Practice Competition and Treatment of Newly Diagnosed Prostate Cancer

Abstract

Objective: To examine the effect of urology practice market competition on use of treatment in men with newly diagnosed prostate cancer.

Methods: We performed a retrospective national cohort study of 48,067 Medicare beneficiaries with newly diagnosed prostate cancer between 2014 and 2018. The primary exposure was urology practice-level market competition. Markets were established by the flow of patients to a practice using a variable radius approach. Practice level competition was measured annually using the Herfindahl-Hirschman Index. The primary outcome was use of treatment for prostate cancer (ie, surgery, radiation, or cryotherapy) stratified by 10-year risk of noncancer mortality.

Results: Between 2014 and 2018, there was a decrease in the total percent of urologists practicing in small single-specialty groups (49%-41%) with an increase in multispecialty practices (38%-47%). After adjusting for demographic and clinical characteristics, a lower percentage of men underwent treatment in practices with low competition relative to those managed in practices with high competition (70% vs 67.0%, P < .001). Among men with the highest risk of noncancer mortality, those managed in practices in the least competitive markets were less likely to receive treatment relative to men managed by practices in the most competitive markets (48% vs 60%, P-value<.001).

Conclusion: Reduction in competition between urology practices is not associated with greater use of treatment in men with newly diagnosed prostate cancer, particularly in those with a high risk of noncancer mortality.

Figure 1.

A) Distribution of urology practices stratified by practices organization (i.e., small, medium, large, and multispecialty groups) and year. B) Number of urologists stratified by practice organization and year. Abbreviations: MSG: multispecialty group

Figure 2.

Adjusted percent of men undergoing treatment for prostate cancer by level of market competitiveness, stratified by 10-year non-cancer mortality risk: A) low B) intermediate C) high, and D) very high. p-values are derived from test for linear trend of adjusted proportions from the logistic regression models. Models are adjusted for age, year of diagnosis, comorbidity, socioeconomic status, race, Medicare dual eligibility, rurality, and practice organization.

Figure 2.

Adjusted percent of men undergoing treatment for prostate cancer by level of market competitiveness, stratified by 10-year non-cancer mortality risk: A) low B) intermediate C) high, and D) very high. p-values are derived from test for linear trend of adjusted proportions from the logistic regression models. Models are adjusted for age, year of diagnosis, comorbidity, socioeconomic status, race, Medicare dual eligibility, rurality, and practice organization.