• prostatectomy;
  • adjuvant therapy;
  • utilization;
  • expenditures;
  • outcomes


  1. Top of page
  2. Abstract


We sought to identify the costs of adjuvant therapies following radical prostatectomy (RP) and factors associated with their receipt.


We used SEER-Medicare data from 2004-2006 to identify 4247 men who underwent RP, of whom 600 subsequently received adjuvant therapies. We used Cox regression to identify factors associated with receipt of adjuvant therapies. Health care expenditures within 12 months of diagnosis were compared for RP alone versus RP with adjuvant therapies.


Biopsy Gleason score, prostate-specific antigen, risk group, and SEER region were significantly associated with receipt of adjuvant treatments (all P<.001). Higher surgeon volume was associated with lower odds of receiving adjuvant therapies (hazard ratio [HR], 0.60; 95% confidence interval [CI], 0.46-0.78 [P<.001]). Factors associated with increased receipt of adjuvant therapies were positive surgical margins (HR, 3.02; 95% CI, 2.55-3.57 [P<.001]), high-risk group versus low-risk group (HR, 7.65; 95% CI, 5.64-10.37 [P<.001]), lymph node–positive disease (HR, 5.36; 95% CI, 3.71-7.75 [P<.001]), and treatment in Iowa (HR, 1.93; 95% CI, 1.12-3.32 [P = .019]) and New Mexico/Georgia/Hawaii (HR, 1.92; 95% CI, 1.09-3.39 [P = .025]) versus San Francisco SEER regions (baseline). Age, race, comorbidities, and surgical approach were not associated with use of adjuvant therapies. The median expenditures attributable to postprostatectomy hormonal therapy, radiation therapy, and radiation with hormonal therapy versus were $1361, $12,040, and $23,487.


Men treated by high-volume surgeons were less likely to receive adjuvant therapies. Regional variation and high-risk disease characteristics were associated with increased receipt of adjuvant therapies, which increased health care expenditures by 2- to 3-fold when radiotherapy was administered. Cancer 2011. © 2011 American Cancer Society.

Prostate cancer remains the most commonly diagnosed solid organ tumor among men in the United States, with approximately 192,000 incident cases in 2009.1 The majority of these tumors are localized, and radical prostatectomy (RP) remains the most popular treatment option.2 However, 21%-37% of men experience biochemical recurrence (BCR) after radical prostatectomy.3 Recent studies have shown that postprostatectomy radiotherapy improves prostate cancer–specific survival4 and significantly decreases overall mortality when used in the adjuvant5 or salvage setting in selected men with high-risk disease.6 Furthermore, the benefit of hormonal therapy needs to be carefully balanced against the significant inherent risks of cardiovascular and thromboembolic disease, along with the substantial health care costs of implementing this treatment.7-9 Hormonal therapy as it pertains to the adjuvant setting, either alone or in combination with radiotherapy, has been less extensively evaluated, with no definitive guidelines on who should receive treatment or when to initiate it.8, 9

Although there are few contemporary characterizations of secondary therapies,6, 10, 11 a study of Medicare beneficiaries from the early 1990s demonstrated that 35% of men receive secondary therapies following RP.12 However, this may not reflect contemporary practice patterns due to the downward stage migration that followed the advent of prostate-specific antigen (PSA) screening.13 The purpose of our population-based study was to evaluate factors associated with the use of adjuvant cancer therapies following RP and estimate the associated health care expenditures of these treatments.


  1. Top of page
  2. Abstract

The study was approved by the Brigham and Women's Institutional Review Board. Patient data were de-identified and the requirement for consent was waived. We used Surveillance, Epidemiology, and End Results (SEER)-Medicare data for analysis, which comprise a linkage of population-based cancer registry data from 16 SEER areas with Medicare administrative data and cover approximately 26% of the United States population. The Medicare program provides benefits to 97% of Americans aged ≥65 years.14

