Blood Inventory Management: Hospital Best Practice

Sebastian H.W. Stanger, Nicola Yates, Richard Wilding, and Sue Cotton

Blood is a perishable product, and hence good manage- sion laboratory managers in the selected hospitals were
ment of inventories is crucial. Blood inventory manage- interviewed in 7 case studies with the aim of identifying
ment is a trade-off between shortage and wastage. The drivers for low wastage and good inventory manage-
challenge is to keep enough stock to ensure a 100% ment practice. The findings from the case studies were
supply of blood while keeping time expiry losses at a compared with the literature. The extant literature
minimum. This article focuses on inventory management asserts that the drivers for good inventory performance
of red blood cells in hospital transfusion laboratories to are the use of complex inventory models and algorithms.
derive principles of best practice and makes recommen- This study has found this not to be the case. Instead,
dations that will ensure losses due to time expiry are good performance is driven by the quality of transfusion
kept to a minimum. The literature was reviewed to laboratory staff, who must be skilled, regularly trained,
identify available models for perishable inventory man- and experienced. Electronic crossmatching, transpar-
agement. Historical data from the UK blood supply chain ency of the inventory, and simple management proce-
was analyzed to identify hospitals with good inventory dures also facilitate good performance.
management practice and low wastage levels. Transfu- © 2012 Elsevier Inc. All rights reserved.

LOOD COMPONENTS REMAIN a scarce used to control precious resources such as blood
B and precious resource [1-4], this is despite the
fact that over 1.9 million units of blood are donated
[10]. Wasting a unit of blood is also a waste of the
donors' time, effort, and contribution which in the
in England and North Wales every year, 0.9 million United States, Scotland, England, and Wales, for
in Canada [5], and between 4.3 and 4.9 million example, is made on an entirely voluntary basis [7].
whole blood donations have been collected in In a study carried out by the World Health
Germany annually over the last 10 years. In the organization (WHO), of the 124 countries investi-
United States, buying blood accounts for about 1% gated, only 49 (39.5) had reached 100% unpaid
of total hospital spend, as blood is utilized in many voluntary blood donation [11]. Wastage can occur
procedures [6]. The nature of the blood supply at many points across the blood supply chain [8,12];
chains established across the world is not consistent. however, for example, in the United Kingdom and
They differ in the structure of hospitals (private vs Germany, in recent years, wastage in hospitals has
state), the type of supply (free vs paid donations), been significantly higher than wastage in blood
pricing for blood, the distribution of blood, and the centers as shown in Figure 1.
handling of shortages [7]. All systems, however, Inventory management and distribution of blood
have the same objectives, to provide sufficient are seen as major components of the cost for
supply while keeping wastage to a minimum. blood [13]. Hence, both efficient management of
Using blood units before they time expire allows blood inventories and logistics can contribute to a
them to be used for treatments and hence reduces reduction in the overall cost of blood. Various
unnecessary costs. Improving blood inventory approaches have been proposed to reduce the use of
management practice reduces losses due to time blood, and hence its cost; these have been discussed
expiry and facilitates the efficient use of blood. in the wider literature along with a number of
Maintaining the critical balance between shortage studies looking at the storage of blood and its
and wastage is the key to good blood stocks impact on supply [14-16]. Therefore, this research
management. Good inventory management perfor- article will focus on a review of inventory practice
mance entails carrying enough stock to guarantee
100% availability while at the same time minimiz- From the Friedrich-Alexander-University, Erlangen-Nuremberg,
ing time expiry [8,9]. Failure to meet these 2 Germany; Cranfield School of Management, UK, and NHS BT, UK.
objectives can result in a hospital transfusion Address reprint requests to Sebastian H. W. Stanger, Friedrich-
laboratory not being able to meet clinical demand. Alexander-University Erlangen-Nuremberg, Lange Gasse 20, 90403
While excess stock will lead to unnecessary costs Nuremberg. Germany.
E-mail: [email protected]
for the health-care system; in hospitals across the 0887-7963/$ - see front matter
world, economic considerations are becoming © 2012 Elsevier Inc. All rights reserved.
increasingly important and are therefore being doi:10.1016/j.tmrv.2011.09.001

Transfusion Medicine Reviews, Vol 26, No 2 (April), 2012: pp 153-163 153

154 STANGER ET AL

Fig 1. Comparison of annual wastage of red blood cell units in the United Kingdom and Germany. (Data from the Blood Stocks
Management Scheme [12] and the Paul Erlich Institute. [4])

