Ethnic variation in return to baseline values of patient-reported outcomes in older prostate cancer patients

Authors


Abstract

BACKGROUND.

Ethnic variation in patient-reported outcomes such as health-related quality of life (HRQoL) and satisfaction with care are understudied areas in the management of elderly prostate cancer (PCa) patients.

METHODS.

In this prospective cohort study, between the years 2002 and 2004, the authors recruited 214 older (≥65 years) men with newly diagnosed PCa from an urban academic hospital and a Veterans Administration hospital. Participants completed generic (SF-36), prostate-specific (UCLA-PCI) HRQoL, and satisfaction with care (CSQ-8) surveys at baseline and at 3, 6, and 12-months follow-up. Clinically significant difference was used to compute return to baseline. The authors compared time to return to baseline HRQoL after controlling for confounding variables by using ANOVA and log-linear models. Survival curves were used to compare time to return to baseline across ethnicity.

RESULTS.

Regression analysis revealed that age and marital status, not ethnicity, were independent predictors of radical prostatectomy, rather than radiation treatment. African Americans reported lower HRQoL scores at diagnosis and required a longer time to return to baseline. Log-linear analysis indicated that African-American ethnicity was associated with lower 12-month scores for role physical (odds ratio [OR], 0.46; standard error [SE], 0.40), role emotional (OR, 0.37; SE, 0.43), bodily pain (OR, 0.74; SE, 0.10), urinary function (OR, 0.90; SE, 0.11), and urinary bother (OR, 0.72; SE, 0.17). Both groups reported comparably high levels of satisfaction with care.

CONCLUSIONS.

African-American elderly exhibited poorer outcomes and required a longer time to return to baseline HRQoL. These differences highlight the need for discussion with patients and families prior to treatment about expectations and the need for support services post-treatment. Cancer 2007. © 2007 American Cancer Society.

Prostate Cancer (PCa) is the leading cancer diagnosed among elderly men in the USA.1 Race/ethnicity and age influence PCa diagnosis, treatment, and outcomes.2–12 For localized and regional disease stages, Caucasian men are more likely to receive radical prostatectomy and African-American men are more likely to receive radiation.4 African-American men who receive radical prostatectomy have often exhibited more adverse pathological features than Caucasians.4, 13 After adjusting for age, prostate-specific antigen (PSA) level, grade, and stage, an ethnic variation was observed in progression-free survival of PCa patients with localized disease who were treated with radical prostatectomy, with African Americans reporting shorter progression-free survival.2, 4–8

Because of uncertainty about the value of screening and the timing of treatment of elderly PCa patients, outcomes such as health-related quality of life (HRQoL) and satisfaction with care have assumed increasing importance.2–9 Post-treatment recovery of generic and prostate-specific HRQoL is a major concern for post-treatment planning and care. However, despite expanding literature on disparity in treatment and survival, little is known about the disparities in HRQoL (generic and prostate-specific), satisfaction with care among elderly PCa patients, and the relation of those disparities to treatment type. Because satisfaction with PCa care offers a patient-centered approach to assess the perceived quality of treatment, measuring patient satisfaction with care has gained increased attention in PCa research. However, studies of PCa outcomes have contained a variety of weaknesses—recollection bias, absence of baseline data, incomplete follow-up, retrospective study design, and failure to measure HRQoL—that limit understanding the significance of these quality of life and satisfactionoutcomes. Therefore, we designed a prospective cohort study to determine the association of ethnicity with post-treatment recovery of HRQoL outcomes (generic and prostate-specific) and satisfaction with care among African-American and Caucasian elderly men with newly diagnosed PCa. We hypothesized that African-American elderly men exhibit poorer presentation at baseline and recovery pattern on prostate-specific and generic HRQoL, after adjusting for demographic and clinical factors.

MATERIALS AND METHODS

An observational prospective cohort design was used to recruit older (≥65 years) newly diagnosed PCa patients from a large urban academic hospital and a Veterans Administration (VA) medical center between 2002 and 2004. The study was approved by local institutional review boards. All study personnel completed subject-protection training and met appropriate health information portability and accountability act (HIPAA) education requirements before engaging in this research. The study included self-identified African-American or Caucasian men of age ≥65 years at the time of diagnosis, newly diagnosed for PCa in the prior 4 months, and in whom treatment had not been initiated. The diagnosis of PCa was based on prostate biopsy complemented by prostate-specific antigen (PSA) level and staging. Patients were excluded if they were diagnosed with metastatic cancer, had visited these clinics for a second opinion only, were unable to communicate in English, were cognitively impaired, and/or were unavailable by mail or telephone.

