A 33-year-old female underwent hysterectomy for refractory endometriosis. For pain post-operatively, the patient was placed on a Patient-Controlled Analgesia (PCA) pump containing morphine sulfate. Three hours after her transfer to a gynecology floor, the patient began complaining of severe frontal headache, nausea, and vomiting. Nurses attributed these symptoms to post-operative pain and the effects of the morphine. The PCA pump was continued, despite her continued complaints of headache and relatively mild incisional discomfort. The patient became progressively drowsier and her respiratory status declined. Luckily, the pulse oximeter alarm was activated when her oxygen saturation fell, and the PCA pump was discontinued. At the time of the discontinuation, the patient's O2 saturation level was 76%. At that point, clinicians realized that she was actually suffering from an adverse reaction to morphine. Further investigation of the patient's prior medical history revealed similar complaints with prior administration of meperidine hydrochloride (Demerol), but these had not been noticed in her chart, elicited on pre-operative history, nor flagged as an adverse reaction or “allergy.”
What does this case tell us about creating a safe environment for the control of post-operative pain? The nurses initially attributed the patient’s problems to effects of the morphine and did nothing about it until the pulse oximeter signaled that the patient was in distress. Similarly, the person responsible for taking the history might be faulted for not probing sufficiently to uncover that the patient previously had trouble with Demerol. Those problems most likely would be considered errors of omission by the individual nurses, rather than allowing a case like this to challenge the underlying system for pain management.
If this incident were attributed simply to errors of omission, what would it tell us about medical error? What information does it provide about why the incident occurred that could be used to prevent similar incidents from occurring? The answer to each of those questions is, “Nothing.” What it does indicate is that the cause of an incident usually is attributed to the individual associated with it, generally the care provider.(1) In this case, the patient’s use of the PCA pump is another implied cause, which will be discussed later.
Although errors have been described by a variety of terms,(2,3) error needs to be understood as an act or behavior. As a behavior, error is subject to the empirical findings and theory of psychology. This understanding is critical to the study of error and patient safety because a basic tenet of psychology is that behavior reflects the interaction between the person and the environment.(4) Thus, identifying the individual care provider as the sole source of error is both incomplete and misleading. Trying to tackle the all-encompassing environment to determine factors that contribute to error, however, can present a daunting, if not impossible, task.
The systems approach (5-7) dissects the environment into manageable segments and serves as a tool for identifying factors that contribute to error. In this approach, factors (evidence-based when possible) are clustered into categories that meet the criteria for a system [a complex of interacting factors(8)] and affect the behavior of the task performer (here the care provider). The eight categories are: (i) ambient conditions, (ii) physical environment, (iii) social environment, (iv) organizational factors, (v) the overarching system of legal, regulatory, national culture, and reimbursement factors (these first five relate to the environment or context of care), as well as the basic care-providing system comprised of (vi) the care provider, (vii) the means of providing care, and (viii) the patient. The context of care categories are represented as concentric circles with the care providing systems in the center.(9-11) In an effort to promote understanding among providers for whom systems-based thinking represents a paradigm shift from their individual-oriented training and socialization, the systems approach is likened to an artichoke, thus adding another error-related food model to the smorgasbord of Swiss cheese (2) and onion.(12) The leaves of the systems-approach artichoke (Figure) are the context of care components, while the basic care-providing systems are the artichoke’s.(13) Because the application of this model forces us to consider each category for error-provoking factors, it is a powerful tool for incident analysis.
Consideration of the basic care-providing systems in this case provides useful insights. The nurse’s focus appears to be on the means of providing care, the PCA pump, and ensuring that the patient was using this piece of technology appropriately. If that focus were on the patient—with the PCA pump considered merely as a means to an end—she would have been more likely to heed what the patient was saying and be alert to the adverse drug reaction. The pervasive attitude that technology is the solution to all problems can compromise patient safety because it focuses attention and reliance on the means of providing care, rather than on the patient. That tends to minimize the vital focus on listening to the patient and following up on her complaints, and overemphasize the role of the machines. The availability of a PCA pump does not mean it should be used in all pain management situations; PCA pumps may not be appropriate for post-operative use, as in this case. Mindless use of PCA pumps in every case driven in part by the “technological imperative”(14) is dangerous, particularly for patients whose cognitive or physical abilities are impaired by illness, medication, or anxiety. However, PCA pumps can be highly beneficial, particularly in an alert, motivated patient.
In examining this case, we also need to consider why the nurse did not act promptly to address the patient’s problems. The nurse could indicate that fatigue was a factor and investigation might reveal that the workload was heavy or the nurse was assigned double shifts. Why did that occur? Most analyses of error determine that the hospital, by assigning the work schedule, was responsible and stop the analysis at this point (the organizational level). The systems approach drives us to address overarching societal level factors. Why did the hospital assign such a demanding work schedule? The answer identifies the actual perpetrator of the incident--the fiscal constraints created by reimbursement policies.
