention group for individual-based interventions (72 participants, SMD -0.59, 95% CI -1.06 to -0.11, P = 0.02) but not group-based interventions (362 participants, SMD -0.06, 95% CI -0.38 to 0.25). However, this result should be treated with caution because some of the coefficients and standard errors in the single individual-based intervention study in this analysis (Weber 2007) were extreme (Chambers 2011).

Depression at six months post-intervention

Carmack Taylor 2006 measured the effects of two psychosocial interventions at 12 months post-baseline. In effect, this was six months post-intervention, as the intervention lasted six months. Data from two studies (Carmack Taylor 2006; Lepore 2003) with a total of 364 participants showed no significant difference in depression between psychosocial intervention and usual care control groups at six months post-intervention (SMD -0.17, 95% CI -0.44 to 0.09) (Analysis 1.25). Both studies assessed group-based interventions only.

Depression at 12 months post-intervention

Only one study (Lepore 2003) provided extractable data from a total of 250 participants for comparison at 12 months post-intervention. Analysis 1.26 shows no difference in effect on depression between psychosocial intervention and usual care groups at this time point (SMD -0.02, 95% CI -0.29 to 0.24). The intervention in Lepore 2003 was group-based.

Finally, of the three studies with extractable data for depression, only two provided details on disease stage: T1, T2 and T3 (Lepore 2003); all stages (Carmack Taylor 2006). In Giesler 2005, participants were at T1a and T2c stages, and in Berglund 2007, 20% had metastasis. Weber 2007 did not provide data on disease stage but reported that "stage of disease was controlled by study design" (p. 157). Subgroup analysis in relation to stage of disease was not possible because of the lack of separate scores for each disease stage.

Discussion

This review shows that psychosocial interventions for men with prostate cancer were beneficial for some outcomes at certain times but had no significant effect on most of the outcomes in this study. Evidence of a small effect of psychosocial intervention on the physical component of general health–related quality of life and on cancer-related quality of life at end of intervention should be treated with caution, as the quality of the evidence was rated low and very low for each of these outcomes, respectively, according to GRADE. In the case of the physical component of general health–related quality of life, the 95% CI crosses the minimal important difference, thereby reducing the clinical significance of these findings.

The small positive effects of the psychosocial interventions on these two aspects of quality of life may be accounted for by the fact that in some of the included studies, participants had similar (Northouse 2007) or higher (Thornton 2004) levels of quality of life at enrolment as compared with the general population. Therefore to expect significant changes in quality of life may have been too optimistic. The same may apply to the other outcomes on which psychosocial interventions had no effects. This raises the question as to whether these psychosocial interventions should target only those with low levels of quality of life, low self-esteem and high levels of uncertainty, distress or depression. The margin for improvement is narrow when participants already experience quality of life similar to the general population (Northouse 2007; Thornton 2004) or low distress or pain (Carmack Taylor 2006). Authors of many studies included in this review have recommended that those in distress who have low quality of life and are experiencing urinary, sexual and other problems should be targeted (Carmack Taylor 2006; Giesler 2005; Lepore 2003; Northouse 2007; Weber 2007). Questions may arise regarding the appropriateness, content and delivery of interventions, as well as the efficacy of tools to measure these outcomes. Researchers and clinicians should aim to clarify, at the start of their study, the minimal clinically important differences that would be acceptable to both clinicians and participants.

Evidence of a positive effect of psychosocial interventions on prostate cancer knowledge at end of intervention and at three months thereafter should be treated with caution. Data related to knowledge came from two studies in which researchers developed their own measures of knowledge. Additional validated and standardised instruments are needed to provide robust evidence related to this outcome. Data on attrition at the different time points in these studies are also lacking.

One possible reason why participants in these studies may have had levels of quality of life or distress similar to those of the general population is that those who participated were self-selected. Evidence showed high numbers of participants who refused to participate and high attrition rates in some studies (Berglund 2007; Campbell 2007; Parker 2009; Penedo 2007). These are indications that this population is difficult to recruit. Those who agreed to take part may have been those who want to help and who probably need these interventions less than those who refused to take part in, or dropped out of, these studies. These issues need further investigation.

Heterogeneity in the included studies

Evidence of considerable heterogeneity was found in these studies, in particular in the characteristics of participants, the types of interventions provided and the tools used to measure outcomes relevant to this review. These studies differed in many aspects in terms of disease stage, treatment stage, age and ethnicity. Some participants were in watchful waiting, some were about to receive treatment and others had completed treatment. Campbell 2007 included only African Americans and Zhang 2006 recruited only those who were experiencing urinary incontinence during the week before the start of the study. Interventions also differed with regard to their content, format, frequency, duration and quantity and in how and by whom they were delivered.

For each of the outcomes in this review, researchers could choose from a variety of tools or could develop their own. For example, 10 scales were used to measure several types of distress such as symptom distress, cancer-specific distress, psychological distress and management of treatment side effects. Sometimes the same scale was used to measure different, but related, outcomes. For example, the Perceived Stress Scale was used to measure ‘psychological well-being' and ‘stress’. Although it is unlikely that there could be a consensus for using one particular scale to measure one particular outcome, better justification of the choice of scales based on evidence would reduce the choice to those that are well validated and widely used. A structured review (Morris 2009) of participant-reported outcome measures for prostate cancer recommended two generic instruments (SF-36 and the EuroQol EQ-5D) for measuring general quality of life. Of two general cancer and nine prostate cancer–specific quality of life scales identified in the review, three were reported by Morris 2009 as having supportive evidence for use in relation to prostate cancer. These were EORTC QLQ–C30 and PR25, FACT-P (including the four domains from the FACT-G) and UCLA-PCI and EPIC. Apart from the EQ-5D, all other tools recommended in the review by Morris 2009 were used to measure quality of life in the included studies in this review. In an integrative review of literature on the effectiveness of psychosocial interventions for cancer patients, Raingruber 2011 found that each study included clients with several types and stages of cancer and that each study used a different type of supportive care and relied on a spectrum of different outcome measures.

Sample size and attrition

Four of the 19 included studies were described as pilot studies. In several of the other studies with larger samples, multiple interventions in the same study combined with high attrition meant that power to carry out the intended analysis was ‘limited’ or ‘decreasing’ (Berglund 2007; Carmack Taylor 2006; Giesler 2005). ‘Simple designs’ (intervention and control) and ‘high power’ were the recommendations in Berglund 2007, which included three intervention groups and one control group. In the future, researchers should have a strong rationale for including more than one intervention, in particular in studies with several follow-up time points, recruitment difficulties and potential attrition issues. Feasibility studies may be more appropriate for cases in which multiple interventions are provided in the same study.

Data on the number of eligible participants who refused to take part were mostly missing in these studies. When this information was provided, it was clear that the proportion of eligible participants who refused to take part was quite high. The end of study attrition rate was greater than 15% in nine of 17 studies that provided this data, and in five of them, it was greater than 20%. This problem was compounded in studies in which the attrition rate was unequally distributed between groups. The smaller the samples, the more likely were differences at baseline between the control group and the intervention group.

Most participants in this review were white. This restricts generalising the findings to all men with prostate cancer. Evidence in some studies revealed efforts to recruit a greater number of ethnic minority men (Campbell 2007; Mishel 2002; Mishel 2009; Penedo 2006; Penedo 2007). Greater effort should be made to include men from a variety of backgrounds in future studies. Targeting those in distress and in need of support, as recommended by several authors of studies included in this review, can have implications for recruitment. The ‘pool’ of participants may be reduced. Focusing more on 'hard to recruit' populations, such as men from ethnic minorities and men from lower social classes, will present challenges that are likely to have time (it may take longer to recruit) and funding implications.

Quality of studies and quality of reporting

The lack of important details in some studies precluded an adequate assessment of their quality. Figure 2 shows that the overall risk of bias was unclear or high in the included studies. Although blinding of participants to group allocation in these types of studies is not possible, it is likely that some of the positive effects of the interventions were due to ‘interventionist effect’ (Giesler 2005). In only four studies was an attempt made to address the interventionist effect. In Penedo 2006 and Penedo 2007, control group participants met for a four-hour seminar. In Thornton 2004, a 45-minute session was held with those in the control group, and in Zhang 2006, control group participants carried out pelvic floor exercises but did not receive the social support that the psychosocial intervention participants received. Devising strategies for giving similar attention to psychosocial intervention and control group participants is a challenge that must be met by researchers.

With regard to missing data, two issues that need to be highlighted here are reporting of outcome data and provision of data on the number of participants in each group at baseline and at all follow-up time points. Such information was missing in a number of studies—a fact that impeded meta-analysis. At the very least, researchers should report the mean scores and standard deviations for each outcome at different time points before reporting data from subgroup analyses. In less than 58% of the excluded studies, a CONSORT or CONSORT-type flow diagram was provided. Some studies with no flow diagram were published recently.

This review highlights some of the issues that clinicians and researchers need to consider as they continue their search for the best ways to support these men. The implications for practice and research are discussed further in their relevant sections.

Summary of main results

Small but statistically significant effects on the physical component of general health–related quality of life favoured psychosocial interventions but lacked clinical significance when compared with usual care in the short term, and for cancer-related quality of life favoured psychosocial interventions when compared with usual care in the short term, but these effects were not seen on the mental component nor on symptom-related quality of life. No significant difference in effect was observed between groups with regard to self-efficacy, uncertainty, distress and depression at any of the time points assessed. However, a significant, moderate increase in knowledge was noted to occur as a result of the psychosocial interventions when compared with usual care.

Subgroup analysis by type of intervention showed that improvement in the physical component of general health–related quality of life at end of intervention with psychosocial interventions remained statistically significant in the subgroup of group-based interventions but not in the subgroup of individual-based interventions. Improvement in knowledge at end of intervention with psychosocial interventions was observed in the subgroups of group-based interventions and individual-based interventions; at three months post-intervention, data were available only for individual-based interventions, and a statistically significant improvement in knowledge was reported with psychosocial interventions. For depression at end of intervention, significant improvement was observed with psychosocial interventions when compared with usual care only in the subgroup of individual-based interventions. The remaining subgroup analyses by type of intervention showed no evidence of a difference in effect between groups.

No data were extractable for subgroup analysis in relation to disease stage or treatment with ADT. Similarly, it was not possible to perform subgroup analysis by ethnicity or age group.