Subject Selection and Recruitment

Recruitment

Newly diagnosed PCa cases were identified and recruited at the urology clinics of an academic medical center and a Veterans Administration medical center. Potential participants received study information from their urologists during clinic visits, and interested patients were contacted by study research assistants. Also, attendees to a weekly prostatectomy orientation class organized at these urology clinics were contacted. At this stage, a potential participant could agree to participate and complete the consent form. In case a person wanted to be contacted later, the research assistant did so. During the telephone contact, if the potential participant agreed to participate, he was mailed a consent form and a prepaid return envelope. Participants were encouraged to discuss the consent form with the research assistant before signing.

Retention Plan and Follow-up

After obtaining baseline data, participants were contacted by mail at 3, 6, and 12 months for follow-up data. Nonrespondents were contacted by telephone after 10 days. A second mail survey was sent to nonrespondents within 4 weeks of the first mailing. During study enrollment, the importance of active participation was emphasized. Also, during scheduled clinic visits, the urologists encouraged participants to continue their participation.

Outcome Measures and Data Collection

After obtaining informed consent and HIPAA authorization from participants, baseline data on generic and prostate-specific HRQoL was obtained before initiation of treatment. A self-report questionnaire was used to obtain data on ethnicity, education, marital status, living arrangement, and income. A structured medical chart review was used to collect data on age, date of PCa diagnosis, health insurance, treatment, PSA scores, Gleason score, TNM stage, and comorbidities. Prostate cancer treatment was classified as radical prostatectomy, external beam radiation therapy, hormone therapy, and no treatment.

Prostate-cancer specific HRQoL was assessed by using the University of California at Los Angeles' Prostate Cancer Index (UCLA-PCI). This comprehensive, self-administered, twenty-item questionnaire quantifies prostate-specific HRQoL in 6 domains (urinary function, urinary bother, sexual function, sexual bother, bowel function, and bowel bother). PCI has performed well in older populations, has demonstrated good psychometric properties, and is easy to understand and complete.19 Generic HRQoL was measured by using the Medical Outcomes Study Short Form (SF-36). It is a single multi-item scale that assesses 8 health domains (physical limitation caused by health problems, limitations on social activities caused by physical and/or emotional problems, role limitations caused by physical health problems, and emotional problems, bodily pain, general mental health, vitality, and general health perceptions). It was constructed for self-administration or for administration by a trained interviewer, either in person or by telephone, and it has been tested for reliability and validity.20 The range of possible scores for each subscale is 100% to 0%. A high score on SF-36 and PCI indicates a higher quality of life.

Satisfaction with care is defined as a pleasant feeling caused by the fulfillment of expectations21 and was measured by using a self-administered Client Satisfaction Questionnaire (CSQ-8). This questionnaire has been extensively studied and has good psychometric properties.22 A high score on CSQ-8 indicates greater satisfaction with care. Baseline Charlson comorbidity index23 was computed using ICD9 codes for all inpatient and outpatient events during 3 months prior to PCa diagnosis. These data were obtained from hospital-based administrative databases.

Statistical Analysis

Demographic and clinical variables were compared across ethnicity by Student t-test and chi-square analyses. Baseline clinical and demographic predictors of treatment were evaluated by logistic regression. Mean HRQoL at baseline, 3, 6, and 12 months were compared. Post-treatment satisfaction with care was compared by chi square. Return to baseline for follow-up HRQoL scores was defined in 2 ways. First, a change of 7 points or less, which is considered to be a clinically significant difference,20, 24 was used as return to baseline. During follow-up, a participant was considered to have returned to baseline for an HRQoL domain if differences in scores between baseline and follow-up was ≤7 points. Alternatively, we defined “return to baseline” as a “minimally important difference” of 0.5 times the standard deviation (SD) for each HRQoL subscale.25

We compared the proportion of participants returning to baseline at 3, 6, and 12 months for all HRQoL subscales by the chi-square method. Mean time to return to baseline was determined by survival analysis. Proportion of participants who never returned to baseline was also compared by chi- square.

