International Journal of
Radiation Oncology biology
physics
www.redjournal.org
Clinical Investigation: Genitourinary Cancer
Receipt of Guideline-Concordant Treatment in Elderly Prostate Cancer Patients Ronald C. Chen, MD, MPH,*,y,z William R. Carpenter, PhD,y,z,x Laura H. Hendrix, MS,* John Bainbridge, MS,y Andrew Z. Wang, MD,*,z Matthew E. Nielsen, MD, MS,y,z,k and Paul A. Godley, MD, PhDy,z,{ *Department of Radiation Oncology, ySheps Center for Health Services Research, zLineberger Comprehensive Cancer Center, xGillings School of Global Public Health, kDepartment of Urology, and {Division of Hematology-Oncology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina Received Aug 6, 2013, and in revised form Oct 31, 2013. Accepted for publication Nov 4, 2013.
Summary We examined treatments received by elderly prostate cancer patients using the Surveillance, Epidemiology and End Results-Medicare linked database. One-third to one-half of high-risk patients received treatment discordant with the National Comprehensive Cancer Network guidelines. Discordance was high even in patients with minimal comorbidities and a >10-year life expectancy. This study demonstrates an undertreatment of elderly patients with aggressive prostate cancer.
Purpose: To examine the proportion of elderly prostate cancer patients receiving guidelineconcordant treatment, using the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked database. Methods and Materials: A total of 29,001 men diagnosed in 2004-2007 with localized prostate cancer, aged 66 to 79 years, were included. We characterized the proportion of men who received treatment concordant with the National Comprehensive Cancer Network guidelines, stratified by risk group and age. Logistic regression was used to examine covariates associated with receipt of guideline-concordant management. Results: Guideline concordance was 79%-89% for patients with low- or intermediate-risk disease. Among high-risk patients, 66.6% of those aged 66-69 years received guideline-concordant management, compared with 51.9% of those aged 75-79 years. Discordance was mainly due to conservative managementdno treatment or hormone therapy alone. Among the subgroup of patients aged 76 years with no measured comorbidity, findings were similar. On multivariable analysis, older age (75-79 vs 66-69 years, odds ratio 0.51, 95% confidence interval 0.50-0.57) was associated with a lower likelihood of guideline concordance for high-risk prostate cancer, but comorbidity was not. Conclusions: There is undertreatment of elderly but healthy patients with high-risk prostate cancer, the most aggressive form of this disease. Ó 2014 Elsevier Inc.
Reprint requests to: Ronald C. Chen, MD, MPH, University of North Carolina at Chapel Hill, Department of Radiation Oncology, 101 Manning Dr, CB #7512, Chapel Hill, NC 27516. Tel: (919) 966-0400; E-mail: [email protected] Presented as an oral presentation at the Annual Meeting of the American Society of Clinical Oncology, June 4-8, 2010, Chicago, IL. Conflict of interest: none. Int J Radiation Oncol Biol Phys, Vol. 88, No. 2, pp. 332e338, 2014 0360-3016/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ijrobp.2013.11.004
Supplementary material for this article can be found at www.redjournal.org AcknowledgmentdThe authors thank the Centers for Medicare and Medicaid Services and the SEER program tumor registries for the creation of the SEER-Medicare database. The interpretation and reporting of these data are the sole responsibility of the authors.
Volume 88 Number 2 2014
Introduction Prostate cancer is the most common malignancy in men and the second leading cause of cancer deaths (1). With median age of diagnosis at 67 years, this is a prevalent disease for the Medicare population. There is significant controversy regarding the potential benefit of prostate cancer screening and treatment of elderly patients with indolent disease (2). However, although the majority of elderly patients with low-risk prostate cancer will die from noncancer-related causes, high-risk cancer is much more lifethreatening and represents the most aggressive form of this disease. In a large population-based study, Albertsen et al (3) reported that with conservative management (observation or hormone therapy alone), the majority of patients with high-risk disease will die from prostate cancer and not from competing causes (3). At 10 years, 60% of patients with high-risk disease diagnosed at age 65-69 years died from prostate cancer, 20% died from competing causes, and 20% were still alive. Further, it is in this group of patients where the available data (ie, randomized trials) provide the strongest level of evidence supporting definitive treatment (4, 5). Treatment patterns in elderly patients with high-risk prostate cancer are not well described and may be important because, in other cancers, elderly cancer patients are less likely to receive curative treatment; even when treatment is given, elderly patients are more likely to receive less-aggressive therapy. Prior population-based studies have documented these trends in bladder, breast, and colorectal cancers, demonstrating potential undertreatment of elderly patients with aggressive cancers (6-9), perhaps due to concerns about lack of benefit or decreased treatment tolerance. Therefore, the goal of this study was to describe the types of initial treatments received by prostate cancer patients on a population level using the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked database and to assess treatment concordance with the published National Comprehensive Cancer Network (NCCN) guidelines. We hypothesized that a significant proportion of elderly prostate cancer patients receive minimal or no treatment, even those with minimal comorbidities and high-risk disease.
