Forecasting methods in health services
Forecasting is a mathematical prediction process that uses the data from the past records to predict the future outcomes. In the case of the health services, it uses the number of patients visiting a given health facility for a given period of time, like, say, for a whole year or a month in order to predict the estimated number of expected patients in the following year. This helps the health facility managers to make a prior plan of financial and material expenditure for the year or month to come.
In addition, forecasting helps to monitor the operational efforts in the health sector, the need to increase the number of staffs in relation to the number of expected patients and monitoring the cost of operation. It also helps a given health facility to evaluate its performance in relation to the expected health norms and standards (Burns, 2002).
Steps in the forecasting process
The following steps are followed when forecasting: first is to identify the forecast goal, that is, the reasons for conducting the forecast. Secondly, develop a time horizon. This is the period for which you want to conduct the forecast. This can be per hour, day, week, and month or per year. This paper will focus on a monthly time horizon. Thirdly, select the technique to be used in the forecast. Next is to carry out the forecast and finally, evaluate the forecast conducted.
Forecasting approaches
There are two approaches to forecasting. One is judgmental forecast, which relies on opinions of managers and staff, and the other is a time series approach of forecasting. The time series approach uses observations taken over a given period of time such as hourly, daily, weekly, monthly or yearly. This paper will use the time series method to analyze the number of visits in the Northern College Health Services for the month of November, 2008.
Forecasting methods
There are four common methods of forecasting. These include seasoned forecast, percent adjustment, trendline and 12-month moving average. This paper compares the last two methods, that is, the trendline and the 12-month moving average in order to determine which one works best.
Trendline forecasting
This is a quantitative method of forecasting based on tangible numbers from the past data. This data is then plotted on the graph, which can then be extrapolated to predict future trends and behavior of the case under study. This kind of time series uses several types of trend patterns such as constant patterns, linear patterns and exponential patterns.
12-month moving average
This is a time series forecasting method where data are analysed for 12 months and then used to predict the coming months.
Moving averages (MA)
In the moving average, data obtained from the most recent period are used.
For example, the data below is taken from the OB clinical analysing the number of visits that the patients make to the clinic for 5 years (Childs, 1989). The data are then used to determine the trend for the 6th year in order to approximate the volume of the expected work in that year. This helped them to consider sufficient staffing and ensure enough finance to take care of the operations in the 6th year.
The data above was used by the hospital to calculate the moving average as below
Ft=MAn=Ain
In which Ft=forecasting for a time t
MAn=moving average for a period n
Ai=actual value for a given age i
i=age
n=number of periods within the moving average
Thus for year 6, F6=MA3=14,271+13,175+10,0223=12,489
Thus the OB clinic expected 12,489 in the 6th year. Though this is an approximation, it shows an increase from the fifth year (year 5) and so the management in the clinic had to prepare for this swelling number of patients.
Northern college health visit
Using the same trend, we can determine the number of visits expected in November 2008 in the Northern college health services.
The data given for the month of November for the three previous years is as follows
November 200530
November 200629
November 200746
November 2008?
Using the above formula for the moving average,
Fnov,2008=(30+29+46)3=35
Thus the expected number of visits in November 2008 is 35.
Conclusion
Considering the forecasting methods above, moving average tends to be more accurate as compared to the trendline. This is because the percentage error in this calculation is minimum since it uses exact figures in calculation. In contrast, the trendline only depends on the graph which in the process of plotting, is accompanied by a series of errors. In addition, the value is only determined through extrapolation of the graph, which is further accompanied by approximation. Thus the moving average is a better method. Using the moving average, the managers of the health service are able to incorporate several data points and finally to predict as many values for the months to come as possible. Trendlines uses regression analysis in order to approximate linear correlation that exists within a given set of data.
References
Burns, L. R. (2002). The health care value chain. San Francisco: Jossey-Bass.
Childs, B. (1989). Bedside Terminals: A new player. U.S Healthcare , pp. 6-7.