Study Cohort

We identified 4247 men aged ≥65 years who were diagnosed with prostate cancer in 2004 and 2005 and underwent RP through 2006 based on the Physician's Current Procedural Terminology Coding System, 4th edition, (CPT-4): codes 55840, 55842, 55845 for open RP and code 55866 for minimally invasive RP. CPT-4 code 55899 (unspecified male genitourinary procedure) may sometimes be used with an open RP administrative code to specify minimally invasive RP with robotic assistance for private health plans,15 but Medicare does not recognize this coding schema, and very few men had this combination of codes; therefore, this schema was not used to identify minimally invasive RP. We excluded men not enrolled in both Medicare Part A and B, or who were enrolled in a Medicare health maintenance organization (because their claims are not reliably submitted). Because SEER only captures positive margin characteristics for American Joint Commission on Cancer pathological T2 and T3a disease, we excluded 292 men with pathological stage T3b, 63 men with pathological stage T4, and 412 men with missing margin status from our cohort. Patients with lymph node–positive disease (n = 45) were included in the study. In addition, to increase the sensitivity for detecting additional postoperative radiation therapy, we restricted our cohort to patients with prostate cancer diagnosed as their only cancer. A total of 204 patients with other cancers, including nonmelanoma skin cancers, were excluded from the analysis.


We examined the utilization of adjuvant therapy (radiation and/or hormonal) after RP in patients with pathological T2 and T3a disease.12, 16 According to the American Urological Association 2007 guidelines, additional radiation and/or hormonal therapy should be administered to patients with adverse pathological features and/or positive surgical margins.17

Control Variables

Age was obtained from each patient's Medicare file; race, census tract measures of median household income and high school education, region, population density (urban vs rural), and marital status were obtained from SEER registry data. Comorbidity was assessed using the Klabunde modification of the Charlson index during the year before surgery.18 The Klabunde modification uses comorbid conditions identified by the Charlson comorbidity index and incorporates the diagnostic and procedure data contained in Medicare physician (Part B) claims. Variables were categorized as in Table 1. Additionally, we used PSA, Gleason grade, and clinical stage to stratify men to low, intermediate, and high-risk disease.19 However, tumor stage was missing/unknown for almost one-third of our patients, and we therefore used a modified risk stratification without clinical stage, resulting in a low-risk designation for 29% of our cohort. Therefore, we used a modified risk classification defined as follows: PSA <10 and biopsy Gleason score <7 = low; PSA 10-20 or Gleason score 7 = intermediate; PSA >20 or Gleason score >7 = high.

Table 1. Demographics of the Study Population
CharacteristicCategoriesTotalNo Adjuvant TherapyHormonal or RadiationPHormonal TherapyRadiation Therapy
  1. SEER indicates Surveillance, Epidemiology, and End Results; PSA, prostate-specific antigen.

