three such events: hospital-acquired diabetic ketoacidosis (DKA), hyperosmolar hyperglycemic syndrome (HHS), and severe hypoglycemia. They reported that during a one-year period there were 15,848 documented such events. Of these, 72.4% were episodes of DKA that began during an inpatient stay, 20.5% were significant cases of HHS, and 7.1% were cases of severe hypoglycemia resulting in coma. In 2008, CMS announced it would not pay for hospital stays in which those never events occurred. Yet, some data from several states show that these never events in glycemic management are still occurring in U.S. hospitals at a rate of close to half of the 2007 rates.7
Few experts in hospital medicine were surprised that the threat of nonpayment by CMS did not have a greater effect on reducing the frequency of these so-called never events. Hospitals are extraordinarily complex structures, and the complexity of care needed for patients who need improvement of their glycemic control often stresses the systems of care present in the hospitals and reveals their shortcomings.8 Changes in the present hospital systems of care will be needed if we are to make in-hospital care safe for the patient with diabetes.
Systemic Issues in the Development of Errors in Insulin Therapy
To understand why some of the errors occur so often and why it is so hard to prevent them, it is important to look at systemic issues that play an important role in the development and persistence of errors in insulin therapy over time. It may seem counterintuitive, but some decisions made far from the bedside, often termed the “blunt end of care,” have a profound effect on the chance that errors will occur. The shortage of nursing personnel is one such example. Errors involving nurses at the point of care, also termed the “sharp end of care,”9 are more likely to occur when shortages in nurses result in the individual nurse being overworked,10 and spending too little time focusing on the many crucial tasks involved in accurate use of insulin therapy. These include (1) assessing whether the order for insulin is reasonable and appropriate for the patient at that point in time; (2) checking to ensure that the appropriate insulin dose of the correct type of insulin is administered to the right patient by the right route at the right time; and (3) verifying that the patient’s blood glucose and clinical status are being monitored sufficiently so the effect of the insulin can be measured and, if needed, the therapy can be altered. The support staff responsible for glucose monitoring also need to have an appropriate workload and training to provide reliable and timely glucose testing. In addition, it is important for all care providers to listen to the patient’s perspective and to use that information to ensure that the care being given is patient centered, effective, efficient, equitable, timely, and safe.
Electronic Health Records
Many hospitals are moving from the traditional paper-based hospital charts to fully operative electronic health records (EHR). This key system change has many benefits, including the ability to share information more rapidly and widely among members of the health-care teams.11 As a systems tool, EHRs have great potential. The clarity and lack of ambiguity of data sources, particularly physician and nursing notes and orders, reduces the chance for medical errors.12 If properly designed, the EHR can provide decision support and forcing functions that may reduce certain types of errors of insulin therapy. But flaws in any electronic record can introduce other errors—for example, if the screen routinely accessed by the providers does not contain crucial information needed at the time they are ordering insulin therapy, or when the decision support tools are judged to be unhelpful or burdensome and, as a result, are routinely bypassed by the user of the EHR.
Computerized Physician Order Entry (CPOE) systems are extremely important12 and have great value in the care of the patient with diabetes, particularly when the algorithms used for specific insulin orders can be reduced to a validated, evidence-based order set and provided to the user of the CPOE system. Examples of these widely used types of order sets are shown in Table 4.1.
Table 4.1—Examples of Order Sets Widely Used in CPOE Systems
• Intravenous (IV) insulin infusion
• Subcutaneous (SQ) basal-bolus insulin
• Hypoglycemia recognition and treatment
• Treatment of diabetic ketoacidosis
• Treatment of hyperosmolar hyperglycemic syndromes
• Discharge insulin therapy
• Transition from IV to SQ insulin therapy
• Insulin therapy for patients on IV hyperalimentation
• Insulin therapy for patients on high-dose corticosteroids
• Insulin when there is an interruption of nutritional therapy (IV or oral)
• Consideration of patient self-use of SQ insulin pumps in the hospital setting
• Criteria for continuing or discontinuing patient self-use of their insulin pumps.
The more often a validated and complete set of tasks can be put into an order set that is available to the physician and hospital staff, preferably in a CPOE set but alternatively in a paper-based order set, the easier it is to administer therapy with fewer errors, to monitor the quality of the work that has been done, and to identify opportunities for improvement. It would be a mistake to believe that a set of often-complex orders for insulin therapy will improve care by itself. The process of introducing it and teaching health-care providers and staff how and when it should be used is key. Subsequent proficiency testing is central to the success of any new CPOE, particularly in the case of insulin therapy, in which the effects of a momentary slip or omission in an order sequence can result in harm to the patient.
Culture of Safety
Many experts in the area of patient safety believe that sponsoring a culture of safety is important for nearly any institution involved in patient care, and mandatory for a high-alert therapy, such as insulin therapy. A culture of safety is created when a collection of individuals decide that they will work together in a collaborative way to eliminate injurious errors and to promote patient safety.8 A culture of safety in hospitals is much more effective when the hospital leadership is strongly supportive. The group engaged in the culture of safety should develop a nonhierarchical approach, which includes checking not only their own work, but also that of others. Back-up checks should be routine. The entire group must focus on and measure their performance and use nonpunitive methods to improve performance. A successful team strives to work in collaboration and to develop a sense of both collective and individual mindfulness.13 Organizations that adopt such a philosophy are typically ones that value education and quality improvement, including constant measurement of relevant metrics, examination of their own performance, and continued collaboration to reduce the errors in care.4,14–16 Human error is part of the human condition and is to be expected, but injury to patients can and should be prevented.
Both the patient and their family as well as other key supportive figures should be included in the culture of safety. In the inpatient setting, an engaged patient and family can be an important set of observers of the patient’s clinical status, and a highly motivated group of helpers to prevent errors in insulin therapy.
Education
Education is often overshadowed by the acute medical need and not a priority in the inpatient setting. From a systems perspective, however, education is an important tool in the prevention of insulin errors.17,18 For example, recent data show that the faculty caring for patients in academic centers may not be up to speed when queried about their knowledge of basic principles of modern management of insulin in the inpatient setting.19,20 Several organizations have developed excellent tools to educate health-care providers and patient