However, all these results should be treated with caution mainly because of the small number of studies included for some outcomes, the small sample sizes (apart from Northouse 2007 and Chambers 2013), the high rate of attrition in some studies and the heterogeneity of measuring scales used. Differences in type, content, format, delivery, frequency, duration and quantity of the interventions in this review, differences in the stage of prostate cancer among these men and lack of detailed figures related to treatment modalities, in particular ADT, restrict the generalisability of these findings.

Overall completeness and applicability of evidence

An adequate number of studies measured quality of life and depression. However, selective reporting and lack of extractable data significantly reduced the number of studies that could be included in the meta-analysis of these two outcomes. Studies that measured self-efficacy, knowledge, uncertainty and distress were too few for robust conclusions to be drawn. These findings and the quality assessment of the included studies show the need for more rigorous studies before any recommendations can be made to clinicians.

Quality of the evidence

Summary of findings for the main comparison shows that the quality of evidence for the seven main outcomes at end of intervention (rated by the GRADE framework) was as follows: physical component of general health–related quality of life: low; mental component of general health–related quality of life: moderate; self-efficacy: very low; knowledge: very low; uncertainty: very low; distress: very low; and depression: very low. The overall risk of bias in the included studies was unclear or high. Evidence of bias was found in a number of studies (see Figure 2), in particular lack of 'blinding of participants and personnel'. Psychosocial interventions, by their very nature and process, cannot be concealed from participants and those who administer the interventions. Therefore, these types of interventions start with a disadvantage when the 'Risk of bias' tool is applied.

Heterogeneity in duration of interventions and in disease stage could have introduced bias in these studies. For example, one intervention lasted two weeks and another six months. Disease- or treatment-related changes in a two-month period may differ from those in a six-month period. Although the interventions in most cases were adequately described, overall very little information was provided on what control group participants received other than information handouts or ‘usual care’.

The number of participants in the 19 included studies (n = 3204) was more than adequate for robust conclusions to be drawn, but lack of extractable data and selective reporting did not allow this to happen. Basic information about group sizes and mean scores at different time points was missing. Some studies had small samples and high rates of attrition.

All included studies were published in English, although no language restriction was applied. All but three studies were set in the USA, thereby restricting generalisability of findings to other settings because of differences in cultural and health systems between countries.

Potential biases in the review process

It is possible that work in languages other than English and in journals not included in the electronic database searches could have been missed. No grey literature was identified, although authors were contacted when indications suggested that a relevant study was undertaken. The interpretation of the term ‘psychosocial intervention' is a subjective one, and we adopted a more, rather than less, inclusive approach. This decision may have ‘diluted’ the findings. All studies were reviewed by two review authors independently, and advice was received from a third review author when necessary. Data input for meta-analysis was carried out by the first review author (KP) and and was checked by another review author (CS). Only one study addressed adverse effects (Penedo 2006), which were not among the included outcomes in this review.

Agreements and disagreements with other studies or reviews

Four relevant reviews were identified (Chambers 2011; Fritzsche 2008; Galway 2012; Weber 2005). Weber 2005 reviewed nine studies of psychosocial interventions and reported that they did not all use common measures of psychosocial outcomes, but all were theory-driven and aimed at reducing the negative emotional responses following prostate cancer and its treatment. This was a descriptive review aimed at highlighting the needs of men with prostate cancer and the need for innovative strategies to address them. Fritzsche 2008 carried out a systematic review of 47 cross-sectional and longitudinal studies of psychological distress and 13 controlled studies of mostly psychoeducational interventions from 1980 to 2007. No meta-analysis was performed because of the broad heterogeneity of interventions and measuring instruments used. Fritzsche 2008 reported that “even low-threshold, brief interventions may lead to small to moderate effects in disease-specific problems” (p. 329). In a systematic review of 21 trials of interventions aimed at improving adjustment in men with prostate cancer, Chambers 2011 reported that “group-based cognitive-behavioural and psycho-education interventions appeared helpful in promoting better psychological adjustment and quality of life for men with prostate cancer". No meta-analysis was carried out. Trial quality was low and no improvement in quality was noted over the study time frame (Chambers 2011). In a recent Cochrane review of psychological interventions provided to improve quality of life and emotional well-being for patients recently diagnosed with cancer (all types), Galway 2012 reported a small increase in cancer-related quality of life but not in general health–related quality of life at six months. The interventions had a positive effect on mood but not on depression. Together, the findings of these three reviews (Chambers 2011; Fritzsche 2008; Galway 2012) reported small improvements in disease-related quality of life. This review shows that psychosocial interventions had small, statistically significant positive effects (at end of intervention) on general health-related and cancer-related but not on prostate cancer–related or symptom-related quality of life. Apart from knowledge, the psychosocial interventions in this review showed no significant effects on all other outcomes at the time points assessed. These previous reviews pointed out wide variations in interventions as well as the low quality of studies included in their reviews. These observations were also noted in this review.

Authors' conclusions

Implications for practice

The evidence from this review is not strong enough to permit meaningful conclusions about the effects of psychosocial interventions for men with prostate cancer. Although small improvements in some aspects of quality of life were reported, the evidence was rated as low or very low quality, according to GRADE. In the case of the physical component of general health–related quality of life, the improvements were too small to have clinical significance. Researchers need to provide more robust evidence of the types of interventions that are effective and the types that are not effective. A consensus of the minimal clinically important difference that would be acceptable to clinicians is needed for each of the outcomes.

Implications for research

Research on psychosocial interventions for men with prostate cancer continues to evolve. Of note is the increasing number of studies, in recent years, that include partners. Improved access to technology, increased numbers of survivors of prostate cancer and a substantial need for cost-effective use of resources are drawing health and social care professionals towards the Internet as a setting and medium by which these men and their partners can be supported. Studies evaluating these ‘new’ approaches to psychosocial interventions will, in time, have to be reviewed for assessment of the evidence that practitioners require to meet the needs of their clients effectively.

Researchers made more use of a general health–related quality of life scale (11 studies with SF-36 or MOS-PF) than a cancer-related (three studies with FACT-G) or prostate cancer–related (three studies with FACT-P) scale, although some studies used more than one of these. Both Lepore 2003 and Penedo 2006 concluded that future research should address disease-specific quality of life. This will require increased use of cancer-related and prostate cancer–related scales. The FACT-P can be used to screen participants with cancer-related symptoms who require help and support. The SF-36 is useful to compare prostate cancer patients’ quality of life with that of the general population or other disease groups. It may not be sensitive enough to measure symptom-related quality of life. The adverse effects of these interventions were reported in only one study (Penedo 2006). Future research should ensure that this outcome is investigated.

What has become more obvious is the need to focus more on those men and their partners who most require help and support. The ‘broad brush’ approach of offering the same interventions to well-adjusted, well-educated men who experience the same quality of life as the normal population and to those experiencing urinary, sexual or marital difficulties is unlikely to be effective in terms of desired outcome and cost. As survival rates improve, a large number of men will be living in the aftermath of the disease and its treatment. Many will not require intensive input from health professionals. The challenge is to develop and test interventions aimed at helping these men to help themselves.

A wide range of interventions was reported among the included studies. Treating psychosocial interventions as one form of treatment does not do justice to the efforts of professionals and researchers in developing and testing creative and sensitive programmes designed to help these men with complex and changing needs as they embark and travel on their cancer and post-cancer journey. It is unlikely that a 'one size fits all' intervention will meet the needs of such a varied population with changing disease and treatment pathways and support needs. According to Badger 2011, "future research needs to address the question, 'who benefits from which intervention at what point in their cancer journey'" (p. 842).

A minority of men who experience symptom-related difficulties that cause discomfort, distress or depression would need person-centred interventions that address the particular problems they face. A review of psychosocial interventions addressing sexual or relationship functioning in men with prostate cancer (Chisholm 2012) concluded that “there is a need for further research to target improving and measuring men and their partner’s sexual and relationship functioning in the context of prostate cancer” (p. 1246). Chisholm 2012 also called for more research to examine the effectiveness of tailoring interventions to the specific needs of men and to their stage of cancer.

Too few studies and insufficient data in this review prevented subgroup analyses related to stage of cancer. The proportion of men at the advanced stage of the disease and with metastasis was very low in the studies included in this review. Another group that was not featured to any meaningful extent consists of younger men with prostate cancer. Those at greatest risk of adverse outcomes following prostatectomy are often young men not accustomed to changes or alterations in physical functions that define their masculinity, such as erectile dysfunction (Weber 2007). Interventions to meet the support needs of these population subgroups should be offered and tested.

A vast majority of participants in the included studies were white and lived in the USA. Only one study (Berglund 2007) was conducted in Europe. The generalisability of findings from these studies is limited to white, North American populations. Cultural factors that can affect the efficacy of psychosocial interventions include patterns of relating to authority figures, time orientation, interpersonal relations, family interdependence and minority status (Penedo 2007). Life priorities can differ among different social class groups. All these factors can affect recruitment and attrition rates. This is an area that remains a challenge to health professionals and researchers, and more should be done to include men from different social and ethnic backgrounds in large enough numbers to obtain results of statistical and clinical significance.

Finally, the reporting of studies needs to improve. In this review, 11 studies measured quality of life with the SF-36 or the MOS-PF, yet data from only six studies were available for meta-analysis. As a minimum requirement, authors should also report mean scores and standard deviations for each outcome for each group at each follow-up time point. Researchers and journal editors should ensure that such basic information is provided when randomised controlled trials are reported.

Acknowledgements

We would like to thank Dr Chris Bleakley and Dr Marialena Trivella for their valuable support and advice. We are grateful to all the study authors who have responded to our queries and requests for additional data.

