Search

This study examines the proportion of men treated primarily with active surveillance across practices and among urologists in the Michigan Urological Surgery Improvement Collaborative.

MUSIC is a statewide consortium of 42 urology practices that aims to improve the quality of prostate cancer care in Michigan. As an initial priority, we examined patterns of care in the radiographic staging of men with newly diagnosed prostate cancer. We determined whether collaborative-wide data review and performance feedback would decrease the imaging rate in men with low risk prostate cancer.Practices submitted standardized data, including the use and results of staging computerized tomography and bone scan, to a web based clinical registry of all men with newly diagnosed prostate cancer. We identified all patients with low risk prostate cancer and compared imaging use patterns before and after practice level performance feedback and guideline review, which were provided at collaborative-wide meetings.In MUSIC 813 patients were newly diagnosed with low risk prostate cancer during the 19-month study period. Of 410 patients diagnosed in the prefeedback period (phase I) 15 (3.7%) and 21 (5.2%) underwent bone scan and computerized tomography, respectively. Of 403 patients diagnosed after feedback (phase II) radiographic staging was done in 5 men (1.3%) with bone scan and in 13 (3.2%) with computerized tomography (p = 0.03 and 0.17, respectively).The overall rate of radiographic staging in men with newly diagnosed low risk prostate cancer was appropriately low. The imaging rate decreased even further after collaborative education and performance feedback. MUSIC appears to be a successful tool for quality improvement, affecting practice patterns and increasing efficiency of care.

INTRODUCTION AND OBJECTIVES: Nationally endorsed prostate cancer quality of care measures are thought to be reflective of quality prostate cancer care. Therefore, it is possible that healthcare systems performing well on these measures will have better outcomes. However, some feel that these measures represent a low bar and thus may not be tightly linked with outcomes. We assessed relationships between healthcare system performance on nationally endorsed prostate cancer quality of care measures and outcomes of prostate cancer treatment. METHODS: We used Surveillance Epidemiology and End Results e Medicare linked data on 48,050 Medicare beneficiaries diagnosed with localized prostate cancer between 2004 and 2009. Based on a composite quality measure incorporating five nationally endorsed individual measures, we categorized healthcare systems into 1-star (bottom 20%), 2-star (middle 60%), and 3-star (top 20%) systems. We assessed relationships between star-ranking and outcomes of treatment with logistic and Cox proportional hazards regression, adjusting for patientand market-level covariates. RESULTS: Patients who underwent prostatectomy in 3-star versus 1-star healthcare systems had a lower risk of perioperative complications (odds ratio 0.80, 95% confidence interval [CI] 0.64-1.00). However, these patients were more likely to undergo a procedure addressing treatment-related morbidity (e.g., 11.3% vs. 7.8% treated for sexual morbidity, p1⁄40.043, Figure). Among patients who underwent radiotherapy, healthcare system quality was not associated with undergoing procedures addressing treatment-related morbidity (Figure). Among all patients, star-ranking was not significantly associated with all-cause mortality (Hazard Ratio [HR] 0.99, 95% CI 0.84-1.15) or secondary cancer therapy (HR 1.04, 95% CI 0.91-1.20). CONCLUSIONS: We found no consistent associations between healthcare system quality and outcomes. Our findings highlight the need for development of quality measures that are more tightly linked to outcomes. Source of Funding: American Cancer Society PF-12-118-01CPPB.

performed to model the most parsimonious and accurate combination of clinical characteristics predicting the presence of UD at final pathology. RESULTS: Overall, 320 (15.4%) patients met the criteria to be enrolled. Mean number of cores taken was 16 (median 14). However, 89 patients (27.8%) harboured UD at final pathology. When PSA, PSA density or clinical stage were excluded from the criteria, only few more patients could be enrolled [+39 pts (+12.2%), +25 (+7.8%) or +2 (+0.6%), respectively] and virtually the same prevalence of patients harboured UD (26.5% vs. 28.1% vs. 27.6%, respectively; p1⁄40.8). When GS was excluded, 360 more patients could be enrolled [+40 (+12.5%)], although this led to a higher prevalence of UD (119/360, 33.1%; p1⁄40.03). Similarly, when number of positive cores was excluded, 887 patients [+567, +177.2%] patients fulfilled the criteria but up to 35.7% (317/887) harboured UD. In the most parsimonious model (biopsy GS 6 and percentage of positive cores

