A logic model to delineate the outcomes, consequences, and QI implications of CMS’s National Partnership

Context: The National Partnership to Improve Dementia Care in Nursing Homes (NP) engages multiple stakeholder groups and impacts others. To facilitate ongoing improvement in long-stay dementia care, we set out to assess the perspectives of every change stakeholder in this quality-improvement (QI) initiative. We were unable to find, however, a model describing QI change applicable to the NP that explains outcomes, provides actionable advice, and adequately emphasizes the critical importance of all stakeholders.

Objective: Develop a logic model to organize and explain our diverse NP findings across multiple stakeholder groups.

Dataset/Population: 10 years of mixed-methods data regarding NP impacts, engaging as many stakeholders as possible.

Study Design: Applying a 5-pass horizontal process of immersion and crystallization, we created a logic model based on the Theory of Planned Behavior that best explains our diverse NP data.

Results: This logic model for QI change positions all stakeholders as essential components in a dynamic and contextual change process and emphasizes the three overarching attributes of effective QI. Effective QI is adaptive, contextual, and inclusive.

Discussion: The NP was created with ‘broad design and scope’ where all are invited to participate. However, initial stakeholder inclusion in the design process was incomplete, particularly patients and their families. This was a missed opportunity, as was focusing on certain stakeholder groups over others. The NP’s development incorporated tests of change to detect and remediate harm. Notwithstanding, our results suggest a diversity of ongoing undesirable policy effects that could be moderated through a more adaptive process. Finally, the NP gives little consideration to context. Since QI measures engage stakeholders in a contextual environment with inherent biases advantaging and disadvantaging certain groups, sensitivity to individual idiosyncratic impacts could improve NP outcomes.

Conclusions: This logic model highlights 3 opportunities for QI optimization: maximal stakeholder engagement, recognizing the critical role of individual/community context, and adaptation. It is plausible that insufficient attention to these factors is the common thread connecting the preponderance of QI shortcomings. This model may have application for policy makers seeking a stakeholder-centric approach to better design, measure, and improve QI initiatives.