Data and analyses

Download statistical data

Comparison 1. Psychosocial intervention versus usual care
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 General health–related quality of life: physical component at end of intervention61414Std. Mean Difference (IV, Random, 95% CI)0.12 [0.01, 0.22]
1.1 Group-based intervention2362Std. Mean Difference (IV, Random, 95% CI)0.24 [0.02, 0.47]
1.2 Individual-based intervention41052Std. Mean Difference (IV, Random, 95% CI)0.11 [-0.06, 0.28]
2 General health–related quality of life: physical component at four to six months51357Std. Mean Difference (IV, Random, 95% CI)0.02 [-0.09, 0.13]
2.1 Group-based intervention2363Std. Mean Difference (IV, Random, 95% CI)0.17 [-0.05, 0.40]
2.2 Individual-based intervention3994Std. Mean Difference (IV, Random, 95% CI)-0.03 [-0.16, 0.10]
3 General health–related quality of life: physical component at eight to 12 months41215Std. Mean Difference (IV, Random, 95% CI)0.14 [-0.03, 0.32]
3.1 Group-based intervention1250Std. Mean Difference (IV, Random, 95% CI)0.15 [-0.12, 0.42]
3.2 Individual-based intervention3965Std. Mean Difference (IV, Random, 95% CI)0.19 [-0.07, 0.46]
4 General health–related quality of lIfe: mental component at end of intervention61416Std. Mean Difference (IV, Random, 95% CI)-0.04 [-0.15, 0.06]
4.1 Group-based intervention2362Std. Mean Difference (IV, Random, 95% CI)-0.02 [-0.24, 0.20]
4.2 Individual-based intervention41054Std. Mean Difference (IV, Random, 95% CI)-0.05 [-0.17, 0.07]
5 General health–related quality of life: mental component at four to six months51357Std. Mean Difference (IV, Random, 95% CI)-0.08 [-0.19, 0.03]
5.1 Group-based intervention2363Std. Mean Difference (IV, Random, 95% CI)-0.01 [-0.23, 0.21]
5.2 Individual-based intervention3994Std. Mean Difference (IV, Random, 95% CI)-0.11 [-0.23, 0.02]
6 General health–related quality of life: mental component at eight to 12 months41215Std. Mean Difference (IV, Random, 95% CI)-0.04 [-0.15, 0.08]
6.1 Group-based intervention1250Std. Mean Difference (IV, Random, 95% CI)0.01 [-0.25, 0.28]
6.2 Individual-based intervention3965Std. Mean Difference (IV, Random, 95% CI)-0.05 [-0.17, 0.08]
7 Cancer-related quality of life (FACT-G) at end of intervention3497Std. Mean Difference (IV, Random, 95% CI)0.21 [0.04, 0.39]
7.1 Group-based intervention2262Std. Mean Difference (IV, Random, 95% CI)0.30 [-0.04, 0.64]
7.2 Individual-based intervention1235Std. Mean Difference (IV, Random, 95% CI)0.16 [-0.09, 0.42]
8 Cancer-related quality of life (FACT-G) at eight and 12 months1 Std. Mean Difference (IV, Random, 95% CI)Subtotals only
8.1 Eight months post-intervention1228Std. Mean Difference (IV, Random, 95% CI)0.01 [-0.25, 0.27]
8.2 12 months post-intervention1218Std. Mean Difference (IV, Random, 95% CI)0.03 [-0.24, 0.29]
9 Prostate cancer–specific quality of life (FACT-P)1 Std. Mean Difference (IV, Random, 95% CI)Subtotals only
9.1 End of intervention1219Std. Mean Difference (IV, Random, 95% CI)0.13 [-0.14, 0.39]
9.2 Eight months post-intervention1219Std. Mean Difference (IV, Random, 95% CI)-0.07 [-0.34, 0.19]
9.3 12 months post-intervention1219Std. Mean Difference (IV, Random, 95% CI)0.14 [-0.13, 0.40]
10 Symptom-related quality of life (EPIC) at end of intervention3 Std. Mean Difference (IV, Random, 95% CI)Subtotals only
10.1 Urinary domain3946Std. Mean Difference (IV, Random, 95% CI)0.18 [-0.16, 0.52]
10.2 Bowel domain3946Std. Mean Difference (IV, Random, 95% CI)-0.02 [-0.15, 0.10]
10.3 Sexual domain3946Std. Mean Difference (IV, Random, 95% CI)0.03 [-0.10, 0.16]
10.4 Hormonal domain2265Std. Mean Difference (IV, Random, 95% CI)0.17 [-0.36, 0.70]
11 Symptom-related quality of life (EPIC) at six to eight months post-intervention2 Std. Mean Difference (IV, Random, 95% CI)Subtotals only
11.1 Urinary domain2887Std. Mean Difference (IV, Random, 95% CI)-0.02 [-0.19, 0.14]
11.2 Bowel domain2887Std. Mean Difference (IV, Random, 95% CI)-0.06 [-0.26, 0.15]
11.3 Sexual domain2887Std. Mean Difference (IV, Random, 95% CI)0.09 [-0.04, 0.22]
12 Symptom-related quality of life (EPIC) at 12 months post-intervention2 Std. Mean Difference (IV, Random, 95% CI)Subtotals only
12.1 Urinary domain2864Std. Mean Difference (IV, Random, 95% CI)-0.07 [-0.21, 0.06]
12.2 Bowel domain2864Std. Mean Difference (IV, Random, 95% CI)0.00 [-0.14, 0.14]
12.3 Sexual domain2864Std. Mean Difference (IV, Random, 95% CI)0.03 [-0.11, 0.16]
13 Symptom-related quality of life (UCLA-PCI) at end of intervention3 Std. Mean Difference (IV, Random, 95% CI)Subtotals only
13.1 Urinary function3428Std. Mean Difference (IV, Random, 95% CI)0.14 [-0.06, 0.34]
13.2 Urinary bother3428Std. Mean Difference (IV, Random, 95% CI)0.01 [-0.19, 0.21]
13.3 Sexual function3428Std. Mean Difference (IV, Random, 95% CI)0.00 [-0.20, 0.20]
13.4 Sexual bother3428Std. Mean Difference (IV, Random, 95% CI)0.17 [-0.03, 0.37]
13.5 Bowel function3430Std. Mean Difference (IV, Random, 95% CI)-0.07 [-0.27, 0.13]
13.6 Bowel bother3428Std. Mean Difference (IV, Random, 95% CI)0.02 [-0.18, 0.22]
14 Symptom-related quality of life (UCLA-PCI) at six months2 Std. Mean Difference (IV, Random, 95% CI)Subtotals only
14.1 Urinary function2357Std. Mean Difference (IV, Random, 95% CI)-0.01 [-0.24, 0.21]
14.2 Urinary bother2357Std. Mean Difference (IV, Random, 95% CI)-0.01 [-0.23, 0.21]
14.3 Sexual function2357Std. Mean Difference (IV, Random, 95% CI)0.11 [-0.11, 0.33]
14.4 Sexual bother2357Std. Mean Difference (IV, Random, 95% CI)0.20 [-0.02, 0.43]
14.5 Bowel function2357Std. Mean Difference (IV, Random, 95% CI)-0.20 [-0.42, 0.02]
14.6 Bowel bother2357Std. Mean Difference (IV, Random, 95% CI)-0.04 [-0.26, 0.18]
15 Symptom-related quality of life (UCLA-PCI) at 12 months2 Std. Mean Difference (IV, Random, 95% CI)Subtotals only
15.1 Urinary function2351Std. Mean Difference (IV, Random, 95% CI)0.11 [-0.12, 0.33]
15.2 Urinary bother2351Std. Mean Difference (IV, Random, 95% CI)-0.05 [-0.27, 0.18]
15.3 Sexual function2351Std. Mean Difference (IV, Random, 95% CI)0.10 [-0.13, 0.32]
15.4 Sexual bother2351Std. Mean Difference (IV, Random, 95% CI)0.21 [-0.08, 0.49]
15.5 Bowel function2351Std. Mean Difference (IV, Random, 95% CI)-0.05 [-0.28, 0.17]
15.6 Bowel bother2351Std. Mean Difference (IV, Random, 95% CI)0.00 [-0.22, 0.22]
16 Self-efficacy at end of intervention3337Std. Mean Difference (IV, Fixed, 95% CI)0.16 [-0.05, 0.38]
17 Self-efficacy at eight and 12 months1 Std. Mean Difference (IV, Random, 95% CI)Subtotals only
17.1 Self-efficacy at eight months1228Std. Mean Difference (IV, Random, 95% CI)-0.05 [-0.31, 0.21]
17.2 Self-efficacy at 12 months1218Std. Mean Difference (IV, Random, 95% CI)-0.07 [-0.34, 0.20]
18 Prostate cancer knowledge at end of intervention2506Std. Mean Difference (IV, Random, 95% CI)0.51 [0.32, 0.71]
18.1 Group-based intervention1250Std. Mean Difference (IV, Random, 95% CI)0.68 [0.41, 0.95]
18.2 Individual-based intervention1256Std. Mean Difference (IV, Random, 95% CI)0.35 [0.08, 0.62]
19 Prostate cancer knowledge at three months post-intervention1256Std. Mean Difference (IV, Random, 95% CI)0.31 [0.04, 0.58]
19.1 Individual-based intervention1256Std. Mean Difference (IV, Random, 95% CI)0.31 [0.04, 0.58]
20 Uncertainty at end of intervention2916Std. Mean Difference (IV, Random, 95% CI)-0.05 [-0.35, 0.26]
21 Uncertainty at six to eight months post-intervention2887Std. Mean Difference (IV, Random, 95% CI)0.04 [-0.09, 0.17]
22 Uncertainty at 12 months post-intervention2864Std. Mean Difference (IV, Random, 95% CI)0.07 [-0.06, 0.21]
23 Distress2 Std. Mean Difference (IV, Random, 95% CI)Subtotals only
23.1 Distress at end of intervention2916Std. Mean Difference (IV, Random, 95% CI)0.02 [-0.11, 0.15]
23.2 Distress at six to eight months post-intervention2887Std. Mean Difference (IV, Random, 95% CI)0.06 [-0.07, 0.19]
23.3 Distress at 12 months post-intervention2864Std. Mean Difference (IV, Random, 95% CI)0.02 [-0.11, 0.15]
24 Depression at end of intervention3434Std. Mean Difference (IV, Random, 95% CI)-0.18 [-0.51, 0.15]
24.1 Group-based intervention2362Std. Mean Difference (IV, Random, 95% CI)-0.06 [-0.38, 0.25]
24.2 Individual-based intervention172Std. Mean Difference (IV, Random, 95% CI)-0.59 [-1.06, -0.11]
25 Depression at six months post-intervention2364Std. Mean Difference (IV, Random, 95% CI)-0.17 [-0.44, 0.09]
25.1 Group-based intervention2364Std. Mean Difference (IV, Random, 95% CI)-0.17 [-0.44, 0.09]
26 Depression at 12 months post-intervention1250Std. Mean Difference (IV, Random, 95% CI)-0.02 [-0.29, 0.24]
26.1 Group-based intervention1250Std. Mean Difference (IV, Random, 95% CI)-0.02 [-0.29, 0.24]
Analysis 1.1.

