SEARCH

SEARCH BY CITATION

Keywords:

  • radiotherapy;
  • brachytherapy;
  • prostatic neoplasms;
  • Health Services Research;
  • epidemiologic factors

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES

BACKGROUND

Prostate cancer treatment choices have been shown to vary by physician and patient characteristics. For patients with low-risk, clinically localized prostate cancer, the authors examined the impact of their clinical, sociodemographic, and radiation oncologists' (RO) characteristics on the likelihood that the patients would receive combined external beam radiotherapy and brachytherapy, a treatment regimen that is at variance with clinical guidelines.

METHODS

The Surveillance, Epidemiology and End Results (SEER)-Medicare linked database and the American Medical Association Physician Masterfile were used in a retrospective analysis of 5531 patients with low-risk, clinically localized prostate cancer who were diagnosed between 2004 and 2007, and the 708 ROs who treated them. Hierarchical logistic regression analyses were used to evaluate the relationship between patient and RO characteristics and the use of combined therapy within 6 months of diagnosis.

RESULTS

Overall, 356 patients (6.4%) received combined therapy. Nonclinical factors were found to be associated with combined therapy. After adjusting for patient and RO characteristics, the odds of receiving combined therapy for patients residing in Georgia were found to be significantly greater than for all other SEER regions. Black patients were significantly less likely to receive combined therapy (odds ratio, 0.62; 95% confidence interval, 0.40-0.96 [P = .03]) compared with white patients. In addition, ROs accounted for 36.6% of the variation in patients receiving combined therapy.

CONCLUSIONS

Geographic and sociodemographic factors were found to be significantly associated with guideline-discordant combined therapy for patients diagnosed with low-risk, clinically localized prostate cancer. Which RO a patient consults is important in determining whether they receive combined therapy. Cancer 2013;119:3619–3628. © 2013 American Cancer Society.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES

Due to the prolonged natural history of prostate cancer, most men diagnosed at an older age with clinically localized prostate cancer face competing health risks and are most unlikely to benefit from aggressive medical treatment (ie, surgery or radiotherapy [RT]).[1, 2] There are also growing concerns about prostate cancer overtreatment. Multiple reports have indicated an increased prevalence of aggressive therapy among patients with low-risk, clinically localized prostate cancer.[2-4] At the same time, patterns of care for patients with prostate cancer are also shifting.[4, 5] In 1991, approximately 54% of Medicare patients with prostate cancer underwent surgical treatment and 33% received RT. By 2002, only 24% of patients were treated surgically, whereas 47% received RT.[5] The increasing use of RT comes with a diverse array of delivery modalities. However, retrospective studies and nonrandomized trials to date have suggested that cancer-specific outcomes are similar across treatments for men with low-risk prostate cancer.[2, 6]

The clinical benefit of combining external beam RT (EBRT) and brachytherapy (BT) for the treatment of patients with low-risk, clinically localized prostate cancer has yet to be proven.[2, 7] In fact, there exist multiple studies that demonstrate increased rates of adverse side effects associated with genitourinary and gastrointestinal toxicity, and a reduction in health-related quality of life when EBRT is supplemented with BT.[2, 8] Consequently, the combination of EBRT with BT for patients with low-risk, clinically localized prostate cancer is not supported by clinical practice guidelines issued by the American Brachytherapy Society (ABS),[2] the American Urological Association (AUA),[2] and the National Comprehensive Cancer Network (NCCN).[2] The costs associated with combined EBRT and BT, hereinafter referred to as combined therapy, are considerably more than for BT when used as monotherapy.[9] In 2005, the mean costs for each Medicare patient with prostate cancer ranged from US dollar (US$)26,000 to US$37,000 for combined therapy compared with US$17,000 for BT alone.[9] Current costs of intensity-modulated RT, the most common form of EBRT, have approached US$50,000 per patient,[10] whereas BT costs have remained approximately the same. Despite these differences in cost, there remains a persistent trend of patients with low-risk, clinically localized prostate cancer receiving combined EBRT and BT.[2, 4, 11] This may be a reflection of socioeconomic and geographic variation as indicated in previous studies,[2, 12] or may be associated with practice site[13] and physician characteristics.[14]

Using clinical guidelines issued by the ABS, AUA, and NCCN, the current study investigated the determinants of the combined use of EBRT and BT for patients with low-risk, clinically localized prostate cancer, which is a specific case of guideline-discordant care. We investigated the association between patient and radiation oncologist characteristics on the patient's receipt of combined therapy.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES

Data Sources

The Surveillance, Epidemiology and End Results (SEER)-Medicare database, which links cancer registry information in selected geographic areas of the United States with claims data for covered health care services of Medicare beneficiaries,[2] was used to create the analytical cohort using the criteria in Figure 1. During the study period, incident cancer cases were available from 16 SEER registries from 2004 through 2007; the SEER program covered 26% of the US population whereas Medicare covered health services for approximately 97% of individuals aged ≥65 years. Approximately 94% of SEER patients aged ≥65 years were successfully linked with their Medicare claims.[2] Data from the Louisiana SEER registry were not used in the current study because of missing data for 2005 after Hurricane Katrina. Data from metropolitan Atlanta and Rural Georgia SEER registries are hereinafter classified as Georgia.

