Society for Maternal‐Fetal Medicine Special Statement: Tools for patient safety and quality case review and response: Obstetric sepsis

Abstract

Sepsis‐related obstetric morbidity has significant consequences and can be a precursor to maternal mortality. Diagnosing and managing obstetric sepsis is uniquely challenging but a critical target for improving maternal outcomes. This document provides considerations and tools for breaking down a case of obstetric sepsis at the clinical unit level. In addition, this document outlines special issues with performing quality improvement projects as they relate to obstetric sepsis: (1) encouraging frontline implementers to consider the wide array of stakeholders involved in sepsis cases, (2) how to harness the electronic health record for enhancing care delivery, (3) challenges around common sepsis‐related metrics, (4) considerations around locations and transitions in obstetric sepsis care, and (5) considerations for ensuring equity in sepsis‐related projects.

1. CASE PRESENTATION

Patient SE is a 35‐year‐old G2P2002 Hispanic Black woman on postoperative day 5 (POD#5) who underwent an uncomplicated low transverse cesarean delivery (LTCD) for arrest of descent. There were no intrapartum or postpartum complications. She was discharged from the hospital on POD#3 in stable condition. SE calls with worsening abdominal pain despite the use of acetaminophen and non‐steroidal anti‐inflammatory drugs. The clinician covering phone triage instructs the patient to try opioid pain medication at home, attributing her pain to normal recovery after her second cesarean.

The patient calls back the next day on POD#6 with a fever of 101.7°F. Abdominal pain has continued despite utilization of opioids, and she now also reports bilateral breast engorgement and tenderness. A different responding triage clinician orders antibiotics to treat mastitis and informs the patient she can stay home.

That night, she presents to the emergency department (ED) with hypotension, tachycardia, and fever of 103.8°F. On the ED exam, the clinician notes “abdominal tenderness appropriate for the postoperative state.” A consult to the obstetrician‐gynecologist (OBGYN) is called, but the consulting physician is delayed due to labor floor acuity. Three hours after presentation, an OBGYN comes to evaluate the patient, noting worsening vital signs and significant fundal tenderness. She makes the diagnosis of endometritis and orders imaging, labs, fluid resuscitation, and targeted antibiotics. The patient is transferred to the postpartum floor for further evaluation and management. She makes it to the postpartum floor after a lapse of an additional 2 h due to staffing and coverage issues with postpartum nursing.

Upon evaluation by the clinical team on the postpartum floor, her hypotension and tachycardia are noted to worsen despite initial fluid resuscitation and antibiotic treatment. In consultation with the infectious disease team, a decision is made to transfer the patient to the intensive care unit (ICU). The patient arrives at the ICU for management of septic shock after an additional 2 h. In the ICU, infectious disease is consulted, and recommends broadening antibiotic coverage given signs of end‐organ damage in lab work. However, this note remains unread by the primary ICU team. By the next morning, the patient has increasing pressor requirements and acidosis. She is taken to the operating room (OR) where a hysterectomy is required for source control. Antibiotics are broadened, and after several days in the ICU, she recovers and is discharged home.

2. SEPSIS‐RELATED MORBIDITY AND MORTALITY IN PREGNANCY

The World Health Organization (WHO) defines maternal sepsis as organ dysfunction resulting from infection during pregnancy, childbirth, post‐abortion, or the postpartum period (up to 42 days) [1]. Sepsis and septic shock represent 2 of 21 indicators of severe maternal morbidity (SMM), as defined by the Centers for Disease Control and Prevention (CDC) [2]. Sepsis‐related morbidity in pregnancy and postpartum has significant consequences for both the birthing person and fetus (if applicable), including maternal organ dysfunction, disseminated intravascular coagulation (DIC), preterm labor and delivery, fetal compromise including growth restriction and stillbirth, and neonatal sepsis [3, 4]. SMM events can represent precursors to maternal mortality. In addition, there are notable racial and ethnic disparities in the incidence and severity of obstetric sepsis [5].

