BACKGROUND: The transition to ICD-10-CM for administrative medical claims on October 1, 2015 led the Kansas Department of Health and Environment to explore enhancing their drug poisoning surveillance system. This transition includes a significant increase in the number of diagnosis codes and changes in medical coding instructions used to describe a drug poisoning that will affect public health surveillance. This project was conducted to build an ICD-10-CM drug poisoning surveillance system in Kansas using the Seventh Injury and Surveillance Workgroup (ISW7) Consensus Recommendations for National and State Poisoning Surveillance.
METHODS: The ISW7 report's ICD-10 mortality matrix was used to map ICD-10-CM diagnosis codes based on their drug poisoning groupings. Analysis was completed using ICD-10-CM diagnosis codes from Kansas acute care inpatient hospital discharges occuring on or after October 1, 2015. Drug poisonings due to an adverse effect, under dose, or sequala were excluded, except for data quality assessment or in the case of some potentially illicit drugs, such as cocaine, cannabis derivatives, or unspecified hallucinogens.
RESULTS: During the fourth quarter of 2015, there were 4,123 total mentions of any ICD-10-CM T36-T50 diagnosis codes for 3,430 hospital discharges. However, only 1,572 drug poisoning mentions among 1,154 hospital discharges did not involve an adverse effect or under dose. Drug poisoning diagnosis codes were mostly medical encounters for initial encounters (93%). Subsequent and sequelae were only mentioned in less than 1% of drug poisoning diagnosis codes. Nearly 6% of drug poisoning diagnosis codes were missing a seventh character specifying the type of encounter. Remarkably, there were over 80 different ICD-10-CM drug poisoning groupings identified (e.g., T42.4, T40.2, T39.1). There was also nearly a threefold decrease in unspecific drug poisonings when compared to ICD-9-CM fourth quarter averages. Only 70 (0.4%) ICD-10-CM drug poisoning diagnosis codes included the most unspecific cause of a drug poisoning (e.g., any T50.90, T50.99). There were very few to no cases of adverse effects for potentially illicit drug poisonings caused by cocaine, cannabis derivatives, or unspecified hallucinogens.
CONCLUSIONS: There is evidence that the ICD-10-CM transition is sufficient for surveillance using the ISW7 ICD-10 mortality matrix as a guide. Future work involves obtaining additional ICD-10-CM data for 2016 and signal-to-noise validation of drug poisoning cases to enhance the Kansas-National Syndromic Surveillance Program, a near real-time syndromic surveillance system for emergency department admission. Presented results are limited to only fourth quarter 2015 data in Kansas and continued monitoring will occur as new data is accessed.