Outcome measures for early phase clinical trials

Abstract

As the number of potential neuroprotective agents for ALS increases, the need for early phase trials that screen drugs before proceeding to efficacy trials also grows. However, it is not known which outcome measures perform best and also provide the most meaningful information in brief small trials. We assessed the performance of different outcome measures for use in early phase clinical trials in ALS, and determined what degree of change in the ALSFRS-R that patients could perceive. Thirty patients underwent six monthly ALS Functional Rating Scale (ALSFRS-R), forced vital capacity, manual muscle testing (MMT) and quality of life assessments. Patients rated their perceived level of change with algorithm scales. Linear mixed effects models assessed the associations among variables and Cox proportional-hazards models examined the ability to predict survival. The quantity of missing data was assessed using descriptive statistics. Correlations were found between all variables. The ALSFRS-R provided the most complete data (99.5%), showed a large within-subject correlation (0.91), and best predicted survival (p = 0.002). One-unit change in patient-perceived clinical function paralleled a 9-point decrease in the ALSFRS-R (p = 0.025; 95% CI 8, 10). This trial assessed just 30 patients over six months, but the standard outcome measures each performed dependably; all could be used in short-duration, early phase trials. The ALSFRS-R most strongly predicted survival and provided the most complete data, but large changes may be necessary before patients perceive treatment effects.