Yet, diagnosing and managing obstetric sepsis are challenging. Pregnant individuals may present with atypical symptoms of sepsis, which can lead to underrecognition or misdiagnosis. Clinicians may overlook signs and symptoms such as tachycardia and shortness of breath, attributing them to normal physiologic changes of pregnancy. Some laboratory and clinical indicators such as leukocytosis and fever can be altered in pregnancy or the labor process, making a definitive diagnosis challenging. Delays in treatment can occur due to hesitancy to prescribe medication in pregnancy or during breastfeeding. Identifying and overcoming these challenges with timely diagnosis and early intervention presents a unique opportunity for preventing the downstream cascade that can result in SMM or maternal mortality from a case of maternal sepsis.

In response to the profound consequences of obstetric sepsis, The Alliance for Innovation on Maternal health (AIM) published a consensus bundle on “Sepsis in Obstetric Care.” The overarching goal of this document is to promote coordinated care for birthing people that recognizes and treats infection before sepsis develops. AIM provides guidance on a multitude of recommended practices for quality improvement (QI) to promote readiness, recognition, and response to sepsis [6]. However, implementation of any consensus bundle requires significant time and resources from the organization working to put it into place.

Commonly, the drive for unit change comes out of a clinical case, which could represent a near‐miss or even a sentinel event. The AIM consensus bundle includes a recommendation to conduct multidisciplinary reviews for systems improvement of each sepsis case to assess the screening program, the quality of care provided to patients with sepsis, and whether instances of bias may have impacted care [6]. This document is intended to provide insight into meeting the AIM recommendation, with a focus on the unique aspects of performing a quality patient safety review of a case of obstetric sepsis. Tools are described specific to this clinical scenario. Here, we also provide initial guidance on selecting a QI project in response, including options regarding metric selection.

3. CASE REVIEW IN OBSTETRIC SEPSIS: UNIQUE CONSIDERATIONS AND QUALITY IMPROVEMENT TOOLS

Reviewing a case of sepsis in pregnancy requires a systematic evaluation of multi‐level barriers to care in order to identify avenues for enhancing care delivery. Several tools for performing a QI‐ focused case review are applicable to a case of obstetric sepsis and are described in the Society for Maternal‐Fetal Medicine Special Statement: Curriculum outline on patient safety and quality for maternal‐fetal medicine fellows [7]. In addition, the application of such tools to a clinical case focused on maternal health equity will be described in a forthcoming document.

As an example of the application of QI methodologies to a case of maternal sepsis, use of Fishbone diagrams could evaluate key categories of root causes, such as the role of clinicians, the role of patients, the role of setting and equipment, and policies which are or are not in place regarding management of patients with obstetric sepsis. Similarly, Pareto charts could be used to evaluate multiple cases of sepsis and determine where contributing factors emerge as themes across patients. Beyond the utilization of general QI tools, the following describes unique considerations for a case review specific to obstetric sepsis [7].

3.1. Need for coordinated review across clinical subspecialties and interdisciplinary professional teams

Managing a case of sepsis during pregnancy and postpartum hinges upon prompt and thorough care coordination between interdisciplinary teams of professionals across various clinical subspecialties. Based on the acuity of clinical presentation, management paradigms will entail seeking and providing multidisciplinary team‐based care that includes clinicians with subspecialty training and expertise in emergency medicine, primary care physicians, maternal‐fetal medicine, infectious diseases, critical care medicine, obstetric anesthesia and gynecologic oncology, depending on the clinical circumstances involved. In addition, care for these complex patients requires support from nursing, clinical pharmacists, respiratory therapists, and staff in the OR and ED. Hence, case reviews should be conducted by a multidisciplinary panel with representation from each of the subspecialties and clinician roles that were or should have been involved in the clinical management of the case being discussed. Nearly all of these disciplines were involved in the care of the case as presented above. The goal of such reviews is to identify areas for improvement not only in medical decision‐making, but also in terms of care coordination, communication, team‐based care, and clinical decision support.