Comparison 1 Psychosocial intervention versus usual care, Outcome 1 General health–related quality of life: physical component at end of intervention.

Analysis 1.2.

Comparison 1 Psychosocial intervention versus usual care, Outcome 2 General health–related quality of life: physical component at four to six months.

Analysis 1.3.

Comparison 1 Psychosocial intervention versus usual care, Outcome 3 General health–related quality of life: physical component at eight to 12 months.

Analysis 1.4.

Comparison 1 Psychosocial intervention versus usual care, Outcome 4 General health–related quality of lIfe: mental component at end of intervention.

Analysis 1.5.

Comparison 1 Psychosocial intervention versus usual care, Outcome 5 General health–related quality of life: mental component at four to six months.

Analysis 1.6.

Comparison 1 Psychosocial intervention versus usual care, Outcome 6 General health–related quality of life: mental component at eight to 12 months.

Analysis 1.7.

Comparison 1 Psychosocial intervention versus usual care, Outcome 7 Cancer-related quality of life (FACT-G) at end of intervention.

Analysis 1.8.

Comparison 1 Psychosocial intervention versus usual care, Outcome 8 Cancer-related quality of life (FACT-G) at eight and 12 months.

Analysis 1.9.

Comparison 1 Psychosocial intervention versus usual care, Outcome 9 Prostate cancer–specific quality of life (FACT-P).

Analysis 1.10.

Comparison 1 Psychosocial intervention versus usual care, Outcome 10 Symptom-related quality of life (EPIC) at end of intervention.

Analysis 1.11.

Comparison 1 Psychosocial intervention versus usual care, Outcome 11 Symptom-related quality of life (EPIC) at six to eight months post-intervention.

Analysis 1.12.

Comparison 1 Psychosocial intervention versus usual care, Outcome 12 Symptom-related quality of life (EPIC) at 12 months post-intervention.

Analysis 1.13.

Comparison 1 Psychosocial intervention versus usual care, Outcome 13 Symptom-related quality of life (UCLA-PCI) at end of intervention.

Analysis 1.14.

Comparison 1 Psychosocial intervention versus usual care, Outcome 14 Symptom-related quality of life (UCLA-PCI) at six months.

Analysis 1.15.

Comparison 1 Psychosocial intervention versus usual care, Outcome 15 Symptom-related quality of life (UCLA-PCI) at 12 months.

Analysis 1.16.

Comparison 1 Psychosocial intervention versus usual care, Outcome 16 Self-efficacy at end of intervention.

Analysis 1.17.

Comparison 1 Psychosocial intervention versus usual care, Outcome 17 Self-efficacy at eight and 12 months.

Analysis 1.18.

Comparison 1 Psychosocial intervention versus usual care, Outcome 18 Prostate cancer knowledge at end of intervention.

Analysis 1.19.

Comparison 1 Psychosocial intervention versus usual care, Outcome 19 Prostate cancer knowledge at three months post-intervention.

Analysis 1.20.

Comparison 1 Psychosocial intervention versus usual care, Outcome 20 Uncertainty at end of intervention.

Analysis 1.21.

Comparison 1 Psychosocial intervention versus usual care, Outcome 21 Uncertainty at six to eight months post-intervention.

Analysis 1.22.

Comparison 1 Psychosocial intervention versus usual care, Outcome 22 Uncertainty at 12 months post-intervention.

Analysis 1.23.

Comparison 1 Psychosocial intervention versus usual care, Outcome 23 Distress.

Analysis 1.24.

Comparison 1 Psychosocial intervention versus usual care, Outcome 24 Depression at end of intervention.

Analysis 1.25.

Comparison 1 Psychosocial intervention versus usual care, Outcome 25 Depression at six months post-intervention.

Analysis 1.26.

Comparison 1 Psychosocial intervention versus usual care, Outcome 26 Depression at 12 months post-intervention.

Appendices

Appendix 1. Cochrane Central Register of Controlled Trials (CENTRAL) search strategy

ID SearchHits
1. prostate cancer (5409)
2. Prostatitis (395)
3. #1 or #2 (5771)
4. educational intervention (5980)
5. cognitive intervention (7726)
6. psycho-social intervention (155)
7. psychosocial intervention (3465)
8. behavioral intervention (9120)
9. behavioural intervention (9120)
10. #4 or #5 or #6 or #7 or #8 or #9 (18635)
11. #3 and #10 (197)

Appendix 2. MEDLINE search strategy

Database: Ovid MEDLINE(R) <1946 to October Week 1 2013>

1 psychological.mp. (338921)

2 psychosocial.mp. (58263)

3 psycho-social.mp. (2523)

4 educational.mp. (150949)

5 psychoeducational.mp. (1428)

6 psycho-educational.mp. (467)

7 behavioural.mp. (47681)

8 behavioral.mp. (194685)

9 prostate cancer.mp. or exp Prostatic Neoplasms/ (103286)

10 prostatitis.mp. or exp Prostatitis/ (5671)

11 intervention.mp. (322256)

12 therapy.mp. (1732183)

13 program.mp. (328099)

14 programme.mp. (61636)

15 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 (717014)

16 9 or 10 (107920)

17 11 or 12 or 13 or 14 (2289150)

18 15 and 16 and 17 (689)

19 trial.mp. (865988)

20 18 and 19 (193)

Appendix 3. EMBASE search strategy

Database: EMBASE <1974 to 21 October 2013>
Search Strategy:
1 psychological.mp. (665978)
2 psychosocial.mp. (88298)
3 psycho-social.mp. (4429)
4 educational.mp. (149142)
5 psychoeducational.mp. (2105)
6 psycho-educational.mp. (851)
7 behavioural.mp. (65005)
8 behavioral.mp. (226566)
9 prostate cancer.mp. or exp Prostatic Neoplasms/ (156516)
10 prostatitis.mp. or exp Prostatitis/ (9068)
11 intervention.mp. (457217)
12 therapy.mp. or therapy/ (3252629)
13 program.mp. (598184)
14 programme.mp. (85374)
15 trial.mp. (1276593)
16 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 (1082193)
17 9 or 10 (163781)
18 11 or 12 or 13 or 14 (4113763)
19 15 and 16 and 17 and 18 (387)

Appendix 4. PsycINFO search strategy

Database: PsycINFO <1806 to October Week 2 2013>

1 psychological.mp. (280339)

2 psychosocial.mp. (82314)

3 psycho-social.mp. (2390)

4 educational.mp. (188548)

5 psychoeducational.mp. (4712)

6 psycho-educational.mp. (1569)

7 behavioural.mp. (29379)

8 behavioral.mp. (224189)

9 prostate cancer.mp. or exp Prostatic Neoplasms/ (1691)

10 prostatitis.mp. or exp Prostatitis/ (74)

11 intervention.mp. (156426)

12 therapy.mp. (291862)

13 program.mp. (172778)

14 programme.mp. (14414)

15 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 (719502)

16 9 or 10 (1761)

17 11 or 12 or 13 or 14 (555720)

18 15 and 16 and 17 (192)

19 trial.mp. (66424)

20 18 and 19 (30)

Contributions of authors

KP conceived of the review, drafted the proposal and undertook the review.

SM provided supervision and statistical advice, appraised papers, extracted data and commented on the drafts.

EM conceived of the review, helped to draft the proposal, provided clinical expertise, appraised papers, extracted data and commented on the drafts.

JN provided advice, appraised papers and helped to resolve discrepancies in appraisals between other pairs of review authors.

CS appraised papers, extracted data, checked data input for analysis and commented on the drafts.

EJH appraised papers, extracted data and commented on the drafts.

MMN contributed to development of the summary of findings table and critically reviewed the review.

PD contributed to development of the summary of findings table and critically reviewed the review.

Declarations of interest

KP and CS received a grant that included support for travel from the Health and Social Care Research and Development Office (Northern Ireland) as part of the Cochrane Training Fellowship Scheme. JN received travel support unrelated to the review from the National Institute for Social Care and Health Research (NISCHR) Steering Group Wales and The Cochrane Collaboration to attend meetings, and received support from grants unrelated to the review from the National Institute for Health Research (NIHR), the NISCHR, The Cochrane Collaboration and Together for Short Lives.

Sources of support

Internal sources

  • The University of Ulster, UK.

External sources

  • This review was funded by the Health and Social Care Research & Development Office (Northern Ireland) as part of the Cochrane Training Fellowship Scheme, UK.

Differences between protocol and review

This review took longer than originally expected; therefore the search for studies was extended to October 2013. We did not search CancerLit and CINAHL, as we had proposed in the protocol, but in the review, we added a search of EMBASE and handsearching of the journals European Journal of Cancer Care, Journal of Advanced Nursing and Journal of Clinical Nursing. Five review authors, rather than two as proposed in the protocol, working independently in pairs were involved in selection of studies, data extraction and assessment of risk of bias to involve more review authors in these review stages. We used the standardised mean difference (SMD) instead of the mean difference (MD) for meta-analysis. Subgroup analyses for ethnicity and age were not conducted because separate data on these variables were not available. Subgroup analysis for stage of disease was not possible because the data were insufficient. In accordance with the recommendations of peer referees at the review stage, we attempted to conduct post hoc subgroup analyses related to treatment with androgen deprivation therapy. The decision to include a summary of findings table was made after publication of the protocol.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Ames 2011

Methods

Study design: randomised controlled trial

Follow-up: week nine (end of treatment) and six months post-treatment

Participants

Fifty-seven men at the biochemical recurrent stage of prostate cancer were randomly assigned to an intervention (n = 30) or to a control group (n = 27). The median age of participants was 76 years. The ethnic composition was 89% white, 10% black and 1% other

Eligibility criteria included men ≥ 18 years old, expected survival ≥ one year, a diagnosis of prostate cancer with prostate-specific antigen (PSA) recurrence post-surgery or post-radiation without clinical or X-ray evidence of metastases and a Mini-Mental Status Examination score ≥ 24. Participants were recruited from two medical centres in Northeast Florida, USA

