出版社：International Association for Social Science Information Service & Technology
摘要：The Data Documentation Initiative (DDI) specification has gone through significant development in recent years.Most Health and Demographic Surveillance System (HDSS) researchers in Low and Middle Income Countries(LMIC) are, however, unclear on how to apply it to their work. This paper sets out considerations that LMIC HDSSresearchers need to make regarding DDI use. We use the Kisesa HDSS in Mwanza Tanzania as a prototype.First, we mapped the Kisesa HDSS data production process to the Generic Longitudinal Business Process Model(GLBPM). Next, we used existing GLBPM to DDI mapping to guide us on the DDI elements to use. We thenexplored implementation of DDI using the tools Nesstar Publisher for the DDI Codebook version and ColecticaDesigner for the DDI Lifecycle version.We found the amounts of metadata entry comparable between Nesstar Publisher and Colectica Designer whendocumenting a study from scratch. The majority of metadata had to be entered manually. Automatically extractedmetadata amounted to at most 48% in Nesstar Publisher and 33% in Colectica Designer. We found ColecticaDesigner to have stiffer staff training needs and software costs than Nesstar Publisher.Our study shows that, at least for HDSS in LMIC, it is unlikely to be the amount of metadata entry that determinesthe choice between DDI Codebook and DDI Lifecycle but rather staff training needs and software costs. LMICHDSS studies would need to invest in extensive staff training to directly start with DDI Lifecycle or they could startwith DDI Codebook and move to DDI Lifecycle later.