image

Figure 1. Definition of a study cohort of 5531 men with low-risk, clinically localized prostate cancer is shown. SEER indicates Surveillance, Epidemiology and End Results; PSA, prostate-specific antigen; HMO, health maintenance organization; AMAPM, American Medical Association Physician Masterfile. *Resulting from the adaption of criteria developed by Shahinian et al.[14]

Download figure to PowerPoint

The characteristics of the physicians who treated the SEER-Medicare patients were obtained from the American Medical Association (AMA) Physician Masterfile. This file contains physician information collected from multiple sources including medical schools, residency training programs, state licensing boards, and the American Board of Medical Specialties (ABMS).[2]

Institutional Review Board approval was obtained from the Morehouse School of Medicine and Emory University.

Patient Characteristics

Patient characteristics used in this study are presented in Table 1. Clinical characteristics that corresponded to prostate cancer at low risk of disease recurrence as defined by NCCN (specifically a prostate-specific antigen level < 10 ng/mL at the time of diagnosis, clinical stage T1 to T2a disease, and low-to-intermediate tumor grade [low: Gleason grade 2-4; intermediate: Gleason grade 2-6])[2] were extracted from SEER files. In addition, race, ethnicity, age, marital status, rural status, SEER region of residence at the time of diagnosis, and year of diagnosis were determined from the SEER files. Education and income levels were based on US Census tract data from 2000, with categories chosen to ensure a reasonable distribution with cutoff values approximately corresponding to quartiles. A modified Charlson Comorbidity Index from Medicare Part A and Part B claims[2] was calculated for each patient using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes consistent with comorbidities of interest. Claims were examined from 1 year before to 1 month after cancer diagnosis for eligible codes.

Table 1. Percentage of Patients Receiving Combined Therapy Within 6 Months After Prostate Cancer Diagnosis According to Their Clinical and Sociodemographic Characteristics and the Characteristics of Their Radiation Oncologists
CharacteristicNo.Percentage of Patients Receiving Combined TherapyPa
  1. Abbreviations: DO, Doctor of Osteopathy; MD, Doctor of Medicine; RT, radiotherapy; SEER, Surveillance, Epidemiology and End Results; USD, US dollars.

  2. a

    P values were calculated from 2-sided chi-square tests for heterogeneity in percentage of patients receiving combined therapy across different patient sociodemographic and clinical characteristics and their radiation oncologists' characteristics.

  3. b

    Comorbidity index was based on a modification of the Charlson Comorbidity Index.

All patients55316.4 
Patient sociodemographic characteristics
Age at diagnosis, y  .227
  66–6917446.5 
  70–7422016.5 
  75–7912676.9 
  80–842834.2 
  ≥85365.6 
Race/ethnicity  .259
  Non-Hispanic white44536.1 
  Non-Hispanic black4227.6 
  Hispanic3568.7 
  Non-Hispanic Asian/Pacific Islander2167.4 
  Other/unknown847.1 
Marital status  .155
  Married42116.7 
  Not married9865.1 
  Unknown3346.6 
SEER region of residence  <.0001
  Georgia24633.7 
  California16843.6 
  Connecticut5062.6 
  Detroit4196.2 
  Hawaii963.1 
  Iowa2216.8 
  Kentucky4925.1 
  New Jersey13238.7 
  New Mexico816.2 
  Seattle3202.8 
  Utah1430.7 
Rural status  .625
  Not rural54546.4 
  Rural777.8 
Census tract: percentage of adults with <high school education.067
  <7.413986.9 
  7.4 to <12.7013616.5 
  12.70 to <21.1013835.0 
  ≥21.1013897.3 
Census tract: median income (USD).038
  <39,00013845.6 
  39,000 to <52,00013715.3 
  52,000 to <71,00014057.4 
  ≥71,00013717.4 
Patient clinical characteristics
Comorbidity indexb  .377
  035386.0 
  113766.8 
  23827.9 
  ≥32357.7 
Clinical tumor classification  .083
  T148406.7 
  T2a6914.9 
Tumor grade  .130
  Low601.7 
  Intermediate54716.5 
Y of diagnosis   
  200414196.4.227
  200513385.3 
  200614556.9 
  200713197.1 
Radiation oncologist characteristics
Age, y  .905
  <4113726.7 
  41–4615966.1 
  47–5213046.3 
  >5212596.7 
Sex  .343
  Male51636.4 
  Female3687.6 
Board certification  .060
  Yes40196.8 
  No15125.4 
US trained  .520
  Yes47696.5 
  No7625.9 
Degree type  .041
  MD54436.5 
  DO881.1 
No. of y after medical school graduation  .246
  <1312996.5 
  14–1815867.2 
  19–2413696.6 
  >2412775.3 
No. of patients  .043
  <1313436.0 
  13–2714537.9 
  28–4814335.4 
  >4813026.4 
Medical school affiliation  <.0001
  Major7664.3 
  Mixed28355.0 
  None16658.6 
  Non-institutional26514.7 

Patients who underwent radical prostatectomy or bilateral orchiectomy (as defined elsewhere[2]) were removed from the study cohort. Thus, our intent was to create a relatively homogenous sample of patients with low-risk, clinically localized prostate cancer who would not be candidates for combined EBRT and BT, according to ABS, AUA, and NCCN guidelines.

