Health-related quality of life and psychosocial factors in patients with prostate cancer scheduled for radical prostatectomy or external radiation therapy
To assess whether baseline health-related quality of life (HRQOL) and psychosocial profiles differ in patients with prostate cancer scheduled for radical prostatectomy (RP) or external radiation therapy (ERT), as there is evidence that HRQOL is influenced by psychosocial factors (PFs), so that any variation at baseline should be considered when comparing the effect of therapy on HRQOL.
PATIENTS AND METHODS
Before receiving therapy, HRQOL and PFs were assessed in 65 patients scheduled for RP and in 73 scheduled for ERT. To measure HRQOL (generic and disease-specific) and PFs, an extended questionnaire was constructed, using validated and standardized instruments. Clinical data were collected from patients’ medical records. Comparisons adjusted for age and socio-economic status (SES) were analysed using Student's t-test and univariate analyses of variance and covariance.
Patients scheduled for ERT were 7.9 years older and had a lower SES (both P < 0.001), more often had stage T3 and T4 disease, had poorer histopathological differentiation and higher levels of prostate-specific antigen (all P < 0.01). They also reported a worse physical, role, cognitive and social function, more fatigue, more pain, a lower overall HRQOL and worse sexual function than patients scheduled for RP. There were no differences in urinary and bowel function, nor in the PFs assessed.
The baseline HRQOL profile of patients scheduled for RP is better than in those scheduled for ERT. These results are in line with those from the few other studies on this subject. Knowing the impact of RP and ERT on HRQOL should therefore be based mainly on longitudinal studies including baseline measures, the analyses of which should be adjusted for age and SES. In the present small study, baseline PFs did not differ between the treatments.
Radical prostatectomy (RP) and external radiation therapy (ERT) are widely accepted treatment options for men with locally (advanced) prostate cancer. In the region of the Comprehensive Cancer Center Amsterdam in the Netherlands, RP and ERT are offered to men in good condition with localized disease and a life-expectancy of > 10 years, while patients with locally advanced disease, comorbidity or a restricted life-expectancy are treated primarily with ERT. If patients are able to choose between these treatment options, the choice will be based on personal preferences, combined with the expected side-effects and the effect on the health-related quality of life (HRQOL) of both options. It is therefore essential to provide adequate information about the effect on HRQOL. A PubMed literature search on the keywords ‘prostate cancer’ and ‘quality of life’ up to 1 January 2002 found 674 articles, abstracts or citations in English. Of these, 61 reported on HRQOL measures in RP, ERT or both. Only a few of these studies using validated standardized questionnaires and comparing HRQOL in men treated with RP or ERT were prospective, while the others were all cross-sectional [1–3]. The comparison of the impact of two treatments on HRQOL in a cross-sectional study is suboptimal because it is not possible to take the pretreatment HRQOL scores into account; therefore, the results should be interpreted cautiously. This also applies to factors influencing the measured HRQOL, e.g. psychosocial factors (PFs). Therefore, we also addressed the role of PFs in the assessment of HRQOL; important PFs representing the psychological and social health status in oncological patients are stress, social support, emotional expression and coping strategies . Stress concerns the level and number of experienced stressful life events, e.g. the loss of a partner. Social support means the level and quality of support of relatives, friends and formal caregivers. Emotional expression represents the ability to acknowledge and express emotions, while coping style involves the efforts to appraise and manage a situation (e.g. distraction from it, active coping, denial, and cognitive adjustment). There is evidence in areas other than prostate cancer that PFs influence the experienced HRQOL [5–7]; these correlations are also found in patients with prostate cancer . Based on this evidence, the influence of PFs should be considered when measuring the HRQOL at baseline. To our knowledge, no studies have been published about the possible variation in baseline psychosocial profile of patients with prostate cancer receiving different treatments. It might therefore be important to know if there is a difference between the baseline psychosocial profile of patients selecting RP or ERT. As argued previously, a valid measure and analysis of baseline HRQOL is needed to evaluate the real effect of treatment on the HRQOL in a longitudinal design. Thus the objective of the present study was to investigate whether the HRQOL and psychosocial profile differ in patients primarily treated with RP or ERT.