  2. Data are presented as No. (%).

Year of surgery200417791503 (84.49)275 (15.46).028138 (7.76)221 (12.42)
 200520581776 (86.30)282 (13.70) 139 (6.75)214 (10.40)
 2006410367 (89.51)43 (10.49) 17 (4.15)39 (9.51)
Age (years)65-6926202240 (85.50)379 (14.47).624177 (6.76)310 (11.83)
 70-7413321154 (86.64)178 (13.36) 95 (7.13)137 (10.29)
 ≥75295252 (85.42)43 (14.58) 22 (7.46)27 (9.15)
Charlson comorbidity index029562543 (86.03)413 (13.97).501194 (6.56)343 (11.60)
 11018865 (84.97)153 (15.03) 85 (8.35)106 (10.41)
 ≥2273238 (87.18)34 (12.45) 15 (5.49)25 (9.16)
RaceWhite33662893 (85.95)473 (14.05).328226 (6.71)384 (11.41)
 Black307265 (86.32)42 (13.68) 22 (7.17)31 (10.10)
 Hispanic356310 (87.08)45 (12.64) 25 (7.02)30 (8.43)
 Asian186150 (80.65)36 (19.35) 20 (10.75)25 (13.44)
 Other3228 (87.50)4 (12.50) 1 (3.13)4 (12.50)
Marital statusUnmarried605523 (86.45)82 (13.55).63235 (5.79)67 (11.07)
 Married34692971 (85.64)497 (14.33) 247 (7.12)393 (11.33)
Education: % of census tract with at least a high school degree<75785672 (85.61)112 (14.27).07457 (7.26)83 (10.57)
 75-84.99785682 (86.88)103 (13.12) 60 (7.64)76 (9.68)
 85-89.99791656 (82.93)135 (17.07) 63 (7.96)113 (14.29)
 ≥9018851635 (86.74)250 (13.26) 114 (6.05)202 (10.72)
Median income in census tract of residence<$35,0001106938 (84.81)168 (15.19).36788 (7.96)124 (11.21)
 $35,000-44,000975842 (86.36)132 (13.54) 64 (6.56)102 (10.46)
 $45,000-59,0001072912 (85.07)160 (14.93) 76 (7.09)134 (12.50)
 ≥$60,0001093953 (87.19)140 (12.81) 66 (6.04)114 (10.43)
SEER regionSan Francisco171151 (88.30)20 (11.70).02910 (5.85)17 (9.94)
 Detroit303273 (90.10)30 (9.90) 16 (5.28)20 (6.60)
 Iowa195156 (80.00)39 (20.00) 20 (10.26)30 (15.38)
 Seattle352312 (88.64)40 (11.36) 19 (5.40)33 (9.38)
 Utah284255 (89.79)29 (10.21) 8 (2.82)24 (8.45)
 Connecticut127108 (85.04)19 (14.96) 10 (7.87)18 (14.17)
 San Jose10382 (79.61)21 (20.39) 8 (7.77)18 (17.48)
 Los Angeles569496 (87.17)73 (12.83) 38 (6.68)51 (8.96)
 Greater California1171987 (84.29)183 (15.63) 89 (7.60)149 (12.72)
 Kentucky215181 (84.19)34 (15.81) 18 (8.37)28 (13.02)
 Louisiana316276 (87.34)40 (12.66) 23 (7.28)29 (9.18)
 New Jersey265226 (85.28)39 (14.72) 20 (7.55)32 (12.08)
 New Mexico/Georgia /Hawaii176143 (81.25)33 (18.75) 15 (8.52)25 (14.20)
Population densityMetropolitan39893430 (85.99)558 (13.99).292271 (6.79)443 (11.11)
 Rural258216 (83.72)42 (16.28) 23 (8.91)31 (12.02)
Clinical stageT1c22181938 (87.38)279 (12.58)<.001133 (6.00)224 (10.10)
 T2737619 (83.99)118 (16.01) 60 (8.14)94 (12.75)
 T33922 (56.41)17 (43.59) 13 (33.33)10 (25.64)
Gleason grade≤616871599 (94.78)88 (5.22)<.00129 (1.72)71 (4.21)
 720731752 (84.52)320 (15.44) 143 (6.90)259 (12.49)
 ≥8469280 (59.70)189 (40.30) 120 (25.59)143 (30.49)
PSA<1031412764 (88.00)377 (12.00)<.001173 (5.51)303 (9.65)
 10-20495391 (78.99)104 (21.01) 49 (9.90)84 (16.97)
 >20170117 (68.82)53 (31.18) 33 (19.41)39 (22.94)
Risk stratificationLow12421188 (95.65)54 (4.35)<.00117 (1.37)41 (3.30)
 Intermediate22651950 (86.09)314 (13.86) 130 (5.74)260 (11.48)
 High637408 (64.05)229 (35.95) 146 (22.92)171 (26.84)

Because surgeon rather than hospital volume is the more significant determinant of outcomes following open RP,20 we determined surgeon volume for each type of procedure by aggregating the number of procedures performed from 2004-2006. Surgeon volume was categorized into quartiles, consistent with a prior study.21

