SEARCH

SEARCH BY CITATION

Keywords:

  • cancer;
  • oncology;
  • distress;
  • screening;
  • psychosocial needs

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Objectives: We evaluated screening for distress in terms of its ability to uncover unmet need for psychosocial services in cancer patients. Correlates of distress, need for services and met and unmet need for services were investigated.

Methods: Immediately after cancer treatment (T1) and 2 months later (T2), 302 patients completed the Hopkins Symptom Checklist-25 (HSCL-25) and a single question assessing the need for services. All distressed patients (HSCL-25⩾39) and non-distressed patients endorsing a need for services were then called (n = 99) to assess their need.

Results: Thirty-seven percent (T1) and 31% (T2) of patients were distressed and 31% (T1) and 18% (T2) expressed the need for services. Both time points showed higher distress in younger patients and females and lower distress in prostate cancer and patients treated by radiotherapy only. Less need for services was found in prostate cancer (T1), greater need was related to being single (T1) and younger (T2). Distress and need for services were positively related (p<0.001). The HSCL-25 showed modest sensitivity (T1: 0.59, T2: 0.65) and specificity (T1: 0.75, T2: 0.78) as an indicator of need for services. Interviews at T2 revealed that 51% of distressed patients needed no psychosocial services and 25% were already receiving services. At T2, regardless of distress level, 10% of all screened patients reported an unmet need for psychosocial services.

Conclusions: Depending on the clinical context, screening might be more efficient if it assessed the unmet need for services rather than distress. More attention should be concentrated on directing patients with meetable unmet needs to available services. Copyright © 2011 John Wiley & Sons, Ltd.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Diagnosis and treatment of cancer have the potential to generate emotional distress including fear, uncertainty and depressive symptoms. The prevalence of emotional distress in cancer patients ranges from 30 to 45%, depending on the point of assessment 1–4. Distress decreases after diagnosis and many cancer patients ultimately recover well 5–9. Yet, improvement in distress is not necessarily linear, and some specific illness-related events, such as end of treatment, can increase the level of distress 8, 9. Patients may lose the support and the sense of security associated with treatment and their regular contact with healthcare professionals. They may have to deal with side effects of treatment and may have difficulties reintegrating into former family and social roles. Elevated distress has been associated with a number of negative outcomes, including poorer adherence to treatment recommendations and lower satisfaction with care and quality of life 10 and so routine screening for distress is widely advocated in cancer care 11, 12. Yet, often overlooked is that uncovering distress is not an end in itself. Rather than just categorizing whether patients are distressed, screening is seen as a means toward the goal of identifying patients with an unmet need for psychosocial services. However, there are few evaluations of whether screening for distress actually meets this objective with regard to receipt of services or leads to improved patient outcomes 13–16. For example, in an intensive screening program of Carlson et al. 16 with additional offers of psychosocial services, less than one-third of patients requested and received a referral and patients with referrals showed less improvement in distress than patients who did not receive a referral.

Importantly, much of the screening literature consists of examinations of whether screening is efficient with respect to identifying psychiatric disorder as assessed by gold standard psychiatric interviews. Such studies implicitly assume that screening positive for distress represents a need for psychosocial services. This assumption has not been adequately tested; some studies in cancer patients find a significant relationship between elevated distress and a need for services 17–20, yet others do not 2, 21–24. Graves et al. 17 found that among 333 lung cancer patients, only 30% of the distressed patients wanted help with their problems. In a cross-sectional study in 277 screened cancer patients, Tuinman et al. 19 found that only 14% of patients who were distressed definitely wanted and 29% possibly wanted to be referred for further help. It is noteworthy that 5% of patients low in distress also definitely wanted services and 13% considered a referral. Sollner et al. 18, 21 also found that a substantial proportion of patients without elevated distress reported a need for services, from 17% (skin cancer) to 45% (breast cancer).

