OBJECTIVES:

Real-world data is valuable when researching chronic illnesses such as osteoporosis (OP)—a disease with a complex multi-faceted care pathway. State Sequence Analysis (SSA) and Process Mining (PM) have recently been adopted to study healthcare utilization across multiple services. SSA methodology uses a patient’s sequence of interactions within the healthcare system to develop a quantitative measure of care pathways. This quantitative measure can be used to cluster patients based on their pathway (dis)similarities. PM provides transitional state visualizations and statistics on how individuals within each cluster move through the healthcare system. The purpose of the study was to demonstrate a combined SSA-PM approach to develop a patient-level model of clinical pathways in patients with, or at risk of, osteoporotic fractures.

METHODS:

This cohort study used administrative health data to identify healthcare interactions and pathways for individuals in the three years following an OP diagnosis. A stepwise process was used: (1) creating a data set including system-wide interactions related to OP and fragility fractures (FF); (2) calculating dissimilarities between state sequences (care pathways); (3) clustering individuals based on state sequence dissimilarities; and, (4) using PM and healthcare resource use (HCRU) metrics to identify differences in pathways across clusters.

RESULTS:

Using a combination SSA-PM approach, four distinct groups were observed. Those: 1) with high rates of healthcare use (HCU) and few poor health outcomes, 2) with high HCU rates and a high number of poor health outcomes, 3) who died within three years of an OP diagnosis, and 4) with low HCU rates and few poor health outcomes. Differences in demographic characteristics, diagnostic imaging, pharmaceutical interventions, health outcomes, FF events, and overall HCRU associated with OP were found.

CONCLUSIONS:

The SSA-PM approach provides valuable insights into OP care pathways and offers a methodology for studying the complex and heterogeneous care pathways of other diseases.