Does Walking Help to Generate a Differential Diagnosis?

TheoryIn Medicine, arriving at the correct diagnosis is of paramount importance for patient health and safety, yet is a difficult task especially when a patient presents with symptoms that do not fit typical patterns of disease. This task can be further complicated by errors of judgment, with the failure to consider all possible diagnoses being the most common of such errors. In this study, we investigated the process of differential diagnosis generation within the growing evidence that diagnostic performance can be increased by activities such as walking as was previously shown in Oppezzo and Schwartz's 2014 study. Hypotheses: It was hypothesized that an increase in performance, as expressed by a greater number of plausible differential diagnoses, would be seen in the walking group. Method: Eighteen medical students in their last two months of pre-clerkship training and eighteen second year family medicine residents were shown four different lists of a constellation of signs and symptoms. Participants were asked to generate differential diagnoses over five minutes per each list. All participants sat when completing the first two lists (pretest phase), and then were equally and randomly assigned to sitting versus walking on a treadmill when completing the last two lists (post-test phase). The number of total and unique differential diagnoses generated was determined, before being submitted to a three-member expert panel who identified appropriate unique differential diagnoses. Results: Two-way mixed ANOVAs were conducted to investigate the impact of exercise on the number of total, unique, and appropriate unique ideas generated and compared between pretest and post-test phases. Conclusions: We conclude that there is neither an increase nor a decrease in the number or quality of differential diagnoses generated by the sitting and walking groups within a population that has acquired some level of expertise.

Relationship quality in higher education and the interplay with student engagement and loyalty.

BACKGROUND: To date, studies that have investigated the bonds between students and their institution have emphasized the importance of student-staff relationships. Measuring the quality of those relationships (i.e., relationship quality) appears to help with investigating the relational ties students have with their higher education institutions. Growing interest has arisen in further investigating relationship quality in higher education, as it might predict students' involvement with the institution (e.g., student engagement and student loyalty). So far, most studies have used a cross-sectional design, so that causality could not be determined. AIMS: The aim of this longitudinal study was twofold. First, we investigated the temporal ordering of the relation between the relationship quality dimensions of trust (in benevolence and honesty) and affect (satisfaction, affective commitment, and affective conflict). Second, we examined the ordering of the paths between relationship quality, student engagement, and student loyalty. Our objectives were to gain a deeper understanding of the relationship quality construct in higher education and its later outcomes. SAMPLE: Participants (N = 1649) were students from three Dutch higher education institutions who were studying in a technology economics or social sciences program. METHODS: Longitudinal data from two time points were used to evaluate two types of cross-lagged panel models. In the first analysis, we could not assume measurement invariance for affective conflict over time. Therefore, we tested an alternative model without affective conflict, using the latent variables of trust and affect, the student engagement dimensions and student loyalty. In the second type of model, we investigated the manifest variables of relationship quality, student engagement, and student loyalty. The hypotheses were tested by evaluating simultaneous comparisons between estimates. RESULTS: Results indicated that the relation between relationship quality at Time 1 with student engagement and loyalty at Time 2 was stronger than the reverse ordering in the first model. In the second model, results indicated that cross-lagged relations between trust in benevolence and trust in honesty at Time 1 and affective commitment, affective conflict, and satisfaction at Time 2 were more likely than the reverse ordering. Furthermore, cross-lagged relations from relationship quality at Time 1 to student engagement and student loyalty at Time 2 also supported our hypothesis. CONCLUSIONS: This study contributes to the existing higher education literature, indicating that students' trust in the quality of their relationship with faculty/staff is essential for developing students' affective commitment and satisfaction and for avoiding conflict over time. Second, relationship quality factors positively influence students' engagement in their studies and their loyalty towards the institution. A relational approach to establishing (long-lasting) bonds with students appears to be fruitful as an approach for educational psychologists and for practitioners' guidance and strategies. Recommendations are made for future research to further examine relationship quality in higher education in Europe and beyond.

Does walking improve diagnosis of skin conditions at varying levels of medical expertise?

