Deep Phenotyping and Prediction of Long-term Cardiovascular Disease: Optimized by Machine Learning

BackgroundPrediction of cardiovascular disease (CVD) is important in clinical practice. Machine learning (ML) may offer an improved alternative to current CVD risk stratification in individual patients. We aim to identify important predictors and compare ML models with traditional models according to their prediction performance in a large long-term follow-up cohort.MethodsThe Atherosclerosis Risk in Communities (ARIC) study was designed to study the progression of subclinical disease to cardiovascular events over a 25-year follow-up period. All phenotypic variables at visit 1 were obtained. All-cause death, CVD, and coronary heart disease were the outcomes for analysis. The ML framework involved variable selection using the random survival forest (RSF) method, model building, and 5-fold cross-validation. Model performance was evaluated by discrimination using the Harrell concordance index (C-index), accuracy using the Brier score (BS), and interpretability using the number of variables in the model.ResultsOf the 14,842 participants in ARIC, the average age was 54.2 years, with 45.2% male and 26.2% Black participants. Thirty-eight unique variables were selected in the RSF top 20 importance ranking of all 6 outcomes. Aging, hypertension, glucose metabolism, renal function, coagulation, adiposity, and sodium retention dominated the predictions of all outcomes. The ML models outperformed the regression models and established risk scores with a higher C-index, lower BS, and varied interpretability.ConclusionsThe ML framework is useful for identifying important predictors of CVD and for developing models with robust performance compared with existing risk models.     

Suprachoroidal Injection of Triamcinolone Acetonide Injectable Suspension for the Treatment of Macular Edema Associated With Uveitis in the United States: A Cost-Effectiveness Analysis

ObjectivesSuprachoroidal injection of triamcinolone acetonide is the first Food and Drug Administration–approved treatment for macular edema associated with uveitis. A cost-effectiveness analysis was performed comparing this treatment with best supportive care (BSC) for the management of this indication from US Medicare and commercial payer perspectives.MethodsA patient-level simulation was developed per the patient characteristics and changes in best-corrected visual acuity letter scores observed in a phase III study of triamcinolone acetonide (PEACHTREE). The wholesale acquisition cost of triamcinolone acetonide was $1650/injection; suprachoroidal injection cost was assumed at $200/injection. Healthcare costs were informed by a US claims–based analysis. Mortality risk associated with severe vision loss and blindness was modeled by applying a hazard ratio to all-cause mortality rates of the US general population. Health-related quality of life weights, obtained from a regression model fitted to the Visual Function Questionnaire-25 data from PEACHTREE, were applied based on the best-corrected visual acuity scores of both eyes. Costs (2020 US dollar) and benefits were discounted at 3% annually. Incremental cost-effectiveness ratios were estimated over a 10-year horizon.ResultsIn the base-case, the incremental cost-effectiveness ratio comparing triamcinolone acetonide with BSC was $28 479 per quality-adjusted life-year gained. The wholesale acquisition cost for triamcinolone acetonide for suprachoroidal use was ～68%, ～56%, and ～27% below the willingness-to-pay thresholds of $150 000, $100 000, and $50 000 per quality-adjusted life-year gained, respectively. Results were robust in sensitivity and scenario analyses.ConclusionsTriamcinolone acetonide for suprachoroidal use is cost-effective compared with BSC for patients with macular edema associated with uveitis.     

基于HS-GC-IMS技术分析六神曲炮制前后（炒、焦）挥发性物质的变化

目的 分析六神曲Massa Medicata Fermentata炮制前后（炒、焦）的挥发性有机物特征及其变化规律。方法 采用顶空-气相色谱-离子迁移质谱（headspace-gas chromatography-ion migration mass spectrometry,HS-GC-IMS）测定六神曲生品、炒品和焦品的挥发性有机物,构建HS-GC-IMS指纹图谱,采用VOCal软件对检测到的成分进行定性和定量分析,运用主成分分析（principal component analysis,PCA）、偏最小二乘-判别分析（partial least squares-discriminant analysis,PLS-DA）、t检验等对样品进行差异性分析。结果 基于HS-GC-IMS技术从六神曲生品、炒品和焦品中共获得了80种挥发性有机物,定性鉴别出60种;构建了六神曲生品、炒品和焦品的HS-GC-IMS指纹图谱,通过PCA、PLS-DA、t检验等分析方法可将三者进行区分,己醛-D(hexanal-D)、乙醇（ethanol）2种物质可作为六神曲生品的特征性成分;丁醛（butanal）...     

浙江省海岛农村社区成年人代谢综合征发病率及其危险因素

目的估计浙江省海岛农村社区成年人代谢综合征发病情况, 探索该地区代谢综合征发病的相关影响因素。方法 2018年6-12月在浙江省玉环市对2012年开展的代谢综合征基线调查人群中无代谢综合征调查对象进行随访调查, 获取相关调查表信息以及实验室数据, 描述代谢综合征发病情况, 使用logistic回归探索发病危险因素及调整风险比(aRR)和95%CI。结果 3 162名随访对象中新发代谢综合征522例, 6年累积发病率为16.5%, 男女性累积发病率分别为12.3%、20.6%, 女性高于男性(P<0.001)。无业、吸烟、饮酒调查对象中代谢综合征发病率较高。女性(aRR=1.96, 95%CI:1.50～2.58)和高血压家族史(aRR=1.31, 95%CI:1.04～1.63)为代谢综合征发病独立危险因素。结论海岛农村社区成年人代谢综合征发病率相对较高, 其中女性以及有高血压家族史者发病风险更高。     

