Age and Choice in Health Insurance Evidence from a Discrete Choice Experiment

Background: A uniform package of benefits and uniform cost sharing are elements of regulation inherent in most social health insurance systems. Both elements risk burdening the population with a welfare loss if preferences for risk and insurance attributes differ. This suggests the introduction of more choice in social health insurance packages may be advantageous; however, it is widely believed that this would not benefit the elderly. Objective: To examine the relationship between age and willingness to pay (WTP) for additional options in Swiss social health insurance. Methods: A discrete choice experiment was developed using six attributes (deductibles, co-payment, access to alternative medicines, medication choice, access to innovation, and monthly premium) that are currently in debate within the context of Swiss health insurance. These attributes have been shown to be important in the choice of insurance contract. Using statistical design optimization procedures, the number of choice sets was reduced to 27 and randomly split into three groups. One choice was included twice to test for consistency. Two random effects probit models were developed: a simple model where marginal utilities and WTP values were not allowed to vary according to socioeconomic characteristics, and a more complex model where the values were permitted to depend on socioeconomic variables. A representative telephone survey of 1000 people aged >24 years living in the German-and French-speaking parts of Switzerland was conducted. Participants were asked to compare the status quo (i.e. their current insurance contract) with ten hypothetical alternatives. In addition, participants were asked questions concerning utilization of healthcare services; overall satisfaction with the healthcare system, insurer and insurance policy; and a general preference for new elements in the insurance package. Socioeconomic variables surveyed were age, sex, total household income, education (seven categories ranging from primary school to university degree), place of residence, occupation, and marital status. Results: All chosen elements proved relevant for choice in the simple model. Accounting for socioeconomic characteristics in the comprehensive model reveals preference heterogeneity for contract attributes, but also for the propensity to consider deviating from the status quo and choosing an alternative health insurance contract. Conclusion: The findings suggest that while the elderly do exhibit a stronger status quo bias than younger age groups, they require less rather than more specific compensation for selected cutbacks, indicating a potential for contracts that induce self-rationing in return for lower premiums.