Development and Validation of a Computerized Adaptive Test of Fatigability

PROJECT ABSTRACT Fatigue is a common debilitating symptom that inhibits activity and functional recovery in hospitalized older adults. While clinicians consider treating fatigue to mitigate its adverse effects, there are no data to guide them on the comparative effects of different interventions. This is in part because fatigue is a subjective self-reported symptom that does not account for patient’s activity level. This is a critical limitation since fatigue and activity influence each other, and interventions that reduce fatigue may increase activity, which in turn may offset initial reductions in fatigue. As a result, trials measuring fatigue only will miss important changes in patient activity, and incompletely estimate the effect of the intervention. A solution to this problem is to measure patient’s fatigue in the context of activity, a measure known as fatigability. Fatigability describes the level of fatigue a patient experiences at any given activity level, and measuring fatigability in hospitalized older adults would be useful for: 1) identifying patients at high risk for functional decline because their fatigue is inhibiting activity, 2) estimating the effect of interventions on changes in fatigue and/or activity, and 3) monitoring longitudinal changes in fatigue and/or activity during hospitalization and after discharge. However, currently available fatigability measures are costly and difficult to operationalize in the inpatient setting (treadmill testing, 6-minute walk), or have been poorly validated using rudimentary statistical methods. The proposed research will address these limitations by building and validating the Computerized Adaptive Test-Fatigability (CAT-F), using multidimensional item response theory (MIRT) to improve the precision, accuracy, and rapidity by which clinicians and scientists can measure fatigability in hospitalized older adults. Aim 1 will apply MIRT to previously collected self-reported fatigability, fatigue, and activity data from hospitalized older adults to develop an adaptive fatigability test that has increased precision while using fewer items compared to currently available fatigability measurement methods. Aim 2 will validate the CAT-F against the current gold standard measures of fatigability in a subset of 75 hospitalized older adults. Future phase 2 work will integrate the validated CAT-F into the EHR at the University of Chicago Medical Center, pilot the CAT-F in a sample of clinicians and patients that are part of an ongoing study of hospitalized older adults, and refine the CAT-F based on user feed-back and additional data collected. Innovation and impact: The CAT-F will integrate data from multiple domains (fatigue and activity) to measure a common latent variable (fatigability) that has significant downstream clinical consequences in hospitalized older adults. The CAT-F will directly improve clinical care by providing clinically relevant information at the point of care through the EHR for clinicians to access and act on, and will increase the efficiency of conducting studies and trials of interventions aimed at reducing fatigue and/or increasing activity. As a result, the CAT-F will improve the care and outcomes of hospitalized older adults, and advance the science of fatigue through efficient high-quality measurement and data generation that informs the optimal treatment strategies for fatigue.