To assign a principal radiation oncologist to each patient, criteria were adapted from Shahinian et al.[14] Patients who did not see at least 1 radiation oncologist within the year after diagnosis on at least 2 separate days were excluded. If a patient saw ≥2 radiation oncologists, he was assigned to the radiation oncologist who saw him for at least 75% of his radiation oncologist visits within the year after diagnosis. If no single radiation oncologist accounted for at least 75% of all the radiation oncologist visits associated with that patient, he was excluded.

Radiation Oncologist Characteristics

Radiation oncologists who treated patients with prostate cancer were identified using either the Health Care Financing Administration specialty codes in Medicare claims, the physician specialty (primary or secondary) information from the AMA Physician Masterfile, or the radiation oncologist board certification identified through the ABMS information within the AMA Physician Masterfile. Other radiation oncologist characteristics obtained from the AMA Physician Masterfile included age, sex, years after medical school graduation, location of training (US or otherwise), type of degree (Doctor of Medicine [MD] or Doctor of Osteopathy [DO]) and radiation oncology board certification (Table 1). Radiation oncologists' patient volume is defined as the number of unique Medicare patients with prostate cancer that each radiation oncologist saw during the study period. The categories chosen for the radiation oncologists' patient volume, age, and years after medical school graduation were prespecified to ensure a reasonable distribution with cutoff values approximately corresponding to quartiles.

Radiation Oncologists' Medical School Affiliation

The source files for key variables used to categorize the radiation oncologist's medical school affiliation are summarized in Figure 2. Medical school affiliations were derived from the SEER-Medicare Hospital (HOSP) file and adapted from methods previously described elsewhere.[2, 14] Radiation oncologists were categorized as having a major medical school affiliation if all their inpatient (from the Medicare Provider Analysis and Review [MEDPAR] file) and institutional outpatient (from the Medicare Outpatient Standard Analytical File [OUTSAF]) claims during the study period were submitted from hospitals with a major medical school affiliation (as defined within the HOSP file). Conversely, radiation oncologists were categorized as having no medical school affiliation if all their inpatient and institutional outpatient claims were from hospitals with no medical school affiliation. Because claims from the MEDPAR file do not contain Unique Physician Identifier Numbers (UPINs), MEDPAR claims are assigned UPINs associated with National Claims History (NCH) file claims (consisting of mostly noninstitutional physician/supplier claims) if 1) the patient associated with both types of claims matched, 2) the place of service of the NCH claim was institutional,[2] and 3) the NCH claims dates fell between the MEDPAR admission and discharge dates. Radiation oncologists whose claims could only be found in the NCH file and whose claims could not be matched with the MEDPAR file (as described previously) were categorized as having a non-institutional affiliation. All other radiation oncologists were categorized as having a mixed medical school affiliation.

image

Figure 2. Locations of key variables across different files of the Surveillance, Epidemiology and End Results (SEER)-Medicare database and the American Medical Association (AMA) Physician Masterfile (AMAPM) are shown. PEDSF indicates Patient Entitlement and Diagnosis Summary File; OUTSAF, Outpatient Standard Analytical File; NCH, National Claims History records; MEDPAR, Medicare Provider Analysis and Review File; HOSP, Hospital File; ABMS, American Board of Medical Specialties; ID, identifier.

Download figure to PowerPoint

Measurement of Treatment and Outcomes

EBRT and BT were identified from Medicare's MEDPAR, OUTSAF, and NCH files based on the presence of Current Procedural Terminology (fourth edition) codes and ICD-9-CM codes, as defined elsewhere.[2, 9] The primary outcome was the receipt of combined EBRT and BT within the first 6 months after a diagnosis of low-risk, clinically localized prostate cancer. Therefore, the current study was limited to investigating whether combined therapy was used as a form of initial therapy among patients with low-risk, clinically localized prostate cancer. In cases in which the use of combined EBRT and BT was observed, it was not supported by clinical evidence and represented discordance with major clinical practice guidelines.[2]