Repeated-measure ANOVA was used to compare change in HRQoL from baseline to 3, 6, and 12 months across ethnicity, after controlling for age, Charlson comorbidity score, treatment, and hospital type. Log-linear regression was used to determine the association of ethnicity with 12-month HRQoL scores for all subscales, after controlling for age, education, marital status, Charlson comorbidity score, baseline PSA, baseline score, treatment, TNM stage, and hospital type. The following variables were dichotomized, race (1 = African American; 0 = Caucasian), marital status (1 = married; 0 = other), education (1 = high school (HS) or less; 0 ≥ HS), treatment group (1 = radical prostatectomy; 0 = radiation therapy), TNM stage (1 = T1a to T2a; 0 = T3a to T3b), and hospital type (1 = non-Veterans Administration; 0 = Veterans Administration).

RESULTS

We recruited 214 (72 African Americans and 142 Caucasians) older (≥65 years) patients with newly diagnosed PCa. Of these, 195 (65 African Americans and 130 Caucasians) completed 3-month follow-up, 184 (62 African Americans and 122 Caucasians) completed 6-month follow-up, and 182 (61 African Americans and 121 Caucasians) completed 12-month follow-up. Comparisons of demographics and clinical characteristics are presented in Table 1. African-American men were older (mean = 71.25; SD = 4.1), compared with Caucasian men (mean = 69.87; SD = 4.5). The majority of Caucasian men were college educated, married, and had an annual income of $40,000 or more. Caucasian participants were mostly from the non-Veterans Administration hospital, whereas the majority of African-American participants were from the Veterans Administration hospital. Their mean Charlson comorbidity scores were comparable. A higher proportion of African-American men reported having to urinate too often and had pain or aches in the back, hips, or legs.

Table 1. Demographics and Clinical Characteristics of Participants
CovariatesCaucasians (n = 142)AA (n = 72)P
  1. SD indicates standard deviation.

Age (mean±SD)69.87 ± 4.571.25 ± 4.1.05
Charlson comorbidity (mean±SD)1.22 ± 2.22.18 ± 2.8.109
Education (%)  <.001
 HS or less31.4369.49 
 College or more68.5730.51 
Marital status (%)  .004
 Single/Widowed/Div22.4642.37 
 Married77.5457.63 
Employment Status (%)  .103
 Full-time15.226.78 
 Part-time/other84.7893.22 
Income Level (%)  <.001
 ≤$40,00037.2178.18 
 >$40,00062.7921.82 
Hospital type  .001
 Non-VA57.8632.41 
 VA42.1467.59 
Signs and symptoms (%)   
 Difficulty or discomfort urinating21.5825.86.514
 Having to urinate too often45.0061.40.036
 Weak urinary stream36.2340.00.625
 Infection of bladder or prostate9.3512.50.512
 Blood in urine5.8010.53.244
 Pain or aches in back, hips or legs26.0947.46.003
 More tired or worn out than usual25.7433.33.283
PSA-at diagnosis, ng/mL, mean±SD7.25 ± 5.98.94 ± 7.8.362
Gleason score (total, ±SD)6.34 ± 0.846.0 ± 1.6.192
TNM stage (%)  .093
 T1a1.300.00 
 T1b1.300.00 
 T1c66.2368.42 
 T2a14.2921.05 
 T2b3.900.00 
 T2c5.190.00 
 T3a7.795.26 
 T3b0.005.26 
Treatment received  .033
 Radical prostatectomy54.721.05 
 External beam radiation therapy44.378.95 

Table 1 also shows clinical characteristics of participants at diagnosis and treatment. Clinical and pathologic stages ranged from American Joint Commission on Cancer staging classification T1N0M0 (clinically inapparent tumor not palpable or visible by imaging [T1], no regional lymph-node metastasis [N0], and no distant metastasis [M0]) to T3bN0M0 (tumor extends through the prostate capsule [T3], no regional lymph-node metastasis [N0], and no distant metastasis [M0]). The majority of participants were between stages T1c and T2a. Tumors were moderately differentiated with a mean Gleason score of 6.34 (SD = 0.84) for Caucasians versus 6.0 (SD = 1.6) for African Americans (P = .19). PSA score and stage of cancer at diagnosis were comparable between groups (P = .36). Treatment pattern differed by ethnicity; a higher proportion of African Americans received radiation, whereas a higher proportion of Caucasian received prostatectomy (P = .03). However, logistic regression showed that not ethnicity but age (OR, 0.65; CI, 0.51–0.82) and marital status (OR, 5.5; CI, 1.2–26.4) were predictors of radical prostatectomy treatment (data not shown).