PATIENTS AND METHODS
In all, 138 patients were entered in this prospective study, which was initiated in 1998; they were referred to the urologist because of urinary symptoms, elevated PSA, or a suspicious DRE; generally patients in our region are not screened for prostate cancer. Eligibility criteria included histologically confirmed carcinoma of the prostate, stages T1-4, G1-3, N0 and M0, and patients scheduled for RP or ERT. Patients receiving neoadjuvant or adjuvant hormonal treatment were excluded. HRQOL and PFs were assessed within the 2 weeks before treatment started (baseline) and again after 1 year. The results of the baseline assessment are presented here. As objective urological variables, the clinical tumour characteristics and PSA level were recorded. To measure HRQOL and PFs, a self-administered questionnaire was completed by the patients, consisting of the following.
(i) Sociodemographic characteristics, i.e. the patients’ age, level of education (six items, ranging from 1, elementary school, to 6, university), marital status (married or living together; never married nor lived together; divorced - living alone; divorced - living together with new partner; widower living alone; widower living together with new partner), ethnicity (open question), and employment status (level in job hierarchy; ranging from 1, part-time paid employment in small company, to 6, full-time self-employed in large company). The socio-economic status (SES) was measured as the total score of education and employment status, ranging from 2 to 12, with both indices correlating (r= 0.80, P < 0.001) .
(ii) Social desirability: this refers to the tendency that subjects are guided by what they consider socially desirable in responding to test questions, and do not give the same answers as they would if they tried to be as honest as possible. A standardized Dutch version of the social desirability scale of Crowne and Marlowe  was used, which contains 15 items on a 2-point scale (agree vs disagree). The total scale score is 0–15, with a higher score indicating a tendency to a more socially desirable response.
(iii) HRQOL; the European Organization of Research and Treatment of Cancer Quality of Life Questionnaire (EORTC QLQ) C30 consists of 30 items that list the functioning and symptoms of patients with cancer . Six multi-item function scales were scored: physical, role, emotional, cognitive and social function, and overall HRQOL. Furthermore, nine single-item symptom scales were scored, including fatigue, pain, dyspnoea and gastrointestinal problems. The scales were, according to the EORTC guidelines, linearly transformed such that all scales were 0–100, and in which a higher scale score represents a higher level of functioning. For the single-item scales a higher score indicates more symptoms or problems.
(iv) Voiding symptoms were measured using the IPSS , which consists of seven items scored on a 0–5 scale (0 = never and 5 = always), with a total score of 0–35, a higher score indicating more symptoms.
(v) Incontinence was assessed by severity, using a Dutch instrument  comprising four items with a 4-point scale and one item with a 5-point scale, giving a total score of 0–21, where a higher score indicates more disorders. The Cronbach α coefficient was 0.94.
(vi) Sexual functioning was measured by a newly developed scale consisting of items from the Sexual Behaviour Questionnaire  and items adopted from a questionnaire developed by the Dutch Social Sexuological Research Institute . The scale measures satisfaction with actual sexual functioning (four items), decrease of sexual activities (four items), ability to achieve and complaints about erections (five items), and the feeling of being rejected by the sexual partner (three items), mostly scored on a 5-point Likert scale. The subscales can be qualified as reliable (Cronbach α of 0.72–0.86 for the various scales). The score of the different subscales depends on the number of items included; a higher score indicates a worse sexual function.
(vii) Gastrointestinal symptoms; there was no Dutch questionnaire available for measuring such symptoms at the time the study was initiated and therefore a self-constructed questionnaire was used, based on the Dutch version of the disease-specific EORTC colorectal module (CR38), which was being developed at that stage and has since been validated. The questionnaire used in the present study contains 12 questions about, e.g. frequency, tenesmus, blood, diarrhoea, pain and urge, scored on a 4-point Likert scale (1, no complaints, to 4, very many complaints). The total score is 12–48 and a higher score indicates more disorders (Cronbach α, 0.75).
(viii) The coping style was measured using the standardized Utrecht Coping List (UCL) . The dimensions are; active coping (seven items), palliation (eight), distraction (seven), seeking social support (five), denial (seven), expression of emotions (three), and cognitive adjustment (five), measured on a 4-point Likert scale. The total score varies for the different dimensions, at 3–12 and 7–28, with a higher score meaning a higher level of coping style.
(ix) To measure the psychological distress, the Profile of Mood States (POMS) was used , in the form of a Dutch standardized validated shortened version, including the following dimensions; depression, anger, fatigue, vigour, tension, and a total mood disturbance score (items on a 5-point scale), with a higher score indicating more distress, except for vigour.