Methods and Materials Data source The SEER-Medicare data have been described extensively elsewhere (10). Briefly, SEER is composed of 17 population-based cancer registries representing approximately 28% of the US population. The SEER registry data are linked to Medicare administrative and health care claims data, which cover 97% of Americans aged 65 years and older. This study received a waiver from the institutional review board.
Study cohorts We included African American and Caucasian patients aged 66-79 years diagnosed with clinically localized prostate cancer between 2004 and 2007, who had no additional cancers and had month and year of diagnosis in the database, and we excluded those diagnosed at autopsy. Analysis was restricted to African American and Caucasian patients owing to very small numbers in other racial
Guideline concordance in prostate cancer
333
groups. Prostate-specific antigen (PSA) and Gleason score information were not available in SEER-Medicare for those diagnosed before 2004, precluding risk stratification. To allow assessment of baseline comorbidity, the patient population was restricted to men with at least 12 months of claims data before diagnosis (11). Further, to ensure complete capture of health services, we excluded men who did not have complete claims because they were enrolled in a health maintenance organization or were not enrolled in both Medicare Part A and Part B for at least 12 months after prostate cancer diagnosis. This resulted in 41,311 patients potentially eligible for inclusion. We further excluded patients with locally advanced (clinical stage T3b-T4) disease or with metastasis to the lymph nodes (N1) (nZ3923), because the ability to cure these advanced cancers is limited. We also excluded patients with missing information for risk stratification (nZ7746), or missing treatment information or error (eg, treatment claims occurred before diagnosis date, indicating data error, nZ641). This resulted in an analytic cohort of 29,001 patients.
Outcome and control variables Using Current Procedural Terminology/Healthcare Common Procedure Coding System procedure codes and Medicare claims data, we identified treatments received within 12 months of diagnosis, including prostatectomy, external beam radiation therapy (RT), brachytherapy, and hormone therapy (Table e1, available online). Patients who received none of these treatments within 12 months were classified as undergoing initial “no treatment.” Treatment concordance was determined by comparison with the NCCN guidelines in 2004-2007, which for purposes of this study did not change significantly during these years. The SEER registry provided demographic variables including race, age and year at diagnosis, and marital status; census tract measure of education; SEER region; population density (metropolitan vs. nonmetropolitan); and diagnostic information to allow classification of patients into D’Amico risk groups (12). Medicare data provide information regarding enrollment, including Medicaid dual-eligibility as an individual-level socioeconomic status indicator. Claims from the 12 months before prostate cancer diagnosis were used to calculate a claims-specific validated comorbidity index score derived from the Charlson Index (11).
Statistical analysis We described the initial management received by patients stratified by risk group and age. Then, in each stratum, the proportion of patients receiving management concordant with the NCCN guidelines was calculated. Differences in guideline concordance among the age groups were assessed using the Fisher exact test. We further examined guideline concordance in a subgroup of patients who were aged 76 years and had no comorbidity. The NCCN guidelines recommends use of the Social Security Administration’s actuarial Period Life Tables to estimate life expectancy, then modifying this estimate on the basis of an assessment of overall health. On a population level in the United States, men aged 76 years have more than 10 years life expectancy (13). On the basis of this and following NCCN recommendations, we created the subgroup of patients in this age group
334
International Journal of Radiation Oncology Biology Physics
Chen et al.
