Cristancho, Sayra M. PhD; Apramian, Tavis MA, MSc; Vanstone, Meredith PhD; Lingard, Lorelei PhD; Ott, Michael MD, MSc; Novick, Richard J. MD, MSc
Increasingly, medical education is recognizing that training novices for competence in routine situations is insufficient; we also need to know they are equipped to contend with the myriad, unpredictable, nonroutine situations they will confront in clinical practice.1 Intensifying this recognition is the growing evidence that safety may be at risk when experts venture beyond routine acts of practice2 into domains of uncertainty.3,4 But how do we train for dealing with uncertainty?
Surgery is a productive site to study uncertainty because surgeons encounter uncertainty on a daily basis, facing unforeseen complications and peculiar anatomies, which necessitate changing from the planned course during procedures.5 In a previous study, we proposed a model that describes how experienced surgeons conceptualize and approach intraoperative challenging moments.6 The model is framed as a three-stage cycle through which surgeons may travel many times over the course of a complex case. While traveling the cycle, surgeons usually described the imminence of a challenging moment in terms of their tolerance for uncertainty. Although uncertainty was part of the picture in our previous study, it was not the focus, nor did it receive detailed analytical attention. Similarly, in the literature there has been little attention focused on uncertainty itself; rather, it is treated as a tacit element of a decision-making event.7–10 Redressing this gap, our study sought to gather empirical data to support the development of a conceptual framework and a descriptive language of uncertainty and its impact on decision making in surgical practice. We anticipate that this language is the first step in providing a base of evidence for improving training and assessment of adaptive thinking in the face of uncertainty.
Decision making and uncertainty
In the early years of research on decision-making processes, uncertainty existed as a tacit precursor to the study of decision making and was generally left unexplored. Classical decision-making research began evolving in the 20th century via new theories derived from in situ observational research of proficient decision makers.7–9 The naturalistic decision making (NDM) framework10 was one theory that began to inform research into the role of uncertainty in decision making, particularly in Lipshitz and Strauss’s11 work in the late 1990s, which conceptualized uncertainty as characterized by inadequate understanding (i.e., a sense of not having sufficiently coherent situation awareness), incomplete information (i.e., a sense of having incomplete, ambiguous, or unreliable information), and conflicting alternatives (i.e., a sense that available alternatives are insufficiently differentiated). This NDM conceptualization of uncertainty now gives researchers an opportunity to build on a tradition of interdisciplinary research in medical uncertainty. Interdisciplinary researchers laid the foundations of today’s understanding of the ubiquity of uncertainty to medical practice12,13 by building on seminal research into medical uncertainty by sociologist Renee Fox14 in the 1950s. However, although their work provides insight into the nature of uncertainty in medicine, they13,15,16 do not directly address the behavioral or systems mechanisms17 by which uncertainty affects decision making.
Uncertainty and safety
The evolution of decision-making research has allowed space for a highly productive conceptual link between decision making and patient safety. The trend in safety research has been to approach uncertainty as a likely initiator of decision-making behaviors that deviate from standard practices. Safety researchers have focused on the negative outcomes of uncertainty by theorizing the mechanisms that lead to institutional drift outside the space of safe practice.18,19 Amalberti20(p66) calls behaviors that lead to such drift “violations,” or “deliberate deviations from standard instructions,” which may on occasion result in adverse events. Violations may be conceptualized as necessary deviations employed as individuals attempt to cope with conflicting demands in complex and uncertain situations.18 Unfortunately, frequent violations can accrue and shift the accepted zone of safe practice in an institutional culture. Amalberti defines this shift as “migration,” where the boundaries of safe practice are distorted and team members begin to routinely accept violations. This model reflects the dual effects of violations: (1) saving time and benefiting individuals and systems, as well as (2) exposing patients and staff to risk. Safety researchers have also begun to recognize that, in professions where uncertainty is the norm of everyday practice, short-term violations may be mechanisms for long-term innovation21–24; however, this theoretical concept has not been sufficiently explored.
