6: Errors of Clinical Reasoning and Decision-making in Veterinary Anesthesia

Errors of Clinical Reasoning and Decision-making in Veterinary Anesthesia

What we observe is not nature in itself but nature exposed to our method of questioning.

Werner Heisenberg, Physics and Philosophy (Penguin Classics, 2000)

Removing the stigma of bias clears the way toward accepting the capricious nature of decision making, and perhaps goes some way toward exculpating clinicians when their diagnoses fail.

Pat Croskerry (2003)

Clinical reasoning and decision-making are critical skills in all areas of medicine including anesthesia. They allow clinicians to efficiently and effectively make a diagnosis and establish a treatment plan. In doing so, information must be gathered from multiple sources, sorted, assimilated, and then formed as a coherent narrative or pattern so that the information makes sense. This information must be cross-referenced against knowledge pertaining to physiological and pathophysiological processes so a diagnosis can be made and treatments chosen.

For some complications encountered during anesthesia the diagnosis (cause) and treatment (intervention) are clear; the complication is straightforward and requires little more than pattern matching, memory, and action. For example, bradycardia following the administration of a potent opioid is a known side effect, one that is easily treated by the administration of an anticholinergic drug such as atropine or glycopyrrolate. These skills derive from the anesthetist’s education, training, experience and, to a certain extent, the amount of preparation and planning that has gone into the anesthetic protocol. The latter two factors––preparation and planning––are important as many complications can be predicted and specific monitoring and interventions strategies can be set in place even before any anesthetic drugs have been administered. However, where there is doubt and uncertainty about a diagnosis there is a requirement for clinical reasoning and decision-making.

Anesthetists face some unique stresses in the process of making clinical decisions, especially when dealing with anesthetic complications. The acute and sometimes immediately life-threatening nature of anesthetic complications impose real and significant pressures of time upon anesthetists to identify the cause of a problem and select and institute an appropriate intervention(s). But the manner in which an anesthetist’s thought processes develop in response to clinical information, typically acquired through their senses and an array of monitoring equipment, can be easily influenced by a variety of external and internal factors (Croskerry 2013; Reason 2004) that make these processes vulnerable to error.

The diagnostic process has been extensively studied and reported on in the literature, probably because it is the foundation of medical practice (Croskerry 2005). Many cognitive factors influence how diagnoses are made, and the summation of their effects result in an action consisting of a diagnosis and plan for intervention (treatment). According to Croskerry, the action can take one of three decision outcome modes (Croskerry 2005):

The correct decision-making style must be matched to the clinical situation for appropriate care to be delivered to the patient. Good clinical judgment can be defined as a clinician’s ability to choose the appropriate decision-making style in any given situation. Diagnostic errors, on the other hand, are those errors caused by faulty clinical reasoning or decision-making, for instance, when a diagnosis is unintentionally delayed (sufficient information was available earlier), a diagnosis is wrong (another diagnosis was made before the correct one), or a diagnosis is missed (no diagnosis was ever made) (Balla et al. 2012).

Specific data on diagnostic error in anesthesia are not readily available. However, depending on the specialty in human medicine, diagnostic errors may account for up to 20% of in-hospital patient deaths (Graber et al. 2005; Graber 2013). Graber identified cognitive factors as accounting for 58% of diagnostic errors, while organizational factors account for 39%, and technical factors for only 3% (Graber et al. 2005). It is important to note that cognitive factors alone or when combined with organizational factors, have a greater impact on patients than do organizational factors by themselves (Graber et al. 2005).

A study of claims submitted to the leading veterinary indemnity insurer in the UK indicates that cognitive limitations were an important source of error in individual clinicians, with slips and lapses identified as the most frequent types of errors (Oxtoby et al. 2015). Indeed, the bulk of the human medical literature suggests that most diagnostic errors are cognitive in origin, occurring in the intuitive mode of cognitive processing and as a consequence of one or more cognitive biases (Norman & Eva 2010). Croskerry (2003) compiled and defined the most comprehensive list of heuristics and cognitive biases pertaining to diagnostic error to date.2 This list was refined by Stiegler et al. (2012) to create a catalogue of anesthesia-specific cognitive errors. These biases are outlined under “Pattern matching and biases” in Chapter 2, and we often refer to them in the following cases and near miss vignettes.


