Background

Patient blood management programs are gaining popularity as quality improvement and patient safety initiatives, but methods for implementing such programs across multihospital health systems are not well understood. Having recently incorporated a patient blood management program across our health system using a clinical community approach, we describe our methods and results.

Methods

We formed the Johns Hopkins Health System blood management clinical community to reduce transfusion overuse across five hospitals. This physician-led, multidisciplinary, collaborative, quality-improvement team (the clinical community) worked to implement best practices for patient blood management, which we describe in detail. Changes in blood utilization and blood acquisition costs were compared for the pre– and post–patient blood management time periods.

Results

Across the health system, multiunit erythrocyte transfusion orders decreased from 39.7 to 20.2% (by 49%; P < 0.0001). The percentage of patients transfused decreased for erythrocytes from 11.3 to 10.4%, for plasma from 2.9 to 2.2%, and for platelets from 3.1 to 2.7%, (P < 0.0001 for all three). The number of units transfused per 1,000 patients decreased for erythrocytes from 455 to 365 (by 19.8%; P < 0.0001), for plasma from 175 to 107 (by 38.9%; P = 0.0002), and for platelets from 167 to 141 (by 15.6%; P = 0.04). Blood acquisition cost savings were $2,120,273/yr, an approximate 400% return on investment for our patient blood management efforts. Conclusions Implementing a health system-wide patient blood management program by using a clinical community approach substantially reduced blood utilization and blood acquisition costs. What We Already Know about This Topic • Blood transfusion is the most common procedure performed in U.S. hospitals and is often overutilized • Encouraging evidence-based transfusion practice can improve blood utilization What This Article Tells Us That Is New • Methods are described for implementing a patient blood management program across a multiinstitutional healthcare system as a quality improvement and patient safety effort • Promoting best practices for patient blood management across a health system can reduce unnecessary transfusions, overall blood utilization, and costs, with a 400% return on financial investment GIVEN that blood transfusion is the most common procedure performed in U.S. hospitals,1 and one of the top five overused procedures according to The Joint Commission,2 it seems prudent that hospitals implement patient blood management (PBM) programs.3,4 Curtailing the overuse of blood transfusions fits well with the Choosing Wisely Campaign, which has focused on opportunities to reduce unnecessary tests and procedures since its launch in 2013. Now five national societies, including the American Society of Anesthesiologists (ASA), have adopted Choosing Wisely aims to reduce unnecessary transfusions.5,6 PBM programs also share common goals with the Perioperative Surgical Home, introduced and defined by the ASA as a “patient-centered, physician-led interdisciplinary team-based system of coordinated patient care.”7 The Perioperative Surgical Home aims to enhance value and achieve a better patient experience, improve health care, and lower costs. Moreover, the cost-saving potential is substantial with PBM because, in the current U.S. healthcare environment, blood is poorly or not reimbursed.8 In the five adult hospitals in our health system, we spend almost$30 million/yr for blood acquisition; however, the true cost is three to four times that,9  owing to the tremendous overhead involved in bringing blood from donor to recipient. Blood is a valuable and often scarce resource; therefore, avoiding unnecessary transfusions will increase blood availability for patients who truly need this lifesaving therapy.

In this article, we describe the implementation of our health system–wide PBM program, using a “clinical community” approach, first described by Professor Mary Dixon-Woods et al.10  at Leicester University as a physician-led, self-governing network of stakeholders who work together to identify and achieve goals related to quality in health care.11  This approach is particularly suited to encourage participation in quality improvement efforts by connecting frontline providers across a multiinstitutional healthcare system and creating a densely networked community with strong horizontal links that exert pressure to conform to initiatives and goals. Such an approach was successfully used in a national effort to reduce central line–associated bloodstream infections in intensive care units.12  Clinical communities achieve success in a sustainable way by supporting clinicians to drive quality improvement. The clinical communities rely on accountable physician leaders to engage, enable, and excite their clinical teams. Anesthesiologists make natural team leaders in a PBM program because they are on the front line, administering blood to patients in the operating rooms and intensive care units, and collaborating with multiple other specialties providing care to patients. Thus, anesthesiologists can lead by education and promotion of best practices across the healthcare system.

Several authors have described the methods and results of PBM programs used in their own institutions.13–15  One area in the PBM field that lacks adequate information relates to implementing a PBM program across a multihospital health system. Because hospitals are merging at a rapid pace, or being bought out by health system corporations, the need for system-wide PBM programs is increasing. Having established a PBM program over the past 5 yr, and expanded it throughout five hospitals in our health system during the past 3 yr, we are able to report on the methods used and lessons learned. With this information, other health systems may choose to adopt clinical communities into their quality improvement efforts to promote patient safety, reduce risk and cost, and thus increase the value of health care delivered.

The Johns Hopkins Health System includes six hospitals, five of which are included in the current report; the sixth is a pediatric specialty hospital. Two are academic centers with residency training programs in multiple specialties, and the other three are community hospitals that cover a wide variety of medical and surgical specialties. Approval from the Johns Hopkins Institutional Review Board (Baltimore, Maryland) was obtained to assess all blood utilization data.

### PBM Program Startup

The largest of the five hospitals is the Johns Hopkins Hospital (JHH), where a PBM program was started in January 2012. An educational campaign was begun with Grand Rounds lectures to each department, covering all methods used to improve blood utilization in a PBM program (listed in table 1). The focus was initially on surgical services and erythrocyte (RBC) hemoglobin (Hb) transfusion triggers (the Hb before the transfusion) and targets (the Hb after the transfusion) as we have previously described.16,17  Emphasis was given to randomized trials, and now eight landmark studies support a restrictive over a liberal RBC transfusion strategy.18–25  The educational campaign also involved spreading awareness of the hospital transfusion policy guidelines for plasma and platelets. Exceptions to laboratory value-based transfusion thresholds were made for patients who are actively bleeding or hemodynamically unstable. For data acquisition early on, we chose a commercially available system, IMPACT Online (Haemonetics Corp., USA), as a blood management intelligence portal, given its unique focus on all variables related to PBM. We have previously described using this system for data acquisition.16,26  Although we have now obtained institutional support to develop our own data systems, we believe that early adoption of a commercial system enabled us to demonstrate sufficient success to justify continuing program efforts.

