Slips and lapses are almost inevitable in our daily lives, and the same is true, for example, in a patient with diabetes. Nearly all of us have experienced leaving our home and noting that we had forgotten something we intended to take with us. For a patient with diabetes, however, if it was their insulin pen, or glucose meter, or glucose tablets, there may be adverse consequences. Slips or lapses are far more common when people are distracted, under stress, or interrupted in their daily routine, as for example, when the person drawing up a dose of insulin receives a phone call in the middle of their sequence of steps. Another example occurs when a person preparing to administer the insulin is engaged in a conversation separate from the task at hand. Then the probability of a slip or lapse increases and an error in insulin administration is much more likely.
Slips and lapses
The remedy offered for such slips or lapses may be straightforward. In our clinical setting, we make clear to patients and family that it is important never to draw up insulin or administer other medications when trying to do something else, and to double-check to be sure they are giving the right type of insulin, at the right dosage, in the correct location, and for the right reason.
Similar types of mechanisms of errors occur frequently with medical and nursing staff. In one large study, reported by Bates and his colleagues, of the non-surgical errors in a hospital, 37% were adverse drug events due to errors in prescribing or administering medication, and errors regarding insulin use were one of the most likely agents to lead to an adverse drug event.
Although many slips and lapses are easily corrected, in a system without checks they can easily become injurious. Checklists are often used to reduce the likelihood that such errors cause injury in critical settings.
Rule-based errors are quite different than slips or lapses, and can be much more difficult to remedy. An example of a rule-based error is the use of sliding-scale insulin (SSI) monotherapy in inpatient diabetes management. There are abundant data that show sliding-scale insulin monotherapy for diabetes management is inferior to basal-bolus insulin therapy and, at times, is dangerous. Deaths have been reported where SSI monotherapy alone in a hospitalized patient with diabetes has led to in-hospital diabetic ketoacidosis (DKA) or hyperglycemic hyperosmolar syndrome (HHS).
Another more recent example of a rule-based error is a diabetic patient’s use of a defective insulin dose calculator app for their Android device. Unfortunately, as a recent analysis showed, many of the insulin dose calculators present on both Android and IOS devices appear to be poorly designed and their output is often incorrect, another potentially dangerous example of a rule-based error.
Rule-based errors survive because their use may simplify the work flow of the person who uses it, but the user may not be aware that the rule is faulty. Or, as common in the case of sliding-scale insulin algorithms, an alternative approach may not be an approach with which the person using the SSI algorithm is familiar. It is very important to ask the patient or provider, if they are using an algorithm for insulin or drug adjustment, about the specifics of the algorithm used and to determine whether the algorithm, or rule, is appropriate for the situation.
Diagnostic errors, while not the most common of types of medical errors, are the most serious in many respects. They are most often associated with serious and injurious errors, in large part because most people have a much more difficult time spotting their own errors in reasoning than they do identifying errors made by others.
A number of common cognitive errors tend to make diagnostic errors more difficult to identify by the diagnostician. There is a tendency to favor information that favors ones point of view and not to emphasize alternative explanations. In a study by Graber on diagnostic errors in medicine, in addition to the strong influence of systems issues in contributing to diagnostic errors, Graber found in 100 cases of diagnostic errors, faulty synthesis (82.6%) was the most common error, and faulty data gathering (14%) was also important. Surprisingly, inadequate knowledge or skill appeared to have a much lesser role (3.6%). An alternate way of expressing this is that the diagnostic errors most commonly made were due to how the physician collected and put together the data to formulate the diagnosis.
Premature closure of the diagnostic inquiry is a very common cognitive error which often plays a significant role in such errors. Without a culture of safety in which a colleague may review the evidence to support the diagnosis, a busy clinician may use pattern recognition as a rule-of-thumb to guide them to a diagnosis, which may be in error.
Cognitive scientists often distinguish two patterns of cognition, termed Type 1 and Type 2 cognition. Type 1, as used by Kahneman, refers to the rapid, almost automatic type of cognition that we use extensively in ordinary existence. It is what we do while driving a motor vehicle, or a jazz musician does while playing with his colleagues, how we size up whether a situation appears dangerous or safe. Pattern recognition falls into this category. But Type 1 thinking is quite prone to cognitive biases, is not at all quantitative, and is relatively error prone.
Type 2 thinking, in contrast, is much slower and deliberate. We would be more likely to use this method in performing mathematics or solving a problem in physics. We would hope our accountant uses this type of thinking. Type 2 thinking, however, is also prone to cognitive biases and is too slow to accomplish many important tasks. Typically, a person uses both kinds of thinking, but diagnostic errors are more likely if only Type 1 thinking is used.
In diabetes care, diagnostic errors can lead to serious delay in therapy, leading to death. It is for this reason that I recommend that, especially in critical care situations, the diabetes care team should always carefully check that the evidence to support the diagnosis is present and to ask whether other diagnostic possibilities might also cause the same clinical presentation.
An example of the danger of premature closure of the diagnostic process is when a busy emergency room physician assumes that the abdominal pain in a young person is appendicitis instead of checking further and finding that the child has diabetic ketoacidosis which is the true cause of the pain.
TABLE 1: Classification of Medical Errors (Click to enlarge)A more common error occurs when elderly patients with focal neurological signs due to hyperglycemic hyperosmolar states do not receive appropriate treatment because a faulty diagnosis of CVA is made, based on the neurologic findings alone, without inspecting the laboratory data. This delay in therapy in these critically ill patients may commonly result in death due to medical error. Likewise, diverse symptoms of hypoglycemia are often also missed, also with serious consequences.
Reason J. Human Error. Cambridge, UK: Cambridge University Press; 1990
Umpierrez GE et al. Management of Hyperglycemia in Hospitalized Patients in Non-Critical Care Setting: An Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab. 2012;97(1):16-38
Graber ML, Gordon R. Diagnostic error in internal medicine. Arch Intern Med. 2005;165(13):1493-1499
Kahneman, D. Thinking, fast and slow. New York: Farrar, Straus and Giroux; 2011
Huckvale K et al. Smartphone apps for calculating insulin dose: a systematic assessment. BMC Medicine. 2015;13:106 (1-10)
Moghissi ES et al. American Association of Clinical Endocrinologists and American Diabetes Association consensus statement on inpatient glycemic control. Diabetes Care. 2009;32:1119-1131
Umpierrez GE et al. Randomized study of basal bolus insulin therapy in the inpatient management of patients with type 2 diabetes undergoing general surgery (rabbit surgery). Diabetes Care. 2011;34:256-261
Umpierrez GE et al. Randomized study of basal-bolus insulin therapy in the inpatient management of patients with type 2 diabetes (Rabbit 2 Trial). Diabetes Care. 2007;30:2181-2186