in hospitals. This article has 2 key objectives, to periods of activity, the 1970s and the 2000s. In
review the available literature on generic perishable 1973, in the first article of note, Jennings [20]
inventory management and in doing so evaluate described the fundamentals of how the blood
potential models that can be applied to the blood supply chain operates and identified the 3 key
supply chain. Second, the article will identify the measures of performance: shortage, outdating/
key drivers for good blood inventory performance wastage, and cost of information and transportation.
by investigating practice in hospitals that have Brodheim et al [21] went on to develop an
minimized their wastage and compare these with inventory model based on the average age and
the findings from the literature. average wastage of blood units using a Markov
chain approach. Cumming et al [22] subsequently
LITERATURE REVIEW developed a planning model for the collection of
Literature on inventory management within the donations and a basic model for issuing units to
blood supply chain is rather limited. However, hospitals. Prastacos and Brodheim [23,24] pub-
blood is a perishable and deteriorating product, and lished a prototype computer-based regional distri-
therefore, more general perishable inventory theory bution model implemented in the United States
can potentially be applied to its management. called Programmed Blood Distribution System
Techniques used in industrial environments, for (PBDS). Taking a different approach, Cohen and
example, just in time, are not suitable for the blood Pierskalla [25] developed target stock levels for
supply system due to the consequences of an hospital blood banks derived from data provided by
inventory shortage [17]. The majority of the extant 1 US hospital and 1 blood center. In a later article,
literature is specifically related to the development the last of this period of research activity, Prastacos
of inventory theory within the blood supply chain [9] reviewed the literature looking at various
management context and has concentrated on the models from an operations research point of view.
development of a variety of increasingly complex There is then a gap of nearly 20 years before
inventory models based on a range of analytical and Owens et al [26] analyzed the impact of the average
simulation techniques. This has been done on the age of blood units on the inventory performance and
premise that, by improving the complexity and, by found that the average age varied from blood group
extension, the accuracy of the inventory models to blood group. They concluded that an extension of
used to manage the blood supply chain, wastage shelf life had the potential to yield significant reduc-
can be reduced. tions in wastage. Hence, inventory management is
Research in blood inventory management dates not the only factor that impacts on the wastage of
back to the 1960s [18,19]; there have been 2 major blood components. In another study from a supply
BEST PRACTICE INVENTORY MANAGEMENT 155

chain perspective, Spens [27] identified collabora- literature. As has been mentioned previously, this
tion and “doing things together” as a driver for good literature is rather limited; however, as a perishable
performance in the blood supply chain. and deteriorating product, blood represents a subset
The establishment of the Blood Stocks Manage- of the broader field of perishable inventory
ment Scheme (BSMS) in the United Kingdom and management literature and therefore should be
the instigation of a large database monitoring stock considered in this context. Table 1 summarizes
levels and wastage rates in hospitals and blood the major models identified in a review of the
centers has opened new possibilities in blood more general perishable inventory literature. In
inventory management research [28]. The avail- each case, it provides a summary of the model
ability of these new data has led to greatly improved together with the model pros and cons. Pros are
transparency and consequently an increased under- indicated with the + symbol and cons are listed with
standing of blood inventory management and the – symbol.
improved visibility of the blood supply chain. It is found that, due to the added complexity of
This has been coupled with improvements in com- the presence of both assigned and unassigned
puter technology, which have led to the availability inventory, most of the general models from
of simulation tools with sufficient power to build perishable inventory theory discussed in Table 1
meaningful models of this extremely complicated are not applicable. In summary, the research to
process. Various approaches to employ simulation date in blood inventory is dominated by opera-
techniques (such as discrete event simulation and tions research specialists who develop mathema-
level crossing techniques) have been carried out, tical models and use them to derive policies
concluding that simulation can help decision [23,25,31,39,50,52]. This academic work creates
makers to make less risky decisions regarding the impression that the adoption of more advanced
changes in the supply chain [29-31]. A recent study and complex inventory management models leads
by Perera et al [32] based on a survey of 265 UK directly to improved performance within the
hospitals identified a number of additional factors blood supply chain by determining optimal order
for efficient inventory management. The research quantities and additionally or alternatively optimi-
showed that reducing the reservation period for zing distribution.
reserved units (assigned inventory) leads to lower
inventory levels and reduced wastage. In addition, METHODOLOGY
computer-assisted ordering processes and training The objective of this exploratory research is to
programs were identified as having a significant identify how good inventory practice can support a
impact on the amount of stock held. reduction in wastage of red blood cells (RBC) units
It is important to note that all of the models in hospitals and how it compares with the findings
and approaches identified in the academic litera- from the literature. Data covering 277 hospitals in
ture have limitations and weaknesses, as, by their England and North Wales was provided by BSMS
very nature, all models are based on assumptions [28] in 2009. These data were analyzed to identify
and require defined inputs. For example, various hospitals with low wastage levels of RBC. The
inventory models consider the so-called costs of database used for analysis contains information
lost sales or shortage costs in case of a stock out; relating to inventory management such as
however, due to the fact that it is hard to obtain
these costs, simple ordering policies are often • hospital type and profile
applied [33]. Mattsson [34] argued that simplifying • stock levels in hospitals and blood centers
assumptions made in inventory control models • age distributions and remaining shelf lives of
have a negative effect on their validity, and hence, red cell units
systems built on such models cannot be expected • issues to hospitals
to perform as well as anticipated. Småros et al • wastage in hospitals and blood centers (num-
[35] confirm this by admitting that imposing strict ber of units and reason)
assumptions leads to distorted results; therefore, the
models do not depict reality and cannot be applied. A performance indicator, “wastage as percentage
This section has summarized the findings from of issues” (WAPI) [53], was used to compare and
the review of the blood inventory management rank hospitals. WAPI is a metric compiled of
156 STANGER ET AL

