Approximately 15% of children admitted to a pediatric intensive care unit (PICU) receive packed red blood cell (pRBC) transfusions during their admission.1–3 Literature published in recent decades has shown that a restrictive pRBC transfusion threshold, defined as a hemoglobin less than 7 g/dl, is safe for most critically ill pediatric patients4,5 and that pRBC transfusions are usually not necessary when the hemoglobin is greater than 7 g/dL.6–9 Further research in pediatric populations has shown that restrictive transfusion thresholds are appropriate for critically ill patients in septic shock after stabilization of the patient, critically ill children admitted to the PICU after a general surgery, and pediatric patients with respiratory conditions including acute respiratory distress syndrome (ARDS).4,5,9 In addition, there are associations of decreased morbidity and mortality in certain critically ill populations,10–12 including pediatrics,4,6,13 when a restrictive transfusion threshold is used. However, a recent survey of pediatric critical care physicians reported that a restrictive transfusion strategy was adopted only by a fraction of respondents: 55.7% of respondents adopted restrictive transfusion strategies for bronchiolitis, 8.3% for septic shock, 38.1% for trauma, and 16% for tetralogy of Fallot repair.14
In adult studies, restrictive transfusion strategies have reduced unnecessary pRBC transfusions by up to 39%.15 However, compliance with restrictive transfusion guidelines has been challenging and often inadequate.16 Successful interventions to improve adherence to guidelines include computerized decision support with institutional protocols,17–20 audits followed by educational interventions,21 monthly educational meetings,22 and blood management programs.23 Interventions that improve adherence to restrictive transfusion strategies in the adult population can result in significant cost reductions ranging from $120,000 to $700,000 annual cost savings.19,20,24 Similar improvements made in the PICU could decrease the frequency of transfusions and improve resource utilization. However, there has been limited research in the pediatric population on effective interventions to improve adherence to restrictive transfusion guidelines, and none to date have used a multi-industry innovation approach.
The objective of this study was to determine the pRBC transfusion rates in our PICU, assess compliance with current evidence-based transfusion guidelines, understand patient-level variables that affect transfusion practices, and use a unique cross-industry innovation platform to implement a practice strategy addressing compliance challenges faced previously by more traditional problem-solving approaches.
This was a pre–post study of all patients admitted to a 40-bed PICU at an urban academic medical center. The study was approved by our institutional review board (2015-601). Baseline data were collected for 12 months on pRBC transfusions (September 2014 through August 2015). Data collection was performed retrospectively through a chart review and included the percentage of patients transfused, transfusion events, units transfused with each transfusion event, blood bank interventions such as irradiation or washing of blood, and patient demographics including sex, age at the time of transfusion, reason for the transfusion, most recent hemoglobin before the transfusion, length of stay (LOS) at the time of transfusion, PICU LOS, and mortality before PICU discharge/transfer.
Once baseline data were collected and analyzed, the study investigators organized a multi-industry innovation platform to generate novel solutions to improve adherence to transfusion guidelines. This innovative platform has been used as an alternative to traditional internal problem-solving approaches at our institution to generate novel solutions to challenging problems.25 Approximately 25 participants participated in the platform and were recruited from communities outside healthcare, representing a wide array of industries, including engineering, law, automotive, marketing, and arts. To foster new perspectives and collaboration from those who may address this problem differently than a healthcare provider commonly would, less than 10% of the participants invited to join this multidisciplinary group were from the healthcare industry. The innovation platform generated more than 125 ideas to decrease unnecessary blood transfusions and conserve blood products. Distillation and prioritization of the generated proposed solutions by the study investigators led to the development of an awareness campaign to effect ordering behavior. The campaign primarily focused around the slogan “7 is the New 8,” with a logo displaying a pint of blood on a scale, designed during the innovation event. Education and marketing strategies were developed to widely disseminate the campaign, including clothing worn by nurses and physicians in the PICU, magnets on computer workstations, and educational sessions targeting both nurses and physicians. The educational efforts included presentations at nursing, fellow, and faculty conferences about the initiative and supporting research, along with emails and posted flyers around the PICU reminding staff of the initiative. Nurses were encouraged to initiate discussions with physicians if they noticed transfusions ordered when the hemoglobin was >7 g/dl. Transfusion practices were monitored for 12 months after the intervention of the campaign (May 2016 through April 2017), with data collection consisting of the same variables from the preintervention period.
