Physician and nurse supply in Serbia using time-series data: A Case Study Abstract Background: Unemployment among health professionals in Serbia has risen in the recent past

Background: Unemployment among health professionals in Serbia has risen in the recent past and continues to increase. This highlights the need to understand how to change policies to meet real and projected needs. This study identifies variables that were significantly related to physician and nurse employment rates in the public healthcare sector in Serbia from 1961 to 2008 and used these to develop parameters to model physician and nurse supply in the public healthcare sector through 2015. Methods: The relationships among six variables used for planning the physician and nurse employment in public healthcare sector in Serbia were identified for two periods: 1961-1982 and 1983-2008. Those variables included: the annual total national population; gross domestic product adjusted to 1994 prices; inpatient care discharges; outpatient care visits; students enrolled in the first year of medical studies at public universities; and the annual number of graduated physicians. Based on historic trends, physician supply and nurse supply in the public healthcare sector by 2015 (with corresponding 95% confidence level) have been modeled using Autoregressive Integrated Moving Average (ARIMA) / Transfer function (TF) models. Results: The ARIMA/TF modelling yielded stable and significant forecasts of physician supply (stationary-R squared = 0.71) and nurse supply (stationary-R squared = 0.92) in the public healthcare sector in Serbia through 2015. The most significant predictors for physician employment were the population and GDP. The supply of nursing staff was, in turn, related to