Improving Outcome Measures for Rett Clinical Trials: The Development of Rett-Specific Anchors for the Clinical Global Impression Scales

Background: High quality outcome measures are a critical component to well-designed clinical trials for individuals with neurodevelopmental disorders. The Clinical Global Impression Scale (CGI) (Guy, 1976) is a measure of global clinical change with strong face validity that has been widely used as an outcome measure in CNS clinical trials, including trials in ASD. The CGI is a 7-point Likert rating scale that reflects expert clinical judgment. It includes a severity of illness (CGI-S) scale and an improvement(CGI-I) scale.  Despite its favorable assay sensitivity in clinical trial settings, a disadvantage of the CGI is its lack of focus on the specific signs and symptoms of the disorder under study (Busner et al. 2009).  Development of key anchors for the scale, specific to the disorder being assessed holds promise for enhancing the validity and reliability of the CGI for specific disorders.

Objectives:   We describe the development of novel anchors specific to Rett syndrome (RTT) signs and symptoms for the CGI-I/CGI-S. This effort is part of an on-going clinical trial involving adolescent and adult females with RTT (ClinicalTrials.gov Identifier: NCT01703533), which is the first industry-sponsored, multi-site clinical trial in this population.  RTT is a severe developmental disorder with features of ASD, and our experience using the CGI as an evaluation for RTT signs and symptoms may provide useful insights for the further development of anchors for ASD.

Methods: Utilizing information obtained from the RTT Natural History Study (ClinicalTrials.gov ID: NCT00299312), a classification grid of symptom severity was created, accompanied by anchors providing a description of progressive levels of impairment in core signs and symptoms.  

Results: Based on score ranges from the RTT Natural History Clinical Severity Scale (CSS), a grid was developed defining gradations in the severity of the core signs and symptoms of RTT including: language/communication, ambulation, hand use, eye contact (social), autonomic function, seizures, and attentiveness.   Each gradation was associated with a severity rating number from the CGI-S. For example, a CGI-S rating of 4 was associated with a CSS score range of 10-20. Specific exemplars from the grid were developed into descriptive anchors for the CGI, outlining signs and symptoms observed for each severity level. The anchors also provide descriptive guidance for differentiating ratings between CGI-S levels, when symptoms are not uniformly impaired across domains. For the CGI-I, anchors provide examples of symptom change as well as a framework for considering the duration, onset, durability of change, and context of symptom change across these domains.

24 of the targeted 48 study participants have been enrolled to date (mean age: 25 years, 10 months). Data presented from the full cohort will show the psychometric properties and feasibility of this RTT-specific version of the CGI scales in the context of this clinical trial.

Conclusions: The rating scheme captures clinically relevant gradations in severity and improvement of RTT-related signs and symptoms, offering the prospect of more consistent and relevant administration across research sites and studies, and it provides a framework for the further refinement of developmentally appropriate CGI anchors in ASD.