Such a coordinated review is not without challenges in a busy clinical practice. Furthermore, clinicians might feel more cautious participating in a quality review in a new environment, outside of their home department. Including staff members involved in the case might pose issues with the hierarchical structure of medicine. As a result, the primary reviewer of the case may elect to hold individual or small group meetings with clinicians involved in order to accommodate schedules. The peer‐protected nature of case reviews should be explained to individuals who might not be as familiar with the process. Some staff members may prefer to have supervisors present for such meetings. Consideration should also be given to asking involved departments outside of obstetrics to perform their own case reviews during QI meetings or through their own departmental “Morbidity and Mortality (M&M)” conferences. Likewise, if an obstetric M&M is planned, inviting representatives and participation across disciplines can foster a culture of safety across the care continuum.

3.2. Evaluating for recognition of evolving pregnancy and postpartum physiology

Reviewing a case of maternal sepsis requires an in‐depth understanding of physiological changes in pregnancy and postpartum that can potentially alter clinical presentation and thereby impact the ability to establish an accurate clinical diagnosis. An appreciation of changes in vital signs, laboratory parameters and findings on imaging studies is vital to such considerations. Hence, reviewing a case of maternal sepsis needs to account for changes in pregnancy physiology, including the transition to postpartum physiology. Importantly, peripartum patients with infections may present to a variety of initial locations—outpatient offices, the ED, obstetric triage, etc. Therefore, a heightened level of vigilance for maternal infectious processes or underlying causes of sepsis by many types of clinicians and clinical settings is essential for sepsis treatment success.

For example, as in the case above, SE's initial reported pain was attributed to normal postpartum recovery, when it was actually indicative of pathology. Normal breast engorgement was called mastitis by the triage provider and identified as a source of infection, when in fact the actual source was elsewhere. The emergency room clinician did not recognize the severity of SE's presentation to the ED. An example of utilizing a fishbone diagram to evaluate the root causes of delayed diagnosis in SE's case is shown in Figure 1. If such gaps in the identification of sepsis are found during the case review process, targeted interventions should be undertaken to prevent future misdiagnosis when proceeding with QI initiatives.

FIGURE 1.

Example of a fishbone diagram to evaluate contributing factors to the delay in sepsis diagnosis.

3.3. Evaluating for timelines of critical clinical interventions

A critical aspect of care delivery in managing and reviewing cases of maternal sepsis is identifying the initial “golden” hour upon presentation. Data support that early diagnosis, utilizing blood cultures, laboratory workup (including lactate), and imaging as appropriate, as well as initiation of interventions like fluid resuscitation and antibiotic administration, is critical to improving sepsis outcomes [8, 9]. Specific recommendations regarding when to begin fluid resuscitation and when to initiate vasopressor use can be found in Figure 1 of the associated SMFM Consult [10]. It is recommended that laboratory evaluation, including blood cultures, be obtained prior to administration of broad‐spectrum antibiotics when there is clinical suspicion of maternal sepsis. The subsequent choice of antibiotics should then be guided by the results of such investigation in conjunction with the local antibiogram, as may be appropriate for the clinical scenario under consideration. Hence, reconstructing and evaluating timelines for obtaining laboratory testing and delivering critical interventions including but not limited to antibiotics, fluid resuscitation, pressor support (if needed) as well as source control are essential for performing thorough case reviews with a focus on identifying opportunities for improvement.

A novel tool like an antibiotic map may be useful in evaluating where gaps in antibiotic coverage may have contributed to an obstetric sepsis case. Such gaps may include issues with communication of recommendations, selected microbial coverage, time from recommendation to ordering or administration, or lapses in antibiotic coverage during the patient's course. A timeline of antibiotic management over the course of the patient's care can be constructed, utilizing Figure 2. A sample antibiotic map form is shown in Figure 3, with an example applied in Figure 4.