Case 6.1

A young male polar bear weighing 318 kg was housed in a zoological exhibit that used a 6 m deep moat to separate observers from the observed. The bear was housed with a female polar bear, and he had a habit of jumping on her when she least expected it. The bear’s problems started one early spring day when he tried to jump on the female as she was walking close to the moat’s edge. The female saw the start of his jump and, having grown tired of these male shenanigans, she ducked; the bear became airborne (albeit briefly) and flew over her and down into the moat.

Remote examination of the bear in the moat confirmed that his hindlegs were fractured, but he was alert and attentive. Although housed in a zoo he was a dangerous animal, which made it impossible to physically examine him more thoroughly at that point in time. The attending veterinarian, recognizing that the bear’s nature precluded a full examination, was rightfully concerned about internal injuries. It was assumed that the bear would not eat if there were significant internal injuries, so he was offered an apple which he promptly ate, evidence confirming that the bear did not have internal injuries. The bear was scheduled for an orthopedic examination and possible surgery the following day at a tertiary care facility. As often happens with zoo’related events, the press and public were keenly interested in the bear’s condition and his surgical management.

At the tertiary care facility the bear was lightly anesthetized with etorphine and xylazine delivered via a pole syringe. Forty minutes later he was removed from the cage and masked to a deeper plane of anesthesia with halothane in oxygen delivered from a large-animal anesthesia machine. During induction a concern was expressed about the bear’s thick fur and the possibility that he would develop hyperthermia during anesthesia. After 10 minutes of mask induction the bear was intubated and the endotracheal tube was attached to the breathing circuit. Because of concerns about etorphine-induced hypoventilation, intermittent positive pressure ventilation was started at 5 breaths per min. After about 30 minutes the vaporizer dial setting was turned down. Radiographs (Figure 6.1) showed that the right tibia had a complete mid-shaft fracture and the distal end of the left femur was fractured. A team of orthopedic surgeons believed the fractures were repairable and the bear was prepared for surgery.

Image described by caption.

Figure 6.1 Leg fractures in a polar bear. Radiograph on the left is of the left femur showing a fracture of the distal end of the femur. The radiograph on the right shows a mid-shaft fracture of the right tibia.

During this time the bear was instrumented for physiological monitoring and a lingual arterial blood sample was obtained for hematocrit (44%), total solids (6.5 g dL−1), and blood gas analysis (pH 7.40, PaCO2 49 mmHg, PaO2 174 mmHg, and base balance +5.2 mEq L−1). The electrocardiogram showed a sinus tachycardia (150 beats per minute), and a temperature probe inserted via the esophagus to the level of the heart, indicated a body temperature of 39.5 °C. The tachycardia was attributed to both the slightly elevated PaCO2 and the hyperthermia, which in turn was attributed to the hospital’s warm environment and the bear’s dense fur coat. Once the bear was moved into the operating theater, a catheter was inserted into a lingual artery for monitoring blood pressure and sampling arterial blood for blood gas analysis. Throughout anesthesia the mean arterial blood pressure ranged from an initial high of 153 mmHg to a low of 72 mmHg with the majority of readings between 100 and 110 mmHg. Four arterial blood samples subsequently collected over the course of 4 hours of anesthesia showed an average PaO2 of 470 mmHg (range: 445–491 mmHg); an average PaCO2 of 45 mmHg for the first 3 hours of anesthesia and an average of 28 mmHg (range: 26–30 mmHg) for the fourth hour of anesthesia; base balance averaged +5.8 mEq L−1 (range: +5.2 to +6.5 mEq L−1). During the last two hours of anesthesia the bear was breathing spontaneously and hyperventilating, a response attributed to hyperthermia. Body temperature, which had increased from 39.5 °C to 41.1 °C, had decreased to 40 °C and remained at that temperature throughout the remainder of anesthesia. This decrease in temperature was a result of packing ice around the bear’s chest and forelimbs and eventually around the breathing circuit. At one point during anesthesia the anesthetist noted that the abdomen seemed “fluidy,” but no further observations were made or actions taken.

