Forecasting for Healthcare - Research Proposal Example

FORECASTING FOR HEALTHCARE By Student’s name Course code and name Professor’s name University name City, State Date of submission Table of Contents Rationale of Study 3 Objectives of Study 4 Literature Review 5 Background 5 Basic Forecasting Methodology 5 Pitfalls of Generic Methodologies in Forecasting for Healthcare 8 Models and Methodologies to be Deployed 9 Examples and Results 12 Conclusion 14 Introduction Forecasting is becoming increasingly important in solving problems facing supply chain management within the healthcare cadre. Apart from medical supply chain management, forecasting has demonstrated usefulness in modelling for operations such as inventory control, personal investments and budgeting in various government agencies. Inasmuch as this process has been utilised over years to solve problems arising in predicting for optimization of service or goods availed, the inconsistency noted raises a lot of concerns as to what is being done to eliminate these worries. This however seems to increase with years as the factors affiliated with this important topic continue getting complex by the day. The strategic approaches applied by various fields need to be investigated with a focus to supply chain management in order to improve prediction of patient’s needs as a means of improving lives. Modelling the right processes in demand management can translate to better capabilities in matching the supplies proactively without unnecessary disruptions. This research proposal is therefore motivated by the need to evaluate the upcoming corporate interfaces that can be utilised by the healthcare supply management sectors across the globe in order to optimize supplies of goods and services through a quantitative method. The results shall go towards improving operations in supply management chains and/ manufacturing outputs for correct budgeting of resources such as capital, labour, and component production among others. Rationale of Study Forecasting in healthcare supply chain management serves to curb the high cost of inventories, shortage of supplies and even deaths. The total average cost involved in producing or acquiring inventories is proportional to the shortage amount. Some of the problems that are posed by this relationship in operations management is the fact that this kind of approach as deployed for piecewise linear functions in advanced mathematics may provide a closer forecast for the supply chains. The buffer between the excess inventory or even lack of it may and the limited planning space require active input from supply chain managements in order to ensure smooth running of affairs. This shall therefore have to deploy forecasting solutions as a means of eliminating waste that emanates from excess inputs. On filling the buffer, the excess quantities may be reserved within a company’s precincts to avoid excess usage of production resources or withholding capital for that matter. The problem also exists when it comes to forecasting for shortages since coming up with optimum quantities or qualities require careful consideration through forecasting procedures. The contemporary inventory control methodologies seem to be uncertain in terms of forecasting for years away since the relationship fades with time. This paper is motivated by the need solve problems related to supply management i.e. high cost of inventories and shortages through adaptation of the most recent forecasting methodologies. Research shall be carried out in recent journals on forecasting with respect to supply chain management in order to come to a consensus of approach. Objectives of Study The main objectives of this study are thus established as follows: i. To establish the most recent forecasting methodologies that are applicable to resolving issues of excess inventories and shortages within the healthcare supply chain. ii. To establish the optimization importance of the forecasting methodologies in (i) above towards improvement of healthcare supply chain management. iii. To carry out modelling as a means of establishing the most viable method towards resolving the issues faced by healthcare supply chain management. Literature Review Background Forecasts are usually carried out in various fields in order to guide decision making. The focus of this proposal is with regard to medical supply chain management meant to utilise the productive resources to an optimal level. Since industries are always producing, is therefore necessary to establish how to ensure that whatever is produced shall be exhausted by the market demand without necessarily straining the sales department efforts. According to Fomby (2008), the supply management of a medical company must be able determine the forecasts properly using the upcoming methodologies in order to coexist well in the competitive environment. This utilises both the estimated output amount and the expected demand at any given time with respect to the market share. Basic Forecasting Methodology In order to demonstrate the basic forecasting methodologies utilised in medical supply chain management, the fixed proportions production function has been utilised widely. In this method of forecasting, denotes the number of output units produced by the medical supplies company, represents capital that is invested in the running of basic business activities and represents labour. The fixed proportion function is therefore constituted as follows; (1) In the equation (1) above, the basic values and simply denote the fixed capital and labour proportions that are meant to produce one unit of output be it a certain type of medicine or herb. This basic forecasting methodology further follows the activity ray in order to be more specific. Considering the isoquant shown below, it is therefore easy to visualise the fixed proportions production function as implied by Fomby (2008). Figure 1: Fixed proportion production function (Fomby). In the above laid isoquant, there is evident that the main properties of basic forecasting involve capital and labour as the main requirements to determine the output levels. The implied fixed