(x) Impact of having prostate cancer: a Dutch version of the Impact of Event Scale (IES) was used to measure the emotional impact of having prostate cancer . The subscale measuring re-experience (six items) and denial (seven items) are scored on a 4-point Likert scale. The total score is 0–35, with a higher score indicating more disorders.
(xi) ‘Not expressing emotions’ was measured using a Dutch validated questionnaire , with measured dimensions of rationality, ‘emotions-in’, ‘emotions-out’, and emotional control. The total score is 4–64, with a higher score indicating more active expression of emotions.
(xii) Social support was measured using the Dutch version of the shortened Social Support Questionnaire (SSQ) , which consists of six items on a 4-point Likert scale, with a total score of 4–24 and a higher score indicating a higher level of social support.
(xiii) Life events were assessed using the Dutch version of the Life Experience Survey (LES) , which concerns 40 potential life events in the last 12 months, with the total number of negative life events used as a total score.
The items of the EORTC QLQ C30 are highly inter-correlated and therefore if ≤ 20% of items are missing, it is possible to substitute values for missing data with values derived from the personal (sub)scale mean. The single-item scales were considered missing if they were not completed. For all remaining questionnaires the mean-procedure within the statistical package was applied with a limit of 20% of the items.
Sociodemographic characteristics and objective clinical variables were analysed using descriptive statistics, chi-squared tests, Student's t-tests, and by nonparametric tests (Mann–Whitney). HRQOL and psychosocial measures were evaluated using univariate anova and covariance. All analyses of HRQOL and PFs were adjusted for age and SES to eliminate their confounding effects. The results from patients scheduled for RP were compared with those scheduled for ERT.
The sociodemographic characteristics and objective clinical variables are shown in Table 1. There were significant differences between patients treated with RP and ERT for age and SES; patients receiving ERT were older (7.9 years) and had a lower SES (both P < 0.001). Those having RP more often had an active work role than did the ERT patients, and this could be explained by the age differences. There were no differences in country of origin and civic status. Thus age and SES should be considered as confounders when comparing the samples. Furthermore, the patients treated with ERT more often had tumour stages T3 and T4, a poorer histopathological differentiation and higher levels of PSA than those treated with RP (Table 1).
Table 1. Baseline patient characteristics (sociodemographic and clinical) and EORTC QLQ C30 (age- and SES-adjusted scores) with treatment
|Age, years|| || ||< 0.001|
|Mean (sd)||62.1 (5.8)||70.0 (6.2)|| |
|Median (range)||62.0 (49–72)||71.0 (55–82)|| |
|Mean (sd) SES|| 7.3 (2.7)|| 6.1 (3.0)||< 0.00|
|Work status, %|
|Actively working||28.6|| 5.3||0.01|
|Not actively working||71.4||94.7|| |
|Country of origin, %|
|Not the Netherlands|| 6.3|| 5.3|| |
|Marital relationship and/or living together, %|
|No relation||11.1|| 9.3|| |
|Tumour stage, %|| || ||0.01|
|Tx|| 0|| 1.8|| |
|T1||36.5|| 9.1|| |
|T3|| 9.5||60.0|| |
|T4|| 0|| 3.6|| |
|Histological grade, %|| || ||0.01|
|G3|| 7.9||18.2|| |
|PSA, ng/mL|| || ||0.01|
|Median (range)|| 9.9 (0.4–48.1)||12.7 (0.9–72.0)|| |
| EORTC QLQ C30 |
|Physical function||91 (17)||86 (20)||0.08|
|Role function||98 (11)||92 (20)||0.05|
|Emotional function||74 (19)||80 (21)|| |
|Cognitive function||92 (15)||84 (23)||0.02|
|Social function||89 (17)||85 (24)||0.06|
|Overall HRQOL||83 (15)||73 (17)||0.01|
|Fatigue||16 (23)||24 (23)||0.04|
|Nausea/vomiting|| 2 (5)|| 2 (6)|| |
|Pain|| 7 (15)||11 (19)||0.05|
|Dyspnoea|| 7 (15)||11 (20)|| |
|Insomnia||14 (13)||12 (15)|| |
|Appetite loss|| 4 (18)|| 5 (12)|| |
|Constipation|| 3 (10)|| 6 (28)|| |
|Diarrhoea|| 6 (14)|| 5 (16)|| |
|Financial difficulties|| 5 (15)|| 6 (15)|| |
The results of the general HRQOL assessment are also listed in Table 1. The patients scheduled for RP reported a far better physical, role, cognitive and social function than patients treated with ERT. Moreover, they scored higher for overall HRQOL and had less fatigue and less pain.
