A hematologist, who also co-chairs a hospital Blood Transfusion Improvement Committee in Australia, reports that they are currently working with a group of hospitals to improve the appropriateness of red cell transfusion. If this initial project proves successful, they will eventually move on to other blood components and products. She adds that the National Health and Medical Research Council and the Australasian Society of Blood Transfusion have developed clinical guidelines for use of blood components. These guidelines include a recommendation that the clinical reason for transfusion should be recorded in the patient's medical record (which should be done as part of good patient management). It is believed that such documentation will also assist in auditing safety and appropriateness of transfusion. To that end, she reports that they have found that documentation of the reason for transfusion needs to be improved in all hospitals where they have audited transfusion episodes. Consequently, they have developed a transfusion checklist sticker (PDF file) that they are evaluating. However, auditing transfusions needs to be done manually by retrieving records, which is time-intensive. It is too early in their evaluation process to conclude how well the sticker has been received and if it is useful. The Australian colleague would really appreciate any help that colleagues could offer them on the issue of improving clinician documentation of the reason for a transfusion. She is interested in learning if lack of such documentation is a problem for other colleagues. If it is a problem, how have others tried to solve it, or if it is not a problem, how have others managed to make clinicians document their reasons for transfusion in the medical notes?

The following responses were received.

ADDENDA April 30, 2003

1. A colleague from a Sunbelt State is of the opinion that documentation of transfusions has been a widespread problem, in her opinion and from discussions with other blood bankers. She likes the suggestion for a "transfusion sticker" or even a two-ply physician order form that is used only for transfusions. She adds (verbatim) "It could include the product requested, reason for transfusion, special requirements of the product and all of the other useful and required information. Anything to result in better documentation of transfusions. The form could have boxes where they check/mark what they are wanting. A copy of the order could be sent to the blood bank". In regards to physician penmanship, she comments that "Many times, with computer systems, it is the unit secretary who is trying to interpret the order and many mistakes are made because of this." Finally she is of the opinion that auditing of charts takes a long time and it is often difficult to find the reason for transfusions. Auditors just have to guess at what the reasons for transfusion were to see if it meets criteria set at their hospital. This is certainly an area that needs development and improvement!

2. A colleague in Texas says that at her hospital she firmly believes that her transfusion service does not have the information available to know if a product has been ordered, transfused, or not. She writes (verbatim) "The physician writes the order to transfuse on the chart, nursing service is responsible for getting the consent signed and they have a checklist to follow before picking up the blood from the transfusion service. I am not aware that anyone at our facility has been transfused without orders to do so in the last 10 years." She is of the opinion that their checklist must be working, since they are unaware of problems. (Editor's note: The Texan did not comment on doing any spot checks or audits of charts to document the absence of problems related to transfusions of patients without orders to do so.)

ADDENDA May 1, 2003

3. A colleague at a hospital in Oakland, California reports that they use the Meditech Hospital Information System, specifically the Order-Entry and Blood Bank modules to capture the reasons for transfusion. The hospital has switched from hand written orders to electronic orders from the Physician Order Entry module. A transfusion order can not be entered by the physician unless the reason for transfusion is stated. [Web master's note: the responding colleague did not comment if audits have been performed to assess how accurately the doctors enter into the computer the reasons for transfusion]. There are reminders in the system to alert the physician when the entered indication is not appropriate (according to their local practice) for the product ordered. The orders transmit to the Blood Bank module where the Clinical Lab Scientist reviews the order and gives a concurrent audit of appropriateness. Any orders that vary from their accepted list of appropriate indications are referred to the Blood Bank Director, pathologist on call, or member of the Transfusion Committee to discuss with the physician before the transfusion is given. (Editor's Note: The responding colleague did not address what happens if a request for transfusion is urgent, but the reason for the transfusion varies from their approved list of indications for transfusion].