To be effective across conditions, efforts to reduce the likelihood of an incident such as this one should be directed to the source of the problem, reimbursement. Although reimbursement policies and other factors in the overarching system are the most difficult to change, instances like this should be documented and data assembled and provided to those that might facilitate change. In that way, the error-inducing impact of reimbursement policies can be made known to those affected by error, such as companies that insure hospitals and care providers, professional organizations that have means of influencing policy makers, and the general public who can influence policy makers by their votes. Recognizing that such changes are slow to occur and that care continues in the meantime, knowledge that the incident’s source may be fiscal constraints and not exclusively inherent to the care provider is incentive to devise ways to reduce other identified error-provoking conditions.
Efforts can be undertaken to reduce excessive reliance on technology (the PCA pump and the pulse oximeter in this case) and increase the exercise of clinical skills such as listening to the patient. The importance of responding to what the patient says could be emphasized by training. Such education might include role-playing, in which staff members “become” a patient whose complaints are ignored, and/or a program that identifies and publicly acknowledges staff members who demonstrate effective clinical skills by listening to patients and responding to their concerns.
Thus, the application of the systems approach enables us to understand that the sources of the error in this case are multidimensional: involving a prevailing reverence for technology over listening to the patient, and the impact of reimbursement policies. Because this analysis was limited to the material in the case report, it is less comprehensive than it would be in a real-life situation. Even so, we see that the tool yields significant insights by identifying error-provoking factors, which would be unknown in its absence. Such factors can promote the enhancement of patient safety. The orientation toward error needs to shift from reacting to the clinical aspects of a case--those that are the most immediately compelling, but tend to distract from the real issues--to preventing similar situations in the future.
The message of the artichoke–systems approach is that an act by an individual care provider, which is judged to be an error, reflects factors in the context of care that contribute to, if not provoke, the act. The context of care is analogous to the script in a play in which the care provider is an actor. Actors who fit the role can be changed and yet the outcome of their performance will be the same, because actors respond to the script. To change the outcome of their performance, the script must be changed. Thus, when analyzing an incident, rather than focusing solely on the individual care provider-actor, the systems-approach artichoke should be peeled to identify error-inducing factors in the context of care script. Enhancing patient safety by effectively reducing the likelihood of error necessitates that actions be taken. Factors identified as contributing to error are the focus for such actions.
Marilyn Sue Bogner, PhD President and Chief Scientist Institute for the Study of Human Error, LLC Bethesda, Maryland
1. Bogner MS. Understanding Human Error. In: Bogner, MS, ed. Misadventure in health care: inside stories. Mahwah, NJ: Lawrence Erlbaum Associates; 2003.
2. Reason JT. Human error. New York: Cambridge University Press; 1990.
3. Rasmussen J. Human errors: A taxonomy for describing human malfunction in industrial installations. Journal of Occupational Accidents. 1982;4:311-333.
4. Lewin K. Principles of topological psychology. New York: McGraw-Hill; 1936/1966.
5. Bogner MS. Human error in medicine: a frontier for change. In: Bogner, MS, ed. Human error in medicine. Mahwah, NJ: Lawrence Erlbaum Associates; 1994:373-383.
6. Bogner MS. Error: it’s what, not who. Trauma Care. 1998;8:82-84.
7. Bogner MS. A systems approach to medical error. In: Vincent C and De Mol B, eds. Safety in Medicine. Amsterdam: Pergamon; 2000:83-101
8. von Bertalanffy L. General systems theory. New York: George Braziller; 1968.
9. Bogner MS. Stretching the search for the “why” of error: the systems approach. Journal of Clinical Engineering. 2002;27:110-115.
10. Bogner MS. Identifying error provoking factors: the systems approach. Keynote address to: The conference on Minimizing the risk of medical errors: focus on the system; November 2002; Toronto, Canada.
11. Bogner MS. The systems approach analysis of error. Plenary address to: Workshop on the Investigation and reporting of incidents and accidents; July 2002; Glasgow, Scotland.
12. Moray N. Error reduction as a systems problem. In: Bogner MS, ed. Human error in medicine. Mahwah NJ: Lawrence Erlbaum Associates; 1994:67-91.
13. Bogner MS. Understanding human error. In: Bogner MS, ed. Misadventures in health care: inside stories. Mahwah, NJ: Lawrence Erlbaum Associates; 2003.
14. Fuchs VR. Who shall live? Health, Economics, and Social Choice. Basic Books; 1983.
Figure. The Systems Approach: Artichoke Model
The leaves of the systems approach artichoke are the constituent systems of the context of care, and the basic care-providing systems are the heart of the artichoke.