Healthcare & Clinical Research

Client Context

A clinical research unit at a teaching hospital was preparing a National Institute for Health Research (NIHR) grant application for a multi-site trial evaluating a digital health intervention for managing Type 2 diabetes. They needed a robust sample size calculation, a statistical analysis plan, and a data management strategy that satisfied both the funder and the NHS Research Ethics Committee.

The Challenge

The trial design was a pragmatic, two-arm, parallel-group RCT with a primary outcome of HbA1c change at 6 months. Recruitment targets needed to account for an expected 20% attrition rate, and the analysis had to adjust for baseline HbA1c and site effects. The research team also required an interim analysis framework for the Data Monitoring Committee.

Our Approach

We calculated the sample size using ANCOVA-based methods, accounting for the correlation between baseline and follow-up HbA1c. With an effect size of 0.35, 90% power, and two-sided alpha of 0.05, we determined that 280 participants per arm were needed, inflated to 350 per arm after accounting for attrition. We drafted the SAP with pre-specified primary and secondary analyses, handling of missing data under both complete-case and multiple-imputation approaches, and an interim analysis plan using O'Brien-Fleming boundaries. We also designed the case report form and data validation procedures in REDCap.

Results

The NIHR funded the trial. Recruitment reached 92% of target within the planned timeframe, and the Data Monitoring Committee approved the interim analysis plan without modifications. The research team credited the "exceptional statistical rigour" of the application as a key factor in the successful funding bid.

Client Testimonial

"AnalyticsScholar turned our rough idea into a fundable trial protocol in under three weeks. Their sample size justification and SAP were described by reviewers as exemplary."

— Professor J. Morton, Chief Investigator

[Explore More Case Studies → /case-studies/]

[Discuss Your Healthcare Project → /contact/]