A novel method for detecting inpatient pediatric asthma encounters using administrative data

Area of Science:

• Health Services Research
• Epidemiology
• Medical Informatics

Background:

• Accurate detection of asthma encounters is vital for public health surveillance, quality reporting, and clinical research.
• Current methods may fail to capture the full scope of asthma, impacting the evaluation of interventions and healthcare.
• Asthma is often associated with other respiratory conditions, complicating detection based on principal diagnosis alone.

Purpose of the Study:

• To develop and validate an algorithm for detecting asthma encounters using principal discharge diagnosis and asthma diagnosis code position.
• To improve the sensitivity and specificity of asthma encounter identification for public health surveillance and research.

Main Methods:

• A medical record review (n=191) served as the gold standard for identifying asthma encounters.
• An algorithm was developed and tested, considering asthma diagnosis code position in relation to principal respiratory diagnoses.
• Statistical analysis, including odds ratios and covariate adjustment, was used to evaluate the algorithm's performance.

Main Results:

• Asthma coded in the second or third position showed no significant difference in odds compared to the principal diagnosis for certain respiratory conditions.
• Coding asthma in the fourth or fifth position significantly differed from the principal diagnosis (P<.001).
• The proposed algorithm, including asthma in the first three code positions, increased sensitivity to 0.84 and reporting accuracy to 0.83, outperforming principal diagnosis alone (0.55) and all asthma codes (0.66).

Conclusions:

• An algorithm incorporating specific respiratory principal diagnoses and asthma diagnosis code position reliably enhances asthma encounter detection.
• This improved detection is critical for accurate population-based health impact measurement and comparative effectiveness research.
• The algorithm significantly increased the reporting accuracy and identified a 64% increase in asthma-related bed days compared to using the principal diagnosis alone.