The efficiency of screening for distress can be reconceptualized in terms of whether screening uncovers unmet needs, which can be met by appropriate psychosocial treatment (meetable unmet need) 25, 26. Reduced likelihood that heightened distress represents a meetable unmet need could come about as a result of distressed patients not wanting services, already receiving services or having a need that cannot be met by available services. Thus, one alternative to screening for distress becomes simple to ask patients directly whether they have an unmet need for services, as was also proposed by others 27–29. This becomes a more attractive alternative if heightened distress were to prove a poor criterion for ruling in a need for psychosocial services.

The overall aim of the present study is to evaluate screening for distress in terms of its ability to uncover unmet need for psychosocial services in cancer patients. Specifically, the study examined: (1) the rate and course of distress and need for services of adult cancer patients in the months following the end of medical treatment; (2) the types of patients being identified when screening on the level of distress or need for services; (3) the predictive value of distress level as an indicator of need for services and (4) the reasons why patients did not want services and the met and unmet psychosocial need of patients in need for services.

Method

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Participants

Participants were recruited from three cancer centers in the Netherlands: University Medical Center Groningen (UMCG, Groningen), Radiotherapy Institute Friesland (RIF, Leeuwarden) and Martini Hospital (MH, Groningen). Eligibility criteria were: ⩾18 years of age, having completed curative cancer treatment, a positive life expectancy of at least 1 year, ability to speak and write Dutch and willingness to complete the screening questionnaire. Patients were being screened as part of recruitment for an intervention study.

Procedure

Patients completed the HSCL-25 as part of routine clinical care (except for the RIF, where the patients had first to consent). Patients were screened from December 2008 through March 2010 and were requested by either a cancer nurse (UMCG/MH) or radiation oncologist (RIF) to complete the screening questionnaire directly after completion of medical treatment (T1). The same screening questionnaire was sent to the patients 2 months later by post or by email, as preferred by the patient (T2). If questionnaires were not returned, patients were reminded by phone. The screening procedure served as a basis for recruitment for an intervention study; patients who had elevated levels of distress and a need for services were offered an intervention. However, during the screening process, patients were not yet informed about this intervention, so the results for the present study were not biased by the intervention study.

After the first screening questionnaire, all patients were sent a short letter indicating whether they had screened positive for distress and provided with a phone number that could be called for assistance. After the second screening, the researcher and a research assistant contacted three groups of patients by phone. First, patients with high distress and a need for services were interviewed about their unmet needs. They were offered the possibility to participate in the intervention study, or, if ineligible, referred to other services. Second, patients with high distress and no need for services were asked about their motives for not wanting any services or if they already received services. Third, patients with low distress, but with a need for services, were asked about their unmet need and, if wanted, referred to available services. On the basis of the second screening, 83 distressed patients (with or without need for services) and 16 non-distressed patients who expressed a need for services were called. Eleven patients who met one of these criteria were not reached.

Measures

HSCL-25

The HSCL-25 30 is a self-report measure of psychological distress. It includes the anxiety and depressive subscales of the HSCL-58, originally developed by Derogatis et al. 31, 32. The HSCL-25 is closely related to the Brief Symptom Inventory-18, which is a widely used screening instrument among cancer patients 4, 33, 34. The instructions inquired about the intensity of symptoms in the previous week. Answers are scored on a scale from 1 (never) to 4 (always). A higher score indicates greater distress and the total score is the sum of items. Separate studies have indicated scores of either ⩾39 or ⩾44 as the optimal cutpoint for identifying ‘cases’, as measured by psychiatric interview 35, 36. The present study adopted the lower cutpoint ⩾39 to allow the opportunity to invite moderately distressed patients to participate in the intervention study. Secondary analyses were reported for the higher (⩾44) cutpoint. Internal consistency for the present study was excellent (Cronbach's α = 0.92), suggesting psychological distress was being assessed, rather than specifically anxiety or depression.