Table 1. Summary and Evaluation of Models Identified in Perishable Inventory Literature

Author Summary Pro (+)/Con (−)

Early research and reviews

Cumming Imbalances in demand and supply − Focus on collection
et al [22]
Brodheim Inventory model based on average age and average waste − Assumption: shelf life 21 days
et al [21] using Markov chain approach
Prastacos and Prototype blood distribution system based on simulation + Findings derived from simulation
Brodheim including a set of SOPs. − Outdated model, more refined models available
[23] + Implementation in New York verified findings
Cohen and FIFO system delivers lowest wastage. Model based on + Seminal work
Pierskalla EOQ theory trading off wastage against shortage − Limited sample size due to limited computing capacity
[25] + Results can be statistically verified
Kendall and Goal programming model proving that stock rotation + Seminal work
Lee [36,37] reduces the wastage + First work focusing on stock sharing and redistribution
Nahmias [38] Review of perishable inventory theory + Proves that FIFO is very hard to achieve in reality
Classifies models into fixed and random lifetime + Approximations are more fruitful than exact
calculations
Sirelson and Review of different policies for platelet distribution + Uses simulation to evaluate effects of weekend
Brodheim Developed a model using regression to identify the deliveries
[39] optimum order quantities
Raafat [40] Review of inventory models for production and − Blood components classified as fixed lifetime
replenishment policies + Reviews based on simple and applicable models
Categorized products into fixed lifetime and decaying
products
Goyal and Review on inventory models for perishable goods + Seminal work
Giri [41] Categorized goods into random and fixed lifetime − Blood inventory management not mentioned
Available models
Wu et al [42], News vendor model + Some similarities with blood supply
Jammernegg − Based on profit maximization not service quality
and Kischka
− Specialized models are too complex with too many
[43], Tekin
et al [44] assumptions
Kopach Revisited queuing models and used level crossing + Tracks only oldest unit
et al [31] techniques to identify optimal quantity − Complex model
+ Splits urgent and nonurgent demand
+ Very good approach
Lodree and Reviewed the impact of deterioration on demand and the + Shows weakness of FIFO models
Uzochukwu impact of customers on FIFO
[45]
van Donselaar Compared different approaches with how supermarkets + Many similarities to blood supply chain
et al [46] manage perishable goods using automated store ordering + Main strength is the simplicity
systems
Goh et al [47] Model focusing on batch sizes for donations and the − Only very limited focus on hospitals and hospital
impact on inventory levels and wastage inventories
Lin et al [48] Model with time-varying demand, deterioration, equal + Many similarities with blood supply chain
review periods, and a fixed planning horizon − Allows backlogs
Hariga [49] Inventory policy that identifies optimal replenishment − Model is too rigid and does not allow ad hoc and
schedules emergency deliveries
− Assumes same shelf life for all products
van Dijk Target stock level model for platelet production based on + Verification of findings
et al [50] real data from a Dutch blood bank + Reduction of time expiry by 15% to 20%
Broekmeulen Inventory policy based on target stock levels with a fixed + Assumptions meet requirements of blood inventory
and van review period including demand profiles in weekly patterns management
Donselaar + Simplicity of the model
[51]
BEST PRACTICE INVENTORY MANAGEMENT 157

Table 2. Types of Wastage and Wastage Distribution in Hospitals Table 3. Nationwide Baseline WAPI Values for all hospitals
in 2008
Year
Contribution
2005 2006 2007 2008
Code/ to Total
Abbreviation Explanation Wastage (%) Nationwide WAPI (%) 2.33 2.06 1.98 2.27
TIMEX Time expired: Units exceeding their 70 This information is used for the evaluation of hospitals.
shelf life have to be discarded and are
recorded with the TIMEX code.
OTCOL Out of temperature control–outside 22
laboratory: Blood units being exposed factors will affect the processes in the hospital
to ambient for more than 30 min transfusion laboratory.
cannot be returned to the main stock To understand what good performance means,
refrigerators due to regulatory issues
the system-wide WAPI for all of the hospitals was
in the United Kingdom and have to be
discarded.
calculated on an annual basis for the last 4 years to
MISC Miscellaneous wastage is wastage 6 establish a base line measure for comparison, this is
not classified in the other categories, shown in Table 3.
such as dropping a unit of blood or Using the baseline WAPI as benchmark, 7 hos-
damaging the packaging, which
pitals with exceptionally low wastage levels have
results in wastage.
FF Refrigerator failure: In case of a major 2
been selected for the case studies. Table 4 shows
failure of the refrigeration, the content the WAPI for the 7 hospitals.
of the refrigerator has to be discarded. Once the best performing hospitals were identi-
fied the transfusion laboratory managers of these
different types of wastage (time expiry, out of hospitals were interviewed to identify what inven-
temperature control, refrigerator failure, and mis- tory policies and practices were utilized.
cellaneous) and is calculated as follows [53]:
WAPI ðin %Þ Interview Methodology
sum of wasted blood units for time t Transfusion laboratory managers in the selected
= × 100
sum of blood units issued to the hospital for time t
hospitals were contacted by the BSMS in May 2009
WAPI shows the percentage of units wasted over to request their participation in the study. The
the period analyzed: the lower the ratio, the better purpose and the methodology of the study were
the performance. It provides a comparable metric explained. Once their agreement was obtained, an
independent of the hospital type and size. interview protocol was circulated via email. The
Table 2 shows the different wastage classifica- interview protocol was designed to identify how
tions along with their associated wastage codes. blood inventory is managed. This incorporated the
The table also presents the typical proportion of evaluation of the models, tools, algorithms, and
wastage allocated to each code at an aggregate level approaches applied in the selected hospitals to
for all hospitals. It can be seen from the data that identify drivers for good performance. The open-
time expiry (TIMEX) is the most common cause ended questions focused on the rationale behind
of wastage. decisions about stock levels, replenishment orders,
To prevent bias in the analysis, the hospitals automated inventory management processes, allo-
were clustered by size, and each cluster was cation of blood units to patients, and order patterns.
analyzed independently. This approach will en- A format that focused on open-ended, uncoded
sure that a comprehensive picture of the differ- questions was selected for the interviews, as they
ences and commonalities in stock management in
hospitals of different sizes is created. Hospital
Table 4. Inventory Performance Indicator for the Hospitals
size is an important factor. In larger hospitals, the
Selected for Interview Based on WAPI
order volumes are larger and deliveries are often
Hospital
more frequent. Additionally, larger hospitals tend
1 2 3 4 5 6 7
to have more complex medical and surgical
services and hence differing demands for blood WAPI 0.302 0.471 0.386 0.980 0.262 0.372 0.267
(%)
when compared with smaller hospitals [8]. These
158 STANGER ET AL

