Big data analytics framework for childhood infectious disease surveillance system using modified mapreduce algorithm: a case study of Tanzania

Abstract

Tanzania has been affected with a potential emerging and re-emerging of infectious diseases such as diarrhea, acute respiratory infections, pneumonia, hepatitis, and measles. There is an increasing trend for the occurrences of new emerging pandemic diseases such as the coronavirus (Covid-19) in 2020 as well as re-occurrence of old infectious diseases such as cholera epidemic in 2015-2017, chikungunya and dengue fever outbreak in 2010, 2012, 2014, 2018, and 2019 which affected different regions in Tanzania. These diseases by far are the main causes of the high mortality rate for women and children of 0-5 years of age. The traditional disease surveillance system as the foundation of the public healthcare practices has been facing challenges in data collection and analysis using health big data sources to prevent and control infectious diseases. Health big data sources on infectious diseases have been recognized as the potential supplement for the provision of evidence-based decision-making worldwide. Tanzania as one of the resource-limited setting countries has lagged because of the challenges in information technology infrastructure and public healthcare resources. The traditional disease surveillance system is still paper-based, semi-automated, and limited in scope which relies on clinical-oriented patient data sources and leaving out nontraditional and pre-diagnostic unstructured big data sources. This research study aimed to improve the traditional infectious disease surveillance system to employ big data analytics technology in healthcare data collection and analysis to improve decision-making. Big data analytics framework for the childhood infectious disease surveillance system was developed which guides healthcare professionals to streamline the collection and analysis of health big data for infectious disease surveillance. The framework was then fairly compared with the existing framework in its performance using infrastructures, data size and transformation, and running-time execution of the systems. The experimental results indicate the efficiency of the framework system performance with the highest running time execution of about 56% quicker over the traditional system. Also, it has the best performance in processing multiple data structures using additional processing units. In particular, the proposed framework can be adopted to improve the prenatal and postnatal healthcare system in Tanzania.