Article Text

This article has a correction. Please see:

Download PDFPDF

Experiential and rational decision making: a survey to determine how emergency physicians make clinical decisions
  1. Lisa A Calder1,3,
  2. Alan J Forster2,3,
  3. Ian G Stiell1,3,
  4. Laura K Carr3,
  5. Jamie C Brehaut3,
  6. Jeffrey J Perry1,3,
  7. Christian Vaillancourt1,3,
  8. Patrick Croskerry4
  1. 1Department of Emergency Medicine, University of Ottawa, Ottawa, Ontario, Canada
  2. 2Department of Medicine, University of Ottawa, Ottawa, Ontario, Canada
  3. 3Department of Epidemiology and Community Medicine, University of Ottawa, Ottawa, Ontario, Canada
  4. 4Department of Emergency Medicine, Dalhousie University, Halifax, Nova Scotia, Canada
  1. Correspondence to Dr Lisa A Calder, The Ottawa Hospital—Civic Campus, 1053 Carling Avenue, F654, Box 685, Ottawa, ON K1Y 4E9, Canada; lcalder{at}ohri.ca

Abstract

Background Dual-process psychological theories argue that clinical decision making is achieved through a combination of experiential (fast and intuitive) and rational (slower and systematic) cognitive processes.

Objective To determine whether emergency physicians perceived their clinical decisions in general to be more experiential or rational and how this compared with other physicians.

Methods A validated psychometric tool, the Rational Experiential Inventory (REI-40), was sent through postal mail to all emergency physicians registered with the College of Physicians and Surgeons of Ontario, according to their website in November 2009. Forty statements were ranked on a Likert scale from 1 (Definitely False) to 5 (Definitely True). An initial survey was sent out, followed by reminder cards and a second survey to non-respondents. Analysis included descriptive statistics, Student t tests, analysis of variance and comparison of mean scores with those of cardiologists from New Zealand.

Results The response rate in this study was 46.9% (434/925). The respondents' median age was 41–50 years; they were mostly men (72.6%) and most had more than 10 years of clinical experience (66.8%). The mean REI-40 rational scores were higher than the experiential scores (3.93/5 (SD 0.35) vs 3.33/5 (SD 0.49), p<0.0001), similar to the mean scores of cardiologists from New Zealand (mean rational 3.93/5, mean experiential 3.05/5). The mean experiential scores were significantly higher for female respondents than for male respondents (3.40/5 (SD 0.49) vs 3.30/5 (SD 0.48), p=0.003).

Conclusions Overall, emergency physicians favoured rational decision making rather than experiential decision making; however, female emergency physicians had higher experiential scores than male emergency physicians. This has important implications for future knowledge translation and decision support efforts among emergency physicians.

  • Decision making
  • emergency medicine
  • questionnaires emergency services
  • hospital
  • emergency department

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

For decades, the discipline of cognitive psychology has extensively explored the cognitive issues behind decision making. The implications of this research for emergency medicine have only recently been described.1–4 To date, no previous study has examined whether emergency physicians share common decision-making characteristics.

The basis for clinical decision making is complex. Dual-process theories suggest that decisions are made by two modes: system 1 and system 2.5–11 System 1, also known as experiential decision making, involves intuitive decisions and the use of heuristics or rules of thumb. Little conscious evaluation is required for system 1. System 2, also known as rational decision making, incorporates critical evaluation of evidence and a structured process that requires more time and conscious effort. Both systems interact with each other; however, it has been demonstrated that people often have affinity for one decision-making style over another.12 It has been postulated that emergency physicians use both systems in their work and one can imagine clinical scenarios whereby experiential decision making is a necessity for rapid, life-saving decision making (eg, the decision to defibrillate), and other scenarios where rational decision making would be appropriate to navigate complicated diagnostic or management issues (eg, the decision to admit or discharge a frail elderly patient with pneumonia and multiple comorbidities).13

It is not known whether emergency physicians as a group tend towards a particular decision-making style. It has been suggested that as much as certain medical specialties attract certain personality characteristics, cognitive styles similarly may be clustered.10 It is possible that this could be because of a given specialty, practice environment or training background. For example, cardiologists conduct most of their clinical practice in outpatient clinics and hospital wards. Decision making in these settings is very different from that of an emergency department, where limited resources, cognitive overloading and the urgency of clinical presentations come into play. Understanding decision-making styles has important implications for those interested in enhancing safe decision making, especially the efficacy of diagnosis and the mitigation of medical error. We hypothesised that emergency physicians would have a common decision-making style distinct from that of cardiologists, the only other population of clinicians who have been assessed for decision-making characteristics. The purpose of this study was to apply a psychometric tool to a sample of emergency physicians to determine whether they perceived their overall clinical decisions to be more experiential or rational and to compare these findings with those of cardiologists assessed by the same tool.

