Descriptive exploration of overdose codes in hospital and emergency department discharge data to inform development of drug overdose morbidity surveillance indicator definitions in ICD-10-CM.

BACKGROUND: In October 2015, discharge data coding in the USA shifted to the International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM), necessitating new indicator definitions for drug overdose morbidity. Amid the drug overdose crisis, characterising discharge records that have ICD-10-CM drug overdose codes can inform the development of standardised drug overdose morbidity indicator definitions for epidemiological surveillance. METHODS: Eight states submitted aggregated data involving hospital and emergency department (ED) discharge records with ICD-10-CM codes starting with T36-T50, for visits occurring from October 2015 to December 2016. Frequencies were calculated for (1) the position within the diagnosis billing fields where the drug overdose code occurred; (2) primary diagnosis code grouped by ICD-10-CM chapter; (3) encounter types; and (4) intents, underdosing and adverse effects. RESULTS: Among all records with a drug overdose code, the primary diagnosis field captured 70.6% of hospitalisations (median=69.5%, range=66.2%-76.8%) and 79.9% of ED visits (median=80.7%; range=69.8%-88.0%) on average across participating states. The most frequent primary diagnosis chapters included injury and mental disorder chapters. Among visits with codes for drug overdose initial encounters, subsequent encounters and sequelae, on average 94.6% of hospitalisation records (median=98.3%; range=68.8%-98.8%) and 95.5% of ED records (median=99.5%; range=79.2%-99.8%), represented initial encounters. Among records with drug overdose of any intent, adverse effect and underdosing codes, adverse effects comprised an average of 74.9% of hospitalisation records (median=76.3%; range=57.6%-81.1%) and 50.8% of ED records (median=48.9%; range=42.3%-66.8%), while unintentional intent comprised an average of 11.1% of hospitalisation records (median=11.0%; range=8.3%-14.5%) and 28.2% of ED records (median=25.6%; range=20.8%-40.7%). CONCLUSION: Results highlight considerations for adapting and standardising drug overdose indicator definitions in ICD-10-CM.

Office-related antibiotic prescribing for Medicaid-enrolled children.

BACKGROUND: Prudent antibiotic prescribing practices are essential to limiting antibiotic resistance. OBJECTIVE: To assess the trend in percentage of office visits for acute respiratory infections (ARIs) linked with an antibiotic prescription. METHODS: Retrospective analysis of Montana Medicaid billing claims data for each year, 1999 to 2010, was done. Participants included continuously enrolled children aged ≤14 years. Primary outcomes were ARI-related office visits and filled antibiotic prescriptions within 10 days of the office visit. RESULTS: Of the 873 244 office visits identified, 116 962 (13%) had an ARI as the primary diagnosis. Among ARI-related office visits, 64 250 (55%) were linked with an antibiotic prescription. From 1999 to 2010, the odds of ARI-related visits being linked with an antibiotic prescription did not change (odds ratio = 1.00; 95% confidence interval = 0.995-1.002). CONCLUSIONS: The percentage of ARI-related visits linked with an antibiotic prescription did not decrease from 1999 to 2010. Further efforts are needed to reduce antibiotic treatment for ARIs.