Effects of Medicare eligibility at age 65 years on affordability of care and food insecurity.
Effects of Medicare eligibility at age 65 years on affordability of care and food insecurity.
复制标题
65 岁医疗保险资格对护理负担能力和粮食不安全的影响。
DOI：
10.1111/jgs.18560
复制
发表时间：
2023
影响因子：
6.3
通讯作者：
Wadhera,RishiK
中科院分区：
文献类型：
--
作者：
Park,Sungchul;Kinsey,ElizaW;Wadhera,RishiK
方法我们使用了2016年，2017年和2020年医疗支出小组调查的数据。我们的样本包括具有全国代表性的美国成年人。我们的主要（二元）结果是四个衡量护理负担能力的指标和两个粮食不安全水平（低和非常低）。我们遵循先前的研究，并根据美国农业部成人粮食安全调查模块的10项内容衡量粮食不安全状况。6我们还将保险范围作为次要结果。为了评估医疗保险资格的影响，我们使用了回归不连续设计，该设计利用了65岁时医疗保险资格的不连续性。使用数据驱动的方法，自动选择平衡偏差和方差的带宽，我们在65岁左右绘制了带宽6，并将58 - 64岁的个体与66 - 72岁的个体进行了比较。回归不连续性方法依赖于这样的假设，即在截止点（65岁）附近的窄带宽内的个体具有相似的基线观察到的和未观察到的特征。因此，我们比较了65岁以下和65岁以上个体的基线特征。然后，我们进行了一个二次年龄趋势的参数回归不连续性模型，同时调整了个人水平的特征（自我报告的种族和民族，性别，就业状况，婚姻状况，教育，家庭收入，居住的人口普查地区，补充营养援助计划福利的接收，以及年份）。5，9我们还使用替代模型规格、研究样本和年龄带宽选择进行了敏感性分析，并排除了2020年的数据。我们使用调查权重来生成具有全国代表性的估计值。样本包括13，314名年龄在58 - 72岁之间的人。65岁以下和65岁以上的个体之间的个体水平特征没有实质性差异（补充表S1）。65岁时的保险覆盖率显著增加：任何覆盖率增加5.7个百分点（95% CI：4.5 - 6.9），医疗保险覆盖率增加79.6个百分点（95% CI：77.8 - 81.5）（表1）。医疗保险资格导致医疗保健和粮食不安全的财政负担大幅下降（补充图S1）。具体而言，医疗保险资格导致年自付费用超过年收入20%的比例大幅下降32.6%（-1.4个百分点; 95%CI：-2.3，-0.4），支付医疗费用有问题的比例大幅下降13.8%（-2.6个百分点; 95%CI：-4.3，-1.0），无法支付家庭医疗费用的比例大幅下降26.2%（-1.7个百分点; 95%CI：-2.8，-0.6）。医疗保险资格也导致报告低粮食安全的人数大幅下降29.2%（-3.3个百分点; 95% CI：-5.0，-1.6）。然而，我们没有发现资格和非常低的粮食安全之间的联系。我们的敏感性分析得出了类似的结果（补充表S2）。我们的亚组分析表明，医疗保险资格对低食品安全的影响在女性和非西班牙裔黑人成年人中更为明显（表2）。
METHODSWe used data from the Medical Expenditure Panel Survey for 2016, 2017, and 2020. Our sample included nationally-representative US adults. Our primary (binary) outcomes were four measures of affordability of care and two levels of food insecurity (low and very low). We followed prior research and measured food insecurity based on the 10-item United States Department of Agriculture Adult Food Security Survey Module. 6 We also included insurance coverage as secondary outcomes. To evaluate the effect of Medicare eligibility, we used a regression discontinuity design, which exploits a discontinuity in eligibility for Medicare at age 65 years. Using a data-driven method that automatically selected a bandwidth that balanced bias and variance, 8 we drew a bandwidth of 6 around age 65 and compared individuals aged 58–64 to individuals aged 66–72. The regression discontinuity approach relies on the assumption that individuals within a narrow bandwidth around the cutoff (age 65 years) have similar baseline observed and unobserved characteristics. Thus, we compared the baseline characteristics of individuals below and above 65 years of age. Then, we conducted a parametric regression discontinuity model with a quadratic age trend while adjusting for individual-level characteristics (self-reported race and ethnicity, sex, employment status, marital status, education, family income, census region of residence, receipt of supplemental nutrition assistance program benefits, and year). 5, 9 We also ran sensitivity analyses using alternative model specifications, study samples, and age bandwidth choices, and excluding the data for 2020. We used survey weights to generate nationally-representative estimates.RESULTSThe sample included 13,314 individuals aged 58–72. Individual-level characteristics did not substantially differ between individuals below and above age 65 years (Supplementary Table S1). There was a significant increase in insurance coverage at age 65: 5.7 percentage points (95% CI: 4.5–6.9) for any coverage and 79.6 percentage points (95% CI: 77.8–81.5) for Medicare coverage (Table 1). Medicare eligibility led to significant decreases in both the financial burden of medical care and food insecurity (Supplementary Figure S1). Specifically, Medicare eligibility led to significant decreases in having annual out-of-pocket expenses exceeding 20% of annual income by 32.6%(− 1.4 percentage points; 95% CI:− 2.3,− 0.4), having problems paying medical bills by 13.8%(− 2.6 percentage points; 95% CI:− 4.3,− 1.0), and being unable to pay family medical bills by 26.2%(− 1.7 percentage points; 95% CI:− 2.8,− 0.6). Medicare eligibility also led to significant decreases in reporting low food security by 29.2%(− 3.3 percentage points; 95% CI:− 5.0,− 1.6). However, we did not find an association between eligibility and very low food security. Our sensitivity analyses yielded similar results (Supplementary Table S2). Our subgroup analysis showed that the effect of Medicare eligibility on low food security was more pronounced among female and non-Hispanic Black adults (Table 2).