Lack of Continuity in VA Primary Care Predicts Lapses in Prescribing Quality

Lund BC, Vaughan-Sarrazin MS, Katz DA. Lack of Continuity in VA Primary Care Predicts Lapses in Prescribing Quality. Paper presented at: VA HSR&D / QUERI National Meeting; 2012 Jul 17; National Harbor, MD.

Related HSR&D Project(s)

• CDA 10-017 – The Validity and Utility of Measures of Inappropriate Prescribing

Objectives: Continuity of Care (COC), defined as a continuous caring relationship with an identified health professional, is a core aspect of primary care quality. Recent healthcare changes may be placing this relationship at risk. Diminished COC has been linked to decreases in preventative services and increased hospitalization and emergency department use. Medication prescribing is an important domain of primary care practice and this study determined whether COC predicted future prescribing quality. Methods: We conducted a retrospective cohort study of VA outpatients assigned to a primary care provider, who had at least 3 primary care visits during FY2007-2008 and were regular VA medication users during FY2009 (N = 2,199,573). COC was estimated during FY2007-2008 using Usual Provider of Continuity (UPC), categorized as low ( < 0.50), intermediate (0.50-0.99), and high (1.0, reference group). The Modified Modified Continuity Index (MMCI) served as an alternate COC measure in sensitivity analyses. Prescribing quality violations were identified during FY2009 using two dichotomous indicators: (1) unnecessary duplicate prescribing, and (2) regimen nonadherence (REG-OUT > 0.20, i.e. adherence < 80%). Multivariable random effects logistic regression models were used to predict each prescribing quality violation, controlling for demographics, comorbidity, and assigned parent station (random effect). Results: The mean age was 66 years and included predominantly men (95%). COC was high for 41% of patients, intermediate for 39%, and low for 20%. Duplicate prescribing was observed for 13% of patients, and regimen nonadherence was present in 25%. In multivariable models, low and intermediate COC were associated with duplicate prescribing, with odds ratios of 1.25 (95% CI: 1.24, 1.27) and 1.24 (1.23, 1.25), respectively. Similar effects were seen for nonadherence, with odds ratios of 1.17 (1.16, 1.18) and 1.10 (1.09, 1.10) for poor and intermediate COC, respectively. These relationships were unaffected by comorbid conditions, and were similar in sensitivity analyses using an alternative COC measure. Implications: Lack of COC was associated with lower prescribing quality, including unnecessary duplicate prescribing and nonadherence. Impacts: Prescribing quality can be added to the list of negative outcomes associated with poor COC. Innovative primary care delivery models (e.g., patient-aligned care teams) should enhance, or at least preserve, COC.