Statistical Analyses

Differences in the percentage of patients receiving combined therapy across radiation oncologist and patient characteristics were evaluated using chi-square tests. The effect of patient and radiation oncologist characteristics on the receipt of combined therapy was evaluated using hierarchical logistic generalized linear mixed models (GLMM)[15] and estimated using the restricted pseudo-likelihood methodology.[2, 15] Hierarchical GLMMs accounted for the clustering of the receipt of combined therapy among patients who had the same radiation oncologist. The unit of analysis was the patient. The radiation oncologist associated with each patient was used as the clustering variable. Univariate and adjusted multivariate odds ratios (ORs) and 95% confidence intervals (95% CIs) for the receipt of combined therapy were estimated for the radiation oncologist and patient variables listed in Table 2. Mean predicted probabilities of the receipt of combined therapy across different SEER regions stratified by the medical school affiliation of the radiation oncologists were also estimated from the final fitted model (Fig. 3).

image

Figure 3. Predicted probabilities of patient receipt of combined therapy for all Surveillance, Epidemiology and End Results (SEER) regions, adjusted for patient sociodemographic and clinical characteristics and their principal radiation oncologists' characteristics, were estimated from hierarchical generalized linear mixed models for each category of the principal radiation oncologists' medical school affiliation. Probability of the receipt of combined therapy for Utah patients treated by radiation oncologists with a non-institutional affiliation cannot be estimated as the original sample did not contain any patient in that category.

Download figure to PowerPoint

Table 2. Unadjusted Univariate and Adjusted Multivariate Multilevel Regression Models Predicting the Odds of Receiving Combined Therapy Among Patients With Low-Risk, Clinically Localized Prostate Cancer
 UnadjustedAdjusted
CharacteristicsOR (95% CI)PaOR (95% CI)Pa
  1. Abbreviations: 95% CI, 95% confidence interval; DO, Doctor of Osteopathy; MD, Doctor of Medicine; OR, odds ratio; RT, radiotherapy; SEER, Surveillance, Epidemiology and End Results; USD, US dollars.

  2. a

    P values were calculated from hierarchical generalized linear mixed models.

  3. b

    Comorbidity index was based on a modification of the Charlson Comorbidity Index.

Patient sociodemographic characteristics
Age at diagnosis, y    
  66–691.00 (Referent) 1.00 (Referent) 
  70–741.00 (0.77–1.29).9720.99 (0.76–1.27).906
  75–791.06 (0.80–1.42).6741.00 (0.72–1.38).989
  80–840.64 (0.35–1.18).1500.56 (0.31–1.00).052
≥850.85 (0.20–3.58).8240.73 (0.26–2.09).559
Race/ethnicity    
  Non-Hispanic white1.00 (Referent) 1.00 (Referent) 
  Non-Hispanic black0.74 (0.48–1.15).1780.62 (0.40–0.96).031
  Hispanic1.29 (0.86–1.94).2261.50 (0.95–2.36).084
  Non-Hispanic Asian/Pacific Islander1.68 (0.67–4.20).2672.39 (0.87–6.56).090
  Other/unknown0.71 (0.25–2.03).5200.72 (0.23–2.29).582
Marital status    
  Married1.00 (Referent) 1.00 (Referent) 
  Not married0.72 (0.53–0.98).0350.76 (0.55–1.05).090
  Unknown1.05 (0.62–1.76).8691.01 (0.55–1.85).976
SEER region of residence    
  Georgia1.00 (Referent) 1.00 (Referent) 
  California0.07 (0.03–0.14)<.00010.05 (0.03–0.11)<.0001
  Connecticut0.05 (0.02–0.12)<.00010.05 (0.02–0.12)<.0001
  Detroit0.12 (0.05–0.32)<.00010.13 (0.04–0.36).0001
  Hawaii0.04 (0.01–0.17)<.00010.02 (0.002–0.16).0003
  Iowa0.12 (0.05–0.30)<.00010.12 (0.05–0.32)<.0001
  Kentucky0.14 (0.06–0.33)<.00010.13 (0.05–0.33)<.0001
  New Jersey0.11 (0.05–0.25)<.00010.11 (0.05–0.25)<.0001
  New Mexico0.16 (0.05–0.49).0010.18 (0.05–0.63).007
  Seattle0.07 (0.03–0.20)<.00010.07 (0.03–0.20)<.0001
  Utah0.02 (0.002–0.29).0040.03 (0.002–0.38).008
Rural status    
  Not rural1.00 (Referent) 1.00 (Referent) 
  Rural1.46 (0.68–3.15).3351.38 (0.60–3.16).449
Census tract: percentage of adults with <high school education  
  <7.41.00 (Referent) 1.00 (Referent) 
  7.4 to <12.700.91 (0.60–1.38).6490.94 (0.56–1.57).807
  12.70 to <21.100.73 (0.50–1.08).1170.81 (0.48–1.35).415
  ≥21.100.94 (0.66–1.34).7141.20 (0.64–2.25).564
Census tract: median income (USD)  
  <39,0001.00 (Referent) 1.00 (Referent) 
  39,000 to <52,0000.87 (0.60–1.26).4531.02 (0.67–1.55).930
  52,000 to <71,0001.34 (0.93–1.93).1141.70 (1.01–2.85).046
  ≥71,0001.12 (0.76–1.67).5631.37 (0.26–2.09).377
Patient clinical characteristics
Comorbidity indexb    
  01.00 (Referent) 1.00 (Referent) 
  11.08 (0.83–1.41).5531.16 (0.88–1.53).284
  21.34 (0.85–2.12).2091.32 (0.81–2.13).266
  ≥31.33 (0.72–2.44).3641.45 (0.77–2.72).250
Clinical tumor classification    
  T11.00 (Referent) 1.00 (Referent) 
  T2a0.86 (0.58–1.26).4400.98 (0.67–1.44).924
Tumor grade    
  Low1.00 (Referent) 1.00 (Referent) 
  Intermediate3.33 (0.47–23.42).2263.09 (0.40–23.68).279
Y of diagnosis    
  20041.00 (Referent) 1.00 (Referent) 
  20050.79 (0.56–1.12).1900.81 (0.56–1.17).258
  20060.98 (0.68–1.41).9160.93 (0.64–1.36).710
  20070.94 (0.66–1.34).7350.88 (0.61–1.26).482
Radiation oncologist characteristics
Sex    
  Male1.00 (Referent) 1.00 (Referent) 
  Female1.56 (0.89–2.71).1181.46 (0.82–2.63).203
Board certification    
  Yes1.00 (Referent) 1.00 (Referent) 
  No0.95 (0.60–1.50).8171.30 (0.70–2.42).413
US trained    
  Yes1.00 (Referent) 1.00 (Referent) 
  No0.95 (0.57–1.58).8341.06 (0.58–1.94).840
Degree type    
  MD1.00 (Referent) 1.00 (Referent) 
  DO0.41 (0.02–7.05).5370.18 (0.01–6.58).350
No. of y after medical school graduation    
  <131.00 (Referent) 1.00 (Referent) 
  14–180.81 (0.44–1.47).4820.75 (0.43–1.31).309
  19–240.80 (0.44–1.47).4790.91 (0.51–1.64).759
  >240.80 (0.44–1.43).4430.72 (0.33–1.61).429
No. of patients    
  <131.00 (Referent) 1.00 (Referent) 
  13–271.03 (0.64–1.64).9111.29 (0.80–2.08).302
  28–480.67 (0.37–1.20).1741.02 (0.52–1.98).959
  >480.73 (0.30–1.81).5000.82 (0.38–1.77).614
Medical school affiliation    
  Major1.00 (Referent) 1.00 (Referent) 
  Mixed1.10 (0.59–2.06).7681.04 (0.54–2.02).911
  None1.61 (0.85–3.05).1431.26 (0.62–2.57).518
  Non-institutional2.79 (1.19–6.55).0191.68 (0.68–4.16).260