Table 2 presents the prostate-specific HRQOL scores; there were no differences between treatments for urinary and bowel function. Before treatment started patients scheduled for RP reported significantly less sexual dysfunction, with more nocturnal erections, more ability to achieve an erection, and greater satisfaction with their sexual functioning.
Table 2. Baseline patient characteristics of urinary, bowel and sexual function by treatment (age- and SES-adjusted scores)
|Mean (sd) IPSS||10.2 (7.0)||11.1 (7.6)|| |
|Frequency of incontinence*, %|
|Once a week||55.6||58.3|| |
|Several times a week||11.1||25.0|| |
|Once a day||16.7||16.7|| |
|Several times a day||11.1|| 0|| |
|Continuous loss of urine|| 5.5|| 0|| |
|Use of pads*, %|
|Sometimes||11.1|| 8.3|| |
|Usually|| 5.6|| 8.3|| |
|Always|| 0|| 0|| |
|Restriction in daily activities*, %|
|Some|| 5.6|| 8.3|| |
|(Very) much|| 0|| 0|| |
|What do you think about your incontinence*, %|
|I don’t care||27.8||41.7|| |
|Difficult and uncomfortable||55.6||58.3|| |
|I feel myself dirty||16.7|| 0|| |
|Mean (sd) gastrointestinal function score||14.3 (7.3)||14.3 (6.9)|| |
| Sexual function (during the last month) |
|Sexually active, %|
|Interest/sexual need|| || ||0.05|
|One till three times||11.3||14.8|| |
|Once a week||46.7||48.1|| |
|Several times a week||24.3||11.2|| |
|Once a day|| 0|| 0|| |
|Frequency of nocturnal erections, %|| || ||0.02|
|1–3 times||29.0||27.3|| |
|Once a day|| 8.1|| 1.9|| |
|Ability to achieve an erection, %|| || ||0.01|
|1–3 times||27.4||23.6|| |
|Once a day||17.7|| 3.6|| |
|Satisfaction with actual sexual function, %|| || ||0.04|
|Sexual dissatisfaction||9.4 (3.6)||10.9 (3.4)||0.01|
|Erectile dysfunction||18.9 (4.5)||21.9 (4.3)||0.02|
|Total sexual dysfunction||35.4 (8.7)||40.6 (8.1)||0.01|
The results of the psychosocial assessments are listed in Table 3. There were no differences between the groups in the psychosocial characteristics of the patients. Analyses on the subdimensions of the POMS showed that ERT patients reported more feelings of fatigue and anger.
Table 3. Baseline patient characteristics for the PFs by treatment (age- and SES-adjusted scores)
| Coping style (UCL) |
|Active coping||18.8 (4.1)||17.1 (4.2)|| |
|Palliation||17.6 (4.3)||16.4 (4.3)|| |
|Avoiding||12.6 (2.9)||13.6 (3.6)|| |
|Social support||12.9 (4.1)||11.1 (3.6)|| |
|Denial||10.4 (2.8)||10.7 (3.5)|| |
|Expression of emotion|| 4.9 (1.4)|| 4.9 (1.7)|| |
|Cognitive adjustment||12.4 (2.1)||12.1 (3.8)|| |
| Mood disturbance (POMS) |
|Depression||12.1 (4.2)||11.4 (5.1)|| |
|Anger|| 8.7 (2.6)|| 9.5 (3.8)||0.01|
|Fatigue|| 8.6 (4.5)|| 9.4 (4.1)||0.02|
|Tension||12.5 (5.0)||10.8 (4.8)|| |
|Vigour||16.4 (4.6)||15.9 (3.8)|| |
|Total Mood disturbance||55.2 (146)||55.5 (16.6)|| |
| Impact of prostate cancer as a life event (IES) |
|Re-experiencing||12.7 (7.5)|| 9.9 (7.8)|| |
|Avoiding||11.8 (7.6)||11.9 (7.9)|| |
|Shock handling||27.1 (14.4)||23.9 (15.7)|| |
| Not expressing emotions |
|Rational orientation||42.3 (6.3)||44.3 (5.9)|| |
|Expression in||12.1 (3.3)||13.0 (3.5)|| |
|Expression out||13.2 (2.6)||11.6 (2.8)|| |
|Emotional control||18.2 (1.8)||18.1 (1.8)|| |
|Social support (SSQ)||20.3 (3.2)||20.2 (3.1)|| |
|Life events (LES)|| 3.7 (2.8)|| 3.8 (3.1)|| |
The present study focused on baseline HRQOL and psychosocial characteristics of patients with prostate cancer who were scheduled for RP or ERT. There were significant differences in baseline characteristics between the groups, with ERT patients being ≈ 8 years older, of a lower SES, a more advanced and worse tumour stage, and a higher PSA level. This was to be expected because, according to the guidelines of the study region, RP is offered primarily to otherwise healthy patients with localized disease and a life-expectancy of ≥ 10 years (e.g. generally younger patients). Why patients scheduled for RP had a significantly higher SES is as yet unclear; possibly patients with a higher SES are better informed about the possibility of developing prostate cancer through the media and literature, resulting in earlier check-ups. This would explain why these patients are younger and that the disease was diagnosed at an earlier stage. Moreover, Saigal et al. found that patients who prefer RP appear to attach a higher value to quantity of life than do older and less fit patients, who attach a higher value to quality of life. There was a significant relationship between age and SES, and with measured HRQOL; older patients with a lower SES assessed the various domains of HRQOL as worse. This is in line with findings of other studies [1,23]. Nevertheless, it is important because it highlights the results of other HRQOL studies that did not correct for age and SES as confounders.
