The applicability of quality improvement research for comparative effectiveness

Presentation Quality improvement research (QIR) and comparative effectiveness research (CER) share a common goal of achieving a higher-performing healthcare system. Both CER and QIR have grown considerably over the past decade, CER out of the need for more relevant information for clinical and policy decisions, QIR as a result of the increasing attention to the uneven quality of healthcare [1]. Although the two fields have often favored different research methods – CER often relies on direct, controlled experimental comparisons while QIR often favors single arm studies in a “real world” context, both of these methodologies are needed for improving patient care. Future studies that incorporate elements of both disciplines will provide a context for understanding the most effective and efficient methods for changing clinical practice and ultimately improving patient outcomes. Better comparative methods for QIR would allow us to select the best quality improvement strategies for a given clinical setting. All QI interventions have costs (including opportunity costs) and none work in all settings and circumstances. As an example, quality improvement methods within the Veterans Administration (VA) have employed a variety of approaches which individually have evidence of effectiveness. These include: • Provider education • Patient education and support • Electronic health records with clinical reminders or clinical decision support • National formulary policies • Performance measurement and reporting • System re-engineering approaches and practice redesign • Patient registries • Change initiatives using collaboratives, champions, and toolkits • Provider and management incentives What we don’t know, however, is which interventions or bundles of interventions are most effective and efficient, with fewest harms or unintended effects, for specific quality improvement aims. Clinical reminders, for example, are easy to implement but their effectiveness varies and overusing reminders can lead to “reminder fatigue” and clinician resentment. In an era when frontline clinicians are feeling pressed by increasing responsibility and decreasing time, it is critical to match QI interventions to the specific needs [2]. Reliable comparisons of QI interventions will also require more complete descriptions of the context in which the improvement efforts are being undertaken. For example, although we know provider incentives can be effective tools for changing provider behavior, we don’t know the best ways to target them or how to set the right level of incentive. Recent work compared incentives aimed at physician groups vs. individuals and incentives targeting clinical teams vs. only physicians [3]. Understanding the context in which individuals are more or less responsive to incentives would inform future implementation efforts. QIR would also benefit from focusing attention on the marginal benefits of additional elements of QI interventions (for example, adding performance measurement to system re-engineering approaches), the unintended consequences of QI interventions, and the more complete assessment of the budget impact and business case for specific QI strategies.