Expenditures Related to the Use of Adjuvant Cancer Therapies

We compared baseline health care expenditures in the 12 months prior to prostate cancer diagnosis for men who underwent RP alone versus those who underwent adjuvant treatment postprostatectomy. To determine the total expense of adjuvant treatment, we summed the total health care expenditures from the beneficiary, Medicare, and supplemental private insurance for inpatient, outpatient, and physician services within 12 months of prostate cancer diagnosis. Approximately 50% of men who received adjuvant therapies did so within 6 months, and we were able to capture costs for 275 of the 600 that received therapy. To ensure that we adequately captured the cost of treatment, we excluded men who underwent RP and adjuvant therapies beyond 6 months following prostate cancer diagnosis. We then subtracted baseline health care expenditures, allowing subjects to serve as their own controls. We considered the difference in health expenditures between men receiving adjuvant treatment versus RP alone to be the health care expenditures attributable to hormonal therapy, radiotherapy, and both treatments in combination. Moreover, the health care expenditures included therapies, consultations, imaging, laboratory tests, and treatment of complications. All costs were adjusted to 2008 dollars using the 2007 Annual Report of the Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Fund.22

Statistical Analysis

Unadjusted analysis using the Pearson chi-square statistic was performed to compare demographic and biopsy tumor characteristics for patients receiving RP and adjuvant treatment versus RP alone, adjusting for clustering by surgeon, surgical approach, surgeon volume, and clinical characteristics.23 A 2-sided result of P<.05 was considered statistically significant. Adjusted analysis was performed with a Cox multivariable regression model to assess the association of the covariates on the use of adjuvant therapies.

All tests were considered statistically significant at α = 0.05. All analyses were performed with SAS version 9.1.3 (SAS Institute, Cary, NC).


  1. Top of page
  2. Abstract

The demographics of our study population are summarized in Table 1. We observed a temporal trend in the administration of adjuvant therapy after RP; patients were more likely to receive adjuvant therapy after RP performed in 2004 versus 2005 or 2006 (15.5%, 13.7% and 10.5%, P = .028). Moreover, whereas age, comorbidities, income, and education were not associated with receipt of adjuvant therapies, there was significant geographic variation for utilization of adjuvant therapies, with the San Jose versus Detroit region having the highest versus lowest utilization rates (20.4% vs 9.9%, P<.001). Furthermore, more aggressive tumor characteristics (higher Gleason grade, preoperative PSA, clinical stage, and risk stratification) were associated with receipt of adjuvant cancer therapy (all P<.001).

In assessing the effect of surgical approach, surgeon volume, and pathological features on the use of adjuvant therapies (Table 2), patients undergoing minimally invasive RP versus retropubic RP were less likely to receive additional cancer therapy (10.9% vs 15.3%, P<.001), and higher surgeon volume was associated with lower utilization of adjuvant cancer therapy (P = .001). Moreover, patients with pathological stage T3a versus T2 disease were more likely to receive additional therapy (36.4% vs 9.7%, P<.001), and patients with positive versus negative surgical margins were more likely to receive adjuvant cancer therapy (31.5% vs 10.0%, P<.001). Finally, patients with positive lymph nodes were more likely to receive adjuvant therapy (75.6% vs 13.5%, P<.001).

Table 2. Adjuvant Therapy by Surgeon Volume, Surgical Approach, Pathological Stage, and Surgical Margin
Independent VariableCategorynAdjuvant TherapyHormonal or RadiationPHormonal TherapyRadiation Therapy
  • MIRP indicates minimally invasive radical prostatectomy; RRP, retropubic radical prostatectomy.

  • Data are presented as No. (%).

  • a

    The actual number of MIRP surgeons is not presented because the National Cancer Institute precludes the reporting of table cells of n<11.

Pathological stageT235473201 (90.25)345 (9.73)<.001148 (4.17)275 (7.75)
 T3a700445 (63.57)255 (36.43) 146 (20.86)199 (28.43)
Positive surgical marginYes822563 (68.49)259 (31.51)<.001129 (15.69)213 (25.91)
 No34253083 (90.01)341 (9.96) 165 (4.82)261 (7.62)
Surgeon volume in quartiles (no. of surgeons by approach)Low (MIRP, 85; RRP, 396)1027867 (84.42)159 (15.48).00163 (6.13)134 (13.05)
 Intermediate (MIRP, 21; RRP, 169)1130944 (83.54)186 (16.46) 94 (8.32)149 (13.19)
 High (MIRP, 12; RRP, 91)1159998 (86.11)161 (13.89) 90 (7.77)120 (10.35)
 Very high (MIRP, <11a; RRP, 37)931837 (89.90)94 (10.10) 47 (5.05)71 (7.63)
Surgical approachMIRP1120998 (89.11)122 (10.89)<.00159 (5.27)97 (8.66)
 RRP31272648 (84.68)478 (15.29) 235 (7.52)377 (12.06)
Positive lymph nodesYes4511 (24.44)34 (75.56)<.00131 (68.89)11 (24.44)
 No42013635 (86.53)566 (13.47) 263 (6.26)463 (11.02)