proportion production function is implied when the perpendicular isoquants i.e. and are drawn to imply the equation (1) above. For example, when looking forward to producing, the approach that can be used is by basically imposing an input vector of where and. Based on the moving nature of forecasting, if a combination of figures like and are used the results would amazingly be the same. The fact is that the fixed proportions production function gives constant returns to scale which is the basis of medical supply chain management forecasting. However there are some other factors such as the population, number of diseases that are in existence within a given region and the number of competing firms may set in resulting to a need to diminish the output (Fomby). , (2) For medical supplies companies that seek to manage their inventories towards an optimized levels, them the forecast demand shall usually lie within the output with an immediate forecast resulting to and while later forecast shall give gives and . The minimum forecast production for any given time in the future becomes as shown in relation (3) below. (3) The situation described in relation (3) above is meant for absolute supply chain management where the first forecast is made for medical supply by utilizing a fixed production function. Therefore, the input required in terms of resources be it funding or capital/ labour shall be and . The profitability levels or achievement of optimum quantities within a medical supply chain scenario aims at more accurate forecasts therefore requiring greater efforts when determining , and . For multiple inputs such as population or other factors that may be deemed as affecting the profitability of a medical supply chain, function (4) below may be utilised. (4) The observable pitfalls from the old forecasting methodologies include the fact that there are that there is no possibility of determining the future supply chain metrics within the medical fraternity. This is because; each site is purely managed through joint performance which is basically inefficient. It is therefore recommended that since there are such shortcomings in the supply chain management, there be devised means of improving the existing methodologies in order to cater for the problems that are being currently faced. According to Lee and Billington (2012), inadequate customer definition is the largest failure that most existing forecasting methodologies contain. The available evidence in the article “Effective demand forecasting in the health care supply chain” by Callahan et al. (2004), health unlike other forms of businesses cannot be considered as a retail. Although the methodologies that have been applied over time reflect the probability of near accuracy, the occurrence of diseases is quiet elusive making it difficulty to forecast for medical services in an effective manner (Fomby). Pitfalls of Generic Methodologies in Forecasting for Healthcare The lack of predictability is common across all medical facilities and their respective supply management departments owing to difficulties in determining the potential diseases outbreaks in their zones. The efforts to automate the forecasting models within this field have also been met by a very serious pitfall categorised as inaccurate delivery status data. The fluctuating nature of disease occurrence has been observed in most areas although this is not the biggest contention. Therefore the biggest debate comes in when the impacts of risks and uncertainties are ignored by the forecasting models. The uncertainties are said to emanate from such sources as the supplier lead time, delivery performances, manufacturing process times, the quality of incoming materials among others. Another uncertainty catergorised within this group is the change in inventory stocking policies which are noit actually taken care of in the simpler methodologies. The forecasting methodologies of hey days have also been found to focus only on internal operations without readily considering the competition and other factors such as inefficiency. In some areas, ignorance has been considered as the most demeaning factors due to inadequate understanding of the operation environment (Sethi, Yan and Zhang). Models and Methodologies to be Deployed Box et al. (2013) in a detailed study meant to demistify the problems faced in forecasting for supply chain management. The trio came up with a very interesting and rather important methodology that utilises time series in order to give more accurate forecasts. Anderson et al. (2011) in a similar study undertaken quoted a very powerful case study that was observed in a Nevada Occupational Health Clinic. The clinic that had been operating at the same site for over 20 years, offered insurance to the patients who were incapacitated during business operations. In order to determine the value of income lost, the clinic utilised a linear forecasting model which has been since improved by the authors in order to come up with a final methodology of advanced time series in order to come up with projections in healthcare. Figure 2: Components of a time series. The time series methodology that is aimed by this study is based on a sequence of observations which vary depending on successive points in time. These observations are made periodically in order to generate a regular interval under which the time series shall be established on. In order to come up with a time series pattern, these series are constructed with the time variable being ploted in the x-axis while the other quantities are plotted on the y-axis. This is to aid in the differentiation in time series regression and cross-sectional regression. Stationary time series are also generated while working with this model in order to exercise the time independent statistical probabilities. This model generates data containing a constant mean and variability over time for series with variate data. A trend pattern is hence generated to exhibit the random fluctuations that are also known for forecasting long term factors that are responsible for a certain behavior. This translates to inclusion of all possible determinants