To estimate the percentage of total variance in the combined use of EBRT and BT attributable to the radiation oncologist, the intraclass correlation coefficient (ICC) from hierarchical GLMM using the threshold method[16] was estimated. Both a null model, which excluded patient and radiation oncologist characteristics, and adjusted models, which included all these characteristics, were constructed. From the adjusted models, the residual ICC, representing the percentage of variance attributable to the radiation oncologist after adjustments, was calculated. All statistical testing was 2-sided, performed at the 5% significance level, and used SAS statistical software (version 9.3; SAS Institute Inc, Cary, NC).

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES

A total of 5531 patients in the SEER-Medicare database were diagnosed with incident low-risk, clinically localized prostate cancer from 2004 through 2007 and assessed as meeting the eligibility criteria defined in Figure 1. A total of 708 principal radiation oncologists treated these patients. The radiation oncologists were predominantly male (85.0%), board certified (72.2%), trained in the United States (84.0%), and had a patient volume of fewer than 13 patients (68.9%) throughout the study period (Table 3).

Table 3. Characteristics of Principal Radiation Oncologists Who Treated Patients Diagnosed With Low Risk, Clinically Localized Prostate Cancer Between 2004 and 2007
 No.%
All radiation oncologists708100.0
Age, y
  <4121129.8
  41–4617624.9
  47–5215121.3
  >5217024.0
Sex
  Male60285.0
  Female10615.0
Board certification
  Yes51172.2
  No19727.8
US trained
  Yes59584.0
  No11316.0
No. of y after medical school graduation
  <1320228.5
  14–1817925.3
  19–2414119.9
  >2418626.3
No. of patients
  <1348868.9
  13–2713018.4
  28–48588.2
  >48324.5
Medical school affiliation
  Major12217.2
  Mixed26337.2
  None23533.2
  Non-institutional8812.4

Table 1 compares the percentages of patients who received combined therapy classified by their sociodemographic, clinical, and principal radiation oncologist characteristics. Overall, 6.4% of the patients with low-risk, clinically localized prostate cancer were treated with combined EBRT and BT. Combined therapy was more commonly administered to patients who resided in Georgia and in census tracts with a median income greater than US$52,000. The principal radiation oncologists associated with patients who received combined therapy were more likely to be practicing in non-institutional settings, to be MDs, and/or to have a patient volume size between 13 and 27 patients.

