Delays in recognition of risk factors and escalation of care are known to be associated with preventable maternal morbidity and mortality.13 These include delays in assessing clinical warning signs, providing accurate diagnoses, implementing optimal treatment, and coordinating care with multidisciplinary teams.1,13,14 For example, for deaths related to maternal hemorrhage, recognition of hemorrhage could be improved through awareness of risk factors, accurate assessment of blood loss, and identification of early signs of hypovolemia.15
The sidebar presents a composite case (combined elements from multiple cases with identifying features, including all personal identifiers, removed to ensure patient confidentiality). The case has been adapted from the Obstetric Hemorrhage Quality Improvement Initiative of the California Maternal Quality Care Collaborative. The case demonstrates how deficiencies in multiple dimensions of the diagnostic process can lead to the escalation of a low-risk pregnancy and birth to an emergent situation and preventable death.
Figure 1 shows the key gaps in the diagnostic process and opportunities for improvement. The diagram is an adaptation of the multidisciplinary Safer Dx framework for measuring, monitoring, and improving diagnostic error.16 The boxes on the left of the diagram are in alphabetical order and note the diagnostic process dimensions within the Safer Dx Framework. The dimensions link to the Quality Improvement Opportunities and Learning Points noted with superscript letters (e.g., a, b) in the Maternal Hemorrhage Case.
The Safer Dx framework adapts the structure-process-outcome framework16 to describe the foundations of safe and timely diagnosis:
- Structure: characteristics of care providers, their tools and resources, and the physical/organizational setting;
- Process: both interpersonal and technical aspects of activities that constitute healthcare; and
- Outcome: change in the patient’s health status or behavior.
The Safer Dx framework emphasizes system-level learning and improvement as a continuous feedback loop. Measurement and learning from diagnostic error (and circumstances without diagnostic error) must include all aspects of the diagnostic process that may evolve over time and should not be limited to what occurs during a single patient encounter. As AHRQ states, “Diagnostic performance is the outcome of these processes within a complex, adaptive sociotechnical system. Safe diagnosis (as opposed to missed, delayed, or wrong) is an intermediate outcome compared with more distal patient and healthcare delivery outcomes.”17
The most common types of preventable events after a patient has accessed care in the hospital system for childbirth are related to:
- Inadequate diagnosis or delay in recognition of acuity,
- Inappropriate or inadequate treatment, and
- Inadequate documentation (poor charting or failure to chart) of risk factors.2
Cognitive biases, such as confirmation bias and implicit biases, can also affect a provider’s diagnostic accuracy. Communication breakdowns can occur between patients and providers and within the healthcare team in an often-fragmented healthcare system.18,19 A number of different approaches can address these known process vulnerabilities, including:
- Implementing hospitalwide safety bundles.
- Addressing implicit bias, structural racism, and social determinants of health.
- Using healthcare simulation, a unique tool for assessing and addressing gaps in care.
Case Discussion: Maternal Hemorrhage20
Composite Case Example: A healthy 24-year-old presented at 39 weeks gestation and gave birth vaginally to an 8-pound, 6-ounce infant after 10 hours in labor.
After placental delivery, the patient had an episode of uterine atony (inability of the uterus to contract) that firmed with massage and described feeling “tired and thirsty” to the healthcare team.a The nurses called the physician 30 minutes later to report a heart rate of 105b and more bleeding,c for which methylergonovine (a medication to help promote uterine contraction) was ordered.d
Sixty minutes after the initial call, the physician performed a dilation and curettage (D&C), with removal of minimal tissue from within the uterus, thus excluding retained placental tissue as the cause of increased bleeding. The patient received more methylergonovine.e
Forty-five minutes later, a second D&C was performed, again with minimal return of placental tissue. Estimated blood loss at this point was >2,000 mL.c Delays in blood transfusion occurred due to the inability to find proper tubing for the rapid infuser.f Arrival of an anesthesiologist was delayed due to another emergency.g
Vital signs became markedly abnormal, with a pulse of 144 beats per minute and blood pressure 80/30 mm Hg. The patient received 2 units of red blood cells but a massive transfusion protocol was not activated due to the lack of an organized standard approach.h The patient had a cardiac arrest from severe blood loss and died 3 hours later despite intensive resuscitative efforts.
Figure 1. Using the Safer Dx Framework16 as a Model for Improvement of Diagnostic Error and Contributions to Maternal Mortality and Severe Maternal Morbidity With Hemorrhage as a Case Example
* Includes eight technological and nontechnological dimensions.
† Includes external factors affecting diagnostic performance and measurement, such as payment systems, legal factors, national quality measurement initiatives, accreditation, and other policy and regulatory requirements.
Patient-Provider Encounter & Initial Diagnostic Assessment
- Healthcare team did not incorporate patient’s self-reported symptoms of tiredness and thirst into assessment.
- Underappreciation of tachycardia in the context of pregnancy changes.
Diagnostic Test Performance & Interpretation
- Visual estimation of blood loss - no quantification of actual blood loss.
- Limited physical exam and no lab tests drawn to clarify extent of bleeding.
Followup and Tracking of Diagnostic Information
- Only a few treatments tried, and then repeated even when ineffective, instead of escalating to additional medications or interventions.
- Delays in blood administration.
Subspeciality/Consultation Referral Issues
- Delay in response from key multidisciplinary team members.
- Lack of an organized standardized team approach.