The Development of a Recruiting-Drawing-Inventory Model for a Community Blood Bank System B. M. SMACKEY From the Department of Management, Lehigh University, and the Miller Memorial Blood Center. Bethlehem, Pennsylvania
cent efforts to minimize the amount of blood supplied from paid donors.2A related consequence of such efforts is the potential modification of nonreplacement fee structures. The financial impact of nonreplacement fees on the Miller Memorial Blood Center is significant as indicated by the fact that 50 per cent of its cash receipts is derived from this s o u ~ c e . ~ An intensive study of the Blood Center’s operations was done. The study was to focus on the efficiency of the existing inventory control system and because of the importance of the study, an outside consultant was engaged to work with permanent staff members.
A community blood bank system is a multiorganizational program that is designed to supply the blood needs of a community. Participating in such a program a r e hospitals, a central blood bank, industrial donor groups, the American Red Cross, advisory committees, and the community at large. The underlying determinant of the community’s success or failure with its blood program is the degree of cooperation among the various organizations. Intertwined with organizational considerations are the management problems associated with the operation of a responsive and efficient inventory control system. This paper reports on the development of a system for a community blood bank that is in its third year of operation. The system that has been developed can be operated manually by a part-time clerk. Details of the model include an integration of the donor scheduling function and the inventory control function. Simulated testing of the model has been conducted and full-scale implementation is awaiting the expansion of the known donor base.
As ONE of the approximately 150 community blood banks throughout the United States, the Miller Memorial Blood Center serves the Lehigh Valley of eastern Pennsylvania. The Blood Center is located in Bethlehem, Pennsylvania and processes 15,000 units of blood annually for six participating hospitals. In its third full year of operation, the Blood Center is one of several cooperative programs of the Hospital Central Services, Inc. (HCS), a communitysupported effort that seeks ways to effect economies in the delivery of health care. The Blood Center has an annual operating budget of $500,000, and supplies the blood needs of a population of 500,000. Community blood banks supply 30 per cent of the total domestic blood ~ o l l e c t e d . ~ This figure is likely to increase in light of re-
Project Outline The overall purpose of the project was defined to be the improvement in the Blood Center’s operations from both a service capability and a cost control viewpoint. The consultant and a staff assistant developed the following project outline: 1) documentation of existing systems; 2) analysis of major cost elements; 3) development of integrated scheduling/inventory control system; 4) simulated testing of new system; 5) pilot test; 6) implementation; and 7) final report. A brief description of each step is given below. Documentation of Existing Systems
This step focused on the integrated nature of the Blood Center and the member hospitals. A flow chart was developed for identi-
Received for publication August 10, 1974; accepted October 7, 1974.
287 Transfusion May-June 1975
Volume I5 Nurnbcr 3
288
SMACKEY
fying the important parameters of the blood supply and demand system. Included in the preliminary investigation were all the major tasks of the Blood Center: the recruitment of donors, the drawing-processing-storage of blood, and the transportation of blood units to and from member hospitals. By flow charting the physical movement of blood through the system, the relevant data flow could be identified and analyzed. Analysis of Major Cost Elements
In order to examine the various Blood Center functions from a productivity standpoint, e.g., a cost-volume relationship, the Blood Center’s chart of accounts was separated into fixed and variable costs. By using data from a 12-month period, a unit variable cos’t and a unit total cost was developed with other cost data. A study of donor room and laboratory labor costs was also concluded. Develop men t of Int egrated Scheduling1 Inventory Control System
Although monthly financial statements were available for analyzing the Blood Center’s operations, no performance measures had ever been specified for evaluating operating decisions. Such decisions affected the Blood Center either directly through its internal operations or indirectly through decisions made at the hospital laboratory level. Without such performance measures, any management evaluation of the Blood Center would be highly subjective and would be of limited value for achieving long-term improvements. Therefore, specific performance measures were identified prior to any model development. The model that has been developed is to be operated manually but can be computerized if the need exists. In conjunction with the model development, the internal organization of the Blood Center was examined and the functional responsibilities of the newly created position of Director of Adminis-
Transfusion May-June 1915
tration was expanded to include donor scheduling and inventory control. These functions had been handled by the Technical Director. This change was to enable the Blood Center to improve control of inventory levels since all inventory and donor data processing will be consolidated under one function. Simulated Testing of New System
Using the most current three months data, the recruiting-drawing-inventory model was utilized to simulate the Blood Center’s operations over the corresponding period. Actual transfusions and blood mobile supplies were used as exogenous variables to “drive” the model. Recruitment, drawings, and shipments were scheduled according to decision rules that will be discussed below. In addition to generating potential “benchmarks” for the performance measures, the simulation was useful for training the Director of Administration and an inventory control clerk in operating the model. Pilo t Test
Since the success of the project was highly dependent upon the cooperation of both the hospital laboratories and the donorluser groups, a series of meetings with the Operating Committee and the Medical Advisory Committee was initiated by the Manager of the Blood Center. Both Committees responded favorably and the Blood Center is now in the process of further implemen tation. Implementation
Complete implementation of the model requires an expansion of the Blood Center’s known donor base. Prior to the project, the Technical Director had been scheduling a fixed number of untyped donors daily regardless of the inventory of blood in the hospitals or in the Blood Center. This policy led to frequent panic calls for specific types
Volume IS Number 3
289
RECRUITING-DRAWING-INVENTORY MODEL
of blood as well as a relatively high level of attrition and blood shipped out of the area. In order to operate the model on a typespecific basis, cooperation had to be secured from the American Red Cross and the Industrial Member Groups and the names of previously typed donors of the blood mobile programs were obtained. At the time of the writing of this paper, the Manager of the Blood Center was in the process of obtaining lists of previous donors and was establishing procedures for obtaining first time donor blood types.
