Accuracy of external cause of injury codes reported in Washington State hospital discharge records
- 1Washington State Department of Health, Injury Prevention Program, Olympia, Washington
- 2Harborview Injury Prevention and Research Center and the Department of Epidemiology, University of Washington, Seattle, Washington
- 3University of Texas, Southwestern Medical School, Dallas, Texas
- Correspondence to: Mary LeMier, Office of Community Wellness and Prevention, Injury Prevention Program, Mail Stop 7832, Olympia, WA 98504, USA
Objective—To evaluate the accuracy of external cause of injury codes (E codes) reported in computerized hospital discharge records.
Methods—All civilian hospitals in Washington State submit computerized data for each hospital discharge to a file maintained by the Department of Health. In 1996, 32 hospitals accounted for 80% of the injury related discharges in this file; from these hospitals, we sampled 1260 computerized records for injured patients in a stratified, but random, manner. An expert coder then visited the 32 study hospitals, reviewed the medical records that corresponded to each computerized record, and assigned an E code for that hospitalization. The computerized E code information was compared with codes provided by the expert reviewer.
Results—The incidence of hospitalization for injury based upon computerized hospital discharge data was very similar to that based upon chart review: incidence rate ratio 1.0 (95% confidence interval 1.00 to 1.02). Computerized hospital discharge data correctly ranked injuries in regard to both mechanism and intent. Overall agreement on coding was 87% for mechanism of injury, 95% for intent of injury, and 66% for the complete E code. The sensitivity of computerized hospital discharge data for identification of falls, motor vehicle traffic injuries, poisonings, and firearm injuries was 91% or better. The predictive value positive of coding for these four categories of injury ranged from 88% for motor vehicle traffic injuries to 94% for poisonings. The amount of agreement for intent coding ranged from 84% for firearm injuries to 99% for falls. Agreement on coding of the complete E code ranged from 57% for firearm injuries to 72% for poisonings.
Conclusions—Computerized hospital discharge data can be used with confidence to determine how many injuries are treated in a hospital setting and the relative magnitude of various categories of injury. E codes reported in hospital discharge data are a reliable source of information on the mechanism and intent of injury, the two types of information most often used for injury related analyses and priority setting. The detail codes (complete E codes) reported in hospital discharge codes are less reliable and must be used with caution.