Developing a utility index for the Aberrant Behavior Checklist (ABC-C) for fragile X syndrome

Abstract

This study aimed to develop a utility index (the ABC-UI) from the Aberrant Behavior Checklist-Community (ABC-C), for use in quantifying the benefit of emerging treatments for fragile X syndrome (FXS). The ABC-C is a proxy-completed assessment of behaviour and is a widely used measure in FXS. A subset of ABC-C items across seven dimensions was identified to include in health state descriptions. This item reduction process was based on item performance, factor analysis and Rasch analysis performed on an observational study dataset, and consultation with five clinical experts and a methodological expert. Dimensions were combined into health states using an orthogonal design and valued using time trade-off (TTO), with lead-time TTO methods used where TTO indicated a state valued as worse than dead. Preference weights were estimated using mean, individual level, ordinary least squares and random-effects maximum likelihood estimation [RE (MLE)] regression models. A representative sample of the UK general public (n = 349; mean age 35.8 years, 58.2 % female) each valued 12 health states. Mean observed values ranged from 0.92 to 0.16 for best to worst health states. The RE (MLE) model performed best based on number of significant coefficients and mean absolute error of 0.018. Mean utilities predicted by the model covered a similar range to that observed. The ABC-UI estimates a wide range of utilities from patient-level FXS ABC-C data, allowing estimation of FXS health-related quality of life impact for economic evaluation from an established FXS clinical trial instrument