Faculty surgeons were sampled purposively25 to include rich informants from a variety of surgical specialties (general, orthopedic, cardiac, urological, vascular, and neurosurgery) at a medical school (unnamed to protect the anonymity of the participants) who were willing to be observed and to reflect on their intraoperative experiences. Seven surgeons (six male and one female) with 5 to 20 years of faculty experience participated in the study. Observations and interviews were conducted by one of us who is not a surgeon (S.C.) with experience observing surgical cases for research purposes. Twenty-six cases were observed in total. Cases in which participants anticipated challenges were selected in order to maximize the potential for surgeons to encounter multiple instances of uncertainty.
Interviews took place immediately after the observed procedures. Field notes captured specific utterances by surgeons and various discussions between surgeons and the residents or operating room staff. These field notes were used to enable interview probing of those moments that surgeons did not explicitly reflect on during their accounts of the operation but that were noted by the nonparticipant researcher as potentially uncertain instances. The interview technique followed the critical decision method (CDM).26 The CDM was particularly effective in response to the high-risk, high-complexity nature of medium to long surgical procedures (4–14 hours) where many challenging and uncertain moments may occur. Observation-informed probes allow the surgeon to expand and further reflect on the details of the situation of interest.
Given our desire to contextualize these results within a rich but mostly nonsurgical literature on uncertainty, we combined template27 and inductive25 analysis. The theoretical construct of “uncertainty,” as defined by Lipshitz and Strauss,11 was used as the template to identify instances of uncertainty in the dataset. The NDM framework was selected because of its emphasis on studying how expert practitioners make decisions in real-world settings characterized by time pressure, uncertainty, ill-defined goals, and high stakes.28
Two of us (S.C. and M.V.) performed an initial “open coding”29 for both template codes and emergent themes through iterative stages of coding, and followed a constant comparative technique to elaborate and ensure intercoder agreement of the categorical constructs and content. The final thematic coding structure and a conceptual model of the relationship between coding themes were developed after several rounds of coding and feedback from the entire research team. Throughout the analysis, qualitative rigor criteria were used, including member-checking through respondent feedback, investigator and participant triangulation, and the formation of an audit trail of the analytical process.30–34 NVivo 9.2 software was used to manage the data.
The study was approved by the institutional research ethics board at the study institution; all participants provided informed consent.
Following NDM theory, an “instance of uncertainty” was characterized as an incident during the surgery in which the surgeon experienced a sense of doubt while trying to make a decision for which there was no clear “best” answer. Across our dataset, 241 instances of uncertainty were identified through template analysis.
Multiple rounds of inductive analysis did not reveal new uncertainty instances over the original set; however, they did reveal new subcategories and better reflected the relationships among these categories than in the template analysis. The template analysis confirmed that existing concepts of uncertainty from the NDM literature, including “inadequate understanding,” “incomplete information,” and “conflicting alternatives,”11 are relevant to the surgical domain. We present together the final results, combined from both analytical phases, in which existing NDM concepts are integrated into a more elaborate model of uncertainty specific to our surgical data. This model consists of two key themes from the inductive analysis—recognizing uncertainty and responding to uncertainty—which contain a number of subthemes, including refined versions of the template concepts. These themes and subthemes are described below, with illustrative excerpts from the dataset.
The process of “recognizing uncertainty” involved surgeons appreciating cues related to patient characteristics, incomplete information, or external factors. “Incomplete information” was kept from the template concepts.
Patient-related characteristics were mostly related to problematic anatomy, comorbidities, medications, prognosis, medical history, social situation, or the reactions or reports of the patient. As a neurosurgeon described:
His exposure was difficult because he has a thick neck and he has a lot of adipose tissue there. We were having to identify the artery amidst all this fat tissue and there were a lot of veins, a lot of branches of veins that normally aren’t there. (S5-I14)
This instance is a representative example of the uncertainty created by the coalescence of patient characteristics and the issue of working with difficult anatomy.
