Castel Amanda D, Terzian Arpi, Opoku Jenevieve, Happ Lindsey Powers, Younes Naji, Kharfen Michael, Greenberg Alan
Department of Epidemiology and Biostatistics, Milken Institute School of Public Health, The George Washington University, Washington, DC, United States.
HIV/AIDS, Hepatitis, STD, and TB Administration, The District of Columbia Department of Health, Washington, DC, United States.
JMIR Public Health Surveill. 2018 Mar 16;4(1):e23. doi: 10.2196/publichealth.9221.
Triangulation of data from multiple sources such as clinical cohort and surveillance data can help improve our ability to describe care patterns, service utilization, comorbidities, and ultimately measure and monitor clinical outcomes among persons living with HIV infection.
The objective of this study was to determine whether linkage of clinical cohort data and routinely collected HIV surveillance data would enhance the completeness and accuracy of each database and improve the understanding of care patterns and clinical outcomes.
We linked data from the District of Columbia (DC) Cohort, a large HIV observational clinical cohort, with Washington, DC, Department of Health (DOH) surveillance data between January 2011 and June 2015. We determined percent concordance between select variables in the pre- and postlinked databases using kappa test statistics. We compared retention in care (RIC), viral suppression (VS), sexually transmitted diseases (STDs), and non-HIV comorbid conditions (eg, hypertension) and compared HIV clinic visit patterns determined using the prelinked database (DC Cohort) versus the postlinked database (DC Cohort + DOH) using chi-square testing. Additionally, we compared sociodemographic characteristics, RIC, and VS among participants receiving HIV care at ≥3 sites versus <3 sites using chi-square testing.
Of the 6054 DC Cohort participants, 5521 (91.19%) were included in the postlinked database and enrolled at a single DC Cohort site. The majority of the participants was male, black, and had men who have sex with men (MSM) as their HIV risk factor. In the postlinked database, 619 STD diagnoses previously unknown to the DC Cohort were identified. Additionally, the proportion of participants with RIC was higher compared with the prelinked database (59.83%, 2678/4476 vs 64.95%, 2907/4476; P<.001) and the proportion with VS was lower (87.85%, 2277/2592 vs 85.15%, 2391/2808; P<.001). Almost a quarter of participants (23.06%, 1279/5521) were identified as receiving HIV care at ≥2 sites (postlinked database). The participants using ≥3 care sites were more likely to achieve RIC (80.7%, 234/290 vs 62.61%, 2197/3509) but less likely to achieve VS (72.3%, 154/213 vs 89.51%, 1869/2088). The participants using ≥3 care sites were more likely to have unstable housing (15.1%, 64/424 vs 8.96%, 380/4242), public insurance (86.1%, 365/424 vs 57.57%, 2442/4242), comorbid conditions (eg, hypertension) (37.7%, 160/424 vs 22.98%, 975/4242), and have acquired immunodeficiency syndrome (77.8%, 330/424 vs 61.20%, 2596/4242) (all P<.001).
Linking surveillance and clinical data resulted in the improved completeness of each database and a larger volume of available data to evaluate HIV outcomes, allowing for refinement of HIV care continuum estimates. The postlinked database also highlighted important differences between participants who sought HIV care at multiple clinical sites. Our findings suggest that combined datasets can enhance evaluation of HIV-related outcomes across an entire metropolitan area. Future research will evaluate how to best utilize this information to improve outcomes in addition to monitoring them.
整合来自多个来源的数据，如临床队列和监测数据，有助于提高我们描述护理模式、服务利用、合并症的能力，并最终测量和监测艾滋病毒感染者的临床结局。
本研究的目的是确定临床队列数据与常规收集的艾滋病毒监测数据的关联是否会提高每个数据库的完整性和准确性，并增进对护理模式和临床结局的理解。
我们将来自一个大型艾滋病毒观察性临床队列——哥伦比亚特区（DC）队列的数据，与2011年1月至2015年6月期间华盛顿特区卫生部（DOH）的监测数据进行了关联。我们使用kappa检验统计量确定了预链接数据库和后链接数据库中选定变量之间的一致性百分比。我们比较了护理保留率（RIC）、病毒抑制（VS）、性传播疾病（STDs）和非艾滋病毒合并症（如高血压），并使用卡方检验比较了使用预链接数据库（DC队列）与后链接数据库（DC队列+DOH）确定的艾滋病毒门诊就诊模式。此外，我们使用卡方检验比较了在≥3个地点接受艾滋病毒护理的参与者与<3个地点接受护理的参与者之间的社会人口学特征、RIC和VS。
在6054名DC队列参与者中，5521名（91.19%）被纳入后链接数据库，并在单个DC队列地点登记。大多数参与者为男性、黑人，以男男性行为者（MSM）作为艾滋病毒风险因素。在后链接数据库中，发现了619例DC队列先前未知的性传播疾病诊断。此外，与预链接数据库相比，RIC参与者的比例更高（59.83%，2678/4476对64.95%，2907/4476；P<.001），而VS参与者的比例更低（87.85%，2277/2592对85.15%，2391/2808；P<.001）。几乎四分之一的参与者（23.06%，1279/五岁21）被确定在≥2个地点接受艾滋病毒护理（后链接数据库）。使用≥3个护理地点的参与者更有可能实现RIC（80.7%，对62.61%，2197/3509），但实现VS的可能性较小（72.3%，154/213对89.51%，1869/2088）。使用≥3个护理地点的参与者更有可能住房不稳定（15.1%，64/424对8.96%，380/4242）、拥有公共保险（86.1%，365/424对57.57%，2442/4242）、患有合并症（如高血压）（37.7%，160/424对22.98%，975/4242），并患有获得性免疫缺陷综合征（77.8%，330/424对61.20%，2596/4242）（所有P<.001）。
将监测数据与临床数据相链接提高了每个数据库的完整性，并增加了可用于评估艾滋病毒结局的可用数据量，从而使艾滋病毒护理连续体估计得以细化。后链接数据库还突出了在多个临床地点寻求艾滋病毒护理的参与者之间的重要差异。我们的研究结果表明，合并数据集可以加强对整个大都市地区艾滋病毒相关结局的评估。未来的研究将评估除了监测这些信息之外，如何最好地利用这些信息来改善结局。