Interventions

The multidisciplinary quality of life intervention (MQOL) consisted of eight weekly group sessions, each lasting one hour. As well as providing information regarding prostate cancer, the sessions dealt with goal setting, problem solving, relaxation training, nutrition, physical activity and conditioning, mood management, social support and maintenance of positive health behaviour. The sessions were delivered mainly by a clinical health psychologist with input from a medical oncologist, a dietician and a psychiatrist, as appropriate

Control group participants did not receive any form of quality of life intervention during the study period but were offered the MQOL following completion of the study

Outcomes

Quality of life (general healthrelated)

MOS 36-Item Short Form Survey (SF–36)

Quality of life (prostate cancerspecific)

Functional Assessment of Cancer Therapy–Prostate (FACT-P)

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskMethod of randomisation not given
Allocation concealment (selection bias)Unclear riskNot stated
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNot stated
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot stated
Incomplete outcome data (attrition bias)
All outcomes
High riskNumber of participants in each group at follow-up time points not given
Selective reporting (reporting bias)Low riskAll outcome data reported
Other biasLow riskNone

Bailey 2004

Methods

Study design: randomised controlled trial

Follow-up: 10 weeks post-baseline (five weeks after intervention was completed)

ParticipantsForty-one men were randomly assigned to an intervention group (n = 21) or a control group (n = 20). Thirty-nine men, with a mean age of 75.4 years, completed the study at 10 weeks post-baseline. Eighty-six per cent were Caucasian and 14% African American. Most had T1 or T2 tumours (97%). Inclusion criteria were patients with a diagnosis of B1, B2 and C1 prostate cancer who had elected watchful waiting in consultation with their physician, with watchful waiting time from one to 124 months. Participants were recruited from a central hospital in North Carolina, USA
Interventions

The intervention consisted of five weekly telephone calls by a male nurse, lasting an average of 13 minutes (range five to 26 minutes). The intervention, which was tailored to each participant's needs, was designed "to help men integrate uncertainty into their lives by teaching them to cognitively reframe the way in which they viewed their illness and the uncertainty it produced" (p. 341). The intervention had four components. The first encouraged participants to reframe uncertainty in a positive light. The second helped participants to incorporate uncertainty into their life structure. The third was designed to support participants’ belief that future treatment would be better. Finally, the fourth component encouraged self-monitoring and vigilance. Information about the disease and about treatment was given

Participants in the control group received usual care and "had access to naturally occurring sources of support" (p. 341)

Outcomes

Quality of life

Two items asking respondents to rate their current quality of life and the quality of life they expect in six months, on a scale of zero to 10

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskUsed a table of random numbers
Allocation concealment (selection bias)Unclear riskNot stated
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNot stated
Blinding of outcome assessment (detection bias)
All outcomes
High risk"data collectors did not deliver the intervention but were aware of group assignment at the time of enrollment" (p. 342)
Incomplete outcome data (attrition bias)
All outcomes
Low riskMissing data accounted for
Selective reporting (reporting bias)Low riskAll data outcomes reported
Other biasLow riskNone

Berglund 2007

Methods

Study design: randomised controlled trial

Follow-up: six months and 12 months post-baseline

Participants

Two hundred eleven men with prostate cancer were randomly assigned to one of the following four groups: physical training (n = 53), information (n = 55), physical training and information (n = 52) and control (n = 51)

One hundred fifty-eight completed the study at 12 months post-baseline. The number that completed in each group was 34 in physical training (Phys), 46 in information (Info), 42 in physical training and information (Phys and Info) and 36 in control. The average age of participants was 69 years, with a range of 43 to 86 years. No ethnic composition of the sample was given. The inclusion criterion was six months after diagnosis of prostate cancer. Overall, 20% of the final sample had metastasis. Patients were excluded if they had severe hearing or vision impairment, were not Swedish speaking or were physically and mentally disabled. Participants were recruited from a university hospital in Uppsala, Sweden

Interventions

The physical training group (Phys) met weekly for seven weeks. A physiotherapist led 60-minute physical training sessions (movement and fitness training, relaxation, breathing exercises and pelvic floor exercises). A booster session was held every two months after the conclusion of the training exercises. Opportunities were provided to discuss exercises and ‘other important issues’

The information (Info) group, led by a nurse, consisted mainly of sharing information on prostate cancer and its treatment and side effects. Opportunities were provided to discuss with the group leader participants' experiences regarding diagnosis and communication. Each session lasted 60 minutes and was delivered once weekly for seven weeks

The information plus physical training (Phys and Info) group, which lasted 135 minutes, took place once weekly for seven weeks and was a combination of the Phys group and Info group programmes. Physical training was given before the information session was held

The control group received standard care. Participants received information, which consisted of two leaflets. They also had the opportunity to phone a nurse if they had any questions but did not partake in group information or in physical training

Outcomes

Depression

14-Item Hospital Anxiety and Depression Scale (HADS)

Quality of life (cancer-related)

EORTC QLQ-C30 version 3.0

NotesThe non-completion rate was 25.1% at one year post-baseline, resulting in an even smaller sample size in the three comparison groups and the one control group
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk

Consecutive "patients were randomly assigned to one of four groups" (p. 84)

Method of randomisation not explained

Allocation concealment (selection bias)Unclear riskNot stated
Blinding of participants and personnel (performance bias)
All outcomes
High riskParticipants informed of group assignment; consequently some dropped out
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot stated
Incomplete outcome data (attrition bias)
All outcomes
Low riskMissing data explained
Selective reporting (reporting bias)High riskData on depression and quality of life collected at six months and not reported
Other biasHigh risk"in addition to relatively small randomisation groups at the beginning, 20% of the patients dropped out during the first year of the study, which resulted in decreasing power at each assessment period" (p. 88)

Campbell 2007

Methods

Study design: randomised controlled trial

Follow-up: six weeks post-intervention

ParticipantsForty men were randomly assigned to an intervention group (n = 20) or a control group (n = 20). Thirty men completed the study (12 in the intervention group and 18 in the control group). All participants were African Americans with a mean age of 60.9 years (SD 8.0) in the intervention group and 62.9 years (SD 9.7) in the control group. Ninety-three per cent had undergone radical prostatectomy. Eligibility criteria were African American men who were beyond the acute diagnosis and treatment phase for prostate cancer and with a Karnofsky Performance Status score ≥ 60, along with their identified intimate partners. Participants were recruited from a urology clinic in North Carolina, USA
Interventions

The intervention was a telephone-based educational intervention designed to teach cognitive and behavioural techniques that increase self-efficacy for symptom management, which can reduce symptom severity and emotional distress. The protocol consisted of providing information about prostate cancer and side effects and teaching problem-solving skills and specific cognitive and behavioural coping skills (e.g. communication skills, relaxation training, activity pacing). The intervention, which was delivered by African American doctoral level medical psychologists, lasted one hour each week for six weeks

The control group received routine care provided through the medical outpatient programme

Outcomes

Self-efficacy

Self-efficacy for Symptom Control Inventory (SESCI)

Quality of life (symptom-related)

50-Item Expanded Prostate Cancer Index Composite (EPIC)

Quality of life (general healthrelated)

Short Form Health Survey (SF-36)

NotesDescribed as a pilot study by the authors. Dropout rate was 25% overall and 40% in the intervention group. The intervention was also delivered to men with prostate cancer and their intimate partners
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskUsed a 'computerized random number generator' (personal correspondence)
Allocation concealment (selection bias)Low riskUsed 'opaque sealed envelopes' (personal correspondence)
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNot stated
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot stated
Incomplete outcome data (attrition bias)
All outcomes
Low riskMissing data reported
Selective reporting (reporting bias)Low riskAll outcome data reported
Other biasHigh riskHigher attrition rate in the psychosocial intervention group (40%) compared with the control group (10%)

Carmack Taylor 2006

Methods

Study design: randomised controlled trial

Follow-up: six and 12 months post-intervention

ParticipantsOne hundred thirty-four men with prostate cancer were randomly assigned to one of three groups: lifestyle programme (n = 46), educational support programme (n = 50) and a standard care group (n = 38). One hundred thirteen completed the programme (35 in lifestyle programme, 44 in educational support programme and 34 in standard care). The average age of participants was 69.2 years (range 44.8 to 89.0 years), and ethnic composition was 73.1% white, 20.1% African American and 6.7% ‘other’. Recruited participants resided within one hour travelling distance from a cancer centre in Texas, USA. Eligibility criteria included men with a diagnosis of prostate cancer, regardless of stage, who were receiving continuous androgen ablation therapy with the expectation to continue therapy for one year and be able to speak and understand English. Exclusion criteria were participating in 30 minutes of purposeful physical activity (three or more times a week for six months) and conditions for which physical activity would be contraindicated. At baseline, participants had been on androgen ablation therapy for an average of 32.7 months. No details on stage of prostate cancer were reported
Interventions

This study tested two psychosocial interventions: the lifestyle programme (LP) and the educational support programme (ESP). Participants were compared with a standard care (SC) group. Participants in the LP and ESP attended small group meetings (approximately eight men) for six months. Both programmes included an orientation session, 16 weekly sessions and four bi-weekly sessions, each lasting 1.5 hours. The LP aimed to build self-efficacy and cognitive behavioural skills to adapt and maintain regular physical activity. Each session covered a specific cognitive behavioural skill for increasing physical activity, including self-monitoring, goal setting, problem solving, overcoming barriers, cognitive restructuring and rewarding oneself. The ESP consisted of a series of preestablished questions covering topics such as diet and prostate cancer, side effects of androgen ablation and sexuality. This programme used a discussion format but provided neither skills training nor instructions on behaviour to improve quality of life. The LP and the ESP were led by group facilitators supervised by a licensed clinical psychologist

Participants in the standard care group did not attend group support meetings but received one mailing of educational material about community resources. These materials were also given to participants in the intervention groups

Outcomes

Quality of life (general healthrelated)

36-Item MOS Short Form Health Survey (SF-36)

Depression

Center for Epidemiological Studies-Depression (CES-D)

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskMinimisation used
Allocation concealment (selection bias)Unclear riskNot stated
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNot stated
Blinding of outcome assessment (detection bias)
All outcomes
Low risk"Research staff conducting assessments were blind to the participants' study program assignment" (p. 851)
Incomplete outcome data (attrition bias)
All outcomes
Low riskMissing data explained
Selective reporting (reporting bias)Low riskAll outcomes data reported
Other biasLow riskNone

