ISS mapped from ICD-9-CM by a novel freeware versus traditional coding: a comparative study.

Area of Science:

• Trauma research
• Injury severity scoring systems
• Health informatics

Background:

• Injury severity is measured using Abbreviated Injury Scale (AIS) or International Classification of Diseases (ICD).
• ICD codes are convenient for administrative data collection.
• ICDPIC trauma software maps ICD-9-CM to AIS codes, offering a potential advantage over manual AIS coding.

Purpose of the Study:

• To compare the Injury Severity Score (ISS) calculated by ICDPIC trauma software (ICDPIC-ISS) with ISS derived from direct, prospective AIS coding (dAIS-ISS).
• To assess the agreement and discriminative power for survival between ICDPIC-ISS and dAIS-ISS.

Main Methods:

• Retrieved administrative records of 289 major trauma cases from Udine, Italy.
• Calculated ICDPIC-ISS and dAIS-ISS.
• Assessed agreement using Cohen's Kappa and Bland-Altman analysis.
• Analyzed differences based on the ratio of ICD-9-CM to AIS codes (DIARATIO).
• Compared survival discriminative power using ROC curves.

Main Results:

• Poor agreement between ICDPIC-ISS and dAIS-ISS (12.1% exact match, k=0.07; 22.4% categorized match, k=0.09).
• Significant differences observed, increasing with DIARATIO, indicating issues with administrative coding completeness.
• ICDPIC-ISS showed lower survival discriminative power (AUC 0.63 vs. 0.76, p=0.02).

Conclusions:

• ICDPIC-ISS demonstrated poor agreement and lower survival discriminative power compared to dAIS-ISS.
• Incomplete ICD-9-CM coding significantly impacted ICDPIC-ISS performance.
• Further research is needed on trauma scores based on ICD data, especially in European countries with similar coding practices.