Table 1.

Steps for Implementation of the Blood Management Clinical Community

A formal business plan was drafted for expanding the program across the health system. This included salary support for three physicians: an anesthesiologist as the medical director of the program at 50% effort, and two transfusion medicine specialists at 25 and 10% effort. These physicians coordinated and carried out the programmatic methods outlined in table 1. The business plan also included support for a nurse coordinator and a data manager, both at 50% effort. We continued using the IMPACT Online data portal for JHH ($36,000/yr); however, the Epic electronic record (Epic, USA) became the primary source of PBM data across the entire health system. Including miscellaneous expenses, the total cost to support the PBM program was approximately$400,000, not including existing infrastructure resources from the Division of Clinical Analytics to set up the data dashboards. The return on investment (ROI) with a 5, 10, and 15% reduction in blood acquisition cost was calculated to be 120, 340, and 560%, respectively.

### Expansion to All Five Hospitals

The steps taken to expand our PBM program across the health system are summarized in table 1. In July 2014, the Clinical Community was formed, which consisted of a multidisciplinary team of individuals organized around a central goal of quality improvement. To improve the structure and impact of the program, we obtained support from the Johns Hopkins Armstrong Institute for Patient Safety and Quality Clinical Communities Project Management Office (Baltimore, Maryland), which formally helped transition the blood management program into a Clinical Community. This provided clinicians access to resources, such as technical experts and safety and quality researchers, who were otherwise not easily accessible or available. Support also was obtained from the health system leadership. Then a multidisciplinary group of providers was recruited from several departments (e.g., Surgery, Obstetrics/Gynecology, Anesthesiology, Perfusion, Pathology [Transfusion Medicine], Internal Medicine, Hematology, Bloodless Medicine, and Surgery) and from all five hospitals. We began having two meetings each month—one for the JHH PBM program and one for the Johns Hopkins Health System PBM program. At these monthly meetings, the collaborative spirit of the clinical community embraced the interdisciplinary membership and facilitated discussion of shared experiences and best practices used to promote quality improvement goals.

After adopting (harmonizing) standardized transfusion guidelines across the health system, we began an educational campaign at the other four hospitals, again covering all methods outlined in table 1. We also implemented the Epic electronic record across all five hospitals between 2014 and 2016, which included customized transfusion order sets for adults, pediatrics, and neonates, for the four different blood components, or 12 order sets in all. Considering that each of these includes a “prepare” and a “transfuse” order, 24 order sets were created.

### Computerized Provider Order Entry with Clinical Decision Support

Modeling the system described by Goodnough et al.,27  we implemented a best practice advisory (BPA) in the computerized provider order entry (CPOE) system at the main campus (JHH) for RBCs in April 2013. This advisory was built with logic to check the last measured Hb value, with an alert triggering if the Hb was greater than or equal to 8 g/dL or if Hb had not been measured within the past 24 h. A year later, the BPA was changed to trigger with a Hb greater than or equal to 7 g/dL. This change was implemented after we recognized that about one third of all RBC transfusions in the hospital were occurring with preceding Hb concentrations between 7 and 8 g/dL. The alert message also included a Choosing Wisely recommendation to administer single-unit RBC transfusions to patients who are not actively bleeding and are hemodynamically stable.6

In September 2015, BPAs were implemented across the entire health system, including the operating rooms and procedural areas, for RBCs, plasma, and platelets. The BPA for RBCs and the indications to proceed with the RBC order are shown in figure 1. Active bleeding was a choice to transfuse outside of typical guidelines, which, for example, allowed intraoperative transfusion to occur unimpeded after acknowledging this reason. The BPA only fires one time per day, for any given blood component ordered by any one provider, to allow unimpeded ordering upon subsequent orders of the same product by the same provider. The BPAs included a laboratory value threshold for utilization review, based on our harmonized transfusion guidelines. These BPAs also served an educational purpose by showing the ordering provider the health system’s approved transfusion guidelines. Our plasma BPA is based on the last measured international normalized ratio (INR) results. It calls for a threshold of greater than or equal to 1.5 for actively bleeding patients, with a recommendation threshold of greater than or equal to 1.7 for patients undergoing surgery or invasive procedures.28  A platelet BPA based on a platelet count threshold of greater than or equal to 50 K also was implemented.

Fig. 1.

Computerized Provider Order Entry alerts. (A) When an erythrocyte (RBC) order is placed, an interruptive best practice advisory (BPA) is triggered if the preceding hemoglobin (Hb) value is greater than or equal to 7 g/dL or has not been measured in the previous 24 h. If the preceding Hb is less than 7 g/dL, then no BPA is triggered and the order is placed unimpeded. After acknowledging the BPA, the provider can either cancel the order or proceed by choosing a reason (B) to override the BPA and transfuse RBCs. These reasons are taken from the health system’s transfusion guideline policy.

Fig. 1.

Computerized Provider Order Entry alerts. (A) When an erythrocyte (RBC) order is placed, an interruptive best practice advisory (BPA) is triggered if the preceding hemoglobin (Hb) value is greater than or equal to 7 g/dL or has not been measured in the previous 24 h. If the preceding Hb is less than 7 g/dL, then no BPA is triggered and the order is placed unimpeded. After acknowledging the BPA, the provider can either cancel the order or proceed by choosing a reason (B) to override the BPA and transfuse RBCs. These reasons are taken from the health system’s transfusion guideline policy.

Close modal

### Data and Value Analytics: Dashboards and Transfusion Guideline Compliance Reports

Under the guidance of the PBM Clinical Community, we created dashboards and reports that we have recently described.29  Dashboards show the big-picture changes in blood utilization for the entire health system and for each hospital, whereas reports show individual providers’ transfusion guideline compliance rates and allow comparison to peers within their own specialty (fig. 2). We created these provider-specific reports using blood management dashboards built on a Qlikview (Qlik, USA) business intelligence platform.29  We found the method of greatest impact for peer comparison to be the rank-order bar graph with provider names visible within departments, but not between departments.30  A green-yellow-red color scheme correlated with preceding Hb values of less than 7 g/dL, 7 to 7.9 g/dL, and greater than or equal to 8 g/dL, respectively.31  The reports also compared the percentages of 1-unit and 2-or-more-unit RBC orders among providers.29  For plasma, green and red were used for preceding INR values of greater than or equal to 1.5 and less than 1.5, respectively.32  For platelets, green and red were used for preceding platelet counts of less than 50,000, and greater than or equal to 50,000, respectively.33

Fig. 2.

