Postpartum hemorrhage protocols and benchmarks: improving care through standardization

Abstract

Postpartum hemorrhage remains a leading cause of maternal morbidity and mortality in the United States. Several state maternal morbidity and mortality committees have reviewed areas of opportunity concerning postpartum hemorrhage management and found that common patterns include delays in recognition and response to hemorrhage. Hospital systems and state perinatal quality collaboratives have found that comprehensive, interdisciplinary response to postpartum hemorrhage care improves patient outcomes and, in some instances, reduces racial disparities. A key component of this focus involves the implementation of stage-based hemorrhage protocols for postpartum hemorrhage management. Stage-based hemorrhage protocols are designed to reduce delays in the diagnosis and management and avoid the pitfalls of cognitive biases. These protocols are complex, and their effectiveness is tied to the quality of their implementation. Systematic benchmarking and development of quality metrics for adherence to postpartum hemorrhage bundles would be expected to improve clinical outcomes, but evidence regarding the effectiveness of this practice in the literature is limited. Here, key features of stage-based interventions and evidence regarding the use of quality metrics for postpartum hemorrhage protocol adherence have been outlined.

Evidence for postpartum hemorrhage management protocols as systems issues

Postpartum hemorrhage (PPH) management involves a team of clinicians working in a coordinated fashion to obtain hemostasis and manage the effects of acute blood loss. Thus, when opportunities in outcomes after PPH are found, they are often related to systems-level factors.^1,2 For example, a PPH event might raise any of the following potential quality issues:

1. Access: Has the patient had the opportunity to seek timely prenatal care to have risk factors optimized (eg, identification and treatment of peripartum anemia)? Did they deliver in a nonhospital setting (eg, home) because of the lack of access to hospital care, resulting in delayed management?

2. Structure: Does the hospital have the equipment in place to address hemorrhage, such as the provision of standardized hemorrhage carts and ready access to uterotonics? Are hospital teams prepared to respond to PPH via participation in team training and simulations for PPH management? Does the hospital have a standardized stage-based management protocol that teams have been trained to follow?

3. Process: How is blood loss measured and communicated to team members during acute hemorrhage?

Several state maternal morbidity and mortality committees have reviewed areas of opportunity concerning PPH management and found that common patterns include delays in recognition and response to hemorrhage.^1,2 These findings were consistent with previous work demonstrating that preventable maternal mortality is often a function of delays in seeking care, arriving at a health facility, and provision of adequate care.⁴

Delays in postpartum hemorrhage recognition and postpartum hemorrhage response

Providers find it difficult to precisely estimate blood loss. Obstetrical literature indicates certain patterns exist, including overestimation in low-volume blood loss events, and underestimation in higher blood loss deliveries (ie, provider estimates are biased toward the median blood loss). Underestimation of blood loss during PPH can result in delays in the diagnosis and management, which have been associated with maternal mortalities.¹ Quantitative blood loss (QBL) has been proposed as an alternative approach intended to reduce these cognitive biases. QBL can be performed using traditional gravimetric and volumetric techniques, such as graduated under buttock drapes, and weighing of laparotomy pads and irrigation canisters; it can also be performed using newer techniques, such as adding real-time imaging of laparotomy pads and irrigation canisters.⁵ These techniques have been shown to be more accurate than estimated blood loss compared with “gold standard methodologies,” such as hemoglobin change, but disappointingly have not yet been shown to alter clinical outcomes.⁶ There are several potential pitfalls for QBL in practice, which may limit its effectiveness. Practitioners can focus on the attainment of an “accurate” QBL value at the expense of a timely QBL value, resulting in time delays in the recognition of hemorrhage. QBL that does not facilitate early recognition of the onset or severity of hemorrhage is not useful. Providers may resist the activation of PPH protocols until the blood loss threshold is “proven” by QBL, and the effort to measure and maintain QBL may distract the attention from other aspects of acute PPH management; thus, if not well implemented, QBL could paradoxically delay rather than accelerate PPH response. Patients may have very different responses to a given level of QBL, based on their baseline hemoglobin, height, and medical comorbidities. Finally, QBL is a diagnostic intervention, and diagnostic interventions can only improve clinical outcomes when paired with effective therapeutics.⁷