The theme of incomplete information took many forms, including ambiguous or difficult to interpret information, information which was inadequate for decision making, unanticipated information, or conflicting sources of information. For instance, the preop diagnosis was not always a reliable source of information, as illustrated in this statement:
You’ve got this weird mass in the ventricular septum. And the preop diagnosis is wrong. (S3-I12)
The theme of external factors represents those issues that were outside both the surgeon’s control and the patient’s condition, such as scheduling-related issues of being on call, staffing, time pressures, and equipment-related issues such as the availability and function of tools. For example, one surgeon explained that having to remain attentive to multiple patients while operating because of the on-call schedule may increase the perceived level of uncertainty:
If you have to break that flow and go do something else and you come back and you don’t know what somebody’s done and you’re looking in and trying to figure out what it is. It is a bit unnerving. (S5-I21)
In summary, the theme of “recognizing uncertainty” described how, when, and/or why a surgeon realized that he or she was dealing with a potentially challenging situation due to patient characteristics, incomplete information, or external factors.
Responding to uncertainty
The theme of “responding to uncertainty” captures strategies that surgeons used in their attempt to manage an uncertain situation. This category included the following subthemes: prioritizing alternatives, reevaluating and adapting the plan, creating innovative solutions, and seeking advice. Two of these were refined template concepts of “conflicting alternatives” and “inadequate understanding.”
The theme of prioritizing alternatives described the issues at play when the surgeon was uncertain about distinguishing or redistinguishing the most appropriate solution after the emergence of new information. During this process, the surgeon considered risk management, potential outcomes, and technical issues. For instance, the following surgeon discusses managing risk and optimizing outcomes when weighing alternatives:
So [the decision] was are we going to give him an ostomy which is the safe thing to do or are we going to remove the rest of the colon and do an anastomosis which is maybe slightly riskier but would give him a better quality of life? (S1-I16)
Technical issues often complicated these decisions and contributed to additional uncertainty.
When surgeons referred to reevaluating and adapting the plan, they described uncertainty mainly in terms of interpreting the contextual cues the case was presenting:
When I poked into that pocket, I put my finger in and it was all slick surface and there was no visible plane and I tried to make a plane but I couldn’t and I wasn’t feeling safe enough playing with the nerves and vessels around. (S2-I4)
This interpretation also included understanding how to work with the anatomy, or using cues from the anatomical environment to anticipate contextual information that was not available:
I was trying to find a tendon that is still attached and still intact that I could use as a landmark. (S2-I4)
When surgeons talked about creating innovative solutions, they emphasized the need for improvisation in the face of uncertainty, the pressure to think creatively, and the need to come up with innovative solutions or ways to see the problem differently. Reflecting on his own improvisation during the observed case, one surgeon remarked:
I don’t know where I got the idea from but I decided that it would be useful to use a laparoscopic stapling device that we would normally use to divide blood vessels in a laparoscopic surgery and that we would use it here. (S1-I16)
The final subtheme of “responding to uncertainty” was seeking advice, which was described in many instances as the last resource when the other strategies didn’t seem to improve the situation or to provide a sufficient sense of control for the surgeon. Advice was sought by either working collaboratively with the resident to understand the situation or by consulting a more senior colleague. For instance, one surgeon reported:
I was well aware of the other approach I just had never technically done it since I’ve been in practice. So I had actually requested that the other surgeon come just in case I felt that my judgment or my technique was not proper. (S7-I26)
Such instances illustrate that “seeking advice” was not exclusively about getting more information; sometimes in the case of experienced surgeons, it was also a reassurance mechanism, particularly in rarely occurring situations.
In the category of “responding to uncertainty,” although “prioritizing alternatives” and “reevaluating and adapting the plan” were more often identified, surgeons acknowledged that depending on the situation, they may be prompted to “create innovative solutions” or to “seek advice” from their colleagues.
Our results confirm that uncertainty is common in the operating room, even for experienced surgeons who, in our study, accepted it as an unavoidable part of practice. Faced with its ubiquity, experienced surgeons come to recognize the cues that signal uncertainty and develop recurrent responses to uncertain situations. In this section, we consider the implications of these results for both an improved understanding of the relationship among uncertainty, safety, and innovation, and an explicit language for teaching clinical novices about approaching uncertainty.