allow flexibility regarding the direction of the vation period, and maximum surgical blood
questions and do not exclude important areas by ordering schedules (MSBOS).
restricting or predefining the possible answers [54].
Before circulating the questions, the interview As this research project was exploratory in
protocol was pretested with independent NHS nature, it was important to ensure that the case
staff to ensure that all logistics and transfusion studies generated valid, reliable, and generalizable
terminology was correct and did not lead to results. Once the interviews were completed,
misunderstanding and hence biased results. interview transcripts were reviewed by the research
Seven formal case study interviews took place in team, and key themes and insights were identified.
the first half of July 2009, 4 were undertaken face This was undertaken by clustering the material from
to face and 3 were undertaken by telephone all sources into different categories for each case
conference. The duration of the interviews ranged individually. The questions and subheadings from
between 35 and 150 minutes. Typically, the the interview protocol were used as a starting point
interviews lasted 60 minutes. The interview in building categories. The interview transcripts,
guideline was divided in 4 major sections: information from site visits, and archival data were
then clustered into these categories. While cluster-
1. Stock levels: Processes, models, equipment, ing the information, the categories have been
and tools that are used to capture and define discussed iteratively within the research team to
stock levels. The aim was to identify how get multiple perspectives on the facts preventing
target stock levels are set and adjusted and the biased conclusions. This ensured that the themes
corresponding methods, tools, and monitoring are unbiased and no information was neglected or
processes used. The interviews also discussed underrated or overrated. These themes were then
safety stock policies, safety stock levels in corroborated by independent academic and NHS
operation, and the methods used to set and personnel to ensure their validity.
adjust them. Strategic decisions to not stock
certain blood groups and the reasoning behind FINDINGS—KEY THEMES AND DISCUSSION
such decisions were also explored. The following section details the findings from
2. Replenishment orders: Processes to replenish the case studies.
blood components and how decisions are
made in this context were discussed. This Human Resources and Training
included the methods used to calculate order Human resources and training was identified as a
quantities and how these methods are significant theme in all of the top performing
reviewed and adjusted over time. The ques- hospitals. Human resource was mentioned specif-
tions also investigated order patterns, review ically by 5 of the 7 hospitals reviewed. All hospitals
periods, and the triggers for placing orders. confirmed that educating their transfusion labora-
3. Inventory management principles: The ques- tory staff, thereby increasing the level of awareness
tions focused on standard operating pro- of what good blood inventory management means,
cedures (SOPs), training of staff, and the use is an essential element of good performance. Staff
of special incentives and initiatives aimed at must be aware of the impact of their decisions on
the reduction of wastage in hospitals. This part the whole blood supply chain. The blood supply
of the interview reviewed the process from chain as a system is very complex and dynamic; a
receipt of the unit of blood by the hospital wrong decision can have a serious impact on the
until transfusion to a patient. quality of service and directly affects patients, and
4. Inventory management tools and equipment: hence, awareness of this impact is therefore vital.
Strategic and general questions were asked Regular staff training and briefings together with
about the equipment and techniques used the use of detailed and up to date SOPs is essential.
to manage and monitor the inventories in the An SOP comprises a detailed description of a
hospitals including how units are tracked and process; they ensure that staff execute the process
how remote storage refrigerators are monitored. correctly. The experience of staff also impacts on
5. Allocation of units to patients: The questions performance. One hospital ensures that, when
included crossmatching techniques, the reser- staff rotation is required, only experienced staff
BEST PRACTICE INVENTORY MANAGEMENT 159

