Real-time record linkage between demographic surveillance and health facility data for monitoring access and utilization of services in rural Tanzania

Christopher Rentsch, London School of Hygiene and Tropical Medicine (LSHTM)
Georges Reniers, London School of Hygiene and Tropical Medicine (LSHTM)
Richard Machemba, National Institute for Medical Research, Tanzania
Baltazar Mntenga, National Institute for Medical Research, Tanzania
Chodziwadziwa Whiteson Kabudula, University of the Witwatersrand
Denna Michael, National Institute for Medical Research, Tanzania
Mark Urassa, National Institute for Medical Research, Tanzania
Jim Todd, London School of Hygiene and Tropical Medicine (LSHTM)
Basia Zaba, London School of Hygiene and Tropical Medicine (LSHTM)

Most analyses of health service use and its outcomes are limited to patients who are enrolled into clinical care and lack a population perspective. In contrast, health and demographic surveillance systems (HDSS) often rely on self or proxy reports of health services use and usually lack the necessary detail and accuracy about clinical events. Since June 2015, we have been linking patients of three clinics to their HDSS record in real-time, to produce an augmented data source for better monitoring access to and utilization of health facility services in a rural community in Tanzania. So far, we have consented 1,360 patients, and among those who reportedly live in the area covered by the HDSS, we have matched 78 percent to their HDSS record. In this manuscript, we describe the record linkage methodology, report record linkage statistics, and discuss some of the patient attributes that are correlated with (un)successful record linkage.