The surgery was completed within 5 hours and anesthesia was stopped soon thereafter. The bear was given an analgesic, transferred to his cage, and given diprenorphine (M50-50) to reverse any remaining effects of the etorphine; he seemed to be recovering as expected. Approximately one hour later the bear died while being transported to the zoo. At necropsy the bear was found to have a ruptured stomach and a complete diaphragmatic hernia (Figure 6.2).

Intraoperative photo of the lungs and remnant of diaphragm as taken from cranial abdominal cavity.

Figure 6.2 Abdominal and thoracic cavity of the polar bear at necropsy. View is from the cranial abdominal cavity toward the ruptured diaphragm and open thoracic cavity. A remnant of the diaphragm is hanging from the thoracic wall.

Initial analysis of the case

The referring veterinarian, a well-respected, knowledgeable, and competent zoo veterinarian, explained to the anesthesia and surgical team his thought processes and diagnostic approach to evaluating this polar bear’s injuries. Even though the bear was seriously injured, he was still a very dangerous animal, a fact that made hands-on physical examination impractical. The feeding of the apple to the bear seemed like a reasonable approach to determine if internal injuries existed. After all, what animal with serious internal injuries would eat? For the anesthesia and surgical team to question this assessment and conclusions of the zoo veterinarian did not seem collegial as he had more experience with zoo species than they did. Questioning his diagnostic approach and conclusions seemed to imply, at least in the collective mind of the team at the time, that they were questioning his medical competency. This mindset, of course, is exactly what Reason cautions against: professional courtesy must not get in the way of checking colleagues’ knowledge and experience, particularly when they are strangers (Reason 2004).

The error in this case was two-fold: (1) failure to request a more thorough physical examination of the bear for fear of offending the zoo veterinarian or at least seeming to be uncollegial; and (2) assuming there were barriers, whether real or imagined, to doing a more thorough medical examination. After all, once the bear was anesthetized, all concerns about examining a dangerous animal were moot and all barriers to performing a thorough physical examination were removed. But other biases and perceived barriers persisted in the minds of the team members, but these were never openly discussed.

Investigation of the case

The circumstances surrounding this case, the context in which it occurred but not mentioned in the case presentation, explain many of the biases that set the stage for the errors in decision-making that were the basis of this diagnostic error. Some of the factors were non-medical in nature and external to the practice but significantly affected the management of this patient. There were also patient-related factors.

A perceived barrier to conducting more diagnostic techniques was that a definitive diagnosis had already been made (premature closure), and the admonition that none of the bear’s fur was to be clipped for minor techniques, such as needle aspirates of the abdominal cavity, until it was determined if the fractures could be repaired. Once it was decided that the fractures could be repaired, other diagnostic tests became irrelevant in the minds of the team members because other injuries were no longer considered; attention was focused (coning of attention) on surgical repair of the fractures.

Soon after the bear was injured the local press picked up on his condition and closely followed his progress. Thus the team came to believe that there was a great deal of pressure to present as positive a picture as possible concerning the bear’s condition and surgical management. In addition, it had been made clear to the team that if the fractures could not be repaired, the bear would be euthanized and its body prepared for public display. No member of the team wanted to euthanize the bear, one that the public knew and loved; as one team member stated, “Who wants to euthanize a bear with a name?” Although euthanasia is a humane procedure, one that is performed by veterinarians for the best of medical reasons, not infrequently it is viewed as a sign of failure. In this particular case none of the team members wanted to be the veterinarian recommending euthanasia of the bear. As a result, there developed a “can do” attitude amongst team members in terms of the bear’s surgical management. This mindset created blinders—tunnel vision—on the team’s thinking and made it difficult for them to consider options other than repairing the fractures.

There were two other subtle influences on the anesthesia and surgery team regarding euthanasia. Public sentiment surrounding the creation of this teaching hospital a few years previously was best described as polarized; many supported and many opposed its creation. This led to a perception, whether real or not, that the teaching hospital was under public scrutiny, especially by the state’s legislature.