are in charge of crucial tasks, such as placing orders Transparency of Inventories and
with the blood service (NHSBT) and the handling Simple Procedures
of deliveries, as the learning curve effect impacts A third driver, essential for good inventory
both on the performance and the quality of performance, is transparency and visibility of
decisions made. inventories. This means that it is crucial to know
A key element confirmed by 6 hospitals is the stock levels in all inventory locations in the
ensuring that staff are aware of the impact of hospital and the status of all RBC units. All
wasting a unit of blood, not only of the financial hospitals stressed that it is vital that inventory
consequences, but also the impact on surgical levels are checked frequently. This inventory
schedules and on patients. This view of the wider includes any remote refrigerators in addition to
implications of wasting a unit of blood, such as the main blood bank inventory location. Benefits
cancellation of treatments, the financial and oper- can only be derived from this information if the
ational impact, and the moral duty toward the information is included in the process of order
donors, was found to be a motivating factor for staff quantity calculation.
in making better decisions. Overall awareness of In contrast with the discussions within the
staff and regular training in handling RBC units academic literature simple “rule-of-thumb” inven-
combined with detailed SOPs was mentioned by all tory procedures were seen as important to success.
hospitals as the key to managing red cell inventory For example, 2 hospitals make use of a so-called
and reducing wastage. “standing order.” This is a fixed order that will be
Stock Levels and Order Patterns fulfilled by the blood service automatically. One
hospital has a fixed standing order in the morning
In direct opposition to what is claimed in the and uses a second routine delivery to adjust the
literature, none of the hospitals surveyed used stock in the afternoon. This reduces complexity and
complex models or equations to readjust target workload in the morning while ensuring blood
stock levels on a frequent basis. All of the supply and, furthermore, gives enough flexibility to
interviewed hospitals were unanimous with regard react to changes in demand with a second order
to the main objective of managing blood inventory: later in the day. However, not all hospitals get 2
to maintain low stock levels to keep time expiry to deliveries per day, and hence, this approach is only
a minimum while ensuring that stock levels are applicable to certain hospitals. To prevent too high
high enough to ensure supply. To achieve this, all a stock level, it is important to prevent panic orders.
of the hospitals have established target stock levels Therefore, inexperienced staff must be aware of all
and maintain them by following an order-up-to of the different stock locations and assigned
policy. These target stock levels have been inventories in the hospital. One hospital reduced
established based on the experience of staff and the physical storage capacity to ensure that nobody
are adjusted continuously over time. In all of the placed panic orders because stock refrigerators
hospitals, the stock levels currently in use have appear to be empty.
evolved over many years of making small incre-
mental changes to achieve optimal levels. However,
levels are not entirely rigid, and 3 of 7 hospitals Focus on Freshness of Stock
adjust them dynamically on a daily basis using their Using a strict oldest-unit-first-out (OUFO)
experience as the rationale behind these changes. policy or first-in-first-out (FIFO) policy was
The sources of information used in the adjustment seen as the most important discipline in reducing
of daily orders are the number of scheduled wastage, and simple procedures are used to
surgeries and recurring top-up transfusions. This support an efficient implementation of these
information, however, is not available in all principles. Therefore, all hospitals sorted their
hospitals in the required detail or data format; inventories in the refrigerators by age (ie,
additionally, different demand profiles may affect remaining shelf life) so that the oldest units are
the availability of the required data. The experience at the front of the shelves and available for easy
of staff plays an important role in the interpretation access. Additionally, in 1 hospital, units close to
of this information to make the right decision for the expiry were highlighted with a red card to enable
ordering process. visual distinction.
160 STANGER ET AL

The FIFO principle has been applied at various hence wastage. This is in line with the findings of
points within the supply chain to good effect. Two Spens [27].
hospitals make use of the blood center stock level
data provided by the BSMS. With this information,
they draw conclusions about the age of the units in Electronic Crossmatching
the blood centers and adjust their orders accord- Another efficient tool in reducing the number of
ingly to prevent them from receiving too many old units in the assigned inventory is electronic cross-
units. As blood centers work on a strict FIFO matching [55]. Electronic crossmatching is widely
policy, high stock levels indicate that a blood center used in the larger hospitals, who confirmed that
is currently issuing older blood when compared a reduction in assigned units can be achieved and
with lower stock levels. the number of time expired units could be reduced
All managers split larger orders into a number due to more flexible allocation of units to patients.
of smaller orders with the purpose of ensuring that One hospital uses the benefits of electronic cross-
the units received have a range of different shelf matching to enhance internal collaboration. The
lives. This is done to avoid increased pressure on increased flexibility meant they could offer a
the transfusion laboratory if a large number of service level agreement guaranteeing a maximum
units time expire on the same date. This also lead time for crossmatched blood. This dramatically
reduces the risk of receiving a large delivery of reduced the number of units in the reserved
units having a very short remaining shelf life from inventory and increased trust between the de-
the blood center. partments in the hospital.

