Zambart Data Department

The majority of the 600 plus Zambart employees currently have access to a computerized device ranging from desktop and laptop computers for staff in head- and satellite offices to PDA’s (Personal Digital Assistants) and android/windows tablets for field staff. The data department supports all of Zambart research and administrative activities with IT-services, data management and statistics.

The data team offers expertise in IT-computer networking, software engineering, geographical information systems, data management, laboratory logistics, medical statistics, epidemiology, monitoring and evaluation, etc.

The data department is involved in all stages of Zambart’s studies, including study design, sampling and field procedures, quality control, database development, data analysis, and supervision and training of field staff.

The community randomized trials that Zambart conducts are massive in size. For example, the ZAMSTAR Zambia, South Africa Tuberculosis and AIDS (ZAMSTAR) study covered a population of 800,000 to study the impact of community and household based interventions on prevalence and transmission of TB, while the Better Health Outcomes through Mentoring and Assessment (BHOMA) trial studies the impact of improved health services on mortality and morbidity in population covered by 42 rural health centers. The HPTN 071/ PopART study focuses on the reduction of HIV transmission as the result of universal testing and treatment in a population of 600,000 people.

We have therefore shifted from paper-based surveys, to electronic data capturing by field staff. Some examples can be found below.

A state of art infrastructure has been deployed at Zambart with a local area network supporting 60-80 office computers in Zambart house and a wide area network to connect district offices and field staff.

Data data1 data2

In 2010, we conducted a TB prevalence survey by selecting 80,000 participants from 16 communities. For consenting participants a questionnaire was administered, a sputum sample taken and optionally body measurements and HIV testing was done in the household. Data collection was done using ruggedized PDA’s (Personal Digital Assistant).

data3

With Google Earth we draw around the clinic and subdivide the area into squares (picture coming soon). The squares are put in random order and visited. All members of the household in a square are asked to participate.

The BHOMA trial (Better Health Outcomes through Mentoring and Assessment) studies the impact of health care strengthening and deployment of Community Health Workers on mortality and morbidity in rural communities in Zambia. Evaluation of the intervention is done through DHS-like surveys. Due to the lack of reliable census information in these areas and uncertainty around the catchment area of the clinic we have chosen to use an alternative sampling method:

  • With Google Earth we draw around the clinic and subdivide the area into squares (left picture).
  • The squares are put in random order and visited. All members of the household in a square are asked to participate
  • We stop if the target of 300 households per community is met
  • The electronic device will tell the field teams if a household is located outside the square (middle picture above) or inside the square (right picture above)

data4

In the preparatory phase of the PopART trial (hyperlink here), a so-called household census was done by collecting the location (GPS) and few details of every household in the community. This data was used as a sampling frame for the impact survey and for planning purposes of deployment of community health workers in PopART.

data5

Over 400 lay counsellors (CHiPs- Community HIV Prevention – workers) are employed to go door-to-door in 12 communities in Zambia, and offer a comprehensive HIV prevention package (including HIV counselling and testing). Tablet computers are used to log their activities and data is send to a central server on a daily basis.