Blood pressure control and the patient centered medical home --- What can we learn from patients?

Research question

We examine the relationship between the risk of high blood pressure (BP<40/90) (HBP) and both an infrastructure measure and a patient based measure of the “Patient Centered Medical Home” (PCMH).

Background

The PCMH model of care encourages physicians to operate in an environment supporting systematic care management enabled by use of health informatics tools, coordination of care among different providers, and fostering a partnership between patient and physician to produce the best health outcomes. Numerous efforts are underway to demonstrate the effectiveness of PCMH to enhance quality-of-care. Notably, NCQA's PCMH recognition is based on measures of infrastructure related to patient-centered care rather than measures of the patient's perceptions of the quality of provider care.

Methodology

American Board of Internal Medicine's (ABIM) Practice Improvement Module (PIMSM) is used physicians to complete their requirements for maintenance of certification. We use data from the 2008 hypertension and diabetes PIMSM. These PIMsSM include chart audits (7,043), patient surveys (6,793), and a practice-system-survey (592).

Applying these data, we construct a system-level measure (PPC-index-proxy) that closely mirrors an instrument commonly used to qualify practices as medical homes (NCQA's PPC®-index). We also apply a measure of patient perception of physician quality (PPPQ) and a patient-based measure of patient centered care (PPCC) (i.e., measures of communication, access to care, care coordination, quality of staff interaction) drawn from patient surveys.

Controlling for patient and physician demographics, we model HBP as a function of the system-level measure, patients' overall perception of quality-of-care, and our PPCC measures (controlling for patient and physician demographics). To account for the dichotomous nature of our dependent measure and the hierarchical nature of our data, we model these relationships using a random-effects logistic model.

Results

All three of our quality-of-care measures have a clinically and statistically significant association with HBP. For example: a 50% increase in the overall share of patients that believe their care is of high quality is associated with a 12% reduction in HBP (p <.05), doubling of the share of a physician practice whose patients meet all our PPCC standards (from a mean of 31% to 61%) is associated with a 6% reduction in the relative risk of high blood pressure (p <.10) and a 50% increase in the PPC-index is associated with a 7% reduction in HBP (p <.05).

Conclusions

Our findings suggest that both infrastructure measures and patient-based measures are important predictors of blood pressure control.