Safer design
H. Kaplan
DOI: https://doi.org/10.1111/j.1537-2995.2007.01245.x
2007-05-01
Transfusion
Abstract:A hazard is “an inherent physical or chemical characteristic that has the potential for causing harm to people, the environment, or property.” Because a hazard is an intrinsic characteristic, its elimination requires altering the material itself or its conditions of use. Despite significant reduction of transfusion risk, that is, the probability and severity of harm, blood transfusion retains inherent hazards that are well known to transfusion specialists and feared by patients and their physicians. Strategies for safe design in medicine are intended to eliminate or diminish hazard risk. The strategies in order of diminishing probable effectiveness have been categorized as inherent, passive, active, and procedural. Inherent strategies eliminate a hazard or reduce its significance. Passive strategies, for example, a fence, reduce risk by their physical presence alone. Active strategies prevent or mitigate accidents by utilizing devices such as interlocks and alarms. Procedural strategies rely instead on people and their training, operating procedures, and checklists. For purposes of this discussion, only the two ends of the spectrum of probable effectiveness will be briefly considered—inherent and procedural strategies. Although a great deal of effort may be expended in attempting to optimize a process, it is preferable, when possible, to eliminate or trivialize its perceived problems. Inherent design strategies eliminate a hazard or reduce its significance. For example, in circumstances where the risk of error is perceived to be very high, such as transfusion in mass casualty situations, the exclusive use of group O blood eliminates concerns regarding possible ABO mismatch. Procedural strategies, probably the least reliable for managing hazards, are the most often employed in transfusion practice and in nursing practice. Redundancy is often utilized to increase procedural reliability. Transfusion safety at the bedside relies heavily on redundancy, particularly the second person check of the accuracy of another—the “two-person-three-way check.” Both high reliability theory (HRT—the underlying philosophy of high reliability organizations that hazardous circumstances can be effectively managed) and normal accident theory (NAT—the underlying belief that it is when, not whether, highly complex and closely coupled systems will fail), two fundamental and divergent safety research perspectives, have been used to evaluate the efficacy of redundant checking in medication administration. Both of them, while recognizing the potential utility of redundancy checking, underscore more strongly its inherent weaknesses, for example, a reduced mindfulness of its safety implications—when checking is perceived as a “routine task” and diffuse responsibility for safety checking with reliance solely on the checker. How a two-person check is carried out is critical to its effectiveness. The error rate of a passive check missing an error may be as high as 10-1.7 Recognizing the limitations of the usual procedural approaches to safe transfusion practice, two separate groups of investigators provide differing strategies for hazard management in this issue of TRANSFUSION. Murphy and coworkers describe a multicenter trial of a novel “low-tech” intervention to reduce errors in the pretransfusion bedside check. Their appealingly simple approach combined a sticker-tag serving as both a visual reminder to the transfusionist to stop and check the patient’s wristband and a barrier by requiring its removal to spike the unit. The elegantly designed study did not demonstrate efficacy of the intervention however, even though the sticker acted as a behavior shaping device and forced a stop. Perhaps the action ultimately forced was to stop to remove the sticker rather than to check the patient’s wristband. Ideally a forcing function would eliminate the possibility of proceeding if the required step was not first carried out. The second article, by Sandler and coworkers, demonstrates the feasibility of a “high-tech” radiofrequency tagging approach for computerized control of information and processes from donor collection to recipient transfusion. Here tight control of the steps of the process is apparently achievable. The current work, although focusing primarily on collecting and generating a radiofrequencytagged and identifiable blood component, also includes a demonstration of an essentially seamless transition to within-hospital control. Taken together with this group’s prior work on the in-hospital phase of transfusion, this admittedly not-here-yet strategy provides an interesting glimpse into a possible future. Despite the recognized advantages of such automation, framing views provided by HRT and particularly NAT provide another cautionary note, especially where the tight coupling of processes by automation might, for example, create the potential for widespread unintended consequences of an unlikely but possibly flawed decision rule. “Too often, engineers accept the hazards in a system and immediately look for systems and procedures to control and manage them. A better approach is to ask first TRANSFUSION 2007;47:758-759.