Implementation of the TRANSFoRm evidence service supporting diagnosis in primary care
The problem
The nature of diagnosis in primary care demands knowledge of a wide range of clinical conditions. Formulation of a working diagnostic hypothesis requires consideration of many differential diagnoses that may present in the form of a particular patient complaint. There follows a process of refinement of the differentials to consider, ruling in or out each candidate differential based on the confirmed presence or absence of diagnostic cues elicited during patient consultation. This process can fail and diagnostic error has been shown to be a major threat to patient safety in the primary care setting. The knowledge base available to any clinician may be limited by their own case experience, making diagnosis more difficult in the case of rare or unfamiliar conditions. The patient safety implications of diagnostic error are potentially severe for both patient and clinician.The approachWe describe the design of an ICT intervention implementing a clinical evidence service for the TRANSFoRm project, an EU funded FP7 collaborative project. The knowledgebase implements a general model of clinical evidence to represent diagnostic evidence for any chosen diagnosis. The model contains diagnostic cues representing clinician observed signs, patient reported symptoms, risk factors and clinical tests. The models are made available through a web based clinical evidence service. This allows querying and refinement of candidate differentials to consider based on coded diagnostic evidence cues extracted from an electronic patient record. An iterative process takes place submitting the patient presenting problem along with consultation diagnostic cues. This allows third party decision support tools to provide coded evidence-based recommendations are integrated with an electronic health record (EHR), supporting the process of diagnostic hypothesis formulation in primary care.
Findings
A separately developed diagnostic decision support interface has been integrated with the Vision 3 EHR and driven by the web based evidence service populated to support diagnosis of conditions presenting as chest pain, abdominal pain or dyspnoea. The tool is designed to support both bottom-up and top-down diagnostic reasoning based on a presenting patient complaint. Suggested differentials to consider are presented and clinicians may ask questions relating to diagnostic evidence for each differential presented through the tool. This is currently undergoing evaluation with GPs and actors-as-patients in a simulated surgery.
Consequences
The growing use and maturity of electronic patient records in primary care provides an opportunity to apply patient data in new ways to support more advanced functionality beyond tracking patient history. Aggregating and mining this data provides opportunities to generate usable clinical evidence that can be applied in practice. The development of computable models of clinical evidence that are interoperable and integrated with different systems are an important prerequisite for unlocking that potential and improving patient safety.
Credits
- Derek Corrigan, Trinity College, Dublin, Ireland
- Gary Munnelly, Kings College, London, UK
- Talya Porat
- Samhar Mahmoud
- Olga Kostopoulou
- Brendan Delaney