Ambulatory Care Fragmentation and Incident Stroke

Abstract

Background More fragmented ambulatory care (ie, care spread across many providers without a dominant provider) has been associated with excess emergency department and inpatient care. We sought to determine whether more fragmented ambulatory care is associated with an increase in the hazard of incident stroke, overall and stratified by health status and by race. Methods and Results We conducted a secondary analysis of data from the REGARDS (Reasons for Geographic and Racial Differences in Stroke) study (2003-2016), including participants aged ≥65 years who had linked Medicare fee-for-service claims and no history of stroke (N=12 510). We measured fragmentation of care with the reversed Bice-Boxerman index. We used Poisson models to determine the association between fragmentation and adjudicated incident stroke. The average age of participants was 70.5 years; 53% were women, 32% were Black participants, and 16% were participants with fair or poor health. Overall, the adjusted rate of incident stroke was similar for high versus low fragmentation (8.2 versus 8.1 per 1000 person-years, respectively; P=0.89). Among participants with fair or poor self-rated health, having high versus low fragmentation was associated with a trend toward a higher adjusted rate of incident strokes (14.8 versus 10.4 per 1000 person-years, respectively; P=0.067). Among Black participants with fair or poor self-rated health, having high versus low fragmentation was associated with a higher adjusted rate of strokes (19.3 versus 10.3 per 1000 person-years, respectively; P=0.02). Conclusions Highly fragmented ambulatory care is independently associated with incident stroke among Black individuals with fair or poor health.

Keywords: ambulatory care; medicare; stroke.

Figure 1. Adjusted rates of incident strokes per 1000 person‐years with 95% CIs, stratified by healthcare fragmentation, self‐rated health, and race.*

*Fragmentation is defined using the reversed Bice‐Boxerman index (high ≥0.85, low <0.85) and modeled as a time‐varying exposure. Adjusted rates were derived from multivariable Poisson models; see Table 3 for more details. The lower bound of the CI for low fragmentation—Black race—excellent self‐rated health was equal to −1.6