Using logistic hierarchical GLMM analyses, we investigated the determinants of receipt of combined therapy (Table 2) while controlling for the fact that multiple patients may be treated by the same principal radiation oncologist. As the unadjusted ORs demonstrate, combined therapy was found to be significantly associated with a patient's SEER region of residence and unmarried status, and the non-institutional practice affiliation of the patient's principal radiation oncologist.

After adjusting for known potential patient and radiation oncologist characteristics, significant regional variation remained. In particular, patients who resided in the SEER region of Georgia were found to be significantly more likely to receive combined therapy compared with all other SEER regions (eg, the odds of receiving combined therapy for patients residing in Georgia [OR, 1.0; referent] was approximately 20 times that of patients residing in California [OR, 0.05; 95% CI, 0.03-0.11 (P < 0.0001)]). In addition, patients residing in census tracts with a median income between US$52,000 and US$70,999 were significantly more likely to receive combined therapy (OR, 1.70; 95% CI, 1.01-2.85 [P = .046]) when compared with patients residing in census tracts with a median income < US$39,000. Non-Hispanic black patients were found to be significantly less likely to receive combined therapy (OR, 0.62; 95% CI, 0.40-0.96 [P = .031]) compared with non-Hispanic white patients.

To further investigate the effects of geographic variation on the receipt of combined therapy, the predicted probabilities from the logistic hierarchical GLMM were estimated. When analyzing by category of the patients' principal radiation oncologists' medical school affiliation, the mean predicted probabilities of receiving combined therapy for patients residing in Georgia were the highest compared with all other SEER regions; in particular, patients in Georgia who consulted radiation oncologists with a non-institutional affiliation were found to have the highest probability (46.4%) of receiving combined therapy (Fig. 3). Conversely, the mean predicted probabilities of receiving combined therapy for patients residing in Utah were the lowest across all categories of principal radiation oncologists' medical school affiliation when compared with all other SEER regions (Fig. 3).

As a measure of the overall influence of the radiation oncologist on the patient's receipt of combined therapy, we estimated the ICC. In the null model, with no predictors included, it is assumed that the probability of the receipt of combined therapy does not vary by individual patient or radiation oncologist characteristics. From the null model, we estimated that 42.4% of the variance in receiving combined therapy could be attributed to the radiation oncologist. After adjusting for the patient and radiation oncologist characteristics listed in Table 2, the variance attributable to the radiation oncologist decreased to 36.6%.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES

After the adjustment for other patient and radiation oncologist characteristics, the hierarchical multivariable analysis suggested important regional differences in the use of combined therapy in patients with low-risk, clinically localized prostate cancer. This treatment regimen is not consistent with patient-centered clinical guidelines,[2] is more costly than monotherapy,[9] has not been shown to have a significant survival benefit, and has been demonstrated to result in increased radiation-related toxicities in several studies.[2, 7, 8] The findings of the current study are consistent with, and complement, those from prior studies that observed variations in treatment selection across SEER regions with a more generalized prostate cancer population[4, 17]; 1 previous study found significantly increased odds of receiving BT with or without EBRT in Atlanta (OR, 13.8; 95% CI, 2.9-64.7) when compared with California among patients with clinically localized prostate cancer.[4] Such geographic variations in treatment patterns may reflect local practice patterns.[18]

Several studies have documented lower rates of overall RT[2, 17] and BT[2, 19, 20] in black men compared with white men. The results of the current study demonstrate that this disparity may have led to increased concordance with clinical guidelines that recommend radiation monotherapy (ie, either BT or EBRT) for patients with low-risk, clinically localized prostate cancer. Similarly, the significant income effects that were also observed in a different study[17] may have been adversely related to the increased odds of receiving combined therapy.

Clinical guidelines from the ABS (published in 1999)[2] and the NCCN (published annually since 2000)[2] have, throughout the study period, consistently recommended radiation monotherapy for patients with low-risk, clinically localized prostate cancer. However in 1995, the AUA published no specific recommendations regarding the use of combined therapy.[2] This may have potentially influenced the behavior of radiation oncologists regarding the use of combined therapy. Nonetheless, in 2007, the AUA did specifically recommend monotherapy for the treatment of low-risk, clinically localized prostate cancer.[2] Although the results of the current study generally support findings from other studies indicating that a higher quality of care and better guideline compliance may be offered in academic settings versus non-academic settings,[2, 21] the lack of adherence to clinical guidelines may also be the consequence of other barriers that were not directly controlled for in the current study.[22]

We acknowledge important limitations of our SEER-Medicare–based study.[23] Patterns of RT use among the elderly may not be generalizable to other patient populations (eg, younger, privately insured patients and/or those receiving care in health maintenance organizations) and may not be representative of the radiation oncologist's entire practice. Nonetheless, approximately 79% of men with incident prostate cancer were aged ≥65 years[2] and in 2010, approximately 76% of Medicare beneficiaries were covered under the traditional fee-for-service program.[24] Individual patient selection biases that were not assessed in the current study may have led to variability in the observations. It is unclear if the results of the current study reflect the decisions of radiation oncologists or individual choices made by patients. Previous analyses of the major role physicians play in advising and influencing patients with prostate cancer regarding best treatment options[2, 25] suggest that patient preference is unlikely to entirely explain the observed treatment patterns of radiation oncologists.