Internal Collaboration within the Hospital CONCLUSIONS AND RECOMMENDATIONS

Collaboration with medical and surgical staff, Interviews with blood inventory managers in the
awareness of issues with units reserved for patients, top performing hospitals revealed 6 key themes that
and critical questioning of orders and requests are together drive good performance in blood stocks
all very important factors in enhancing performance inventory management. These themes are human
and were mentioned by 5 hospitals in this study. resources and training, stock levels and order
The size of the current assigned inventory is a key patterns, transparency of inventories, simple inven-
factor in this. If the stock levels in the assigned tory procedures, focus on freshness, and internal
inventory can be reduced, then flexibility in using collaboration within the hospital. All of these
and allocating available units to other patients can themes highlight the importance of having high-
be increased and wastage of valuable shelf life quality, trained, and experienced staff. These staff
avoided. For example, some transfusion laborato- must be aware of the wider implications of wasting
ries adjust internal orders and requests according to a unit of blood, be able to apply simple inventory
their predefined MSBOS, which defines the management techniques, and collaborate with
maximum number of units issued for each standard clinical staff outside of the hospital laboratory.
treatment. The aim of this is to reduce the number of These findings are contrary to what is claimed in the
just-in-case orders by medical staff and thus academic literature, which suggests that it is
attempts to reduce the number of units going into complex policies, models, and techniques that are
the assigned inventory. The MSBOS quantities are the key drivers to low wastage levels in the blood
based on experience of the number of blood units supply chain. The results of the case studies
required for certain surgeries and treatments. The presented in this study reveal that these approaches
main concern of the transfusion laboratory man- are not applied in reality and that before imple-
agers was that most orders and requests are just-in- menting complex processes for the calculation of
case orders. The blood may or may not be required order quantities and target stock levels, wastage can
but is always placed in the assigned inventory be reduced by focusing on other factors.
waiting to be either transfused or returned after the Based on the findings of this exploratory
crossmatch release period with a lower remaining research, the following recommendations for
shelf life. By collaboration with key stakeholders, it improvements in blood inventory management
was possible to reduce the just-in-case orders and are made.
BEST PRACTICE INVENTORY MANAGEMENT 161

Human Resources and Training OUFO policy for issuing refrigerators in hospital
Decisions regarding order quantities and stock blood banks is applied; this makes certain that the
levels should be based on experience rather than oldest units closest to expiry are used first and
complex equations. This study emphasized that keeps the remaining stock as fresh as possible.
experienced and skilled staff are the key to low Organize refrigerators so that units are collated by
wastage. Ensure that staff handling critical process- remaining shelf life and the oldest units are
es, such as ordering and issuing, are trained physically highlighted and placed at the front.
regularly and that fluctuation and rotation rates in
these staff positions are kept low. Table 5. Summary of Key Findings and Recommendations From
the Interviews Carried Out With Transfusion Laboratory Managers,
Stock Levels and Order Patterns Grouped Into the Six Key Themes Identified

Order patterns have a major impact on wastage. Stock levels

• Target stock levels based on experience are adjusted
When large quantities of blood are required, use a
continuously
number of smaller orders rather than a single large • Demand profiling to adjust on daily demand patterns
order. This prevents delivery of large numbers of • Use information about scheduled treatments
units with the same expiry date and hence allows ○ Careful handling of recurring orders
more flexibility in the usage of these units. Monitor • Full transparency of stock levels including remote and issue
refrigerators
stock levels carefully and regularly. Adjust target
○ Frequent monitoring
stock levels when required, incorporating available • Consideration of recurring and planned transfusions
information about scheduled surgeries. To prevent Order patterns
excess inventory, critically review storage capacity, • Avoid panic orders
as large stock refrigerators may appear empty, and ○ Keep storage capacity low
○ Train staff
this could lead to panic orders. In addition, provide
• Split big orders into several small orders to get different shelf-
guidance for inexperienced staff. This could be in lives
the form of visible standard order quantities on the • Make use of standing orders to reduce workload and
stock refrigerators. replenish using “top-up orders”
• Use action levels and predefined order quantities for “out-of-
Transparency of Inventories hours” time to prevent panic orders
Issuing
Transparency of inventory is a key lever for low • Strict OUFO principle
wastage, as better information about the locations • Store units sorted by age and use visual highlighting for units
and quantities of blood components in the hospital close to expiry
• Try to keep assigned inventories as low as possible
leads to more accurate decisions about order
• Question and challenge internal requests for blood to keep
quantities and hence lower wastage. Therefore, assigned inventories low and reduce just-in-case requests
the status and location of all blood units in the • Electronic crossmatching reduces assigned inventories
hospital must be visible to decision makers at all Remote refrigerators
times. This can be achieved either with IT support • Monitor stock levels frequently
• Allow the removal of reserved blood units from remote
or using paper-based lists registering the locations
refrigerators and assigned inventory when a patient is found
of units; attention to detail is vital no matter which for instant transfusion.
method is used. • Check regularly and return units to main storage
Human resources
Simple Inventory Procedures • Train staff and make staff aware of the financial impact of
wasting a unit
Simple processes are highly effective in manag- • Regular training and refreshing courses
ing blood inventories efficiently. For example, a • Motivate staff to keep wastage low
piece of paper, on which the standard order • Ensure that experienced staff are placing orders and handling
quantities are printed, pinned onto issue refrigera- incoming deliveries
Collaboration
tors reduces the chance of making mistakes.
• Motivate and incentivize hospitals to share knowledge with
other hospitals
Focus on Freshness • Reduce mistrust between hospitals
A major lever in reducing wastage is a strict • Reduce mistrust between departments in hospitals
• Use internal service level agreements to generate trust.
focus on the freshness of blood units. Ensure a strict
162 STANGER ET AL

