Methods to include persons living with HIV not receiving HIV care in the Medical Monitoring Project

The Medical Monitoring Project (MMP) is an HIV surveillance system that provides national estimates of HIV-related behaviors and clinical outcomes. When first implemented, MMP excluded persons living with HIV not receiving HIV care. This analysis will describe new case-surveillance-based methods to identify and recruit persons living with HIV who are out of care and at elevated risk for mortality and ongoing HIV transmission. Stratified random samples of all persons living with HIV were selected from the National HIV Surveillance System in five public health jurisdictions from 2012–2014. Sampled persons were located and contacted through seven different data sources and five methods of contact to collect interviews and medical record abstractions. Data were weighted for non-response and case reporting delay. The modified sampling methodology yielded 1159 interviews (adjusted response rate, 44.5%) and matching medical record abstractions for 1087 (93.8%). Of persons with both interview and medical record data, 264 (24.3%) would not have been included using prior MMP methods. Significant predictors were identified for successful contact (e.g., retention in care, adjusted Odds Ratio [aOR] 5.02; 95% Confidence Interval [CI] 1.98–12.73), interview (e.g. moving out of jurisdiction, aOR 0.24; 95% CI: 0.12–0.46) and case reporting delay (e.g. rural residence, aOR 3.18; 95% CI: 2.09–4.85). Case-surveillance-based sampling resulted in a comparable response rate to existing MMP methods while providing information on an important new population. These methods have since been adopted by the nationally representative MMP surveillance system, offering a model for public health program, research and surveillance endeavors seeking inclusion of all persons living with HIV.

医疗监测项目（Medical Monitoring Project，简称MMP）是一套艾滋病监测系统，可提供全国范围内艾滋病相关行为与临床结局的评估数据。该项目最初启动时，未纳入未接受艾滋病诊疗的人类免疫缺陷病毒（Human Immunodeficiency Virus，简称HIV）感染者。本研究分析将介绍基于病例监测的新型方法，用于识别并招募脱离诊疗、面临较高死亡风险及持续HIV传播风险的HIV感染者。
2012至2014年间，研究团队从5个公共卫生辖区的全国艾滋病监测系统中，抽取所有HIV感染者的分层随机样本。通过7类数据源与5种联络方式对抽中对象进行定位与联系，以开展访谈及病历摘要采集工作，并针对无应答及病例报告延迟情况对数据进行加权调整。
改良后的抽样方法共完成1159份访谈（调整后应答率为44.5%），并为其中1087名对象（占比93.8%）匹配到了对应的病历摘要数据。在同时拥有访谈与病历数据的对象中，有264人（占比24.3%）若采用原MMP方法将无法被纳入研究。
研究确定了成功联络（如保持诊疗随访，调整后比值比（adjusted Odds Ratio，简称aOR）=5.02；95%置信区间（Confidence Interval，简称CI）=1.98~12.73）、完成访谈（如迁出所属辖区，aOR=0.24；95% CI=0.12~0.46）及病例报告延迟（如居住于农村地区，aOR=3.18；95% CI=2.09~4.85）的显著预测因素。
基于病例监测的抽样方法与原MMP方法的应答率相当，同时还获取了重要的新增人群相关数据。目前该方法已被具备全国代表性的MMP监测系统采纳，为公共卫生项目、研究及监测工作中纳入所有HIV感染者提供了可推广的实践范式。