Fax: (310) 206 5343
Mental health in men treated for early stage prostate carcinoma†
A posttreatment, longitudinal quality of life analysis from the cancer of the prostate strategic urologic research endeavor™
Article first published online: 28 JUN 2002
Copyright © 2002 American Cancer Society
Volume 95, Issue 1, pages 54–60, 1 July 2002
How to Cite
Litwin, M. S., Lubeck, D. P., Spitalny, G. M., Henning, J. M. and Carroll, P. R. (2002), Mental health in men treated for early stage prostate carcinoma. Cancer, 95: 54–60. doi: 10.1002/cncr.10651
CaPSURE™ is sponsored by TAP Pharmaceutical Products Inc. and managed by The Urology Outcomes Research Group at the University of California, San Francisco.
- Issue published online: 28 JUN 2002
- Article first published online: 28 JUN 2002
- Manuscript Accepted: 4 FEB 2002
- Manuscript Received: 30 NOV 2001
- quality of life;
- prostate carcinoma;
- mental health;
- radical prostatectomy;
- pelvic radiation
The current study was conducted to assess posttreatment changes in the mental components of health related quality of life in prostate carcinoma patients during the two years following diagnosis and management with radical prostatectomy, pelvic irradiation, or watchful waiting.
The authors studied the mental domains of general health related quality of life in 452 men recently diagnosed with early stage prostate carcinoma and treated with radical prostatectomy, pelvic radiation, or watchful waiting. Outcomes were assessed with the RAND 36-Item Health Survey, a validated health-related quality of life instrument that includes four mental domains. To minimize the influence of potentially confounding factors, the authors adjusted for age, comorbidity, prostate specific antigen (PSA) at diagnosis, and biopsy Gleason score. All subjects were drawn from CaPSURE™, a national, longitudinal cohort.
By 6–12 months after treatment, the active treatment groups began to show differences in mental health and vitality. By 15 months, surgery and radiation patients scored differently in all four mental domains. Over time, the gaps between mental domain scores grew wider among the treatment groups, with surgery patients performing the best, radiation patients performing the worst, and watchful waiting patients falling in between.
The mental health profiles differ for patients undergoing surgery, radiation, or watchful waiting for early stage prostate carcinoma. Men with more serious disease, as evidenced by higher PSA levels or more aggressive histology, tended to worry more about it. Older men performed better, while sicker men performed worse, even though the older men tended to be sicker. Cancer 2002;95:54–60. © 2002 American Cancer Society.
Health related quality of life (HRQOL) outcomes have a substantial impact on men facing the difficult choice for treatment for early stage prostate carcinoma,1–4 particularly for those with early stage tumors, largely because treatment often affects various dimensions of HRQOL.5–10
In this context, mental health and its associated domains are especially relevant for men with prostate carcinoma. Although treatment may lead to successful long-term control of the tumor, anxiety and depression are well-known in patients with any type of cancer. This is true even though most men learn to adapt to specific impairments they experience.11 An important component of medical outcomes research in prostate carcinoma has been highlighting not only survival outcomes but also patient-centered outcomes, such as HRQOL. Ultimately, the dual goals of cancer cure and quality of life preservation are critical.12 In fact, some studies have suggested that men with prostate carcinoma are willing to make decisions that optimize quality of life rather than maximize survival.13–15
The objective of the current study was to document longitudinal changes in the mental components of quality of life in prostate carcinoma patients during the two years following diagnosis and treatment with radical prostatectomy, pelvic irradiation, or watchful waiting. These components include mental health, role limitations due to emotional problems, vitality, and social function. To minimize the influence of other factors, we adjusted for age, comorbidity, prostate specific antigen (PSA) at diagnosis, and biopsy Gleason score.
Subjects were drawn from the Cancer of the Prostate Strategic Urologic Research Endeavor (CaPSURE™) database, a project initiated in 1995 as a longitudinal, observational registry of men with biopsy proven prostate carcinoma. Patients are invited to join the database in a consecutive fashion as they present for outpatient care, regardless of tumor stage or treatment choice. Data are collected from 35 academic and community based clinical sites across the United States.