The use of walking workstations in educational and work settings has been shown to improve cognitive abilities. At the same time, it has been repeatedly shown that medical residents around the world do not meet exercise guidelines, mainly due to a scarcity of available free time. Our study investigates the boundaries of the previously observed phenomenon of improved cognitive performance with physical activity using materials that represent real life tasks. Participants had different level of expertise and involved second year psychology students, medical students, and family medicine residents. We examined the effect of being physically inactive (i.e., sitting) or active (i.e., walking) while diagnosing multiple complex presentations of four skin conditions. We assumed that being physically active, irrespective of the level of expertise, will bolster diagnostic performance. Our findings show, however, that being physically active does not change the performance level of participants with different levels of medical expertise. Implications for medical education and suggestions for further research will be discussed.

The testing effect for mediator final test cues and related final test cues in online and laboratory experiments.

BACKGROUND: The testing effect is the finding that information that is retrieved during learning is more often correctly retrieved on a final test than information that is restudied. According to the semantic mediator hypothesis the testing effect arises because retrieval practice of cue-target pairs (mother-child) activates semantically related mediators (father) more than restudying. Hence, the mediator-target (father-child) association should be stronger for retrieved than restudied pairs. Indeed, Carpenter (2011) found a larger testing effect when participants received mediators (father) than when they received target-related words (birth) as final test cues. METHODS: The present study started as an attempt to test an alternative account of Carpenter's results. However, it turned into a series of conceptual (Experiment 1) and direct (Experiment 2 and 3) replications conducted with online samples. The results of these online replications were compared with those of similar existing laboratory experiments through small-scale meta-analyses. RESULTS: The results showed that (1) the magnitude of the raw mediator testing effect advantage is comparable for online and laboratory experiments, (2) in both online and laboratory experiments the magnitude of the raw mediator testing effect advantage is smaller than in Carpenter's original experiment, and (3) the testing effect for related cues varies considerably between online experiments. CONCLUSIONS: The variability in the testing effect for related cues in online experiments could point toward moderators of the related cue short-term testing effect. The raw mediator testing effect advantage is smaller than in Carpenter's original experiment.

The Testing Effect and Far Transfer: The Role of Exposure to Key Information.

Butler (2010: Experiment 3) showed that retrieval practice enhanced transfer to a new knowledge domain compared to rereading. The first experiment of the present study was a direct replication of Butler's third experiment. Participants studied text passages and then either reread them three times or went through three cycles of cued recall questions (i.e., retrieval practice) with feedback. As in Butler's (2010) experiment, an advantage of retrieval practice on the final far transfer test emerged after 1 week. Additionally, we observed an advantage of retrieval practice on the final test administered after 5 min. However, these advantages might have been due to participants in the retrieval practice condition receiving focused exposure to the key information (i.e., the feedback) that was needed to answer the final test questions. We therefore conducted a second experiment in which we included the retrieval practice condition and the reread condition from our first experiment, as well as a new reread-plus-statements condition. In the reread-plus-statements condition, participants received focused exposure to the key information after they had reread a text. As in Experiment 1, we found a large effect on far transfer when retrieval practice was compared to rereading. However, this effect was substantially reduced when retrieval practice was compared to the reread-plus-statements condition. Taken together, the results of the present experiments demonstrate that Butler's (2010) testing effect in far transfer is robust. Moreover, focused exposure to key information appears to be a significant factor in this far transfer testing effect.

Age-related decline and diagnostic performance of more and less prevalent clinical cases.

Since cognitive abilities have been shown to decrease with age, it is expected that older physicians would not perform as well as their younger counterparts on clinical cases unless their expertise can counteract the cognitive effects of aging. However, studies on the topic have shown contradictory results. This study aimed to further investigate the effect of aging on physicians' diagnostic accuracy when diagnosing prevalent and less prevalent cases based on clinical vignettes. A mixed design was used to assess the influence of case prevalence (high vs. low) as a within-subjects factor, and age group as a between subjects factor (<30; n = 23, 30-39; n = 19, 40-49; n = 27, >50 years old; n = 19) on the diagnostic accuracy of 65 family physicians and 25 residents. Repeated Measure ANOVA revealed a significant effect of case prevalence (p < .001) and age group (p < .001). Post-hoc analyses revealed that younger physicians showed the best performance. This study did not demonstrate the positive effect of experience in older physicians. In line with previous studies on expertise development, findings of the present study suggest that skills should be actively maintained to assure a high performance level throughout one's lifespan. If not, performance level could gradually decline with age.

