Pooled Data Analysis of Long-Term Treatment Effect
Our client had a treatment for kidney disease that had been shown to slow kidney function decline in various short-term clinical studies. They wanted to assess their treatment compared with standard of care on the annual rate of decline in kidney function over a longer period, using existing clinical trial data.
Our client had data from approximately 10 clinical studies, including single-arm phase 2 studies, randomized phase 3 studies, and observational studies. While each individual study had a relatively short treatment duration, they had patient data from nearly 10 years. Our team had to address database structure changes and treatment gaps between studies.
- We constructed a pooled database that harmonized relevant data from all studies and linked data from multiple studies for longer treatment durations.
- Our researchers requested data from NIH-sponsored studies to obtain the long-term historic control of patients receiving standard of care.
- We adjusted for confounding using covariate matching and regression adjustment to obtain a valid treatment comparison without treatment randomization.
- The team employed piecewise mixed models to account for treatment gaps.