Two provider-specific departmental blood utilization reports are shown for a 1-month period for the Department of Internal Medicine (top two panels) and the Department of Orthopedic Surgery (bottom two panels). On the x-axis is the number of erythrocyte (RBC) units attributed to each attending physician. Physicians are identified by their numeric identification code on the y-axis. This type of report is sent to each department within the hospital. The proportion of RBC orders is shown according to whether they were for 1 or 2 or more units (left figures) and by the preceding hemoglobin value (right figures) by color coding (green = Hb < 7 g/dL; yellow = Hb 7 to 7.9 g/dL; red = Hb ≥ 8 g/dL). Both departments have a low rate of 2-or-more-unit RBC orders; however, the Department of Internal Medicine has a much larger proportion of RBC orders with a Hb trigger < 7 g/dL. Data were acquired from the Epic (USA) computerized provider order entry system. HGB = hemoglobin; Prov = provider.

Fig. 2.

Two provider-specific departmental blood utilization reports are shown for a 1-month period for the Department of Internal Medicine (top two panels) and the Department of Orthopedic Surgery (bottom two panels). On the x-axis is the number of erythrocyte (RBC) units attributed to each attending physician. Physicians are identified by their numeric identification code on the y-axis. This type of report is sent to each department within the hospital. The proportion of RBC orders is shown according to whether they were for 1 or 2 or more units (left figures) and by the preceding hemoglobin value (right figures) by color coding (green = Hb < 7 g/dL; yellow = Hb 7 to 7.9 g/dL; red = Hb ≥ 8 g/dL). Both departments have a low rate of 2-or-more-unit RBC orders; however, the Department of Internal Medicine has a much larger proportion of RBC orders with a Hb trigger < 7 g/dL. Data were acquired from the Epic (USA) computerized provider order entry system. HGB = hemoglobin; Prov = provider.

Close modal

In July 2015, we began sending these provider-level reports out monthly at the main campus (JHH), and quarterly at the other four hospitals, to the Vice Presidents of Medical Affairs (VPMAs) and to each department chair, who disseminated the reports to attending physician members in their department. An example of this report is shown in figure 2.

### “Why Give Two When One Will Do?” Choosing Wisely Campaign

After closely monitoring Hb transfusion thresholds for RBC transfusions for approximately 2 yr at the main campus, we recognized that the threshold (Hb trigger) was often guideline-compliant but that many orders were routinely placed for two RBC units, resulting in a higher-than-evidence–based Hb target after the transfusions. This presented a substantial opportunity to reduce RBC utilization. In April 2015, a “Why Give Two When One Will Do?” campaign was launched in an effort to reduce RBC overuse.34  A communications campaign was begun with newsletter articles and screensaver messages showing the American Association of Blood Banks (AABB) Choosing Wisely aims,6  which advocate single-unit RBC transfusions in hemodynamically stable, nonbleeding patients. This campaign was deemed necessary given that multiunit RBC orders were extremely common in three of our hospitals (~66%), and somewhat common in the other two (20 to 40%).

### Other Interventions

The various methods used to reduce unnecessary transfusions are summarized in table 1. The educational curriculum encouraged use of antifibrinolytic medications (e.g., tranexamic acid) for orthopedic surgeries, the early diagnosis and treatment of preoperative anemia before elective surgery (when feasible), and other changes in practice such as minimizing phlebotomy-related blood loss for laboratory tests by using smaller blood-collection tubes, using in-line blood return devices, and reducing routine daily tests in stable hospitalized patients.35  Autologous blood salvage (cell salvage) was also encouraged for specific cases such as transplant, cardiac, and major orthopedic and spine surgeries. Point-of-care testing also was encouraged (e.g., thromboelastography) to reduce blood utilization in cardiac, transplant, and major vascular surgery at the main campus (JHH), but the other four hospitals did not have thromboelastography available. Anesthetic methods (normothermia, controlled hypotension, autologous normovolemic hemodilution) and surgical techniques (topical hemostatic agents, newer cautery systems, and minimally invasive and robotic approaches) also were encouraged.36,37

### Data Analysis

The number of patients included in this report was based on available data and no a priori power calculation was conducted. All measured outcomes of interest were assessed at monthly intervals, and the changes over time were analyzed. For statistical analysis, the baseline pre-PBM time period was considered to be FY2014 (July 1, 2013 to June 30, 2014), and the post-PBM time period was considered to be FY2015, FY2016, and the first six months of FY2017 (July 1, 2014 to December 30, 2016). Because some 2014 data were missing for Hospital 5 owing to a change in the electronic health record (EHR), we used the first four months of available data (July to October 2014) as the baseline, after annualizing by a threefold multiplier. We assessed changes in blood utilization using all units/all patients (including those not transfused), because nontransfused patients are important in a PBM program. Results are reported as number of units per 1,000 discharged patients.27  Blood and cost savings were calculated by comparing the most recent year (FY2017) to the FY2014 baseline. We used actual acquisition costs for the three blood components and not the activity-based costs,9  which include estimated overhead costs. Proportions were compared by chi-square tests. Changes in blood utilization were analyzed by one-way ANOVA. P < 0.05 (two-tailed design) was considered statistically significant. All analyses were done using JMP ver. 12.1.0 (SAS Institutes, USA).

The characteristics of the five hospitals, including the yearly number of blood components transfused, are given in table 2. Hospitals 1 and 2 have major teaching affiliations and a greater number of residents and medical students. Hospital 1 is the largest center and utilized 67.9% of all health system RBC units, 83.3% of all plasma units, and 89.7% of all platelet units during the 2014 baseline year. Four of the five hospitals have a labor and delivery service.

Table 2.

