JMIR Publications

ABSTRACT

Background:

This work details the outbreak and factors affecting the spread of novel coronavirus (COVID-19) in the Indian state of Maharashtra, which is considered as one of the most massive and deadly pandemic outbreaks. Observational data collected between 14 March 2020 and 4 May 2020 is statistically analyzed to determine the nonlinear behaviour of the epidemic. It is followed by validating predicted results with real-time data. Proposed model is further used to obtain statistical summaries in which Grubbs tests for outlier detection have justified high values of evaluation metrics. Outliers are found to be pilot elements in an outbreak under considered region. Statistically, a significant correlation has been observed between dependent and explanatory variables. Transmission pattern of this virus is very much different from the SARS-CoV-1 virus. Key findings of this work will be predominant in maintaining environment conditions at healthcare facilities to reduce transmission rates at these most vulnerable places.

Objective:

Study is aimed to make aware the government sources and public about the tentative outbreak by mid July in Indian state of Maharshtra to help them in prevention of spread. Study is not registered as it involves forecasting and data is not about diagnostic treatment of patient

Methods:

Data collection from government sites, Data cleaning, sorting, Non-Linear regression analysis, Easy fit tool of AI, Oultlier detection, Evaluation Metric, forecasting, mathematical modelling using multiregression analysis

Results:

The general trend of data in confirmed cases vs humidity plot is observed as an increase in confirmed cases with an increase in humidity in different regions under observation in Maharashtra. This shows a positive correlation between humidity and confirmed cases, as shown in Fig.12. As we traverse from the western coastal area towards the eastern part of the state, a significant drop in humidity is observed with the remarkable lower number of confirmed cases. Whereas in the case of confirmed cases vs temperature, the dependent variable drops significantly low with an increase in temperature of air showing a negative correlation between temperature and confirmed cases as shown in Fig.13

Conclusions:

Third-order polynomial nonlinear regression is evaluated by statistical modelling and validated against real data obtained. Nonlinear regression is carried out to assess the model of an outbreak of novel coronavirus COVID-19 in the state of Maharashtra. Machine learning tool- random forest test is carried out to detect the presence outliers in the data and asses the model as a complex nonlinear model due to retaining outliers, which cause high errors and deviations. Third-order polynomial predicted the outbreak correctly. Based upon this, advanced measures can be taken to have adequate medical facilities in place to withstand a possible outbreak. It is found that despite having unfavorable environment conditions of extremely higher humidity and significantly higher temperatures ( for earlier virus strains), an outbreak of COVID-19 is extensive in densely populated areas. This signifies that social distancing, which includes negligible surface contacts, is of utmost importance in public places where humidity and temperatures cannot be controlled. Whereas in specialized COVID-19 dedicated hospitals, clinics, housing facilities, and other indoor areas where surface contacts are significantly more & temperature, humidity can be controlled, these variables must in the precise range which will prevent both; viruses from becoming airborne and also from drying out on the surface and forming salt casing. Humidity less than 60% and temperature higher than 250 C is to be preferred in hospitals to prevent the spread. Dense areas should be vacated, and the corresponding population needs to be relocated to a different region of the state without affecting the concerned natives. Clinical Trial: Study is not registered as it involves forecasting and data is not about diagnostic treatment of patient