Baseline HRQoL

A comparison of pretreatment assessment of generic HRQoL (Table 2) showed that African-American men reported significantly lower scores for all generic HRQoL subscales, except bodily pain. For prostate-specific HRQoL, groups were comparable for sexual function, urinary bother, bowel bother, and sexual bother subscales.

Table 2. Comparison of Baseline HRQoL by Ethnicity (mean ± SD)
VariableCaucasian (n = 142)AA (n = 72)P
  1. HRQoL indicates health-related quality of life; SD, standard deviation.

RAND 36 item Generic HRQoL
 Physical function62.4 ± 20.552.0 ± 25.5.006
 Role physical76.8 ± 37.261.2 ± 44.9.013
 Role emotional85.9 ± 31.170.2 ± 40.7.004
 Vitality67.3 ± 22.160.6 ± 23.7.060
 Mental health81.8 ± 14.375.8 ± 18.4.016
 Social function89.3 ± 19.179.8 ± 26.2.004
 Bodily pain82.7 ± 21.679.9 ± 23.1.430
 General health70.0 ± 20.258.9 ± 20.1.005
UCLA prostate cancer specific HRQoL
 Urinary function90.2 ± 17.784.1 ± 18.2029
 Bowel function88.9 ± 14.283.2 ± 19.3.022
 Sexual function39.9 ± 29.937.6 ± 26.7.601
 Urinary bother85.6 ± 22.584.2 ± 23.9.704
 Bowel bother88.9 ± 20.584.3 ± 27.4.202
 Sexual bother56.1 ± 40.557.9 ± 34.9.761

Post-treatment HRQoL and satisfaction with care.

Differences were seen in the rate and pattern of recovery post-treatment; however, both ethnic groups reported a comparable high level (>90%) of satisfaction with care at 12-months post-treatment (Table 3). Table 4 presents the comparison of proportion of participants returning to baseline, proportion not returning to baseline, and mean number of days to return to baseline. For generic HRQoL, at 3 months post-treatment, the proportion of participants returning to baseline was comparable by ethnicity. A lower proportion of African Americans returned to baseline for bodily pain and general health at 6 months. A higher proportion of African Americans did not return to baseline by 12 months for physical function, role emotional, mental health, social function, bodily pain, and general health. Also, African Americans took longer to return to baseline for physical function, role emotional, mental health, social function, bodily pain, and general health. With the exception of the sexual bother subscale, the proportion returning to baseline on all other PCa-specific HRQoL subscales were comparable at all times. A higher percentage of African Americans did not return to baseline for sexual bother over the 12-month period. They also took longer to return to baseline. We repeated the analysis using “minimally important difference” (0.5 × SD) as the criterion for “return to baseline.” The results (data not shown) were comparable to those obtained by using a change score of 7 points, a difference score deemed to be clinically important.

Table 3. Comparison of Satisfaction With Care (CSQ-8) by Ethnicity
VariableCaucasian (n = 121)AA (n = 61)P
Post-treatment satisfaction with care
How would you rate the service you have received?  .941
Poor/Fair4.104.35 
Good/Excellent95.9095.65 
Did you get the kind of service you wanted?  .852
No7.386.52 
Yes92.6293.48 
To what extent has treatment met your needs?  .453
None/Only a few9.0913.04 
Most/Almost all90.9186.96 
If a friend were in need of similar help, would you recommend our program to him?  .412
No4.962.13 
Yes95.0497.87 
How satisfied are you with the amount of help you have received?  .583
Dissatisfied6.614.35 
Satisfied93.3995.65 
Have the services you received helped you to deal more effectively with your problems?  .721
No5.834.44 
Yes94.1795.56 
In an overall sense, how satisfied are you with the service you have received?  .663
Dissatisfied4.966.67 
Satisfied95.0493.33 
If you were to seek help again, would you come back to our program?  .421
No5.002.17 
Yes95.0097.83 
Table 4. Percentage of Patients Returning and Not-Retuning to Baseline Scores During 12 Months of Follow-Up
 3 mo, % n = 1956 mo, % n = 18412 mo, % n = 182% of Patients not returning to baseline valuesMean days to return to baseline values
CaucasianAACaucasianAACaucasianAACaucasianAACaucasianAA
  • HRQoL indicates health-related quality of life; AA, African American.

