Exploring the association of care fragmentation and patient ratings of care quality: A mediation analysis of women Veterans' experience with VA care

Abstract

Objective: To examine the relationship between care fragmentation and patient ratings of care quality and identify potentially actionable mediators.

Data sources/study setting: 2015 telephone survey of 1395 women Veterans with three or more visits in primary care and/or women's health care in the prior year at 12 Veterans Affairs (VA) medical centers.

Study design: Cross-sectional analysis.

Data collection/extraction methods: We operationalized lower care fragmentation as receiving VA-only care versus dual use of VA/non-VA care. Participants rated VA care quality (overall care, women's health care (WH), and primary care (PC)) and three aspects of their patient experience (ease of access to services, provider communication, and gender sensitivity of VA environments). We examined associations between care fragmentation and care ratings and applied the Karlson-Holm-Breen decomposition method to test for mediation by aspects of patients' experience.

Principal findings: Lower care fragmentation was associated with higher ratings of care quality (odds ratios [95% CI] for overall care: 1.57 [1.14;2.17]; WH: 1.65 [1.20;2.27]; PC: 1.41 [1.10;1.82]). Relationships were mediated by patient-rated provider communication and gender sensitivity (26-54 percent and 14-15 percent of total effects, respectively). Ease of access was associated with higher care ratings (odds ratios [95% CI] for overall care: 2.93 [2.25;3.81]; WH: 2.81 [2.15;3.68]; PC: 2.33 [1.63;3.33], in models with the three types of patient care experiences included), but did not mediate the association of care fragmentation and care ratings.

Conclusions: Potential negative effects of care fragmentation on care quality ratings could be mitigated by attention to quality of patient-provider communication and gender sensitivity of VA environments.

Keywords: Veterans; access to care; gender; patient experience; patient-provider communication.

Figure 1

Model for the mediation analysis. We hypothesized that the effect of lower care fragmentation on the patient ratings was mediated by the three experience factors: ease of access to care, provider communication, and feeling welcome at VA as a woman as a proxy measure for the gender sensitivity of VA environments. Table 2 presents supporting evidence for the relationship of lower care fragmentation and patient ratings, as well as the associations of patient experience factors and patient ratings. The relationships between lower care fragmentation and two of the candidate mediators, provider communication and feeling welcome, are supported by the results of the regression analyses presented in Table 3, whereas the role of ease of access as mediator was rejected. Based on those results, mediation of the effect of lower care fragmentation on care ratings via provider communication and feeling welcome at VA was further tested (Table 4)