In adjusted analysis (Table 3), age, race, marital status, and surgical approach (minimally invasive RP vs retropubic RP) were not significantly associated with receipt of adjuvant therapies. However, risk stratification was significantly associated with use of adjuvant therapies as patients with intermediate (hazard ratio [HR], 2.86; 95% confidence interval [CI], 2.14-3.83 [P<.001]) and high-risk (HR, 8.3; 6.13-11.22 [P<.001]) versus low-risk disease were more than 2 and 8 times more likely to undergo adjuvant therapies. Survival estimates are shown in Figure 1 for the various risk groups. Men undergoing RP by very high-volume surgeons were less likely to receive adjuvant therapies (HR, 0.64; 95% CI, 0.49-0.84 [P = .001]). Moreover, patients with positive versus negative surgical margins were 3 times more likely to undergo adjuvant therapies (HR, 3.2; 95% CI, 2.71-3.78 [P<.001]). Men with positive versus negative lymph nodes were 5 times more likely to receive adjuvant therapies (HR, 5.36; 95% CI, 3.71-7.75 [P<.001]). In addition, there was greater use of adjuvant therapies in Iowa (HR, 1.93; 95% CI, 1.12-3.32 [P = .019]) and New Mexico/Georgia/Hawaii (HR, 1.92; 95% CI, 1.09-3.39 [P = .025]) versus San Francisco SEER regions.

thumbnail image

Figure 1. Estimated time to adjuvant therapy for the 3 risk groups with the number of subjects at risk at 1, 2, 3, and 4 years.

Download figure to PowerPoint

Table 3. Unadjusted and Adjusted Model for Predictors of Adjuvant Cancer Treatment
Covariate (Referent)CategoriesUnivariate HR (95% CI)Multivariate HR (95% CI)Multivariate P
  1. HR indicates hazard ratio; CI, confidence interval; RRP, retropubic radical prostatectomy; MIRP, minimally invasive radical prostatectomy.

Age (≥75 years)65-690.98 (0.71-1.34)1.12 (0.81-1.55).477
 70-740.9 (0.65-1.26)0.96 (0.69-1.35).823
Race (white)Black1.01 (0.74-1.39)1.11 (0.79-1.55).555
 Hispanic0.91 (0.67-1.24)0.85 (0.62-1.17).316
 Asian1.47 (1.05-2.06)1.26 (0.88-1.8).203
SEER region (San Francisco)20 = Michigan0.89 (0.5-1.56)0.90 (0.51-1.6).723
 22 = Iowa1.76 (1.01-3.06)1.93 (1.12-3.32).019
 25 = Seattle1.82 (1.06-3.11)1.10 (0.64-1.89).738
 26 = Utah1 (0.58-1.7)1.16 (0.65-2.08).612
 2 = Connecticut0.94 (0.53-1.65)1.37 (0.73-2.58).323
 31 = San Jose1.32 (0.71-2.48)1.71 (0.92-3.17).089
 35 = Los Angeles1.82 (0.98-3.35)1.30 (0.79-2.14).307
 41 = Greater California1.15 (0.7-1.89)1.48 (0.93-2.36).098
 42 = Kentucky1.39 (0.87-2.2)1.40 (0.8-2.45).233
 43 = Louisiana1.41 (0.81-2.44)1.33 (0.77-2.3).301
 44 = New Jersey1.14 (0.67-1.95)1.51 (0.87-2.61).141
 New Mexico/Georgia/Hawaii1.33 (0.77-2.28)1.92 (1.09-3.39).025
Risk stratification (low)Intermediate3.34 (2.5-4.46)2.86 (2.14-3.83).001
 High10.28 (7.64-13.84)7.65 (5.64-10.37)<.001
Surgical margin (negative)Positive3.65 (3.1-4.29)3.02 (2.55-3.57)<.001
Lymph nodes (negative)Positive12.73 (8.99-18.02)5.36 (3.71-7.75)<.001
Surgical approach (RRP)MIRP0.72 (0.59-0.88)0.89 (0.72-1.1).285
Surgeon volume (low)Intermediate1.06 (0.86-1.31)1.02 (0.82-1.27).855
 High0.89 (0.72-1.11)0.86 (0.69-1.08).203
 Very high0.64 (0.49-0.82)0.60 (0.46-0.78)<.001
Year (2004)20051 (0.85-1.19)0.99 (0.83-1.18).903
 20060.85 (0.61-1.18)0.86 (0.62-1.19).356