  • *

    P < .05.

Generic HRQoL
 Physical function70.4367.4479.1369.7782.1876.329.4522.64*148181*
 Role physical65.7976.3282.1471.7985.0073.537.0914.89150158
 Role emotional84.3576.9289.2980.0092.8664.71*2.4012.50*123153*
 Vitality61.7458.5478.0769.0570.8754.0512.6019.61165187
 Mental health74.7865.8577.1961.9083.5071.055.5119.61*143185*
 Social function57.3954.7674.3465.1277.4562.1611.9030.77*171205*
 Bodily pain60.0048.7866.6745.24*72.5545.46*15.7531.57*179215*
 General health72.4165.9176.5260.49*71.8455.2614.8726.42*151191*
Prostate cancer specific HRQoL
 Urinary function53.5760.4760.7164.2955.0054.0523.3922.64195193
 Bowel function71.0566.6769.9173.1774.2678.3814.407.69156166
 Sexual function46.7342.5041.1246.1536.8454.2947.9038.00208239
 Urinary bother46.8560.4765.7755.8169.7055.2619.5026.42199207
 Bowel bother78.0769.0582.3075.6182.1878.387.2011.54136167
 Sexual bother64.4242.11*62.3834.29*60.8731.25*20.6945.83*177243*

Table 5 presents results of repeated-measure ANOVA, which was used to compare mean changes in scores. Lower values indicate better post-treatment function. The pattern of mean change scores for physical function was significantly different across ethnicity. For role physical, Caucasians improved over time. The African Americans had improved at 6 months, but their function score had declined at 12 months. For role emotional, both groups exhibited reverse trends. For vitality, African Americans improved over time. Caucasians had lower than baseline levels at 3 months, improved by 6 months, and had somewhat declined by 12 months. By 3 months, Caucasians reported a decline in mental health scores that improved by 12 months. On the other hand, for African Americans, mental health score stayed higher than baseline during follow-up, although the difference was smaller at 12 months. Social function at 3 months was lower than baseline level but improved by 6 months for Caucasians and African Americans. The pattern of change for scores on bodily pain was significantly different by ethnicity. General health declined slightly for Caucasians over time, whereas it improved for African Americans over time.

Table 5. Comparison of Mean Change in Scores by Ethnicity (Mean±SE)
SubscalesBaseline, 3 moBaseline, 6 moBaseline, 12 moP
CaucasianAACaucasianAACaucasianAABetween group effect (Ethnicity)Within group effect (Time)
  1. HRQoL indicates health-related quality of life; AA, African American.

Generic HRQoL
 Physical function4.5 (2.3)4.8 (3.7)1.1 (1.9)6.1 (3.1)1.9 (1.9)1.5 (3.1).0425.2284
 Role physical19.7 (4.9)12.6 (8.2)5.6 (3.7)7.6 (6.0)5.3 (3.6)12.4 (5.9).9174.0881
 Role emotional5.9 (4.8)1.8 (7.9)4.3 (4.5)5.8 (7.4)−.24 (4.2)14.1 (6.9).6005.7160
 Vitality6.9 (2.2)10.9 (3.6)−0.04 (2.2)6.3 (3.6)1.04 (2.4)6.1 (3.9).1791.0071
 Mental health3.4 (1.7)−2.3 (2.8)0.59 (1.8)−1.9 (2.9)−.91 (1.8)−1.7 (2.9).3228.4405
 Social function11.7 (3.0)8.04 (25.7)6.5 (2.6)3.4 (4.1)3.1 (2.3)5.5 (3.7).7767.0379
 Bodily pain8.5 (2.7)11.9 (4.3)3.7 (2.3)14.3 (3.8)1.3 (2.3)12.1 (3.8).0320.3309
 General health1.18 (1.8)5.7 (2.9)1.99 (1.8)3.08 (2.9)2.9 (1.8)1.74 (2.9).6299.7320
Prostate cancer-specific HRQoL
 Urinary function14.61 (3.8)20.9 (6.2)12.2 (3.6)19.4 (5.8)11.05 (3.2)16.7 (5.2).3093.3171
 Bowel function0.82 (2.1)9.5 (3.6)1.5 (2.4)6.3 (4.0)−0.28 (2.1)3.09 (3.6).1588.0988
 Sexual function14.9 (3.2)19.3 (5.9)16.5 (3.6)18.3 (5.9)13.8 (3.4)19.4 (5.5).5669.8941
 Urinary bother13.4 (3.8)29.5 (6.0)11.2 (4.0)28.3 (6.3)7.2 (2.9)17.8 (4.6).0175.0057
 Bowel bother0.94 (3.4)15.80 (5.6)−0.46 (3.3)7.12 (5.5)0.69 (2.9)3.89 (5.0).1312.0554
 Sexual bother11.34 (5.2)26.9 (8.5)16.7 (5.8)35.06 (9.3)10.17 (5.9)32.71 (9.5).0412.3495

