Multimorbidity measures from health administrative data using ICD system codes: A systematic review.

Area of Science:

• Health Services Research
• Epidemiology
• Biostatistics

Background:

• Multimorbidity, the co-occurrence of multiple chronic conditions, is a growing public health concern.
• Accurate measurement of multimorbidity is crucial for understanding population health and informing healthcare policy.
• Health administrative data, utilizing International Classification of Diseases (ICD) codes, are increasingly used to identify diseases.

Purpose of the Study:

• To systematically review and characterize adult population-based multimorbidity measures derived from health administrative data.
• To assess the methodological quality and heterogeneity of existing multimorbidity measures.
• To identify areas for improvement in the development and application of multimorbidity assessment tools.

Main Methods:

• A narrative systematic review was conducted on studies developing or validating multimorbidity measures.
• Key aspects compared included the number of diseases, case definition processes, and validation strategies.
• Methodological robustness was evaluated using eight criteria, including general indicator assessment (AIRE instrument) and multimorbidity-specific criteria.

Main Results:

• Twenty-two distinct multimorbidity measures were identified, incorporating 5 to 84 diseases (median 20).
• Most measures (19/22) included both physical and mental health conditions, primarily using ICD codes from inpatient/outpatient data (18/22).
• Only six measures met at least six of the eight robustness criteria, indicating significant heterogeneity and variable quality.

Conclusions:

• A substantial portion (approximately one-third) of multimorbidity measures derived from administrative data are of low to moderate methodological quality.
• Significant heterogeneity exists in the composition and validation of these measures.
• Developing a more consensual approach to defining the number and types of diseases included in multimorbidity measures is recommended to enhance comparability and reduce confounding in research.