Results Donor Room and Laboratory Productivity Because of the relative newness of the Blood Center, staffing levels had been established on an entirely judgmental basis. A staff assistant and the consultant conducted a productivity study of the donor room and laboratory. Data were collected for a 24-day period and the following calculations were made: Donor Room: Total Units Drawn in Month Highest No. of Units Drawn Per Nurse Hour 1102 = - = 245NurseHours 4.5 Actual Hours
=
734
Upper Bound of Improved Productivity = 734 - 245 =
489 Nurse Hours
Laboratory: Total Units of Whole Blood Processed in Month Highest No. of Units Processed Per Technologist Hour - -1447
4.5
= :289
Technologist Hours
Actual Hours
=
693
Upper Bound of Improved Productivity =
=
693 - 289
404 Technologist Hours
Although the upper bound estimates do not take into account scheduling fluctuation, minimum hours for employment, and required hours of Blood Center operation, the figures were indicative of underutilization of personnel. As a result, certain staffing changes were approved that were expected to produce a minimum savings of $15,000 per annum.
Analysis of Major Cost Elements The entire set of HCSC-Blood Center accounts were subdivided into two categories: 1) expense accounts, and 2) revenue and/or contra accounts. The expense accounts were further classified as to whether they represented a fixed or variable cost element. Based on the staffing changes that were approved, direct labor was treated as a variable cost element. An abbreviated version of the actual expense data is shown in Table 1. Additional data were developed to determine the cost impact of processing a unit of whole blood as whole blood versus a unit of whole blood processed as components. The following data are relevant: Incremental Labor Cost Per Component = 2) Incremental Supplies Per Component = Total Incremental Cost Per Component = 3) Average Number of Components Made from a Unit of Whole Blood = Variable Cost Per $18.40 Component = 2.33 Total Variable Cost Per Component = 1)
$ 0.27
1.86 $ 2.13
2.33 7.90 7.90 +2.13 $10.03
Transfusion
S MAC KEY
May-June 1975
Table 1. Analysis of Costs for 7556 Units of Whole Blood Drawn During Firsr Seven Months of 1973 Expense Category
Total Cost
Fixed Portion
Variable Portion
Salarieslwages Expendable supplies Building rental Amortization of start-up All other (42 ”other” categories)
$102,000 44,000 15,000 8,000 48,000
$17,000
$ 85,000 44,000
15,000 8,000 38,000
10,000
$21 7,000
$78.000
$139,000
Totals
-
-
Figures do not include component preparation costs. Total cost per whole blood unit drawn =
$217 000 = $28.72 7556
Variable cost per whole blood unit drawn = $’ 39’000 - $18.40. 7556 ~
Although the incremental cost figures for components can be refined further, the data indicate that the use of packed red blood cells instead of whole blood can have a significant impact on the Blood Center’s operations. System Objectives and Performance Measures In order to establish performance measures relevant to the integrated operations of the Blood Center and the participating hospital laboratories, explicit management objectives had to be set forth: 1) prompt availability of needed units; 2) efficient use of resources; 3) appropriate use of components; and 4) minimum age of unit prior to transfusion. Thus, the Blood Center must attempt to meet multiple objectives which can be in conflict with one another, e.g., 1) with 4) and 2). The problem of multiple objectives has been partially approached by an indifference curve technique which trades off outdating with shortages as a function of desired inventory level.’ This technique defines “shortage” as a request upon unassigned blood inventory that could not be met. Although this definition may have applicability for individual hospital inventory policies, it is not appropriate for the Miller Memorial Blood Center system which has seven
