Related Experiment Videos
Applying sequence clustering techniques to explore practice-based ambulatory care pathways in insurance claims data.
Verena Vogt1, Stefan M Scholz2, Leonie Sundmacher3
1Department of Health Care Management, Berlin Centre of Health Economics Research (BerlinHECOR), Technische Universität Berlin, Berlin, Germany.
European Journal of Public Health
|October 18, 2017
Summary
Sequence clustering identified distinct heart failure (HF) patient treatment pathways. Analysis revealed variations in care but no significant impact on HF hospitalization rates.
Area of Science:
- Health Services Research
- Data Science
- Cardiology
Background:
- Care pathways aim to standardize high-quality patient care.
- Empirical data on actual patient treatment sequences are limited.
- Understanding real-world treatment patterns is crucial for optimizing care.
Purpose of the Study:
- To explore sequence clustering for identifying typical ambulatory heart failure (HF) treatment sequences.
- To compare the effectiveness of identified treatment sequences.
Main Methods:
- Utilized routine data from 982 incident HF patients (Germany, 2009-2011).
- Categorized events by physician specialty, service/procedure, and medication.
- Applied longest common subsequence (LCS) for similarity and k-medoids for clustering.
- Used logistic regression to assess sequence effectiveness in avoiding hospitalizations.
Main Results:
- Identified three specialty, four procedure, and four prescription sequence clusters.
- Clusters varied in care timing, intervals, comorbidities, and hospitalization rates.
- No significant association was found between cluster membership and HF hospitalization.
Conclusions:
- Sequence clustering is a valuable explorative tool for analyzing patient treatment sequences.
- This method allows systematic extraction, description, and comparison of care patterns.
- Further research is needed to link identified sequences to clinical outcomes.