Berkowitz Seth A, Fabreau Gabriel E, Raghavan Sridharan, Kentoffio Katherine, Chang Yuchiao, He Wei, Atlas Steven J, Percac-Lima Sanja
Division of General Internal Medicine, Massachusetts General Hospital, 50 Staniford St, 9th Floor, Boston, MA, 02114, USA.
Diabetes Population Health Unit, Massachusetts General Hospital, Boston, MA, USA.
J Community Health. 2016 Dec;41(6):1274-1281. doi: 10.1007/s10900-016-0216-4.
To determine the difference in risk of developing diabetes for refugees, immigrants, and American-born participants living in the same communities, and to explore potential mediators of that difference. Retrospective longitudinal cohort from January 1, 2003 and December 31, 2013. Refugees aged ≥18 years were matched in a 1:3 ratio by age, gender, and date of care initiation to (1) Spanish-speaking non-refugee immigrants, and (2) English-speaking controls receiving care in the same community health center. We used proportional hazards regression to estimate the risk of incident diabetes. We tested whether differences in education or baseline obesity mediated diabetes risk using counterfactual mediation analysis. We included 3174 participants. Among refugee participants, the most common countries of origin were Somalia (17.8 %), Iraq (16.7 %) and Bhutan (8.8 %). Diabetes incidence rate was 1.94, 1.91, and 1.22 cases per 100 person-years follow-up for refugees, immigrants, and controls, respectively. In adjusted models, both refugee (HR 2.08 95 % CI 1.32-3.30) and immigrant (HR 1.51 95 % CI 1.01-2.24) statuses were associated with increased diabetes risk compared with controls. Risk between refugees and immigrants did not differ (adjusted HR for refugees 1.37 95 % CI 0.91-2.06). In mediation analyses, educational attainment mediated 36 % (p = 0.007) of the difference in diabetes risk between refugees/immigrants and controls. Baseline obesity did not mediate difference in diabetes risk (proportion mediated 1 %, p = 0.84). Refugees and immigrants had significantly increased risk for diabetes, partially mediated by education. Education-based lifestyle interventions may be a promising strategy to prevent diabetes for these vulnerable patients.
为了确定生活在同一社区的难民、移民和美国本土参与者患糖尿病风险的差异，并探索造成这种差异的潜在中介因素。回顾性纵向队列研究，时间跨度为2003年1月1日至2013年12月31日。年龄≥18岁的难民按年龄、性别和开始接受治疗的日期以1:3的比例与（1）讲西班牙语的非难民移民，以及（2）在同一社区健康中心接受治疗的说英语的对照者进行匹配。我们使用比例风险回归来估计糖尿病发病风险。我们使用反事实中介分析来检验教育程度或基线肥胖是否介导了糖尿病风险差异。我们纳入了3174名参与者。在难民参与者中，最常见的原籍国是索马里（17.8%）、伊拉克（16.7%）和不丹（8.8%）。难民、移民和对照者每100人年随访的糖尿病发病率分别为1.94、1.91和1.22例。在调整模型中，与对照者相比，难民（风险比2.08，95%置信区间1.32 - 3.30）和移民（风险比1.51，95%置信区间1.01 - 2.24）状态均与糖尿病风险增加相关。难民和移民之间的风险没有差异（难民调整后的风险比1.37，95%置信区间0.91 - 2.06）。在中介分析中，教育程度介导了难民/移民与对照者之间糖尿病风险差异的36%（p = 0.007）。基线肥胖并未介导糖尿病风险差异（介导比例1%，p = 0.84）。难民和移民患糖尿病的风险显著增加，部分由教育介导。基于教育的生活方式干预可能是预防这些弱势群体患糖尿病的一种有前景的策略。
