2005 HSR&D National Meeting Abstract
1067 — Variation in Inappropriate Prescribing for Older Veterans: The Impact of System-Level Factors
Pugh MV (Veterans Evidence-based Research, Dissemination, and Implementation Center)
Chang BH (Center for Health Quality, Outcomes, and Economic Research)
Burke M (Pharmacy Benefit Management Strategic Health Group)
Amuan M (Center for Health Quality, Outcomes, and Economic Research)
Fincke BG (Center for Health Quality, Outcomes, and Economic Research)
Bierman A (University of Toronto)
Rosen AK (Center for Health Quality, Outcomes, and Economic Research)
Cunningham F (Pharmacy Benefit Management Strategic Health Group)
Berlowitz DR (Center for Health Quality, Outcomes, and Economic Research)
Recent research indicates that 18% of older veterans receive medications that are inappropriate for the elderly, which expose them to risk for falls, hip fractures, cognitive impairment, diminished independence, and death. This study seeks to identify system-level factors associated with receiving inappropriate medications.
We merged data from the National Patient Care Database (inpatient and outpatient) and national outpatient pharmacy data for veterans >65 years (FY00), and identified veterans receiving inappropriate medications (per Beers expert consensus panels), while accounting for appropriate use defined by researchers from the Agency for Healthcare Research and Quality. We calculated case-mix adjusted rates of inappropriate prescribing for facilities and conducted hierarchical Bayesian linear models (HBLM) to identify system-level predictors (VISN, teaching status, urban vs. rural, availability of geriatric care) of inappropriate prescribing, and qualitative analysis of readily available Pharmacy Benefit Management Strategic Health Group documents to help understand these findings.
HBLM analysis identified significant VISN level variation in inappropriate prescribing (lowest 20%; highest 35%). Nine VISNs had significantly higher rates than the lowest rate VISN (posteria probability <.025). Furthermore, facilities without medical school and residency program affiliations had lower risk-adjusted inappropriate prescribing than facilities with medical school and residency program affiliations (posteria probability=.02). Our qualitative analysis revealed that the VISN with the lowest likelihood of risk-adjusted inappropriate prescribing removed propoxyphene (the most commonly used inappropriate medication) from its formulary in FY99, and the two VISNs that added propoxyphene in FY98 had significantly higher risk-adjusted inappropriate medication use (29% and 35% vs. 20%).
Even after facility effects were accounted for, variation in inappropriate prescribing by VISN was substantial. These differences were associated with variation in VISN formularies. However, formulary variation did not account for all differences in inappropriate prescribing. Furthermore, non-teaching hospitals had lower risk-adjusted inappropriate medication use than teaching hospitals.
The association of inappropriate prescribing with teaching hospitals and formulary differences suggests that system-level interventions are likely to have a large impact on reducing PIPE. Further research is required to identify other system-level barriers amenable to interventions.