Although patient and clinical factors influence treatment decisions, there is growing evidence that physician characteristics are also important determinants of cancer care.[2, 14, 26] To our knowledge, the current study is the first to explore patient-related and physician-related determinants of guideline-discordant care for patients with low-risk, clinically localized prostate cancer. Previous studies examining variance in treatment attributable to different oncology physician specialties have found ICCs to be lower (ie, 21%-23%)[2, 27] than in the current study. One interpretation of the current study's findings is that changing radiation oncologist behavior would likely lead to a larger impact on clinical practice patterns[2, 28] for patients with low-risk, clinically localized prostate cancer. Although the adjusted hierarchical GLMM yielded nonsignificant associations for all the radiation oncologist characteristics used in the current study, future research is needed to explore other radiation oncologist characteristics (eg, intrinsic motivation, professionalism) that may account for treatment variation.

The reasons behind the eventual decision to undergo combined therapy are likely to be complex and multifactorial.[2, 29] There is growing evidence documenting that physicians with ownership interests refer patients for procedures more frequently than those without potential financial conflicts. Published studies, across multiple diagnostic and therapeutic areas, have demonstrated that this generally leads to increased health care use and associated costs, especially within a fee-for-service health care system.[30-36] In particular, a prior study found that RT facilities with ownership interests by non-radiation oncologists performed 58% more procedures than did facilities without ownership conflicts.[37] Additional research is needed to investigate whether the substantially higher probability of discordance by radiation oncologists practicing in noninstitutional settings may be associated with the potential for self-referral among urology-radiation oncology–integrated practices. Prior literature has suggested the potential for the overtreatment of patients with prostate cancer.[2, 10]