Faced with uncertainty, surgeons’ decision making commonly involved an adaptation of their surgical plan, sometimes to the point of what they considered “innovation.” Viewed through the lens of current safety theory, particularly Amalberti’s migration model, such adaptations might be considered violations given that they constitute “deviations from standard practice that individuals employ to cope with conflicting demands in complex situations.”18–22 However, these adaptations often allowed the surgical procedure to continue when “standard practice” could no longer offer a way forward, suggesting also a positive impact.
We do not uncritically accept our participants’ “positive spin” on their nonstandard practices. However, the many examples of positive outcome from nonstandard practices in our dataset suggest that we need to grapple with ways to characterize these events in the light of current safety theory. If uncertainty functions as a mechanism for positive innovation as well as negative “drift,” we require a theoretical model for thinking about and teaching around this complex issue. Towards this end, we propose extending Amalberti’s model into a spectrum in which “innovations” and “migrations” anchor opposite ends and the middle point corresponds to the “standard approach” to practice. Using this new representation, situations that deviate from the “standard approach” towards the “innovation” end point may be referred as “adaptation,” whereas situations that deviate from the “standard approach” to the “migration” end point may be referred as “violations.” Introducing a debate about whether uncertainty may have both negative and positive long-term implications has both merit and precedent. More recently, others have argued the need for investigating the implications of performing under uncertainty as a mechanism for promoting adaptive thinking.21,23 For example, McKenna et al22 have described ways by which improvising when dealing with uncertain situations may be an effective mechanism to promote physicians’ creativity. The findings of McKenna and colleagues’ study also suggests that the nature of improvising goes beyond the cognitive to also include the social interdependencies and that there is a pressing need for further exploration of these issues.
Our proposed notion of an innovation/migration spectrum offers a useful heuristic and vocabulary, which presents three advantages for medical education. First, it could assist with current educational initiatives to cultivate innovation as a learnable skill among experts. As expertise researchers have described, what separates “routine” from “adaptive” experts is their ability to display innovative, flexible, and creative thinking when working at the “edge of their competence”35–37; our findings nicely operationalize some of the features of adaptive expertise in the surgical setting. Second, the spectrum could be a valuable teaching tool to help surgeons and trainees learn to explicitly and critically assess any nonstandard practice for its migration dangers as well as its innovation potential. Finally, the spectrum allows teachers and learners to discuss the critical issue of context specificity: A specific response to uncertainty may fall on either side of the violation–adaptation spectrum. Depending on the situation, innovating to improve patient outcomes in one situation may constitute a dangerous migration that threatens patient safety in another. Seeing past the outcome of today’s action was not a pattern in our data; in fact, reflections from some of our participants suggested that the immediate outcome of their actions usually dictated whether they would regard them as adaptations or violations. However, as Amalberti et al18 have persuasively argued, a good outcome can follow from a violation—until the day when it doesn’t anymore. We propose that this outcome orientation requires critical attention in teaching discussions, and that the innovation/migration spectrum could be used as a heuristic to teach residents to assess their actions independent of single-instance outcomes. Through such conversations, experts and novices can acknowledge and discuss the fine line distinguishing “being innovative” from “being lucky,” discussions which were not in evidence in our study data.
Beyond the issue of the “spectrum,” our thematic concepts for describing the recognition of and response to uncertainty offers a language for surgeons to have explicit conversations with their residents about how they function outside the comfortable space of routine practice. This has become increasingly important, with the recognition that we train novices mostly for the routine, and that duty hours reforms may be limiting their opportunity to experience a range of uncommon, nonroutine clinical situations.38–40 Our description of the cues expert surgeons used to recognize uncertainty constitutes a vocabulary that can be used by teachers to help novices “see” what the expert is seeing in the situation. As Moulton et al41,42 have described, expert surgeons withdraw from conversation with team members and trainees when they encounter uncertain situations; this may be due to the cognitive load involved in effortful practice, or it may also be due to a lack of language to describe what they are experiencing. In the absence of such a language, novices may miss out on critical opportunities to acquire legitimate peripheral experience recognizing and responding to those unusual situations, which don’t arise in every case.