At enrollment the urologist completes an extensive prostate carcinoma history based on the existing medical record and the patient's current disease and treatment profile. Additional data are recorded at the time of each subsequent office visit. Every three months, patients complete a self-administered survey that includes HRQOL, satisfaction with care, resource utilization, and use of drugs, devices, and other treatments. Patients are followed until death or withdrawal from the study. Approval was obtained from the University of California, San Francisco (UCSF) Human Subjects Protection Committee, which serves as the umbrella organization for each clinical site; written informed consent is given by each patient before enrollment. Data collection, management, and quality control are conducted and centrally managed by the Urology Outcomes Research Groups at UCSF. Additional methodologic details on the database have been published previously.16 For inclusion in the current study, subjects had to have undergone radical prostatectomy or radiation or have actively selected watchful waiting within six months of diagnosis, and they had to have completed at least two HRQOL surveys during the followup period. However, HRQOL was not assessed before diagnosis or treatment. A total of 452 men met these criteria.
Patients had chosen either radical prostatectomy (RP; n = 282; age: 62.1 ± 7.1), pelvic irradiation (XRT; n = 104; age: 70.8 ± 6.0 years), or watchful waiting (WW; n = 66; age: 71.3 ± 8.1) within six months of a diagnosis of prostate carcinoma and completed at least two HRQOL assessments during the followup period. Mean duration of followup was 18.9 months (range, 4.6 to 25.1).
The mental components of quality of life were measured with the RAND 36-Item Health Survey 1.0 (SF-36),17–19 a self-administered, 36 item questionnaire that quantifies general HRQOL with 8 multi-item scales, including 4 mental domains which cover emotional well-being, role limitations due to emotional problems, vitality, and social function. In addition, a mental component summary scale may be calculated to provide a more global assessment of HRQOL in the mental domains. The SF-36 has been extensively tested and shown to be reliable and valid with test-retest reliability coefficients of 0.78 or greater and internal consistency Cronbach α coefficients20 of 0.78 to 0.93 in various populations.18 We have previously documented the performance of the SF-36 in CaPSURE participants,21 in men with early21, 22 and late23 stage prostate carcinoma, and in healthy older men without prostate carcinoma.24
Each domain is scored on a scale of 0 to 100 points, with higher values representing better outcomes. Hence, a higher emotional well-being score indicates better mental health, and a higher role-emotional score indicates less role limitation due to emotional problems. Differences of 5–10 points on the 0–100 scales of the SF-36 scales are thought to be clinically meaningful.25–27 Comorbidity was collected with an 11 item medical history checklist based on an established comorbidity rating scale.28 The checklist includes conditions such as arthritis, hypertension, heart disease, stroke, diabetes, lung disease, other cancers, kidney disease, blood disease, and gastrointestinal disease. In this analysis, patients were classified by the number of comorbidities they reported. This method of adjusting case mixes does not account for severity of illness. Gleason score and PSA at diagnosis were provided by each participant's urologist at the time of enrollment.
All analyses were performed using SAS Version 6.12 (SAS Institute, Cary, NC). In the current study we used the SAS Mixed procedure to model the four HRQOL mental domains outcomes as a function of treatment choice (XRT, RP, or WW) and time since treatment. The model included covariates of comorbidity count, PSA at diagnosis, Gleason score on biopsy, and age at end of treatment. Time was treated as a linear effect, and an interaction of time with treatment choice was included in the model. Model intercept and slope over time were treated as random effects. Least-squares means were computed for each treatment group at selected time intervals. Two-tailed tests of least-squares means differences between treatment choices were adjusted by the Tukey-Kramer method. Demographic and clinical variables among treatment groups were compared with the chi-square test (education, marital status, Gleason score) and one-way ANOVA (age, PSA, number of comorbidities). We did not control for nerve-sparing or for surgical approach to radical prostatectomy.
