Abstract
Drug-distribution errors at a Calgary hospital pharmacy that killed two patients hit the headlines in mid-March. The news reports stated that potassium chloride was used instead of sodium chloride in a dialysis solution.
Pharmacists and other health professionals will recognize “look-alike packaging” as a well-known risk factor — a frequently identified potential source of “systems error” in health care settings. When the Canadian Adverse Event Study is published as anticipated this spring, it will no doubt spark greater public scrutiny of the hidden epidemic of patient harm in Canada — including a significant frequency of drug error.
If results from comparable studies in the United States are any indication, this Canadian study will show that preventable medical error is responsible for 5000 to 11,000 deaths annually in Canadian hospitals and also results in patient injuries that add billions of dollars every year to our health care costs. 1 Most Canadian pharmacists will know that American studies show that adverse drug events are the single leading cause of medical injuries, accounting for 19.4% of all adverse events. 2 Furthermore, preventable adverse drug events lead to extended hospital stays that have been estimated to cost $2 billion annually in the US, not including costs of injuries to patients or malpractice costs. 3
Order-entry systems
Many people believe that information technology (IT) is a key to reducing medication errors, and indeed, some computer-based physician-order-entry (CPOE) systems have reduced the overall number of non-intercepted serious medication errors compared to no such computer support. 4 I agree that IT can play a crucial role in improving patient safety, but the success of such systems is
The benefit of computerization is a function of the degree of fit between the technology and the users
contingent on a point that is still not fully or widely appreciated: the benefit of computerization is a function of the degree of fit between the technology and the users, their work practices, and their organizational structure. 4 Therefore, an explicit user needs assessment should be conducted at an early stage to inform the design of IT — a step that is sometimes ignored or conducted late and hastily in the budget-constrained rush to get a system up and running.
Checkered history
The importance of identifying user needs is illustrated by the checkered history of CPOE systems. For example, while Bates and colleagues observed an overall reduction in medication errors, they also found that the number of such errors for drug problems that were not addressed by the computer database doubled. 5 Computer technology is only as good as the requirements embedded in it, and inadequate requirements can lead to a performance decrement.
A more distressing example is the attempted implementation of a multi-million-dollar CPOE system at Cedars-Sinai Medical Center in California. Shortly after it was introduced, usage of the system was suspended in 2003 because hundreds of doctors complained that it was difficult to use and was threatening patient safety. This example shows that IT is not a silver bullet. Any one CPOE can lead to success in one setting and failure in another.
The reasons for these types of problems have been known for some time: “Past experience suggests that efforts to introduce clinical information systems into practice settings will result in failures and unanticipated consequences if their technical aspects are emphasized and their social and organizational factors are overlooked…. The critical issues in the implementation of these systems are social and organizational, not solely technical.” 6 Thus, the technology should be tailored to the human and organizational setting in which it will be used, which requires a deep understanding of context that is made possible by an analysis of user needs. This lesson may seem obvious, but it is not always heeded, as the relatively recent (and costly) example of Cedars-Sinai indicates.
If IT is to fulfill its significant potential to reduce medication errors, then we need to change the process by which such systems are developed. People and their needs must take precedence over computers and their bits, otherwise we will not reduce the number of Canadians who die annually from preventable medical error.