The transition from the International Classification of Disease (ICD)-9 to ICD-10 coding system has important implications on clinical research as the prevalence of common neurologic diagnoses may be significantly impacted. Evidence suggests a dramatic reduction in prevalence of status epilepticus while the prevalence of idiopathic intracranial hypertension, critical illness myopathy, intracerebral hemorrhage, ischemic stroke, and Parkinson disease increased significantly due to the coding transition, according to study results published in Neurology.
The United States changed its coding system from the ICD-9 to ICD-10 system on October 1, 2015. The objective of the current study was to explore the impact of ICD-9 to ICD-10 coding transition on the prevalence of 16 common neurologic diagnoses by comparing diagnoses before (using the ICD-9 coding system) and after (using the ICD-10 coding system) the ICD coding transition.
Using data from the National Inpatient Sample, study researchers identified 16 common neurologic diagnoses from January 2014 to September 2015 and from October 2015 to December 2017. They then compared the prevalence of each diagnosis before vs after the ICD coding transition.
With the exception of a small increase in the monthly rate of change associated with the coding transition for subarachnoid hemorrhage (from 4.32 to 24.32 diagnoses per month), the monthly rate of change in other neurologic diagnosis frequency remained unchanged over time between ICD-9 and ICD-10 coding systems.
Significant changes in cross-sectional prevalence were evident for 6 of 16 neurologic diagnoses, including a dramatic decrease in diagnoses of status epilepticus (odds ratio [OR], 0.30; 99.7% CI, 0.26-0.34) between ICD-9 and ICD-10. On the other hand, the coding transition was associated with an increase in diagnoses of idiopathic intracranial hypertension (OR, 1.2; 99.7% CI, 1.05-1.37), critical illness myopathy (OR, 1.14; 99.7% CI, 1.01-1.3), intracerebral hemorrhage (OR, 1.08; 99.7% CI, 1.03-1.14), ischemic stroke (OR, 1.04; 99.7% CI, 1.02-1.06), and Parkinson disease (OR, 1.18; 99.7% CI, 1.14-1.22).
The study had several limitations, including data limited to 2 years of complete ICD-10 coding data, potential secular trends unrelated to the coding transition, no available patient-level clinical data, and inability to assess the accuracy and diagnostic validity of the diagnostic codes.
“[O]ur data suggest that changes in administrative coding can systematically bias estimates of neurologic disease prevalence, which have important implications for epidemiology, outcomes research, and clinical trial recruitment efforts that rely on healthcare databases,” concluded the study researchers.
Hamedani AG, Blank L, Thibault DP, Willis AW. Impact of ICD-9 to ICD-10 coding transition on prevalence trends in neurology. Neurology. Published online January 29, 2021. doi:10.1212/CPJ.0000000000001046