Versatility, value and limitations of using Health and Demographic Surveillance Site data for secondary analyses: guidance for researchers using examples from existing analyses

Health and Demographic Surveillance sites (HDSS) are geographic open cohorts operating in countries with absent/incomplete vital registration. Data on demographic events, socio-demographic indicators, and certain health conditions are regularly gathered on the whole population of a small area, sometimes for decades. In the same countries there are often also nationally-representative demographic data available from roughly quinquennial Demographic and Health Surveys (DHS). This paper uses a comparison with DHS data, and an in-depth review of complex HDSS analyses to demonstrate the utility of using HDSS data for secondary analyses, and to provide guidance on the conduct and reporting of these analyses. DHS data has advantages in terms of representativeness and data access, and HDSS in terms of scope for complex longitudinal analyses which may take household and familial contexts into account. HDSS also have issues which make interpretation of data and conclusions challenging: lack of data on in- and out-migrants when they are outside of the area, and repeatedly collected data may result in inconsistencies and/or more reliable data for longer-term residents. Despite these challenges, the reviewed HDSS data analyses demonstrate the flexibility and unique strengths of HDSS data. HDSS data users are recommended to clearly state their methods, particularly how they approached the issues specific to HDSS analyses of handling repeated data, migration and missing data: while there were interesting ways used to approach these issues, they were often not discussed. HDSS data producers are further encouraged to ensure that the data is being used to its full potential.