Table 1
Results
Demographic and clinical characteristics Characteristic
Age (y) 66-69 70-74 75-79 Race Caucasian African American National Cancer Institute comorbidity index score 0 >0 Marital status Married Unmarried/unknown Diagnosis year 2004 2005 2006 2007 Geographic region Northeast South Midwest West Proportion of non-high school graduates in census tract (%) Quartile 1 (worst) Quartile 2 Quartile 3 Quartile 4 (best) Medicaid dual eligibility No Yes Metropolitan residence Yes No Prostate specific antigen (ng/mL) 40 Missing Prostate cancer risk category Low Intermediate High
n (%) 9801 (34) 11,201 (39) 7999 (28) 25,752 (89) 3249 (11)
19,638 (68) 9363 (32) 20,975 (72) 8026 (28) 7151 6954 7484 7412
(25) (24) (26) (26)
7000 5996 3452 12,553
(24) (21) (12) (43)
7280 5930 6645 9146
(25) (20) (23) (32)
26,515 (91) 2486 (9) 24,522 (85) 4479 (15) 20,705 4449 1555 1399 893
(71) (15) (5) (5) (3)
9499 (33) 9548 (33) 9954 (34)
with no comorbiditydwho are most likely to have at least 10 years life expectancy. Multivariable logistic regression analysis was used to examine potential covariates associated with receipt of concordant management in each risk group. Covariates included age category (66-69, 70-74, 75-79 years), comorbidity score (stratified by the median value of 0), race, diagnosis year, marital status, SEER region, Medicaid eligibility, census-tract education, and metropolitan residence. Statistical significance was set at .05; all tests were 2-tailed. Analyses were performed using SAS, version 9.2 (SAS Institute, Cary, NC).
Patients were well distributed in the age and risk groups (Table 1). Eleven percent of the men were African American, and 72% were married. Figure 1 and Table 2 summarize the management received by patients in each risk category, stratified by age group. Overall, 24% of patients received treatment discordant with guidelines. For low- and intermediate-risk disease, the proportions of patients receiving guideline-concordant management were relatively high (79.1%-89.0%), although concordance decreased in older age groups (P10-year life expectancy. To our knowledge this is the first population-based study to examine guideline concordance in prostate cancer management in Medicare patients, stratified by prostate cancer risk category. SEER-Medicare data before 2004 did not contain the necessary elements for risk stratification. Risk stratification is an important part of this study because it allows for comparison of treatments received with published guidelines and places our findings in a clinical context of cancer aggressiveness. Prostate cancer is often considered an indolent disease, and recent studies have suggested that there is overtreatment of patients with low-risk, and potentially clinically insignificant, prostate cancer (3, 14). Although the majority of patients with low-risk prostate cancer will die from non-cancer-related causes, high-risk disease is much more life-threatening (3). Therefore, definitive treatment for high-risk prostate cancer has the most potential for benefit. It is also in the high-risk patients that the majority of prospective clinical trials have focused, providing guidance on management of these patients.
Volume 88 Number 2 2014
Fig. 1.
Guideline concordance in prostate cancer
Concordant and discordant management received by patients in each risk category, stratified by age group. the largest in all of oncology. In the subgroup of patients with PSA >20 ng/mL (high-risk cancer), the absolute benefit from RT in cancer-specific survival was 17.3%. Importantly, this trial showed
In one trial that compared hormone therapy alone vs. radiation therapy (RT), patients who received RT achieved an absolute overall survival benefit of 10% (5). This absolute benefit is one of Table 2
335
Management received by patients in each risk category, stratified by age group 66-69 y
70-74 y
Risk category
Discordant
Concordant
Discordant
Concordant
Low risk (nZ9499) No treatment HT only Brachytherapy RT RT þ HT Prostatectomy Prostatectomy þ HT Intermediate risk (nZ9548) No treatment HT only Brachytherapy RT RT þ HT Prostatectomy Prostatectomy þ HT High Risk (NZ9954) No treatment HT only Brachytherapy RT RT þ HT Prostatectomy Prostatectomy þ HT
388 (11.0)
3148 (89.0) 594 (16.8)
612 (16.3)
3135 (83.7) 765 (20.4)
76 (2.2)
277 (7.8)
473 (12.6)
2853 (86.5) 365 (11.1)
110 (3.3) 254 (7.7)
16 (0.4) 564 (15.5)
(2.4) (33.4) (11.6) (7.6) (4.4) (5.9)
119 (4.0)
3074 (84.5) 445 (12.2)
138 (4.7)y 869 (29.3) 971 (32.7)
68 (3.1)
Radiation Oncology biology
physics
www.redjournal.org
Clinical Investigation: Genitourinary Cancer
Receipt of Guideline-Concordant Treatment in Elderly Prostate Cancer Patients Ronald C. Chen, MD, MPH,*,y,z William R. Carpenter, PhD,y,z,x Laura H. Hendrix, MS,* John Bainbridge, MS,y Andrew Z. Wang, MD,*,z Matthew E. Nielsen, MD, MS,y,z,k and Paul A. Godley, MD, PhDy,z,{ *Department of Radiation Oncology, ySheps Center for Health Services Research, zLineberger Comprehensive Cancer Center, xGillings School of Global Public Health, kDepartment of Urology, and {Division of Hematology-Oncology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina Received Aug 6, 2013, and in revised form Oct 31, 2013. Accepted for publication Nov 4, 2013.