FIGURE 2.

Antibiotic timeline. This form can be used for each day of a patient's case.

FIGURE 3.

Blank antibiotic map.

FIGURE 4.

Antibiotic map with a case example. Here, the value in tracking antibiotic administration can be seen, especially as the patient traverses locations within the health system, as this identifies specific targets for improvement.

Once such a gap is identified, a process map may be helpful in further identifying where change is required. For example, if the antibiotic map finds gaps in antibiotic coverage secondary to a delay from ordering a medication to administration, a process map can detail the exact steps that need to occur to ensure the patient receives treatment. Importantly, the actual process is often different from the assumed process and requires speaking to frontline staff and interrogating electronic systems to concretely elucidate delays. Once the true process is determined, steps susceptible to error or delay can be targeted for intervention. An example of such a process map specific to SE's case is shown in Figure 5.

FIGURE 5.

Sample process map evaluating the workflow from order to administration of an antibiotic.

3.4. Care transition across clinical settings, hand‐off communication, and the impact of fourth trimester care in shaping sepsis‐related morbidity

As presentation of the peripartum patient with infectious processes may occur in a variety of locations, appropriate time to and location of patient disposition is also critical to track during the review. Patients may need to be transferred from outpatient to inpatient settings, to the ICU, to a step‐down or higher‐acuity special care unit, or to the inpatient obstetric floors. In the case of SE, there was a significant delay in transfer from the ED to the inpatient labor unit, and then a quick transition to the ICU. Delays like this one likely contributed to her cumulative morbidity.

Furthermore, postpartum patients may experience significant difficulties interacting with the healthcare system, especially in the context of social determinants of health. Often, inpatient systems of care, particularly after delivery, do not place enough stress on connecting back to the outpatient setting, such as continuing care with the primary obstetrician, primary care provider, or other subspecialist [11]. When there are risk factors for or early signs of infection, communication to and follow‐up by outpatient clinicians are integral to the early diagnosis and treatment of sepsis. Thus, evaluation of follow‐up after hospital discharge, outpatient clinical care in the fourth trimester, and timely recognition and management of postpartum complications should constitute an integral part of case reviews.

Finally, within hospital communication and healthcare handoffs are another target for evaluation around obstetric sepsis, especially when consulting services become involved in a patient's care. For example, in SE's case, a recommendation was made by an infectious disease consultant in an electronic note, but this recommendation was not recognized, and care was not changed in response for several hours. Case reviews in obstetric sepsis should particularly focus on identifying these multidisciplinary gaps in communication which become more pronounced in the setting of complex patients with obstetric sepsis.

3.5. Impact of social drivers of health, systemic and institutional racism, and cultural biases on obstetric sepsis case review

With the implementation of bundles, clinical guidelines, and protocols, standardization of care management may decrease disparities in care and address implicit biases that impact decision‐making [12, 13, 14, 15, 16, 17, 18]. Case reviews in obstetric sepsis should aim to objectively evaluate if patient and clinician level barriers to care, systemic and institutional biases, and the influence of social drivers of health impact clinical decision‐making, care delivery, and clinical outcomes. With SE, there is concern that individual and systemic biases led to dismissiveness towards her pain, therefore delaying the diagnosis of infection. Evaluating for bias playing a role in quality case review is a challenging task. However, some methods include evaluating the medical record for linguistic features of disbelief, as well as assessing the role of cognitive bias when discussing the case with involved clinicians.