stocking points, i.e., six hospitals and the Blood Center, in close proximity to one another. In addition to the inapplicability of this definition of “shortage,” no records were available for such a measurement. It was therefore decided not to develop or use any measurement related to potential shortages. The complete set of system performance measures are as follows: 1) median of days inventoried before transfusion; 2.1) total cost per transfused unit; 2.2) variable cost per transfused unit; 3) attrited units per transfused unit; and 4) components used per transfused unit. The pertinent management report formats are shown in Table 2. Description of Recruiting-DrawingInventory Model Each day a decision must be made concerning whether or not inventory levels throughout the system are adequate. To make this decision, projected demand (transfusion) and projected supply (current inventory plus scheduled drawings) must be compared. If the net projected inventory position for the entire system is above a specified safety stock level, no additional donors will be scheduled. If the net projected inventory position for the entire system is below a specified safety stock level, addi-
Volume IS Number 3
29 1
RECRUIT I NG- DRA W I NG- I NVENTORY MODEL Table 2. Formats for Management Reports
Components Used per Transfused Unit Component Category
WB PRC PC CRY0 FFP Totals
Units Transfused
445 328 192 61 45 1071
Median of Days Blood on Hand Before Transfusion
Attrition
Per Cent
41.5 30.6 17.9 5.8 4.2 loo
Component Category Units transfused Attrited units in hospital inventory more than 10 days Inventoried more than 10 days and returned with less than 5 days remaining Total hospital attrition Per Cent attrition
tional donors will be scheduled to make up the deficit. Thus, daily inventory levels (at each hospital and the Blood Center) and the number of donors scheduled to be drawn are used in conjunction with a statistically based projection of demand for operating the model. The complete description of the model is contained in the Appendix. Consisting of nothing more than several blackboards (control boards) and work sheets, the model is quite simple and inexpensive to operate. Using the inventory operating rules shown in the Appendix, a half-time clerk should be able to perform the scheduling and dispatching calculations. An illustrative example is shown on the work sheet of Table 3. In the example, the total inventory of group O+ whole blood that should be maintained in the system (S) is 35 and the projected transfusions over the next 14 days (PT) is 60. For the complete set of calculations necessary for whole blood and packed red cells, control boards similar to Table 4 and Table 5 should be used. The stepwise system to be used for each type of blood is as follows: 1) record the combined actual inventory level in the entire blood system (I); 2) sum up the projected 14day drawings (D); 3) add I + D; 4) add latest blood system safety stock level (S) to projected 14-day transfusion (PT); 5) determine the net projected inventory position by
WB
445 43
PRC
Type
WB
PRC
328
0’
11
A+ B+ AB+
7 6 8 4 6 2 6 1
7 6 7 2 6 5 1 1
01 44 9.8
0 11
3.4
A-
BAB-
subtracting (I + D) from (S + PT). If the quantity (S + PT) - (I + D) is positive, it equals Bi,j,the additional number of units to be recruited on the current day “i” for drawing on any future day “j.” If the quality is zero or negative, Bij should be equal to zero. 6) if I is below one half PT, schedule B;j as soon as possible; 7) record actual units transfused; 8) record actual units to be added to inventory, 9) calculate ending inventory and set as beginning inventory (I). The inventory levels can be updated twice a day or more frequently if needed.
