Errors in management
In the last two decades, both the medical profession and the public have noted with dismay that two essential properties of health care, quality and safety, are often found to be suboptimal. Although the domains of quality and safety overlap, most authorities think that highest priority should be that the health care delivered is safe and free of errors that cause injury. Medicine can be considered an endeavor with high opportunity for error, but diabetes care, because of its complexity and the unique vulnerabilities of patients with diabetes, carries an especially high risk for errors in care. In the following entries we will explore numerous aspects of errors in diabetes management and present strategies of how best to protect our patients against these errors.
As diabetes care increases in both sophistication and complexity, it has become increasingly clear that there is frequently a considerable gap between what important clinical outcomes are possible for an individual patient, and what actually occurs.
Problems due to reduced access to timely care, limited financial resources, the negative impact of other comorbid conditions, and unrelated accidents often play a significant role in diminishing the likelihood of a favorable clinical outcome for the patient. But errors in diabetes management, both self-management and provider errors, also diminish the likelihood of favorable clinical outcomes and lead to increased morbidity, complications, disability, and death.
Only recently have we been able to effectively measure the frequency with which injurious errors occur in diabetes management. Even though the data is limited, most authorities now think that injurious errors in both diagnosis and management of diabetes care are significantly underreported in the medical literature. Although we have only imperfect estimates of the frequency of significant medical errors in diabetes management, there are credible estimates from a variety of sources of the impact of medical errors on mortality rates in general. In 2001 the Institute of Medicine in the United States published a report which provided an estimate that between 44,000-98,000 people die yearly in the U.S. as a result of medical errors. But a later paper suggested that the actual number may be more than twice as high annually (249,000).
In 1997, we reported outcome data regarding a prospective cohort study of 780 patients with diabetes, examining the effect of intensive therapy upon the outcomes of death and end-stage renal failure over a 15-year period. Subsequent analysis of the 283 patients in the comparison cohort who died during the 15-year study period revealed that 20.49% of those who died had a major error in their care which contributed to excess morbidity, or complications, or disability, or death. Of those who died, 4.24% died within 24 hours of a catastrophic error in diabetes management. In this group the reviewers found that 33% of the catastrophic errors which led to death within 24 hours were errors in insulin management. However, much larger studies of mortality data are needed. We hope that more data will become available as registries with richer clinical data sources become accessible.
Given the current underreporting of adverse medical events that occur both in hospitals and in the outpatient setting, it is likely that new strategies will be needed to get a truer picture of the scope of the problem. Increasing the use and scope of registries is likely to be a key advance. Other promising methods include looking for indirect evidence that an error has occurred, such as diagnostic codes that are indicative of an earlier error.
Such data became available in the U.S. in 2007, when the Centers for Medicare and Medicaid Services (CMS) published what they called “never events” in diabetes care (Figure 1). They were episodes of diabetic ketoacidosis (DKA), hyperosmolar hyperglycemic syndrome (HHS), or severe hypoglycemia, all occurring de novo in a hospital. They found 15,848 events in 2007, 72.4% of these events were due to DKA, 20.5% were due to HHS, and 7.1 were due to
Figure 1: “Never Events” reported by CMS in 2007 (click to enlarge)severe hypoglycemia. It is clearly unacceptable that an episode of DKA has its onset during a hospitalization, and the CMS data have shown clearly that serious errors in hospital settings that cause increased morbidity and mortality are much too frequent.
Unfortunately, much of the data on medical errors is difficult to obtain, in large part because self-reporting of errors has been well-documented to grossly underestimate the frequency of medical errors and adverse events in general. For example, although it is well-accepted that the rate of adverse events in hospitals world-wide is ~10% of hospitalizations (of which more than 50% of these adverse events are preventable), Bates and colleagues have shown that the self-reported adverse drug events were only 1% of the total adverse drug events found in a thorough review at a leading U.S. hospital.
Many hospitals are utilizing clinical pharmacists to review hospital records daily, identify adverse drug events and other undesirable outcomes, and to play a more active role in providing active feedback to the providers. Others have developed inpatient multidisciplinary diabetes care teams, to review EHR data daily, to identify potential clinical issues and work with other clinicians to reduce the frequency of injurious errors in diabetes care.
Origins of Medical Errors
We should not be surprised that errors in diabetes management are both common and injurious. Unfortunately, the amount of relevant clinical and scientific information that is relevant for each person with a specific medical condition is very large and rapidly increasing. In 50 years, medicine has gone from a profession in which we had too little information to satisfactorily treat many conditions, to one where the flow of information has increased exponentially. Our task now is often to separate the wheat from the chaff, the relevant from the irrelevant.
