Abstract
Using the officially published data and aware of the uncertain source and insufficient number of samples, we present a first and (for the moment) unique attempt to study the first two months spread of the pandemic COVID-19 in Madagascar. The approach has been tested by predicting the number of contaminated persons for the next week after fitting the inputs data collected within 7 or 15 days using standard least χ(2)-fit method. Encouraged by this first test, we study systematically during 67 days, 1-2 weeks new data and predict the contaminated persons for the coming week. We find that the first month data are well described by a linear or quadratic polynomial with an increase of about (4-5) infected persons per day. Pursuing the analysis, one note that data until 46 days favours a cubic polynomial behaviour which signals an eventual near future stronger growth as confirmed by the new data on the 48th day. We complete the analysis until 67 days and find that the data until 77 days confirm the cubic polynomial behaviour which is a remarkable feature of the pandemic spread in Madagascar. We expect that these results will be useful for some new model buildings. A comparison with some other SI-like models predictions is done. These results for infected persons may also be interpreted as the lowest values of the real cases due to the insufficient number of samples (about 12,907 for 27 million habitants on 05/06/20). The data analysis of the absolute number of cured persons until 67 days shows an approximate linear behaviour with about 3 cured persons per day. However, the number of percentage number of cured persons decreases above 42-46 days indicating the limits of the hospital equipment and care to face the 2nd phase of the pandemic for the 67th first days. Some comments on the social, economical and political impacts of COVID-19 and confinement for Madagascar and, in general, for Worldwide are shortly discussed.