Psychometric evaluation and statistical analysis of COAs

Psychometric evaluation

Our psychometricians apply many traditional and modern psychometric methods according to the specific needs for each unique project. We can offer strategic advice during the whole clinical outcome assessment (COA) development cycle including the writing of full psychometric analysis plans, analyses, recommendations for refining the COA, advice on scoring and interpretation, and full report/manuscript writing. Our team have experience with many different types of COAs across different disease areas including daily diaries and computer-adaptive testing. Our programming team ensure that analyses for regulatory purposes is fit for submission where required.

COA aspects of clinical trials

Our statisticians can advise on COA strategy early in the planning stage of product development and advise on implementation including the timing of assessments to optimize comparisons between treatments in randomised controlled trials (RCTs) or to optimize information gathered in single arm trials. We can offer advice on the development of COA endpoints including implementation of the estimand framework, aligning with the International Council for Harmonisation (ICH) E9 guidance and Food and Drug Administration (FDA) Patient-Focussed Drug Development (PFDD) 4 guidance documents.

Our subject knowledge of both statistics and experience in a wide range of diseases and COAs enables us to personalize our advice for each study, optimizing the choice of endpoint and analysis methods for the comparison between treatments or observation of trends over time.

We offer a full service around design, analyses and reporting of COA endpoints in clinical trials including full specification of the analyses in a supplemental Patient-Reported Outcome (PRO) statistical analysis plan, analysis with proper programming (using SAS, R, and MPlus) to the required regulatory standard, interpretation of results, contribution to the clinical study report as well as report and manuscript writing. Our experts not only apply their statistical knowledge but knowledge of the COA scores to aid interpretation, including the derivation of meaningful change thresholds where required.

Score interpretation

Our statisticians and psychometricians work alongside our qualitative researchers to inform the interpretation of scores using mixed methods research pioneered by Adelphi Values. We can offer derivation of thresholds from blinded data cuts from clinical trial data or bespoke studies. We focus on distinguishing between methods for individual change versus those for between or within-group change to derive appropriate thresholds. Using a combination of anchor-based and distribution-based methods to inform us about the limitations of the scale alongside the likely range of meaningful change. We also offer a robust approach to triangulation of multiple estimates based on the strength of the evidence from each anchor.

Methodological research and development

Our team are actively involved in various international initiatives to improve methodology in our field. We can offer additional bespoke services including literature reviews and recommendations on best practice and tailor-made training courses/workshops on a wide range of topics. Previous workshops have included introduction to statistical analysis methods for COAs, application of the estimand framework to COAs, dealing with missing data and derivation of meaningful change.