Methods

Study design

This was a postal survey using a previously published psychometric tool, the Rational Experiential Inventory (REI-40), and a demographic panel of questions administered to all actively practising emergency physicians in Ontario.5 The study was approved by the Ottawa Hospital Research Ethics Board.

Study population

Inclusion criteria

We included all actively practising English-speaking emergency physicians in Ontario who were registered with the provincial licensing body, the College of Physicians and Surgeons of Ontario, according to their website in November 2009. The search results listed 1065 emergency physicians. We accepted surveys from retired emergency physicians as long as they had completed the survey in its entirety.

Exclusion criteria

We excluded emergency physicians whose mailing address was not in Ontario or who responded that they were no longer practising emergency medicine.

Survey instrument

We used a psychometric tool, REI-40, which has been validated in many diverse populations, to determine decision-making style.5 ,12 ,14 ,15 These validation studies have found that the tool is internally consistent, highly reliable (Cronbach's α ranging from 0.74 to 0.91) and consistent with other dual-process theory assessment tools.5 ,11 ,12 ,14 The survey consisted of 40 statements where respondents indicated their response on a 5-item Likert scale, ranging from Definitely False (1) to Definitely True (5). We modified the statements to refer to ‘my clinical work’ rather than ‘my life’. An example of a statement used was: “I like to rely on my intuitive impressions in my clinical work.” We appended eight demographic questions at the end of the survey (eg. gender, age, practice setting). We piloted the survey among 22 emergency residents to assess face validity.

Data collection procedures

We administered the survey using a modified Dillman procedure.16 We mailed an anonymised REI-40 to the sample of emergency physicians in January 2010. The survey included a cover letter outlining the goal of the study, contact information and assuring respondents of the confidentiality of their responses and their anonymity for the analysis and data reporting. The cover letter stated that consent was implied if the survey was returned. We enclosed a prepaid postage envelope. Three weeks after the mailing of the initial survey, we sent out reminder cards, and another three weeks later we sent a second copy of the questionnaire to a random sample (n=200) of non-respondents. We selected a random sample using an online random generator (http://www.random.org). In addition, research staff had a booth set up at a local emergency medicine conference (2010 National Capital Conference on Emergency Medicine—NCCEM, February 2010) to encourage emergency physicians to respond. No incentives were provided for completing this survey.

We entered the returned survey responses into a Microsoft Excel database and performed quality assurance checks on every 20th survey entered to ensure accuracy of data entry (proportion of data entry errors=0.7%).

Data analysis

We analysed demographic data using descriptive statistics and calculated medians for skewed data as appropriate. We coded the responses on the basis of an unpublished coding manual provided by Dr Epstein, who devised the REI-40. We calculated the mean scores and standard deviations for rational and experiential decision making according to Pacini and Epstein's protocol and those of other studies after confirming that the variables were normally distributed.5 ,11 Each of the 40 items (20 items per domain) had a score between 1 and 5. The scores for each domain (rational and experiential) were averaged to provide variables ranging from 1 to 5. A higher score reflected a greater tendency to endorse the decision-making style being measured.11 We calculated the 95% CI for the mean difference between mean rational and experiential scores. We used an independent Student t test to assess differences between mean scores and analysis of variance by comparing the mean scores between demographic categories with multiple levels. We compared our rational and experiential mean scores with published scores from other previous studies using the REI-40 tool, including the only published study of a physician population.5 ,11 We analysed all data using SAS V.9.1 (SAS Institute Inc.).

Sample size

We did not perform a formal sample size calculation for this study because our intent was to survey the entire list of emergency physician members of the College of Physicians and Surgeons of Ontario. We did, however, determine that for an anticipated response rate of 50%, a 95% two-sided CI around a single mean would yield a margin of error of 0.03, assuming the SD was 0.5 and accounting for the finite population correction factor.

Results

We received surveys from 434/925 emergency physicians (response rate 46.9%) between January and June 2010 (figure 1). Participants' characteristics are described in table 1.

Figure 1

Flow of postal survey responses from 434 emergency physicians.

