Issue | Recommendation |
---|---|
1. Use of frameworks, theories, and models | • Use theoretical model and include those construct definitions in manuscripts |
• Work towards consensus language as a field | |
• Consider use of the CFIR Wiki for construct definitions | |
2. Need to establish instrument psychometric properties | • Identify, perform, and report results of most appropriate reliability assessments when possible |
• Identify, perform, and report results of most appropriate validity assessments when possible | |
• Attempt to establish psychometric properties while simultaneously investigating factors involved in the implementation process e.g., [52] | |
3. The use of `home-grown’ and adapted instruments | • Utilize SIRC IRP and NCI GEM projects to identify existing instruments for constructs under evaluation |
• Utilize proper test development procedures [27],[33],[34] | |
• Report adaptations (changes in length, language, structure) and updated psychometrics to assess effect of adaptations | |
4. Choosing the most appropriate evaluation method and approach | • Consider utilizing mixed-methods approaches |
• When appropriate, consider utilizing multi-informant, direct observation, and administrative data in addition to or instead of self-report | |
5. Practicality | • Keep instrument costs as low as possible |
• Keep item numbers low (perhaps 10 or fewer) | |
• Provide the instrument items in the published manuscript | |
• Provide the instrument to the NCI GEM project or SIRC IRP | |
• Make the language accessible | |
6. Need for decision-making tools | • Utilize the SIRC IRP and NCI GEM projects to identify instruments |
• Report on any and all results of psychometric analyses |