A three-step health services research approach to improve prescribing.

Medications are often prescribed suboptimally; some effective medications are underused, some ineffective medications are overused, and some medications that should be received by a few are instead given to many. The underlying causes of suboptimal prescribing likely differ for each medication, and therefore must be understood anew, although previous studies can help generate hypotheses. This perspective sets forth a 3-step research agenda, which has worked well for us in several recently completed and ongoing projects. The three steps are to 1) demonstrate variation in suboptimal prescribing for the targeted medication; 2a) use mixed methods to understand the patient-, provider-, and system-level causes of suboptimal prescribing for this medication; 2b) develop a justification for improving the use of this medication, often involving a business case analysis; and 3) develop and implement interventions to improve prescribing of the targeted medication, informed by what has been learned in Steps 1 and 2 and relying on the principles of implementation science. Previous efforts have focused disproportionately on Step 1, or documenting gaps in practice, and Step 3, or deploying and evaluating efforts to improve practice. Our contention is that addressing all three steps sequentially, while effort-intensive, will maximize the chances of deploying a more effective intervention that will impact population health. We commend this three-step approach to health services researchers who wish to maximize impact by basing their research on a natural progression from documenting problems, to understanding their causes, to formulating and deploying a solution.