Self-regulated learning and academic performance in medical education.

CONTENT: Medical schools aim to graduate medical doctors who are able to self-regulate their learning. It is therefore important to investigate whether medical students' self-regulated learning skills change during medical school. In addition, since these skills are expected to be helpful to learn more effectively, it is of interest to investigate whether these skills are related to academic performance. METHODS: In a cross-sectional design, the Self-Regulation of Learning Self-Report Scale (SRL-SRS) was used to investigate the change in students' self-regulated learning skills. First and third-year students (N = 949, 81.7%) SRL-SRS scores were compared with ANOVA. The relation with academic performance was investigated with multinomial regression analysis. RESULTS: Only one of the six skills, reflection, significantly, but positively, changed during medical school. In addition, a small, but positive relation of monitoring, reflection, and effort with first-year GPA was found, while only effort was related to third-year GPA. CONCLUSIONS: The change in self-regulated learning skills is minor as only the level of reflection differs between the first and third year. In addition, the relation between self-regulated learning skills and academic performance is limited. Medical schools are therefore encouraged to re-examine the curriculum and methods they use to enhance their students' self-regulated learning skills. Future research is required to understand the limited impact on performance.

The development of self-regulated learning during the pre-clinical stage of medical school: a comparison between a lecture-based and a problem-based curriculum.

Society expects physicians to always improve their competencies and to be up to date with developments in their field. Therefore, an important aim of medical schools is to educate future medical doctors to become self-regulated, lifelong learners. However, it is unclear if medical students become better self-regulated learners during the pre-clinical stage of medical school, and whether students develop self-regulated learning skills differently, dependent on the educational approach of their medical school. In a cross-sectional design, we investigated the development of 384 medical students' self-regulated learning skills with the use of the Self-Regulation of Learning Self-Report Scale. Next, we compared this development in students who enrolled in two distinct medical curricula: a problem-based curriculum and a lectured-based curriculum. Analysis showed that more skills decreased than increased during the pre-clinical stage of medical school, and that the difference between the curricula was mainly caused by a decrease in the skill evaluation in the lecture-based curriculum. These findings seem to suggest that, irrespective of the curriculum, self-regulated learning skills do not develop during medical school.

Non-cognitive selected students do not outperform lottery-admitted students in the pre-clinical stage of medical school.

Medical schools all over the world select applicants using non-cognitive and cognitive criteria. The predictive value of these different types of selection criteria has however never been investigated within the same curriculum while using a control group. We therefore set up a study that enabled us to compare the academic performance of three different admission groups, all composed of school-leaver entry students, and all enrolled in the same Bachelor curriculum: students selected on non-cognitive criteria, students selected on cognitive criteria and students admitted by lottery. First-year GPA and number of course credits (ECTS) at 52 weeks after enrollment of non-cognitive selected students (N = 102), cognitive selected students (N = 92) and lottery-admitted students (N = 356) were analyzed. In addition, chances of dropping out, probability of passing the third-year OSCE, and completing the Bachelor program in 3 years were compared. Although there were no significant differences between the admission groups in first-year GPA, cognitive selected students had obtained significantly more ECTS at 52 weeks and dropped out less often than lottery-admitted students. Probabilities of passing the OSCE and completing the bachelor program in 3 years did not significantly differ between the groups. These findings indicate that the use of only non-cognitive selection criteria is not sufficient to select the best academically performing students, most probably because a minimal cognitive basis is needed to succeed in medical school.

Conscious versus unconscious thinking in the medical domain: the deliberation-without-attention effect examined.