Characteristics of the Five Hospitals in the Health System

Changes in blood utilization for all five hospitals combined are shown in table 3. Across the health system, comparing the most recent year (2017) to the baseline time period before PBM (2014), the changes in blood utilization are as follows. The percentage of patients transfused decreased for RBCs from 11.3 to 10.4%, for plasma from 2.9 to 2.2%, and for platelets from 3.1 to 2.7%, (P < 0.0001 for all three). The number of units transfused per 1,000 discharged patients decreased for RBCs by 19.8% (P < 0.0001), for plasma by 38.9% (P = 0.0002), and for platelets by 15.6% (P = 0.04). Multiunit RBC transfusion orders decreased by 49% (P < 0.0001), which coincided with our “Why Give Two When One Will Do?” campaign. The percentage of RBC orders with a preceding Hb greater than or equal to 8 g/dL decreased by 34.7% over time (P < 0.0001), while plasma orders with an INR less than 1.5 decreased by 9.1% (P < 0.0001), and platelet orders with a platelet count greater than or equal to 50 K decreased by 3.4% (P = 0.02). Some 41.1% of all plasma (30,212 of 73,564 units) and 49.9% of all platelets (35,342 of 70,770 units) were given to those patients receiving 10 or more RBC units over their hospital stay. For the five-hospital health system, changes in blood utilization over the entire study period are illustrated in figure 3.

Table 3.

Blood Utilization by Year for the Five Health System Hospitals Combined

Fig. 3.

Across the entire health system, for each of the three blood components, changes in utilization (number of units per 1,000 patients) are shown over time. FFP = fresh frozen plasma; PLTS = platelets; RBC = erythrocyte.

Fig. 3.

Across the entire health system, for each of the three blood components, changes in utilization (number of units per 1,000 patients) are shown over time. FFP = fresh frozen plasma; PLTS = platelets; RBC = erythrocyte.

Close modal

Considering only the patients who were transfused each specific blood product, the average number of RBC units per patient given RBCs decreased from 4.0 ± 6.5 to 3.5 ± 4.9 (by 14.6%; P < 0.0001). The average number of plasma units per patient given plasma decreased from 6.1 ± 14.9 to 4.6 ± 13.5 (by 25.1%; P = 0.0014). The average number of platelet units per patient given platelets decreased from 5.1 ± 8.1 to 4.8 ± 8.1 (by 4.9%; P = 0.002).

For each of the five individual hospitals, changes in blood utilization are shown in the Supplemental Digital Content (http://links.lww.com/ALN/B527; tables 1–5). The most notable changes were the decrease in multiunit RBC transfusion orders and RBC utilization in units per patient in all hospitals, except for Hospital 2. Plasma utilization decreased significantly in two of the five hospitals, including the largest hospital (Hospital 1). Platelet utilization decreased in four hospitals (only one with statistical significance), and was unchanged in the fifth hospital. Compared to the baseline year 2014, the annualized blood acquisition cost savings for 2017 was \$2,120,273/yr, representing an approximate 400% return on investment for our PBM efforts.

Using the methods described here and a health system–wide clinical community approach to promote a PBM program, we effectively reduced overall blood utilization and blood acquisition costs. Although methods of establishing PBM programs have been reported previously,15,27,38,39  to our knowledge, this is the first description of methods used to implement a PBM program across multiple hospitals within a health system. Our methods also illustrate how anesthesiologists can be leaders in quality improvement, providing value both inside and outside of the operating rooms.

Although some have described PBM as giving the right dose of the right product to the right patient for the right reason, others believe that PBM is simply avoiding unnecessary transfusions. In reality, good blood management is just the good practice of medicine with regard to preventing and managing anemia, optimizing hemostasis to reduce or prevent hemorrhage, and promoting optimal blood conservation with a patient-centered focus.40–44  PBM is also focused on patient safety. Although transfusion-transmitted viral infections are much less common than they were 20 or 30 yr ago, new emerging infectious agents (e.g., Zika virus and babesiosis) are of great concern.45,46  Furthermore, the top two causes of transfusion-related death—transfusion-related acute lung injury and transfusion-associated circulatory overload—occur at least 100-fold more commonly than HIV or viral hepatitis,47  and more commonly than has been historically reported.48,49

As previously mentioned, the clinical community concept we describe as the structure behind our PBM efforts shares common principles with both the Choosing Wisely and the Perioperative Surgical Home campaigns, which aim to reduce unnecessary procedures by standardizing care and reducing variation in practice. The clinical community provided the infrastructure through which we accomplished these aims. Three critical features of the clinical community are the physician leadership, project management support, and robust data analytics. The Clinical Community Project Management Office (PMO) supports the clinical community with a project manager, data analytics, lean support, and other resources as appropriate to promote quality improvement. For example, even seemingly simple tasks of creating agendas, scheduling meetings, providing call-in numbers, and video conferencing login access, can be challenging without a dedicated support staff.

The other important feature that cannot be overemphasized in our program is the collection and analysis of high-quality data and its dissemination back to the providers to improve practice.16,17  Physicians are a competitive group, and the rank-order bar graph with names promotes a competition among physicians to do better than their peers. There is some resistance, however, to the green, yellow, and red data representation, as clinicians claim their patients are exceptions to guidelines because of active bleeding, symptomatic anemia, or hypovolemia. For this reason, we carefully acknowledge on the reports that our transfusion guidelines apply specifically to nonbleeding, hemodynamically stable patients. In fact, our dashboards have a filter to exclude transfusion episodes when the ordering provider indicates active bleeding, such as those episodes that commonly occur in the operating rooms. In our PBM program, we opted to invest time and money into data collection, analysis, and audits with feedback to providers rather than hiring a full-time transfusion safety officer for each of the five hospitals. We recently counted more than 600 physicians who receive our transfusion guideline compliance reports at regular intervals. Given the approximately 100,000 units of blood components that are transfused annually across the five hospitals, a large-scale data-driven approach with guideline compliance reports sent at regular intervals to providers was thought to be more effective and less expensive than hiring a TSO for each institution to review individual transfusions.