Demographic and clinical characteristics

Questions assessed the patients' age, sex, marital status, educational level, cancer diagnosis, disease status (first occurrence, recurrence, second tumor), time of diagnosis and medical treatment.

Need for services

One question inquired about the patients' need for services: ‘Would you like to talk to a care provider about your situation?’ answered with a Yes, Maybe, or No response.

Interviews

Telephone interviews evaluated more in depth patients' need for psychosocial services; whether this need was already met or still unmet and whether this unmet need was meetable. Patients not wanting services were asked about their motives for not wanting any services.

Data analysis

Analyses were performed using SPSS for Windows, version 16.0. Standard descriptive statistics were generated to characterize the demographic, clinical and distress variables in the sample. T-tests, chi-square analysis, ANOVA and correlations were used to examine the course of distress and need for services, possible correlates of distress and need for services and the relationship between distress and need for services. To establish the performance of the HSCL-25, we used a web-based calculator to define sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) 37.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Research participants

Figure 1 displays patient flow. Three hundred eighty-eight first screening questionnaires were returned (i.e. 77% of all patients approached at UMCG and Martini Hospital). Subsequently, 302 second screening questionnaires were returned (i.e. 78% of all patients approached for the second questionnaire). The characteristics of the 302 patients returning both questionnaires are summarized in Table 1.

thumbnail image

Figure 1. Flowchart of the study

1Due to a difference in collecting the questionnaires: the RIF did not give questionnaires numbers beforehand

Download figure to PowerPoint

Table 1. Descriptives
VariableNumber of patientsValid percentageMean±SD (range)
  • RT, radiotherapy; CT, chemotherapy; HT, hormone therapy.

  • a

    aTestis, bladder and penis cancer.

Age  61.0±11.7 (29–87)
Gender
 Men12641.9 
 Women17558.1 
Marital status
 Married/living together25684.8 
 Single/widowed4615.2 
Educational level
 Low11437.7 
 Average13143.4 
 High5718.9 
Cancer diagnosis
 Breast12742.1 
 Prostate6019.9 
 Gastrointestinal4414.6 
 Lung227.3 
 Gynecologic82.6 
 Other urologica82.6 
 Hematologic82.6 
 Other258.3 
Disease status
 First occurrence25183.1 
 Recurrence278.9 
 Second tumor247.9 
Treatment received
 RT only5118.7 
 Surgery and RT8230.0 
 Surgery, RT and HT248.8 
 Surgery and CT269.5 
 Surgery, RT and CT5018.3 
 Surgery, RT, CT and HT2910.6 
 Other114.0 
Time since diagnosis  (at T1) (months)  5.4±3.3 (0–25)

Rates and course of distress, and need for services

At T1, the mean distress level (±SD) was 36.9 (±9.2), with 37% being highly distressed (i.e. scoring above the cutoff ⩾39). At T2, the mean distress level (±SD) was 35.3 (±8.7), with 31% of the patients being highly distressed. The decrease in mean scores of distress over time was significant (t(300) = 4.41, p<0.001), and the decrease in percentage of patients being distressed over time was not significant. At T1, 31% of the patients reported a need for services (5% yes and 26% maybe), at T2 18% of the patients (8% yes and 10% maybe) (χ2 = 1.10, p<0.001). In secondary analysis, using a cutpoint ⩾44 of the HSCL-25, 21% of patients were highly distressed at T1 and 17% of patients at T2 (χ2 = 84.8, p<0.05).

Patients who were identified with distress

At T1 and T2, younger patients (t(298) = 3.4, p = 0.001; t(294) = 4.3, p<0.001) and women showed more distress than men (t(298) = −2.8, p<0.01; t(278) = −2.7, p<0.01). ANOVA with post hoc test showed that patients with prostate cancer were less distressed than patients with other types of cancer at T1 (F(7,293) = 4.6, p<0.001) and less distressed than patients with other urological cancers at T2 (F(7,292) = 2.9, p<0.01). Patients receiving only radiotherapy were less distressed than patients receiving all other medical treatments at T1 (F(6,266) = 4.9, p<0.001) and less distressed than patients receiving a combination of surgery, radiotherapy, chemotherapy and hormone therapy at T2 (F(6,265) = 3.9, p<0.001) (Table 2). However, there should be caution in interpreting these results for treatment variables. They are confounded with cancer site. Educational level, marital status and disease status were not related to distress at both time points.