Internal Collaboration Within the Hospital Based on the above, it is clear that further
Improving the collaboration within the hospital research is required to explore the reasons why
has a huge potential for reducing wastage. Moti- the complex models proposed in the literature
vating staff to share information and collaborate are not used in practice, and if used, any further
leads to lower wastage rates and increased flexibil- improvements in wastage could be achieved in
ity in the transfusion laboratory. Clinical staff must the “real world.”
share information related to planned surgeries to
allow the transfusion laboratory to use it in their ACKNOWLEDGMENTS
ordering decisions. Using electronic crossmatching This research would not have been possible
and setting up internal service level agreements without the support of the Blood Stocks Manage-
reduces the number of units reserved crossmatched ment Scheme (BSMS). Detailed data about the UK
in the assigned inventory on a just-in-case basis blood supply chain was made available for research
wasting valuable shelf life. Table 5 summarizes purposes for the first time. Access to these data
the key findings and recommendations. allowed us the carry out this research.

REFERENCES
[1] Goodnough LT, Brecher ME, Kanter MH, AuBuchon JP. [13] Participants of the Cost of Blood Consensus Conference.
Transfusion medicine—Blood transfusion. N Engl J Med The cost of blood: Multidisciplinary consensus conference
1999;340:438-47. for a standard methodology. Transfus Med Rev 2005;19:
[2] Reynolds E, Wickenden C, Oliver A. The impact of 66-78.
improved safety on maintaining a sufficient blood supply. [14] Hess JR, Greenwalt TG. Storage of red blood cells: New
Transfus Clin Biol 2001;8:235-9. approaches. Transfus Med Rev 2002;16:283-95.
[3] Vamvakas EC. Epidemiology of red blood cell utilization. [15] Barshtein G, Manny N, Yedgar S. Circulatory risk in the
Transfus Med Rev 1996;10:44-61. transfusion of red blood cells with impaired flow properties
[4] Provisional report on notifications pursuant to Section 21 induced by storage. Transfus Med Rev 2011;25:24-35.
TFG (German Transfusion Act) for 2010. 2011. Accessed [16] Holovati JL, Hannon JL, Gyongyossy-Issa MIC, Acker JP.
08.08.2011, 2011, at http://www.pei.de/cln_170/ Blood preservation workshop: New and emerging trends in
nn_155724/EN/infos-en/21tfg-en/berichte-21-en/berichte- research and clinical practice. Transfus Med Rev 2009;23:
en/8-21tfg-report-2010-en.html. 25-41.
[5] A Report to Canadians—Canadian Blood Services [17] Chapman JF, Hyam C, Hick R. Blood inventory manage-
2008/20. 2010. Accessed 05.08.2011, 2011, at http:// ment. Vox Sang 2004;87:143-5.
www.bloodservices.ca/CentreApps/Internet/UW_V502_ [18] Millard DW. Industrial inventory models as applied to the
MainEngine.nsf/resources/Annual+Reports/$file/CBS- problem of inventorying whole blood. Ohio State Univer-
Annual-Report-2008-2009-en.pdf. sity Engineering Experiment Station, Bulletin No 180,
[6] Pierskalla WP. Supply chain management of blood banks. Columbus, OH, March 1960; 1960.
In: Sainfort F, Brandeau ML, Pierskalla WP, editors. [19] Silver A, Silver AM. An empirical inventory system for
Operations Research and Health Care: A Handbook of hospital blood banks. Hospitals 1964;38:56-9.
Methods and Applications. Boston (MA): Kluwer Aca- [20] Jennings JB. Blood bank inventory control. Manage Sci
demic; 2005. p. 103-45. 1973;19:637-45.
[7] Rock G, Åkerblom O, Berséus O, et al. The supply of blood [21] Brodheim E, Derman C, Prastacos G. On the evaluation of
products in 10 different systems or countries. Transfus Sci a class of inventory policies for perishable products such as
2000;22:171-82. blood. Manage Sci 1975;21:1320-5.
[8] Cobain TJ. Fresh blood product manufacture, issue, and [22] Cumming PD, Kendall KE, Pegels CC, Seagle JP, Shubsda
use: A chain of diminishing returns? Transfus Med Rev JFA. Collections planning model for regional blood
2004;18:279-92. suppliers: Description and validation. Manage Sci 1976;
[9] Prastacos GP. Blood inventory management: An over- 22:962-71.
view of theory and practice. Manage Sci 1984;30: [23] Prastacos GP, Brodheim E. Computer-based regional blood
777-800. distribution. Comput Oper Res 1979;6:69-77.
[10] Isbister JP. Risk management in transfusion medicine. [24] Prastacos GP, Brodheim E. PBDS: A decision support
Transfus Med Rev 1996;10:183-202. system for regional blood management. Manage Sci
[11] World Blood Donor Day 2006. 2006. Accessed 1980;26:451-63.
08.08.2011, at http://www.who.int/mediacentre/news/ [25] Cohen MA, Pierskalla WP. Target inventory levels for a
releases/2006/pr33/en/index.html. hospital blood bank or a decentralized regional blood
[12] Annual Report 2009-10:Full Report. 2011. Accessed banking system. Transfusion 1979;19:444-54.
15.05.2011, 2011, at http://www.bloodstocks.co.uk/pdf/ [26] Owens W, Tokessy M, Rock G. Age of blood in inventory
BSMS1.pdf. at a large tertiary care hospital. Vox Sang 2001;81:21-3.
BEST PRACTICE INVENTORY MANAGEMENT 163