Yet, together with the explosion of medical and scientific information, we still have a problem accessing the key information, often buried in the “silo” of an unsearchable medical chart or journal, and making sure the right information is transmitted to the right person at the right time so the decision is the correct one. Unfortunately, as a species, humans are very error-prone. The slips and lapses that we experience every day are a reminder to all of us that the cognitive flexibility of humans is also accompanied by a quite variable, but relatively high incidence of errors in memory, judgment, calculation, and analysis. Such cognitive errors are often one of the causes of injurious errors in the management or self-management of diabetes.
Our patients with diabetes are particularly vulnerable, because not only is modern diabetes care exceptionally complex, but what is required of the patient is complex as well. The patient has to deal not only with the distractions of ordinary life, but also deal with diabetic dysglycemia severe enough to lead to either hypoglycemia or hyperglycemia of sufficient degree to cause cognitive dysfunction and errors in self-management. Catastrophic accidents may result, for example, if severe hypoglycemia occurs when the diabetic patient is driving a vehicle. In addition, the stresses of diabetes self-care and the fear of both complications and hypoglycemia often generate anxiety and/or depression, which in turn frequently increase the likelihood that the depressed or anxious person will make errors in their self-care.
Cognitive dysfunction occurs more frequently in patients with diabetes, particularly those with higher HgbA1c levels, and this too may be associated with more frequent errors, and ultimately may take away the patient’s ability to provide adequate self-care.
However, we should draw a distinction between medical errors that are injurious and the much more common errors which do not lead to injury. The majority of errors in care, either self-care or care given by a provider, are errors which are slips or lapses, and often are corrected immediately. These errors can be reduced in frequency, but are likely never to be completely eliminated. On the other hand, both errors that are immediately injurious and ones that will eventually lead to patient injury should be our focus. Most experts agree that our focus should be to redesign diabetes care so as to eliminate injurious medical errors during diabetes care.
But improving the safety of care of those with diabetes has proven to be difficult, and the barriers in achieving patient safety are often not clearly visible. Sometimes people confuse the term quality of care with patient safety. Both are very important, but not synonymous. Rather, they overlap. Also, not all injuries or adverse events that occur are due to medical errors, and not all are preventable. As an example, a patient may receive optimal diagnosis, management, and therapy for a diabetic foot infection and still have a poor outcome.
A useful strategy is to identify weaknesses in the system of care which increase the risk of errors. These are commonly termed “latent errors”. An example of a latent error is when one places bottles containing U-500 insulin next to bottles of U-100 insulin in a clinical area. This latent error will increase the probability that a busy nurse may pick the wrong bottle and give an overdose of insulin to the patient.
Figure 2: Relationship between quality, errors, and safety (Click to enlarge)The following diagram (Figure 2) shows the complex relationship between quality of care, medical errors, patient safety, and non-preventable injuries, which also occur. Please note that latent medical errors, often due to systems issues, play a most important role and must also be accounted for in order to increase the safety of care.
Kohn LT et al. To Err Is Human: Building a Safer Health System. Washington, DC: National Academies Press; 2000.
Bogner MS. Human Error in Medicine. Hillsdale, NJ: L. Erlbaum Associates;1994
Hellman R et al. Patient safety and inpatient glycemic control: Translating concepts into action. Endocr Pract. 2006;12(Suppl 3):49-55
Hellman R, Regan J, Rosen H. Effect of intensive treat¬ment of diabetes on the risk of death or renal failure in NIDDM and IDDM. Diabetes Care. 1997;20:258-264.
Jha AK et al. Identifying Adverse Drug Events: Development of a Computer-based Monitor and Comparison with Chart Review and Stimulated Voluntary Report. JAMIA. 1998;5:305-314
Bates D et al. American College of Endocrinology and American Association of Clinical Endocrinology and American Association of Clinical Endocrinologists position statement on patient safety and medical system errors in diabetes and endocrinology. Endocr Pract. 2005;11:197-202
Bates DW et al. Improving safety with information technology. N Engl J Med. 2003;348:2526-2534
Leape LL et al. What practices will most improve safety? Evidence-based medicine meets patient safety. JAMA. 2002;288:501-507
Aspden P et al. Patient Safety: Achieving a New Standard for Care. Washington DC: National Academies Press, 2004.
Bates DW et al. Effect of computerized physician order entry and a team intervention on prevention of serious medical errors. JAMA. 1998;280:1311-1316