BACKGROUND New prescriptions unfilled by patients (primary non-adherence) contribute to the overwhelming problem associated with medication non-adherence. Little research has been conducted in this area relative to refill non-adherence. The overall impact of unclaimed prescriptions is potentially significant depending on the medication and its indication. Because primary nonadherence has been recognized as an issue within our organization, we are conducting a retrospective study to evaluate fills up to twelve months following a new prescription in order to assess patient and clinical variables associated with primary non-adherence. OBJECTIVE: Our objectives are to compute prescription fill rates among subgroups defined by patient and drug attributes, and to determine how subsequent health care utilization differs for those who do and do not fill their prescriptions
METHODS: This pilot study is a retrospective review of electronic medical record (EMR), prescription orders and fills 1/1/2009-6 30/2009 for patients aged 18+with 12 months of continuous medical and prescription coverage 1/1/2008-6/30/2010 who have a clinic or emergency room visit and a new prescription order for a new diagnosis. We will evaluate this cohort for prescription fill data over the following 12 months (through 6/30/2010) to gather EMR, census, and patient visit data linked to the prescription order.We will quantify the number of patient visits (primary, specialty, and urgent care, ER and inpatient) for members who fill the original prescription, those who ultimately fill with the same or similar drug, and those who never fill the prescription (within one year of prescription order).
RESULTS The data dictionary has been developed and project programming is underway. We anticipate we will complete data collection by January 2011 and analysis by March 2011.
CONCLUSION AND IMPLICATIONS We expect primary non-adherence to occur within our integrated health system and to vary by diagnosis. Variables predicted to vary among those who fill and those who do not include plan design, benefit structure, gender, race, age and SES. This pilot study will be used to identify target populations for future interventional studies.