Variation in stroke care at the hospital level: A cross-sectional multicenter study

Abstract

Introduction: Stroke is one of the leading causes of mortality and disability. Improving patient outcomes can be achieved by improving stroke care and adherence to guidelines. Since wide variation in adherence rates for stroke guidelines still exists, we aimed to describe and compare stroke care variability within Belgian hospitals.

Materials and methods: An observational, multicenter study was performed in 29 Belgian hospitals. We retrospectively collected patient characteristics, quality indicators, and time metrics from the last 30 consecutive patients per hospital, diagnosed with ischemic stroke in 2019 with structured questionnaires. Mean adherence ratios (%) ± SD (minimum - maximum) were calculated.

Results: We analyzed 870 patient records from 29 hospitals. Results showed large inter- and intrahospitals variations in adherence for various indicators. Almost all the patients received brain imaging (99.7%) followed by admission at a stroke unit in 82.9% of patients. Of patients not receiving thrombolysis, 92.5% of patients were started on antithrombotic drugs. Indicators with moderate median adherence but large interhospital variability were glycemia monitoring [82.3 ± 16.7% (26.7-100.0%)], performing clinical neurological examination and documentation of stroke severity [63.1 ± 36.8% (0-100%)], and screening for activities of daily living [51.1 ± 40.3% (0.0-100.0%)]. Other indicators lacked adequate adherence: swallowing function screening [37.0 ± 30.4% (0.0-93.3%)], depression screening [20.2 ± 35.8% (0.0-100%)], and timely body temperature measurement [15.1 ± 17.0% (0.0-60%)].

Conclusion: We identified high adherence to guidelines for some indicators, but lower rates with large interhospital variability for other recommendations also based on robust evidence. Improvement strategies should be implemented to improve the latter.

Keywords: adherence; care process; guidelines; stroke; variation.

Figure 1

The hospitals' randomization numbers are plotted in columns and QI is depicted in rows of the figure. From top to bottom, the figure is sorted according to total mean adherence ratio for each indicator, which is shown in the right vertical axis. From left to right, the figure is sorted by total mean QI adherence rate for each hospital, which is depicted at the bottom horizontal axis of the figure. The darker red the color of the cells, the poorer the adherence ratio. If no information was found in the record, this was analyzed in a similar way as “not performed”. For all the QI, patient records of all the 30 patients were included in the denominator (n = 30), except for antithrombotic administration. For this indicator, only the patients, for whom antithrombotic administration was the primary therapy, were included for the calculation of this QI. FeSS, Fever, Sugar, Swallowing protocol; ADL, activities of daily living; QI, quality indicator.

Figure 2

(A) Time from ED arrival to glycemia measurement, to imaging, and door-to-needle time, if applicable, for early presenters. Recommendation for door-to-needle time, according to guidelines. Recommendation for time to brain imaging, according to guidelines Hospitals 28 and 29 are excluded from this figure as they had no data available for door-to-needle time. Numbers in the figure indicate the number of patients who were treated with thrombolysis. N = 326. (B) Time to SU admission and time to antithrombotic treatment for late presenters. ED, emergency department; SU, stroke unit Hospitals 3,7, 27 and 28 are excluded from this figure as they had no data available for time ED to aspirin and time to SU. ^*No SU available No data available on time from ED to aspirin administration N = 216.

Figure 3

The hospitals' randomization numbers are plotted in the y-axis, sorted from long LOS at the top of the figure, to short LOS at the bottom of the figure. The median days from admission are plotted in the x-axis.