b iframe width height src https www youtube com embed
federal health policy changes has been to better align pay-ment with quality of care in order to foster the ‘triple aim’
of healthcare reform: improving care and health, while reducing cost.13-17,28,29 Despite the widespread adoption of electronic medical record systems, understanding practice patterns for these evolving models within RIPA Lysis Buffer urologic practice has remained challenging due to the persistently siloed and unstructured nature of healthcare data. Practiced-based research and implementation science have been cited as important ways to improve our understanding of how healthcare is delivered.16 Accordingly, the Large Urology Group Practice Association Active Surveil-lance Benchmarking Project, through an unrestricted grant by Genomic Health, Inc., sought to review and describe contemporary practice patterns in 9 geographi-cally distributed large community urology group practi-ces to provide the insight necessary to improve practice patterns.
MATERIALS AND METHODS
The primary objective of the study was to assess the utilization of AS among men with lower risk prostate cancer (very low and low risk combined). A secondary objective was to examine utili-zation rates of surgery and radiotherapy interventions for a large contemporaneous cohort of newly diagnosed localized prostate cancer patients within a diverse US geographical distribution. Nine member practices form Large Urology Group Practice Association participated in this study: Atlantic Urology Clinics, Myrtle Beach, SC; Carolina Urology Partner, Charlotte, NC; Genesis Healthcare Partners, San Diego, CA; Oregon Urology Institute, Springfield, OR; Skyline Urology, Los Angeles, CA; Urology of Indiana, Indianapolis, IN; The Urology Center of Colorado, Denver, CO; The Urology Group, Cincinnati, OH; and Virginia Urology, VA.
We conducted a retrospective chart review of all men with newly diagnosed prostate cancer between January 1, 2013 and March 31, 2014. AS was defined by documentation of AS within 6 months of diagnosis, or absence of curative therapy with continued follow-up of at least 1 PSA and office visit beyond 6 months. Each practice had its own protocol for AS. Only men diagnosed by prostate biopsy at the participating practices were included, while those with high-risk disease or who were 75 years or older were excluded from the study. Patients without documented primary therapy or without fol-low-up of at least 6 months were excluded. Each participating practice assigned 1 or 2 employees to perform chart abstrac-tion. All abstractors were trained by the principal investigator (JS) and abstraction was overseen by practice clinician cham-pions and JS. We employed the NCCN definitions of prostate cancer disease risk group from 2013 (very low risk = T1c, Gleason 6, PSA <10, 1-2 biopsy cores positive with <50% cancer in any 1 core, and PSA density <0.15 ng/mL; low risk = T1-T2a, Gleason 6, PSA <10; intermediate risk = T2b-T2c or Gleason 7 or PSA 10-20).18 Practice characteristics were assessed by surveying practice leadership. Abstracted data was validated by an independent professional medical abstrac-tion firm (CIOX, Atlanta, GA) through a random sample of 20 cases from 6 of the 9 participating practices. In 3 practices, administrative and legal concerns about Health Insurance Por-tability and Accountability Act of 1996 and related law, and data ownership prevented outside/remote access to the medical record. Only deidentified data was analyzed and IRB approval was obtained.
Descriptive statistics including Chi-square (Fisher’s exact tests when necessary) and median tests were used to characterize the study cohort and to compare the use of initial AS to use of cura-tive therapy and, among those who chose curative therapy, to compare the use of surgery to radiotherapy according to relevant patient and histopathologic characteristics. The association between predictor variables and use of AS, and choice of cura-tive therapy was further evaluated with a multivariable random effects regression model adjusting for age, race, and risk group (Table 2). Sensitivity testing of risk group classification vs its individual components showed that Gleason score and PSA were both independently associated with choice of AS (odds ratio [OR] 2.95, P <.001 for 3 + 3 vs 3 + 4; OR 1.93, P =.002 for PSA <10 vs PSA 10-20. To limit colinearity, only the summa-tive risk group classification was used in the final regression model. To assess variation in primary management between practices we derived the adjusted proportion of patients undergo-ing AS, surgery and radiation in each practice from multivariable model. All statistical testing was performed using SAS v.9.4 (SAS Institute Inc., Cary, NC) or Stata v.13.1 (StataCorp, Col-lege Station, TX) at the 5% significance level. This study was funded in part by an unrestricted grant from Genomic Health, San Francisco, CA.