The characteristics of the study sample are presented in Table 1. Most patients were white, although 8.2% were African-American. Most were married or living with a partner, and 38% had completed college. On average, patients choosing RP were almost nine years younger (P < 0.0001), had slightly fewer comorbidities (P < 0.0001), scored 9–10 points higher on the SF-36 General Health Perceptions scale (P < 0.0001), and had PSA levels that were 3.8 ng/mL lower (P < 0.0001). Patients undergoing XRT were more likely to have Gleason scores from 7 to 10 (P < 0.001).
|Radical prostatectomy (n = 282)||Pelvic irradiation (n = 104)||Watchful waiting (n = 66)||Total sample (n = 452)|
|Age at treatment in years||62.1 ± 7.1||70.8 ± 6.0||71.3 ± 8.1||65.5 ± 8.3|
|White||248 (88%)||95 (91%)||134 (90%)||403 (89%)|
|African-American||23 (8%)||8 (8%)||10 (7%)||37 (8%)|
|Other||10 (4%)||1 (1%)||0 (0%)||11 (2%)|
|Married or living with spouse/partner||250 (92%)||79 (85%)||52 (83%)||381 (89%)|
|Grade school||7 (3%)||6 (6%)||3 (5%)||16 (4%)|
|Some high school||28 (10%)||13 (14%)||8 (13%)||49 (11%)|
|High school graduate||74 (27%)||28 (30%)||17 (27%)||119 (28%)|
|Some college||43 (16%)||23 (24%)||19 (30%)||85 (20%)|
|College graduate||65 (24%)||9 (10%)||8 (13%)||82 (19%)|
|Graduate school||56 (21%)||15 (16%)||9 (14%)||80 (19%)|
|PSA at diagnosis, ng/mL||8.8 ± 9.9||12.6 ± 12.5||11.6 ± 13.4||10.1 ± 11.2|
|2–4||23 (8%)||10 (10%)||11 (18%)||44 (10%)|
|5–6||195 (70%)||47 (46%)||35 (56%)||277 (63%)|
|7||42 (15%)||29 (28%)||10 (16%)||81 (18%)|
|8–10||17 (6%)||16 (16%)||6 (10%)||39 (9%)|
|Mean||5.9 ± 1.1||6.2 ± 1.4||5.7 ± 1.4||5.9 ± 1.2|
|Comorbidity count||1.3 ± 1.1||1.8 ± 1.1||2.1 ± 1.4||1.5 ± 1.2|
|General health perceptionsa||75.1 ± 18.4||65.7 ± 20.8||64.5 ± 16.5||71.4 ± 19.3|
|Clinical tumor stage|
|T1 N0 M0||80 (28%)||33 (32%)||23 (35%)||136 (30%)|
|T2 N0 M0||197 (70%)||66 (64%)||35 (53%)||298 (66%)|
|T3/4 or N+ or M+||5 (2%)||5 (5%)||8 (12%)||18 (4%)|
Table 2 and Figures 1–4 present least-squares mean scores for the four SF-36 mental domains (mental health, role limitations due to emotional problems, vitality, and social function) in tabular and graphic format. At 1.5 months, the three treatment groups scored similarly in each of the four domains, but, by 6-12 months after treatment, the RP and XRT groups began to show differences in mental health and vitality. By 15 months, RP and XRT patients scored differently in all four mental domains. By 18 months, differences were also evident between RP and WW patients in the social function domain. Over time, the gaps between mental domain scores grew wider among the three treatment groups, with RP patients performing the best, XRT patients performing the worst, and WW patients falling somewhere in between. By 24 months after treatment, the four domain scores differed by 10 to 14 points.
