Multi-specialist contact may be beneficial for patients diagnosed with advanced prostate cancer (PCa) and ensure that they receive comprehensive care and select the most appropriate, evidence-based treatment. Little is known about specialist visit patterns among patients with advanced PCa in the period immediately following diagnosis. We investigate whether metastatic (M1) PCa is associated with increased variation in the types of physicians seen following diagnosis compared to stage IV non-metastatic PCa.
Methods:
We used the Surveillance, Epidemiology and End results (SEER) cancer registry data linked with Medicare claims data for stage IV PCa cases diagnosed between 2000 and 2007. Multi-specialist contact was measured using the physician visit index (PVI), which is derived from the Herfindahl index and used to measure concentration of care across physician types. The PVI was calculated using claims indicating a face-to-face visit with a urologist, radiation oncologist, medical oncologist, nuclear medicine physician, or primary care physician in the six months (6m) following diagnosis. The PVI ranged from 0 (many different types of physicians seen) to 1 (a single type of physician seen). To identify the clinical and demographic factors associated with a lower PVI, we estimated covariate-adjusted marginal effects using a beta regression model.
Results:
The sample included 9,826 men diagnosed with incident stage IV PCa of whom 74% (N=7,301) were diagnosed with M1 PCa. The average age in the sample was 77yrs and 13% (N=1,291) were African American (AA). Based on a multivariable beta regression, an incident diagnosis of M1 PCa was associated with a decrease in the 6m PVI of 0.03 (95% CI: -0.04 to -0.02). The 6m PVI also was statistically significantly lower (more types of physicians seen) among individuals residing in an urban area, with prior use of preventive services or with a non-zero CCI score compared to individuals in the reference categories. The PVI was statistically significantly higher (fewer types of physicians seen) among AA men, individuals above 85yrs, and individuals with poor functional status.
Conclusions:
The number of physician types seen following PCa diagnosis varies according to metastatic disease status as well as patient race and comorbidity status. Given the relationship between physician contact and treatment receipt, the downstream implications of these clinical and demographic differences in PVI score are worth further investigation.