Assessing PRISM context domains and RE-AIM outcomes: Data from use of the Iterative PRISM Webtool

TL;DR

This study evaluates the PRISM framework using a webtool to assess how context domains and outcomes relate in implementation science.

Contribution

The study provides empirical validation of PRISM constructs and identifies areas for refinement in the iPRISM webtool.

Findings

Seven of 11 hypotheses about PRISM context domains and RE-AIM outcomes were supported.

Equity ratings were consistently lower than general ratings on RE-AIM dimensions.

Fewer hypotheses about PRISM context items were supported, possibly due to limited items per domain.

Abstract

Use of an implementation science (IS) theory, model, or framework (TMF) is one of the hallmarks of a well-executed IS study. While these days TMFs are almost aways used for IS studies, the TMFs themselves are seldom evaluated. Understanding the relationships between the constructs within an IS TMF and their effect on the implementation and effectiveness outcomes can help to refine the TMF, advance IS, and assist implementers.

We evaluated several hypotheses pertinent to the context domains and Reach, Effectiveness Adoption, Implementation, and Maintenance (RE-AIM) outcomes from the Practical, Robust Implementation and Sustainability Framework (PRISM). Data for these evaluations emerged from the use of the Iterative PRISM (iPRISM) webtool, which includes 21 assessment questions that operationalize PRISM context and outcomes constructs. We tested 11 a priori hypotheses including relationships within and between different framework constructs and considered whether some of the webtool’s assessment questions could be refined or changed to make the assessment more pragmatic.

A total of 348 clinical, community, and public health respondents completed the iPRISM survey as part of using the publicly available webtool. They reported on projects from a wide variety of clinical, community, and public health settings; in English and Spanish; and in different project phases. Seven of the 11 hypotheses were fully or partially supported (e.g., that ratings for Maintenance would be lower than other RE-AIM outcomes). One exception was that the hypothesis that the correlation between Reach and Effectiveness would be the lowest among RE-AIM outcomes was not supported. As hypothesized, scores on the equity items on the various RE-AIM dimensions (e.g., Reach) were consistently lower than general ratings for that dimension. Fewer of the hypotheses about the PRISM context items were supported, possibly due to there being only one item per contextual domain.

The webtool questions provide a standardized way to operationalize PRISM constructs and initial norms for different items. Future research, including qualitative evaluation, is needed to replicate, explore, and understand the complex relationships found within RE-AIM outcomes, within PRISM context domains, and between PRISM context ratings and RE-AIM outcomes.

Figures