Question: A middle-aged woman developed cellulitis after sustaining multiple mosquito bites in her lower left leg. The area of infection did not appear to reach the knee, which housed a prosthesis implanted there 3 years earlier. Over the next few days, she had significant knee pain, which was attributed to the surrounding cellulitis. Her pulse rate reached 105 beats per min, and her temperature was 101° F, but she was continued on oral antibiotics as an outpatient. Later that evening, she collapsed at home.
Which of the following is best?
A. Fever and tachycardia alone are enough to make the diagnosis of systemic inflammatory response syndrome (SIRS) and should have raised sepsis as a cause.
B. In septic patients, even a short delay in antibiotic administration can significantly affect morbidity and mortality.
C. Failure to diagnose is the most common basis for a medical malpractice claim.
D. The doctor may have anchored his diagnosis on the mosquito-bite incident, and should have considered a septic joint and/or sepsis in the differential.
E. All are correct.
Answer: E. “Failure to diagnose” is a legal term, whereas in medical usage we tend to use terms such “missed diagnosis,” “overlooked condition,” or “diagnostic error.” If such failure is shown to be a breach of the standard of care and is proven to be a proximate cause of the patient’s injury, then a case for medical negligence is made out. Even if the situation is atypical or complex, there still is the duty to refer, if customarily required, to an appropriate specialist, and failure to do so may also constitute negligence.
Diagnostic errors tend to cause the most severe injuries, especially in hospitalized patients. Roughly 5% of autopsies uncover a diagnostic error that was amenable to appropriate treatment, and some 50,000 annual hospital deaths may be the result of a delayed, incorrect, or overlooked diagnosis.
Failure to diagnose occurs in both outpatient and in-hospital settings, recurring examples being myocardial infarction, dissecting aneurysm, pulmonary embolism, appendicitis, ectopic pregnancy, meningitis, cancers, and fractures.
In a records-review study covering a large urban Veterans Affairs facility and an integrated private health care system in a primary care setting, the authors reported that pneumonia, heart failure, acute renal failure, cancer, and urinary tract infections (UTIs) were frequently missed diagnoses.2 They identified 190 diagnostic errors over a 12-month period in 2006-2007, and they attributed them to “process breakdowns” involving the practitioner-patient clinical encounter, referrals, patient factors, follow-up and tracking of diagnostic information, and interpretation of test results. Deficiencies in bedside history taking, physical exam, and test ordering were common; significantly, there was no documentation of an initial differential diagnosis in 80% of misdiagnosed cases.
Malpractice carriers regularly compile data regarding the nature of their covered losses, and their reports on diagnostic errors, although not subject to the usual scientific peer review, have generally corroborated the published literature.
For example, data from 2009 to 2013 collected by , a large malpractice mutual insurance company on the West Coast, impute almost half of all general medicine claims to diagnostic errors.3 The cases, frequently involving cancer and heart disease, resulted in high-severity injuries and death. Lapses in clinical judgment, communication, and patient-related behavior issues were the primary contributing factors that affected the diagnosis-related claims. Pitfalls included errors in patient assessment, diagnostic processing, provider follow-up, and referral to specialists.
Recent reports have drawn attention to sepsis, an example of SIRS, as an important missed diagnosis, often with deadly consequences. It has been pointed out that if a sepsis case goes to trial, jurors will immediately learn that mortality rates are increased if antibiotics are delayed, even for a short period. In one study, each hour’s delay increased mortality by 7.6%, mortality being 21.1% if antibiotics were given in the first hour, compared with 58% if delayed by more than 6 hours.4
To avoid suits, physicians should be alert to seemingly minor vital sign changes, such as new tachypnea or tachycardia. Notably, patients can have severe sepsis and septic shock without fever or hypothermia. Uncomplicated sepsis is common and can quickly progress to severe sepsis, with organ failure and septic shock. The has estimated that more than 750,000 individuals develop severe sepsis in North America each year, with mortality around 50%.
The Sullivan Group,which comprises a team of professionals dedicated to perfecting a system solution that reduces medical error and improves patient safety, recently published a wrongful-death narrative from undiagnosed sepsis.5 The decedent gave birth to her first child after 24 hours of labor, sustaining severe vaginal and rectal tearing. Three days later, she began experiencing chills, nausea, worsening vaginal pain, and fever. Her temperature reached 101.9° F (38.8° C). The following day, 4 days after delivery, she was seen by a nurse practitioner in the emergency department with symptoms of nausea, abdominal and back pain, and fever. She was tachycardic at 115 per minute.
The presence of fever plus tachycardia should have raised the diagnosis of SIRS, especially in view of her abdominal pain and a recent complicated delivery. Instead, the practitioner diagnosed a UTI and discharged her on antibiotics. That same afternoon, she collapsed and was admitted for sepsis. Despite an emergency hysterectomy, her condition worsened, she developed multiple organ failure and septic shock, and died the next day. The source of her sepsis was endometritis, and the jury returned a $20 million verdict for the plaintiff.
Observers of medical errors point to our recurring failure to continue to consider alternatives after forming an initial tentative diagnosis, and warn us about the various cognitive biases familiar to behavioral economists but ignored by many doctors.6 These include anchoring bias, in which one is locked into an aspect of the case; framing bias, in which there is misdirection because of the way the problem was posed; availability bias, in which things are judged by what comes readily to mind, such as a recent experience; and confirmation bias, in which one looks for confirmatory evidence of one’s preferred diagnosis while ignoring evidence to the contrary.
In the case outlined earlier, the Sullivan Group noted that the practitioner did not consider sepsis because of cognitive bias, anchoring, and premature closure. The trial documents indicated that the urinalysis did not show bacteria, but the practitioner may have settled – prematurely – on the UTI diagnosis, based on the presence of WBCs in the urine and her obstetrics history. Having anchored on that thought process and prematurely closed her decision making, the practitioner then ignored the elevated white blood cell count with a left shift, and a depressed platelet count of 50,000. Perhaps UTI was a reasonable consideration in the differential, but the working diagnosis of sepsis should have been first and foremost.
Dr. Tan is emeritus professor of medicine and a former adjunct professor of law at the University of Hawaii, Honolulu. This article is meant to be educational and does not constitute medical, ethical, or legal advice. Some of the materials may have been published in earlier columns in Internal Medicine News, and can be accessed at