We would like to emphasize that without financial support, a PBM program is unlikely to succeed. This type of work cannot be accomplished on nights and weekends, and the people doing the primary work need salary support for dedicated nonclinical time. To justify the financial support for a PBM program, it is helpful to set targets for decreasing blood use and a return on investment to show the health system leadership that the resources can pay for themselves. This is especially true in light of the poor or nonexistent reimbursement for blood and blood components. Our current return on investment is about 400%, in terms of savings on blood acquisition costs.

One challenge that we recognize is the large number of massively transfused patients seen in an urban tertiary care center. This likely explains why two-thirds of all RBC units and 80 to 90% of all plasma and platelet units in the healthcare system were given at Hospital 1, which had the most massive transfusions. We also have shown that these patients consume a huge percentage of overall blood within the hospital.50  Even with education and the use of CPOE with decision support (BPAs), the ability to alter blood utilization for services such as transplant, vascular, and cardiac surgery is less, especially when patients require major procedures such as extracorporeal membrane oxygenation, left ventricular assist device, and heart and lung transplants.50  Even with these cases, however, autologous blood salvage (cell salvage) thromboelastography, and careful monitoring of Hb targets—not just triggers—are important. For example, we have seen massively transfused patients leave the hospital with a Hb level higher than is necessary. Accordingly, it is just as important to know when to stop giving blood in a massive transfusion as it is to know when to initiate the protocol. We also have recently reported that outcomes can be favorable even after massive transfusions.51  The heterogeneity across hospitals for changes in blood utilization is to be expected, based on differences in case mix, as well as the proportion of patients who require massive transfusions.

Other challenges that we faced in setting up a multihospital PBM program included the differing EHR systems that were in use at the institutions, the transition to new EHR systems, and the ability to maintain “legacy” data from the old EHR system after the transition. Now that we have a complete transition to the Epic EHR across the entire system for ordering all transfusions, our data capabilities are improved. Another challenge was a reliable method of report distribution down to the provider level. We elected to delegate report distribution by using a chain-of-command approach, by sending them to the VPMA, who sends to the department chairs, and then to their departmental attending physician members. This top-down approach worked well because the VPMA has oversight and authority to support such quality improvement efforts. Another issue was how to determine cost avoidance when blood utilization decreased. We chose to use blood acquisition cost primarily, rather than the roughly fourfold higher activity-based cost of blood,9  because we recognized that some, but not all, overhead costs are borne by the hospital. We do, however, recognize that these overhead costs exist and sometimes refer to both methods of cost assessment.

When we compared our PBM results with those from other institutions, for example the results of Goodnough et al.52  at Stanford University Hospital (Stanford, California), our decrease in RBC utilization was less (19.8 vs. 42%). It should be recognized, however, that the baseline degree of RBC overuse was likely higher at Stanford, where 60 to 65% of RBC transfusions had a preceding Hb greater than 8 g/dL in 2008, and they were able to reduce it to 35 to 40% by 2011.40  Our pre-PBM baseline degree of RBC overuse by these same criteria (table 3) was similar to Stanford’s post-PBM overuse percentages. The likely explanation for these differences is that five of the eight randomized controlled trials supporting a restrictive transfusion strategy were published during or after 2011.18–25  With this degree of high-quality evidence supporting PBM practices, blood utilization was likely already curtailed by the time our program began. Rates of massive transfusions and double-unit RBC orders also may have differed between institutions.

A few limitations to our study should be recognized. First, PBM programs are usually introduced in stages, as we described. It is therefore possible that we underestimated the actual blood savings because we had already introduced some degree of PBM to the main campus during what we call the pre-PBM period. It is also difficult to determine the relative contribution of individual blood conservation methods, when some methods began simultaneously (i.e., BPAs and guideline compliance reports). Second, the progress we report is that from a single health system, and results in other centers may be different. The amount of blood savings likely depends on the baseline degree of transfusion overuse, which can vary substantially among centers. Third, most previous reports focus on RBCs, whereas we report all three major blood components. The issue here is that most of the evidence supporting transfusion triggers relies on studies comparing liberal and restrictive RBC transfusion.18–25  No such randomized controlled trial has been published for plasma, and only one, in the setting of oncology, for platelet transfusion.53  Therefore, compared with Hb transfusion thresholds for RBCs, we have less evidence to support the INR and platelet count thresholds associated with adequate hemostasis. Finally, the percentage of plasma and platelets given outside the laboratory value thresholds changed by small amounts, while plasma and platelet utilization decreased substantially. It is possible that active bleeding or massive transfusion was the indication for these transfusions rather than abnormal laboratory values. In fact, almost half of all our plasma and platelet units went to patients given more than 10 RBC units, and the vast majority of these components were given at the main campus, where trauma, transplant, and major cardiac cases are prevalent.

In summary, we describe methods for implementing a comprehensive PBM program across five hospitals in our healthcare system using a clinical community approach. By working together collaboratively across disciplines, sharing best practices, and using high-quality data collection and feedback to promote evidence-based practice, we can reduce unnecessary transfusions and associated risks while decreasing costs. Such changes in healthcare delivery are examples of how we can increase the value of care we provide to our patients.

The authors would like to acknowledge both financial and personnel support from the Johns Hopkins Health System Armstrong Institute for Patient Safety and Quality (Baltimore, Maryland). They would also like to thank Claire F. Levine, M.S., E.L.S. (Scientific Editor, Department of Anesthesiology/Critical Care Medicine, Johns Hopkins Medicine, Baltimore, Maryland), for editorial assistance.

Support was provided solely from institutional and/or departmental sources.

Dr. Frank has received consulting fees from Haemonetics (Braintree, Massachusetts), Medtronic (Minneapolis, Minnesota), and Zimmer/Biomet (Warsaw, Indiana). All other authors declare no competing interests.

