Adoption of the Hospitalist Model by Primary Care Providers: An Analysis of Medicare Beneficiaries in Texas

Background:

How and why primary care providers (PCPs) adopt the hospitalist model is not well defined. We sought to characterize the patterns by which PCPs transition from personally caring for their hospitalized patients to referring them to hospitalists. We also examined PCP and patient factors that are associated with adopting the hospitalist model.

Methods:

We conducted a retrospective cohort study using a 100% sample of Medicare beneficiaries in Texas > 66 years old for the years 2001–2009. We defined PCPs as general practitioners, family physicians, general internists, or geriatricians who billed > 3 outpatient evaluation and management (E&M) visits for a beneficiary in the year prior to the beneficiary's hospitalization. When assessing rates by which PCPs transitioned to the hospitalist model, we focused on ‘established PCPs who relinquished inpatient care, defined as those who cared for >20 Medicare beneficiaries every year throughout the study period and who submitted inpatient codes for >33% of their patients in 2001 but <4% in 2009. We defined hospitalists as general internists with at least 5 E&M claims and >90% of yearly E&M claims from inpatient services. Analytic models included beneficiaries' demographic variables, Elixhauser comorbid conditions, admission type, hospital type, and PCP characteristics. Analyses were conducted using descriptive statistics and multilevel models.

Results:

Between 2001 and 2009, the proportion of 608,686 Texas Medicare beneficiaries who received inpatient care from a PCP decreased, whereas the proportion receiving hospitalist care grew. These beneficiaries were associated with 1172 PCPs, of whom 228 were established PCPs who relinquished inpatient care during the study period. The number of established PCPs adopting the hospitalist model in any given year ranged from 18 (8.7%) in 2003 to 58 (54.2%) in 2007. When established PCPs gave up hospital care, they did so rapidly: the proportion of inpatient charges submitted by these PCPs decreased from >45% to <10% within 1 to 2 years. PCPs associated with the same primary hospital were more likely to relinquish inpatient care in the same year (ICC, 8.17–33.84). PCP factors associated with adoption of the hospitalist model include family practice versus internal medicine training (OR, 1.55; 95% CI, 1.35–1.77) and lack of board certification (OR, 1.34; 95% CI, 1.12–1.62). Among patients with established PCPs, patients were more likely to receive inpatient care from their PCP if they were black (OR, 1.13; 95% CI, 1.09–1.16), older (OR, 1.04; 95% CI, 1.03–1.04), Medicaid eligible (OR, 1.15; 95% CI 1.13–1.17), or if they had more comorbidities (P < 0.001 for trend).

Conclusions:

PCPs adopt the hospitalist model at different times, but when they transition, they do so rapidly. When PCPs adopt the hospitalist model, other PCPs within the same hospital are more likely to do so, too. Among patients with established PCPs, healthier, less socioeconomically disadvantaged patients are more likely to receive inpatient care by hospitalists.