The Joint Commission (TJC), through its Blood Management Performance Measures Technical Advisory Panel, has identified 19 proposed blood-related performance measures that can be uniformly implemented in health care organizations.
TJC is seeking stakeholder review and public comments in an effort to assess the proposed measures. These measures were developed as part of the Blood Management Performance Measures Project, which was initiated this past January to identify, develop and test a set of standardized measures for blood management in the hospital setting. According to TJC, compliance with each individual performance measure, as well as when taken together, could help health care organizations, government agencies and business communities assess quantifiable improvements in the care of patients who receive blood products.
AABB members are encouraged to review the proposed measures and submit comments appropriate for their respective facilities. To aid members in preparing comments and providing feedback, two AABB committees — the Clinical Transfusion Medicine Committee (CTMC) and the Scientific Section Coordinating Committee (SSCC) — have reviewed these measures detailing issues, concerns, advantages and disadvantages for each one of the proposed 19 Performance Measures. These comments reflect the expertise of the committee members but are not necessarily the opinion of AABB. The full text of the Blood Management Performance Measures can be viewed by clicking on the following link (Blood Management Performance Measures). Each measure is then listed below along with the comments prepared by the committees.
Note: The Joint Commission has provided an electronic survey to facilitate the receipt of feedback. The survey is available at Request for Public Comment on Blood Management Performance Measures. The survey must be completed by August 19, 2008.
Questions regarding this summary analysis should be sent to the AABB Public Policy Department.
BM-1 Transfusion Indication
· This measure is generally reasonable, appropriate, and already tracked in many settings.
· Implementation may lead to outcomes improvement by facilitating easier review of orders and indications.
o Depending on the health care system, this measure may be very difficult and labor intensive to monitor. Few institutions would be able to capture these data electronically. In addition, some order-entry systems include standing orders (e.g., ECMO, heme-onc) that are separated from the blood orders and laboratory tests. Manual review would be required to retrospectively tie the pre-post lab tests with the time of issue in many settings.
o It would be difficult to monitor this measure in operating rooms.
o The category of “patients that receive blood emergently” must be better defined, and may also require some corresponding documentation (e.g., trauma protocol).
BM-2 Blood Transfusion Informed Consent
· Documentation and compliance for consent is appropriate, measurable, and in some facilities can be audited via the electronic medical records under consent documents.
o It would be better to document that either informed consent was obtained or that the urgency of the clinical situation warranted transfusion without obtaining informed consent first. What is desired is that clinicians document something (informed consent or urgency of situation).
o This measure should also take into account how often a consent needs to be signed for chronic transfusion patients (i.e., sickle cell anemia patients).
o This measure must consider whether the facility uses a consent for blood that is separate from other consents (e.g., consent associated with surgery)?
o Because informed consent is often documented in different ways, in some institutions monitoring of this parameter would require manual chart review.
o To some, this is a major patient care issue.
BM-3 Blood Administration Documentation
· Documentation and compliance are desired and achievable, and may lead to outcomes improvement.
o This measure combines too many elements (e.g., patient identification, monitoring and vital signs, and adverse reactions and interventions) into a single measure to be useful for outcomes improvement.
o Definition of active bleeding as 4 units of RBCs within 1 hour is too stringent, does not include other emergency situations, and may lead to additional difficulties in identifying patients that fit these criteria.
o In pediatrics, “patient education” needs to be parent education. This is consistent with an overall lack of attention to pediatric transfusion patients in these draft measures.
o Blood administration documentation is done many different ways in different care settings and in different institutions. Capturing these data will require a lot of work, most of it manual.
BM-4 Time to Crossmatched Blood Administration
o None identified.
o This measure misses the important point. Time to crossmatched blood administration is only one aspect of emergency transfusion. The real point is adequacy and timeliness of transfusion. As written this would be of no value.
o It would be difficult to identify, track, and measure this small patient subset.
o Outcomes for this measure are significantly affected by technologies that are differentially available and/or applied in different settings (e.g., use of electronic crossmatch, availability of emergency release unit in the emergency department).
o The specifics of the proposed measure do not take into account currently available technology. Most automated testing requires 45 minutes at a minimum, excluding labeling, dispensing, transport, etc.
o How do you correct the results for patients who, following initial evaluation, did not need crossmatched blood after all?
o Numerator of patients transfused with crossmatched blood within 60 minutes of arrival in ED who receive uncrossmatched blood is not meaningful. This presumes that all uncrossmatched blood in ED is requested before or at arrival which is not always the case. Measure should be 60 minutes after receiving uncrossmatched blood.
o This measure might be improved by focusing on turn around time from receipt of stat emergency department samples in the blood bank to completion of ABO type and antibody screen. Also, one could separately focus on time from entry of patient into the emergency department to the time the specimen arrives in the blood bank.
