Understanding caseload and practice through analysis of therapeutic state transitions

Increasing uptake of electronic medical records (EMRs) offers opportunities for analysis and quality assurance of clinical practice; but making appropriate inferences about chronic disease management is non-trivial. We propose a state-transition model for analysis of chronic disease therapy by community practices based on clinical practice guidelines. We abstract guideline recommendations to key therapeutic states, and map electronic prescribing data into a series of states and transitions, which can then be interpreted with respect to the recommendations of the guideline. By applying our approach to a patient cohort diagnosed with both hypertension and diabetes mellitus, we demonstrate that the method identifies cases at risk of suboptimal care, deserving further review by General Practitioners (GPs).