| Objective: To investigate the allocation situation of human resource in community health service organizations in Shihezi city from 2008~2014 so as to find out the current problems which have existed in the health workforce allocation. Establishing the forecasting model that suitable for the actual situation of health human resource in Shihezi City. The model was used to predict the number of community health workforce in 2015~2017 and provide scientific basis for health administrative departments at all levels to formulate training and development planning of human resource for health.Methods: The data of health workforce in community health service organizations from 2008~2014 was collected. The indicators of proportion, average speed of increase and the descriptive study of epidemiology was used in analyzing proportion of the number, professional, age, gender, education, technical post and the dynamic trends of the health workforce for its development rate in 2008~2014 of health workforce. The demands of community health manpower from 2015~2017 were projected by model of GM(Grey Dynamics Model), Multifactorial Grey Dynamics Model, ARIMA(Autoregressive integrated moving average), BP(Back propagation) neural network model and the indicators consist of MAE(mean absolute error), MAPE(Mean absolute percentage error), RMSE(Root-mean-square error) were used to evaluate the prediction results. The Microsoft office 2007, statistical package of SPSS Statistics 17(IBM SPSS Statistics 17) and MATLAB 7.1(R2010b) were applied to modeling and statistical analysis.Results: 1.Through investigating and analyzing the allocation situation of human resource and general medical training in community health service organizations, the results showed that:(1)From 2008 to 2014 the total number of staff in community health service organizations was 614 people per year. There was 565 health workers, accounting for 92.02% of the total staff, that including 39.65% of general practitioners, 37.70% of nurses, only 5.84% of public health physicians. And in 284 physicians, there was 88.38% of the physicians engaged in the medical work, doctors engaged in public health was only 11.58 %.(2)From 2008~2014 the number of worker with junior professional position and middle professional position account for 43.00%, 50.44% respectively, and the senior professional position account for only 4.96%.The proportion of senior, middle and junior professional position was 1:10.18:8.68.(3)The proportion of community physicians with undergraduate study and above accounted for 39.73% and graduate study, secondary technical study and below account for 41.07%, 18.75% respectively. The junior professional position and have accounted for 32.59 percent of physicians, middle professional position accounted for 57.14%, with senior professional position accounted for only 8.48 %. The proportion of community nurses with graduate study, secondary technical study and below, undergraduate study and above accounted for 43.19%, 47.88%, 7.98% respectively. community nurses with senior, middle, junior professional position accounted for 0.94%, 46.48%, 51.64% respectively.(4)The proportion of general practitioner that participate in the job training was 71.10%, and the community nurses accept job training accounted for 80.19%. Due to the number of acceptance standardized training of general practitioners and passing the national professional qualification of intermediate technology exam was small, the proportion of community health workers that received job training and high-quality training(such as general practitioners standardized training) was only 57.83%.2.The number of community health workers, doctors and nurses all belong to the time series data, so the model of GM(1,1) model, Multifactorial grey dynamics model, ARIMA model and BP neural network model was used to prediction.(1)To predict the community health workforce, the values of forecasting accuracy of GM(1,1) model showed that the mean absolute error(MAE) was 6.5664, with mean absolute percentage(MAPE) as 0.01191, Root-mean-square error(RMSE) as 8.7410. The predicting results of Multifactorial grey dynamics model showed that the MAE, MAPE, RMSE was 10.5379, 0.01872, 13.06670 respectively. The ARIMA model showed that the MAE was 7.4111, with MAPE as 0.01332, RMSE as 10.2453. The BP neural network model showed that the MAE, MAPE, RMSE was 5.1580, 0.009581, 6.9950 respectively. The prediction error of BP neural network model was smaller and with highly accuracy for prediction. The number of nurses of community health service organizations from 2015 to 2017 are expected to be 628, 638, 646 respectively.(2)To predict the community general practitioners, the values of forecasting accuracy of GM(1,1) model showed that the mean absolute error(MAE) was 2.5464, with mean absolute percentage(MAPE) as 0.01082, Root-mean-square error(RMSE) as 3.0232. The predicting results of Multifactorial grey dynamics model showed that the MAE, MAPE, RMSE was 13.0766, 0.05741, 15.8691 respectively. The ARIMA model showed that the MAE was 10.1262, with MAPE as 0.04491, RMSE as 12.1094. The BP neural network model showed that the MAE, MAPE, RMSE was 8.8250, 0.03861, 10.8228 respectively. The prediction error of GM(1,1) model was smaller and with highly accuracy for prediction. The number of nurses of community health service organizations from 2015 to 2017 are expected to be 235, 237, 240 respectively.(3)To predict the community nurses, the values of forecasting accuracy of GM(1,1) model showed that the mean absolute error(MAE) was 3.0407, with mean absolute percentage(MAPE) as 0.01468, Root-mean-square error(RMSE) as 3.9546. The predicting results of Multifactorial grey dynamics model showed that the MAE, MAPE, RMSE was 5.2169, 0.02512, 5.5416 respectively. The ARIMA model showed that the MAE was 3.4690, with MAPE as 0.01662, RMSE as 4.4137. The BP neural network model showed that the MAE, MAPE, RMSE was 3.0558, 0.01463, 3.5489 respectively. The prediction error of BP neural network model was smaller and with highly accuracy for prediction. The number of nurses of community health service organizations from 2015 to 2017 are expected to be 253, 262, 269 respectively.3.There was 506 health workforce in community health service organizations in 2008. According to the predicting result, it was increased to 646 and the annual average growth rate was 1.03%. In 2008, the number of general practitioners, nurses were 208, 180 respectively, and would increased to 240, 269 respectively with an annual average growth rate was 1.02%, 1.05% in 2017. Conclusion: The community health workforce still has major problems about irrational structure of profession and titles, such as community public health workforce was fewer, the education of health workforce was low, quality and level of medical staff was not higher in Shihezi city. On the analysis of number of human resources, the BP neural network model was highly accurate for the prediction of health workforce and nurses in community health service organizations. The GM(1,1) model with high accuracy for projecting the community general practitioners. The number of community health workforce showed an increased trend in Shihezi city and the growth rate of community nurses grew relatively fast. As the relevant government departments in efforts to increase the number of community health workers, should be improve the quality of General Medicine training as well, so as to lay the foundation for building a high-quality community health services contingent. |
PDF Full Text Request