Data demonstrate an association between timely treatment of recognized PPH and outcomes. In a large Canadian cohort of vaginal deliveries, each 5-minute additional delay in the administration of a second uterotonic (after Pitocin) from recognition of PPH was associated with a 31% higher odds of hypotension and a 34% higher odds of transfusion. Similar associations were found in a French cohort between delay in oxytocin administration and manual uterine cavity examination and rates of severe PPH⁸ and in a joint Beninese and Malian cohort between oxytocin administration times and severe PPH rates. Among second-line interventions, early use of an intrauterine tamponade balloon seems to reduce the risk of needing subsequent procedures.⁹ Concerning adjunctive interventions, the landmark WOMAN trial (World Maternal Antifibrinolytic) for PPH treatment demonstrated that tranexamic acid (TXA) given early in the course of PPH was more effective in reducing the risk of PPH-related maternal morbidity or mortality.¹⁰ Therefore, further evaluation of blood loss recognition strategies that specifically lead to earlier management of hemorrhage is important. This may include creating validated quality measures that focus on the connection between recognition and response to hemorrhage and not just the process of blood loss estimation.

Cognitive biases

Several forms of cognitive bias have been described in the literature.¹¹ In the management of PPH, many of these can adversely influence patient care and result in deviations from optimal medical and team-based management. Anchoring bias can result in providers becoming committed to 1 etiology for PPH (eg, atony), which can delay a search for other potential etiologies (eg, laceration or retained products of conception). Normalcy bias leads people to disbelieve or minimize threat warnings and therefore fail to recognize signs of danger as being abnormal (such as bleeding volume is “normal” or tachycardia is related to “anxiety”). Availability bias reflects the human tendency to overweight recent experiences when evaluating a new situation (eg, a recent case of amniotic fluid embolism may cause a provider to suspect amniotic fluid embolism in subsequent patients more commonly, even when other explanations for hemorrhage may be more likely). Implicit biases occur when attitudes and stereotypes affect our understanding, actions, and decisions in an unconscious manner (unconscious discrimination). Implicit bias is a well-recognized cause of healthcare inequities leading to racial and ethnic disparities in health outcomes and is suspected that this bias may contribute to differences in response to PPH resulting in the known disparity in severe maternal morbidity (SMM) rates related to PPH.¹² Many organizations, researchers, and advocacy groups have published about the effects of implicit bias and recommended strategies, such as implicit bias training, development of respectful care models, and including people with lived experiences in quality improvement teams. However, there is less evidence-based literature about other cognitive biases, including strategies to help healthcare teams identify during the case review process when biases result in patient adverse outcomes and how to develop corrective action strategies to reduce them.

Postpartum hemorrhage protocols improve systems of care but are dependent on implementation

PPH protocols are intended to standardize the recognition and management of PPH (both for obtaining hemostasis and for resuscitating the patient), reduce cognitive biases by standardizing response, and help teams have a common understanding of the pathway to stop hemorrhage. We noted that PPH protocols are only 1 component of larger PPH bundles. By “hemorrhage protocols,” we referred to the stage-based process by which PPH risk is assessed and appropriate peripartum anticipatory and prophylactic actions are taken and guides the management of patients who do experience a PPH. Using structured blood loss measurement techniques and objective vital signs threshold to define PPH and stage its severity is designed to reduce delays in diagnosis, although recommendations to implement interventions at different thresholds of blood loss (such as additional uterotonics) are intended to reduce delays in management. In addition, PPH protocols provide structured guidance on considering a wide differential for PPH, which is designed to reduce the risk of providers developing anchoring bias and similar cognitive biases. Because PPH protocols emphasize the importance of objective measures of blood loss and vital signs in identifying and staging PPH and prescribe a recommended response at each stage of hemorrhage, the application of PPH protocols may minimize implicit biases.

