Reason for hospitalization contrasting adjudication versus ICD-10-CM coding among persons with HIV, 2016-2019

Area of Science:

• Medical Informatics
• Public Health
• Infectious Disease Epidemiology

Background:

• People with HIV (PWH) experience high hospitalization rates, with reasons often unclear.
• Discharge diagnosis codes, such as International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM), may misclassify hospitalization causes in PWH.
• Accurate classification of hospitalization reasons is crucial for understanding healthcare needs and improving outcomes for PWH.

Purpose of the Study:

• To evaluate the accuracy of ICD-10-CM codes in representing the primary reason for hospitalization in people with HIV (PWH).
• To compare ICD-based hospitalization reasons with expert adjudication.
• To identify patient characteristics associated with discrepancies between ICD coding and expert review.

Main Methods:

• A random sample of 302 hospitalizations (21%) from 1428 hospitalizations between 2016-2019 among PWH in the UNC CFAR Clinical Cohort Study (UCHCC) was selected.
• The ICD-based reason for hospitalization was determined using the first or second listed ICD discharge diagnosis.
• ICD-based reasons were compared against expert adjudication, categorizing outcomes as complete agreement, partial agreement, disagreement, or inadequate ICD coding.

Main Results:

• Expert adjudication completely agreed with the ICD-based reason in 73% of hospitalizations, with partial agreement in 14%, disagreement in 8%, and inadequate coding in 6%.
• Patient factors associated with lower agreement included recent CD4 count <200 cells/mm³, nadir CD4 count <50 cells/mm³, ≥15 discharge diagnoses, length of stay ≥4 days, and HIV as a principal diagnosis.
• The study population comprised 204 PWH (69% male), with a median age of 51 years and median CD4 count of 539 cells/mm³ at hospitalization.

Conclusions:

• ICD-10-CM codes effectively capture the reason for hospitalization in the majority of cases for people with HIV (PWH).
• Expert adjudication is recommended for specific patient subgroups, including those with poorly controlled HIV (e.g., low CD4 counts) and complex hospitalizations.
• Refined coding practices and targeted adjudication can improve the accuracy of hospitalization data for PWH, aiding in better healthcare management and research.