Patient demographics and the primary reason for transfusion were summarized using descriptive statistics. Linear mixed models were run using a random effect of subject IDs to control for multiple observations. Linear mixed models were also run to create adjusted models for preintervention to postintervention changes. A first-order autoregressive correlation structure was used. A p value < .05 was considered statistically significant. Odds ratios (ORs) and confidence intervals were calculated. Data were analyzed in SAS 9.4 (SAS Institute Inc, Cary, NC).
There were 1,543 admissions to the PICU before the intervention, with 14.3% (220/1,543) of admissions receiving a pRBC transfusion. There were 1,566 admissions in the postintervention period, with 15.3% (239/1,566) of admissions transfused. Nineteen transfusion events for 11 patients were excluded from the study analysis; events were excluded if the patient was on extracorporeal membrane oxygenation (ECMO) at the time of the transfusion, in the operating room at the time of transfusion, or if the pRBC transfusion had been ordered by a nonintensivist (such as a surgeon ordering blood on-hold to the operating room). This resulted in a total of 517 pRBC transfusion events for 192 patients before the intervention (56.3% male) and 492 transfusion events for 199 patients after the intervention (55.3% male), for a total of 1,009 transfusion events for 382 unique patients included in the study. Nine patients received transfusions in both the pre- and postintervention periods.
The median age of each patient for the 517 transfusions in the preintervention period was 7.6 years (interquartile range [IQR] 1.7–13.3 years), and for the postintervention period, it was 3.3 years (IQR 1.2–15.0 years). In the preintervention group, median PICU stay at the time of transfusion was 4 days (IQR 1–18 days); median PICU stay at the time of transfusion in the postintervention group was 5 days (IQR 1–17 days). Median LOS in the PICU in the preintervention group was 16 days (IQR 5–79 days), whereas median LOS in the postintervention group was 19 days (IQR 6–48 days). There were no significant differences in age, PICU hospital stay at the time of transfusion, or LOS between the pre- and postintervention groups. The most common reason for transfusion in the preintervention period was an oncologic diagnosis, whereas the most common reason after the intervention was respiratory failure. Approximately 17% of the patients who were transfused died during each respective study period (Table 1).
In the preintervention group, 41.4% of transfusion events (214/517) had a pretransfusion hemoglobin less than or equal to 7.0 g/dl, with 41.4% (214/517) transfusions occurring for a hemoglobin between 7 and 8 g/dl and 17.2% (89/517) transfusions occurring when the hemoglobin was greater than 8.0 g/dl. In the postintervention group, 53.3% of transfusions (262/492) occurred with a hemoglobin less than or equal to 7.0 g/dl, with 33.9% (167/492) of transfusions occurring for a hemoglobin between 7 and 8 g/dl (167/492) and 12.8% (63/492) of transfusions occurring for a hemoglobin greater than 8.0 g/dl. When comparing preintervention with postintervention, transfusions were more likely to adhere to a restrictive strategy (hemoglobin less than or equal to 7.0 g/dl) in the postintervention group (OR 1.66, 95% confidence interval [CI] 1.21–2.28).
Adjusted models assessing intervention to postintervention changes for transfusion-related variables are given. Restrictive transfusions remained more likely to occur in the postintervention group compared with the preintervention group after adjusting for age (OR 1.66, 95% CI 1.21–2.27), PICU LOS (OR 1.70, 95% CI 1.24–2.34), overall mortality (OR 1.70, 95% CI 1.25–2.33), mortality within 24 hours of admission (OR 1.67, 95% CI 1.21–2.29), and mortality within 48 hours of admission (OR 1.69, 95% CI 1.23–2.32) (Table 2). There were no statistically significant differences in adherence to restrictive transfusions when age groups were compared with a reference age group. Patients with a PICU LOS greater than or equal to 20 days were less likely to be adherent to a restrictive transfusion threshold compared with patients with an LOS less than 5 days (OR 0.66, CI 0.45–0.97). Patients who died during their PICU admission were less likely to receive a pRBC transfusion within the parameters of the restrictive transfusion guidelines compared with patients who had not died (OR 0.49, 95% CI 0.3–0.71). Patients who died within 24 or 48 hours of receiving a transfusion were not found to have a statistically significant difference in adherence to restrictive transfusion guidelines (Table 2).
The interventions led to improvements in adherence to a restrictive transfusion threshold for nearly all medical and surgical problems, with the exception of patients with a primary problem of uncomplicated anemia, cardiac arrest, or among patients transfused for postoperative indications. Overall, the highest adherence to restrictive transfusion thresholds was seen in patients with the primary problem of uncomplicated anemia, followed by patients with liver disease, gastrointestinal disease, oncologic disease, and shock. The largest improvement in adherence was in the oncology population, with 29% of patients meeting restrictive transfusion guidelines before the intervention and 56% of patients meeting transfusion guidelines after the intervention (Table 3).
