Abstract
Objective
To examine the effect of an employer‐mandated switch to high‐deductible health plans (HDHP) on emergency department (ED) low‐value imaging.
Data Sources
Claims data of a large national insurer between 2003 and 2014.
Study Design
Difference‐in‐differences analysis with matched control groups.
Data Collection/Extraction Methods
The primary outcome is low‐value imaging during ED visits for syncope, headache, or low back pain. We included members aged 19‐63 years whose employers offered only low‐deductible (≤$500) plans for one (baseline) year and, in the next (follow‐up) year, offered only HDHPs (≥$1000). Contemporaneous members whose employers offered only low‐deductible plans for two consecutive years served as controls. The groups were matched by person and employer propensity for HDHP switch, employer size, baseline calendar year, and baseline year quarterly number of total and imaged ED visits for each condition. We modeled the visit‐level probability of low‐value imaging using multivariable logistic regression with member‐clustered standard errors. We also calculated population level monthly cumulative ED visit rates and modeled their trends using generalized linear regression adjusting for serial autocorrelation.
Principal Findings
After matching, we included 524 998 members in the HDHP group and 5 448 803 in the control group with a mean age of approximately 42 years and 48% female in both groups. On visit‐level analyses, there were no significant differential changes in the probability of low‐value imaging use in the HDHP and control groups. In population‐level analyses, compared with control group members, members who switched to HDHPs experienced a relative decrease of 5.9% (95% CI − 10.3, −1.6) in ED visits for the study conditions and a relative decrease of 5.1% (95%CI −9.6, −0.6) in the subset of ED visits with low‐value imaging.
Conclusion
Though HDHP switches decreased ED utilization, they had no significant effect on low‐value imaging use after patients have decided to seek ED care.