MLS 3302 BIOSTATISTICS RESEARCH METHODS LABORATORY PRACTICES Unit
MLS 3302 BIOSTATISTICS, RESEARCH METHODS, & LABORATORY PRACTICES Unit 4 – Financial Management Section 5 – Inventory Forecasting Matthew Nicholaou, Dr. PH, MT(ASCP)
Unit 4 – Overview Section 1 � The Section 2 � US Laboratory Management Index Program (LMIP) Healthcare System & Reimbursement Section 3 � Cost / Benefit Analysis Section 4 � Laboratory Budgeting Section 5 � Inventory Forecasting
Unit 4. 5 – Inventory Replenishment FIXED-QUANTITY REPLENISHMENT – the lab always has 10 units of reagent X on hand. When inventory drops below 10, submit a purchase requisition (purchase order) for more. FIXED-PERIOD REPLENISHMENT – weekly, or monthly submit a purchase requisition for more reagent X. MIXED REPLENISHMENT – Weekly or monthly check supply of reagent X if you have 10 units you don’t have to order, if you have less than 10 submit a requisition.
Unit 4. 5 – Inventory Forecasting The most difficult part of managing an inventory is knowing ‘how much’ to order. Many factors need to be considered: � What is the unit cost? � Is there a discount with a bulk purchase? � What does it cost to store? � Is there a restocking fee? � What is the current need (volume)? * � Will your volume increase in the future? * � What is the shelf-life?
Unit 4. 5 – Inventory Forecasting: Tools There a number of methods forecasting inventory, each is suited to particular volume conditions. Moving Average Exponential Smoothing Base Index Linear Regression Polynomial Regression
Unit 4. 5 – Inventory Forecasting
Unit 4. 5 – Inventory Forecasting: Moving Average When the volume of tests remain relatively constant a MOVING AVERAGE is appropriate forecasting inventory MOVING AVERAGE � Is a mean of the last 12 months � If forecasting how much inventory would be needed in January, calculate the mean of tests ordered from last January to December. � The mean is the predicted volume of tests needed in January
Unit 4. 5 – Inventory Forecasting
Unit 4. 5 – Inventory Forecasting: Linear Regression When a constant trend is observed it may be useful to ‘fit’ a line describing this trend using LINEAR REGRESSION � Perform a regression analysis to obtain the equation of the best-fit line � Use the regression line to estimate the next months volume y = mx + b Volume = slope(month) + intercept Note: Months need to be replaced with integers (dummy variables) The month you start with number 1, then 2, and so on…
Unit 4. 5 – Inventory Forecasting
Unit 4. 5 – Inventory Forecasting: Polynomial Regression Sometimes there will be a trend that follows a curvilinear rather than a linear relationship. In this instance a POLYNOMIAL REGRESSION can be used to ‘fit’ a line to predict the next months volume. POLYNOMIAL REGRESSION � Identical to performing a linear regression except variables now enter as polynomials Linear Regression Polynomial Regression y = mx + b y = mx 2 + mx + b Note: You must convert the months to dummy variables the same way you did for the linear regression analysis.
Unit 4. 5 – Inventory Forecasting
Unit 4. 5 – Inventory Forecasting: Exponential Smoothing When there appears to be a recent trend in test volume (an upward trend in the example) an EXPONENTIAL SMOOTHING is appropriate for inventory forecasting. EXPONENTIAL SMOOTHING � The exponential smoothing is a moving average that assigns unique weights to historical experience. With the most current experience receiving the greatest weight � Is more responsive to upward or downward trends than a moving average � Will always lag behind the actual trend
Unit 4. 5 – Inventory Forecasting
Unit 4. 5 – Inventory Forecasting: Base Index When there appears to be a trend in the volume of tests, seasonal trend, a BASE INDEX is appropriate forecasting inventory. BASE INDEX � The Base Index technique will compensate for seasonal trends but requires a massive amount of historical data (need multiple years) � A weight is given to seasonal peaks in volume � This weight will be used to correct for the seasonal average
Unit 4. 5 – Inventory Forecasting There are many approaches to forecasting inventory with ever increasing complexity. In general all suffer from the same weaknesses � Forecasting inventory from too long of a time period will introduce statistical ‘noise’ � Forecasting from too short of a time period will potential miss cyclical trends and potential over/under estimate based on recent aberrant trends
Unit 4. 5 – Inventory Forecasting
- Slides: 17