4. QUALITY IMPROVEMENT INTERVENTION IN OBSTETRIC SEPSIS

Utilization of QI tools as described above may help a site identify one or more targeted projects to address gaps in obstetric sepsis‐related care, especially where implementing the entire AIM bundle simultaneously may not be feasible. Examples of sepsis‐associated data points and metrics to target include utilization of sepsis identification screening tools, timeliness of lab draws, markers of timely initiation of antibiotic therapy, timeliness of source control procedures, metrics around antibiotic stewardship, and clinical outcome measures, such as readmission, length of stay, and mortality rates. A comprehensive list of obstetric sepsis associated metrics is available in the AIM bundle Supplemental Files [19]. Such metrics help healthcare facilities assess their performance in sepsis management, and monitor the effectiveness of interventions and QI initiatives over time.

As an example of a QI project that could emerge from the issues identified in the case of SE, a site may decide to target improving the time from presentation with obstetric sepsis to antibiotic administration. When considering how to successfully move ahead with a focused QI project, there are again unique considerations for obstetric sepsis that parallel many of the considerations for case review.

4.1. Stakeholder engagement across clinical subspecialties, interdisciplinary professional teams, and patients

Form a collaborative team encompassing all clinician and staff types who may engage with a patient who presents with symptoms of infection. In this case, members might include ED nursing, ED physicians, OBGYN clinicians, infectious disease clinicians, hospital administration, and pharmacy staff. Attention should always be given to including trainees and other front‐line staff in the development and implementation of quality interventions. The patient voice is also a critical component to successful QI. However, identifying a patient with lived experience of obstetric sepsis to participate may pose a challenge due to the relative rarity of the diagnosis. The stakeholder team should also be continuously engaged in project development, implementation, and sustainment, with regular review of data metrics and discussions of implementation adaptation in order to inform improvement.

4.2. Utilizing the electronic health record (EHR) and data science to improve identification of obstetric sepsis

While there remains a lack of consensus on the best sepsis screening algorithm or early warning system to implement [19], there are increasing data on the performance characteristics of these tools [20, 21]. Clearly, a process that harnesses the power of the EHR has high potential value specific to obstetric sepsis [22, 23]. With the goal of improving time from meeting sepsis criteria to antibiotic administration, a site may decide that an obstetric sepsis identification tool is the best targeted intervention. However, implementing an obstetric sepsis identification system in the EHR poses challenges similar to many complex EHR builds, including difficulty identifying key, discrete variables; need for personnel with expertise and time to create and test the build; and lack of coding reciprocity across sites despite similar record systems. Other EHR modifications that could be beneficial in the management of obstetric sepsis include standardized order sets, including all recommended laboratory evaluations, as well as options for initial antibiotic treatment. A sample order set is provided on the California Maternal Quality Care Collaborative website: Appendix M: Sample Sutter Health Sepsis Order Set [24].

4.3. Identifying “time zero” for timeliness measures

When considering a project to improve the time from diagnosis to antibiotic administration, such as in response to SE's case, one needs to define the time of sepsis diagnosis in order to calculate the time from diagnosis to antibiotic administration. An EHR identification tool to help recognize that a patient meets criteria for obstetric sepsis, based on vital signs and/or laboratory findings, could provide an automated “time zero” if built into the tool's workflow. Specifically, an EHR‐based tool could identify an initial positive sepsis screen as detailed by the SMFM Consult Series [10], defined by two of the following four criteria: maternal temperature <36°C (96.8F) or ≥38°C (100.4°F), heart rate >100 bpm and sustained for 15 min, respiratory rate >24 breaths/min and sustained for 15 min, and white blood cell count >15,000 mm³ or <4000 mm³ or >10% immature neutrophils.

However, when an EHR‐based tool is not feasible or available, identification of “time zero” is more challenging, and the need for individual medical record review could limit the ability to sustain data collection on the impact of the QI intervention. Alternative data points that could be used for “time zero”, each with their own downsides, include: (1) time of presentation for evaluation, (2) time an abnormal lab resulted (leukocytosis, etc.), (3) time of first abnormal vital signs or documented fever, (4) timestamp of a clinician note, and (5) time of antibiotic order placement. Importantly, each of these selected “time zeros” truncates a portion of the workflow. For example, if the time of antibiotic order placement is utilized as “time zero,” this metric ignores the delay in time to recognition of the need for antibiotics. Finally, the difficulty in identifying a perfect “time zero” to track should not prevent teams from attempting improvement, with an understanding of the limitations of the selected metric.