Table 2. Demographic and clinical variables, distress and need for services
VariableDistress T1 (mean±SD)Distress T2 (mean±SD)% Need for services (T1)% Need for services (T2)
  • p<0.05 significance level for differences between rows.

  • *

    *p<0.05 significance level for differences in column.

  • a

    aBreast, lung, urological and other cancers more distressed than prostate.

  • b

    bProstate less need for services than other diagnoses.

  • c

    cUrological more distressed than prostate.

  • d

    dRadiotherapy only less distressed than other treatments (except ‘other’).

  • e

    eRadiotherapy only less distressed than surgery, RT, CT and HT.

Total distress36.9±9.235.3±8.73118
Age (years)
 <6538.4±10.237.1±9.63437
 ⩾6534.9±7.5*33.0±6.5*2712*
Gender
 Men35.1±8.333.7±8.23019
 Women38.1±9.7*36.4±8.8*3118
Marital status
 Partner36.5±8.934.9±8.22919
 No partner39.2±10.537.6±10.643*20
Educational level
 Low37.0±8.835.1±8.22618
 Average36.5±9.135.1±8.73319
 High37.4±10.536.1±9.63719
Cancer diagnosis
 Breast37.3±9.235.6±8.02717
 Prostate31.7±5.2a31.7±6.219b22
 Gastrointestinal37.4±9.535.8±10.14120
 Lung40.0±9.536.8±8.53614
 Gynecologic40.9±13.940.4±12.26338
 Other urological43.0±7.841.9±12.6c6350
 Hematological40.9±13.835.9±9.1250
 Other38.9±9.636.3±9.74317
Disease status
 First36.8±9.235.3±8.63019
 Recurrence37.6±10.635.2±9.93019
 Second tumor37.3±8.934.8±8.42113
Treatment received
 RT only31.7±5.8d31.0±5.6e1612
 S and RT36.8±10.135.5±9.03218
 S, RT and HT34.0±5.034.5±7.32521
 S and CT39.6±9.536.0±10.92719
 S, RT and CT37.1±10.035.6±9.14220
 S, RT, CT and HT41.1±9.040.0±8.13531
 Other38.6±9.134.1±6.6279

Patients who were identified in need for services

At T1, patients being single were more in need for services (χ2 = 6.6, p<0.05). The least need for services was found in patients with prostate cancer (19%) and the highest need for services in patients with gynecological and other urological cancers (63%) (χ2 = 24.3, p<0.05). At T2, younger patients were more in need for services than older patients (χ2 = 7.8, p<0.05). Gender, disease status, type of treatment, time since diagnosis and educational level were not related to need for services at both time points (Table 2).

Predictive value of distress level

Table 3 shows the association between distress and need for services over time. The three groups of patients differed significantly in need for services, at both T1 (F(2,296) = 45.8. p<0.001) and T2 (F(2,298) = 57.0, p<0.001). Patients with no need reported the lowest level of distress and patients with definitely a need the highest level of distress.

Table 3. Relationship between patients' distress and need for services
 After finishing medical treatment (T1)Two months later (T2)
 Distress level (mean±SD)n (%)Distress level (mean±SD)n (%)
  • *

    *p<0.001 significant differences in distress level between three groups of need for services (ANOVA).

Need for services
 Yes52.4±14.0*14 (5%)48.9±10.9*24 (8%)
 Maybe41.0±8.779 (26%)40.4±7.331 (10%)
 No34.3±7.3206 (69%)33.3±7.0246 (82%)

Similarly, distressed patients were more in need for services (response of yes or maybe) than non-distressed patients, both at T1 (52 versus 20%) and T2 (40 versus 9%). Nonetheless, 48% (T1) and 60% (T2) of the distressed patients did not have a need for services and 20% (T1) and 9% (T2) of non-distressed patients had a need for services.

