Estimation of Oral Disease Burden from Claims and Self-Reported Data

Abstract

Objective: To examine the use of Medicaid, commercial claims, and self-reported survey data to estimate the prevalence of oral disease burden. Methods: We analyzed 2018 Medicaid claims from IBM Watson Medicaid Marketscan database, commercial claims from the IBM Dental Database, and Medical Expenditure Panel Survey (MEPS) data. The estimate of oral disease burden was based on standard metrics using periodontal and caries-related treatment procedure codes. Examples are restorations: D2000 to D2999, root Canals: D3230 to D3334, periodontics: D4000 to D4999, prosthodontics: D5000 to D6999 and extractions: D7000 to D7251. A direct comparison between the data sets was also done. Enrollees from the different databases were broken down by gender, race/ethnicity, and into age groups. Results: Medicaid and commercial enrollees were 11.6 million and 10.5 million. The weighted proportion from MEPS for Medicaid and commercial plans ranged from 80-208 million people. Prevalence of caries-related treatments was estimated for IBM Watson and MEPS for total enrollees for Medicaid (13% vs. 12%); and commercial claims (25% vs. 17%), respectively. Prevalence of periodontal related treatments was estimated for IBM Watson and MEPS total enrollees for Medicaid (0.7% vs. 0.5%) and commercial claims (7% vs. 1.6%), respectively. Prevalence of dental diseases was higher in patients with at least one visit for Medicaid, commercial plans, and MEPS. Prevalence based on specific procedures were higher in commercial plans than in Medicaid. Conclusions: Claims data has the potential to serve as a proxy measure for the estimate of dental disease burden in a population. In addition, in rare events, claims data provides a better estimate of disease burden because it is based on a larger dataset.