Validity of Nonspecific Stroke Location ICD-10 Subcodes: A Descriptive Study.

Area of Science:

• Neurology
• Medical Informatics
• Health Services Research

Background:

• Investigating imaging characteristics of stroke hospitalizations with nonspecific/unspecified vascular region subcodes.
• Assessing for systematic bias in the utilization of these subcodes within stroke coding.

Purpose of the Study:

• To describe imaging findings in stroke hospitalizations coded with nonspecific/unspecified vascular region subcodes.
• To evaluate potential systematic bias in the application of these ICD-10 subcodes.

Main Methods:

• Analysis of 5,234 first ischemic stroke hospitalizations from 2018-2022.
• Imaging review of 200 randomly selected hospitalizations (100 nonspecific, 100 specific subcodes).
• Multilevel logistic regression to assess bias in subcode use, considering primary provider.

Main Results:

• 43% of ischemic stroke hospitalizations received nonspecific/unspecified subcodes.
• Of reviewed cases, 85% had imaging-confirmed infarcts in specific locations (45% anterior, 40% posterior circulation).
• Lower NIHSS scores and non-neurological specialist involvement were associated with nonspecific subcode use.

Conclusions:

• Most nonspecific stroke codes correspond to specific infarct locations, particularly posterior circulation.
• Coding bias is suggested by associations with lower NIHSS and specialist type, though variance explained is modest.
• Researchers should acknowledge limitations and use sensitivity analyses when utilizing these subcodes due to potential bias.