Table 4 shows the performance of the HSCL-25 in identifying a need for services using both cutpoints (39 and 44) chosen a priori. Using the cutpoint of 39, the HSCL-25 showed modest sensitivity and specificity as an indicator for need for service. Although specificity and NPV were both good at the higher cutpoint, the HSCL-25 was inefficient in indicating a need for services. Taking these percentages into consideration, distress is an inefficient indicator for need for psychosocial services, with most distressed patients not needing services, and some patients who are not distressed nonetheless needing services.

Table 4. Both cutpoints of the HSCL-25 as an indicator of need for services
  Sensitivity95% CISpecificity95% CIPPVNPV
T1HSCL⩾390.590.48–0.690.750.69–0.810.350.65
T1HSCL⩾440.440.34–0.550.890.84–0.930.210.79
T2HSCL⩾390.650.51–0.770.780.72–0.830.300.70
T2HSCL⩾440.450.32–0.590.900.85–0.930.170.83

Patients' reasons for not wanting services and patients met and unmet psychosocial need

Distressed patients with no need for services

Of the 83 distressed patients who were interviewed at T2, 42 patients (51%) reported during the interview no need for services. Reasons for not wanting services were ‘I receive adequate social support from family and friends’ (n = 15), ‘I wish to improve by myself’(n = 6), ‘I don't want to talk about my problems’ (n = 12), ‘I am physically impaired’ (n = 3), ‘I want to wait, maybe I want services later in time’ (n = 4) and ‘I don't think talking will help’ (n = 2). Sixty-five percent of these patients were women, 54% were younger than 65 years, 56% were low educated and they were diagnosed with various types of cancer.

Distressed patients receiving services

Twenty-one of 83 distressed patients (25%) mentioned that they already received psychosocial services; they visited a psychologist/psychiatrist (n = 18), a psycho-oncological rehabilitation center (n = 1) or received other services (from creative therapy/a nurse) (n = 2). Most of these patients were women (91%), younger than 65 years (86%), well-educated (moderate/high = 76%) and diagnosed with breast cancer (81%).

Distressed patients in need for services

Twenty of 83 distressed patients (24%) had an unmet need for psychosocial services. They were asked to participate in an intervention study (n = 11) or, if not eligible or not wanting to take part, referred to other psychosocial services (n = 9). Of these patients, 62% were women, 76% were younger than 65 years, 86% were well-educated (moderate/high) and they were diagnosed with various types of cancer.

Non-distressed patients in need for services

Of the 16 non-distressed patients with need for services, 2 were satisfied with a short talk with the researcher (a psychologist), 1 wished for a referral for a psychologist and 7 patients wished to consult their medical specialist about their uncertainty about physical symptoms. In addition, one patient had already visited a psychologist and five did not have a meetable psychosocial need (e.g. back pain). Thus, 10 (63% of the 16) non-distressed patients indicated an unmet need for services. Most of these 10 patients were men (72%), younger than 65 years (60%), low educated (60%) and diagnosed with prostate cancer (50%).

In conclusion, 30 cancer patients (i.e. 20 distressed and 10 non-distressed patients) reported an unmet need for psychosocial services, which is 10% of the total sample of 302 screened patients.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This study is novel in evaluating screening for distress in terms of its ability to uncover unmet need for psychosocial services in cancer patients. Results showed that distress level and need for services were positively related. However, about half of the distressed patients reported no need and about 10–20% of non-distressed patients nonetheless reported a need for services, consistent with past research 2, 18, 19, 22. Interviews with patients revealed that 10% of all cancer patients had an unmet psychosocial need. For the purpose of identifying need for services, our screening for distress showed a modest sensitivity and specificity and a low PPV. When using the higher cutpoint for the HSCL-25, we found an even lower sensitivity and PPV. This seeming poorer performance of a more stringent cutpoint to identify patients in need for services was due to a smaller proportion of distressed and a greater proportion of non-distressed patients in need for services.

