What the Substitution Test can tell us about negligence


What the Substitution Test can tell us about negligence


Except in the rare instance of criminal negligence or intention, responsibility for harm is not likely to be a black and white issue, and asking whether another person would have done the same thing can shed useful light on the situation, says Dan Cohen of Datix.

When something goes wrong, when a patient is harmed as a result of what is commonly called “error in medicine”, as in Lucian Leape’s seminal article of the same title (JAMA, 1994), there is a natural inclination to find someone to blame. Although numerous analyses of patient safety incidents have revealed that healthcare systems and process issues contribute to, or are cause factors in, most harmful incidents, the fact is that in the realm of diagnostic error, someone is likely to be held accountable.

Clinical reasoning—the process of diagnosis—‚Ä®begins with determining a list of likely diagnoses and then constructing an investigation plan to refute and/or confirm a particular diagnosis. Usually one physician is the primary decision-maker, the one responsible for care and the one held accountable.

All too frequently, harmful safety incidents relate directly to errors in diagnosis, ie, incorrect, missed and/or delayed diagnoses, and these categories are not mutually exclusive by any means. They interface, interlace and run together.

It seems easy and logical to hold someone accountable and the professional liability industry is focused on establishing personal and/or institutional responsibility for errors and harm. For healthcare professionals being held accountable may result in actions by supervisors, credentials committees and state boards of medicine that can affect careers and reputations, so matters relating to responsibility and accountability must be closely examined and approached with respect and caution.

Accountability may seem like a straightforward concept, but on deeper reflection, dissecting the contributing and causal factors that cumulatively result in error should give pause for thought. Our intentions may be benevolent, but human factors and the challenges of diagnostic cognition, often encumbered by external factors and constraints, all contribute to diagnostic error.

An important principle to keep in mind when addressing the issue of accountability, and the professional ramifications of being found wanting, is whether other physicians of similar background, training and experience, working under the same environmental and situational circumstances might, or would, have made the same errors resulting in incorrect, missed and delayed diagnoses.

This query is commonly framed as the Substitution Test, named by J. Reason in Managing the Risks of Organizational Accidents in 1997. If others, substituted into the same clinical scenario and confronted with the same human factors liabilities, contributing and causal factors, would, or might, have committed the same errors, then the professional actions of supervisors and credentials committees should be moderated.

In such circumstances it is apparent that problems within the systems and processes of care are, at least in part, substantially responsible for the errors. The factors cumulatively resulting in errors become understandable and thus steps can be put in place to ameliorate or modulate circumstances contributing to error. This is how we learn from errors, how we improve—it’s what matters most.

Individuals should be held accountable for their actions, but understanding the texture and complexities of factors contributing to clinical decision-making should affect just what professional accountability actually means. Except in the rare instance of criminal negligence or intention, accountability will never be a black and white issue.

Administrative actions should be appropriate to circumstances and severe punishment should be reserved for exceptional instances of negligence determined only after thorough, unbiased reviews.

Case study

A pediatrician involved in a delayed diagnosis incident was severely admonished by his hospital leadership. He was responsible for the care of a child whose delayed diagnosis was associated with clinical deterioration and a prolonged hospitalization, but whether he should be held accountable for all of the errors or even his specific insufficiencies on the day in question is another matter.

A four-year-old girl, a recent immigrant from Haiti, presented in mid-winter during influenza season to a very busy metropolitan emergency department. She’d had 48 hours of nasal congestion, modest fever, vomiting, lassitude, anorexia, and a history of recurrent urinary tract infection.

Initially she was seen by a newly graduated nurse, who was assigned to triage for the first time and who ordered laboratory work per an “infectious illness” protocol. The patient then was put in the queue for non-urgent physician evaluation. Laboratory results in this ER were usually available within 30 to 45 minutes after the specimens were obtained.

The child had been brought to the ER by her mother, was unknown to the healthcare system, had no medical records available and, although her mother spoke some English, French was her native language. After a 2½-hour wait, the child was seen by a pediatrician who obtained additional history of recent onset, frequent urinary incontinence and nocturnal urination and a history of maternal sickle cell trait. The father’s health history was unknown.

On examination, the child appeared to be very pale and dehydrated but was responsive and engaging during the exam. Her respiratory rate was slightly elevated but her lungs were clear, and she did not cough during the exam. The pediatrician concluded that most likely the child had a upper respiratory infection/viral syndrome, but felt it best to wait until the laboratory results returned to rule out a urinary tract infection.

He ordered a chest X-ray because of concerns about silent pneumonia. As the previously ordered lab results were not yet in the computer, he asked a desk clerk to check on them and went on to another patient awaiting an urgent intervention.

The child was sent to the X-ray department and did not return for over an hour. On return from the X-ray department, the nurses noted that she had worsened considerably. She was now alternatively somnolent and then combative. Her temperature had risen to 39oC, she was breathing very rapidly and her blood pressure had dropped to 80/60. In addition, her bed sheets were soaked with urine as she had become incontinent, and the nurses noted a fruity odor to her breath.

The pediatrician reassessed the child immediately, noted the fruity odor to her breath and obtained a bedside blood glucose test confirming severe hyperglycemia. He was then informed for the first time that the previously ordered laboratory work had never been obtained. The chest X-ray was clear, however the clinical picture now was more consistent with diabetic ketoacidosis and sepsis.