4. The Texan who submitted reply #2 above wishes to add that at her hospital they have had a comprehensive transfusion documentation process in place for the last 5 years, which is what she based her initial reply upon. In 2003 they started a more prospective review of finding out appropriateness of the order before or within 24 hours of the transfusion. She admits that without computerization of the medical record, it would be an impossibly large task. Here is what they now do (verbatim) "We review 16% of all products transfused. Every Wednesday we print out the names of everyone who received products on Tuesday. We then look up their Hgb. and if its below 8, they meet criteria. If that isn't the case, we look up blood pressure, amount of blood loss, symptomology, etc. In other words, we go through all our criteria until we find one that the patient met. If they didn't meet any, we take the name to a Pathologist who gives a friendly call to the patient's physician to discuss the case. Our computer system prints the patient's Hgb, Hct, PT and PTT if available at the time we crossmatch so we are aware right away of those parameters. During surgery, we do rely on the judgment of the physicians in the OR, however, there is usually documentation in the computer of the amount of blood loss. We have some physicians who send patients for outpatient transfusions with the diagnosis of "Anemia". We have called their offices and found office personnel who will fax us the lab results at the time they send the patient to the hospital for his/her transfusion. This works particularly well with platelet counts. This process takes about 3 hours a week for 1 person to put this on a spreadsheet and graph it for a quarterly presentation to the Blood Utilization Review committee. The other thing we do is have the computer print a report each day of all the Hgb's below 7. We check these patients to see if they are transfused, or if not, why. Sometimes, it's legitimate: a post-partum with no symptoms, a Jehovah's witness, or a refusal. At other times, our Pathologists can intervene to get the patient transfused if they need it. This last report came about from a JCAHO request that we not only look at patients who got blood that didn't need it but also look at patients who needed blood and didn't get it. We'd like to expand this part to look at FFP and Platelet needs but haven't been able to formulate the report since there are heparinized, Coumadin "ized" and ITP patients. During the 1st quarter of 2003 we looked at 128 patients, 281 products of 1,748 transfused. Only 2 of those had to go to the BUR physicians for peer review. We also looked at 71 patients with low Hgb's. 60 of those were transfused, 7 were Post Partum with a Hgb between 6 and 7 and 1 was a DNR patient, 1 was a refusal, 1 was a Jehovah's Witness and only one was a peer review case."

ADDENDA May 2, 2003

5. A transfusion medicine colleague who practices in the Midwest also believes that physician documentation for the reason to transfuse is sorely lacking, and that this is a widespread problem. This colleague is of the opinion that while accrediting agencies require a written order for transfusion, these same agencies do not require a written reason for transfusion. The responding colleague performs blood administration audits at a large academic institution in the Midwest. Most times, when he looks in the patient chart for the physician's written orders to transfuse, he does NOT see a documented REASON for the transfusion. At his institution, the patient care unit is responsible for sending a "release form" to the transfusion service when they are ready to transfuse. This form has checkboxes that need to be filled in, stating the reason for transfusion. However, most of the time it is the unit secretaries who fill out the release forms, and he believes that they take their best shot at checking the appropriate reason for transfusion. Prospective auditing by Transfusion Service technologists reveals that the reason that is checked is not always accurate, leading to the conclusion that more responsibility needs to be placed on the physician to justify (in writing) his/her reason for ordering a transfusion.

ADDENDA May 5, 2003

6. A colleague in the Pacific Northwest reports that they implemented a system of prospective review about a year ago. She reports (verbatim) "All orders to transfuse must have a documented indication on the generic physician order form. When that is missing the technologist calls the nursing unit and the physician is notified. This process is not used for ED or surgery or a bleeding patient at any location. All plasma, platelet and cryo products are prepared only when there is a transfusion trigger. If there are no lab results or no trigger then the physician is called for test order or indication. We have many success stories using this process. Our process for review of red cell transfusion actually turns out to be retrospective review in many cases because we do not withhold red cell transfusion if there is no transfusion trigger. There are too many clinical situations that the technologist cannot evaluate. We do not have the luxury of residents or a Medical Director whose focus is Clinical Pathology so all of this is done by the blood bank technologists. All of the patients with orders to transfuse are documented on our prospective review log as the orders come in. The indication and product order are documented. This log is reviewed monthly by the technical specialist. The charts for the cases requiring further review are pulled. A primary review is performed by the technical specialist. When there is no obvious indication on the chart then the chart is reviewed by a transfusion committee physician and possibly discussed at the executive session of the transfusion committee which meets quarterly. If it is determined that there was no justification then a letter is sent to the physician with references related to the transfusion of the particular product. This process works well. We review approximately 200 cases a month and most of the review is done at the time of order. This process has prevented the transfusion of much FFP and cryo that would have otherwise been transfused needlessly. We have been disappointed in the documentation of the indication. In some cases we have to make a phone call and in most cases the indication is not on the order form but in the progress note which is OK but requires that the unit secretary search longer for the indication. In June we will implement a physician order form. It is to be used for orders to transfuse blood/components. It contains indications for each product and requires that the physician check a box next to the appropriate indication. Many physicians are eagerly awaiting the implementation of this form, not to mention the blood bank technologists and the unit secretaries."

