- Analytics use case(s): Population-Level Estimation
- Study type: Methods Research
- Tags: -
- Study lead: Martijn Schuemie
- Study lead forums tag: schuemie
- Study start date: October 9, 2025
- Study end date: -
- Protocol: -
- Publications: -
- Results explorer: -
Extending our calibration framework to a multi-database setting, combining it with OHDSI's Bayesian random-effects meta-analysis framework. Because systematic error is likely correlated between databases (e.g. the same confoudning by indication will likely occur in many places), we cannot simply meta-analyse the calibrated estimates. Instead both calibration and evidence synthesis must be combined in a single approach that accounts for these correlations.