into the forecasting activity in order to draw an accurate conclusion (Box, Jenkins and Reinsel). In case of seasonal diseases, the healthcare supply chain management may identify and analyse multiyear historical data that has been collected in the past despite the technological advancments that seemingly make it difficult for projections. For diseases having cyclical patterns, the procurement departments or manufacturers should use alternating sequences in order to curb uncertainties. This cyclicity may be brought about by such factors as fluctuating climate conditions that actually bring about the long term trends. It is therefore evident that when it comes to time series methodology of forecasting, there is need to have an accurate approach through choice of best method. This is because forecast accuracy differs from method to method depending on the disease causing microorganisms or death patterns in patients that are treated or attended to by physicians over a given period of time (Anderson, Sweeney and Williams). Discrete time series formulated by Box et al. have been employed widely for advanced or collective forecasting taking the fixed interval form and usually are denoted by x(1), x(2),…, x(N). This only brings the essence of sophistication by introducing simple and complex stastical concepts that require more attention when it comes to healthcare. Although this methodology does not dwell on the cause and effect approach as a means of instilling simplicity into this proposal, the accuracy of the latter is very high. The Box Jenkins model utilised for time analysis appears to be stochastic by nature and usually records observations for a given time or a probability density function . For a joint time model with more that two variables, the random variables become , with joint probabilities denoted as . In case of a distributed probability, the Box-Cox transformation can be given by the series shown in (4) below; (4) Where the transformation parameter is represented by . The Box-Cox transformation is meant to ensure that the linear transformations improve forecasts. When it comes to ensuring that clean health care data finds its way to the analysis level, editing which includes modifying of outliers, correcting of evident errors and filling in of any missing information. This helps in defining the trend that makes it easy to predict the long term change for purposes of prediction. The trend is presented in both linear and nonlinear forms which may be deterministic (simple linear) or also stochastic by nature (Chatfield). Examples and Results Since diseases outbreak is a seasonal phenomenon, the methodology that shall be deployed is the seasonal pattern in carrying out the main research. A study shall be sunctioned within the prevailing healthcare record in local hospitals in order to model for the final methodology to be implemented in order to fulfill the objectives that are indicated in the introduction section. The seasonal pattern is a time series although the example illustrated herein shall not dwell on the error calculation but shall indicate the basics of what is supposed to be done in way of fulfilling this studi. To begin with, the seasonal trend can be analysed by multiyear movements using historical data availed by local clinics or hospitals. This example indicates that during cold months of winter it is indeed common trend that cough lozenges and syrup sales will pick – a trend that is common. Table 1: Number of patients with cold attended to. YEAR QUARTER PATIENTS 1 1 125 2 153 3 106 4 88 2 1 118 2 161 3 133 4 102 3 1 138 2 144 3 113 4 80 4 1 109 2 137 3 125 4 109 5 1 130 2 165 3 128 4 96 Considering the number of patients who were attended to in a local clinic for cold occurences, the peak is realized to take place during months with extremely cold temperatures. In not a so surprising trend, the the time series of this plot exhibits a pattern that repeats itself year after the other. This series also shows the seasonal patterns that are going to happen in the next few years for purposes of forecasting within the medical field. The long term trend of sales and supplies for lozenges and syrups meant to cure the cold shall therefore be availed by the supply chains managers in different locations – this might as well guide the medication manufacturers. Figure 3: Graphical presentation of the seasonal phenomenon to be used in forecasting of number of patients with cold who will be attended in a local dispensary. Conclusion The trends in forecasting for healthcare are dynamic depending on the determining factors. While this research proposal is specific on demystifying the erroneous trends that are faced in forecasting for healthcare services and supplies, it emerges that there are new upcoming methodologies that can be of extreme help. It is therefore suggested that the improved version of Box-Cox transformation and time series forecasting be studied as a way of coming up with tailored solutions depending on the problem at hand. Upon successful completion of this study, the following objectives; error resolution through application of modern technologies and establishing importance of modelling through exhaustive modelling. Works Cited Anderson, David, et al. An Introduction to Management Science: Quantitative Approaches to Decision Making. New York: Cengage Learning, 2011. Box, George E. P., Gwilym M. Jenkins and Gregory C. Reinsel. Time Series Analysis: Forecasting and Control. Hoboken, New Jersey: John Wiley & Sons, 2013. Callahan, Timothy J., David R. Guzman and Mark A. Van Sumeren. "Effective Demand Forecasting In the Health Care Supply Chain." 16 July 2004. Revenue Performance. 7 November 2014. . Chatfield, Chris. Time-Seriek Forecasting. London: CRC Press , 2000. Fomby, Thomas B. "Some applications of forecasting." 2008. pdf. Lee, Hau L. and Corey Billington. "Managing Supply Chain Inventory: Pitfalls and Opportunities." 15 April 2012. MITsloan Management Review. 6 November 2014. . Sethi, Suresh Prakash, Houmin Yan and ‎Hanqin Zhang. Inventory and Supply Chain Management with Forecast Updates. New York: Springer Science & Business Media, 2006. Read More