Our work responds to recent claims in the literature for studies of nontechnical skills in surgery that rely on empirical evidence.1 However, our research design decisions necessarily limit the insights arising from this work. Limitations include the recruitment of surgeons from a single educational institution, the reliance on participants’ assessment of anticipated challenges for case selection, and the use of a single observer and the possibility of hindsight bias in participant interviews in the CDM. The use of a research team, including three surgeons, during analysis helps to broaden the analytical results beyond the single observer’s point of view. We aimed to guard against hindsight bias in interviews by using the observation field notes to inform the probing questions.
Our results confirm the relevance to surgery of the broad concepts used by the NDM tradition to conceptualize uncertainty. Refining these concepts and integrating them into an expanded catalog of features involved in uncertainty has produced a context-specific theoretical language for describing how surgeons perceive and respond to uncertainty. Future research will explore the notion of an innovation/migration spectrum, particularly in relation to unique features of surgery such as maintaining the “principle of progress” and the difficulty of predicting outcome. This spectrum-based concept of uncertainty confirms the utility of current research into adaptive expertise35,43 and teaching strategies that emphasize creative and critical thinking. Moreover, it will help in understanding how sociopolitical and cognitive constructions of uncertainty influence the full spectrum of decision-making practices. Toward this end, the language provided in this study may be used in faculty development research, to explore its ability to structure effective discussions among experts and novices about the positive and negative implications of intraoperative decision making.
1. Hull L, Arora S, Aggarwal R, Darzi A, Vincent C, Sevdalis N. The impact of nontechnical skills on technical performance in surgery: A systematic review. J Am Coll Surg. 2012;214:214–230
2. Dismukes K, Berman BA, Loukopoulos LD The Limits of Expertise: Rethinking Pilot Error and the Causes of Airline Accidents. 2007 Aldershot, UK Ashgate
3. Tubbs EP, Elrod JA, Flum DR. Risk taking and tolerance of uncertainty: Implications for surgeons. J Surg Res. 2006;131:1–6
4. Farnan JM, Johnson JK, Meltzer DO, Humphrey HJ, Arora VM. Resident uncertainty in clinical decision making and impact on patient care: A qualitative study. Qual Saf Health Care. 2008;17:122–126
5. Fioratou E, Pauley K, Flin R. Critical thinking in the operating theatre. Theor Issues Ergon Sci. 2011;12:241–255
6. Cristancho SM, Vanstone M, Lingard L, LeBel ME, Ott M. When surgeons face intraoperative challenges: A naturalistic model of surgical decision making. Am J Surg. 2013;205:156–162
7. Lipshitz R, Klein G, Carroll J. Introduction to the special issue. Naturalistic decision making and organizational decision making: Exploring the intersections. Organ Stud. 2006;27:917–923
8. Lipshitz R Taking Stock of Naturalistic Decision Making. 2001 Chichester, UK Wiley
9. March JG, Simon HA Organizations. 1958 New York, NY Wiley
10. Klein GA Decision Making in Action: Models and Methods. 1993 Norwood, NJ Ablex Pub
11. Lipshitz R, Strauss O. Coping with uncertainty: A naturalistic decision-making analysis. Organ Behav Hum Decis Process. 1997;69:149–163
12. Beresford EB. Uncertainty and the shaping of medical decisions. Hastings Cent Rep. 1991;21:6–11
13. Hall KH. Reviewing intuitive decision-making and uncertainty: The implications for medical education. Med Educ. 2002;36:216–224
14. Fox RC Vision and Reality: The Evolution of Medical Uncertainty. 1979 Chapel Hill, NC School of Medicine, University of North Carolina at Chapel Hill