For both groups, scores on urinary function at 12 months were lower than baseline level, but the mean change score declined most among African Americans. The mean score on bowel function showed improvement over 12 months for African Americans. For both groups, sexual function at 12 months was lower than baseline values. Both groups demonstrated improved scores on urinary bother by 12 months. The scores on bowel bother improved greatly by 12 months for African Americans. Scores on sexual bother declined at 6 months and improved by 12 months for both groups. However, the improvement was lower for African-American men.

Results of log-linear regression to determine association of ethnicity with 12-month HRQoL scores are presented in Table 6. Other covariates were age, education, marital status, Charlson comorbidity score, baseline PSA score, treatment type, TNM stage, baseline score, and hospital type. African-American ethnicity was associated with a lower 12-month score for role physical (OR, 0.46), role emotional (OR, 0.37), and bodily pain (OR, 0.74). As shown in Table 7, for prostate-specific HRQoL, African-American ethnicity was associated with lower 12-month scores for urinary function (OR, 0.90) and urinary bother (OR, 0.72).

Table 6. Predictors of 12-month Generic HRQoL Scores
RAND 36-item health survey
Independent variablesPhysical functionRole physicalRole emotionalVitalityMental healthSocial functionBodily painGeneral health
OR (SE)OR (SE)OR (SE)OR (SE)OR (SE)OR (SE)OR (SE)OR (SE)
  • HRQoL indicates health-related quality of life; OR, odds ratio; SE, standard error; RP, radical prostatectomy; AA, African American; PSA, prostate-specific antigen.

  • *

    P < .05.

Intercept3.3 (1.50)9.9 (3.2)198 (0.34)164 (1.1)40.4 (0.67)18.2 (0.84)10.8 (0.68)1.6 (1.0)*
Baseline values1.02 (0.003)*1.02 (0.004)*1.01 (0.004)*1.01 (0.003)*1.01 (0.002)*1.00 (0.002)*1.02 (0.002)*1.02 (0.003)*
Age at treatment1.03 (0.02)1.02 (0.05)0.97 (0.05)0.97 (0.02)1.01 (0.01)1.01 (0.01)1.02 (0.01)*1.03 (0.02)*
AA ethnicity0.93 (0.22)0.46 (0.40)*0.37 (0.43)*1.04 (0.15)1.07 (0.09)1.02 (0.12)0.74 (0.10)*1.49 (0.14)
Education0.67 (0.18)*0.78 (0.33)0.95 (0.35)0.87 (0.13)0.91 (0.08)0.94 (0.10)1.02 (0.08)1.02 (0.12)
Married1.19 (0.21)0.054 (0.39)0.86 (0.41)0.93 (0.15)0.95 (0.09)0.86 (0.12)0.86 (0.01)1.03 (0.13)
PSA at baseline1.001 (0.02)1.02 (0.04)1.00 (0.04)1.01 (0.01)1.01 (0.008)0.99 (0.01)1.01 (0.01)0.99 (0.01)
TNM stage0.35 (0.31)*0.42 (0.54)0.53 (0.60)0.79 (0.22)1.2 (0.13)0.74 (0.16)0.73 (0.14)*0.99 (0.19)
RP treatment0.85 (0.15)0.64 (0.28)0.77 (0.30)0.87 (0.11)0.94 (0.07)0.92 (0.08)0.91 (0.07)0.85 (0.10)
Charlson comorbidity0.98 (0.03).086 (0.07)*0.93 (0.08)0.97 (0.02)0.98 (0.02)0.98 (0.02)1.01 (0.02)0.97 (0.02)
Non-VA hospital1.04 (0.24)2.2 (0.42)1.75 (0.47)1.11 (0.16)1.09 (0.01)1.35 (0.13)*1.06 (0.11)1.49 (0.15)*
R20.620.530.330.440.290.450.560.62
Table 7. Predictors of 12-Month Prostate-Specific HRQoL Scores
University of California at Los Angeles prostate cancer index
 Urinary functionBowel functionSexual functionUrinary botherBowel botherSexual bother
OR (SE)OR (SE)OR (SE)OR (SE)OR (SE)OR (SE)
  • HRQoL indicates health-related quality of life; OR, odds ratio; SE, standard error; RP, radical prostatectomy; AA, African American; PSA, prostate-specific antigen.