Previous studies have shown that with important decisions, unconscious thought has surprisingly led to better choices than conscious thought. The present study challenges this so-called 'deliberation-without-attention effect' in the medical domain. In a computerized study, physicians and medical students were asked, after either conscious or unconscious thought, to estimate the 5-year survival probabilities of four fictitious patients with varying medical characteristics. We assumed that experienced physicians would outperform students as a result of their superior knowledge. The central question was whether unconscious thought in this task would lead to better performance in experts or novices, in line with the deliberation-without-attention effect. We created four fictitious male 60-year-old patients, each of whom with signs and symptoms related to likely prognosis, from 12 (Complex) or 4 (Simple) categories. This manipulation resulted in objectively different life expectancies for these patients. Participants (86 experienced physicians and 57 medical students) were randomly allocated to the Simple or Complex condition. Statements were randomly presented for 8 s. Next, each participant assessed the life expectancies after either conscious or unconscious thought. As expected, experienced physicians were better in assessing life expectancies than medical students. No significant differences were found in performance between conscious and unconscious thought, nor did we detect a significant interaction between expertise level and mode of thought. In a medical decision task, unconscious thought did not lead to better performance of experienced physicians or medical students than conscious thought. Our findings do not support the deliberation-without-attention effect.

Exposure to media information about a disease can cause doctors to misdiagnose similar-looking clinical cases.

PURPOSE: Anecdotal evidence indicates that exposure to media-distributed disease information, such as news about an outbreak, can lead physicians to errors; influenced by an availability bias, they misdiagnose patients with similar-looking but different diseases. The authors investigated whether exposure to media-provided disease information causes diagnostic errors and whether reflection (systematic review of findings) counteracts bias. METHOD: In 2010, 38 internal medicine residents first read the Wikipedia entry about one or another of two diseases (Phase 1). Six hours later, in a seemingly unrelated study, they diagnosed eight clinical cases (Phase 2). Two cases superficially resembled the disease in the Wikipedia entry they had read (bias expected), two cases resembled the other disease they had not read about (bias not expected), and four were filler cases. In Phase 3, they diagnosed the bias-expected cases again, using reflective reasoning. RESULTS: Mean diagnostic accuracy scores (Phase 2; range: 0-1) were significantly lower on bias-expected cases than on bias-not-expected cases (0.56 versus 0.70, P = .016) because participants misdiagnosed cases that looked similar to a Wikipedia description of a disease more often when they had read the Wikipedia description (mean = 0.61) than when they had not (mean = 0.29). Deliberate reflection (Phase 3) restored performance on bias-expected cases to pre-bias levels (mean = 0.71). CONCLUSIONS: Availability bias may arise simply from exposure to media-provided information about a disease, causing diagnostic errors. The bias's effect can be substantial. It is apparently associated with nonanalytical reasoning and can be counteracted by reflection.

Students' self-explanations while solving unfamiliar cases: the role of biomedical knowledge.

OBJECTIVE: General guidelines for teaching clinical reasoning have received much attention, despite a paucity of instructional approaches with demonstrated effectiveness. As suggested in a recent experimental study, self-explanation while solving clinical cases may be an effective strategy to foster reasoning in clinical clerks dealing with less familiar cases. However, the mechanisms that mediate this benefit have not been specifically investigated. The aim of this study was to explore the types of knowledge used by students when solving familiar and less familiar clinical cases with self-explanation. METHODS: In a previous study, 36 third-year medical students diagnosed familiar and less familiar clinical cases either by engaging in self-explanation or not. Based on an analysis of previously collected data, the present study compared the content of self-explanation protocols generated by seven randomly selected students while solving four familiar and four less familiar cases. In total, 56 verbal protocols (28 familiar and 28 less familiar) were segmented and coded using the following categories: paraphrases, biomedical inferences, clinical inferences, monitoring statements and errors. RESULTS: Students provided more self-explanation segments from less familiar cases (M = 275.29) than from familiar cases (M = 248.71, p = 0.046). They provided significantly more paraphrases (p = 0.001) and made more errors (p = 0.008). A significant interaction was found between familiarity and the type of inferences (biomedical versus clinical, p = 0.016). When self-explaining less familiar cases, students provided significantly more biomedical inferences than familiar cases. CONCLUSIONS: Lack of familiarity with a case seems to stimulate medical students to engage in more extensive thinking during self-explanation. Less familiar cases seem to activate students' biomedical knowledge, which in turn helps them to create new links between biomedical and clinical knowledge, and eventually construct a more coherent mental representation of diseases. This may clarify the previously found positive effect that self-explanation has on the diagnosis of unfamiliar cases.

