Machine readable clinical guidelines and pathways
There has recently been rapid growth in the use of clinical guidelines and pathways for the assessment, diagnosis and management of medical conditions.
Summer of Research
Project by David Bassett, supervised by Dr. Patrick Gladding (University of Auckland)
There has recently been rapid growth in the use of clinical guidelines and pathways for the assessment, diagnosis and management of medical conditions. These guidelines generally contain unstructured free text and are difficult to digitise into electronic health records, limiting the ability to share these pathways.
In the context of rapidly changing clinical guidelines in many disciplines, it is also labour intensive to manually modify existing guidelines and maintain transparent version control practices.
A solution to both of these problems would be a machine readable format for clinical guidelines and pathways, that could auto-populate patient information and then automatically suggest relevant management options based on the information it pulled.
Electronic clinical pathways (ECPs) have been frequently proposed as a way to reduce the time intensity of clinical pathways and improve patient care. However despite a considerable volume of informatics literature on ECPs, there is scant literature regarding the clinical acceptability of them.
This project aims to clarify the potential benefits and challenges of ECPs from a clinical perspective. Semi-structured interviews with seven junior medical staff were undertaken and a thematic analysis completed. Twenty commonly used adult medical and surgical clinical pathways were also assessed for features that could potentially benefit or challenge an electronic format.
Our research found that junior medical staff were largely positive about the introduction of ECPs but saw them primarily as a teaching tool rather than as a rigorous protocol. Guideline discoverability was seen to be an important way to increase productivity and the importance of ECP speed and reliability was noted. The pathway analysis found that medical pathways contained elements most likely to benefit from automation but that all pathways still required clinician input on key variables. The prevalence of teaching points was also noted.
Overall, we found that ECP design should prioritise teaching effectiveness and guidelines discoverability to maximise value to clinicians. Advances in natural language processing may extend the capabilities of ECPs in the future.