However, evidence indicates that the mere presence of a stage-based protocol may not reduce the PPH rate or rate of SMM related to PPH.¹³ The proponents of PPH bundles argue that this is related to the lack of protocol implementation, also known as the “implementation gap,” in other conditions.^14,15 Hospital systems and state perinatal quality collaboratives (PQCs) that have focused on implementation have shown a reduction in morbidity in patients experiencing PPH. Investigators in California compared the rates of severe hemorrhage-related morbidity among deliveries in hospitals preceding and following the implementation of a state PQC focused on PPH and found that participation in the collaborative was associated with a 20.8% reduction in hemorrhage-related morbidity.^3,16 Although the California Maternal Quality Care Collaborative provides extensive guidance on PPH management, local leaders customized their protocols to reflect facility-specific needs and opportunities.^16,17 As an example, a large health system (Dignity Health, a California-based not-for-profit public-benefit corporation that operates hospitals and ancillary care facilities in 3 states—California, Arizona, and Nevada) developed a hemorrhage bundle, which included specific auditing for compliance, including PPH risk assessment, blood product crossmatch according to PPH risk, QBL (not qualitative blood loss) assessment, laboratory results obtained in stages 2 and 3 hemorrhages, permission to give >2 uterotonics without a physician present when necessary, and blood product administration per protocol. When compliance was less than expected, safety nurses visited the hospitals to provide staff education and simulation. The bundle compliance increased from 54% to 80% over this period, during which recognition and correct designation of the stages of hemorrhage also improved and the overall use of blood products decreased (25.9% reduction per 1000 births in blood product use at 10 months after implementation). These successful PPH bundles and stage-based protocols highlight the importance of implementation and verification of practice change.

Guidance on benchmarks for postpartum hemorrhage management

Current PPH quality improvement initiatives incorporate the assessment of care delivery into their efforts. Currently experienced organizations in bundle implementation, including the Alliance for Innovation on Maternal Health and the California Maternal Quality Care Collaborative, divide the metrics into several categories: process, structure, and outcome measures.^21,22 Process measures include components of early recognition, including assessment of hemorrhage risk assessment and quantitative assessment of blood loss. Furthermore, they include the process for fostering a culture of safety and improvement via systems-based learning, including case reviews for all hemorrhage cases that have complications (such as transfusion of ≥4 units packed red cells or intensive care unit [ICU] admissions) and event debriefs for cases of hemorrhage above a specific threshold.

The World Health Organization emphasizes the importance of the administration of a prophylactic uterotonic as an important process metric but also recognizes that locally agreed, more specific indicators may be employed as well. Importantly, none of these guides address directly how to measure the quality of the efforts to obtain hemostasis during PPH—either the interventions performed or the timing in which they are performed.

There is some evidence to suggest that focused attention to the interventions and timing of procedures performed for PPH management may improve outcomes. French investigators from 2 facilities participating in a cluster-randomized trial of a PPH bundle continued ongoing reviews of PPH cases.¹⁸ During that time, cases of PPH were audited to the ensure provision of “optimal care,” which included (1) examination of the uterine cavity and/or manual removal of the placenta within 15 minutes of the PPH diagnosis, (2) call for additional staff, (3) instrumental examination of the vagina and cervix, (4) intravenous administration of oxytocin, and (5) IV administration of sulprostone (prostaglandin E2 agonist; second-line agent) within 30 minutes of initial diagnosis if PPH persisted and due to atony. The findings of the audit were presented during quarterly multidisciplinary meetings in which the treating providers and audit team had the opportunity to discuss both the audit results and other lessons from each PPH. The investigators found that in the 4-year auditing process, there was a statistically significant improvement in the rate of severe PPH (bleeding requiring blood transfusion, surgical intervention, or resulting in a hemoglobin decline of ≥4 g/dL or maternal death) at both facilities (1.52% to 0.96% at 1 facility and 2.08% to 0.57% at the other), although the rates of adherence to their guidelines improved during the study period. Although they did not assess clinical outcomes, investigators in Malawi and Ethiopia also demonstrated that criteria-based clinical audits could improve adherence to PPH process metrics.^19,20

Finally, outcome measures that have been used for PPH include SMM in hemorrhage cases, transfusions of ≥4 units, and hemorrhage rates. Each strategy has limitations. SMM in hemorrhage cases is a measure reported using delivery diagnosis-related group codes in association with many International Classification of Diseases, Tenth Revision, codes validated as being associated with morbidity. The limitations include that coding data does not reliably distinguish between 1 and 100 units of blood, and these codes are not risk stratified. As an example, a hospital that is a referral center for placenta accreta spectrum may have higher rates of transfusion than a level 1 facility that transfers all patients with a high risk of hemorrhage to another center. Moreover, hemorrhage rates are fraught with potential error, as QBL often leads to higher numeric volumes of blood loss being reported simply related to more accurate assessments.