1.
Healthcare Cost and Utilization Project: Most frequent procedures performed in U.S. hospitals
.
2013
.
Available at: http://www.hcup-us.ahrq.gov/reports/statbriefs/sb149.pdf Accessed February 19, 2017
2.
Joint Commission Overuse Summit 2012
.
3.
Yazer
MH
,
Waters
JH
:
How do I implement a hospital-based blood management program?
Transfusion
2012
;
52
:
1640
5
4.
Waters
JH
,
Ness
PM
:
Patient blood management: A growing challenge and opportunity.
Transfusion
2011
;
51
:
902
3
5.
Onuoha
OC
,
Arkoosh
VA
,
Fleisher
LA
:
Choosing wisely in anesthesiology: The gap between evidence and practice.
JAMA Intern Med
2014
;
174
:
1391
5
6.
Callum
JL
,
Waters
JH
,
Shaz
BH
,
Sloan
SR
,
Murphy
MF
:
The AABB recommendations for the Choosing Wisely campaign of the American Board of Internal Medicine.
Transfusion
2014
;
54
:
2344
52
7.
Kain
ZN
,
Vakharia
S
,
Garson
L
,
Engwall
S
,
Schwarzkopf
R
,
Gupta
R
,
Cannesson
M
:
The perioperative surgical home as a future perioperative practice model.
Anesth Analg
2014
;
118
:
1126
30
8.
Mechanic
R
,
Tompkins
C
:
Lessons learned preparing for Medicare bundled payments.
N Engl J Med
2012
;
367
:
1873
5
9.
Shander
A
,
Hofmann
A
,
Ozawa
S
,
Theusinger
OM
,
Gombotz
H
,
Spahn
DR
:
Activity-based costs of blood transfusions in surgical patients at four hospitals.
Transfusion
2010
;
50
:
753
65
10.
Aveling
EL
,
Martin
G
,
Armstrong
N
,
Banerjee
J
,
Dixon-Woods
M
:
Quality improvement through clinical communities: Eight lessons for practice.
J Health Organ Manag
2012
;
26
:
158
74
11.
Gould
LJ
,
Wachter
PA
,
Aboumatar
H
,
Blanding
RJ
,
Brotman
DJ
,
Bullard
J
,
Gilmore
MM
,
Golden
SH
,
Howell
E
,
Ishii
L
,
Lee
KH
,
Paul
MG
,
Rotello
LC
,
Satin
AJ
,
Wick
EC
,
Winner
L
,
Zenilman
ME
,
Pronovost
PJ
:
Clinical communities at Johns Hopkins Medicine: An emerging approach to quality improvement.
Jt Comm J Qual Patient Saf
2015
;
41
:
387
95
12.
Dixon-Woods
M
,
Bosk
CL
,
Aveling
EL
,
Goeschel
CA
,
Pronovost
PJ
:
Explaining Michigan: Developing an ex post theory of a quality improvement program.
Milbank Q
2011
;
89
:
167
205
13.
Oliver
JC
,
Griffin
RL
,
Hannon
T
,
Marques
MB
:
The success of our patient blood management program depended on an institution-wide change in transfusion practices.
Transfusion
2014
;
54
(
10 Pt 2
):
2617
24
14.
Szpila
BE
,
Ozrazgat-Baslanti
T
,
Zhang
J
,
Lanz
J
,
Davis
R
,
Rebel
A
,
Vanzant
E
,
Gentile
LF
,
Cuenca
AG
,
Ang
DN
,
Liu
H
,
Lottenberg
L
,
Marker
P
,
Zumberg
M
,
Bihorac
A
,
Moore
FA
,
Brakenridge
S
,
Efron
PA
:
Successful implementation of a packed red blood cell and fresh frozen plasma transfusion protocol in the surgical intensive care unit.
PLoS One
2015
;
10
:
e0126895
15.
Leahy
MF
,
Roberts
H
,
Mukhtar
SA
,
Farmer
S
,
Tovey
J
,
Jewlachow
V
,
Dixon
T
,
Lau
P
,
Ward
M
,
Vodanovich
M
,
Trentino
K
,
Kruger
PC
,
Gallagher
T
,
Koay
A
,
Hofmann
A
,
Semmens
JB
,
Towler
S
;
Western Australian Patient Blood Management Program
:
A pragmatic approach to embedding patient blood management in a tertiary hospital.
Transfusion
2014
;
54
:
1133
45
16.
Frank
SM
,
Resar
LM
,
Rothschild
JA
,
Dackiw
EA
,
Savage
WJ
,
Ness
PM
:
A novel method of data analysis for utilization of red blood cell transfusion.
Transfusion
2013
;
53
:
3052
9
17.
Frank
SM
,
Savage
WJ
,
Rothschild
JA
,
Rivers
RJ
,
Ness
PM
,
Paul
SL
,
Ulatowski
JA
:
Variability in blood and blood component utilization as assessed by an anesthesia information management system.
Anesthesiology
2012
;
117
:
99
106
18.
Hébert
PC
,
Wells
G
,
Blajchman
MA
,
Marshall
J
,
Martin
C
,
Pagliarello
G
,
Tweeddale
M
,
Schweitzer
I
,
Yetisir
E
:
A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. Transfusion Requirements in Critical Care Investigators, Canadian Critical Care Trials Group.
N Engl J Med
1999
;
340
:
409
17
19.
Hajjar
LA
,
Vincent
JL
,
Galas
FR
,
Nakamura
RE
,
Silva
CM
,
Santos
MH
,
Fukushima
J
,
Kalil Filho
R
,
Sierra
DB
,
Lopes
NH
,
T
,
Roquim
AC
,
Sundin
MR
,
Leão
WC
,
Almeida
JP
,
Pomerantzeff
PM
,
Dallan
LO
,
Jatene
FB
,
Stolf
NA
,
Auler
JO
Jr
:
Transfusion requirements after cardiac surgery: The TRACS randomized controlled trial.
JAMA
2010
;
304
:
1559
67
20.
Carson
JL
,
Terrin
ML
,
Noveck
H
,
Sanders
DW
,
Chaitman
BR
,
GG
,
Nemo