Baseline health care expenditures in the 12 months prior to prostate cancer diagnosis did not differ for patients who underwent RP alone versus adjuvant therapies of hormonal therapy, radiation therapy, and hormone and radiation therapy. However, the 12-month post–prostate cancer diagnosis health care expenditures (Table 4) of patients who underwent RP alone versus adjuvant therapies of hormonal therapy, radiation therapy, and combination hormonal and radiation therapy were significantly greater for adjuvant therapies (P<.001).

Table 4. Cost Analysis of Adjuvant Cancer Treatments
 Radical ProstatectomyRadical Prostatectomy and Hormonal TherapyRadical Prostatectomy and RadiationRadical Prostatectomy and Radiation with Hormonal TherapyP
  • a

    We excluded patients who underwent radical prostatectomy and adjuvant therapies >6 months after initial treatment (radical prostatectomy) to ensure that we fully captured the expense associated with primary and adjuvant therapy.

  • b

    1-year pre–prostate cancer diagnosis expenditures and expenditures of radical prostatectomy alone, respectively subtracted from 12-month postprostatectomy health care expenditures of various adjuvant therapies.

Baseline health care expenditures in the year prior to prostate cancer diagnosis, median$1861$1272$1380$1333.011
1-year postprostatectomy health care expenditures,a median$15,022$17,661
Health care expenditures attributed to adjuvant therapiesb$1367$12,040$23,487<.001


  1. Top of page
  2. Abstract

Approximately 13%-34% of men who undergo prostatectomy will have adverse pathological features such as positive surgical margins or extracapsular extension/pT3a disease.24, 25 There is a lack of consensus regarding when to initiate treatment in such patients; however, 22%-34% of these patients will receive salvage secondary treatments within 3 years of BCR.26, 27 Whereas a recent population-based study demonstrated significantly greater use of additional cancer treatments (eg, radiation and/or hormonal therapy), within 6 months of minimally invasive versus open RP, potential confounders such as surgical margin status and pathological stage and grade were unavailable in this analysis of Medicare beneficiaries.16 In addition, there is an absence of population-based studies that assess use of adjuvant treatments after adjusting for surgical approach and surgeon volume. Aside from the lack of definitive guidelines on when to initiate adjuvant treatments after BCR and the appropriateness thereof, there is also concern of the added health care costs when adjuvant therapies are initiated.

Our paper has several important findings. First, higher surgeon volume was associated with decreased utilization of adjuvant cancer therapy independent of tumor characteristics. These findings would suggest that heterogeneity in practice patterns exist and that there is not uniform standardization of care. More experienced surgeons may prefer to manage positive surgical margins and extracapsular extension conservatively with surveillance versus adjuvant therapy. Similarly, Bianco et al.28 found significant heterogeneity among BCR rates after adjusting for tumor characteristics and surgeon experience, and oncological outcomes vary due to measured and unmeasured characteristics of the treating surgeon. Thus, as Bianco et al. alluded to, there must be unmeasured characteristics of high-volume surgeons that result in decreased use of adjuvant therapies.

