Health Services Research & Development

Veterans Crisis Line Badge
Go to the ORD website
Go to the QUERI website

2007 HSR&D National Meeting Abstract

Printable View

National Meeting 2007

1032 — Harnessing the Power of VA Data for Identifying Medication Adherence for Hypertension

Hogan MM (VA Ann Arbor COE) , Heisler M (VA Ann Arbor COE), Kerr EA (VA Ann Arbor COE), Hofer TP (VA Ann Arbor COE)

Despite VHA quality improvements, more than one-fourth of patients with diabetes do not have good blood pressure (BP) control. VA electronic pharmacy and clinical data systems have the potential to identify BP medication intensification and adherence problems, but this information is not in an easily usable form. The aim of this VA-funded QUERI-related study was to develop methods to make valid information about adherence problems available to providers.

We developed algorithms for an adherence measure for outpatient BP medications based on the Continuous Multiple interval measure of Gaps in therapy (CMG). Our measure is defined as the number of days the patient did not have a medication available (gap days) divided by the total number of days the patient should have been on a medication (should have days). Our refinements added rules to account for VHA hospitalizations, use of leftover pills when there is a dose change, and changes within a medication class. We developed a person-level, by day, by medication class methodology that can be applied to any specified time interval, for example, a calendar year or the year prior to a specific elevated BP event. The procedure identifies for each day and class of medication, whether a medication dose should have been taken, whether a medication dose was available, whether the patient was hospitalized, and whether there was a gap day because there was no medication available. Additionally, rolling counts of gap days before each fill and rolling counts of gap days and should have days over the time interval are produced. Summary records generated for each medication class produce variables, e.g., numbers of should have days and gap days over the specified time interval, which allow for a variety of person or event-level analyses.

To validate the measure, physicians reviewed 103 patient records to identify alternative explanations for gaps besides adherence problems, e.g., non-VA hospitalizations or medications. Other explanations were found for 15 percent of 871 refill gaps identified in pharmacy data, and only 2 percent were accounted for by non-VA prescriptions. In a related pilot, 9 of 11 patients, identified as having a gap from pharmacy data, confirmed problems with medication adherence after a counseling session with a pharmacist. We applied the measure to patients with diabetes in one VISN who had fills for BP medication in 2004 and at least one BP and PCP visit in 2005. Among these 27,365 patients, 36 percent had 20 percent or more gap days relative to should have days in at least one medication class. Applying the methodology to elevated blood pressure events among the same population (28,242 elevated BP events in 2005), we found that 43 percent of elevated BP events had 20 percent or more gap days relative to should have days in at least one class in the prior year.

The vast majority of patients identified from pharmacy records as having refill gaps were likely to be having difficulties with medication adherence. We refined, tested, and used a measure of adherence for hypertension medications using automated VA data sources that produces information usable by clinicians.

Significant gaps in BP medications are common. Ultimately we plan to implement this method to develop clear graphical displays of medication usage within CPRS for BP, lipid, and glycemic management.