ADDENDA May 6, 2003

7. A colleague in Saudi Arabia reports that at his institution the transfusion medicine service (TMS) staff review every order for blood product dispensing (prospective review) (written on specific double-ply form with relevant check-boxes for details relating to indication/diagnosis etc). If the recorded parameters indicate that the order is within the hospital (detailed) guidelines for transfusion of blood components, staff proceed to dispense the requested products. If some information is missing, staff will try to locate the missing data in the Laboratory Information System (LIS) or call the ward for information. If the order proves to be outside established guidelines, the TMS physician is called to discuss the matter with the ordering doctor or to approve as an exception - documented on a separate form. (There is a procedure for emergency release of O -RBCs in case of emerrgency). If there is disagreement with the ordering physician, the component is usually released (if there is any kind of urgency) and the case is referred to the Transfusion Committee. According to the responding colleague, they have very few cases for the Transfusion Committee to review, as most orders appear to be within established guidelines. They dispense around 30,000 blood components annually. They hope to implement a LIS-Patient info system, that will guide the ordering physician through the process and remind the physician if guidelines are being violated. It will be possible to over-ride the system, but a report will be printed for review/quality assurance.

ADDENDA Apr. 24, 2007

8. A Healthcare Standards Consultant is working with a medium-sized hospital laboratory on, among other things, documentation of their transfusion practices. The hospital employs a form that physicians are supposed to check off to indicate the rationale for each patient's transfusion. The form is basically a checklist that includes all the usual reasons for transfusions by product category and is used as part of Transfusion Review process. However, the hospital's experience with the form is that less than half the time it is filled out. Compliance with completing the form is NOT improved by education, memos, or other methods of persuasion. In addition to the strategies already mentioned in this discussion, what are institutions doing to gather this information? The inquiring consultant wants to know if any institutions are doing reviews of medical records to excerpt this information from the patient record?

ADDENDA May 13, 2007

9. Dr. Mark Friedman, Medical Director, Blood Bank and Transfusion Services, St. Luke's-Roosevelt Hospital Center in New York (attribution used with permission) writes:

"I am acutely familiar with the problem regarding poor transfusion documentation, having recently co-authored a publication on this topic (Ref: Friedman and Ebrahim. Adequacy of physician documentation of red blood cell transfusion and correlation with assessment of transfusion appropriateness. Arch Pathol Lab Med. 2006;130:474-9). Our data showed that nearly half (49%) of all medical record-reviewed transfusion events [TE's] (172 TE's involving 257 RBC units transfused to 154 patients over a 2 year period) were poorly documented and that lack of documentation correlated with failure to justify transfusions retrospectively, as would be expected (over 60% of poorly-documented TE's could not be justified compared with less than 10% for TE's with adequate documentation). As such, we have taken steps already mentioned by others in this forum (staff education, formative physician letters) to improve blood utilization and documentation, though we have yet to quantify any significant improvement in the latter aspect (ideally, we should do a follow-up study). Routinely, we audit one-week's worth of red cell transfusions per month by first comparing patient lab data from the hospital information system to our transfusion guidelines; if the transfusion cannot be justified via this initial assessment, then the medical record is reviewed by myself (the Transfusion Service Medical Director). Cases not justified by medical record review are referred to the Transfusion Committee for further review and letter to the responsible physician as deemed necessary."

ADDENDA May 25, 2007

10. Editors' Note: The following article, Rock G et al. Documenting a transfusion: how well is it done? Transfusion 2007 April; 47(4);568–72 is germane to this ongoing discussion.

ADDENDA June 1, 2007

11. A Transfusion Medicine Laboratory Supervisor in Mississauga, Ontario Canada reports that several years ago their community hospital formalized Clinical Practice Guidelines for Transfusion of Red Cells, Plasma, Platelets and Cryoprecipitate. At the same time a patient information pamphlet was developed, as was an Information Sheet for Healthcare Providers, and a Physician Pre-printed Order form (PPO). The PPO is completed when the decision to transfuse is made and provides documentation (via check boxes or "other") of the reason for transfusion. There is provision for nursing to document telephone orders. The form also contains a check box to indicate that consent for transfusion has been obtained.

The PPO form is used to obtain product from the Transfusion Medicine Lab. If the consent box on the PPO form is not completed, follow up is done prior to product issue. The Technologist also reviews the reason for transfusion and consults their Pathologist or the ordering physician when the order may be inappropriate. Situations involving uncrossmatched blood bypass this process.

Introduction of this new process took place over a year at which time the form completion became mandatory. Compliance remains an issue, but ongoing audits allow for review and follow up with specific clinical areas and specific clinicians.

Please submit comments to the e-Network Forum.

Ira A. Shulman, MD
CBBS e-Network Forum Editor & Moderator

W. Tait Stevens, MD
CBBS e-Network Forum Assistant Editor & Moderator