  • *

    P < .05.

Intercept2.6 (0.85)244 (0.63)298 (4.2)16.4 (1.2)*270 (1.6)89 (4.6)
Baseline values1.00 (0.003)1.00 (0.002)1.03 (0.007)*1.00 (0.002)1.00 (0.004)1.02 (0.006)*
Age at treatment1.00 (0.01)0.98 (0.009)*0.93 (0.06)1.00 (0.02)0.97 (0.02)0.98 (0.07)
AA ethnicity0.90 (0.11)*1.04 (0.09)0.77 (0.53)0.72 (0.17)*0.94 (0.22)0.71 (0.61)
Education1.03 (0.09)1.02 (0.07)0.52 (0.44)1.3 (0.15)*1.07 (0.18)0.46 (0.53)
Married1.09 (0.11)1.04 (0.08)0.80 (0.55)1.1 (0.17)1.02 (0.21)5.8 (0.64)*
PSA at baseline1.02 (0.009)*1.02 (0.007)*1.03 (0.04)1.0 (0.01)1.03 (0.02)1.06 (0.05)
TNM stage0.79 (0.15)0.61 (0.12)*0.75 (0.68)0.33 (0.24)*0.17 (0.30)*0.27 (0.86)
RP treatment1.08 (0.07)0.98 (0.06)1.7 (0.33)0.97 (0.11)0.92 (0.15)1.29 (0.44)
Charlson comorbidity0.99 (0.02)1.00 (0.01)1.00 (0.08)0.99 (0.03)1.00 (0.03)1.05 (0.10)
Non-VA hospital0.98 (0.12)1.09 (0.09)0.96 (0.60)1.18 (0.19)1.5 (0.25)0.52 (0.72)
R20.260.320.310.400.430.29

DISCUSSION

Prostate cancer is the most commonly diagnosed cancer in elderly men, and as the population continues to age, PCa will have an increasingly significant influence on healthcare delivery and health outcomes. Racial and ethnic disparities in patient-reported outcomes, such as HRQoL and satisfaction with care, among elderly men with newly diagnosed PCa have not been sufficiently addressed. In this prospective cohort study of 214 elderly men, we evaluated the impact of differential treatments received by African Americans and Caucasians on generic and prostate-specific HRQoL and satisfaction with care. Main findings of this study are as follows:

  • 1)Compared with Caucasian elderly men, African-American elderly men take a longer time to return to their baseline generic and sexual bother HRQoL scores.
  • 2)African Americans reported poorer outcomes by 12 months for some generic and prostate specific HRQoL domains.
  • 3)When baseline and clinical and demographic characteristics are statistically controlled, ethnicity is not associated with treatment type.
  • 4)African-American ethnicity is associated with lower 12-month scores on role physical, role emotional, bodily pain, urinary function, and urinary bother.
  • 5)There is no significant ethnic difference in satisfaction with care.

Other studies have shown that ethnicity influences treatment and affects cancer recurrence and outcome in elderly.8–16, 26–30 African-American men have a higher incidence of PCa, exhibit poorer stage-specific survival than Caucasians, and have a higher rate of presentation with late stage disease.1, 15, 26 A cohort study that used Surveillance, Epidemiology, and End Results (SEER) data showed that African Americans were 64% less likely to receive radical prostatectomy than Caucasians for localized PCa.9 For localized and regional disease stages, Caucasians are more likely to receive prostatectomy, whereas African Americans are more likely to receive radiation.9, 10, 28 Although our results are in accord with earlier results, after controlling for patient-level covariates, not ethnicity, but age and marital status were predictors of radical prostatectomy treatment (data not shown). Thus, most of the observed ethnic variation in treatment may be attributable to patient and provider characteristics.