Discussion The initial results of the simulation demonstrated the importance of expanding the known donor base to be used by the Blood Center. By not operating the model on a type specific basis, the Blood Center will continue to experience sizeable attrition rates and emergency calls for the more common blood types. The inability to control inventories by blood type can be corrected, and necessary steps for achieving the necessary cooperation should be initiated as soon as possible. In particular, current operating policies related to the blood mobile program should be reexamined in light of the large number of units shipped out of the area-approximately 100 units per
49
42
3.3,4
4. 5
4.4
4
4
3
4
0
6
3
15
16
17
18
19
25
27
36
57
13
7.3
12
4
8
11
14
7
10
60
4
9
67
-, 4. 3
-, 8.4
12
8
64
3. 3,4
63
9
9
7
-, 4
60
62
7
6
4
56
3,5. 4
5
5
4
54
65
7
4
2.3
-. 7.3
8
3
49
34
71
7
2
6
-
-, 5,2,3
5
1
Projected 14-Day Cumulative Total (D)
66
21
Day
Total (Net) Units To Be Drawn On Day
5. 5
Units To Be Recruited On "ith" Day for 8, 9, 10, 11, 12, 13 Days I n Future
60
Projected 14-Day Transfusion (PT)
1
(S)
Required Safety Stock
S
t
+ PT
67
62
59
49
43
34
26
25
20
17
19
24
27
28
31
34
41
Beginning Inventory (1)
~~
5
3
4
8
9
10
6
10
12
12
10
10
7
10
4
4
4
4
3
4
9
7
4
12
9
7
5
4
5
a 7
-
Actual Units Added
7
Actual Units Transfused
.-
Table 3. A n Example o f the Work Sheet Used for Group O', Whole Blood ~~
66
67
62
59
49
43
34
26
23
20
17
19
24
27
28
31
34
Ending Inventory
~~~~~
92
89
95
91
92
91
86
87
91
83
86
88
90
88
87
88
90
D+I
~.
3
6
0
4
3
4
9
8
7
4
12
9
7
5
7
8
7
5
21
Units T o Be Recruited
i4 n0 3
;j
i.
25
=D
5 z;'
2
? i x
z
.-
h, v)
Volume
IS
RECRUITING-DRAWING-INVENTORY
Number 3
29 3
MODEL
Table 4. Example of Control Board To Be Used for Inventory and Scheduling Calculations
O+ Hospital
WB
PRC
WB
AB-
0PRC
-
WB
PRC
A.
1. Latest inventory 2. Four-week moving average of transfer B. F.
Totals 1. Latest inventory 2. S u m of four-week moving average of transfusions 3. s 4. PT 5. D 6. S + PT 7. I + D 8. B . . 1.1
Table 5. Example of Control Board Used for Scheduling Drawings
o+
Day
Units T o Be Recruited
WB
Total Units T o Be Drawn
Projected 14-Day Cumulative Total
Units To Be
Recruited
month based on year-to-date figures through October 1973. Further refinements such as variability of demand can be incorporated in the model after an initial learning period. Acknowledgment Many thanks are extended to the following personnel who assisted in the Miller Memorial Blood Center Project: Mr. Joseph Frank, Mr. John Kern, Mr. Theodore Loosbrock, Manager, Mr. Lionel G. Price, Executive Director, and Mr. Carry Schelb.
' 0 PRC Total Projected Units 14-Day To Be Cumulative Drawn Total
-
AB-
PRC
Totals
References 1.
Jennings, J. B.: An analysis of hospital blood bank whole blood inventory control policies. Transfusion 8:335, 1968.
2.
Mason, C.C.: The Illinois blood labeling act and the questions it raises. Lab. Med. 4:13, 1973.
Miller Memorial Blood Center Cash Flow Statement, October and Year-to-Date, 1973. 4. Summary Report: National Heart and Lung Institute Blood Resource Studies, June 30, 1972, DHEW Publication No. (NIH)73-416, p. 24. 3.
294
SM ACKEY
Appendix
Initial Policy for Inventory Levels Each hospital will reorder (by type by component o r whole blood) so a s to maintain an inventory level equal to seven days of use. 2) The Blood Center will maintain its own inventory level at the average of seven days of use of the five largest hospitals.
I)
Terms and Definitions TA, TB, TC, TD. TE, T F are the four-week (28 days) moving averages of transfusions in hospitals A, B, C, D, E, and F (by type, by component or whole blood). IA, IB, IC, ID, IE, IF are thedaily actual inventory levels in the respective hospitals (by type, by component or whole blood) and IR is the inventory level at the Blood Center.
Transfusion May-June 1975
3) S is the combined inventory level that should be maintained in the entire blood system, i.e.. all hospitals plus the Blood Center. S approximately equals 1.2 x (TA + T B + TC + T D + T E + T F ) 4) I is the combined actual inventory level (daily) in the entire blood system. ' is the projected transfusions over the next 14 5) FT days (two weeks). PT = 2 x (TA + T B + TC + T D + T E + TF) 6) D is the latest total of scheduled drawings (including blood mobiles) over the next 14 days. 7 ) Bi, is the number of units (net) to be recruited on the current day i for drawing on any future day j. Note: D is equal to the sum of previous Bij that have not been drawn yet. Bij = (S + PT) (I + D) or zero if the expression is negative. 8) If I is less than one half PT, reschedule D as soon as possible. 9) If a hospital's actual inventory is more than twice its latest moving average of transfusions, hospital should be asked if some units (oldest) could be transhipped to another hospital.
cent efforts to minimize the amount of blood supplied from paid donors.2A related consequence of such efforts is the potential modification of nonreplacement fee structures. The financial impact of nonreplacement fees on the Miller Memorial Blood Center is significant as indicated by the fact that 50 per cent of its cash receipts is derived from this s o u ~ c e . ~ An intensive study of the Blood Center’s operations was done. The study was to focus on the efficiency of the existing inventory control system and because of the importance of the study, an outside consultant was engaged to work with permanent staff members.