BM-5 Time to Uncrossmatched Blood Administration
- The patients represent a very small population in some hospitals, which can confound and skew analysis.
- Manual audits will probably be required to track this measure, and may lead to inaccuracies. Uncrossmatched blood orders are often verbal, and documentation of start times is probably handwritten.
- An additional weakness of this measure is that patients may be receiving uncrossmatched blood inappropriately.
BM-6 Pre-operative Blood Tests
· This is both measurable and may lead to outcomes improvement.
· This is an important patient safety indicator, especially in light of a recent FDA report on transfusion fatalities which shows from 2005 to 2006 there were more fatalities due to acute hemolysis caused by incompatibilities at "minor" blood groups, rather than ABO. For example, after an elective surgery case has started the sample sent to the blood bank may show a positive antibody screen---unfortunately, if there is no time for antibody identification the risk to the patient may be unnecessarily increased.
o The denominator should be elective surgical patients having operations that often require red cell transfusion, not all elective surgical patients (this would include cases that generally do not take blood at all).
o This measure is difficult to control, and is driven by market forces and patient preferences outside control of hospital. For example, many patients prefer not to come in for testing prior to surgery. In addition, this measure is most relevant to patients with a.m. surgeries, not all elective surgical patients.
o The Q-Probe looked at whether screening or crossmatches were completed before the start of surgery, not one day before. Although encouraging preadmission testing is very laudable, the start of surgery (or entering the operating room) is probably the better endpoint.
o This proposed measure needs major revision. What is "elective" surgery? How do you define it and identify this patient subgroup for auditing? The population needs to be defined much more clearly.
BM-7 Pre-operative Hemoglobin Level
· Measureable and may lead to outcomes improvement.
o This measure appears to be based on isolated studies that have not been subjected to systematic review and guideline formulation. These kinds of measures should await analysis, integration, and interpretation of better clinical data.
o There is no clear benchmark.
o Not all elective surgery patients have a hemoglobin level. This would be hard to track since it does not correlate with the audit criteria of a given facility.
o There is still substantial variation in practice patterns. Different patient populations (requiring different transfusion approaches) may predominate in different institutions.
o If adopted in its present form, and at this time, this measure may well create a false standard for transfusion.
BM-8 Blood Salvage Documentation
· Measureable and may possibly improve outcomes.
o This is not currently tracked or monitored by most transfusion services, since it is not often under the auspices of the transfusion service.
o Most of these data are recorded in surgical records, and would require manual compilation.
o Should also include the volume salvaged, especially postoperatively. There should be documentation that the time between beginning of salvage to infusion is less than 8 hours.
o This is not necessarily a major clinical problem.
BM-9 Transfusion Complications
o This measure can be significantly skewed by the differing complexity of patients in different hospital settings.
o The specific codes being proposed are not specifically transfusion related (e.g., hemolysis NEC, sepsis/infection, shock, thromboembolism). To be useful, the reaction rates should reflect the transfusion reaction categories standard to our profession (e.g., hemolytic, fever/chill non-hemolytic, urticarial, anaphylactic, TRALI, sepsis, hypotension associated with ACE inhibition, circulatory overload, non-immune hemolysis, air embolus, hypocalcemia).
o Not all of the complications listed are true measures of performance (i.e., no way for hospital to lower the complication rate). Only incompatible reactions can be improved upon, but their frequency is likely to be too low and should just continued to be treated as sentinel events.
BM-10 Red Blood Cell Administration
· Measurable and may lead to outcomes improvement.
o This measure is based on isolated studies that have not been subjected to systematic review and guideline formulation. These kinds of measures should await analysis, integration, and interpretation of better clinical data.
o There is no clear benchmark.
o What is the evidence for excluding neurosurgical patients?
o Despite the TRICC trial there is a great deal of controversy about the red cell transfusion trigger. This might be meaningful if broken down by specific patient populations such as adult intensive care without cardiac disease. As written it would be meaningless. Also, intraoperative transfusions rarely have an accurate pretransfusion hemoglobin.
BM-11 Multiple RBC Units Transfused
· May be measurable, and possibly lead to outcomes improvement.
o This is based on isolated studies that have not been subjected to systematic review and guideline formulation. These kind of measures should await analysis, integration, and interpretation of better clinical data.
o There is no clear benchmark.
o It may be difficult to identify those patients who receive more than one unit of RBCs and to then look at the pre-transfusion hemoglobin.
o Exclusion criteria are a too stringent.
o While the intension of this measure may be good, as written it would be meaningless. It does not take into account bleeding. Intraoperative transfusions rarely have serial hemoglobins.