The teaching hospital was a short walk from a large regional medical center. One day while waiting in the lunch line at the medical center, two surgical nurses were overheard commenting on how stupid veterinarians were because they did not know that fractured bones in horses could be repaired. This conversation was prompted by a recent on-track euthanasia of a race horse that had fractured its leg during a race. These were subtle but real influences on the team members, influences that contributed to a mindset that considered euthanasia as a sign of failure.

Obviously this bear was a wild animal and in its normal habitat it would be a predator. How might this reality affect its management? The assumption that the bear would not eat the offered apple if it had severe internal injuries failed to consider the nature of the animal. There is no survival advantage for a predator to show signs of disease, for by doing so the predator becomes prey. This reality of the animal’s natural behavior was not considered by members of the team.

To summarize, there were several factors that prevented the entire team from correctly identifying the extent of this bear’s injuries:

  • Failure to question a colleague’s assessment and diagnosis of the bear’s condition.
  • A bias that euthanasia of the bear would be viewed by many, especially the public, as a sign of failure.
  • An assumption that various findings such as tachycardia and hyperthermia were due to the unique nature of the bear when in fact these findings were easily explained by less exotic considerations attributable to the trauma, such as peritonitis and fever.
  • Holding on to admonitions about how the bear was to be managed long after the justifications for those admonitions had passed.

The bear died as a result of undiagnosed and untreated severe internal injuries. Had these injuries been diagnosed at presentation to the tertiary care facility, the bear would have been promptly euthanized. This, of course, was the one thing everyone wanted to avoid, but which would have been the most medically sound management of the bear. The chances of surgically and medically managing the bear’s fractures and internal injuries were nil; it was a wild animal and the type of intensive hands-on care he would have required postoperatively was neither feasible nor safe.

Case 6.2

A 5-month-old 4-kg male Shetland sheepdog was referred to a veterinary teaching hospital because of hindlimb paralysis that was unresponsive to treatment with steroids. Two days previously, the owners’ child had been playing roughly with the dog just prior to the onset of paralysis.

Initial examination revealed a bright and alert young dog with paralysis of both hindlimbs. Heart and respiratory rates were 84 beats per minute and 24 breaths per minute, respectively; rectal temperature was 38.5 °C, hematocrit (Hct) 34%, and plasma total protein concentration 50 g L−1. Lateral and ventrodorsal radiographic views of the vertebral column revealed that the caudal end of lumbar vertebra 5 (L5) was fractured and the distal fragment and vertebral column were displaced cranially and laterally to the left thus fully compromising the neural canal (Figure 6.3). There was no evidence of free fluid in the abdominal cavity. With the owners’ informed consent, the dog was anesthetized for surgical reduction and stabilization of the fracture.

Image described by caption.

Figure 6.3 Lateral radiograph of the spine of the 5-month-old 4-kg male Shetland sheepdog. The fracture is at the caudal end of L5 with the distal fragment and vertebral column displaced cranially and laterally to the left of midline; the neural canal is fully compromised in that it is not aligned.

Reused from: Ludders, J.W., et al. (1998) Anesthesia case of the month. Journal of the American Veterinary Medical Association 213(5): 612–614. With permission of the publisher.

The dog was catheterized intravenously in the ICU and subsequently premedicated by an anesthesia resident with diazepam, butorphanol, and glycopyrrolate, all given intravenously. After premedicating the dog it was anesthetized with thiamylal, intubated, started on isoflurane (1.2%) in oxygen (2 L min−1) and allowed to breathe spontaneously. The dog was monitored with an electrocardiogram, and systolic arterial blood pressure was indirectly measured and recorded using a Doppler device (Model 811b; Parks Medical Electronics Inc., Aloha, OR, USA) with its probe placed over the plantar common digital artery and the cuff placed around the leg above the hock and attached to a sphygmomanometer.

At induction, the dog’s heart rate was 143 beats per minute and systolic blood pressure was 110 mmHg. Throughout the 3 hours of surgery, the heart rate fluctuated between 100 and 170 beats per minute, and systolic arterial blood pressure fluctuated between 100 and 140 mmHg. The isoflurane vaporizer setting varied from 1.0 to 1.5% during anesthesia, and nitrous oxide (50%) was used to supplement isoflurane. Vecuronium bromide (a neuromuscular blocking drug) was administered to provide muscle paralysis and facilitate surgical reduction of the fracture. During the following 2 hours ventilation was controlled with a mechanical ventilator at 11 breaths per minute. Methylprednisolone sodium succinate and cefazolin were given at 50 and 120 minutes, respectively, after the start of anesthesia.