From see one do one, to see a good one do a better one: learning physical examination skills through peer observation.

BACKGROUND: Learning and mastering the skills required to execute physical exams is of great importance and should be fostered early during medical training. Observing peers has been shown to positively influence the acquisition of psychomotor skills. PURPOSE: The current study investigated the influence of peer observation on the acquisition of psychomotor skills required to execute a physical examination. METHODS: Second-year medical students (N=194) learned the neurological physical examination for low back pain in groups of three. Each student learned and performed the physical examination while the other students observed. Analyses compared the impact of the quantity and the quality of observed performances on students' learning of the physical examination skills. RESULTS: Students benefited from observing peers while they executed their examination. Moreover, observing a high-performing peer increased the acquisition of physical examination skills. CONCLUSIONS: Results suggest that group learning activities that allow students to observe their peers during physical examination should be favored.

To observe or not to observe peers when learning physical examination skills; that is the question.

BACKGROUND: Learning physical examination skills is an essential element of medical education. Teaching strategies include practicing the skills either alone or in-group. It is unclear whether students benefit more from training these skills individually or in a group, as the latter allows them to observing their peers. The present study, conducted in a naturalistic setting, investigated the effects of peer observation on mastering psychomotor skills necessary for physical examination. METHODS: The study included 185 2nd-year medical students, participating in a regular head-to-toe physical examination learning activity. Students were assigned either to a single-student condition (n = 65), in which participants practiced alone with a patient instructor, or to a multiple-student condition (n = 120), in which participants practiced in triads under patient instructor supervision. The students subsequently carried out a complete examination that was videotaped and subsequently evaluated. Student's performance was used as a measure of learning. RESULTS: Students in the multiple-student condition learned more than those who practiced alone (81% vs 76%, p < 0.004). This result possibly derived from a positive effect of observing peers; students who had the possibility to observe a peer (the second and third students in the groups) performed better than students who did not have this possibility (84% vs 76%, p <. 001). There was no advantage of observing more than one peer (83.7% vs 84.1%, p > .05). CONCLUSIONS: The opportunity to observe a peer during practice seemed to improve the acquisition of physical examination skills. By using small groups instead of individual training to teach physical examination skills, health sciences educational programs may provide students with opportunities to improve their performance by learning from their peers through modelling.

Experienced physicians benefit from analyzing initial diagnostic hypotheses.

BACKGROUND: Most incorrect diagnoses involve at least one cognitive error, of which premature closure is the most prevalent. While metacognitive strategies can mitigate premature closure in inexperienced learners, these are rarely studied in experienced physicians. Our objective here was to evaluate the effect of analytic information processing on diagnostic performance of nephrologists and nephrology residents. METHODS: We asked nine nephrologists and six nephrology residents at the University of Calgary and Glasgow University to diagnose ten nephrology cases. We provided presenting features along with contextual information, after which we asked for an initial diagnosis. We then primed participants to use either hypothetico-deductive reasoning or scheme-inductive reasoning to analyze the remaining case data and generate a final diagnosis. RESULTS: After analyzing initial hypotheses, both nephrologists and residents improved the accuracy of final diagnoses (31.1% vs. 65.6%, p < 0.001, and 40.0% vs. 70.0%, p < 0.001, respectively). We found a significant interaction between experience and analytic processing strategy (p = 0.02): nephrology residents had significantly increased odds of diagnostic success when using scheme-inductive reasoning (odds ratio [95% confidence interval] 5.69 [1.59, 20.33], p = 0.07), whereas the performance of experienced nephrologists did not differ between strategies (odds ratio 0.57 [0.23, 1.39], p = 0.20). DISCUSSION: Experienced nephrologists and nephrology residents can improve their performance by analyzing initial diagnostic hypotheses. The explanation of the interaction between experience and the effect of different reasoning strategies is unclear, but may relate to preferences in reasoning strategy, or the changes in knowledge structure with experience.

Reflection as a strategy to foster medical students' acquisition of diagnostic competence.