Chambers 2013

Methods

Study design: randomised controlled trial

Follow-up: two, six, 12 and 24 months post-treatment

Participants

Seven hundred forty men, newly diagnosed with localised prostate cancer, were randomly assigned to an intervention group (n = 372) or a usual care group (n = 368). The ethnic composition was not given in the paper. From personal correspondence, the following data were obtained: most of the sample (85%) were of "British/Scottish/Welsh/Irish descent; the remainder were of European, Asian, or other ethnic background"

Eligibility criteria were newly diagnosed with localised prostate cancer, suitable for treatment with curative intent, with no evidence of metastatic disease on scan/X-rays; able to read and speak English; with no history of head injury, dementia or psychiatric illness and no other concurrent cancer. Participants were recruited from the geographic areas of South East and North Queensland, Australia

Interventions

This psychoeducational, telephone-delivered intervention consisted of elements such as decision support, challenging unhelpful conditions, psychoeducation about adjustment to prostate cancer, stress reduction techniques and coaching about problem-solving skills for side effects associated with prostate cancer treatments

The intervention was delivered by nurse counsellors in five sessions: two pretreatment sessions, two sessions at three and seven weeks post-treatment and a final booster session five months after treatment

The usual care group received standard medical management and no other support apart from a small set of already widely available patient education materials delivered by mail

Outcomes

Quality of life (general healthrelated)

Short Form Survey (SF-36)

Quality of life (symptom-related)

UCLA Prostate Cancer Index (EPIC)

Distress

Revised Impact of Events Scale (RIES)

Uncertainty

Decisional Conflict Scale-Revised

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskBlock randomisation
Allocation concealment (selection bias)Low riskSequence undertaken by project manager and concealed from investigators
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskParticipants and counsellors aware of who were receiving the intervention
Blinding of outcome assessment (detection bias)
All outcomes
Low riskAssessors blinded to condition
Incomplete outcome data (attrition bias)
All outcomes
Low riskMissing data explained
Selective reporting (reporting bias)Low riskAll outcomes reported
Other biasLow riskNone

Davison 1997

Methods

Study design: randomised controlled trial

Follow-up: five to six weeks post-baseline

ParticipantsSixty men newly diagnosed with prostate cancer were randomly assigned to an intervention group (n = 30) or a control group (n = 30). Selection criteria included having been told their diagnosis, not having had their initial treatment consultation, able to read, speak and write English and no evidence of mental confusion. The mean age of participants was 67.9 years. All were recruited from one community clinic in Winnipeg, Canada. No details were provided on the ethnic composition of the sample nor on the stage of their cancer, although participants were in the newly diagnosed phase
Interventions

The intervention was based on an empowerment model aimed at increasing self-efficacy in seeking information and making treatment decisions. It consisted of the provision of a written information package containing various types of information about prostate cancer. With the help of a nurse, participants in the intervention group identified the information they needed to assist them in deciding which treatment would be best for them. Additional questions that arose out of the discussion between the participant and the facilitator were added to the list

Participants in the control group received the same information package without the opportunity to discuss their information needs with a facilitator

Outcomes

Depression

Center for Epidemiological Studies Depression Scale (CES-D)

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskRandomisation method not described
Allocation concealment (selection bias)Unclear riskNot stated
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNot stated
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot stated
Incomplete outcome data (attrition bias)
All outcomes
Low riskMissing data accounted for
Selective reporting (reporting bias)Unclear riskAll outcome data reported but full statistical results for depression not presented
Other biasLow riskNone

Giesler 2005

Methods

Study design: randomised controlled trial

Follow-up: four, seven and 12 months post-baseline

ParticipantsNinety-nine men, diagnosed with prostate cancer and within six weeks post-treatment, were randomly assigned to an intervention group (n = 48) or a control group (n = 51). Fourteen participants dropped out at 12 months. Eligibility criteria were men with a diagnosis of stage T1a to T2c prostate cancer; scheduled to undergo or to have undergone surgery, external beam radiation or brachytherapy; 18 years or older and having a spouse or partner willing to participate in the trial. Most participants were Caucasian (n = 87; 90%). Eight per cent (n = 8) were African American. Participants were recruited from Indiana, Michigan and Louisville in the USA. Mean age of participants was 63.8 years
Interventions

The intervention focused primarily on issues related to sexual, urinary and bowel dysfunction, anxiety, relationship adjustment, depression and other common sequelae of cancer. The intervention was delivered by a nurse, whose task consisted of assessing participants’ problems or concerns, helping them identify strategies to address them and monitoring progress with the use of a laptop computer. Intervention visits took place once a month for six months. The first two visits were in person and the other four were over the telephone

Control group participants received standard care. No further detail was provided

Outcomes

Depression

20-Item Center for Epidemiological Studies-Depression Scale

Quality of life (general healthrelated)

Short Form Health Survey (SF-36)

Quality of life (prostate cancer)

Prostate Cancer Quality of Life Instrument (PCQoL)

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskFlip of a coin (personal correspondence)
Allocation concealment (selection bias)High risk"No concealment because the group participants had been assigned to was obvious to all" (personal correspondence)
Blinding of participants and personnel (performance bias)
All outcomes
High riskSee above
Blinding of outcome assessment (detection bias)
All outcomes
Low riskAssessors were blind to group assignment
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskDid not provide the number of participants at each follow-up time point for each group
Selective reporting (reporting bias)Unclear riskDid not give overall scores for each group on prespecified outcomes
Other biasLow riskNone

Lepore 2003

Methods

Study design: randomised controlled trial

Follow-up: two weeks, six months and 12 months post-intervention

ParticipantsTwo hundred seventy-nine men with prostate cancer were randomly assigned to a 'group education' intervention, a 'group education plus discussion' intervention or a control group. The number allocated to each group was not provided. Twenty-nine men (10.4%) did not complete the study. The number of participants (n = 250) who completed the study in each group was as follows: 'group education' intervention: n = 84; 'group education plus discussion' intervention: n = 86; and control: n = 80. The final sample comprised 226 (90.4%) Caucasians, 23 (9.2%) African Americans and one Asian American. Most of these men were at the T2 stage of prostate cancer (n = 173; 69.2%). The rest were at T1 (n = 45; 18%) and T3 (n = 32; 12.8%). All had recently completed treatment before enrolment in the study. They were recruited from 11 Greater Pittsburgh area urology and radiology clinics and hospitals
InterventionsThe two interventions, education group (GE) and education plus discussion group (GED), consisted of six sessions (one hour each week). The GE group received lectures on a range of prostate cancer–related topics from experienced health professionals. Lecturers allowed 10 minutes for questions, but participants were not encouraged to talk with one another. The GED group received the lectures plus 45 minutes of group discussion facilitated by a clinical psychologist. Participants in the control group received nothing beyond their standard medical care
Outcomes

Prostate cancer knowledge

13-Item investigator-developed questionnaire

Quality of life (general healthrelated)

Short Form Survey (SF-36)

Quality of life (symptom-related)

UCLA Prostate Cancer Index

Depressive symptoms

Center for Epidemiological Studies-Depression Scale (CES-D)

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskActual method of randomisation not stated
Allocation concealment (selection bias)Low risk"After the baseline interview, participants received a sealed envelope that revealed their assignment" (p. 445)
Blinding of participants and personnel (performance bias)
All outcomes
Low riskParticipants not informed of the hypotheses of the study
Blinding of outcome assessment (detection bias)
All outcomes
Low riskInterviewers blinded to experimental condition
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNumber of participants at each follow-up time point not reported for each group
Selective reporting (reporting bias)Unclear riskOutcome data for self-efficacy not reported
Other biasLow riskNone

Manne 2011

Methods

Study design: randomised controlled trial

Follow-up: eight weeks post-baseline

ParticipantsSeventy-one men and their partners were randomly assigned to an intervention group (n = 37) or a control group (n = 34). The mean age of the men was 60 years. Sixty-three (88.7%) were described as ‘white’ and eight (11.3%) as ‘non-white’. Eligibility criteria were those who were 18 years of age or older; diagnosed with localised prostate cancer in the year before the study; without a hearing impairment and able to speak English. All participants had an ECOG performance score of ‘0’ (fully active, able to carry on all predisease performance without restriction). Most (n = 60, 84.5%) were at stage II of prostate cancer, and 11 (15.5%) were at stage I. Forty-four (62%) had surgery; 20 (28.1%) had radiation therapy; three (4.2%) had hormone therapy and four (5.6%) had not begun treatment. Participants were recruited from two cancer centres in the Northeastern USA
Interventions

The intimacy-enhancing psychological intervention (IET) used in this study consisted of five 90-minute sessions (within two months) with couples. Its content focused on improving couples’ ability to comfortably share their thoughts and feelings regarding cancer, promoting mutual understanding and support regarding their own and one another’s cancer experiences, facilitating constructive discussion of cancer concerns and enhancing and maintaining emotional intimacy (p. 1199-200). Techniques were derived from cognitive behavioural and behavioural marital therapy. Five therapists provided the intervention

Usual care participants could access a social worker routinely and, if required, were referred to a psychiatrist or a psychologist

Outcomes

Distress (psychological)

24-Item Psychological Distress Scale of the Mental Health Inventory

Distress (cancer-specific)

15-Item Impact of Events Scale

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk"A computerized system was used for allocation" (personal correspondence)
Allocation concealment (selection bias)Unclear risk"Concealed from staff before assignment but not afterwards" (personal correspondence); however, the actual method of concealment not stated
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNot stated
Blinding of outcome assessment (detection bias)
All outcomes
Low riskParticipants sent questionnaires back to data collectors
Incomplete outcome data (attrition bias)
All outcomes
Low riskIntention-to-treat analysis carried out
Selective reporting (reporting bias)High riskOutcome data for psychological distress and cancer-specific distress not reported
Other biasLow riskNone

Mishel 2002

Methods

Study design: randomised controlled trial

Follow-up: four and seven months post-baseline

ParticipantsTwo hundred fifty-two men diagnosed with localised prostate cancer and with two weeks post–catheter removal after surgery and within three weeks into radiation therapy were randomly assigned to two intervention groups or a control group. No data on the number of men allocated to any of the three groups were provided, although 13 (5%) dropped out before the final time point. Mean age of participants was 64. One hundred thirty-four (56%) were Caucasian, and 105 (44%) were African American. Most (61%) had stage T2 tumours, 27% had stage III and 8% had T1 tumours. Some participants (4%) did not know the stage of their cancer. Most had undergone surgery (79%), and 21% had received radiation therapy. Participants were recruited from nine treatment facilities in Central and Eastern North Carolina, USA
Interventions