A community blood bank system is a multiorganizational program that is designed to supply the blood needs of a community. Participating in such a program a r e hospitals, a central blood bank, industrial donor groups, the American Red Cross, advisory committees, and the community at large. The underlying determinant of the community’s success or failure with its blood program is the degree of cooperation among the various organizations. Intertwined with organizational considerations are the management problems associated with the operation of a responsive and efficient inventory control system. This paper reports on the development of a system for a community blood bank that is in its third year of operation. The system that has been developed can be operated manually by a part-time clerk. Details of the model include an integration of the donor scheduling function and the inventory control function. Simulated testing of the model has been conducted and full-scale implementation is awaiting the expansion of the known donor base.
As ONE of the approximately 150 community blood banks throughout the United States, the Miller Memorial Blood Center serves the Lehigh Valley of eastern Pennsylvania. The Blood Center is located in Bethlehem, Pennsylvania and processes 15,000 units of blood annually for six participating hospitals. In its third full year of operation, the Blood Center is one of several cooperative programs of the Hospital Central Services, Inc. (HCS), a communitysupported effort that seeks ways to effect economies in the delivery of health care. The Blood Center has an annual operating budget of $500,000, and supplies the blood needs of a population of 500,000. Community blood banks supply 30 per cent of the total domestic blood ~ o l l e c t e d . ~ This figure is likely to increase in light of re-
Project Outline The overall purpose of the project was defined to be the improvement in the Blood Center’s operations from both a service capability and a cost control viewpoint. The consultant and a staff assistant developed the following project outline: 1) documentation of existing systems; 2) analysis of major cost elements; 3) development of integrated scheduling/inventory control system; 4) simulated testing of new system; 5) pilot test; 6) implementation; and 7) final report. A brief description of each step is given below. Documentation of Existing Systems
This step focused on the integrated nature of the Blood Center and the member hospitals. A flow chart was developed for identi-
Received for publication August 10, 1974; accepted October 7, 1974.
287 Transfusion May-June 1975
Volume I5 Nurnbcr 3
288
SMACKEY
fying the important parameters of the blood supply and demand system. Included in the preliminary investigation were all the major tasks of the Blood Center: the recruitment of donors, the drawing-processing-storage of blood, and the transportation of blood units to and from member hospitals. By flow charting the physical movement of blood through the system, the relevant data flow could be identified and analyzed. Analysis of Major Cost Elements
In order to examine the various Blood Center functions from a productivity standpoint, e.g., a cost-volume relationship, the Blood Center’s chart of accounts was separated into fixed and variable costs. By using data from a 12-month period, a unit variable cos’t and a unit total cost was developed with other cost data. A study of donor room and laboratory labor costs was also concluded. Develop men t of Int egrated Scheduling1 Inventory Control System
Although monthly financial statements were available for analyzing the Blood Center’s operations, no performance measures had ever been specified for evaluating operating decisions. Such decisions affected the Blood Center either directly through its internal operations or indirectly through decisions made at the hospital laboratory level. Without such performance measures, any management evaluation of the Blood Center would be highly subjective and would be of limited value for achieving long-term improvements. Therefore, specific performance measures were identified prior to any model development. The model that has been developed is to be operated manually but can be computerized if the need exists. In conjunction with the model development, the internal organization of the Blood Center was examined and the functional responsibilities of the newly created position of Director of Adminis-
Transfusion May-June 1915
tration was expanded to include donor scheduling and inventory control. These functions had been handled by the Technical Director. This change was to enable the Blood Center to improve control of inventory levels since all inventory and donor data processing will be consolidated under one function. Simulated Testing of New System
Using the most current three months data, the recruiting-drawing-inventory model was utilized to simulate the Blood Center’s operations over the corresponding period. Actual transfusions and blood mobile supplies were used as exogenous variables to “drive” the model. Recruitment, drawings, and shipments were scheduled according to decision rules that will be discussed below. In addition to generating potential “benchmarks” for the performance measures, the simulation was useful for training the Director of Administration and an inventory control clerk in operating the model. Pilo t Test
Since the success of the project was highly dependent upon the cooperation of both the hospital laboratories and the donorluser groups, a series of meetings with the Operating Committee and the Medical Advisory Committee was initiated by the Manager of the Blood Center. Both Committees responded favorably and the Blood Center is now in the process of further implemen tation. Implementation
Complete implementation of the model requires an expansion of the Blood Center’s known donor base. Prior to the project, the Technical Director had been scheduling a fixed number of untyped donors daily regardless of the inventory of blood in the hospitals or in the Blood Center. This policy led to frequent panic calls for specific types
Volume IS Number 3
289
RECRUITING-DRAWING-INVENTORY MODEL
of blood as well as a relatively high level of attrition and blood shipped out of the area. In order to operate the model on a typespecific basis, cooperation had to be secured from the American Red Cross and the Industrial Member Groups and the names of previously typed donors of the blood mobile programs were obtained. At the time of the writing of this paper, the Manager of the Blood Center was in the process of obtaining lists of previous donors and was establishing procedures for obtaining first time donor blood types.