There was a significant difference between the 903 pRBC units transfused before the intervention (mean units per transfusion event 1.75, SD 1.58) and 722 units transfused after the intervention (mean units 1.47, SD 0.92, p = .005). Patients in the preintervention period were transfused an average of 2.35 times per PICU admission (SD 4.28), whereas patients in the postintervention period were transfused an average of 2.06 times (SD 2.59, p = .38). In sum, patients received an average of 4.1 units per PICU admission before the intervention (SD 9.57) and an average of 3.02 units per PICU admission after the intervention (SD 5.04, p = .13). There were fewer irradiated units transfused after the intervention compared with before the intervention (320 irradiated after the intervention compared with 364 irradiated before the intervention).
Using these data, financial costs were compared in the pre- and postintervention periods. The total cost of blood transfusions, inclusive of operating costs, in the preintervention period was approximately $255,594, whereas the total cost in the postintervention period was $239,897. This demonstrates a cost savings of $15,697 annually or a 6% cost reduction in blood transfusions over 1 year, significant for a very low-cost intervention.
There are several limitations to this study. Although the pre- and postintervention periods did account for 12 months of data, it is possible that the illness severity differed between the two groups, which may have affected pRBC ordering practices. In addition, some months in both the pre- and postintervention periods included patients who received more than 10 transfusions; these patients with high transfusion burdens may have confounded the analysis, although with the mixed models analysis, this limitation should be minimized. Next, this was a single-center study; the significant improvements seen using a campaign strategy may not be generalizable to other institutions. This study focused only on intensive care unit (ICU) patients, although we suspect that a similar intervention on other units in our hospital would result in improvement. Given that the data were collected through a chart review, there exists a possibility that blood transfusion events may have been missed. In addition, the reason for each blood transfusion may not have been accurately identified because we did not have knowledge of why providers ordered a blood transfusion. Next, the cost savings are a general estimate of potential savings, but likely underestimate actual savings; although these costs include the estimated labor, time, and sunk cost of equipment, they do not account for other nonfinancial cost savings, such as blood being a limited resource, which may be the most valuable cost to some, or the potential for decreasing adverse reactions or infection risk. Last, the pre–post study design cannot definitively show that the improvement in adherence was due solely to the study interventions.
This study shows that an awareness campaign, developed from a multi-industry innovation platform, can lead to a statistically significant increase in adherence to restrictive pRBC transfusion guidelines and a reduction in the number of blood products transfused for critically ill pediatric patients. Although the total number of blood transfusion events was not significantly different between the pre- and post-intervention groups, it is possible that transfusions were avoided in the postintervention period because of improved adherence to guidelines. In addition, fewer pRBC units were transfused in the postintervention period, which may be due to the intervention's effect on heightened provider awareness regarding safe levels of hemoglobin in critically ill children.
The results showed that patients who died within 48 hours of a pRBC transfusion were more likely to receive a transfusion outside of the guidelines, which may be related to the patient's severity of illness, as well as more liberal physician decision making during these cases. In addition, the interventions led to improvements in adherence for nearly all diagnostic subgroups, with large improvements seen in the oncology population. These subgroup analyses provide direction for future interventions, including targeting patients at high risk of mortality and patients presenting with anemia in the setting of cardiac arrest, oncologic diagnoses, and those in the postoperative period.
This marginal modification of behavior change was the result of deliberately bringing together individuals from different backgrounds and industries, with vastly varying perspectives, to address a common problem in ways we had not previously, or likely would have, considered. The positive outcomes associated with this study were made possible not only from ideation of an innovative strategy to improve adherence but also through empowering all members of the patient care team to uphold restrictive transfusion guidelines; this unexpected secondary value of the intervention is likely a reflection of the social and consumer science perspectives of those engaged in this platform. In our awareness campaign and through educational interventions, nurses were encouraged to remind physicians of the slogan when blood products were being ordered for a hemoglobin greater than 7 g/dl. Although this did occur anecdotally, we did not measure how often nurses advocated for adherence to restrictive blood transfusion guidelines. Previous research has shown that nursing role in transfusion decision making has been effective in improving adherence to these guidelines.16 Given the multi-disciplinary nature of clinical providers in the PICU, the intervention had a broad exposure among training levels, from medical students to attendings, all becoming proponents of the new transfusion practice. The wide geographic dissemination of the campaign throughout the PICU also generated awareness among families, who began to advocate for their child if a transfusion was being ordered outside of the publicized restrictive transfusion range. In addition, as consultants and rotating students and trainees circulate in the PICU, the campaign was disseminated indirectly to other floors and units of the hospital, leading to physicians outside the PICU adopting restrictive transfusion guidelines; this has spurred discussion of policy change on transfusion practices throughout the hospital and a meaningful cultural change with regards to transfusion triggers for stable hospitalized pediatric patients.
