A VALIDATION OF HYSTERECTOMY PROCEDURAL CODES IN THE CANADIAN INSTITUTES FOR HEALTH INFORMATION DISCHARGE ABSTRACT DATABASE

Objectives The Canadian Institute of Health Information (CIHI) Discharge Abstract Database (DAD) is the main source of routinely-collected data for gynecologic surgery in Canada and are increasingly used for research. As these data were originally collected for healthcare administrative purposes, like billing or surveillance, they are prone to error and should be validated prior to their use for clinical research. The objective of this study was to validate hysterectomy codes from the CIHI-DAD at a single institution. Methods We obtained a consecutive sample of all gynecologic procedures performed from April 2016 to March 2017 using the CIHI-DAD at The Ottawa Hospital. Patient data, including diagnosis, procedure type, and surgical approach, were reabstracted from charts. Reabstracted chart data were then compared to CIHI-DAD Canadian Classification of Health Interventions (CCI) codes using sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and kappa coefficient with associated 95% confidence intervals. Results Of 1079 gynecologic procedures, 649 hysterectomies were performed, including 23.3% vaginally, 17.1% laparoscopically, and 14.9% abdominally. The median patient age was 46years (interquartile range:41-54). The sensitivity, specificity, PPV, NPV, and kappa, with associated 95% confidence intervals, for all hysterectomies was94.8%(92.8-96.4%), 88.4%(85.0-91.3%), 92.5%(90.2-94.4%), 91.8%(88.7-94.3%) and 0.84(0.80-0.87%), respectively. For vaginal hysterectomy, sensitivity=88.8%(84.3-92.5%), specificity=99.2%(98.3-99.7%), PPV=97.0%(93.8-98.9%), NPV=96.7%(95.3-97.8%), and kappa=0.91(0.87-0.94). For laparoscopic hysterectomy, sensitivity=91.3%(86.3-95.0%), specificity=92.8%(91.0-94.4%), PPV=72.4%(66.2-78.1%), NPV=98.1%(97.0-98.9%), and kappa=0.79(0.71-0.81). For abdominal hysterectomy, sensitivity= 96.9%(93.9-99.0%), specificity=94.9%(93.2-96.2%), PPV=76.8%(70.4-82.5%), NPV=99.4%(98.7-99.8%), and kappa=0.83 (0.78-0.87). Conclusions Our study suggests a high level of validity for hysterectomy CCI codes in the CIHI-DAD for clinical research purposes.