This study had two intervention groups and one control group. In one group, only the participant received the intervention (treatment direct group), and in the other, both the participant and a family member received the intervention (treatment-supplemented group). The intervention given to both groups was the same. It was an individually tailored, telephone-based psychoeducation intervention delivered during eight consecutive weeks (one hour per week). It consisted mainly of validating and reinforcing the views and behaviours of the participants, providing information, activating resources, teaching symptom management, structuring expectations and teaching problem solving and assertion techniques for communicating with healthcare providers. Both interventions were delivered by nurses

The usual care group received the same non-prostate cancer-specific printed material about health in general

Outcomes

Uncertainty

26-Item Version of Mishel Uncertainty in Illness Scale

Knowledge (cancer)

21-Item, investigator-developed Cancer Knowledge Scale

Distress

15-Item Symptom Distress Scale

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskStated that "men were blocked on ethnicity and randomly assigned to treatment or control groups", but randomisation method not given
Allocation concealment (selection bias)Unclear riskNot stated
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNot stated
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskMentioned that data collectors were not involved in delivery of the intervention (p. 1858), but no information provided on whether they were blinded to group allocation
Incomplete outcome data (attrition bias)
All outcomes
High riskNo information provided about the number of participants in each group at four and seven month follow-up time points
Selective reporting (reporting bias)Low riskAll outcomes reported
Other biasLow riskNone

Mishel 2009

Methods

Study design: randomised controlled trial

Follow-up: four weeks and three months post-baseline

ParticipantsTwo hundred fifty-six men, newly diagnosed with cancer, were randomly assigned to one of three groups: men only (n = 93), men and a primary support person (n = 89) or control group (n = 74). No data were available on the number who completed in each group. The mean age of participants was 62.5 years (SD 7.4), and the ethnic composition was Caucasian 71.5% and African American 28.5%. Participants were recruited from six sites in North Carolina, USA. Eligibility criteria included men at stage T1a, T1b, T1c or T2a or with a Gleason score < 10; PSA levels < 120; no major cognitive impairment; ability to read and access to a telephone
Interventions

The intervention dealt with communication strategies presented through a DVD, a booklet and four telephone calls to participants by a nurse trained in the intervention. The calls were delivered over seven to 10 days and consisted of helping participants to identify and formulate specific questions and to practice communication skills with the nurse on the telephone. The same intervention was delivered directly to the men only group and to the men and primary support person group

Control group participants received only a handout on staying healthy during treatment

Outcomes

Knowledge (prostate cancer)

20-Item, investigator-developed Prostate Cancer Scale

Quality of life

Single item, in a visual analogue format, asking respondents to indicate, on a scale of one to 10, how positive they felt about their quality of life

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskRandomisation method not described
Allocation concealment (selection bias)Unclear riskNot stated
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNot stated
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskData collection done over the phone (p. 350) but not known whether assessors were aware of group allocation
Incomplete outcome data (attrition bias)
All outcomes
High riskNumber of participants in each group at follow-up time points not given
Selective reporting (reporting bias)Low riskAll outcome data reported
Other biasLow riskNone

Northouse 2007

Methods

Study design: randomised controlled trial

Follow-up: four, eight and 12 months post-baseline

ParticipantsTwo hundred sixty-three participant-spouse dyads were randomly assigned to a standard care group (n = 143) or a 'standard care plus a five-session family intervention' group (n = 129). Of the 235 who completed at four months, 84% were Caucasian, 14% African American and 2% Hispanic. The average age of male participants was 63 years (SD 9.78; range 42 to 90). Sixty-five per cent of men were in the newly diagnosed stage, 14% in the biochemical recurrent stage and 21% in the advanced stage of prostate cancer. Eligibility criteria were patients ≥ 30 years of age; a life expectancy ≥ 12 months and a spouse or live-in partner and residence within 75 miles of participating cancer centres. Those with a second primary cancer were excluded. Participants were recruited from three large cancer centres in the Midwest of the USA.
Interventions

The supportive-educative intervention, called the FOCUS programme, comprised five core components: family involvement, optimistic attitude, coping effectiveness, uncertainty reduction and symptom management. Nurses were trained to deliver the intervention via three 90-minute home visits and two 30-minute telephone sessions spaced two weeks apart and delivered between baseline and four months

Participants in the control group received standard clinic care at their cancer centre. Although these centres offered some support groups, no specific psychosocial resources were provided for couples facing prostate cancer

Outcomes

Quality of life (general healthrelated)

Medical Outcomes Study (MOS SF-12 version 2)

Quality of life (cancer-related)

Functional Assessment of Cancer Treatment (FACT-G version 4)

Quality of life (prostate cancerspecific)

Functional Assessment of Cancer Treatment (FACT-P)

Quality of life (symptom-related)

50-Item Expanded Prostate Cancer Index Composite (EPIC)

Uncertainty

28-Item Mishel Uncertainty in Illness Scale

Self-efficacy

17-Item Lewis Cancer Self-efficacy Scale

Distress (general symptom)

16-Item Symptom Scale of the Omega Screening Questionnaire (OSQ)

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk"Used a random number list" (personal correspondence)
Allocation concealment (selection bias)High riskParticipants were told of group allocation as per IRB requirements (personal correspondence)
Blinding of participants and personnel (performance bias)
All outcomes
High riskAs above
Blinding of outcome assessment (detection bias)
All outcomes
Low riskData collection nurses, who were blinded to group assignment, collected the data
Incomplete outcome data (attrition bias)
All outcomes
Low riskRandom-effects regression models were used to analyse all available data rather than dropping participants with missing data (p. 2814)
Selective reporting (reporting bias)Low riskAll outcome data reported or available from the author
Other biasLow riskNone

Parker 2009

Methods

Study design: randomised controlled trial

Follow-up: six weeks, six months and 12 months post-baseline

ParticipantsOne hundred fifty-nine men at the early stage of prostate cancer, undergoing radical prostatectomy, were randomly assigned to one of three groups: stress management (n = 53), supportive attention (n = 54) or standard care (n = 52). Their mean age was 60.5 years. The ethnic composition of participants was as follows: 78% (n = 124) white, 13% (n = 21) African American and 4.4% (n = 7) Hispanic/Latino. Almost three quarters (74.4%) of the men were at stage II prostate cancer; 12.8% were at stage I and 12.8% at stage III. Participants were recruited from three hospitals in Texas, USA
Interventions

Two intervention groups and one control group were included

The stress management (SM) intervention consisted of two 60- to 90-minute and two short booster individual sessions with a clinical psychologist. The sessions were cognitive behavioural in nature with approximately 60% of the time focused on relaxation skills. Problem-solving strategies were discussed, and informational material and audiotapes were provided

The supportive attention (SA) intervention consisted of two 60- to 90-minute and two short booster sessions, provided before and after surgery. The same clinical psychologist (as in the SM group) provided empathy and used reflective listening skills. Participants had the opportunity to discuss their concerns with the psychologist

The SM and SA interventions were delivered approximately one to two weeks before surgery

The standard care group (control) received routine care and did not have meetings with a clinical psychologist

Outcomes

Quality of life (general healthrelated)

Short Form-36 (SF-36)

Quality of life (symptom-related)

UCLA Prostate Cancer Index (PCI)

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskMinimisation used
Allocation concealment (selection bias)High risk"It was not a blinded trial" (personal communication)
Blinding of participants and personnel (performance bias)
All outcomes
High riskBlinding of participants not stated. The same psychologist delivered both interventions
Blinding of outcome assessment (detection bias)
All outcomes
Low risk"A dedicated research assistant, who was blinded to group assignment, collected all measures" (p. 3170)
Incomplete outcome data (attrition bias)
All outcomes
High riskReasons for non-completion not reported
Selective reporting (reporting bias)Low riskAll data on relevant outcomes reported
Other biasHigh riskA high attrition rate reported (> 36%)

Penedo 2006

Methods

Study design: randomised controlled trial

Follow-up: within two to three weeks post-intervention

ParticipantsTwo hundred thirty-five men who had undergone radical prostatectomy or radiation therapy for stage I or II prostate cancer were randomly assigned to a cognitive behavioural stress management intervention (n = 134) or a control group (n = 101). Participants were non-Hispanic white (40%), Hispanic (41%) and African American (18%), and the average age was 65.1 (SD 7.7; range 47 to 84) years. Eligibility criteria included patients over 45 years old, with no prior history of cancer, who had received treatment within the 18 months before study enrolment. A screening interview was used to assess reading level and absence of cognitive impairment. Participants were recruited from a number of sites in Florida, USA
Interventions

The cognitive behavioural stress management (CBSM) intervention for prostate cancer was designed to provide participants with skills to manage daily stressors as well as prostate cancer–associated physical and emotional challenges, including sexual dysfunction, spousal or partner relations and engagement of the medical systems. The intervention provided individuals with effective coping skills and stress management techniques, as well as relevant information about prostate cancer, treatment and recovery. It comprised didactic instruction, discussion and relaxation. It was delivered to groups of four to eight participants for two hours once a week for 10 weeks

Control group participants met once for a four-hour seminar (in groups of four to eight) and received basic educational materials on the stress management techniques presented in the CBSM intervention

Outcomes

Quality of life (cancer-related)

Functional Assessment of Cancer Therapy-General Module (FACT-G)

Notes"No study-related adverse events were reported in either condition" (p. 263)
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskFlipping a coin
Allocation concealment (selection bias)Unclear riskNot stated
Blinding of participants and personnel (performance bias)
All outcomes
High riskStated that several participants were aware of group assignment, although this was not overtly discussed with them
Blinding of outcome assessment (detection bias)
All outcomes
Low riskAssessors and data entry staff blind to participant group assignment
Incomplete outcome data (attrition bias)
All outcomes
Low riskMissing data explained
Selective reporting (reporting bias)Low riskAll outcome data reported
Other biasLow riskNone

