The Long Term Effects of Consumer-Directed Health Plans on Health Care Costs

“Consumer-Directed” Health Plans (CDHPs), those with high deductibles and tax-advantaged personal medical accounts, are a market-based strategy that focuses on the demand-side of health care to reduce costs. They are increasingly popular. CDHPs are offered by more than half of large employers and now make up 18% of all employer-based insurance. Health care reform will likely reinforce the trend towards CDHPs as the “Cadillac tax” will penalize employers that offer generous health plan products. 

Most existing CDHP-costs research focuses on the first-year effects of CDHPs enrollment and the preponderance of evidence is that CDHPs reduce health care spending. However, there is debate on whether this is just a short-term effect or whether CDHPs have longer term effects and can “bend the cost curve”.  Only two studies have compared longer term effects of CDHPs. Both studies were limited to a single employer and they reached conflicting conclusions. Our study is the first, to our knowledge, that examines the longer-term impact of CDHPs on costs in a multi-firm, multi-plan-type setting. 

We analyzed a unique dataset of 5 years of enrollment and claims data for 60 large U.S. firms (about 17 million, person-years of data). Half of the firms began offering a CDHP during the study period. These data are supplemented by detailed insurance benefit design information and geocoded socio-demographic information. The ‘treatment’ firms had varying rates of initial CDHP take-up and there was heterogeneity in the level of the CDHP deductibles (ranges from $1,000 to $5,000 for individual contracts) and whether the CDHP was linked to a Health Reimbursement Arrangement or Health Savings Account (HRA or HSA). The aim of our analysis is to identify the effect of CDHP offer on health care cost trends using an intent-to-treat design. We compare the treatment (CDHP offering) firms to control (offering only traditional plans) firms using a differences-in-differences approach to control for time-invariant differences in firms. To control for differences in treatment and control firm composition over time, we use a rich set of control variables to balance observables within each firm over time. In addition, we implement treatment on the treated propensity score weighting to align control firm observables with treatment firm observables including pre-year levels and cost trends prior to CDHP offer.

Our main outcome is overall spending. Secondary outcomes include costs by service type (inpatient, outpatient, emergency room, and pharmaceuticals). In sub-analyses we examine differential effects by deductible amount and HSAs versus HRAs.

The results of our study will help inform the ongoing debate on the longer-term effects of CDHPs.