The Case

A 28-year-old woman arrived at the Emergency Department (ED) complaining of back pain and bloody vaginal discharge. She reported some missed periods and a positive urine pregnancy test at home but had not received any prenatal care and was unsure of her expected due date. 

The initial physician evaluating the patient was an intern on their first day of service in a busy Emergency Department. The intern suspected complications of early pregnancy including ectopic pregnancy, spontaneous abortion, and gestational trophoblastic disease. After cursory review by a more experienced physician, a pelvic ultrasound was ordered as a “first trimester” examination. Active labor was not suspected by the ED physicians.

The on-call sonographer was trained in abdominal ultrasound but had limited experience in fetal sonography. The patient’s obesity limited the ultrasound assessment; the quality of the ultrasound images was poor. The sonographer did not recognize that the patient was in her third trimester based on a review of the images, and mistakenly identified the fetal head as a “pelvic mass.” The radiologist on call was experienced in abdominal ultrasound but had limited experience in fetal ultrasound. The sonographic images were reviewed remotely, and the radiologist also failed to recognize that the “pelvic mass” was actually a fetal head. The radiologist did not repeat pertinent aspects of the ultrasound exam himself.

Subsequently, a consultation was sought from the obstetrics/gynecology team on call, which was completed four hours after the patient had arrived to the ED. Based on their evaluation, including a pelvic examination and repeated pelvic ultrasound at the bedside, they recognized that the patient was in her third trimester and in active labor. The patient was transferred to Labor and Delivery, where standard procedures for labor management were initiated. The initial fetal heart rate tracing (FHRT) was category 1 (normal), but variable decelerations were later noted (category 2) which deteriorated to fetal bradycardia. The obstetric team proceeded to an emergent cesarean section that was successfully accomplished. Apgar scores at one, five, and ten minutes were 2, 2, and 7, respectively. The infant then suffered a neonatal seizure. Subsequent investigation revealed a perinatal thrombotic infarct of the left basal ganglia that was confirmed on brain MRI. It could not be determined if the perinatal stroke was a result of an antenatal insult or a consequence of the labor and delivery.

The Commentary

By Gary S. Leiserowitz, MD, MS and Herman Hedriana, MD

The primary concern in this case was the delay in recognition that the patient was in her third trimester and in early active labor when she presented to the ED, rather than having first trimester bleeding.  Several patient-related factors and medical diagnostic errors probably contributed to this delay.

• The patient could not recall time-related details of her pregnancy and her health literacy was limited. Thus, she did not recognize how advanced her pregnancy was.
• The patient was obese, which contributed to a failure by the ED physicians to appreciate the fundal height of a third trimester pregnancy on physical exam. The patient’s obesity also made the ultrasound more challenging to perform and interpret by the sonographer.
• The sonographer had limited training in obstetric ultrasound. Since the ultrasound had been ordered as a “first trimester” exam, the sonographer used that context in interpreting the findings and failed to appreciate that the “pelvic mass” was actually a fetal head.
• The radiologist on call had little recent experience in obstetric sonography. He read the ultrasound remotely and therefore did not repeat key components of the exam himself.  Thus, the incorrect diagnostic impression of a “pelvic mass” persisted.

Diagnostic Errors

For various reasons, establishing a definitive diagnosis is not always a clear process. Because clinical reasoning by individual physicians sits at the intersection of science and art in medicine, for example, obtaining balance between missing a diagnosis and ordering too many tests is difficult. Diagnoses also often unfold over time as more information emerges, making putting the story together accurately before all of the relevant facts are gathered and integrated very difficult. These difficulties in diagnosing can contribute to diagnostic errors.

Diagnostic errors are often more challenging to address than other patient safety issues such as medication errors.⁷ Diagnostic errors generally are of two types: underdiagnosis (as illustrated by this case) or overdiagnosis, both of which can have serious adverse consequences.⁷ Diagnostic errors also can be attributed to three broad categories: patient-related factors, cognitive errors, and systems factors,¹ all three of which likely contributed to the delay in diagnosing the third-trimester pregnancy and active labor in this case; and as the diagnostic errors aligned, holes in layers of defense also lined up, a situation often referred to as the “Swiss cheese” effect.² Distinguishing between the contributions to diagnostic errors of systems problems versus cognitive factors, especially in a busy ED, also can be challenging.

In this case, the patient-related factors included limited health literacy and obesity. The patient’s own inability to remember the timing and number of her missed periods and irregular bleeding episodes prevented her from recognizing her advanced state of pregnancy. Her lack of prenatal care represented a missed opportunity to get an early ultrasound for pregnancy dating.

Cognitive errors were made by the intern and the sonographer, who both lacked sufficient knowledge and experience with the diagnosis of pregnancy and examination. Several types of cognitive biases are known to influence clinical decision-making, especially in the setting of limited knowledge and experience. One example relevant to this case is referred to as “framing bias,” which is a tendency to be influenced by how a question is asked or how a problem is presented;² the sonographer and radiologist were presumably influenced by the intern’s request for a “first trimester” ultrasound, so they were not looking for a large fetal head. In addition, the radiologist was likely influenced by “anchoring bias” - the undue influence that an initial impression has on evaluation of subsequently collected information. Anchoring bias may explain how the radiologist interpreted the finding of a fetal head as a “pelvic mass” instead of normal fetal anatomy.

Systems factors appear to have played an important role in these diagnostic errors as well.  One pervasive example was communication among the various providers. The original ultrasound was ordered as a “first trimester” examination. This error was carried through from the intern (and likely the supervising ED physicians), through the sonographer and radiologist. The sonographer recognized that the ultrasound examination was technically suboptimal due to obesity, and this should have been communicated directly to the radiologist, which would have an opportunity to repeat the pertinent portions of the study.  Other systems factors relate to insufficient supervision of the intern due to it being particularly busy in the ED at the time and the likelihood that the supervising physician did not examine the patient but accepted the story by the intern at face value, perhaps because the story appeared very straightforward.