Our log-linear regression demonstrated that ethnicity was an independent predictor of 12-month scores for some generic and prostate-specific HRQoL subscales. By using the Cancer of the Prostate Strategic Urologic Research Endeavor (CaPSURE) database, Lubeck et al demonstrated significant differences in clinical presentation, sociodemographics, and HRQoL between black and white PCa patients. The HRQoL differences persisted at 1-year post-treatment.8 Among men receiving radical prostatectomy, African Americans had poorer outcomes in PSA level than Caucasians.18 In a PCa outcome study, Johnson et al found that among prostatectomy patients, African Americans reported higher sexual and urinary function at 5 years postdiagnosis than Caucasians. However, the ethnic difference in recovery of sexual and bowel function among radiation therapy patients was similar.11 Differences in study design and analytical methods may explain differences with the results found in the present study. In the Lubeck et al study, baseline Gleason score showed significant variation across race or ethnicity. Participants were asked to complete a baseline assessment retrospectively at 6 months post-treatment. Also, generic HRQoL was not reported, and comorbidity was not statistically controlled in the analysis. For some HRQoL subscales, African-American elderly may take more time to recover as shown in our study. Over a 12-month period, African-American elderly took a significantly longer time to recover to baseline for generic (physical function, role emotional, mental health, social function, bodily pain, and general health) and prostate specific (sexual bother) HRQoL. In a prospective study, Knight et al observed similarities in preferences, optimism, involvement in care and, similar to our results, found differences in quality of life measures between black and white veterans.31 Additionally, studies have shown that the effect of ethnicity on HRQoL outcomes is modified by PCa patients' health behaviors and self-efficacy.32, 33 As with HRQoL, satisfaction is an important measure of care. Similar to findings of earlier researchers,34 we observed a high level of satisfaction with care in both ethnic groups.

Our study contained some limitations. Because the sample was selected from clinical and hospital settings affiliated with a tertiary care academic medical center, the patients may not represent patients receiving care in hospitals without a comparable affiliation. The care of patients in a Veterans Administration and in a private university hospital may differ. However, we have adjusted for hospital type in our analysis. Because the follow-up period was 12 months, long-term HRQoL changes remain unknown.

In conclusion, we observed that curative treatments for newly diagnosed PCa had differential generic and prostate-specific HRQoL outcomes across ethnicity. African-American elderly were more likely to take longer to return to their baseline function. Also, a higher percentage of them did not return to their baseline function by 12 months. African-American elderly showed a significant variation in baseline characteristics, generic and prostate-specific HRQoL compared with Caucasian elderly. These baseline differences, although statistically controlled in the multivariate model, may still reflect differences in comorbidities that influence functional status outcomes. Moreover, differences in education, income, and marital status may result in less social support, culminating in loss of function in generic measures but not in prostate-specific measures, with the exception of sexual bother. The latter may reflect an attitudinal difference or difference in expectations, especially because sexual function returned to baseline, whereas sexual bother did not. The appearance of most differences at 6 months or later is consistent with the social support hypothesis, rather than differences in biology or acute care, which would be expected to be seen sooner.

These findings have important implications for effective management and counseling of elderly PCa patients from different race or ethnic groups and merit further research to explain mediating factors and design interventions. These differences highlight the need for pretreatment discussions with patients and families about quality of life expectations post-treatment. In particular, healthcare providers should assertively discuss and thoroughly understand patient expectations of sexual performance after prostate cancer treatment. All patients should be advised that recovery of functional status will often require 5 months or longer, although the majority will return to baseline function. We are not recommending differential advising by ethnicity, as ethnicity is not a substitute for individual characteristics. However, awareness of possible differences across ethnic groups may cause physicians to be alert for additional discussion and services. Additionally, more attention should be given to design and implementation of support services and family counseling beyond a few months during the post-treatment period. Whether these services will be effective in improving and reducing variations in outcomes will require additional research.

Ancillary