Penedo 2007

Methods

Study design: randomised controlled trial

Follow-up: two to three weeks post-intervention

ParticipantsNinety-three monolingual Spanish-speaking men with prostate cancer were randomly assigned to a cognitive behavioural stress management intervention (n = 53) or a control group (n = 40). Of these, 71 completed the study at two to three weeks post-intervention. Thirty-two (45.1%) had undergone radical prostatectomy, and 39 (54.9%) had undergone radiotherapy for localised prostate cancer within 21 months before the study began. Their mean age was 63.5 (SD 7.6) years. The ethnic composition was Cuban/Cuban-American (62%), Colombian/Colombian-American (15%), other Central American countries (10%) and other South American countries (4%). Eight per cent did not specify their country of origin. Average years lived in the USA was 27 (SD 15.5). Participants were recruited in Florida, USA
Interventions

The one-week, group-based cognitive behavioural stress management (CBSM) intervention was tailored for Spanish-speaking men in terms of their specific cultural and linguistic needs. Each session lasted two hours (90 minutes didactic and 30 minutes relaxation training). Topics included cognitive behavioural approaches to stress management (e.g. identification of distorted thoughts, rational thought replacement, effective coping, anger management, assertiveness training and social support development) and information specific to prostate cancer treatment and side effects. Relaxation training included deep breathing, progressive muscle relation and imagery

Control group participants were invited (in groups of four to six) to attend a half-day psychoeducational seminar consisting of a single four-hour session presenting a summary of the content of the CBSM programme. Both groups were co-facilitated by a clinical health psychology graduate student and a doctoral level licensed clinical psychologist, both of whom were fluent in Spanish and trained in the CBSM protocol

Outcomes

Quality of life (cancer-related)

Functional Assessment of Cancer Therapy-General Module (FACT-G) Spanish version 4

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk"Drawing numbers from a box" (p. 166)
Allocation concealment (selection bias)Unclear riskNot stated
Blinding of participants and personnel (performance bias)
All outcomes
High riskStated that several participants were aware of group assignment, although this was not overtly discussed with them
Blinding of outcome assessment (detection bias)
All outcomes
Low riskAssessors and data entry staff blind to group assignment
Incomplete outcome data (attrition bias)
All outcomes
Low riskMissing data accounted for
Selective reporting (reporting bias)Low riskAll outcome data reported
Other biasLow riskNone

Thornton 2004

Methods

Study design: randomised controlled trial

Follow-up: three weeks, six months and one year post-baseline

ParticipantsOne hundred six men with prostate cancer (beginning before radical prostatectomy and proceeding across the first year of surgery) were randomly assigned to an intervention group (n = 53) or a control group (n = 53). Of these, 92 (87%) were Caucasian, five (4.7%) African American and six (6.4%) Latino. Their mean age was 61.2 years. Most (94%) had undergone a nerve-sparing radical prostatectomy for early (stage A or B, 82%) or locally extended prostate cancer (stage C, 17%). Twenty-six (24.5%) participants did not complete the study at the final time point. Participants were recruited from a cancer centre in California, USA
Interventions

The intervention was a 45-minute, one-time psychological intervention aimed at enhancing communication with healthcare professionals and between the participant-partner dyad. It was also designed to help participants obtain relevant information. It was delivered by an MSc trained clinical or counselling psychologist or a clinical social worker

The control group received standard care and information primarily related to surgery (dealing with issues such as sexuality, pain control, incontinence, pelvic floor exercises, medication and support group information). Participants in the intervention group received standard care, as well as the 45-minute intervention session

Outcomes

Quality of life (general healthrelated)

Short Form Health Survey (SF-36)

Quality of life (prostate cancerspecific)

Functional Assessment of Cancer Treatment-Prostate version (FACT-P)

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskMethod of randomisation not stated
Allocation concealment (selection bias)Unclear riskNot stated
Blinding of participants and personnel (performance bias)
All outcomes
High riskPatients were informed of their group assignment as per instruction from the Institutional Review Board (p. 19)
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot stated
Incomplete outcome data (attrition bias)
All outcomes
High riskMissing data on participants who did not complete the study not explained
Selective reporting (reporting bias)High riskData collected at six months not reported. Outcome data for the control group at the three follow-up time points not reported
Other biasLow riskNone

Weber 2007

Methods

Study design: randomised controlled trial

Follow-up: four weeks and eight weeks post-intervention

ParticipantsEighty-one men, who had undergone radical prostatectomy previously, were randomly assigned to an intervention group or a control group. Complete data were available on 72 men (37 in the intervention group and 35 in the control group). Of these, 60 (83.3%) were white, eight (11.1%) African American, two (4.2%) Hispanic and one Asian. Their mean age was 60 years (SD 7), with a range of 47 to 74 years. Inclusion criteria were men with localised prostate cancer; 45 years or older; with no psychiatric disorders and English speaking. Participants were recruited from the urology clinics at two tertiary care medical centres in the Southeastern USA
Interventions

In addition to usual care from their urologist, participants received an intervention that consisted of pairing the participant with a former patient (to form a dyad) who had had a radical prostatectomy at least three years before the start of the study and had experienced similar treatment side effects as those experienced by participants. The dyads met eight times during an eight-week period, when the support partner was "able to vicariously relate successful coping and recovery strategies after prostatectomy" (p. 157)

Men in the control group received usual health care provided by their urologist

Outcomes

Quality of life (general healthrelated)

36-Item Rand Medical Outcomes Study Short Form (SF-36)

Quality of life (symptom-related)

20-Item UCLA Prostate Cancer Index

Depressive symptoms

15-Item Geriatric Depression Scale

Self-efficacy adjustment

38-Item Stanford Inventory of the Cancer Patient Adjustment (SIPCA)

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskTossing a coin (personal correspondence)
Allocation concealment (selection bias)Low riskEnvelopes used (personal correspondence)
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNot mentioned
Blinding of outcome assessment (detection bias)
All outcomes
Low riskData collectors blinded to group allocation (personal correspondence)
Incomplete outcome data (attrition bias)
All outcomes
Low riskMissing data reported
Selective reporting (reporting bias)Unclear riskOutcome data for one of the follow-up time points (four months) not reported
Other biasUnclear riskNone

Zhang 2006

Methods

Study design: randomised controlled trial

Follow-up: three months post-baseline

ParticipantsTwenty-nine men with prostate cancer were randomly assigned to an intervention group (n = 14) or a control group (n = 15). Inclusion criteria were patients at stages I to III prostate cancer; had received surgical treatment longer than six months before the start of the study and were experiencing urinary incontinence during the week before the start of the study. The mean age of the 27 men who completed the study was 61.5 years, and the overall ethnic composition of the sample was 71.9% white and 28.1% African American. The control group had fewer African American participants (13.3%) than the support group (42.9%). Participants were recruited from two major hospitals in Ohio, USA
Interventions

Both intervention and control group participants carried out pelvic floor exercises and received biofeedback via a sensor inserted in the participant’s rectum and connected to a computer monitor

The intervention group additionally received social support at six biweekly meetings held over a period of three months. Each meeting (four to five participants), facilitated by a ‘licensed’ health psychologist, lasted 90 to 120 minutes. During the meeting, the psychologist provided information about anatomy and physiology of urinary incontinence and moderated group discussions regarding experiences and problems with pelvic floor muscle exercises, personal struggles with functional difficulties in daily life, family communication about sexual dysfunction and other psychosocial issues such as stigma, loneliness and social relations

The control group carried out the pelvic floor exercises with biofeedback only

Outcomes

Quality of life (physical function)

10-Item Medical Outcomes Study Physical Functioning Measure (MOS-PF)

Symptom distress

McCorkle and Young Symptom Distress Scale

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskA table of random numbers was used (personal communication)
Allocation concealment (selection bias)High riskNo concealment method used (personal communication)
Blinding of participants and personnel (performance bias)
All outcomes
High riskMentions that one participant dropped out because of dissatisfaction with assignment to the control group. Blinding of personnel not stated
Blinding of outcome assessment (detection bias)
All outcomes
Low riskData collectors blind to group assignment (personal communication)
Incomplete outcome data (attrition bias)
All outcomes
Low riskDetails of missing data given
Selective reporting (reporting bias)High riskTwo prespecified outcomes (symptom distress and general emotional disturbance) not reported in the paper and not available from the author (personal correspondence)
Other biasHigh riskSignificantly more African Americans in the intervention group (42.9%) than in the control group (13.3%)

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
Badger 2011Two interventions and no control (usual care)
Beard 2011Does not meet inclusion criteria for psychosocial intervention
Burgio 2006Does not meet inclusion criteria for psychosocial intervention
Canada 2005No control group. The study compared the effects of the intervention between two groups: men alone and men attending with their partners
Davison 2007Did not measure outcomes relevant to this review
Downe-Wamboldt 2007Includes prostate cancer patients, but data not analysed separately
Goode 2011Does not meet inclusion criteria for psychosocial intervention
Green 2002Does not meet inclusion criteria for psychosocial intervention
Gyomber 2010Does not meet inclusion criteria for psychosocial intervention
Hack 2007Does not meet inclusion criteria for psychosocial intervention
Hellbom 1998Includes prostate cancer patients, but data not analysed separately
Ibfelt 2011Includes prostate cancer patients, but data not analysed separately
Johansson 2001Includes prostate cancer patients, but data not analysed separately
Johansson 2008Includes prostate cancer patients, but data not analysed separately
Johnson 1996Does not meet inclusion criteria for psychosocial intervention
Jones 2006Includes prostate cancer patients, but data not analysed separately
Kim 2002Does not meet inclusion criteria for psychosocial intervention
Kornblith 2006Includes prostate cancer patients, but data not analysed separately
Lee 2009Does not meet criteria for psychosocial intervention
Livingston 2010Includes prostate cancer patients, but data not analysed separately
Mohide 1996Does not meet inclusion criteria for psychosocial intervention
Morey 2009Includes prostate cancer patients, but data not analysed separately
Ramachandra 2009Includes prostate cancer patients, but data not analysed separately
Rottmann 2011Includes prostate cancer patients, but data not analysed separately
Scura 2004Very small sample (n = 17). Outcome data for control and intervention groups not reported separately
Siddons 2013Primary outcome was psychosexual adjustment, and extractable data (for meta-analysis; e.g. standard deviations) not reported
Simeit 2004Includes prostate cancer patients, but data not analysed separately
Stiegelis 2004Includes prostate cancer patients, but data not analysed separately
Templeton 2004Does not meet inclusion criteria for psychosocial intervention
Zakowski 2004Includes prostate cancer patients, but data not analysed separately

Ancillary