While the surgeon was stabilizing the fracture with Steinmann pins, the Doppler’s audible signal (indicating the flow of blood through the artery) stopped and could not be re-established even after the location and position of the probe were checked and adjusted on the hindleg. Another Doppler device was obtained, its probe was placed over the dog’s palmar common digital artery, and a strong signal was obtained; surgery and anesthesia continued for another hour.

Near the end of surgery the dog was prepared for transport from the operating room to radiology. Nitrous oxide was discontinued 10 minutes before transport and the dog was weaned from the ventilator. For most of the anesthetic period, the dog was judged to be at a moderate plane of anesthesia. However, while preparing the dog for transport and during radiography, the dog was judged to be at a lighter plane of anesthesia based on increases in systolic arterial blood pressure (from 110 to 140 mmHg) and respiratory rate (from 12 to 15 breaths per min). To maintain an adequate plane of anesthesia, the isoflurane vaporizer dial setting was increased from 0.9% to 1.5%. The radiographs showed good reduction and alignment of the fracture (Figure 6.4).

Image described by caption.

Figure 6.4 Lateral radiograph showing reduction and fixation of the vertebral fracture in the 5-month-old 4-kg male Shetland sheepdog.

Reused from: Ludders, J.W., et al. (1998) Anesthesia case of the month. Journal of the American Veterinary Medical Association 213(5): 612–614. With permission of the publisher.

To make recovery smoother, the dog was given butorphanol for analgesia and diazepam for muscle relaxation and tranquilization. As the dog recovered from anesthesia, it appeared uncomfortable in that it whined and looked at its hindquarters. Twenty-five minutes later, oxymorphone and acepromazine were given, both intravenously, because the dog seemed more painful and anxious. Despite these drugs, the dog continued seemingly to be uncomfortable and to focus its attention on its hindquarters. At this time the mucous membranes of the penis were noticed to be pale. In general, recovery was not progressing as expected.

While in the radiology suite and then in the recovery room, evaluation of the postoperative radiographs raised questions amongst the anesthesia-surgical team members about the possibility of one or more Steinmann pins affecting blood flow to the hindquarters, specifically flow through the abdominal aorta.

Pulses in the femoral and dorsal pedal arteries could not be palpated. Application of a Doppler probe to either pedal artery did not yield an audible blood-flow signal, but a signal was detected when the probe was placed over a palmar common digital artery in a forelimb. On the basis of these findings and the dog’s behavior, it was re-anesthetized for surgical exploration of its abdomen. The dog was mask-induced to anesthesia with isoflurane in oxygen, and anesthesia was maintained with isoflurane in oxygen and nitrous oxide. Fentanyl was administered for additional analgesia, and vecuronium bromide for muscle paralysis so as to facilitate retraction of the abdominal muscles and abdominal exploration. Exploration of the abdomen found that a Steinmann pin had caught the adventitia of the abdominal aorta approximately 1 cm cranial to the aortic bifurcation, and the aorta was wrapped around the pin, thus occluding aortic blood flow. Before cutting the pin and freeing the aorta, Rummel tourniquets were loosely placed around the aorta cranially and caudally to the obstruction. Shortly after freeing the aorta, blood flow returned to the hindquarters and bleeding occurred within the peritoneal cavity. The hemorrhage was stopped by occluding the aorta with the Rummel tourniquets. The source of hemorrhage was a tear in the vena cava, which was repaired with suture. After removing the Rummel tourniquets the other three pins were trimmed, the abdominal cavity was flushed with warm sterile saline (0.9% NaCl) solution, and the body wall was closed in a routine manner.

Only gold members can continue reading. Log In or Register to continue

Aug 14, 2022 | Posted by in SUGERY, ORTHOPEDICS & ANESTHESIA | Comments Off on 6: Errors of Clinical Reasoning and Decision-making in Veterinary Anesthesia

Full access? Get Clinical Tree

Get Clinical Tree app for offline access