OBJECTIVES: Developing diagnostic competence in students is a major goal of medical education, but there is little empirical evidence on instructional strategies that foster the acquisition of this competence. The aim of this study was to investigate the effects of structured reflection compared with the generation of immediate or differential diagnosis while practising with clinical cases on learning clinical diagnosis. METHODS: This was a three-phase experimental study. During a learning phase, 46 Year 4 students diagnosed six clinical cases under different experimental conditions: structured reflection, immediate diagnosis, or differential diagnosis. This was followed by an immediate test and a delayed test administered 1 week later. Each test consisted of diagnosing four different cases of diseases presented in the learning phase. Performance in diagnosing these new cases was used as a measure of learning. RESULTS: Repeated-measures analysis of variance on the mean diagnostic accuracy scores (range: 0-1) showed a significant interaction between performance moment (i.e. performance in the learning phase and on each test) and instructions followed during the learning phase (p=0.003). Follow-up analyses of this interaction showed that diagnostic performance did not differ between conditions in the learning phase. On the immediate test, scores in the reflection condition (mean=0.48, 95% confidence interval [CI] 0.38-0.58) were significantly lower than scores in the differential diagnosis condition (mean=0.62, 95% CI 0.54-0.70; p=0.012) and marginally lower than those in the immediate diagnosis condition (mean=0.61, 95% CI 0.52-0.70; p=0.04). One week later, however, scores in the reflection condition (mean=0.66, 95% CI 0.56-0.76) significantly outperformed those in the other conditions (differential diagnosis: mean=0.48, 95% CI 0.37-0.58 [p<0.01]; immediate diagnosis: mean=0.52, 95% CI 0.43-0.60 [p=0.01]). Comparisons within experimental conditions showed that performance from the immediate to the delayed test decreased in the immediate and differential diagnosis conditions (immediate diagnosis: p=0.042; differential diagnosis: p=0.012), but increased in the reflection condition (p=0.003). CONCLUSIONS: Structured reflection while practising with cases appears to foster the learning of clinical knowledge more effectively than the generation of immediate or differential diagnoses and therefore seems to be an effective instructional approach to developing diagnostic competence in students.

Accepting diagnostic suggestions by residents: a potential cause of diagnostic error in medicine.

BACKGROUND: Psychological research has shown that people tend toward accepting rather than refuting hypotheses. Diagnostic suggestions may evoke such confirmatory tendencies in physicians, which may lead to diagnostic errors. PURPOSE: This study investigated the influence of a suggested diagnosis on physicians' diagnostic decisions on written clinical cases. It was hypothesized that physicians would tend to go along with the suggestions and therefore would have more difficulty rejecting incorrect suggestions than accepting correct suggestions. METHODS: Residents (N = 24) had to accept or reject suggested diagnoses on 6 cases. Three of those suggested diagnoses were correct, and 3 were incorrect. RESULTS: Results showed the mean correct evaluation score on cases with a correct suggested diagnosis (M = 2.21, SD = 0.88) was significantly higher than the score on cases with an incorrect suggested diagnosis (M = 1.42, SD = 0.97), meaning physicians indeed found it easier to accept correct diagnoses than to reject incorrect diagnoses, t(23) = 2.74, p < .05, d = .85, despite equal experience with the diagnoses. CONCLUSION: These findings indicate that suggested diagnoses may evoke confirmatory tendencies and consequently may lead to diagnostic errors.

Assessing patient management plans of doctors and medical students: an illness script perspective.

INTRODUCTION: Illness script theory offers explanations for expert-novice differences in clinical reasoning. However, it has mainly focused on diagnostic (Dx) performance, while patient management (Mx) has been largely ignored. The aim of the present study was to show the role of Mx knowledge in illness script development and how it relates to diagnostic knowledge in the course of development toward expertise. METHODS: The participants were 10 fourth-year and 10 sixth-year medical students, and 10 experienced physicians (ie, internists). Participants were asked to study 4 written clinical cases and provide management plans. Based on propositional analysis the management plans were examined for accuracy and elaborateness as well as the number of Mx and Dx items. RESULTS: Providing accurate Mx plans that concurrently pay attention to Dx and Mx was a characteristic of doctors' performance. The Mx plans of sixth-year students were as accurate as fourth-year students, but the format and the size of sixth-year students' plans was more similar to those of doctors. While sixth-year students generated plans with an Mx focus, the plans of fourth-year students were characterized by a Dx focus. DISCUSSION: The experienced physicians' accurate management plans are characterized by a high number of the Mx and Dx items. For sixth-year students the management plans are still incomplete, which leads to generic as well as inaccurate Mx orders. For fourth-year students, the Mx focus is lacking, and hence they seem to treat an Mx task as if it were a Dx task.

