Predicting health utilities using health administrative data: leveraging survey-linked health administrative data from Ontario, Canada
Niu Y, Begen N, Zou G, Sarma S. Appl Health Econ Health Policy. 2025; Feb 6 [Epub ahead of print].
Objective — We sought to determine the accuracy of administrative data for identifying mental health service provision in primary care.
Study Design — This was a chart abstraction study measuring agreement between billing data and clinical data on the binary variable "mental health visit." Data were collected from the charts and billing records of 5 academic family practice clinics in Toronto, Ontario (1999 to 2000). Billing claims (n = 952) were selected from the billings for all visits by a stratified random sampling technique. A blinded data abstractor reviewed the clinical charts and assigned diagnostic codes for each patient visit associated with the selected claims. Any visit with at least 1 abstracted mental health diagnostic code was defined as a mental health visit. The test characteristics of 4 administrative measures of mental health service provision, based on different combinations of billing codes, were calculated.
Results — The accuracy of the administrative data was 86.8% when compared with clinical data. The sensitivity of the 4 administrative measures ranged from 22.3% to 80.7%. The specificity ranged from 97.0% to 99.5%.
Conclusions — This is the first study to establish the performance of administrative data in measuring mental health service provision in a primary care setting. In our setting, broadly defined administrative measures of mental health have excellent specificity and adequate sensitivity for exploring and understanding mental health service utilization.
Steele LS, Glazier RH, Lin E, Evans M. Med Care. 2004; 42(10):960-5.
The ICES website uses cookies. If that’s okay with you, keep on browsing, or learn more about our Privacy Policy.