"The pediatrician was managing 10 patients simultaneously, which included three with very complex illnesses."

After drawing appropriate blood work and urine studies, IV fluids, insulin and antibiotics were started and the child was transferred to the pediatric ICU, where over two hours she descended into diabetic coma associated with hyperosmolar syndrome, which lasted for one week. In addition, she was found to have Streptococcus pneumonia septicemia and hemoglobin SC disease (combined hemoglobin S and hemoglobin C), a predisposing factor for severe pneumococcal sepsis. She nearly died.

Her total time in the ER had been six hours.

A quality assurance review of this case found that there had been a delay in diagnosis resulting in harm and related to the following factors:

• Insufficient triage by an inexperienced nurse who failed to recognize the serious nature of the clinical presentation, thus delaying physician evaluation.      

• Insufficient quality control procedures to assure that specimens for laboratory studies have been obtained and processed.  

• Inadequate initial physical examination by the pediatrician, who failed to recognize the fruity odor to the child’s breath, which would have indicated diabetes mellitus.         

• Inadequate attention to detail by the pediatrician, who should have personally called the laboratory to query the results as part of his initial evaluation, as results were usually returned within 45 minutes of specimens being obtained.

• Failure by the pediatrician to consider sickle cell disease with associated sepsis and diabetes mellitus with or without ketoacidosis as alternatives to viral syndrome and urinary tract infection early in his clinical assessment.

The pediatrician was held responsible and accountable for the delayed diagnosis resulting in adverse outcomes of prolonged coma, hospitalization and risk of death. He was sanctioned and reprimanded by the chief of the clinical staff and the credentials committee, although no adjustments were recommended to his privileges.

Further investigation

At the request of the entire emergency department physician staff, an external evaluation was undertaken of the care provided and the following question was posed: given the environmental and situational circumstances on the evening of this child’s evaluation, and the physiological, social and human factors that contribute to human performance, might other pediatricians of similar background and experience have made the same judgements regarding this child’s care?

That is to say, would other pediatricians, if substituted for this pediatrician, have performed similarly?

The external peer review revealed the following contributing factors:

• On the evening of this incident, the ER wait time for patients not deemed urgent via triage was between 2 ½ and 3 hours, as a record number of patients had presented for evaluation. This was coincident with a peak in community illnesses consistent with influenza and other viral respiratory and gastrointestinal syndromes.

• The nurse responsible for triage was inexperienced and should not have been given this responsibility in a high volume, high acuity situation without senior nursing backup.

• French language translators were not available even though this ER served a large Haitian community.

• The pediatrician responsible for this child’s evaluation had been on duty for 10.5 hours and had not had time for a break or for a meal for seven hours.

• The pediatrician was managing 10 patients simultaneously, which included three with very complex illnesses.

• Another patient requiring an urgent arterial line insertion, coincident with impending respiratory failure, was awaiting the pediatrician’s attention and this was on his mind.

• Just prior to evaluating this child, the pediatrician had had a very confrontational telephone conversation with a gynecologist who had not responded in a timely fashion to his request for urgent evaluation of a teenager with fever, pelvic adnexal pain and uterine cervical discharge, consistent with pelvic inflammatory disease requiring admission.

• The pediatrician’s spouse had recently given birth and his sleep pattern at home was disturbed by nocturnal activities related to a colicky baby.

• The pediatrician’s spouse was experiencing post-partum depressive symptoms, so life at home was a bit overwhelming at the time of the incident.                 

• The ER staff had insufficient physician resources on hand to accommodate absences of key personnel during this very busy time of year, so paternity leave was not the norm.

• Given the circumstances at hand, the situational and social/family distractions, the pediatrician applied an availability heuristic in the course of his diagnosis. He reached for a common pattern of historical information and findings and applied it in his reasoning. Although this was an error in thinking, given the confounding contributing factors affecting his performance on the evening in question, degradation in his performance and insufficient attention to details of management would neither be unlikely nor unexpected.

• Clearly, the pediatrician was professionally and socially task-saturated, working beyond normal capacity. 

• There was insufficient evidence to conclude that even if the correct diagnoses had been made two or three hours earlier, this would have made any difference in the subsequent course of hospitalization. The diabetic hyperosmolar coma could not specifically be attributed to the delayed diagnoses. 

• Other pediatricians of similar background and experience, working under similar environmental and social stressors, might easily have come to similar or identical conclusions, and thus the Substitution Test in this instance was affirmed.

• Insufficiencies in management and staffing of the ER should be urgently addressed.

Responsibility, accountability, and negligence … it’s not always as simple as it seems. Substitution may provide clarity where confusion abounds.


Dr Dan Cohen is International Medical Director of Datix. He trained in pediatrics and hematology/oncology at the Boston Medical Center, Boston University and the Boston Children’s Hospital, Dana Farber Cancer Institute, Harvard Medical School. He is a Senior Fellow of the Royal College of Paediatrics and Child Health and a Fellow of the American Academy of Pediatrics. He can be contacted at: dcohen@datixusa.com

Substitution Test, Datix, Dan Cohen ER, Negligence, US