Proposed benchmarks and further directions

We would suggest consideration of a few specific management strategies in stage-based hemorrhage protocols, based on existing literature (Table). We selected these as benchmarks based on the evidence reviewed above—that timely use of uterotonics, TXA, and procedural placement reduces the risk of severe hemorrhage sequela and the need for additional procedures. Hospital systems and PQCs that include auditing in their implementation strategies reported improvements in PPH outcomes, although hospitals that merely have a protocol without measuring compliance do not have similar improvements noted.

TABLE.

Suggested benchmarks for PPH protocol implementation

Benchmarks
For hemorrhage involving uterine atony, administration of a second uterotonic (after Pitocin) without delay once PPH is recognized, typically either methylergometrine or carboprost, unless contraindicated
For patients with PPH, administration of tranexamic acid as rapidly as feasible after diagnosis
For patients whose PPH is due to atony and is refractory to a second uterotonic, early placement of an intrauterine tamponade balloon (eg, Bakri balloon) or intrauterine vacuum device (eg, Jada device) before blood loss exceeds 1.5 L
Auditing of PPH cases to ensure teams are following PPH bundles and stage-based protocols, with corrective action plans to include reeducation and simulation when compliance is suboptimal, with escalation to processes, such as peer review reserved for cases in which specific providers do not alter practice patterns despite reeducation

PPH, postpartum hemorrhage.

Federspiel Postpartum hemorrhage protocols and benchmarks. Am J Obstet Gynecol MFM 2022.

We recognize that the benefits of timely PPH management raise the question of whether specific PPH treatment time standards should be recommended. Timeliness for PPH response matters greatly. However, the authors believe that the implementation of PPH care benchmarks involving specific time standards should include an evaluation of their effectiveness before broad implementation. Despite good intentions, quality metrics in medicine, particularly in obstetrics, remain challenging to define and implement.²¹ There is precedent in the medical literature regarding unintended adverse consequences of mandating process metrics based on time. Based on observational data that provision of antibiotics within 4 hours of hospitalization for patients with community-acquired pneumonia (CAP) was associated with improved outcomes,²² the Infectious Diseases Society of America included this benchmark as a practice guideline, and it was tied to quality reporting and reimbursement.²³ This benchmark created a strong incentive to prescribe antibiotics to any patient potentially at risk of CAP, which is often difficult to diagnose. Consequently, there was a significant increase in antibiotic prescriptions for hospitalized inpatients, an increased rate of incorrect diagnosis of CAP, and no improvement in outcomes for patients in which the CAP diagnosis was confirmed.²⁴ Based on this evidence, this benchmark was removed from the guidelines.

One could imagine that a similarly plausible phenomenon might be possible for PPH benchmarks—that a benchmark might prompt reflexive delivery of these services, even in patients in which it is unlikely to benefit, to ensure the provider gets “credit” for the intervention. Considering the example of uterotonics, for example, all available agents have potential adverse effects (eg, hypertension for methylergonovine, bronchospasm for carboprost, and fever for misoprostol). The need to precisely document times of hemorrhage recognition and management steps and the PPH etiology (as most of these guidelines are primarily pertinent to PPH because of uterine atony) would also significantly increase the data collection burden for participating sites. For this reason, we would advocate for a prospective observational study, ideally with a difference-in-difference design or via a pre- or postdesign, if not possible, to evaluate whether routine measurement and reporting of time-based PPH management benchmarks improve adherence and whether it improves PPH outcomes, considering outcomes, such as transfusion of ≥4 units of blood products, need for additional procedures, and ICU admission.

Conclusion

PPH remains a significant etiology of maternal morbidity and mortality worldwide. Recognition that a hemorrhage is occurring is essential to lead to management via stage-based protocols. We recommend that these protocols include a focus on the early use of uterotonic use, TXA, and intrauterine balloon tamponade. Finally, team training, simulation, and auditing for compliance are essential features to ensure protocols lead to a reduction in hemorrhage management.