Summary We examined treatments received by elderly prostate cancer patients using the Surveillance, Epidemiology and End Results-Medicare linked database. One-third to one-half of high-risk patients received treatment discordant with the National Comprehensive Cancer Network guidelines. Discordance was high even in patients with minimal comorbidities and a >10-year life expectancy. This study demonstrates an undertreatment of elderly patients with aggressive prostate cancer.
Purpose: To examine the proportion of elderly prostate cancer patients receiving guidelineconcordant treatment, using the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked database. Methods and Materials: A total of 29,001 men diagnosed in 2004-2007 with localized prostate cancer, aged 66 to 79 years, were included. We characterized the proportion of men who received treatment concordant with the National Comprehensive Cancer Network guidelines, stratified by risk group and age. Logistic regression was used to examine covariates associated with receipt of guideline-concordant management. Results: Guideline concordance was 79%-89% for patients with low- or intermediate-risk disease. Among high-risk patients, 66.6% of those aged 66-69 years received guideline-concordant management, compared with 51.9% of those aged 75-79 years. Discordance was mainly due to conservative managementdno treatment or hormone therapy alone. Among the subgroup of patients aged 76 years with no measured comorbidity, findings were similar. On multivariable analysis, older age (75-79 vs 66-69 years, odds ratio 0.51, 95% confidence interval 0.50-0.57) was associated with a lower likelihood of guideline concordance for high-risk prostate cancer, but comorbidity was not. Conclusions: There is undertreatment of elderly but healthy patients with high-risk prostate cancer, the most aggressive form of this disease. Ó 2014 Elsevier Inc.
Reprint requests to: Ronald C. Chen, MD, MPH, University of North Carolina at Chapel Hill, Department of Radiation Oncology, 101 Manning Dr, CB #7512, Chapel Hill, NC 27516. Tel: (919) 966-0400; E-mail: [email protected] Presented as an oral presentation at the Annual Meeting of the American Society of Clinical Oncology, June 4-8, 2010, Chicago, IL. Conflict of interest: none. Int J Radiation Oncol Biol Phys, Vol. 88, No. 2, pp. 332e338, 2014 0360-3016/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ijrobp.2013.11.004
Supplementary material for this article can be found at www.redjournal.org AcknowledgmentdThe authors thank the Centers for Medicare and Medicaid Services and the SEER program tumor registries for the creation of the SEER-Medicare database. The interpretation and reporting of these data are the sole responsibility of the authors.
Volume 88 Number 2 2014
Introduction Prostate cancer is the most common malignancy in men and the second leading cause of cancer deaths (1). With median age of diagnosis at 67 years, this is a prevalent disease for the Medicare population. There is significant controversy regarding the potential benefit of prostate cancer screening and treatment of elderly patients with indolent disease (2). However, although the majority of elderly patients with low-risk prostate cancer will die from noncancer-related causes, high-risk cancer is much more lifethreatening and represents the most aggressive form of this disease. In a large population-based study, Albertsen et al (3) reported that with conservative management (observation or hormone therapy alone), the majority of patients with high-risk disease will die from prostate cancer and not from competing causes (3). At 10 years, 60% of patients with high-risk disease diagnosed at age 65-69 years died from prostate cancer, 20% died from competing causes, and 20% were still alive. Further, it is in this group of patients where the available data (ie, randomized trials) provide the strongest level of evidence supporting definitive treatment (4, 5). Treatment patterns in elderly patients with high-risk prostate cancer are not well described and may be important because, in other cancers, elderly cancer patients are less likely to receive curative treatment; even when treatment is given, elderly patients are more likely to receive less-aggressive therapy. Prior population-based studies have documented these trends in bladder, breast, and colorectal cancers, demonstrating potential undertreatment of elderly patients with aggressive cancers (6-9), perhaps due to concerns about lack of benefit or decreased treatment tolerance. Therefore, the goal of this study was to describe the types of initial treatments received by prostate cancer patients on a population level using the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked database and to assess treatment concordance with the published National Comprehensive Cancer Network (NCCN) guidelines. We hypothesized that a significant proportion of elderly prostate cancer patients receive minimal or no treatment, even those with minimal comorbidities and high-risk disease.
