Population Level Analysis of Acute Stroke Care Patterns in Hungary

Stroke is a global health challenge and it represents a
significant economic and social burden. The treatment of stroke
requires urgent and coordinated procedures. In this clinical
data analysis study, our aim is to identify anomalies in the stroke
care system. We analyze the stroke care patterns using case
records of all 281,948 publicly financed cases between 2010 and
2017 in Hungary. The essence of the method is creating care
events of some basic types, organizing event series into episodes,
classifying episodes with respect to relevant care patterns,
computing ‘spectra’ of episode type frequencies for the
providers and forming clusters of the providers based on the
correlations among their spectra. A similar method is applied
for postal code areas. The results show that two clusters can be
defined that divide the 61 clinics into a smaller and a larger
cluster with significantly different care practices. The spatial
analysis also revealed that the clusters of the postal code areas
form geographically co-located patches marking anomalies in
the stroke care system. The novelty of the paper is the proposed
method and its application to stroke care. The findings may be
used for quality management of the national stroke network.