Although our results demonstrate an improvement in adherence to pRBC transfusion guidelines, there are other strategies we can adopt to uphold restrictive transfusion guidelines. The marketing and educational campaign created from our multi-industry innovation platform was the primary intervention for this study. Other ideas from the platform included innovative point-of-care hemoglobin testing, strategies to limit blood draws and develop feasible noninvasive testing, real-time electronic medical record notifications, and technology and incentives to encourage blood donation; these are novel ideas not previously generated through our usual problem-solving forums held internally. In addition, strategies in the adult ICUs, such as transfusion bundles, have been shown to reduce inappropriate red blood cell transfusions.26 A similar bundle could be developed and tested in the pediatric population. Future marketing strategies in our hospital include posters or signage in staff areas such as break rooms and elevators, and future educational initiatives might aim to target other front-line providers such as resident physicians and non-PICU physicians who consult in the ICU.
In conclusion, innovation stemming from a pediatric-focused multi-industry forum, including ideation from those outside of healthcare, led to an improved adherence to restrictive transfusion guidelines in a PICU setting. An awareness campaign can improve adherence to guidelines and lead to cost savings and resource conservation. Future interventions include strategies to avoid notification fatigue, behavioral campaign integration into clinical decision support, and expansion of the interventions to hospital units outside of the PICU.
This study shows that simple, unique interventions such as a behavioral modification marketing campaign augmented by educational sessions can lead to improvements in critical standardized care paradigms, such as in this case, adherence to pRBC transfusion guidelines. These interventions are generalizable to other institutions and to other issues that require behavior change to improve quality and care delivery outcomes. Future research may focus on other successful behavioral modification interventions using multi-disciplinary cross-pollination for quality adherence and improvement in a busy hospital setting.
The authors thank the nursing staff working at the PICU of Ann & Robert H. Lurie Children's Hospital of Chicago for their commitment to this strategy and all members of the innovation platform who made this intervention possible.
1. Kleiber N, Lefebvre E, Gauvin F, et al. Respiratory dysfunction associated with RBC transfusion in critically ill children: A prospective cohort study. Pediatr Crit Care Med. 2015;16(4):325–334.
2. Demaret P, Tucci M, Karam O, Trottier H, Ducruet T, Lacroix J. Clinical outcomes associated with RBC transfusions in critically ill children: A 1-year prospective study. Pediatr Crit Care Med. 2015;16(6):505–514.
3. Armano R, Gauvin F, Ducruet T, Lacroix J. Determinants of red blood cell transfusions in a pediatric critical care unit: A prospective, descriptive epidemiological study. Crit Care Med. 2005;33(11):2637–2644.
4. Lacroix J, Hebert PC, Hutchison JS, et al. Transfusion strategies for patients in pediatric intensive care units. New Engl J Med. 2007;356(16):1609–1619.
5. Valentine SL, Bembea MM, Muszynski JA, et al. Consensus recommendations for RBC transfusion practice in critically ill children from the pediatric critical care transfusion and anemia expertise initiative. Pediatr Crit Care Med. 2018;19(9):884–898.
6. Karam O, Tucci M, Ducruet T, et al. Red blood cell transfusion thresholds in pediatric patients with sepsis. Pediatr Crit Care Med. 2011;12(5):512–518.
7. Tyrrell CT, Bateman ST. Critically ill children: To transfuse or not to transfuse packed red blood cells, that is the question. Pediatr Crit Care Med. 2012;13(2):204–209.
8. Parker RI. Transfusion in critically ill children: Indications, risks, and challenges. Crit Care Med. 2014;42(3):675–690.
9. Lacroix J, Tucci M, Pont-Thibodeau GD. Red blood cell transfusion decision making in critically ill children. Curr Opin Pediatr. 2015;27(3):286–291.
10. Hebert PC, Wells G, Martin C, et al. Variation in red cell transfusion practice in the intensive care unit: A multicentre cohort study. Crit Care. 1999;3(2):57–63.
11. Walsh TS, Boyd JA, Watson D, et al. Restrictive versus liberal transfusion strategies for older mechanically ventilated critically ill patients: A randomized pilot trial. Crit Care Med. 2013;41(10):2354–2363.
12. Vincent JL, Baron JF, Reinhart K, et al. Anemia and blood transfusion in critically ill patients. JAMA. 2002;288(12):1499–1507.
