Striving for a faster diagnosis for patients with rare and difficult to diagnose disease
The CAPTURED trial will evaluate a series of quality improvement programmes each designed to reduce the time to diagnosis for patients with rare and difficult to diagnose conditions. The following steps outline the trial design.
1. Algorithm of disease algorithm and randomisation of practices
The disease algorithm will be run on OPCRD, an anonymised dataset to identify practices that have patients at high probability of specific rare or difficult to diagnose disease. The algorithms use digital biomarkers in patients electronic medical record data such as symptoms, diagnoses and healthcare utilisation to identify patients who would benefit from specific diagnostic testing. Individual projects will have specific cut offs to determine high probability of disease but typically it is a likelihood of 60% or more.
The practices will then be randomised into clusters for a stepped-wedge implementation of the intervention:
The stepped wedge model enables comparison between the intervention and standard care. By the end, all practices will benefit from the intervention. Anonymised patient data will be collected for each intervention time point, as per the stepped wedge design. For example, in Phase 4 above, outcomes for control practices E, F & G are collected. As you move across time all practices move from control to intervention at different points in time, so, all practices contribute some data to both control and intervention periods. All data evaluation and analysis for CAPTURED will occur at the end of the 5-year study period.
2. The Quality Improvement Intervention
The quality improvement support will be delivered by a trained quality improvement team working on behalf of the GP practice and will comprise of fa full review of the electronic health record for patients identified as high-probability for disease, offer of specific diagnostic testing for eligible patients and support for practices and patients for any necessary next steps.