FUNDING SUPPORT

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES

This study was supported by the Intramural Research Department of the American Cancer Society, Atlanta, Georgia.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES
  • 1
    Albertsen PC, Moore DF, Shih WC, Lin Y, Li H, Lu-Yao GL. Impact of comorbidity on survival among men with localized prostate cancer. J Clin Oncol. 2011;29:13351341.
  • 2
    Emory University, American Cancer Society. tinyurl.com/CancerIdea3AddRef. Accessed May 17, 2013.
  • 3
    Cooperberg MR, Lubeck DP, Meng MV, Mehta SS, Carroll PR. The changing face of low-risk prostate cancer: trends in clinical presentation and primary management. J Clin Oncol. 2004;22:21412149.
  • 4
    Hamilton AS, Albertsen PC, Johnson TK, et al. Trends in the treatment of localized prostate cancer using supplemented cancer registry data. BJU Int. 2011;107:576584.
  • 5
    Warren JL, Yabroff KR, Meekins A, Topor M, Lamont EB, Brown ML. Evaluation of trends in the cost of initial cancer treatment. J Natl Cancer Inst. 2008;100:888897.
  • 6
    D'Amico AV, Whittington R, Malkowicz SB, et al. Biochemical outcome after radical prostatectomy, external beam radiation therapy, or interstitial radiation therapy for clinically localized prostate cancer. JAMA. 1998;280:969974.
  • 7
    Merrick GS, Butler WM, Wallner KE, et al. Impact of supplemental external beam radiotherapy and/or androgen deprivation therapy on biochemical outcome after permanent prostate brachytherapy. Int J Radiat Oncol Biol Phys. 2005;61:3243.
  • 8
    Sanda MG, Dunn RL, Michalski J, et al. Quality of life and satisfaction with outcome among prostate-cancer survivors. N Engl J Med. 2008;358:12501261.
  • 9
    Nguyen PL, Gu XM, Lipsitz SR, et al. Cost implications of the rapid adoption of newer technologies for treating prostate cancer. J Clin Oncol. 2011;29:15171524.
  • 10
    Falit BP, Gross CP, Roberts KB. Integrated prostate cancer centers and over-utilization of IMRT: a close look at fee-for-service medicine in radiation oncology. Int J Radiat Oncol Biol Phys. 2010;76:12851288.
  • 11
    Cooperberg MR, Broering JM, Kantoff PW, Carroll PR. Contemporary trends in low risk prostate cancer: risk assessment and treatment. J Urol. 2007;178(3 pt 2):S14S19.
  • 12
    Spencer BA, Fung CH, Wang M, Rubenstein LV, Litwin MS. Geographic variation across veterans affairs medical centers in the treatment of early stage prostate cancer. J Urol. 2004;172(6 pt 1):23622365.
  • 13
    Cooperberg MR, Broering JM, Carroll PR. Time trends and local variation in primary treatment of localized prostate cancer. J Clin Oncol. 2010;28:11171123.
  • 14
    Shahinian VB, Kuo YF, Freeman JL, Orihuela E, Goodwin JS. Characteristics of urologists predict the use of androgen deprivation therapy for prostate cancer. J Clin Oncol. 2007;25:53595365.
  • 15
    Littell RC, Milliken GA, Stroup WW, Wolfinger RD, Schabenberger O. SAS® for Mixed Models. 2nd ed. Cary, NC: SAS Institute Inc; 2006.
  • 16
    Snijders TAB, Bosker R. Multilevel Analysis: An Introduction to Basic and Advanced Multilevel Modeling. Thousand Oaks, CA: Sage; 1999.
  • 17
    Krupski TL, Kwan L, Afifi AA, Litwin MS. Geographic and socioeconomic variation in the treatment of prostate cancer. J Clin Oncol. 2005;23:78817888.
  • 18
    McNaughton Collins M, Barry MJ, Zietman A, et al. United States radiation oncologists' and urologists' opinions about screening and treatment of prostate cancer vary by region. Urology. 2002;60:628633.
  • 19
    Zeliadt SB, Potosky AL, Etzioni R, Ramsey SD, Penson DF. Racial disparity in primary and adjuvant treatment for nonmetastatic prostate cancer: SEER-Medicare trends 1991 to 1999. Urology. 2004;64:11711176.
  • 20
    Underwood W 3rd, Jackson J, Wei JT, et al. Racial treatment trends in localized/regional prostate carcinoma: 1992–1999. Cancer. 2005;103:538545.
  • 21
    Jha AK, Li Z, Orav EJ, Epstein AM. Care in US hospitals–the Hospital Quality Alliance program. N Engl J Med. 2005;353:265274.
  • 22
    Cabana MD, Rand CS, Powe NR, et al. Why don't physicians follow clinical practice guidelines? A framework for improvement. JAMA. 1999;282:14581465.
  • 23
    Warren JL, Klabunde CN, Schrag D, Bach PB, Riley GF. Overview of the SEER-Medicare data: content, research applications, and generalizability to the United States elderly population. Med Care. 2002;40(suppl 8):IV-3IV-18.
  • 24
    The Henry J. Kaiser Family Foundation. Medicare Chartbook, 2010. http://kaiserfamilyfoundation.files.wordpress.com/2013/01/8103.pdf. Accessed May 17, 2013.
  • 25
    Zeliadt SB, Moinpour CM, Blough DK, et al. Preliminary treatment considerations among men with newly diagnosed prostate cancer. Am J Manag Care. 2010;16:e121e130.
  • 26
    Hershman DL, Buono D, McBride RB, Tsai WY, Neugut AI. Influence of private practice setting and physician characteristics on the use of breast cancer adjuvant chemotherapy for elderly women. Cancer. 2009;115:38483857.
  • 27
    Shahinian VB, Kuo YF, Freeman JL, Goodwin JS. Determinants of androgen deprivation therapy use for prostate cancer: role of the urologist. J Natl Cancer Inst. 2006;98:839845.
  • 28
    Hofer TP, Hayward RA, Greenfield S, Wagner EH, Kaplan SH, Manning WG. The unreliability of individual physician “report cards” for assessing the costs and quality of care of a chronic disease. JAMA. 1999;281:20982105.
  • 29
    Zeliadt SB, Ramsey SD, Penson DF, et al. Why do men choose one treatment over another?: a review of patient decision making for localized prostate cancer. Cancer. 2006;106:18651874.
  • 30
    Shen J, Andersen R, Brook R, Kominski G, Albert PS, Wenger N. The effects of payment method on clinical decision-making: physician responses to clinical scenarios. Med Care. 2004;42:297302.
  • 31
    Mitchell JM, Sass TR. Physician ownership of ancillary services: indirect demand inducement or quality assurance? J Health Econ. 1995;14:263289.
  • 32
    Mitchell JM. Do financial incentives linked to ownership of specialty hospitals affect physicians' practice patterns? Med Care. 2008;46:732737.
  • 33
    Gazelle GS, Halpern EF, Ryan HS, Tramontano AC. Utilization of diagnostic medical imaging: comparison of radiologist referral versus same-specialty referral. Radiology. 2007;245:517522.
  • 34
    Mitchell JM. Effect of physician ownership of specialty hospitals and ambulatory surgery centers on frequency of use of outpatient orthopedic surgery. Arch Surg. 2010;145:732738.
  • 35
    Mitchell JM. Urologists' self-referral for pathology of biopsy specimens linked to increased use and lower prostate cancer detection. Health Aff (Millwood). 2012;31:741749.
  • 36
    Nallamothu BK, Rogers MA, Chernew ME, Krumholz HM, Eagle KA, Birkmeyer JD. Opening of specialty cardiac hospitals and use of coronary revascularization in medicare beneficiaries. JAMA. 2007;297:962968.
  • 37
    Mitchell JM, Sunshine JH. Consequences of physicians ownership of health care facilities-joint ventures in radiation therapy. N Engl J Med. 1992;327:14971501.