|Months since treatment||Mental health||Role limitations due to emotional problems||Vitality||Social function|
|1.5||80 ± 0.9||76 ± 1.6||77 ± 2.0||76 ± 2.1||75 ± 3.6||79 ± 4.5||62 ± 1.1||61 ± 2.0||61 ± 2.5||75 ± 1.3||81 ± 2.2||79 ± 2.8|
|6||81 ± 0.9||76 ± 1.5||78 ± 1.9||80 ± 1.8||77 ± 3.1||81 ± 3.9||64 ± 1.1||61 ± 1.9||62 ± 2.3||80 ± 1.1||82 ± 1.9||81 ± 2.4|
|9||81 ± 0.9||75 ± 1.5||78 ± 1.9||82 ± 1.6||77 ± 2.9||82 ± 3.6||65 ± 1.1||61 ± 1.9||63 ± 2.3||83 ± 1.0||83 ± 1.8||83 ± 2.3|
|12||82 ± 0.9||75 ± 1.5||79 ± 1.9||85 ± 1.6||78 ± 2.8||83 ± 3.4||67 ± 1.1||61 ± 1.9||63 ± 2.4||87 ± 1.0||84 ± 1.8||84 ± 2.3|
|15||83 ± 0.9||75 ± 1.6||80 ± 1.9||87 ± 1.6||79 ± 2.9||84 ± 3.5||68 ± 1.1||61 ± 2.0||64 ± 2.5||90 ± 1.0||84 ± 1.9||85 ± 2.4|
|18||84 ± 0.9||75 ± 1.7||80 ± 2.0||89 ± 1.7||80 ± 3.1||84 ± 3.8||70 ± 1.2||61 ± 2.1||65 ± 2.6||93 ± 1.2||85 ± 2.1||86 ± 2.6|
|21||84 ± 1.0||75 ± 1.8||81 ± 2.2||92 ± 1.8||80 ± 3.4||85 ± 4.2||71 ± 1.3||61 ± 2.3||65 ± 2.8||97 ± 1.3||86 ± 2.4||88 ± 2.9|
|24||85 ± 1.0||75 ± 1.9||81 ± 2.4||94 ± 2.0||81 ± 3.8||86 ± 4.7||73 ± 1.4||61 ± 2.5||66 ± 3.1||100 ± 1.4||86 ± 2.7||89 ± 3.3|
Table 3 presents the results of the multivariate regression analysis which simultaneously controlled for the effects of all relevant covariates. Significant effects on various mental health domains were seen for age, treatment choice, PSA at diagnosis, Gleason score, comorbidity, time, and the interaction of time with treatment choice. Of the 20 statistically significant effects in the model, 14 had P values equal to or less than 0.01. Variables with the most positive effects on the mental health domains included age, time since treatment, and treatment choice modified by time. Those with the most negative effects included PSA, Gleason score, and comorbidity.
|Source||Mental health||Role limitations due to emotional problems||Vitality||Social function|
|Effect||P value||Effect||P value||Effect||P value||Effect||P value|
|Age (yrs)||0.43||< 0.0001||0.26||NS||0.24||0.04||0.42||0.0003|
|XRT vs WW||−0.55||−3.63||−0.033||2.03|
|RP vs WW||2.78||−3.71||0.21||−5.32|
|PSA at diagnosis||−0.13||0.038||−0.42||0.0002||−0.17||0.026||−0.21||0.0038|
|Time (days)||0.0066||0.006||0.011||0.0002||0.0068||0.0002||−0.14||< 0.0001|
|Time × treatment||0.005||0.04||0.0002||< 0.0001|
|XRT vs WW||−0.0086||−0.0024||−0.0070||−0.0065|
|RP vs WW||0.0011||0.016||0.0090||0.023|
The current study has three important findings. First, the mental health profiles differ for patients undergoing surgery, radiation, or watchful waiting for early stage prostate carcinoma. Initially, the three groups appear similar; however, meaningful differences begin to appear over time. After surgery, mental health, role limitations due to emotional problems, vitality, and social function all improve substantially and appear to be headed higher as long as 24 months after treatment. After radiation, improvement is slower in role limitations and social function, while mental health and vitality remain fairly constant, despite the passage of time. Men who choose expectant management see modest improvement in all four domains, but their scores consistently remain ensconced between scores of the two active treatments. It is remarkable that at all times from six months after treatment, surgery patients have higher mental domain scores than do radiation patients.