However, after adjusting for age and SES there were also differences in baseline HRQOL; patients scheduled for RP had less pain and fatigue, and had, among other things, better physical function, and they assessed their overall HRQOL as better than did patients treated with ERT. In their screening study, Madalinska et al. also established that patients treated with RP appeared to be a relatively fit group and even tended to score above the general population norm. However, much of what is known about the effect of RP or ERT on HRQOL is based on cross-sectional studies [24–26]. In these studies the HRQOL is measured some time after treatment and from these data, the effects of the various treatments on HRQOL are analysed and compared. Because the HRQOL profile before treatment may differ in groups of patients, such studies can only provide an overall impression of the HRQOL at that time, and will not estimate the exact effect on HRQOL. If valid information is required then reliable knowledge about the effect of treatment on HRQOL is desirable. Therefore, large methodologically sound prospective studies, including pretreatment measures, are needed to confirm the present view.
The role of PFs should be considered; there is a general consensus that HRQOL consists of various independent domains [27,28], e.g. physical, functional and social health, but psychological health is important and is represented by several PFs. Recently there have been studies of the relationship between these PFs and the general and disease-specific HRQOL. The assessed HRQOL is influenced by several PFs [5–8] and this might change the investigator's perception of HRQOL as assessed to date. In oncological research, PFs are not an integral part of the assessment of HRQOL; as with age and SES, the PFs should be corrected for to measure the real effect of a treatment on HRQOL. This applies to unrandomized studies in particular; in randomized studies it might be assumed that various factors that can act as confounders are equally distributed among groups, but this does not hold for unrandomized studies. None of the reports describing the impact of RP and ERT on the HRQOL are of randomized studies. In this category of patients, PFs are therefore important. Moreover, it is very likely that in two groups with not only different baseline HRQOL, but also different age and SES, the psychosocial profile will also differ. In the present study the HRQOL and the psychosocial profile were examined, but for the PFs, age- and SES-adjusted analyses showed no significant differences at baseline between patients scheduled for RP or ERT. Only anger and fatigue, as signs of mood disturbance, were reported more frequently by ERT patients. Coping styles, expression of emotions, social support and the number of experienced life events were equal in both groups. However, the present study is one of very few addressing this subject, and there were relatively few patients. Another limitation of the present study is that for assessing gastrointestinal symptoms a self-constructed but not well validated questionnaire was used.
In conclusion, age and SES are important confounders in measuring HRQOL in patients with prostate cancer scheduled for radical treatment. The HRQOL scores before treatment, as measured with the EORTC QLQ C30, differed significantly between patients scheduled for RP or ERT. Assessing the effect of RP and ERT on HRQOL should therefore be based mainly on longitudinal studies including pretreatment measures, the analyses of which should be adjusted for age and SES. There is evidence that the general and disease-specific HRQOL are also influenced by PFs. In the present study, baseline PFs did not differ between the treatment groups, but whether, as for age and SES, adjustment for PFs is also necessary must be explored in larger and statistically more powerful studies.
health-related quality of life
Utrecht Coping List
Profile Of the Mood Scale
Impact of Event Scale
Social Support Questionnaire
- EORTC QLQ
European Organization of Research and Treatment of Cancer Quality of Life Questionnaire.