Results Donor Room and Laboratory Productivity Because of the relative newness of the Blood Center, staffing levels had been established on an entirely judgmental basis. A staff assistant and the consultant conducted a productivity study of the donor room and laboratory. Data were collected for a 24-day period and the following calculations were made: Donor Room: Total Units Drawn in Month Highest No. of Units Drawn Per Nurse Hour 1102 = - = 245NurseHours 4.5 Actual Hours
=
734
Upper Bound of Improved Productivity = 734 - 245 =
489 Nurse Hours
Laboratory: Total Units of Whole Blood Processed in Month Highest No. of Units Processed Per Technologist Hour - -1447
4.5
= :289
Technologist Hours
Actual Hours
=
693
Upper Bound of Improved Productivity =
=
693 - 289
404 Technologist Hours
Although the upper bound estimates do not take into account scheduling fluctuation, minimum hours for employment, and required hours of Blood Center operation, the figures were indicative of underutilization of personnel. As a result, certain staffing changes were approved that were expected to produce a minimum savings of $15,000 per annum.
Analysis of Major Cost Elements The entire set of HCSC-Blood Center accounts were subdivided into two categories: 1) expense accounts, and 2) revenue and/or contra accounts. The expense accounts were further classified as to whether they represented a fixed or variable cost element. Based on the staffing changes that were approved, direct labor was treated as a variable cost element. An abbreviated version of the actual expense data is shown in Table 1. Additional data were developed to determine the cost impact of processing a unit of whole blood as whole blood versus a unit of whole blood processed as components. The following data are relevant: Incremental Labor Cost Per Component = 2) Incremental Supplies Per Component = Total Incremental Cost Per Component = 3) Average Number of Components Made from a Unit of Whole Blood = Variable Cost Per $18.40 Component = 2.33 Total Variable Cost Per Component = 1)
$ 0.27
1.86 $ 2.13
2.33 7.90 7.90 +2.13 $10.03
Transfusion
S MAC KEY
May-June 1975
Table 1. Analysis of Costs for 7556 Units of Whole Blood Drawn During Firsr Seven Months of 1973 Expense Category
Total Cost
Fixed Portion
Variable Portion
Salarieslwages Expendable supplies Building rental Amortization of start-up All other (42 ”other” categories)
$102,000 44,000 15,000 8,000 48,000
$17,000
$ 85,000 44,000
15,000 8,000 38,000
10,000
$21 7,000
$78.000
$139,000
Totals
-
-
Figures do not include component preparation costs. Total cost per whole blood unit drawn =
$217 000 = $28.72 7556
Variable cost per whole blood unit drawn = $’ 39’000 - $18.40. 7556 ~
Although the incremental cost figures for components can be refined further, the data indicate that the use of packed red blood cells instead of whole blood can have a significant impact on the Blood Center’s operations. System Objectives and Performance Measures In order to establish performance measures relevant to the integrated operations of the Blood Center and the participating hospital laboratories, explicit management objectives had to be set forth: 1) prompt availability of needed units; 2) efficient use of resources; 3) appropriate use of components; and 4) minimum age of unit prior to transfusion. Thus, the Blood Center must attempt to meet multiple objectives which can be in conflict with one another, e.g., 1) with 4) and 2). The problem of multiple objectives has been partially approached by an indifference curve technique which trades off outdating with shortages as a function of desired inventory level.’ This technique defines “shortage” as a request upon unassigned blood inventory that could not be met. Although this definition may have applicability for individual hospital inventory policies, it is not appropriate for the Miller Memorial Blood Center system which has seven
stocking points, i.e., six hospitals and the Blood Center, in close proximity to one another. In addition to the inapplicability of this definition of “shortage,” no records were available for such a measurement. It was therefore decided not to develop or use any measurement related to potential shortages. The complete set of system performance measures are as follows: 1) median of days inventoried before transfusion; 2.1) total cost per transfused unit; 2.2) variable cost per transfused unit; 3) attrited units per transfused unit; and 4) components used per transfused unit. The pertinent management report formats are shown in Table 2. Description of Recruiting-DrawingInventory Model Each day a decision must be made concerning whether or not inventory levels throughout the system are adequate. To make this decision, projected demand (transfusion) and projected supply (current inventory plus scheduled drawings) must be compared. If the net projected inventory position for the entire system is above a specified safety stock level, no additional donors will be scheduled. If the net projected inventory position for the entire system is below a specified safety stock level, addi-
