Identifying longitudinal utilization patterns of medical procedures using national health insurance claims data: a time-series clustering study in South Korea

Health systems need systematic approaches to monitor covered health technologies because their clinical value, utilization, and cost-effectiveness may change over time. However, systematic approaches to identifying technologies for post-coverage review remain underdeveloped. In South Korea, the Second Comprehensive Plan for National Health Insurance has emphasized post-coverage management, highlighting the need for scalable methods to identify technologies for further review. This study aimed to identify heterogeneous longitudinal utilization patterns of covered medical procedures using nationwide administrative claims data and to assess the potential of time-series clustering as a first-stage screening tool for post-coverage management. We analyzed 1,844 classification-level procedure codes across diagnostic tests, diagnostic imaging and radiotherapy, and procedures and surgeries using nationwide claims data from 2008 to 2023. Annual claim volumes were structured as time series and standardized to capture relative changes in utilization trajectories. Procedures with discontinuous claims during the most recent three-year period were pre-classified as limited utilization. Dynamic time warping (DTW)-based hierarchical clustering was then applied to identify distinct utilization patterns. Among 1,844 procedures, 64.6% exhibited increasing patterns, 27.9% exhibited decreasing patterns, and 7.4% showed stable or limited utilization. Across procedure categories, four to five clusters captured distinct longitudinal utilization trajectories. Increasing patterns predominated among diagnostic tests (77.1%), whereas decreasing patterns were more common among procedures and surgeries. These findings suggest that utilization trajectories may reflect technological diffusion, substitution by alternative procedures, changes in clinical practice, or persistent lack of real-world use. Time-series clustering can systematically classify longitudinal utilization patterns of covered medical procedures in a fee-for-service national health insurance system. Increasing, decreasing, and limited-utilization patterns can help identify procedures that warrant further review, with the appropriate management approach determined by the underlying drivers of each utilization trajectory. This approach provides a scalable, claims-based framework for post-coverage surveillance and priority setting in health technology management.

Authors

Institutions

Publication Details

Journal
BMC Health Services Research
Published
2026-11-14
DOI
https://doi.org/10.1186/s12913-026-16060-2
Primary Topic
Healthcare Policy and Management
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Identifying longitudinal utilization patterns of medical procedures using national health insurance claims data: a time-series clustering study in South Korea

Yoon Jung Choi, Jong Youn Moon, Sun Hyun Moon, Su-Jin Cho et al.
BMC Health Services Research
Healthcare Policy and Management
article

Identifying longitudinal utilization patterns of medical procedures using national health insurance claims data: a time-series clustering study in South Korea

Yoon Jung Choi, Jong Youn Moon, Sun Hyun Moon, Su-Jin Cho, Hee Sun Jung
article en

Abstract

Health systems need systematic approaches to monitor covered health technologies because their clinical value, utilization, and cost-effectiveness may change over time. However, systematic approaches to identifying technologies for post-coverage review remain underdeveloped. In South Korea, the Second Comprehensive Plan for National Health Insurance has emphasized post-coverage management, highlighting the need for scalable methods to identify technologies for further review. This study aimed to identify heterogeneous longitudinal utilization patterns of covered medical procedures using nationwide administrative claims data and to assess the potential of time-series clustering as a first-stage screening tool for post-coverage management. We analyzed 1,844 classification-level procedure codes across diagnostic tests, diagnostic imaging and radiotherapy, and procedures and surgeries using nationwide claims data from 2008 to 2023. Annual claim volumes were structured as time series and standardized to capture relative changes in utilization trajectories. Procedures with discontinuous claims during the most recent three-year period were pre-classified as limited utilization. Dynamic time warping (DTW)-based hierarchical clustering was then applied to identify distinct utilization patterns. Among 1,844 procedures, 64.6% exhibited increasing patterns, 27.9% exhibited decreasing patterns, and 7.4% showed stable or limited utilization. Across procedure categories, four to five clusters captured distinct longitudinal utilization trajectories. Increasing patterns predominated among diagnostic tests (77.1%), whereas decreasing patterns were more common among procedures and surgeries. These findings suggest that utilization trajectories may reflect technological diffusion, substitution by alternative procedures, changes in clinical practice, or persistent lack of real-world use. Time-series clustering can systematically classify longitudinal utilization patterns of covered medical procedures in a fee-for-service national health insurance system. Increasing, decreasing, and limited-utilization patterns can help identify procedures that warrant further review, with the appropriate management approach determined by the underlying drivers of each utilization trajectory. This approach provides a scalable, claims-based framework for post-coverage surveillance and priority setting in health technology management.

BMC Health Services Research
Gachon University (KR), Health Insurance Review and Assessment Service (KR)
Good health and well-being
Healthcare Policy and Management
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon