Clinical Transfusion

13. Audit


Audits are an important component of a PBM strategy. In general, an audit involves checking a process, structure of even an outcome itself to ensure that it conforms to an expectation of performance. The audit will reveal whether the process is doing what is desired of it, and in an efficient manner. Audits of transfusion practice can occur both in the blood bank itself and on the hospital wards and operation rooms; audits of the latter are generally performed to ensure that clinicians are following the institutional directives or guidelines for transfusion practice. Simply by starting to look at a process, potential areas for improvement can be found. This text describes important audit topics both inside and outside the transfusion service, and includes some examples that can be adapted for use at your institution.


A simple as it sounds, it is very important for a transfusion service to know exactly how many blood products it issues, as well as the disposition of every unit. In many countries, it is a legal mandate for the transfusion service to be able to account for every unit that it issues, from donor to recipient. Keeping close track of the inventory is essential for the transfusion service to be aware of changes in transfusion practice that could affect the demand on its inventory, unusually high wastage rates on wards, or know exactly which recipients were transfused with blood products. The latter is important to know if a recall of blood products has been issued because of potential problems related to the donor’s health or to a problem with the unit itself. It is also crucial to minimize wastage of blood products on the clinical wards to ensure a stable blood supply.

The crossmatch technique can help with inventory management

Once a transfusion service can account for all of the products that it receives and issues, it can perform several different kinds of internal audits. A novel mechanism for establishing the efficiency of the transfusion service’s crossmatch policy is to calculate the crossmatch:issue (C:I) ratio over a period of time [1].To do this, a transfusion service simply divides the number of RBCs that it crossmatched over, for example, a 1 month period of time by the number of crossmatches that it had performed over the same time period. An ideal ratio is 1. Some crossmatched RBCs can be excluded from this calculation, such as those that are initially issued as uncrossmatched because the crossmatching typically occurs later. Perhaps the most efficient method for crossmatching RBCs is to use the electronic crossmatch for eligible patients (those with a confirmed ABO group, a current valid type and screen sample, and no current or historical antibodies) at the time that the RBCs have been requested to be issued by the clinical team (see Table 1 below). This policy is in contrast to an alternative crossmatching policy, which is to crossmatch the RBC units at the time that the order for the crossmatch is received. At some institutions, the RBCs can be ordered at a time that is distinct from the time at which they are needed for transfusion; effectively these crossmatched but not issued RBCs are on hold in the blood bank for a specific patient, thereby reducing the fluidity of the RBC inventory. The theoretical advantage of this crossmatch policy is that as soon as the clinicians want to transfuse the units, they can be immediately issued without having to take additional time to perform the crossmatch. However, as has been recently demonstrated, when the transfusion service at a large academic hospital switched from their traditional policy of performing the electronic crossmatch at the time that the RBCs were ordered to the time that they were actually requested for transfusion, the C:I ratio improved significantly and the turnaround time (the time from receipt of the order in the blood bank to product issue) for issuing the RBCs increased by only 1 minute [1]. This slight prolongation of the turnaround time is not likely to be clinically significant, and this transfusion service can now optimize the use of their RBC inventory by not having RBC units that might not actually be issued crossmatched on their shelf. Transfusion services that do not routinely use the electronic crossmatch, that is, those that perform the serological crossmatch as their main technique, will likely have a higher C:I ratio than those that perform the electronic crossmatch.

Monitoring the turnaround times for the issuing of all products is another internal audits that can be performed by a transfusion service and comparing their times to the hospital’s benchmarks that are established based on the clinical urgency of the need for the RBCs, controlling the number of units that expire on their shelves by returning them to the blood supplier or sending them to another hospital that will use them immediately (or changing the size of their inventory to better coincide with the number of products issued), as well as keeping track of the complaints that are lodged against the blood bank and the actions taken to resolve them.



Table 1: Comparison of serological and electronic crossmatch technique


Serological crossmatch

Electronic crossmatch

Suitable for all recipients?


No, see text

Time required to perform

>30 minutes (if antihuman globulin phase used)

<5 minutes

Special equipment required


Yes (validated computer system)

Facilitates rapid issuing of RBCs?

No (even immediate spin crossmatch requires 10 minutes)



Audits outside of the transfusion service are an essential part of PBM, and can help to focus novel PBM initiatives on wasteful practice. A simple audit that can be performed under the auspice of the hospital’s transfusion committee is to regularly monitor the crossmatch:transfusion (C:T) ratio of individual blood product prescribers or entire services. Just as the C:I ratio measures the transfusion service’s efficiency of crossmatching RBCs, the C:T ratio is a measure of the clinician’s acumen in using the RBCs that they had crossmatched. Again, the ideal ratio is 1 as this would represent a clinician who uses every unit that they had ordered to be crossmatched. However, as clinical situations can change rapidly, it is not always realistic to expect that a clinician will use every RBC unit that was crossmatched, so a C:T ratio of 2:1 is generally accepted as the standard. Ratios significantly in excess of 2:1 reflect a clinical service that overestimates the number of RBC units that they will ultimately transfuse, and if the transfusion service is crossmatching the RBCs at the time that the crossmatch order is placed then this can cause significant inventory issues in the blood bank and reduce the transfusion service’s flexibility with their units.