INTRODUCTION AND OBJECTIVES: The Michigan Urological Surgery Improvement Collaborative (MUSIC) is a state-wide consortium of 29 urology practices funded by Blue Cross Blue Shield with the goal of improving the quality of prostate cancer care in Michigan. As an initial priority we examined practice-level and temporal variations in the use of radiographic staging among men with newly diagnosed prostate cancer. METHODS: Practices submit standardized data, including use and results of staging computed tomography and bone scans, to a webbased clinical registry for all men with newly-diagnosed prostate cancer. We compared patterns of imaging utilization across patient risk groups before (Phase I, March-October 2012) and after (Phase II, November 2012-June 2013) practice-level performance feedback and guideline review provided at one of our tri-annual collaborative-wide meetings. RESULTS: Across all MUSIC practices in Phase I, imaging studieswereorderedwith significant variability amongpractices (p 0.001) for approximately 5%, 20%, and 70% of low, intermediate, and high risk patients, respectively. After practice-specific feedback there was a statistically significant decreased in imaging for low risk disease (Table 1). CONCLUSIONS: Although variability exists among urology practices in Michigan, patterns of care regarding the use of imaging studies seem to be quite appropriate and significantly improving for low risk disease based on collaborative education and performance feedback. Opportunities exist for further improvement in the rates of imaging and refinement of imaging criteria and risk classifications. Table 1 Rates of radiographic staging in phase I (March 2012 October 2012) compared to phase II (November 2012 June 2013)

There remains significant controversy surrounding the optimal criteria for recommending prostate biopsy. To examine this issue further urologists in MUSIC assessed statewide prostate biopsy practice patterns and variation in prostate cancer detection.MUSIC is a statewide, physician led collaborative designed to improve prostate cancer care. From March 2012 through June 2013 at 17 MUSIC practices standardized clinical and pathological data were collected on a total of 3,015 men undergoing first-time prostate biopsy. We examined pathological biopsy outcomes according to patient characteristics and across MUSIC practices.The average cancer detection rate was 52% with significant variability across MUSIC practices (range 43% to 70%, p0.0001). Of all patients biopsied 27% were older than 69 years, ranging from 19% to 36% at individual practices. Men with prostate specific antigen less than 4 ng/ml comprised an average of 26% of the study population (range 10% to 37%). The detection rate in patients older than 69 years ranged from 42% to 86% at individual practices (p=0.0008). In the 793 patients with prostate specific antigen less than 4 ng/ml the cancer detection rate ranged from 22% to 58% across individual practices (p=0.0065). The predicted probability of cancer detection varied significantly across MUSIC practices even after adjusting for patient age, prostate specific antigen, prostate size, family history and digital rectal examination findings (p0.0001).While overall detection rates are higher than previously reported, the cancer yield of prostate biopsy varies widely across urology practices in Michigan. These data serve as a foundation for our efforts to understand and improve patient selection for prostate biopsy.

While transrectal prostate biopsy is the cornerstone of prostate cancer diagnosis, serious post-biopsy infectious complications are reported to be increasing. A better understanding of the true prevalence and microbiology of these events is needed to guide quality improvement in this area and ultimately better early detection practices.Using data from the MUSIC registry we identified all men who underwent transrectal prostate biopsy at 21 practices in Michigan from March 2012 to June 2013. Trained data abstractors recorded pertinent data including prophylactic antibiotics and all biopsy related hospitalizations. Claims data and followup telephone calls were used for validation. All men admitted to the hospital for an infectious complication were identified and their culture data were obtained. We then compared the frequency of infection related hospitalization rates across practices and according to antibiotic prophylaxis in concordance with AUA best practice recommendations.The overall 30-day hospital admission rate after prostate biopsy was 0.97%, ranging from 0% to 4.2% across 21 MUSIC practices. Of these hospital admissions 95% were for infectious complications and the majority of cultures identified fluoroquinolone resistant organisms. AUA concordant antibiotics were administered in 96.3% of biopsies. Patients on noncompliant antibiotic regimens were significantly more likely to be hospitalized for infectious complications (3.8% vs 0.89%, p=0.0026).Infection related hospitalizations occur in approximately 1% of men undergoing prostate biopsy in Michigan. Our findings suggest that many of these events could be avoided by implementing new protocols (eg culture specific or augmented antibiotic prophylaxis) that adhere to AUA best practice recommendations and address fluoroquinolone resistance.

RESULTS: Surgeons varied widely in their prediction of operative risk (Table) and on average, 40% of surgeons chose to recommend operation and 60% chose not to. For example, for a case of mesenteric ischemia, surgeons without the RC predicted a 64% chance of serious complication; those exposed to the RC predicted a 44% chance (RC prediction of serious complications, 25%). Although the RC changed the surgeons’ perception of risk, it did not change their decision to recommend operation (69% vs 66%; p1⁄40.34). These findings were consistent across all 4 vignettes.