G
,
Dragert
K
,
Beaupre
L
,
Hildebrand
K
,
Macaulay
W
,
Lewis
C
,
Cook
DR
,
Dobbin
G
,
Zakriya
KJ
,
Apple
FS
,
Horney
RA
,
Magaziner
J
;
FOCUS Investigators
:
Liberal or restrictive transfusion in high-risk patients after hip surgery.
N Engl J Med
2011
;
365
:
2453
62
21.
Lacroix
J
,
Hébert
PC
,
Hutchison
JS
,
Hume
HA
,
Tucci
M
,
Ducruet
T
,
Gauvin
F
,
Collet
JP
,
Toledano
BJ
,
Robillard
P
,
Joffe
A
,
Biarent
D
,
Meert
K
,
Peters
MJ
;
TRIPICU Investigators; Canadian Critical Care Trials Group; Pediatric Acute Lung Injury and Sepsis Investigators Network
:
Transfusion strategies for patients in pediatric intensive care units.
N Engl J Med
2007
;
356
:
1609
19
22.
Villanueva
C
,
Colomo
A
,
Bosch
A
,
Concepción
M
,
Hernandez-Gea
V
,
Aracil
C
,
Graupera
I
,
Poca
M
,
Alvarez-Urturi
C
,
Gordillo
J
,
Guarner-Argente
C
,
Santaló
M
,
Muñiz
E
,
Guarner
C
:
Transfusion strategies for acute upper gastrointestinal bleeding.
N Engl J Med
2013
;
368
:
11
21
23.
Holst
LB
,
Haase
N
,
Wetterslev
J
,
Wernerman
J
,
Guttormsen
AB
,
Karlsson
S
,
Johansson
PI
,
Aneman
A
,
Vang
ML
,
Winding
R
,
Nebrich
L
,
Nibro
HL
,
Rasmussen
BS
,
Lauridsen
JR
,
Nielsen
JS
,
Oldner
A
,
Pettilä
V
,
Cronhjort
MB
,
Andersen
LH
,
Pedersen
UG
,
Reiter
N
,
Wiis
J
,
White
JO
,
Russell
L
,
Thornberg
KJ
,
Hjortrup
PB
,
Müller
RG
,
Møller
MH
,
Steensen
M
,
Tjäder
I
,
Kilsand
K
,
Odeberg-Wernerman
S
,
Sjøbø
B
,
Bundgaard
H
,
Thyø
MA
,
Lodahl
D
,
Mærkedahl
R
,
Albeck
C
,
Illum
D
,
Kruse
M
,
Winkel
P
,
Perner
A
;
TRISS Trial Group; Scandinavian Critical Care Trials Group
:
Lower versus higher hemoglobin threshold for transfusion in septic shock.
N Engl J Med
2014
;
371
:
1381
91
24.
Murphy
GJ
,
Pike
K
,
Rogers
CA
,
Wordsworth
S
,
Stokes
EA
,
Angelini
GD
,
Reeves
BC
;
TITRe2 Investigators
:
Liberal or restrictive transfusion after cardiac surgery.
N Engl J Med
2015
;
372
:
997
1008
25.
Robertson
CS
,
Hannay
HJ
,
Yamal
JM
,
Gopinath
S
,
Goodman
JC
,
Tilley
BC
,
Baldwin
A
,
Rivera Lara
L
,
Saucedo-Crespo
H
,
Ahmed
O
,
S
,
Ponce
L
,
Cruz-Navarro
J
,
Shahin
H
,
Aisiku
IP
,
Doshi
P
,
A
,
Neipert
L
,
Waguspack
JM
,
Rubin
ML
,
Benoit
JS
,
Swank
P
;
Epo Severe TBI Trial Investigators
:
Effect of erythropoietin and transfusion threshold on neurological recovery after traumatic brain injury: A randomized clinical trial.
JAMA
2014
;
312
:
36
47
26.
Frank
SM
,
Wick
EC
,
Dezern
AE
,
Ness
PM
,
Wasey
JO
,
Pippa
AC
,
Dackiw
E
,
Resar
LM
:
Risk-adjusted clinical outcomes in patients enrolled in a bloodless program.
Transfusion
2014
;
54
(
10 Pt 2
):
2668
77
27.
Goodnough
LT
,
Shieh
L
,
E
,
Cheng
N
,
Khari
P
,
Maggio
P
:
Improved blood utilization using real-time clinical decision support.
Transfusion
2014
;
54
:
1358
65
28.
Yazer
MH
,
Triulzi
DJ
,
Reddy
V
,
Waters
JH
:
Effectiveness of a real-time clinical decision support system for computerized physician order entry of plasma orders.
Transfusion
2013
;
53
:
3120
7
29.
Wintermeyer
TL
,
Liu
J
,
Lee
KH
,
Ness
PM
,
Johnson
DJ
,
Hoffman
NA
,
Wachter
PA
,
Demski
R
,
Frank
SM
:
Interactive dashboards to support a patient blood management program across a multi-institutional healthcare system.
Transfusion
2016
;
56
:
1480
1
30.
Thakkar
RN
,
Lee
KH
,
Ness
PM
,
Wintermeyer
TL
,
Johnson
DJ
,
Liu
E
,
A
,
Knight
AM
,
Wachter
PA
,
Demski
R
,
Frank
SM
:
Relative impact of a patient blood management program on utilization of all three major blood components.
Transfusion
2016
;
56
:
2212
20
31.
Tobian
AA
,
Heddle
NM
,
Wiegmann
TL
,
Carson
JL
:
Red blood cell transfusion: 2016 clinical practice guidelines from AABB.
Transfusion
2016
;
56
:
2627
30
32.
Roback
JD
,
Caldwell
S
,
Carson
J
,
Davenport
R
,
Drew
MJ
,
Eder
A
,
Fung
M
,
Hamilton
M
,
Hess
JR
,
Luban
N
,
Perkins
JG
,
Sachais
BS
,
Shander
A
,
Silverman
T
,
Snyder
E
,
Tormey
C
,
Waters
J
,
Djulbegovic
B
;
American Association for the Study of Liver; American Academy of Pediatrics; United States Army; American Society of Anesthesiology; American Society of Hematology
:
Evidence-based practice guidelines for plasma transfusion.
Transfusion
2010
;
50
:
1227
39
33.
Kaufman
RM
,
Djulbegovic
B
,
Gernsheimer
T
,
Kleinman
S
,
Tinmouth
AT
,
Capocelli
KE
,
Cipolle
MD
,
Cohn
CS
,
Fung
MK
,
Grossman
BJ
,
Mintz
PD
