The Use of ICD-9-CM Coding to Identify COVID-19 Diagnoses and Determine Risk Factors for 30-Day Death Rate in...

Area of Science:

• Public Health
• Epidemiology
• Health Informatics

Background:

• Accurate quantification of COVID-19 hospital admissions in Italy was challenging due to inconsistent coding in Hospital Information Systems (HIS) across pandemic waves.
• Heterogeneity in International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes complicated reliable data extraction for COVID-19 patient cohorts.

Purpose of the Study:

• To define a specific combination of ICD-9-CM codes for identifying COVID-19 hospitalizations within the Italian HIS.
• To investigate risk factors associated with mortality among COVID-19 patients admitted to Italian hospitals in 2020.
• To analyze temporal changes in mortality rates during the first and second waves across different geographical areas and care settings.

Main Methods:

• Retrospective analysis of hospital discharge records from over 1300 Italian hospitals.
• Development and implementation of an algorithm using specific ICD-9-CM code combinations to identify COVID-19 hospitalizations.
• Multivariable Cox regression models to assess 30-day mortality risk factors and temporal trends.

Main Results:

• Identified 325,810 COVID-19-related hospitalizations; 73.4% were classified as 'due to COVID-19'.
• The 'due to COVID-19' cohort (n=205,048) had a median age of 72, with 60.6% males; overall 30-day mortality was 9.9 per 1000 person-days.
• Lower mortality observed in women and individuals from specific immigrant groups; higher mortality in southern regions, particularly during the second wave and in intensive care units.

Conclusions:

• The developed algorithm represents the first national-level criteria for identifying COVID-19 hospitalizations in HIS.
• The algorithm provides a tool for ongoing pandemic monitoring.
• Future research will assess long-term COVID-19 effects by following up on the 2020 cohort.