Institute of Health Policy and Management, Erasmus University Rotterdam, Burgemeester Oudlaan 50, 3062 PA, Rotterdam, The Netherlands.
Eur J Health Econ. 2018 Jul;19(6):775-784. doi: 10.1007/s10198-017-0921-7. Epub 2017 Aug 3.
Empirical evidence suggests that people are fairly sensitive to cost sharing arrangements in ambulatory mental healthcare. However, pure cost sharing effects are typically hard to measure due to the presence of adverse selection effects. In this paper, we examine the impact of cost sharing on mental healthcare utilization in the context of mandatory health insurance where adverse selection is absent. Using a large proprietary dataset of a Dutch private health insurer, we examine to what extent a new copayment scheme for adult mental healthcare changed healthcare utilization. We exploit the fact that non-adults are exempted from copayments. First, we compare changes in utilization among adults and non-adults using t tests and a difference-in-difference analysis. Second, we highlight differential changes in mental healthcare utilization by treatment (duration and type of mental illness) and individual characteristics (gender and socioeconomic status). Third, we evaluate to what extent anticipatory behavior occurred pending the introduction and subsequent repeal of the new copayment scheme. Our results show a strong and significant (p < 0.01) decrease in outpatient secondary mental healthcare utilization among adults following the introduction of copayments, which is absent among non-adults. This decrease is concentrated among treatments for less severe mental illnesses. Furthermore, the utilization patterns suggest the presence of anticipatory behavior.
实证证据表明,人们对门诊精神保健中的成本分担安排相当敏感。然而,由于存在逆向选择效应,通常很难衡量纯粹的成本分担效应。在本文中,我们在不存在逆向选择的情况下,考察了强制医疗保险背景下成本分担对精神保健利用的影响。我们利用荷兰私人健康保险公司的大型专有数据集,研究了新的成人精神保健共付额计划在多大程度上改变了医疗保健的利用。我们利用未成年人豁免共付额的事实。首先,我们使用 t 检验和差异分析比较了成年人和未成年人之间利用情况的变化。其次,我们突出了按治疗(持续时间和精神疾病类型)和个体特征(性别和社会经济地位)划分的精神保健利用的差异变化。第三,我们评估了在引入和随后废除新共付额计划之前,预期行为发生的程度。我们的研究结果表明,在引入共付额后,成年人的门诊二级精神保健利用率显著下降(p<0.01),而未成年人则没有。这种下降主要集中在治疗较轻的精神疾病上。此外,利用模式表明存在预期行为。