Channelling Algorithm Proficiency To Understand, Recognise and Elucidate Disease

The aim of CAPTURED is to evaluate the efficiency of algorithm technology and tailored primary care quality improvement support in identifying, testing and referring patients at high probability of specific rare or difficult to diagnose disease.

The trial aims to reduce the diagnostic odyssey for rare and difficult to diagnose disease patients and facilitate quicker access to appropriate treatment, utilising algorithms designed to identify patients that are at high probability.

CAPTURED is a series of quality improvement programs (QIPs) that are each delivered in a stepped wedge model, to enable comparison between the QI intervention and standard care. Each QIP, under CAPTURED, will utilise a specific algorithm for rare or difficult to diagnose disease, and a specific diagnostic testing/quality improvement support package. It is planned to run for 5 years (2025-2030). Each QIP will run for a defined period.

CAPTURED is UK-wide multicentre study. All participating practices should be able to accept the intervention (diagnostic test/screening and tailored quality improvement support package) as part of their patient care. CAPTURED will be conducted at an estimated 500 GP practices in the UK. Only practices with patients identified as at high probability of specific rare or difficult to diagnose disease will participate in each project. 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.

Practices with patients with high probability of disease will be randomised into clusters. Randomisation will occur individually for each QIP. The intervention (diagnostic testing/screening and quality improvement support package) will be given at different time-points defined by the stepped wedge design.

The index date (time = 0) will be defined as the date that the first cluster of practices receives communication about having patients identified as at high probability of rare or difficult to diagnose disease. Communication with remaining clusters will be sent out at the exact time points defined by the stepped wedge design but will be specific to each project.

The primary objective of CAPTURED is to demonstrate that the patient testing pathway works defined as:

• To determine if algorithm technology and quality improvement support increases rates of diagnosis of rare or difficult to diagnose disease.

The secondary objectives of the trial are outlined below:

• To determine if the algorithm technology and quality improvement support improves rates of diagnostic testing for patients at high probability of rare or difficult to diagnose disease.
• To evaluate the efficacy of communication with patients regarding the quality improvement intervention.
• To evaluate the rate of acceptance of the quality improvement intervention in the eligible patient population.

The following exploratory objectives will be explored as part of the trial:

• To evaluate if the algorithms identify suitable patients as high probability for rare or difficult to diagnose disease, who are deemed clinically appropriate for further testing by Medical professional judgement .
• To evaluate the outcomes of the diagnostic testing with regard to alternative diagnosis being derived as a direct result of the quality improvement programmes.

Results for all projects (up to 10) will be combined for overall analysis of CAPTURED at the end of the 5-year planned trial period.

The algorithm for each QIP will be run at baseline on the Optimum Patient Care Research Database (OPCRD) and the Optimum Patient Care Research Database (OPCSD), to identify practices with patients who are at high probability of disease. Only practices (and their patients) that fulfil the inclusion criteria and none of the exclusion criteria will be included. Initially, practices with a population of at least 15,000 patients will be included, to maximise likelihood of identifying patients at high probability of disease. This is a flexible cut-off that can be removed depending on numbers of patients identified. Practices that have no current high probability patients for a particular QIP will be excluded from that QIP only but will remain in the CAPTURED overarching trial and eligible for further QIPs.

Participating practices will be offered quality improvement support which will include specific diagnostic testing. Exact details of quality improvement support provided will be determined by each QIP and will all comprise the following core activities:

• Identification (using algorithm)
• Medical review of patients (EHR data) and/or expert review of patients
• Provision of test/screen to patients
• Guidance and support/assistance with patient referrals and follow-on care

These are patient care activities and are NOT research activities.

The specific diagnostic testing process may be remote (e.g. delivered to patients), at practice or at a specified test centre; and will be determined by each QIP.

Once testing has been completed, the patient and GP practice are informed of the result and supported with further steps, as required. Anonymised patient data from participating practices will be collected for trial analysis.

Evaluation of CAPTURED will occur at the end of the 5-year study period.

Practices will be considered as in the control phase in the pre-intervention window. Once randomisation occurs practices move from control to intervention at different points in time. So, all practices contribute some data to both control and intervention periods. Thus, the comparison is those control periods versus those intervention periods.

The proposed study population will be GP practices with patients flagged by the algorithm(s) in the Optimum Patient Care Research Database (OPCRD) and the Optimum Patient Care Service Database (OPCSD). The OPCRD is a REC approved, longitudinal electronic medical record database of over 29 million patients in the UK that is established and managed by Optimum Patient Care Ltd (OPC). OPCSD is pseudonymised patient database used to provide ongoing quality improvement services and programmes to GP practices across the UK.

Inclusion criteria

• General practice in the United Kingdom (England, Scotland, Wales, Northern Ireland) participating in Optimum Patient Care services and contributing deidentified data to OPCRD.
• At least 1 patient identified as high probability of rare or difficult to diagnose disease by one of the quality improvement program algorithms.

Exclusion criteria

• General practices hosting or affected by research, or other aspects of care, which might significantly influence the implementation of CAPTURED.