The National Institutes of Health (NIH) Patient-Reported Outcomes Measurement Information System (PROMIS) measures were developed to provide highly precise and valid measures of patient-reported outcomes, establish a common metric across chronic health conditions, and enable computerized adaptive testing. The PROMIS tools have not been widely tested in the older adult population, especially those aged 80 and older nor among those with Mild Cognitive Impairment (MCI) and Alzheimer's and Related Dementias (ADRD). The reliability and validity of responses provided by those with MCI or ADRD on PROMIS measures is unknown, as are the threshold on cognitive measures where self-reported PROMIS data are no longer representative of underlying health status.
The Specific Aim of this project is to validate the PROMIS Profile measures in older adults and to identify the threshold where cognitive impairment invalidates these measures. We will test the utility of the PROMIS Profile measures (29-, 43-, and 57-item versions) among older adults with varying levels of cognitive impairment to find thresholds (cut points) where patient-reported data can no longer be reliably or validly attained. Cognitive levels will be assessed with the Montreal Cognitive Assessment (MoCA) and the Mini-Mental State Examination (MMSE). A total of 448 subjects will be enrolled from outpatient practices ensuring equal numbers of subjects, ages 65-79 and 80+ years as well as a spectrum of cognitive function from normal cognition to MCI to ADRD. Qualitative cognitive interviews in 48 older adults will examine patient understanding of individual items and capacity to provide valid information. The cognitive interviews will be used to elucidate patterns of response based on cognitive complexity of the items relative to the cognitive capacity of the participants. Quantitative investigation of the thresholds of cognition on the PROMIS profile measures (29-, 43-, and 57-item versions) will be done in an additional sample of 400 older adults to identify the cognitive threshold where responses become inconsistent (unreliable) and invalid. Scale reliability and item functioning will be assessed in three ways: i) investigating differential item functioning on the PROMIS Profiles for the MoCA and MMSE threshold where patient reports become less internally consistent (using Item- Response Theory modeling); ii) split-half reliability of PROMIS Profiles to identify where internal consistency falls below alpha=0.8; and iii), patterns of responses elucidated in Aim 1 will be examined to identify thresholds where subjects use satisficing strategies. Validity will be assessed by examining the consistency of relationships between PROMIS Profiles and other variables (e.g., collateral proxy reports, self-reports on similar measures) cross-sectionally for concurrent validity and longitudinally (e.g., hospitalizations, falls), to show predictive validity. The study will result in practical data for immediate translation in research and clinical practice on the capacity of older adults with varying levels of cognitive impairment to provide accurate responses on PROMIS measures of disability and psychological distress.
This study will look at people with cognitive problems and Alzheimer's disease or other dementias. It will see how well they can answer for themselves on measures of patient outcomes called PROMIS. Results will help researchers refine patient outcome measures for those with cognitive problems.