,
O’Malley
BA
,
Sesok-Pizzini
DA
,
Shander
A
,
Stack
GE
,
Webert
KE
,
Weinstein
R
,
Welch
BG
,
Whitman
GJ
,
Wong
EC
,
Tobian
AA
;
American Association of Blood Banks
:
Platelet transfusion: A clinical practice guideline from the AABB.
Ann Intern Med
2015
;
162
:
205
13
34.
Podlasek
SJ
,
Thakkar
RN
,
Rotello
LC
,
Fleury
TA
,
Demski
RJ
,
Ness
PM
,
Frank
SM
:
Implementing a “Why give 2 when 1 will do?” Choosing Wisely campaign.
Transfusion
2016
;
56
:
2164
35.
Thakkar
RN
,
Kim
D
,
Knight
AM
,
Riedel
S
,
Vaidya
D
,
Wright
SM
:
Impact of an educational intervention on the frequency of daily blood test orders for hospitalized patients.
Am J Clin Pathol
2015
;
143
:
393
7
36.
Frank
SM
,
Wasey
JO
,
Dwyer
IM
,
Gokaslan
ZL
,
Ness
PM
,
Kebaish
KM
:
Radiofrequency bipolar hemostatic sealer reduces blood loss, transfusion requirements, and cost for patients undergoing multilevel spinal fusion surgery: A case control study.
J Orthop Surg Res
2014
;
9
:
50
37.
Grant
MC
,
Resar
LM
,
Frank
SM
:
The efficacy and utility of acute normovolemic hemodilution.
Anesth Analg
2015
;
121
:
1412
4
38.
Abbett
SK
,
Kaufman
RM
,
Gustafson
M
,
Greenberg
JO
:
Multifaceted approach to reducing unnecessary red blood cell utilization.
Healthc (Amst)
2015
;
3
:
67
73
39.
Goodnough
LT
,
Maggio
P
,
E
,
Shieh
L
,
Hernandez-Boussard
T
,
Khari
P
,
Shah
N
:
Restrictive blood transfusion practices are associated with improved patient outcomes.
Transfusion
2014
;
54
(
10 Pt 2
):
2753
9
40.
Goodnough
LT
,
Shander
A
:
Patient blood management.
Anesthesiology
2012
;
116
:
1367
76
41.
Goodnough
LT
,
Shander
A
:
Blood management.
Arch Pathol Lab Med
2007
;
131
:
695
701
42.
Shander
A
:
Preoperative anemia and its management.
Transfus Apher Sci
2014
;
50
:
13
5
43.
Goodnough
LT
:
Blood management: Transfusion medicine comes of age.
Lancet
2013
;
381
:
1791
2
44.
Goodnough
LT
,
Shander
A
,
Brecher
ME
:
Transfusion medicine: Looking to the future.
Lancet
2003
;
361
:
161
9
45.
Dodd
RY
:
Emerging pathogens and their implications for the blood supply and transfusion transmitted infections.
Br J Haematol
2012
;
159
:
135
42
46.
Allain
JP
,
Bianco
C
,
Blajchman
MA
,
Brecher
ME
,
Busch
M
,
Leiby
D
,
Lin
L
,
Stramer
S
:
Protecting the blood supply from emerging pathogens: The role of pathogen inactivation.
Transfus Med Rev
2005
;
19
:
110
26
47.
Carson
JL
,
Grossman
BJ
,
Kleinman
S
,
Tinmouth
AT
,
Marques
MB
,
Fung
MK
,
Holcomb
JB
,
Illoh
O
,
Kaplan
LJ
,
Katz
LM
,
Rao
SV
,
Roback
JD
,
Shander
A
,
Tobian
AA
,
Weinstein
R
,
Swinton McLaughlin
LG
,
Djulbegovic
B
;
Clinical Transfusion Medicine Committee of the AABB
:
Red blood cell transfusion: A clinical practice guideline from the AABB.
Ann Intern Med
2012
;
157
:
49
58
48.
Clifford
L
,
Jia
Q
,
Subramanian
A
,
H
,
Wilson
GA
,
Murphy
SP
,
Pathak
J
,
Schroeder
DR
,
Kor
DJ
:
Characterizing the epidemiology of postoperative transfusion-related acute lung injury.
Anesthesiology
2015
;
122
:
12
20
49.
Clifford
L
,
Jia
Q
,
H
,
Subramanian
A
,
Wilson
GA
,
Murphy
SP
,
Pathak
J
,
Schroeder
DR
,
Ereth
MH
,
Kor
DJ
:
Characterizing the epidemiology of perioperative transfusion-associated circulatory overload.
Anesthesiology
2015
;
122
:
21
8
50.
Zuckerberg
GS
,
Scott
AV
,
Wasey
JO
,
Wick
EC
,
Pawlik
TM
,
Ness
PM
,
Patel
ND
,
Resar
LM
,
Frank
SM
:
Efficacy of education followed by computerized provider order entry with clinician decision support to reduce red blood cell utilization.
Transfusion
2015
;
55
:
1628
36
51.
Johnson
DJ
,
Scott
AV
,
Barodka
VM
,
Park
S
,
Wasey
JO
,
Ness
PM
,
T
,
Frank
SM
:
Morbidity and mortality after high-dose transfusion.
Anesthesiology
2016
;
124
:
387
95
52.
Goodnough
LT
,
Baker
AS
,
Shah
N
:
How I use clinical decision support to improve red blood cell utilization.
Transfusion
2016
;
56
:
2406
11
53.
Slichter
SJ
,
Kaufman
RM
,
Assmann
SF
,
McCullough
J
,
Triulzi
DJ
,
Strauss
RG
,
Gernsheimer
TB
,
Ness
PM
,
Brecher
ME
,
Josephson
CD
,
Konkle
BA
,
Woodson
RD
,
Ortel
TL
,
Hillyer
CD
,
Skerrett
DL
,
McCrae
KR
,
Sloan
SR
,
Uhl
L
,
George
JN
,
Aquino
VM
,
Manno
CS
,
McFarland
JG
,
Hess
JR
,
Leissinger
C
,
Granger
S
:
Dose of prophylactic platelet transfusions and prevention of hemorrhage.
N Engl J Med
2010
;
362
:
600
13