Volume IS Number 3
29 1
RECRUIT I NG- DRA W I NG- I NVENTORY MODEL Table 2. Formats for Management Reports
Components Used per Transfused Unit Component Category
WB PRC PC CRY0 FFP Totals
Units Transfused
445 328 192 61 45 1071
Median of Days Blood on Hand Before Transfusion
Attrition
Per Cent
41.5 30.6 17.9 5.8 4.2 loo
Component Category Units transfused Attrited units in hospital inventory more than 10 days Inventoried more than 10 days and returned with less than 5 days remaining Total hospital attrition Per Cent attrition
tional donors will be scheduled to make up the deficit. Thus, daily inventory levels (at each hospital and the Blood Center) and the number of donors scheduled to be drawn are used in conjunction with a statistically based projection of demand for operating the model. The complete description of the model is contained in the Appendix. Consisting of nothing more than several blackboards (control boards) and work sheets, the model is quite simple and inexpensive to operate. Using the inventory operating rules shown in the Appendix, a half-time clerk should be able to perform the scheduling and dispatching calculations. An illustrative example is shown on the work sheet of Table 3. In the example, the total inventory of group O+ whole blood that should be maintained in the system (S) is 35 and the projected transfusions over the next 14 days (PT) is 60. For the complete set of calculations necessary for whole blood and packed red cells, control boards similar to Table 4 and Table 5 should be used. The stepwise system to be used for each type of blood is as follows: 1) record the combined actual inventory level in the entire blood system (I); 2) sum up the projected 14day drawings (D); 3) add I + D; 4) add latest blood system safety stock level (S) to projected 14-day transfusion (PT); 5) determine the net projected inventory position by
WB
445 43
PRC
Type
WB
PRC
328
0’
11
A+ B+ AB+
7 6 8 4 6 2 6 1
7 6 7 2 6 5 1 1
01 44 9.8
0 11
3.4
A-
BAB-
subtracting (I + D) from (S + PT). If the quantity (S + PT) - (I + D) is positive, it equals Bi,j,the additional number of units to be recruited on the current day “i” for drawing on any future day “j.” If the quality is zero or negative, Bij should be equal to zero. 6) if I is below one half PT, schedule B;j as soon as possible; 7) record actual units transfused; 8) record actual units to be added to inventory, 9) calculate ending inventory and set as beginning inventory (I). The inventory levels can be updated twice a day or more frequently if needed.
Discussion The initial results of the simulation demonstrated the importance of expanding the known donor base to be used by the Blood Center. By not operating the model on a type specific basis, the Blood Center will continue to experience sizeable attrition rates and emergency calls for the more common blood types. The inability to control inventories by blood type can be corrected, and necessary steps for achieving the necessary cooperation should be initiated as soon as possible. In particular, current operating policies related to the blood mobile program should be reexamined in light of the large number of units shipped out of the area-approximately 100 units per
49
42
3.3,4
4. 5
4.4
4
4
3
4
0
6
3
15
16
17
18
19
25
27
36
57
13
7.3
12
4
8
11
14
7
10
60
4
9
67
-, 4. 3
-, 8.4
12
8
64
3. 3,4
63
9
9
7
-, 4
60
62
7
6
4
56
3,5. 4
5
5
4
54
65
7
4
2.3
-. 7.3
8
3
49
34
71
7
2
6
-
-, 5,2,3
5
1
Projected 14-Day Cumulative Total (D)
66
21
Day
Total (Net) Units To Be Drawn On Day
5. 5
Units To Be Recruited On "ith" Day for 8, 9, 10, 11, 12, 13 Days I n Future
60
Projected 14-Day Transfusion (PT)
1
(S)
Required Safety Stock
S
t
+ PT
67
62
59
49
43
34
26
25
20
17
19
24
27
28
31
34
41
Beginning Inventory (1)
~~
5
3
4
8
9
10
6
10
12
12
10
10
7
10
4
4
4
4
3
4
9
7
4
12
9
7
5
4
5
a 7
-
Actual Units Added
7
Actual Units Transfused
.-
Table 3. A n Example o f the Work Sheet Used for Group O', Whole Blood ~~
66
67
62
59
49
43
34
26
23
20
17
19
24
27
28
31
34
Ending Inventory
~~~~~
92
89
95
91
92
91
86
87
91
83
86
88
90
88
87
88
90
D+I
~.