13. Akyildiz B, Ulgen Tekerek N, Pamukcu O, et al. Comprehensive analysis of liberal and restrictive transfusion
strategies in pediatric intensive care unit. J Trop Pediatr. 2018;64:118–125.
14. Du Pont-Thibodeau G, Tucci M, Ducruet T, Lacroix J. Survey on stated transfusion practices in PICUs. Pediatr Crit Care Med. 2014;15(5):409–416.
15. Carson JL, Carless PA, Hebert PC. Transfusion thresholds and other strategies for guiding allogeneic red blood cell transfusion. Cochrane Database Syst Rev. 2012;4:CD002042.
16. Vlaar AP, In der Maur AL, Binnekade JM, Schultz MJ, Juffermans NP. Determinants of transfusion decisions in a mixed medical-surgical intensive care unit: A prospective cohort study. Blood Transfus. 2009;7(2):106–110.
17. Rana R, Afessa B, Keegan MT, et al. Evidence-based red cell transfusion in the critically ill: Quality improvement
using computerized physician order entry. Crit Care Med. 2006;34(7):1892–1897.
18. Rothschild JM, McGurk S, Honour M, et al. Assessment of education and computerized decision support interventions for improving transfusion practice. Transfusion. 2007;47(2):228–239.
19. Corson AH, Fan VS, White T, et al. A multifaceted hospitalist quality improvement
intervention: Decreased frequency of common labs. J Hosp Med. 2015;10:390–395.
20. Saag HS, Lajam CM, Jones S, et al. Reducing liberal red blood cell transfusions at an academic medical center. Transfusion. 2017;57(4):959–964.
21. Petaja J, Andersson S, Syrjala M. A simple automatized audit system for following and managing practices of platelet and plasma transfusions in a neonatal intensive care unit. Transfus Med. 2004;14(4):281–288.
22. Brandt MM, Rubinfeld I, Jordan J, Trivedi D, Horst HM. Transfusion insurgency: Practice change through education and evidence-based recommendations. Am J Surg. 2009;197(3):279–283.
23. Kumar A, Figueroa PI, Gowans KL, et al. An evolution in blood management: Past, present, and future. Qual Manag Health Care. 2011;20(4):311–321.
24. Neri RA, Mason CE, Demko LA. Application of six sigma/CAP methodology: Controlling blood-product utilization and costs. J Healthc Manag. 2008;53(3):183–195;discussion 195-186.
25. Malakooti MR. Igniting innovation
in healthcare: A novel collaborative multidisciplinary approach in a PICU. Pediatr Crit Care Med. 2018;19(7):e374–e377.
26. Borgert M, Binnekade J, Paulus F, et al. Implementation of a transfusion bundle reduces inappropriate red blood cell transfusions in intensive care: A before and after study. Transfus Med. 2016;26(6):432–439.
Colleen M. Badke, MD, MPH, is a pediatric critical care medicine fellow at Ann & Robert H. Lurie Children's Hospital of Chicago in Chicago, IL. She participates in IGNITE Innovation and is involved in several quality improvement initiatives.
Julie A. Borrowman, MSN, RN, CPN, is the Manager of Patient Care Operations in the PICU at Ann & Robert H. Lurie Children's Hospital of Chicago in Chicago, IL. She participates in IGNITE Innovation, is the Co-Chair of House-wide Emergency Response Team, co-leads the PICU Code and Critical Assessment Team, advises the PICU Nursing Quality Council, and co-leads the PICU Multi-disciplinary Quality Committee.
Shannon Haymond, PhD, is the Vice Chair for Computational Pathology at Ann & Robert H. Lurie Children's Hospital of Chicago in Chicago, IL. She is an Associate Professor of Pathology at Northwestern University Feinberg School of Medicine. She directs clinical laboratories and is engaged in the development of programs to aid the effective and efficient use of laboratory tests.
Karen Rychlik, MS, is a senior biostatistician at the Stanley Manne Children's Research Institute's Biostatistics Research Core in Chicago, IL. She supports research at Ann & Robert H. Lurie Children's Hospital of Chicago, including IGNITE Innovation.
Marcelo R. Malakooti, MD, is an assistant professor of pediatrics in critical care medicine at Ann & Robert H. Lurie Children's Hospital of Chicago in Chicago, IL. He is medical director of the Lefkofsky Pediatric Intensive Care Unit, director of innovation and strategy for the PICU, and he initiated and leads IGNITE Innovation which develops solutions for operations, quality, and care delivery through multidisciplinary cross-pollination.