Table 1

Characteristics of 434 emergency physician respondents

The majority of respondents were men (72.6%), with a median age of 41–50 years. Over half (55.1%) of the respondents spent more than 75% of their time in clinical duties. The majority (64.7%) of respondents worked in academic or teaching community settings. Although we do not have specific data on the non-respondents, the College of Physicians and Surgeons at the time had 77.6% male members who were emergency physicians.

Emergency physicians' mean rational score was 3.93 (SD 0.35) and the mean experiential score was 3.33 (SD 0.49). The difference of 0.60 between these scores was statistically significant (p<0.0001) and the estimate is precise (95% CI 0.56 to 0.64). The mean scores were normally distributed on the basis of the SAS V.9.1 test for normality and demonstrated discrimination but also some overlap between the two decision-making domains (figure 2). We compared the mean scores for different demographic categories (table 2).

Figure 2

Distribution of Rational Experiential Inventory (REI-40) scores (experiential and rational) among 434 respondents.

Table 2

Comparison of mean Rational Experiential Inventory (REI-40) scores for 434 respondents on the basis of demographics

Although, overall, female emergency physicians tended more towards rational decision making than towards experiential decision making, their mean experiential scores were significantly higher than those of male respondents (3.40 (SD 0.49) vs 3.30 (SD 0.48), p=0.003), whereas the mean rational scores of male respondents were higher than those of female respondents (3.94 (SD 0.35) vs 3.89 (SD 0.33), p=0.03) We found that emergency physicians who underwent specialty training (FRCPC) for 5 years had higher mean rational scores than those who underwent training in family medicine (CCFP-EM) for 3 years (4.01 (SD 0.34) vs 3.90 (SD 0.35), p=0.0001). Respondents practising in urban centres had significantly higher rational scores than those practising in rural centres (3.95 (SD 0.35) vs 3.90 (SD 0.33), p=0.03).

Our systematic search of the literature found two other studies reporting REI-40 scores in other populations, only one of which used a population of clinicians (cardiologists from New Zealand).11 This latter study used the REI-40 to assess thinking styles of the cardiologists with regard to acute coronary syndrome guidelines. See table 3 for comparisons of mean rational and experiential scores against those published for other populations.

Table 3

Comparison of mean Rational Experiential Inventory (REI-40) scores for 434 emergency physicians with other study samples

The greatest differentiation between the mean rational and experiential scores was found for the cardiologists from New Zealand, followed by emergency physicians and college students.

Sixteen per cent of respondents provided comments in the survey, many of them quite animated and passionate. Several respondents expressed that they used both thinking styles to help aid their clinical decision making (box 1).

Box 1

Sample of survey comments

Comments from respondents

“For me there is sometimes a conflict in my mind after an ER shift. I reflect upon the case and am aware I made a ‘gut’ decision. I sometimes second guess the ‘gut’ decision—I will follow up with the patients and my gut is almost 100% right.”

“Emergency medicine doesn't afford the physician the luxury of pontificating over complex cases. I think a lot of ‘intuition’ comes from knowledge and experience, especially in dealing with undifferentiated problems.”

“It is much easier to make snap judgements and rely on gut feelings for complicated cases. I force myself to use logic to not miss big things. Using logic is tiring but satisfying.”

Je n'aime pas le terme—‘gut feeling’”=“I don't like the term ‘gut feeling’”

Discussion

Interpretation of findings

This is the first study to examine the decision-making styles of emergency physicians. We found that, overall, emergency physicians from Ontario favoured rational decision making, a more analytical and structured process. We found a narrower difference between their mean rational and experiential scores than between the scores for the cardiologists from New Zealand.11 Female emergency physicians favoured experiential, more intuitive decision making, which was consistent with what has been found in previous studies.5 ,17 Respondents indicated that they used both decision-making styles, but the overall results indicated statistically significant differences between those who endorsed rational versus experiential decision making. We found that emergency physicians who underwent 5 years of specialty training were more likely to score higher on rational decision making than those who underwent 3 years of family-medicine-based emergency training. Urban physicians favoured rational decision making when compared to rural physicians. Previous authors have expressed a need to better understand emergency physicians' psychological characteristics to be able to design effective knowledge translation interventions.4 These findings give important insight into how emergency physicians make decisions and could be used in future knowledge translation efforts.