Methods and Materials Data source The SEER-Medicare data have been described extensively elsewhere (10). Briefly, SEER is composed of 17 population-based cancer registries representing approximately 28% of the US population. The SEER registry data are linked to Medicare administrative and health care claims data, which cover 97% of Americans aged 65 years and older. This study received a waiver from the institutional review board.
Study cohorts We included African American and Caucasian patients aged 66-79 years diagnosed with clinically localized prostate cancer between 2004 and 2007, who had no additional cancers and had month and year of diagnosis in the database, and we excluded those diagnosed at autopsy. Analysis was restricted to African American and Caucasian patients owing to very small numbers in other racial
Guideline concordance in prostate cancer
333
groups. Prostate-specific antigen (PSA) and Gleason score information were not available in SEER-Medicare for those diagnosed before 2004, precluding risk stratification. To allow assessment of baseline comorbidity, the patient population was restricted to men with at least 12 months of claims data before diagnosis (11). Further, to ensure complete capture of health services, we excluded men who did not have complete claims because they were enrolled in a health maintenance organization or were not enrolled in both Medicare Part A and Part B for at least 12 months after prostate cancer diagnosis. This resulted in 41,311 patients potentially eligible for inclusion. We further excluded patients with locally advanced (clinical stage T3b-T4) disease or with metastasis to the lymph nodes (N1) (nZ3923), because the ability to cure these advanced cancers is limited. We also excluded patients with missing information for risk stratification (nZ7746), or missing treatment information or error (eg, treatment claims occurred before diagnosis date, indicating data error, nZ641). This resulted in an analytic cohort of 29,001 patients.
Outcome and control variables Using Current Procedural Terminology/Healthcare Common Procedure Coding System procedure codes and Medicare claims data, we identified treatments received within 12 months of diagnosis, including prostatectomy, external beam radiation therapy (RT), brachytherapy, and hormone therapy (Table e1, available online). Patients who received none of these treatments within 12 months were classified as undergoing initial “no treatment.” Treatment concordance was determined by comparison with the NCCN guidelines in 2004-2007, which for purposes of this study did not change significantly during these years. The SEER registry provided demographic variables including race, age and year at diagnosis, and marital status; census tract measure of education; SEER region; population density (metropolitan vs. nonmetropolitan); and diagnostic information to allow classification of patients into D’Amico risk groups (12). Medicare data provide information regarding enrollment, including Medicaid dual-eligibility as an individual-level socioeconomic status indicator. Claims from the 12 months before prostate cancer diagnosis were used to calculate a claims-specific validated comorbidity index score derived from the Charlson Index (11).
Statistical analysis We described the initial management received by patients stratified by risk group and age. Then, in each stratum, the proportion of patients receiving management concordant with the NCCN guidelines was calculated. Differences in guideline concordance among the age groups were assessed using the Fisher exact test. We further examined guideline concordance in a subgroup of patients who were aged 76 years and had no comorbidity. The NCCN guidelines recommends use of the Social Security Administration’s actuarial Period Life Tables to estimate life expectancy, then modifying this estimate on the basis of an assessment of overall health. On a population level in the United States, men aged 76 years have more than 10 years life expectancy (13). On the basis of this and following NCCN recommendations, we created the subgroup of patients in this age group
334
International Journal of Radiation Oncology Biology Physics
Chen et al.
