Forecasting cesarean deliveries with robust time series models in a tertiary care hospital

Abstract

Introduction: Caesarian section rate was 47% at Patan Hospital in 2014 despite the recommendation of World Health Organization for keeping it below 15%. This has become a public health problem and now debated as the human right violation of childbearing women. This study aims to use robust time series model to get valid forecasts of caesarian deliveries for this hospital.

Method: Univariate time series models were used to forecast 3-year caesarean deliveries using 60-month (2010- 2014) data of Patan Hospital. A time series model with low mean average percentage error from validation period and without autocorrelation problem was selected as robust model and it was used to forecast the caesarean deliveries for 2015-2017 periods.

Result: Winter’s additive model had lowest validation forecasting error and showed decreasing trend of caesarian deliveries but it showed autocorrelation problem. Quadratic regression gave similar results but is also had problem of autocorrelation. Artificial Neural Network – Multilayer Perceptron model gave close forecasts but autocorrelation was not assessed. Best Autoregressive Integrated Moving Average model with (0,1,1),(0,0,0) parameters did not show autocorrelation for 60-month data and taken as robust model for doing 3-year forecasts.

Conclusion: Best ARIMA model with one difference stationary and first order of moving average correction gave valid forecasts for Patan Hospital. Advanced univariate and multivariate time series models with large samples can be used to get more precise forecasts of caesarean deliveries in Nepal.