我国农村老年人抑郁倾向与认知功能变化轨迹关系研究

目的 探索我国农村老年人抑郁倾向与认知功能变化轨迹的关系。方法 基于中国老年健康影响因素跟踪调查（CLHLS）中2011-2018年农村≥65岁老年人的数据,分别构建无条件和条件潜变量增长曲线模型,分析我国农村老年人简易精神状态评价量表（MMSE）得分变化轨迹及与抑郁倾向的关系。结果 最终纳入了1 788名2011/2012年开始调查的农村老年人,其中有抑郁倾向者占19.1%;有、无抑郁倾向者基线MMSE得分别为26.62±3.54和27.59±3.17。无条件和条件潜变量增长曲线模型结果显示：MMSE得分下降率为0.52分（P<0.05）;纳入协变量前后,有抑郁倾向者MMSE初始得分比无抑郁倾向者分别低0.84分（P<0.05）和0.81分（P<0.05）,有抑郁倾向者MMSE得分下降速度比无抑郁倾向者分别慢0.33分（P<0.05）和0.40分（P<0.05）。结论 我国农村老年人抑郁倾向与认知功能变化轨迹有关,有抑郁倾向者处于更低水平轨迹,提示应积极关注我国农村老年人的心理健康并及时干预,以延缓或防止老年痴呆的发生发展。     

中国中老年人共病发展轨迹及其对新发失能的影响

目的 探索我国中老年人共病发展轨迹及其对新发失能的影响,以识别具有相似共病发展历程的同质群体,为我国中老年人群的失能风险干预提供依据。方法 基于中国健康与养老追踪调查2011-2018年4期数据,采用组基轨迹模型拟合共病发展轨迹,进而采用含时间依存协变量Cox回归模型分析其对新发失能的影响。结果 共纳入8 580名中老年人,依据中老年共病发展轨迹趋势,可分为无共病型（2 136人,24.90%）、新发-发展型（3 758人,43.80%）、适度-发展型（2 270人,26.45%）和严重-发展型（416人,4.85%）4种类型。适度-发展型和严重-发展型人群中女性、自评健康差、超重/肥胖、单身、居住在农村、家庭年人均支出水平高、发生新发失能的占比较高。调整人口学、行为学协变量后,与新发-发展型相比,严重-发展型（HR=3.132,95%CI：1.884~5.207）的失能风险最高,适度-发展型（HR=1.400,95%CI：1.026~1.909）次之,无共病型（HR=0.631,95%CI：0.424~0.938）最低。结论 我国中老年人的共病发展轨迹存在异质性,共病发展轨迹的失能风险随共病水平的升高而增加,提示共病发展轨迹升高是发生失能的危险因素。     

匹配在观察性研究中的作用——有向无环图视角

匹配是观察性研究中选择研究对象的一种常用方法,具有控制混杂因素、提高统计效率等作用,但其控制混杂因素的作用在不同观察性研究中并不一致,匹配在队列研究中能够消除匹配变量的混杂偏倚,但在病例对照研究中匹配本身并不能消除混杂偏倚。在匹配性病例对照研究选择匹配变量时,研究者可能并不能准确判断该变量是否为混杂变量,若误将真实情况为非混杂因素的变量进行匹配,则会形成过度匹配,造成统计效率下降或引入难以避免的偏倚或增加工作量等;若将真实情况为混杂因素的变量遗漏不予匹配,则会造成混杂偏倚。有向无环图是一种直观的展示不同流行病学研究设计、变量间复杂因果关系的可视化图形语言。本文从有向无环图视角分析匹配在不同观察性研究设计中的作用、匹配性病例对照研究中欲匹配变量的选择标准制定,为今后流行病学研究设计提供一定的参考建议。     

Influenza vaccination hesitancy and its determinants among elderly in China: A national cross-sectional study

BackgroundThe Chinese elderly face a significant threat from seasonal influenza, owing to the consistently low vaccination coverage. This study investigated the prevalence and determinants of influenza vaccination hesitancy among the Chinese elderly.MethodsIn 2019, 3849 elderly individuals from 10 provinces in China were recruited in a cross-sectional survey. Multinomial logistic regression was applied to investigate the determinants of influenza vaccination hesitancy.ResultsAmong the elderly respondents, 37.18% expressed some degree of hesitancy towards influenza vaccination: 19.28% were hesitant, and 17.90% refused influenza vaccination, including 19.28% acceptors with doubts and 17.90% refusers. Only 39.10% of the respondents considered themselves as the priority group for influenza vaccination, and 13.93% reported receiving a recommendation for vaccination from healthcare workers. Respondents with higher education levels and from urban areas had significantly higher odds of vaccine hesitancy than their counterparts. Confidence in the safety of vaccines was negatively associated with vaccine hesitancy, but confidence in vaccine efficacy had no such association. Respondents who perceived themselves as highly susceptible to influenza (AOR = 0.85; 95 %CI = 0.77–0.93) and those aware of the elderly as a priority group for influenza vaccination (AOR = 0.51; 95 %CI = 0.41–0.64) had a significantly lower odds of being refusers.ConclusionThis study found a high prevalence of hesitancy towards influenza vaccination among the Chinese elderly, especially well-educated and urban-dwelling respondents. The government should address vaccine hesitancy through culturally appropriate communication, subsidies for vaccination, and actively promoting vaccines through primary care professionals.