Our results concerning the prevalence of distress and proportion of patients in need for services are both in line with other studies 8, 18, 19, 24. Also consistent with other research, we found that younger patients reported a greater need for services and prostate cancer patients less need 16, 18, 21, 22. However, we found no gender difference, while others found that men were less in need for psychosocial services 22, 23. It is noteworthy that the interview data showed that the majority of distressed patients already receiving services were female and more than expected patients with an unmet psychosocial need were men. These results point out possible gender differences in seeking services and in the type of services wanted.

Overall, relying exclusively on screening for distress to identify patients in need for services is less efficient than has generally been assumed. Presumably, some interaction with clinic personnel would be required to establish that over half of distressed patients did not want services or already were receiving services. Faced with the necessity of a choice because of already strapped clinical resources in some settings, it might be advisable simply to ask directly about the need for services and forgo screening for distress, as was also proposed by Garssen et al. 27, 28.

Furthermore, we are arguing for evaluating screening in terms of whether it succeeds in linking patients with services that they would otherwise not receive, not just in terms of some conventional gold standard. We recommend that future studies evaluating the performance of screening measures routinely incorporate assessment of the extent to which a score above a particular cutpoint identifies patients interested in available services that they are currently not receiving.

Limitations

The following issues should be taken into account regarding our findings. First, as this study focused on the end of medical treatment, no information was available about patients' distress level and need for services at the time of diagnosis, during medical treatment or long-term follow-up. This type of information is important for understanding the course of distress and the possibly changing need for services during the cancer trajectory and should be addressed in future research. Second, our simple means of assessing interest in psychosocial services yielded provocative findings concerning the potential limitations of exclusively relying on screening for distress to define need for services. Future research should design and validate more refined instruments to measure psychosocial needs of cancer patients, as existing questionnaires (see, for example, 38–40) leave the concept still unclear, partly because some regard the experience and report of distress as a direct indicator of a need for service.

For a start, more detailed information might be obtained with a (semi-)structured questionnaire or interview with patients inquiring: (1) Do you have physical or emotional problems related to your cancer or medical treatment at this moment? (2) Would you like to receive services for (one of) these problems? (Answer options: no/needs are already met, maybe, yes) (3) What kind of services would you like (e.g. information about disease, treatment and side-effects; physical services regarding symptom management and physical limitations; psychosocial services regarding coping and emotional adjustment, medical treatment, or a combination of these)? When discussing patients' unmet needs, it should be checked to what extent these needs are meetable by appropriate services. As practice shows that some of the needs uncovered in cancer care that are related to distress may be unmeetable in that context, i.e. inadequate housing, financial problems or functional limitations that persist despite adequate medical treatment, we need to consider that psychosocial services attached to cancer care cannot be expected to resolve all the sources of distress and misery that patients and family members face.

Clinical implications

Our results raise the issue whether screening for distress is the optimal means of identifying unmet needs for psychosocial services and suggest direct inquiry about interest in services might be more efficient. Yet, while relevant to the question of whether to screen for distress, our results do not resolve the question. Rather, ‘to screen for distress, or not to screen?’ is a complex issue to be settled with consideration of the unencumbered clinical resources in a setting and the existing flow of patients into services, in the absence of screening. Without additional resources, introducing routine screening risks the negative effect of withdrawing resources from existing clinical activities 41. Yet, screening may be unnecessary in an exceptionally well-functioning clinical cancer care setting where patients are already routinely accessing services, regardless of whether they are distressed 42.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This study was funded by a grant from the Dutch Cancer Society (RUG-2007-3805). The authors wish to express their gratitude to all study patients and to the participating cancer centers for their contribution to the data collection.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References