Exploring the role of salient distracting clinical features in the emergence of diagnostic errors and the mechanisms through which reflection counteracts mistakes.

BACKGROUND: Flaws in clinical reasoning are present in most diagnostic errors and occur even when physicians have enough knowledge to solve the problem. Deliberate reflection has been shown to improve diagnoses. The sources of faulty reasoning and how reflection counteracts them remain largely unknown. OBJECTIVE: To explore the causes of faulty reasoning and the mechanisms through which reflection neutralises them by investigating the influence of salient distracting clinical features on diagnostic decision-making. DESIGN AND SETTING: In a prior study, 34 internal medicine residents and 50 medical students of the Erasmus Medical Centre, Rotterdam, diagnosed four clinical cases by means of non-analytical reasoning and four by reflective reasoning. In the secondary analysis of the data presented here, five internists independently evaluated the diagnoses and examined the nature of the diagnostic errors in relation to case features that gave rise to these errors. MAIN OUTCOMES: Frequency of incorrect diagnoses caused by salient distracting features made through reflective and non-analytical reasoning. RESULTS: Among residents, reflective reasoning (Mean diagnostic accuracy score (M)=2.09, 95% CI 1.77 to 2.40) led to a significantly higher number of correct diagnoses than non-analytical reasoning (M=1.71, 95% CI 1.37 to 2.04; p=0.03). This higher diagnostic accuracy was associated with fewer incorrect diagnoses triggered by salient distracting clinical features (M=0.47, 95% CI 0.26 to 0.68) compared with non-analytical reasoning (M=0.85, 95% CI 0.59 to 1.11; p=0.02). Students did not benefit from reflection to improve diagnoses. CONCLUSION: Salient features in a case tend to attract physicians' attention and may misdirect diagnostic reasoning when they turn out to be unrelated to the problem, causing errors. Reflection helps by enabling physicians to overcome the influence of distracting features. The lack of effect for students suggests that this is only possible when there is enough knowledge to recognise which features discriminate between alternative diagnoses.

Consistency in diagnostic suggestions does not influence the tendency to accept them.

BACKGROUND: Studies suggest that residents tend to accept diagnostic suggestions, which could lead to diagnostic errors if the suggestion is incorrect. Those studies did not take into account that physicians in clinical practice will mainly encounter correct suggestions. The present study investigated residents' diagnostic performance if they would first encounter a number of correct suggestions followed by a number of incorrect suggestions, and vice versa. It was hypothesized that more incorrect suggestions would be accepted if participants had first evaluated a series of correct suggestions. METHOD: Residents (n = 38) evaluated suggested diagnoses on eight written clinical cases. Half of the participants first evaluated four correct suggestions and then evaluated four incorrect suggestions (C/I condition). The other half started with the four incorrect suggestions followed by the correct suggestions (I/C condition). RESULTS: Our findings show that the evaluation score in the C/I condition (M = 2.87, MSE = 0.14) equaled that in the I/C condition (M = 2.66, MSE = 0.14), F(1,36) = 1.09, p = 0.30, ns, meaning that consistency in preceding suggested diagnoses did not influence the tendency to accept subsequent diagnostic suggestions. There was, however, a significant interaction effect between case order and phase, F(1,36) = 11.82, p = 0.001, η p (2) = 0.25, demonstrating that the score on cases with correct suggestions was higher than the score on cases with incorrect suggestions. CONCLUSION: These findings indicate that consistency in preceding correct or incorrect diagnostic suggestions did not influence the tendency to accept or reject subsequent suggestions. However, overall residents still showed a tendency to accept diagnostic suggestions, which may lead to diagnostic errors if the suggestion is incorrect.