Forecasting Methods In Health Services Organizations

Method 1: Average Change

Varied steps involved in the procedure of carrying out forecasting are hereby mentioned below:

Explanation of the 1^st Method:  Change of Average

The first as well as foremost step involved in forecasting is particularly plotting the definite data prior to any application of any technique of prediction mechanisms. Based on the plotting of the specific data, a definite trend can be witnessed. Essentially, the mechanisms of the average change can necessarily be illustrated using the following steps explained herein below:

Stage 1 refers to computation of the change/alteration every month

Stage 2 refers to computation of average alteration every month

Stage 3 refers to computation of the specific mid-point of the acquired data. In case if the total number of data is particularly even in number, then in that case, the mid point can be determined by the formula (n+1)/2, in which “n” stands for the total number of data. However, in case if the total number of data is essentially odd in number, then in that case, the mid-point can be ascertained by using the formula: n/2 (Mead, 2017). 

Stage 4 indicates towards computation of the forecast for specified or given month. This can be enumerated by the formula:

Average visits per month + (Average change per month) X (mid-point of the data)

Explanation of the 2^nd  Method:  Confidence Interval

Confidence interval refers to a mechanism that presents a specific level of interval for the prediction (Mertler & Reinhart, 2016). In this case, the prediction is founded on confidence interval of particularly 95%. In itself, confidence interval of 95% is presented as average amount ±1.96 x (SD) standard deviation]

Stage 1 of this method includes computation of the mean as well as standard deviation

Stage 2 of this particular method refers to ascertainment of the confidence interval

(that is 95% Confidence Interval = average visits ± 1.96 X (SD) standard deviation of total visits

Explanation of the 3^rd   Method:  Average change of percentage

 This specific method is on the groundwork of calculation of percentage change of average. Various steps involved in this process are hereby mentioned below:

Step 1 refers to change from prior methods and thereafter divided by the value obtained from the prior method (Mertler & Reinhart, 2016)

Step 2 includes enumeration of the percentage alteration

Step 3 refers to enumeration of average of change in percentage

Finally Step 4 indicates towards determination of the forecasted figure. This forecasted figure is calculated by using the following formula:

Method 2: Confidence Interval

Latest Month+ Average change in percentage X Most Current Month

Explanation of the 4^th  Method:  Moving Average Method

There is need to make use of historical data using moving average. Various steps involved in the process of enumeration of moving average method include the following:

Step 1 refers to calculation of the predicted figure for the month of November of the year 2008 as per the example. Therefore, there is need to include data from November of the year 2007 till the period of October of the year 2008

Step 2 involves selection of a specific period. Fundamentally, the period might be between two months to 6 months

Step 3 indicates towards calculation of moving average of n period, for instance, at the time when the n is equal to 2 , then it is feasible to enumerate number of visits to hospital of the prior month and thereafter divide the same by the value of the n (that is to say 2). Essentially, this is the predicted figure for the month

Step 4 refers to enumeration of forecast error (abbreviated as FE). Calculation of forecasted error involves subtraction of actual of the particular month from the predicted figure of them month.

Step 5 indicates towards calculation of the total sum of the forecasted error (Mertler & Reinhart, 2016). Basically, this can be referred to as the absolute deviation also simply indicated as AD.

Step 6 involves calculation of the mean absolute deviation of the data

Step 7 refers to comparison of the figure on mean absolute deviation and selection of the lowest value of the mean absolute deviation

Step 8 involves prediction or forecast based on the n period of the mean absolute deviation. In case if n (that is the number of period =4), then the average amount of the past 4 months reflects the forecast/prediction) of this month.

Explanation of the 5^th Method:  Exponential Smoothing

This specific method exponential smoothing utilizes the formula hereby mentioned below:

In this, F stands for the forecast for the following month

O stands for the observed figure for the most current period

F_t-1 stands for the predicted or forecasted figure of the current period

SC stands for the smoothing constant

In essence, the smoothing constant needs to be ascertained by means of trial as well as error method. Therefore, the process of utilizing multiple constants of smoothing can aid in lessening the overall error and assist in providing a superior forecast.

Method 3: Average Percent Change

Steps involved in carrying out the process are hereby mentioned below:

Step 1 indicates towards calculation of the forecasted error founded on multiple constants of smoothing.

Step 2 involves calculation of the mean absolute deviation (MAD) and thereafter comparison of the same. In this case, the lowest MAD can be thereafter chosen. Essentially, this might perhaps be the best one.

Step 3 refers to selection of smoothing constant that is necessarily the next higher one.

Step 4 indicates towards fraction that can again be divided and the above procedure can again be continued (Mertler & Reinhart, 2016). 

Step 5 refers to selection of the smoothing constants that provides the lowest Mean Absolute Deviation (MAD)

Step 6 indicates towards forecast for the entire month that is carried out by the formula utilizing the best of the smoothing constant

Forecasting can be considered to be a specific tool in particularly the sector of Healthcare. Fundamentally, this mechanism helps diverse providers of healthcare service in the process of carrying out predictions and undertaking fitting dimensions to minimize risks and handle several demands. Essentially, forecasts in the segment of healthcare can prove to be beneficial at the time when they are able to deliver timely warning that in turn can help in undertaking remedial steps (Manly & Alberto, 2016). Remedial actions can be undertaken for delivering potential demand as well as allocation of resource. In itself, there are different mechanisms are utilized in the process of forecasting health. Comparison was carried out between six different mechanisms of forecasting time series in particularly an outpatient clinic.