Additionally, although fetal ultrasound is a well-established, ubiquitous, and very safe diagnostic modality,⁵ diagnostic ultrasound in pregnancy is susceptible to three categories of diagnostic errors: incorrect pregnancy date, technician error, and obesity.^3,4 Errors in each of these categories played a role in this case. Another potentially serious error is misreporting, such as the failure to report a recognized finding in the radiology report.

Systems Change Needed/Quality Improvement Approach

Keys to minimizing diagnostic errors involve processes such as clear communication, teamwork, breaking down silos, and involvement of patients in decision-making.⁷ Better communication, for example, would have informed the radiologist in this case that confidence in the quality of the scan was suboptimal, potentially leading to the sonographer repeating key components of the examination or having the radiologist perform another scan himself.

Strategies to ensure availability of appropriate expertise also are likely to reduce diagnostic errors, since lack of specific knowledge may be a critical component of mistakes. Appropriate training and experience, vigilance and a rigorous systematic approach to radiologic interpretation are all approaches that can be taken to minimize the types of diagnostic errors made in this case.^4,5 Putting processes in place to ensure adequate review of decisions made by less experienced team members, like interns in the ED, could have helped avoid various errors in this case.

Better engagement and communication with patients are likely to contribute to better outcomes. Patient engagement is seen as a potential intervention for error prevention, since patients can recognize if a diagnostic approach is following the correct pathway or not.

Finally, non-judgmental feedback about mistakes is a highly effective tool to prevent future diagnostic errors.^8,9 Hindsight can play an important role in understanding the contributions of various isolated facts to both diagnoses and diagnostic errors. Done well, providing feedback can improve communication, break down professional silos, strengthen work teams and increase the knowledge and experience of team members, thereby helping to ensure that the same errors are not repeated in the future.

Optimal systems to address diagnostic errors are under development. There are relatively few studies that have empirically assessed the effectiveness of different interventions. One systematic review identified 43 articles that included 6 tested interventions and 37 suggestions for possible interventions.¹ Several successful interventions are relevant to this specific case.  One is to have a comprehensive reevaluation within 24 hours to ensure that initial diagnostic impression is reconfirmed over time. In this patient’s situation, the patient had a re-evaluation via a consultation by a separate service (OB/GYN), which re-interpreted the discrepant data, and resulted in a correct assessment and appropriate intervention.

Take-Home Points

• Patient-related factors, cognitive errors, and systems factors are common categories of diagnostic errors, all three of which played a role in the failure to recognize that the patient in this case was in her third trimester of pregnancy and in early labor
• Communication among members of work teams is critical for avoiding perpetuation of cognitive errors
• Appropriate supervision of physician trainees is necessary to ensure high-quality patient care
• Diagnostic errors in the use obstetrical ultrasound can result from inadequate training and experience, and the lack of a systematic approach to examinations and interpretations; patient factors such as obesity can limit diagnostic detection of important findings.

Gary S. Leiserowitz, MD, MS
Professor and Chair
Department of Obstetrics and Gynecology
UC Davis Heath
gsleiserowitz@ucdavis.edu

Herman Hedriana, MD
Professor and Director,
Division of Maternal-Fetal Medicine
Department of Obstetrics and Gynecology
UC Davis Health
hlhedriana@ucdavis.edu

References

1. Singh H, Graber ML, Kissam SM, et al. System-related interventions to reduce diagnostic errors: a narrative review. BMJ Qual Saf. 2012;21(2):160-170. doi:10.1136/bmjqs-2011-000150
2. Itri JN, Tappouni RR, McEachern RO, Pesch AJ, Patel SH. Fundamentals of Diagnostic Error in Imaging. Radiographics. 2018;38(6):1845-1865. doi:10.1148/rg.2018180021
3. Paladini D. Sonography in obese and overweight pregnant women: clinical, medicolegal and technical issues. Ultrasound Obstet Gynecol. 2009;33(6):720-9. doi:10.1002/uog.6393
4. Pinto A, Pinto F, Faggian A, et al. Sources of error in emergency ultrasonography. Crit Ultrasound J. 2013;5 Suppl 1:S1. doi:10.1186/2036-7902-5-S1-S1
5. Pinto A, Reginelli A, Pinto F, et al. Errors in imaging patients in the emergency setting. Br J Radiol. 2016;89(1061):20150914. doi:10.1259/bjr.20150914
6. Casciani E, De Vincentiis C, Mazzei MA, et al. Errors in imaging the pregnant patient with acute abdomen. Abdom Imaging. 2015;40(7):2112-2126. doi:10.1007/s00261-015-0508-7
7. Singh, H. Diagnostic errors: A new chapter in patient safety science, policy, and practice.  Perspectives on Safety. Patient Safety Network. Agency for Healthcare Research and Qualtiy. January 1, 2016. https://psnet.ahrq.gov/perspective/diagnostic-errors-new-chapter-patient-safety-science-policy-and-practice
8. Singh, Hardeep MD, MPH; Upadhyay, Divvy K. MBBS, MPH; Torretti, Dennis MD Developing Health Care Organizations That Pursue Learning and Exploration of Diagnostic Excellence, Academic Medicine: October 29, 2019 - Volume Publish Ahead of Print - Issue - doi: 10.1097/ACM.0000000000003062
9. Sargeant JM, Mann KV, van der Vleuten CP, Metsemakers JF. Reflection: a link between receiving and using assessment feedback. Adv Health Sci Educ Theory Pract. 2009;14(3):399-410. doi:10.1007/s10459-008-9124-4