Permanent URI for this collection
Browse
Recent Submissions
Item Big data analytics model for early infectious disease outbreak prediction in public health.(Tshwane University of Technology, 2023-08-01) Busobozi, Viola Vivian; Prof. Sello Mokwena; Dr. Appolonia I. IlorahThe public health sector is facing challenges of delays in timeous prediction of infectious disease outbreaks. Despite the surveillance systems in place, there are delays in information aggregation and processing times. Real time feedback is indispensable in epidemic disease outbreak situations. Present surveillance systems produce accurate prediction and tracking after unacceptably long periods. Surveillance results are rendered redundant, unless they are provided in a timely manner. Big Data analytics (BDA) could address the lag between data collection and processing. Adopting and implementing BDA may improve public healthcare services. This study proposed to develop a public health Big Data Analytics Predictive Model (BDA-PH-PM) that would be adopted as a strategy to facilitate early prediction of disease healthcare outcome from infectious disease outbreak. To achieve the above goal, this study used contingency theory constructs used in Management Information System (MIS) research that underpinned this study. Contingency variables were used to address non-data management variables like strategy, structure, environment, individual task, and technology. BDA MIS variables addressed data management elements which are technological or technical and these included BDA management, BDA implementation, BDA structure, and BDA development. Data was collected from hospital employees targeting operational staff in data management such as HMIS officers and records officers. The managers like biostaticians and district health teams involved in decision making on disease outbreak management were also targeted. Qualitative responses were integrated, subthemes that were used to formulate questionnaires that were used in a survey. An applied conceptual Architecture of Big Data Analytics framework was used to guide me on the flow of analysis procedures. Quantitative data collected as a source data from questionnaires was pre-processed into desirable formats. Further variable selection was done to select the appropriated variables for model development. The collected data was partitioned into 70% of training data and 30% test data. Training data was used for variable selection and model development whereas testing data used for model evaluation. A threshold of 0.5 which is a default threshold for variable selection and model building was used and 4 variables were selected. The key finding was that multiple regression model (MRM) was the selected model that was used to develop a BDA-PH-PM. The developed model was used to predict the factors or challenges that may influence prediction of public health outcome. The results showed that organisation structure, BDA management, environment and BDA structure were the factors that could influence the prediction of public health outcome if adopted