The outcome of the data reflected an enhanced forecast with sequencing data in time series with normal day clustering (Mertler & Reinhart, 2016). Diverse categories of analysis of time series models were necessarily utilized for predicting different emergency department visit. It was thereby discovered that time series model can necessarily be utilized for the purpose of predicting arrival in department for paediatrics (Mertler & Reinhart, 2016).

Mead (2017) applied regression founded models of forecasting for predicting arrivals in different emergency department. Essentially, their evaluation reflected the fact that analysis of regression was an important mechanism for handling different needs for forecasting. In essence, this replicated the fact that the data need to be updated to arrive at a superior prediction.

In essence, the world health organization published in their bulletin predicts worldwide shortage of particularly physicians. Therefore, it can be hereby observed that methods of forecasting can be specifically utilized at diverse stages of the entire Healthcare system.

Method 4: Moving Averages

Use of method 1 (Moving Average)

A specific plot is drawn on the total number of individuals at the health care service between the year 2005 and year 2008. It can be hereby observed during the period 2006 to 2008, the total number of individuals at the health care service has necessarily diminished particularly during Jun mainly before showing again an increasing trajectory (Meeker & Escobar, 2014).

Table 1: Reflecting Average Change

Based on the calculation presented above, it can be hereby mentioned that the predicted value for the visits during the period November is equal to addition of average number of visits plus the average of the alteration multiplied by the mid point value. Therefore, the average change mechanism predicts the number of visits dutring November to be equal to 34.3

Use of Method 2: (Confidence Interval)

For 95% confidence interval, both the upper limit as well as lower limit can be calculated as presented below:

Based on the above mentioned method, it can be said that the predicted figure for the total number of visits will range between  20.68 – 47.92

Use of the third Method: Average Change in Percentage

Based on the particular method, it can be said that the forecasted figure for the period of November of the year 2008 would be as follows:

Therefore, based on the calculated figure, it can be hereby said that the forecasted figure for the period of November stands at 40.48

Use of 4^th Method: Moving Average

The data selected for the current study is between November of the year 2007 to October of the the year 2008. The table below presents the calculation of moving average and considers n to lioe between 2 to 6

Based on the calculations presented above, it can be hereby said that the lowest MAD is particularly for n equal to 2. Therefore, the prediction for the period of November of the year 2008 shall necessarily be the average figure of the prior two months.

So, the predicted figure for the period of November of the year 2008 stands at 38.

Use of 5^th Method: Exponential Smoothing’

As the mechanism of exponential smoothing utilizes the mechanism of moving average, therefore data for the period of November of the year 2007 to October of the year 2008 has been hereby selected. In this case, the values of forecast utilizing smoothing constants are necessarily 0.1, 0.3, 0.5 as well as 0.9. Thereafter, the mean abdsolute deviation is calculated. It can be hereby observed that at the time when the value of forecast is equal to 0.7, the lowest MAD stands at (5.17). After that, the next higher value of MAD is equal to 5.23 for the specific forecast value of essentially 0.5. Further, MAD is again enumerated for different values of forecast such as 0.5 and 0.55, 0.6 and 0.66. It can be hereby witnessed that the lowest value of MAD stands at 5.11 for the particular predicted value of 0.6. Thus, the prediction for the period November 2008 is

 If we observe the chart for the past 4 years, an increase in the total number of visits can be observed in the year 2005 as well as the year 2007 from the month October to particularly November. However, the same is said to decline during the year 2006.

Again, if we observe the outcomes acquired from confidence interval prediction, it can be hereby be said that prediction presented a range between 20.68-47.92. Essentially, these ranges can be considered to be very broad ranges.

Again, the average method presented a value that was necessarily very low (Hjorth, 2017). In addition to this, the average change in percentage presented a value that is a bit higher than the total number of visits during the period October in the year 2008.

Also, the values acquired from the value of Moving average that is 38 and exponential smoothing were essentially very close to one another (Marcoulides & Hershberger, 2014). Therefore, any one of the two mechanisms might perhaps be used for the purpose of forecasting in particularly health care services.

References

Hjorth, J. U. (2017). Computer intensive statistical methods: Validation, model selection, and bootstrap. Routledge.

Manly, B. F., & Alberto, J. A. N. (2016). Multivariate statistical methods: a primer. CRC Press.

Marcoulides, G. A., & Hershberger, S. L. (2014). Multivariate statistical methods: A first course. Psychology Press.

Mead, R. (2017). Statistical methods in agriculture and experimental biology. Chapman and Hall/CRC.

Meeker, W. Q., & Escobar, L. A. (2014). Statistical methods for reliability data. John Wiley & Sons.

Mertler, C. A., & Reinhart, R. V. (2016). Advanced and multivariate statistical methods: Practical application and interpretation. Taylor & Francis.