4.4. Considering care locations and transitions

Just as the multiple locations a patient may present or be transferred to with obstetric sepsis need to be considered in case review, they must also be considered in any QI project. First, the ED needs the ability to expeditiously recognize pregnant or recently pregnant patients with sepsis to specifically target this population for early recognition and treatment. In addition, ED and obstetric triage workflows must be aligned to effectively target timely diagnosis and antibiotic administration. Stakeholder engagement across departments and an in‐depth understanding of the differences in context could contribute to meaningful QI initiatives.

4.5. Multilevel implementation strategies, with a focus on equity

Throughout healthcare implementation, it has been well‐studied that provider and staff education and training alone is insufficient to promote change [25, 26, 27]. When considering the proposed project example to improve the timeliness of antibiotic administration in obstetric sepsis, we have already addressed other critical implementation strategies, including garnering stakeholder buy‐in and engaging champions among included units, such as the ED, obstetric triage, and ICU. Sites may also find that additional strategies such as audit and feedback may be helpful, at least initially, regarding metric outcomes in order to encourage and engage all clinicians in the QI endeavor. Central to any of the strategies selected to enhance implementation is the consideration of including an equity lens to ensure that the implementation plan will reach all patients [28]. For example, when considering the case of SE, if we initiated a QI intervention that only focused on the processes in obstetric clinic triage, and not the ED, and our patients of historically marginalized backgrounds were more likely to present to the ED due to access barriers, our QI project would inherently worsen disparities in sepsis outcomes.

5. CONCLUSIONS

The magnitude of sepsis‐related morbidity in pregnancy and postpartum highlights the critical need for a multifaceted approach to address this concerning issue. Quality review of obstetric sepsis is uniquely challenging; this paper provides considerations and tools for breaking down a case of obstetric sepsis and outlines special issues with QI projects as they relate to obstetric sepsis. The goal of this work is to enable healthcare facilities to identify areas for enhancement, track the effectiveness of interventions, and ultimately reduce sepsis‐related morbidity and mortality in pregnancy. While this document may aid individual sites in sepsis‐related QI, there remains significant work to be done on a larger scale. The development of improved sepsis‐related quality metrics would bolster QI initiatives in this area. While there is more recent work implementing interventions designed to improve obstetric sepsis [23, 29], future research should focus on how to sustain these improvements.

The use of this information is voluntary, and clinicians should be familiar with and comply with all applicable laws and regulations.

All authors and committee members have filed a disclosure of interests delineating personal, professional, business, or other relevant financial or nonfinancial interests in relation to this publication. Any substantial conflicts of interest have been addressed through a process approved by the Society for Maternal‐Fetal Medicine (SMFM) Board of Directors. SMFM has neither solicited nor accepted any commercial involvement in the specific content development of this publication.

This document has undergone an internal peer review through a multilevel committee process within SMFM. This review involves critique and feedback from the SMFM Patient Safety and Quality and Document Review Committees and final approval by the SMFM Executive Committee. SMFM accepts sole responsibility for the document content. SMFM publications do not undergo editorial and peer review by Pregnancy. The SMFM Patient Safety and Quality Committee reviews publications every 24 to 36 months and issues updates as needed. Further details regarding SMFM publications can be found at www.smfm.org/publications.

SMFM recognizes that obstetrical patients have diverse gender identities and strives to use gender‐inclusive language in all publications. SMFM uses terms such as “pregnant person” and “pregnant individual” and the singular pronoun “they.” When describing study populations used in research, SMFM uses the terminology reported by the study investigators.

All questions or comments regarding the document should be referred to pubs@smfm.org

Reprints will not be available.