Title: |
|
Undercounting of Healthcare Utilization in the Medical Expenditure Panel Survey |
MEPS component: |
|
Household Component, Medical Provider Component |
Description: |
|
Household-reported medical events in the Medical Expenditure Panel Survey (MEPS) may be undercounted. This working paper explores whether person characteristics or aspects of the interview such as use of memory aids are correlated with misreporting among respondents. Multivariate analyses are used to isolate the characteristics associated with misreporting while controlling for other covariates. In addition, the paper assesses whether machine learning algorithms can accurately predict whether a household was likely to overreport or underreport events, as well as to predict the number of events reported in the MEPS Medical Provider Component (MPC). The computer-assisted personal interview (CAPI) instrument used for MEPS household data collection was redesigned in 2018 in an effort to improve user experience for MEPS respondents and thus the quality of the collected data. The methods outlined in this paper could be used as a baseline for comparing reporting accuracy before and after the CAPI redesign, to help assess the success of the survey redesign in improving the collection of household-reported events. |
Author(s): |
|
Emily Mitchell, Ph.D.; Jerrod Anderson, Ph.D.; and Adam Biener, Ph.D. |
Agency: |
|
Rockville (MD): Agency for Healthcare Research and Quality |
Back
Email Updates