Although decision making is complex with many individual, cultural, social, environmental as well as context-specific influences, the dual process theories are well established in the psychology literature.5 ,10–12 ,14 ,15 ,18 Our findings support that emergency physicians use both rational and experiential decision-making styles but that they endorse more strongly the rational approach. This is consistent with a previous study of Australian emergency physicians using the Myers Briggs Type Indicator which found that emergency physicians scored highest for the ‘thinking’ trait, implying that they preferred reasonable, logical and causal decision making.19 Our findings could be explained by the higher proportion of men in the sample, but could also represent a cultural view among emergency physicians that rational decision making is ‘better’ than experiential decision making. This was mentioned in some of the comments we received. It is also possible that this finding is due to social desirability bias. This view contrasts, however, with the findings of Witteman et al12 who describe intuitive thinking to be as powerful and accurate as analysis. Thus, both styles are needed for clinical decision making and one style is not superior to another. Sladek et al10 suggest that people tend to select rational decision making when ‘the stakes are high’. Many would argue that this is a common situation in the emergency department.

Comparison of emergency physicians with other populations

Although several studies have validated the REI-40 as a psychometric tool,5 ,12 ,14 ,15 we were only able to find two other studies that had reported REI-40 scores.5 ,11 Not surprisingly, our study sample contrasted quite significantly with that of American college students, who tend more towards experiential decision making.5 An interesting similarity was the finding in both populations of female respondents identifying more with experiential decision making than male respondents. Compared to the cardiologists from New Zealand, emergency physicians from Ontario had similar decision-making styles, favouring more rational than experiential decision making.11 The difference between mean scores was greater for the cardiologists and that study did not include any female respondents.

Research implications

Although we make decisions in complex ways and differently in certain situations, as clinicians, we know little about our overall decision-making style. This information could be useful when considering change management within the healthcare system. Some authors suggest that people who favour rational decision making may be more open to knowledge translation efforts and evidence-based medicine.10 Our findings suggest that decision support tools should be designed to take into account both decision-making styles. It is possible that female emergency physicians would respond differently from male physicians to certain decision support tools. The same may be true for physicians from different training backgrounds and different practice settings. This might suggest that these tools need careful refinement and specificity. The other implication for this research is that a tool such as the REI-40 could be used for self-assessment. These findings are relevant during the clinical patient encounter if one is aware of the decision-making style being used and its inherent limitations. For example, if a physician is assessing a patient with chest pain and using experiential decision making to diagnose musculoskeletal pain, the physician can then use rational decision making to verify that important differential diagnoses have not been overlooked. If one is aware of one's overall decision-making style, one might be able to better engage in metacognition, the process of “thinking about how we think”, to address abundant cognitive biases.3

Limitations

This study has several limitations. First, as with all voluntary surveys, there is the risk of self-selection bias. We were able to determine that our sample has a similar demographic profile to that of Canadian emergency physicians (19.7% women and 80.3% men).20 Second, it is possible that people who favoured rational decision making were also more likely to respond to a survey. Third, we also had a single physician comparator population in a different country, and in that study REI-40 was used to assess decision-making styles with respect to clinical guidelines in acute coronary syndrome rather than overall clinical practice.11 It is possible that the similarities we found between cardiologists and emergency physicians are unique to these two groups and that marked differences may be noted when comparing these results with other specialties. Fourth, our response rate was low but in keeping with typical response rates for physician surveys.21 And finally, our sample was skewed to more male respondents; however, this is representative of the emergency physician population in Canada as a whole.

Conclusions

We assessed the overall decision-making style of emergency physicians and found that, although both experiential and rational styles are used in clinical decision making, emergency physicians favour rational decision making. This was similar to what was observed in a population of cardiologists. We believe these results are generalisable to all Canadian emergency physicians, and this has important implications for knowledge translation efforts in this population as well as strategies aimed at reducing error in decision making. Future studies should consider assessing the decision-making styles of other physician specialties and scenario-based decision-making styles, and using these data when designing decision support tools.

Acknowledgments

LAC, JJP, CV, JCB, IGS and AJF thank the Ottawa Hospital Research Institute for support. All the authors thank Dr Epstein for giving us permission to use this tool and providing us with the coding manual. Thanks also to Dr Monica Taljaard for assisting us with our questions on the statistical analysis.

References

Footnotes

  • Funding This study was funded by the Emergency Medicine Patient Safety Foundation and the Society for Academic Emergency Medicine, Des Plaines, Illinois. Fellowship funds were provided for research and training in patient safety. The funder was not involved in data analysis or in altering the manuscript.

  • Competing interests None to declare.

  • Ethics approval Ottawa Hospital Research Ethics Board.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles

  • Miscellaneous
    BMJ Publishing Group Ltd and the British Association for Accident & Emergency Medicine
  • Primary survey
    Geoffrey Hughes