BM-12 Platelet Administration A
· Measurable and possibly leads to outcomes improvement.
o Should exclude specific patient populations known to have higher platelet requirements: transplant patients with fever, sepsis, undergoing invasive procedures, etc.
o The excluded populations are too narrow, given that they are including stem cell patients. It is not uncommon to have to raise platelet thresholds to 30-40K or more for hemorrhagic cystitis and other complications.
o ECMO is also another procedure requiring platelets but <4 units/RBC per hour.
o A cut-off of < 20,000 for outpatients, particularly for patients who live in a rural community at a distance from medical care, is not uncommon. Also for patients on weekly outpatient transfusions, platelet counts are not always obtained before transfusion.
BM-13 Platelet Administration B
· Measurable and possibly leads to outcomes improvement.
· The active bleeding definition, while strict, is probably acceptable.
o The platelet thresholds proposed here are not based on well-performed studies but are largely anecdotal.
o Does not take into account patients who are on platelet inhibitors (clopidogrel). The auditing will probably be manual, and the large number of various exclusion criteria makes auditing even more difficult (e.g., exclusion criteria of 4 U of RBCs transfused is too stringent).
o While this is a significant issue, it is also highly controversial.
BM-14 Plasma Administration
· Measurable and should lead to outcomes improvement.
o This is based on isolated studies that have not been subjected to systematic review and guideline formulation. These kinds of measures should await analysis, integration, and interpretation of better clinical data.
o This is weak in terms of appropriate use of FFP. There is no reference range for INR or any directive as an indication for transfusion. It does not address any kind of liver implications as well as when it is appropriate to use FFP to reverse warfarin (as opposed to vitamin K administration).
o Suggest making it the prothrombin time equivalent to INR 1.5, for the coagulation purist.
o This is a significant issue, but the population needs to be more clearly defined. The numerator should be INR <1.5. Children should be excluded as the INR reference range is not defined.
BM-15 Patient RBC Transfusion Rate
o This measure is highly focused on the concept of blood avoidance.
o Many of the procedures in this list are only performed in larger tertiary care centers and are not applicable to the majority of institutions.
o In some institutions, usage rates are routinely generated in cardiac surgery by DRG, with expansion of the data warehouse to other procedures. Whether this is feasible in institutions with limited computer infrastructure and support should also be of concern.
o This measure is too general for specific guidance or use to improve transfusion practice. Also the measure might be a reflection of greater or lesser access to care, or to elective surgeries vs. similar levels of healthcare in other countries but with greater rationing of invasive procedures.
o Too global, broad and subject to intrahospital variation to be meaningful for quality improvement.
o Why exclude exchange transfusions?
o Unless this is broken down by very specific patient populations it will be meaningless. Even breaking it down by DRG may not be sufficient. This needs more thought.
BM-16 DRG Blood Utilization
· Measurable and may lead to outcomes improvement.
· It is possible that stratifying by age would be useful.
o There does not appear to be an obvious advantage to collecting these data
o Collecting data for this measure is not easy or practical. There are many included DRGS so to learn from any data gathered at an individual hospital level may be difficult.
o No specific blood products were identified (RBCS vs. platelets vs. plasma).
o Too ambitious and may not correlate with proper blood use per DRG.
o The calculation as described is confusing. Is the numerator per patient, not per total patients in a DRG?
o Lots of intra-center variation seen in liver transplants and cardiac surgery.
o This measure needs more thought. High blood use in and of itself does not necessarily mean inappropriate use. A heart transplant after VAD usually required more blood than a previously unoperated patient. Liver transplantation for PBC usually uses much less than alcoholic cirrhosis.
BM-17 Specimen Testing Criteria
· Measurement of mislabeled specimens is appropriate, important, and broadly applicable to most hospitals.
· Pretransfusion specimen labeling is very important and it should be fairly easy to obtain relevant data.
o What will the benchmark be -- percentage should be 100%, but is that realistic?
o Too many elements, better to focus on specimen rejection rate instead and to then focus on some categories of rejection. It seems this measure was drafted in an attempt to capture it all in one number, but to do so would hide or obscure specific trends in errors in certain categories of specimen collection.
o This will be hard to collect. Many labs capture specimen rejections but do not split out specific areas. Much of this may have to be collected manually.