[27] Spens K. Integration and performance in a blood supply [41] Goyal SK, Giri BC. Recent trends in modeling of
network. In J Integr Supply Manag 2006;2:231-50. deteriorating inventory. Eur J Oper Res 2001;134:1-16.
[28] Chapman JF, Cook R. The Blood Stocks Management [42] Wu J, Li J, Wang S, Cheng TCE. Mean-variance analysis
Scheme, a partnership venture between the National Blood of the newsvendor model with stockout cost. Omega
Service of England and North Wales and participating 2009;37:724-30.
hospitals for maximizing blood supply chain management. [43] Jammernegg W, Kischka P. Risk preferences and robust
Vox Sang 2002;83:239-46. inventory decisions. Int J Prod Econ 2009;118:269-74.
[29] Rytilä JS, Spens KM. Using simulation to increase [44] Tekin E, Gürler Ü, Berk E. Age-based vs. stock level control
efficiency in blood supply chains. Manage Res News 2006; policies for a perishable inventory system. Eur J Oper Res
29:801-19. 2001;134:309-29.
[30] Katsaliaki K. Cost-effective practices in the blood service [45] Lodree EJ, Uzochukwu BM. Production planning for a
sector. Health Policy 2008;86:276-87. deteriorating item with stochastic demand and consumer
[31] Kopach R, BalcIoglu B, Carter M. Tutorial on constructing choice. Int J Prod Econ 2008;116:219-32.
a red blood cell inventory management system with two [46] van Donselaar K, van Woensel T, Broekmeulen R, Fransoo J.
demand rates. Eur J Oper Res 2008;185:1051-9. Inventory control of perishables in supermarkets. Int J Prod
[32] Perera G, Hyam C, Taylor C, Chapman JF. Hospital blood Econ 2006;104:462-72.
inventory practice: The factors affecting stock level and [47] Goh CH, Greenberg BS, Matsuo H. Perishable inventory
wastage. Transfus Med 2009;19:99-104. systems with batch demand and arrivals. Oper Res Lett
[33] Sezen B. Changes in performance under various lengths of 1993;13:1-8.
review periods in a periodic review inventory control [48] Lin C, Tan B, Lee WC. An EOQ model for deteriorating
system with lost sales: A simulation study. Int J Phys items with time-varying demand and shortages. Int J Syst
Distrib Logist Manag 2006;36:360-73. Sci 2000;31:391-400.
[34] Mattsson SA. Inventory control in environments with short [49] Hariga M. Optimal inventory policies for perishable items
lead times. Int J Phys Distrib Logist Manag 2007;37: with time-dependent demand. Int J Prod Econ 1997;50:
115-30. 35-41.
[35] Småros J, Lehtonen JM, Appelqvist P, Holmström J. The [50] van Dijk NM, Haijema R, Van der Wal J, Sibinga CS.
impact of increasing demand visibility on production and Blood platelet production: A novel approach for practical
inventory control efficiency. Int J Phys Distrib Logist optimization. Transfusion 2009;49:411-20.
Manag 2003;33:336-54. [51] Broekmeulen RACM, van Donselaar KH. A heuristic to
[36] Kendall KE, Lee SM. Formulating blood rotation policies manage perishable inventory with batch ordering, positive
with multiple objectives. Manage Sci 1980;26:1145-57. lead-times, and time-varying demand. Comput Oper Res
[37] Kendall KE, Lee SM. Improving perishable product 2009;36:3013-8.
inventory management using goal programming. J Oper [52] Cohen MA, Prastacos GP. Critical number ordering policy
Manag 1980;1:77-84. for LIFO perishable inventory systems. Comput Oper Res
[38] Nahmias S. Perishable inventory theory: A review. Oper 1981;8:185-95.
Res 1982;30:680-708. [53] Chapman J. Unlocking the essentials of effective blood
[39] Sirelson V, Brodheim E. A computer planning model for inventory management. Transfusion 2007;47:190S-6S.
blood platelet production and distribution. Comput [54] Thietart RA. Doing Management Research: A Compre-
Methods Programs Biomed 1991;35:279-91. hensive Guide. London: Sage Publications; 2001.
[40] Raafat F. Survey of literature on continuously deteriorating [55] Arslan Ö. Electronic crossmatching. Transfus Med Rev
inventory models. Oper Res Soc 1991;42:27-37. 2006;20:75-9.