Use of a statistical computational simulation to predict intensity of malaria transmission from data of clinical symptomatic episodes of malaria and climate

Abstract

Abstract Data on the monthly clinical episodes of malaria and prevalence from laboratory diagnosis of patients for malaria infection was obtained from an array of data gathered from malaria parasite tests conducted on patients clinically diagnosed for malaria in health centers within the study area in Akinyele Local Government Area of Ibadan city in Nigeria, for years 1997, 1998, 1999, 2000, 2001 and 2005 (6years) which falls between years 1997–2005. Also, data was gathered for climatic factors (rainfall, relative humidity, temperature and sunshine hours) for all years between years 1997 and 2005 (9years complete) from Geospatial Laboratory of International Institute of Tropical Agriculture IITA in Ibadan, Oyo State, Nigeria. Thereafter, we engaged statistical methods with computational support from Microsoft Excel version 2007, to generate a climate based- simulation to predict periods of the years for which there were high malaria intensity for malaria. We could not retrieve complete data for prevalence (laboratory positive results for tests) the month for October. So, we proceeded to determine the correlation between clinical episodes and prevalence for the 6 years for which we retrieved data. The Pearson moment correlation coefficient “r” between clinical symptomatic episode and positive outcomes of tests (prevalence of infection) as computed from Microsoft Excel was + 0.986265 This shows a high enough positive correlation, upon which we could use the clinical episodes to compute of simulations to predict periods of high intensity of clinical symptomatic episodes and which can then be related to the intensity for prevalence of malaria. The statistical computations indicated high intensity of clinical episodes to correlate (correspond) with rise for the climatic factors, and low intensities for lowered levels of most of the climatic factors for years 2002 and 2004, as they both recorded positive ranges of correlation “r” values between clinical episode and climatic factors. This can be used to predict periods of the year with high intensity of clinical episodes of malaria as our simulated prediction. Then we conducted two test-runs using two observed variants in the climate based-yearly periods of high intensity (those of years 1998 and 2001). The predictions indicated matches for periods of high intensity transmission using statistical tool of Pearson’s moment correlation analysis derived relationships and other descriptive statistical attributes. These range of correlative value matches were between the precise values of correlation coefficients of the obtained laboratory data and that of calculated predictive ranges of these values. Since the Pearson correlation between clinical episode and prevalence of malaria was high (close to 1.0), these simulation can assist to predict prevalence of infection obtained from the laboratory diagnosis. Our analysis and predictive simulations will require future extraction of more data to input into the simulation and run more tests with other support statistical tools to see how the trends in the output from the simulation perform. If successful, we channel this simulative prediction of malaria transmission intensity into a built algorithm involving use of machine learning platforms.