Table 1
Results
Demographic and clinical characteristics Characteristic
Age (y) 66-69 70-74 75-79 Race Caucasian African American National Cancer Institute comorbidity index score 0 >0 Marital status Married Unmarried/unknown Diagnosis year 2004 2005 2006 2007 Geographic region Northeast South Midwest West Proportion of non-high school graduates in census tract (%) Quartile 1 (worst) Quartile 2 Quartile 3 Quartile 4 (best) Medicaid dual eligibility No Yes Metropolitan residence Yes No Prostate specific antigen (ng/mL) 40 Missing Prostate cancer risk category Low Intermediate High
n (%) 9801 (34) 11,201 (39) 7999 (28) 25,752 (89) 3249 (11)
19,638 (68) 9363 (32) 20,975 (72) 8026 (28) 7151 6954 7484 7412
(25) (24) (26) (26)
7000 5996 3452 12,553
(24) (21) (12) (43)
7280 5930 6645 9146
(25) (20) (23) (32)
26,515 (91) 2486 (9) 24,522 (85) 4479 (15) 20,705 4449 1555 1399 893
(71) (15) (5) (5) (3)
9499 (33) 9548 (33) 9954 (34)
with no comorbiditydwho are most likely to have at least 10 years life expectancy. Multivariable logistic regression analysis was used to examine potential covariates associated with receipt of concordant management in each risk group. Covariates included age category (66-69, 70-74, 75-79 years), comorbidity score (stratified by the median value of 0), race, diagnosis year, marital status, SEER region, Medicaid eligibility, census-tract education, and metropolitan residence. Statistical significance was set at .05; all tests were 2-tailed. Analyses were performed using SAS, version 9.2 (SAS Institute, Cary, NC).
Patients were well distributed in the age and risk groups (Table 1). Eleven percent of the men were African American, and 72% were married. Figure 1 and Table 2 summarize the management received by patients in each risk category, stratified by age group. Overall, 24% of patients received treatment discordant with guidelines. For low- and intermediate-risk disease, the proportions of patients receiving guideline-concordant management were relatively high (79.1%-89.0%), although concordance decreased in older age groups (P10-year life expectancy. To our knowledge this is the first population-based study to examine guideline concordance in prostate cancer management in Medicare patients, stratified by prostate cancer risk category. SEER-Medicare data before 2004 did not contain the necessary elements for risk stratification. Risk stratification is an important part of this study because it allows for comparison of treatments received with published guidelines and places our findings in a clinical context of cancer aggressiveness. Prostate cancer is often considered an indolent disease, and recent studies have suggested that there is overtreatment of patients with low-risk, and potentially clinically insignificant, prostate cancer (3, 14). Although the majority of patients with low-risk prostate cancer will die from non-cancer-related causes, high-risk disease is much more life-threatening (3). Therefore, definitive treatment for high-risk prostate cancer has the most potential for benefit. It is also in the high-risk patients that the majority of prospective clinical trials have focused, providing guidance on management of these patients.
Volume 88 Number 2 2014
Fig. 1.
Guideline concordance in prostate cancer
Concordant and discordant management received by patients in each risk category, stratified by age group. the largest in all of oncology. In the subgroup of patients with PSA >20 ng/mL (high-risk cancer), the absolute benefit from RT in cancer-specific survival was 17.3%. Importantly, this trial showed
In one trial that compared hormone therapy alone vs. radiation therapy (RT), patients who received RT achieved an absolute overall survival benefit of 10% (5). This absolute benefit is one of Table 2
335
Management received by patients in each risk category, stratified by age group 66-69 y
70-74 y
Risk category
Discordant
Concordant
Discordant
Concordant
Low risk (nZ9499) No treatment HT only Brachytherapy RT RT þ HT Prostatectomy Prostatectomy þ HT Intermediate risk (nZ9548) No treatment HT only Brachytherapy RT RT þ HT Prostatectomy Prostatectomy þ HT High Risk (NZ9954) No treatment HT only Brachytherapy RT RT þ HT Prostatectomy Prostatectomy þ HT
388 (11.0)
3148 (89.0) 594 (16.8)
612 (16.3)
3135 (83.7) 765 (20.4)
76 (2.2)
277 (7.8)
473 (12.6)
2853 (86.5) 365 (11.1)
110 (3.3) 254 (7.7)
16 (0.4) 564 (15.5)
(2.4) (33.4) (11.6) (7.6) (4.4) (5.9)
119 (4.0)
3074 (84.5) 445 (12.2)
138 (4.7)y 869 (29.3) 971 (32.7)
68 (3.1)
