Validation of ICD-9-CM and ICD-10-CM Diagnosis Codes for Identifying Preterm Birth and Low Birthweight Infants in Florida Birth Hospitalizations, 2008-2018.

Objective:To estimate the sensitivity and positive predictive value (PPV) of ICD-9-CM and ICD-10-CM diagnosis codes for identifying preterm birth (PTB) and low birthweight (LBW) infants in birth hospital discharge data, using birth certificate records as the reference standard. Methods:We conducted a population-based validation study of Florida live births from 2008-2018 using linked birth certificate and infant hospital discharge records. Birth certificate obstetric estimates and recorded birthweights served as the reference standards. We calculated sensitivity and PPV for International Classification of Diseases, Ninth Edition, Clinical Modification (ICD-9-CM) and Tenth Edition (ICD-10-CM) diagnosis codes for PTB (<37 weeks) and LBW (<2,500 g) overall, by gestational-age and birthweight strata, and across maternal and hospital characteristics. Results:Among 2,166,535 infants, 10.3% were preterm and 8.7% LBW. PPV exceeded 92% for PTB and 97% for LBW in both coding eras. Sensitivity improved from 77.0% to 84.4% for PTB and from 73.3% to 76.9% for LBW after the ICD-10-CM transition. Sensitivity for diagnosis codes was notably lower nearest the binary classification thresholds for PTB (37 weeks) and LBW (2500 grams). Coding accuracy was higher in urban teaching hospitals compared to nonteaching or rural hospitals. Conclusions:Infant hospital discharge data can reliably identify markedly preterm or low-birthweight infants but may under-ascertain cases near classification thresholds. Researchers should quantify and adjust for these limitations, for example through quantitative bias analysis or sensitivity analyses to account for misclassification.