Computerized physician order entry as an audit tool

Thus far, this section has described audits that have been performed at the level of a hospital or clinical service. Audits can also be performed at the level of the individual clinician or blood product prescriber. In order to be able to be this specific, the transfusion service or hospital’s information technology must be able to identify specific prescribers of blood products. One way of obtaining this information is by employing a computerized physician order entry (CPOE) system whereby the blood product prescriber must personally log in to the computerized blood product order entry system before they can order the desired component [3-5]. Using the information gleaned from a CPOE, the hospital system is able to provide specific feedback to their clinicians about their blood product use in relation to that of their colleagues who perform the same procedures. Using what is termed a “bubble report”, which plots for each surgeon the percent of their patients who receive at least 1 unit of RBCs in the x-axis and on the y-axis the average number of RBCs that are transfused to the patients who receive even 1 transfusion, the surgeons can see how their transfusion practice compares to that of their colleagues who perform the same procedures [6]. Bubble reports are provided to each individual surgeon on a monthly basis, and they are anonymous with the exception of the bubble that is specific for the individual surgeon. A fully identified report is also provided to the chief financial officer at each hospital so that they can be kept up to date on the transfusion practice that is occurring at their site. These reports are very useful in identifying outlying practices so that specific interventions can be initiated at the level of an entire service or with specific providers. At the moment at a large hospital system in the USA these reports are being generated for surgeons performing hip and knee replacements, spine and cardiac surgeries, as well as obstetricians who perform hysterectomies. In theory these reports could be generated for any service that uses blood products.

Audits can be carried out on virtually any process in which the blood bank engages

both inside and outside of its walls. Audits can require extensive use of information technology in terms of generating the required data, mining it, and producing an output that can be used to determine if the process is functioning optimally or if it needs to be adjusted. However, even if the technological resources are not available to conduct the types of audits described above, much simpler audits can be carried out in the blood bank by manually measuring inventory levels, turn around times for product orders or the completion of pre-transfusion testing, the ideal time to perform the RBC crossmatch etc. Simply by starting to look at a process, potential areas for improvement can be found.



The obstetric service suddenly began ordering significantly more crossmatches than they had previously been ordering

which was noted at the author’s institution. This increased crossmatch workload placed a burden on the transfusion service’s staff and increased the hospital’s blood charges. The transfusion service, through the hospital’s transfusion committee, discovered that a new guideline that dictated when antepartum patients should have a type and screen performed and when RBCs should be crossmatched had been implemented by the nursing staff in the labour and delivery ward. The transfusion service then conducted an audit to determine if these newly implemented guidelines were actually accurate in predicting the transfusion needs of these patients, and it turned out that many of the patients who had some form of pre-transfusion testing ordered because of the implementation of these guidelines did not require a transfusion at all! Only 2% of the women on which the guidelines predicted a moderate risk of experiencing a postpartum hemorrhage actually required at least 1 RBC, meaning that over 2400 type and screens were unnecessarily performed over a 1 year period amongst the women in this risk category [7]. Amongst the women in the high risk category during the same time period, where the guidelines suggested that not only should a type and screen be ordered but also some RBCs should be crossmatched, 7.3% of the women required 1 RBC unit leading to the transfusion service unnecessarily having performed a significant amount of pre-transfusion testing on nearly 400 additional women [7]. This audit was essential in developing new antenatal pre-transfusion testing guidelines that better reflected the local practice.  This effort resulted in the changing of the pre-transfusion testing ordering habits of the obstetricians thereby relieving the workload burden on the transfusion service, reducing unnecessary testing for the patients, and lowering healthcare costs for the hospital system.

A 2nd 8.5 year long year long audit of the obstetrical service focused on intraoperative cell salvage

(ICS) in obstetrical cases [8]. This audit discovered that the vast majority of patients on which ICS was employed, 79%, did not actually receive a reinfusion of the salvaged blood because these patients did not shed enough blood during their case for ICS processing and reinfusion to occur. Cases where ICS is set up but no blood is reinfused due to the low volume of shed blood are known as stand by cases, and there is a fixed cost to these cases in terms of the hardware and disposables that are needed to collect whatever blood is shed, as well as the costs of the technologists to operate the machine. This audit revealed that the most efficacious use of ICS was in patients who underwent a Cesarean hysterectomy or who had some form of postpartum hemorrhage [8]. The results of this audit have also helped to shape the practice of obstetrics at this hospital.

Mark Yazer

Mark Yazer


United States of America


The content of this resource has been developed and reviewed by members of the ISBT Clinical Transfusion Working Party and should be used at the discretion of healthcare professionals utilising this clinical resource. The authors or the International Society of Blood Transfusion cannot accept legal responsibility for the content of this resource.