Second, we found that risk stratification was a significant predictor of adjuvant therapy use. Intermediate to high risk patients were approximately 3 to 8 times more likely to receive adjuvant therapy. Tumor biology as measured by pathological stage and grade have been previously shown to be powerful predictors for additional cancer therapy, whereas other patient variables including age and comorbidity have not.12 Moreover, rapid PSA doubling time has also been shown to be significant predictors for secondary therapies.29 Unfortunately, these endpoints are not captured in SEER-Medicare.

Third, positive surgical margin status was associated with increased utilization of adjuvant therapies, despite mixed evidence available during our study period regarding the impact of positive surgical margins on cancer recurrence and survival.30 However, recently published randomized control trials demonstrate survival benefit from early adjuvant radiotherapy for positive surgical margins and high-risk features.5, 31 The interpretation of these trials is not without ongoing controversy, and further studies are warranted to clarify which patients would benefit most from adjuvant treatment.32 Furthermore, patients with lymph node–positive disease were more likely to receive adjuvant therapy, an increase that may be explained by prior studies demonstrating improved cancer-specific survival in such patients managed with adjuvant therapy.33, 34 With greater dissemination of evidence in favor of early adjuvant radiotherapy for adverse pathological features, more widespread adjuvant therapy use is expected and our results may underestimate current and future utilization of adjuvant therapies as practice patterns evolve.

Fourth, patient age, comorbidity status, and race were not significant predictors of adjuvant cancer therapy, consistent with previous studies.11, 12, 29 One would expect that patient factors such as older age and more comorbidities would decrease the likelihood of receiving adjuvant therapies if treatment decisions were individualized. Moreover, these findings may highlight the need for guidelines based on life expectancy and postprostatectomy nomograms to better stratify which patients benefit most from adjuvant therapy. In addition, surgical approach was not a significant predictor for adjuvant therapy on multivariate analysis. Our findings contradict other studies that demonstrated greater use of secondary therapies following minimally invasive versus open RP, whereas other studies found no difference.16, 35 This difference may result from differences between the study populations: namely, a 5% random sample of Medicare beneficiaries16 versus 100% of the Medicare beneficiaries in SEER tumor registry regions. Our study captures the entire surgeon Medicare experience in SEER regions versus a national 5% sampling of surgeon Medicare experience.

Finally, health care expenditures were $23,487 higher for combination radiation and hormonal therapy versus no treatment following prostatectomy. The additional expenditures for adjuvant hormonal therapy and radiotherapy were $1367 and $12,040, respectively versus RP alone. In particular, positive surgical margins, a surgeon-dependent variable, may increase the cost of cancer therapy significantly, particularly after level 1 evidence of improved survival from secondary radiation therapy.4–6

Our findings must be interpreted within the context of the study design. First, Medicare is limited to patients ≥65 years of age, and nerve-sparing may be performed more frequently in younger, potent men.36 This factor, combined with the absence of margin status for pathological stage T3b and T4 disease, may lead to underestimation of the overall prevalence of adjuvant cancer treatments in patients undergoing RP.24 Second, the SEER tumor registry does not contain detailed clinical information on PSA or biochemical recurrence, tumor volume, perineural invasion, and tertiary high Gleason grade, factors that increase the likelihood of adjuvant therapy use.37–39 Third, we were unable to determine whether adjuvant radiotherapy was administered in an adjuvant versus salvage fashion, because postprostatectomy PSA data were unavailable. This observation is noteworthy, because initiation of adjuvant therapies is influenced by variation in provider practice patterns, whereas initiation of salvage therapy may be influenced by variations in PSA biochemical recurrence thresholds. Finally, our estimates of adjuvant therapy expenditures are lower than expenditures by private health plans versus Medicare.


Higher surgeon volume and geographic variation was independently associated with decreased use of additional therapy, demonstrating physician and regional practice pattern heterogeneity. Patients undergoing RP were significantly more likely to undergo adjuvant treatments in the presence of higher risk stratification and positive surgical margins. Finally, adjuvant therapies significantly increased cancer-specific health care expenditures by 2- to 3-fold when radiotherapy was administered postoperatively.


  1. Top of page
  2. Abstract

This work was supported by a Department of Defense Prostate Cancer Physician Training Award (W81XWH-08-1-0283) presented to J. C. Hu.


  1. Top of page
  2. Abstract