Forecasting of Malaria in India using Time series analysis

Introduction:

World Health Organization (WHO) has highlighted the serious and increasing threat of vector-borne diseases with the theme "Preventing vector-borne diseases" and also with the slogan "Small bite, big threat" for the year 2014. Among vector borne diseases, malaria poses the biggest threat with about 40% of the world's population at risk of infection. In 2013, 97 countries had ongoing transmission of malaria.¹ Malaria decreases economic growth by more than one percentage point per year in endemic countries. Malaria is a life-threatening blood disease caused by a parasite that is transmitted to humans by the Anopheles mosquito. Out of 400 species of anopheline mosquitoes throughout the world, only 60 species are vectors of malaria. Whereas, in India 45 anopheline species have been reported, out of which 9 species are malarial vectors. Malaria imposes great socio-economic burden on humanity and about 36 per cent of the world population, i.e. 2020 million is exposed to the risk of contracting malaria in approximately 90 countries. World Health Organization estimates 300-500 million malaria cases annually worldwide.² Unlike Africa, where most of the deaths are reported in infants and children, in India, it is seen that malarial mortality is maximum in the economically productive age groups of 15 to 44 years.³ A malaria-stricken family spends an average of over one quarter of its income on malaria treatment, as well as paying prevention costs and suffering loss of income.⁴ Due to the severe health impact of malaria, there is a growing need for methods that will allow forecasting and early warning with timely case detection in areas of unstable transmission, so that more control measures can be implemented effectively.⁵ In India, National Vector Borne Disease Control Programme (NVBDCP) reported around 0.8 million cases of malaria in India in 2017 and there were about 104 deaths, of which 40-50 % of the cases is due to Plasmodium falciparum. Malaria can be cured by effective treatment at its early stages.

Even a century after the discovery of malaria transmission through mosquitoes in India by Sir Ronald Ross in 1897, malaria continues to be one of India's leading public health problems. In the 1930's, a treatise written by Sinton⁶ on 'what malaria costs India recorded that the problem of the very existence in many parts of India was in fact the problem of malaria. There is an increase in the availability of Medical staff over the years. With the improvement in all such facilities, malaria cases will have to be decreased and ultimately we can expect this disease to be eliminated from India like the Polio. Although there has been marked reduction in the number of malaria cases in India under National Vector Borne Disease Control Programme (NVBDCP), malaria still is the leading cause of infectious diseases with the development of drug resistant Plasmodium species and insecticide resistant mosquitoes. Keeping this in view, the present study attempts to forecast the malaria cases in India.

Methodology:

Results:

Data regarding the number of malaria cases in India from 1990 to 2017 was collected from National vector borne disease control program (NVBDC). This data were plotted on a graph to see the trend in individual country (Figure 1).

Close

Figure .1:

No. of malaria cases during the period of 1990-2017

Export to PPT

Auto Correlation Function (Figure 2) and Partial Auto Correlation Function (Figure 3) showed that an irregular increasing pattern in the number of malaria cases.

Close

Figure. 2:

Auto correlation function (ACF)

Export to PPT

Close

Figure .3:

Partial auto correlation function (PACF)

Export to PPT

Henceforth, ARIMA models (p, d, q), for such scenario was applied. The best suitable model was selected based on minimal Bayesian Information Criteria (BIC) value. In this study, the least BIC value is 0.070 (Table 1) and the corresponding model is ARIMA (10, 1, 9) with goodness of fit (R²=93%). The model verification is done by checking the residuals of the model (Figure 4). This is done through examining the autocorrelation and correlation of the residuals of various orders.

Close

Figure .4:

Residual ACF and PACF

Export to PPT

The forecasted values for 2018 to 2025 were in proximity to the actual values (Figure.5 and Table. 2) and there by the validity of the model were ensured.

Close

Figure 5:

Actual and forecasted cases

Export to PPT

Discussion:

As early mentioned above ARIMA model consist of four steps, the first step was the identification of the model. The model identification done by ACF and PACF (Fig 2 & 3), it exposed increasing pattern in the number of cases of Malaria. Model parameters were estimated using SPSS ver. 23.0. In the forthcoming years (2018 to 2025), the model predicted a gradually increased. The malaria case which is 0.84 million in 2017 in expected to reduce to 0.64 millions in the year 2025. This implies that the existing interventions in our country will not be adequate to control the on-going epidemic even in another 8 years. With the current ongoing interventions in the country to control and manage malaria its seems to be a Himalayan task to achieve the target of sustainable developmental goal which has set an target to end the epidemics of AIDS, tuberculosis, malaria by the year 2030.

A rule of the thumb is to have at least 50 data points for developing a good ARIMA model. But in this study we have used 27 data points which is a limitation of this study.

Conclusion:

In this study, forecasting methods were applied to predict the number of malaria cases in India based on cases reported by NVDCP. Model prediction was done using ARIMA models in which the appropriate model was identified using minimum BIC value. The trend in forecasted values from 2018 to 2025 reveals that there is a steady decrease in the number of cases of Malaria. But the decline in the number of cases in not adequate enough to stop the malaria epidemic as stated in SDG. Hence, to prevent the disease from getting established as an endemic in these countries, additional interventions with an increase in the intensity of existing interventions and support of the international community along with WHO is essential.