Second, in multivariate analyses, PSA and Gleason score appear to have a negative impact on the mental domains of HRQOL over time. This is consistent with clinical experience, which suggests that men with more serious disease, as evidenced by higher PSA levels or more aggressive histology, tend to worry more about it. Although it would have been useful to have examined surgical pathology as a predictor of mental HRQOL, this variable was not available for the XRT and WW groups. This observation illustrates the difficulty in controlling for such important factors as prognosis, which is strongly impacted by tumor factors evident only from a pathology report. Mental domains of HRQOL are undoubtedly affected by whether the patient sees himself as cured, this effect enhancing his ability to adapt to impairments he may be experiencing in urinary, sexual, or bowel function.
Third, in multivariate analyses, age and comorbidity were significant predictors for three of the four mental outcomes. Older men performed better, while sicker men performed worse, even though the older men tended to be sicker. While patients with greater comorbidity might be expected to have worse mental domain scores, it is not clear why older men scored better. Nonetheless, this is consistent with findings from other investigators, who have shown that demographic variables tend to affect various outcomes in men with prostate carcinoma.29, 30
The current findings confirm longitudinal trends observed by Talcott et al.31, 32 and Litwin et al.,33, 34 who showed that disease-specific HRQOL evolves over time following surgery or radiation. Men undergoing either treatment have been shown to benefit from strong social support,10 and the current study confirms that social function may be substantially impacted. Despite previous cross-sectional evidence that the general HRQOL domains are less affected than the disease specific domains by prostate carcinoma treatment,22 the current longitudinal study indicates that over time the mental components of general HRQOL are in fact impacted.
The current findings are limited by several important methodologic considerations. First, the demographic homogeneity of the current sample may reduce the generalizability of the findings. The CaPSURE participants represent a convenience sample of men selected from urology private practices; these men tend to be more affluent and better educated than the general population. Second, this analysis did not include longitudinal PSA measurements, which might have provided a useful context in which to interpret the relative differences in mental domain scores. Such an analysis is presently underway. Third, although we measured urinary, sexual, and bowel impairment, we did not include them in the current study. A separate analysis to explore the effects of pelvic dysfunctions on mental domains of quality of life is also underway. Fourth, since we did not document pretreatment HRQOL, we were prevented from drawing firm conclusions about the independent impact of treatment on mental health outcomes. Fifth, the treatment groups may have been inherently different in areas that cannot be adequately adjusted with multivariate methods. This concern is tempered by the statistical similarity of all groups' baseline mental domain scores. Finally, the current study did not include men who had undergone brachytherapy, which may have an entirely different set of effects on mental outcomes following treatment.
Despite these limitations, the prospective, longitudinal design strengthens the value of the current conclusions; hence, the observations may have important implications for men choosing among surgery, radiation, and watchful waiting for early stage prostate carcinoma. Men undergoing surgery have better mental health, fewer role limitations due to emotional problems, more vitality, and higher social function following treatment when compared with men choosing radiation or watchful waiting. These differences are both statistically and clinically meaningful. The gaps between treatment groups continue to widen over time. Patients undergoing radiation see only modest improvements in role limitations and social function over time while showing no improvement in mental health and vitality.
For most men, the treatment choice for early stage prostate carcinoma may not ultimately make a difference in survival;35 however, they must live with the quality of life effects for many years. If the goals of treatment are not only to add years to life, but also to add life to years, then the mental domains of quality of life must be considered important factors when selecting a management plan. Profiling the quality of life differences among treatments will help men who are faced with difficult treatment choices after diagnosis with early stage prostate carcinoma.
- 12Outcomes of cancer treatment for technology assessment and cancer treatment guidelines. American Society of Clinical Oncology. J Clin Oncol. 1996; 14: 671–679.
- 19SF-12: How to score the SF-12 physical and mental health summary scales. Boston, MA: The Health Institute, New England Medical Center, 1995., , .
- 25SF-36 physical and mental health summary scales: a user's manual. Boston: The Health Institute, New England Medical Center, 1994., , .