BM-18 Blood Product Wastage
· Possibly measurable.
o The point of this measure is unclear. Hospitals already have enough incentive to monitor and reduce wastage.
o This measure can be affected by the complexity of the patient mix in each hospital. Furthermore, in this regard, platelets are probably the least helpful type of units to monitor here.
o Would not lead to outcomes improvement.
o This measure is not particularly relevant since there is still variation in practice patterns and since certain patient populations may predominate from institution to institution. For example, it probably would not be valuable to compare red cell transfusion rates of Hospital X, which is a cardiac surgery center, to Hospital Y, which has a large obstetrics patient base; or to compare platelet transfusion rates at a large academic center with stem cell transplant, to someplace that provides little oncology service.
o This measure is not stratified by any demographic distinction: size, scope of practice, etc. The numerator and denominator only address platelets not other products.
o Wastage is an economic issue, not really a quality of care issue.
BM-19 Blood Inventory Level
o None identified.
o Blood shortages are usually a supply problem, not an inventory management problem. They have no exclusions, ignoring the dramatic impact that a few patients can have on the entire hospital system—particularly during times of preexisting short supply.
o Not measurable, nor would it likely lead to outcomes improvement.
o This measure is ill-defined as written. The number of occurrences of adequate inventory is confusing. Better to try and capture number of days inventory was below set requirements for adequacy. However, this requires that transfusion services already have formal inventory level requirements. Also, when would inventory assessment be performed? (at the beginning of the day vs. at the end of the day? ). Which blood products?
o Inventory management depends not only on the hospital but also the blood supplier.
o What is the numerator? What is the denominator? The issues are not under hospital control, so tracking this would be of limited value for improving patient care by the hospital.
o Adequacy of blood inventory is important, but it is very difficult to define. It is common for the inventory to be below an arbitrarily established target, but it is rare that surgery or other patient care is delayed or denied. More thought needs to be given to measuring the clinical impact.
q Clarifications needed :
1. The 19 measures will ultimately be reduced to a more reasonable number. There may be significant concern over the fact that there are 19 measures currently proposed.
2. There should also be clarification that these measures are mandatory only if the hospital chooses to use them as one of its performance measure sets.
3. Participants should be advised that these measures when reported to The Joint Commission do become part of the public domain (i.e., they could be accessible to the general public).
4. There is a possibility that these measures will eventually become part of the core measures which all hospitals must measure.
q Warnings of participation bias:
Since not all hospitals will choose to use these performance measures, there is participation bias as to which hospitals would choose to use these performance measures. However, the results obtained by these performance measures will likely be misinterpreted as benchmarks for all hospitals despite the participation bias.
q Better cooperation between AABB and The Joint Commission:
If Joint Commission requirements for Blood Banks and Transfusion Services are not coordinated with AABB Standards, the value of the uniform nationwide approach to maximizing patient care and safety could be jeopardized. While it is recognized that AABB has had input into drafting of these measures, further coordination is needed.
q Performance Measures vs. Standards in Disguise:
The Joint Commission should only incorporate or choose performance measures that reflect existing Joint Commission or AABB standards. Otherwise these performance measures are essentially Joint Commission standards in disguise, but without going through the more rigorous Joint Commission standards creation and review process. If there is no standard on a particular issue, the an appropriate standard should be created first.
q Unintended consequences of Performance Measures on Electronic Health Record (EHR) or Computer Physician Order Entry (CPOE):
Although well intended, there may be a number of hospitals who are in the process of creating or designing CPOE and EHR systems for their hospitals and may be compelled to create mandatory documentation requirements (hard stops in the ordering or documentation process) that may be compliant by design but may impede life-saving treatments instead of creating electronic auditing or reporting processes that capture data. For example a system could be created where no transfusion order can be transmitted, issued, or started without all required documentation and indications being met despite an exsanguinating bleed from a previously normal hematocrit. The performance measures would be superb in this design (100% because of the hard stops) but there would be avoidable fatalities due to systems flaws (see Han et al. Pediatrics 2005; 116:1506-1512). Clear and unambiguous directives must be given to hospitals customizing their CPOE and EHR to avoid these types of systems flaws with the implementation of any mandatory performance measure.
q Provision of these data would require information from the HIS and the LIS, which for some centers may be difficult and labor intensive to extract. For many of the measures, such as the percentage of patients needing emergency release blood that receive it within 15 minutes of arrival, the necessary information is not in the LIS.
q Renewed attention by The Joint Commission to blood transfusion is beneficial, and carries great impact for hospitals. For measures which are not being done already, the availability of transfusion performance measures could be an inducement to find ways and means to accomplish these studies. On the other hand, many hospitals would probably not have digitized data or software to find data necessary for these analyses. There are also large differences between hospitals (size, types of patients) which would make the comparability of some of these measures difficult.
q The Joint Commission hospital accreditation requirements for performance measures are at APR.04.01.01 in the 2009 standards currently on line. However, it is difficult on The Joint Commission website to identify what the available non-core performance measures are.