A qualitative analysis of medication use variance reports

This report of a process change utilized a qualitative approach to data analysis to improve medication use safety in a large hospital. The two goals were to design a strategy to analyze the qualitative data and to use that strategy to uncover previously unclassified medication use variance patterns that could be prevented. A multidisciplinary team performed the analysis in an effort to improve the quality and yield of the approach.

All medication use variance, incident, and event reports from Yale-New Haven Hospital during April-June 2000 were collected (N = 264). A 20% random sample of the reports was distributed to a five-member evaluation group (a pharmacist, two nurses, and two physicians) for independent qualitative analysis and coding. An initial coding framework was produced using a consensus process. This coding framework was applied to another sample, and the consensus and coding processes were repeated until no new domains were identified.

Ten general medication use variance domains were determined. In addition, 21 subdomains among the various general domains were determined.

Utilizing a multidisciplinary team and a qualitative strategy of analysis improved patient safety efforts. This combination led to the discovery of new variance domains, causes, and opportunities to intervene and ultimately prevent medication use variances. This analytic approach is widely applicable, adaptable, and dynamic. The design and results of this report improve on a strictly quantitative approach to medication use variance analysis. The approach employed by this report will be used to improve medication use safety within the Yale-New Haven Health System.