15. Kasper J, Geiger F, Freiberger S, Schmidt A. Decision-related uncertainties perceived by people with cancer—modelling the subject of shared decision making. Psychooncology. 2008;17:42–48
16. Babrow AS, Kasch CR, Ford LA. The many meanings of uncertainty in illness: Toward a systematic accounting. Health Commun. 1998;10:1–23
17. Han PKJ, Klein WMP, Arora NK. Varieties of uncertainty in health care: A conceptual taxonomy. Med Decis Mak. 2011;31:828–838
18. Amalberti R, Vincent C, Auroy Y, de Saint Maurice G. Violations and migrations in health care: A framework for understanding and management. Qual Saf Health Care. 2006;15:66–71
20. Amalberti R. The paradoxes of almost totally safe transportation systems. Safety Sci. 2001;37:109–126
21. Dekker S. Failure to adapt or adaptations that fail: Contrasting models on procedures and safety. Appl Ergon. 2003;34:233–238
22. McKenna K, Leykum LK, McDaniel RR Jr. The role of improvising in patient care. Health Care Manage Rev. 2013;38:1–8
23. Mestre JP Transfer of Learning From a Modern Multidisciplinary Perspective. 2005 Greenwich, Conn IAP
24. Hollnagel E, Woods DD, Leveson N Resilience Engineering: Concepts and Precepts. 2006 Aldershot, UK Ashgate
25. Lingard L, Kennedy T Qualitative Research Methods in Medical Education. 2007 Edinburgh, UK ASME
26. Crandall B, Hoffman RR, Klein GA Working Minds: a Practitioner’s Guide to Cognitive Task Analysis. 2006 Cambridge, Mass MIT Press
27. Cassell C, Symon G Essential Guide to Qualitative Methods in Organizational Research. 2004 London, UK SAGE
28. Schraagen JM Naturalistic Decision Making and Macrocognition. 2008 Aldershot, UK Ashgate
29. Charmaz K Constructing Grounded Theory: A Practical Guide Through Qualitative Analysis. 2006 London, UK SAGE
30. Pope C, Mays N Qualitative Research in Health Care. 20083rd ed Malden, Massachusetts: Blackwell Publishing
31. Flick U Managing Quality in Qualitative Research. 2007 Los Angeles, Calif SAGE
32. Flick UWE. Triangulation revisited: Strategy of validation or alternative? J Theory Soc Behav. 1992;22:175–197
33. Flick U. Qualitative research—state of the art. Soc Sci Inf. 2002;41:5–24
34. Sandelowski M. Rigor or rigor mortis: The problem of rigor in qualitative research revisited. Adv Nurs Sci. 1993;16:18–8
35. Mylopoulos M, Regehr G. Cognitive metaphors of expertise and knowledge: Prospects and limitations for medical education. Med Educ. 2007;41:1159–1165
36. Bransford J, Brown AL, Cocking RR, et al. How People Learn: Brain, Mind, Experience, and School. 2000 Washington, DC National Academies Press
37. Hatano G, Inagaki KStevenson HW, Azuma H, Hakuta K. Two courses of expertise. Child Development and Education in Japan. 1986 New York, NY W.H. Freeman/Times Books/Henry Holt & Co:262–272
38. Bell RH Jr, Biester TW, Tabuenca A, et al. Operative experience of residents in US general surgery programs: A gap between expectation and experience. Ann Surg. 2009;249:719–724
39. Chikwe J, de Souza AC, Pepper JR. No time to train the surgeons. BMJ. 2004;328:418–419
40. Schwartz A, Pappas C, Bashook PG, et al. Conceptual frameworks in the study of duty hours changes in graduate medical education: A review. Acad Med. 2011;86:18–29
41. Moulton CA, Regehr G, Lingard L, Merritt C, Macrae H. ‘Slowing down when you should’: Initiators and influences of the transition from the routine to the effortful. J Gastrointest Surg. 2010;14:1019–1026
42. Moulton CA, Regehr G, Lingard L, Merritt C, MacRae H. Slowing down to stay out of trouble in the operating room: Remaining attentive in automaticity. Acad Med. 2010;85:1571–1577
43. Mylopoulos M, Regehr G. How student models of expertise and innovation impact the development of adaptive expertise in medicine. Med Educ. 2009;43:127–132