3
6
0
4
3
4
9
8
7
4
12
9
7
5
7
8
7
5
21
Units T o Be Recruited
i4 n0 3
;j
i.
25
=D
5 z;'
2
? i x
z
.-
h, v)
Volume
IS
RECRUITING-DRAWING-INVENTORY
Number 3
29 3
MODEL
Table 4. Example of Control Board To Be Used for Inventory and Scheduling Calculations
O+ Hospital
WB
PRC
WB
AB-
0PRC
-
WB
PRC
A.
1. Latest inventory 2. Four-week moving average of transfer B. F.
Totals 1. Latest inventory 2. S u m of four-week moving average of transfusions 3. s 4. PT 5. D 6. S + PT 7. I + D 8. B . . 1.1
Table 5. Example of Control Board Used for Scheduling Drawings
o+
Day
Units T o Be Recruited
WB
Total Units T o Be Drawn
Projected 14-Day Cumulative Total
Units To Be
Recruited
month based on year-to-date figures through October 1973. Further refinements such as variability of demand can be incorporated in the model after an initial learning period. Acknowledgment Many thanks are extended to the following personnel who assisted in the Miller Memorial Blood Center Project: Mr. Joseph Frank, Mr. John Kern, Mr. Theodore Loosbrock, Manager, Mr. Lionel G. Price, Executive Director, and Mr. Carry Schelb.
' 0 PRC Total Projected Units 14-Day To Be Cumulative Drawn Total
-
AB-
PRC
Totals
References 1.
Jennings, J. B.: An analysis of hospital blood bank whole blood inventory control policies. Transfusion 8:335, 1968.
2.
Mason, C.C.: The Illinois blood labeling act and the questions it raises. Lab. Med. 4:13, 1973.
Miller Memorial Blood Center Cash Flow Statement, October and Year-to-Date, 1973. 4. Summary Report: National Heart and Lung Institute Blood Resource Studies, June 30, 1972, DHEW Publication No. (NIH)73-416, p. 24. 3.
294
SM ACKEY
Appendix
Initial Policy for Inventory Levels Each hospital will reorder (by type by component o r whole blood) so a s to maintain an inventory level equal to seven days of use. 2) The Blood Center will maintain its own inventory level at the average of seven days of use of the five largest hospitals.
I)
Terms and Definitions TA, TB, TC, TD. TE, T F are the four-week (28 days) moving averages of transfusions in hospitals A, B, C, D, E, and F (by type, by component or whole blood). IA, IB, IC, ID, IE, IF are thedaily actual inventory levels in the respective hospitals (by type, by component or whole blood) and IR is the inventory level at the Blood Center.
Transfusion May-June 1975
3) S is the combined inventory level that should be maintained in the entire blood system, i.e.. all hospitals plus the Blood Center. S approximately equals 1.2 x (TA + T B + TC + T D + T E + T F ) 4) I is the combined actual inventory level (daily) in the entire blood system. ' is the projected transfusions over the next 14 5) FT days (two weeks). PT = 2 x (TA + T B + TC + T D + T E + TF) 6) D is the latest total of scheduled drawings (including blood mobiles) over the next 14 days. 7 ) Bi, is the number of units (net) to be recruited on the current day i for drawing on any future day j. Note: D is equal to the sum of previous Bij that have not been drawn yet. Bij = (S + PT) (I + D) or zero if the expression is negative. 8) If I is less than one half PT, reschedule D as soon as possible. 9) If a hospital's actual inventory is more than twice its latest moving average of transfusions, hospital should be asked if some units (oldest) could be transhipped to another hospital.
