Heuristic decision making in medicine

Frequency of medical students' language expressing implicit uncertainty in simulated handovers

OBJECTIVES

The aim of this study was to investigate the number and type of implicit expressions of uncertainty by medical students during simulated patient handovers.

METHODS

Eighty-seven volunteer medical students, a convenience sample collected on a first-come, first-served basis, participated in simulated handovers. They each worked with three simulated patients who presented with different chief complaints and personal conditions. The handovers were video recorded and transcribed. A framework of implicit expressions of uncertainty was used to identify and count modifiers that attenuate or strengthen medical information using MAXQDA lexical search. We analysed the findings with respect to the patients' contexts.

RESULTS

Implicit uncertainty expressions which attenuate or strengthen information occurred in almost equal frequency, 1879 (55%) versus 1505 (45%). Attenuators were found most frequently in the category 'Questionable', 1041 (55.4%), strengtheners in the category 'Focused', 1031 (68.5%). Most attenuators and strengtheners were found in the handover of two patients with challenging personal conditions ('angry man', 434 (23.1%) versus 323 (21.5%); 'unfocused woman', 354 (19.4%) versus 322 (21.4%)) and one patient with abnormal laboratory findings ('elevated creatinine', 379 (20.2%) versus 285 (18.9%)).

CONCLUSIONS

Medical students use a variety of implicit expressions of uncertainty in simulated handovers. These findings provide an opportunity for medical educators to design communication courses that raise students' awareness for content-dependent implicit expressions of uncertainty and provide strategies to communicate uncertainty explicitly.

Use of heuristics during the clinical decision process from family care physicians in real conditions

RATIONALE AIMS AND OBJECTIVES

The available evidence on the use of heuristics and their relationship with diagnostic error in primary care is very limited. The aim of the study is to identify the use of unknown thought and specifically the possible use of Representativeness, Availability and overconfidence heuristics in the clinical practice of primary care physicians in cases of dyspnoea and to analyse their possible relationship with diagnostic error.

METHODS

A total of 371 patients consulting with new episodes of dyspnoea in Primary Care centres in Spain were registered. Based on specific operational definitions, the use of unconscious thinking and the use of heuristics during the diagnostic process were assessed. Subsequently, the association between their use and diagnostic error was analysed.

RESULTS

In 49.6% of cases, the confirmatory diagnosis coincided with the first diagnostic impression, suggesting the use of the representativeness heuristic in the diagnostic decision process. In 82.3% of the cases, the confirmatory diagnosis was among the three diagnostic hypotheses that were first identified by the general physicians, suggesting a possible use of the availability heuristic. In more than 50% of the cases, the physicians were overconfident in the certainty of their own diagnosis. Finally, a diagnostic error was identified in 9.9% of the recorded cases and no statistically significant correlation was found between the use of some unconscious thinking tools (such as the use of heuristics) and the diagnostic error.

CONCLUSION

Unconscious thinking manifested through the acceptance of the first diagnostic impression and the use of heuristics is commonly used by primary care physicians in the clinical decision process in the face of new episodes of dyspnoea; however, its influence on diagnostic error is not significant. The proposed explicit and reproducible methodology may inspire further studies to confirm these results.

Biasing Influence of 'Mental Shortcuts' on Diagnostic Decision-Making: Radiologists Can Overlook Breast Cancer in Mammograms When Prior Diagnostic Information Is Available

When making decisions under uncertainty, people in all walks of life, including highly trained medical professionals, tend to resort to using 'mental shortcuts', or heuristics. Anchoring-and-adjustment (AAA) is a well-known heuristic in which subjects reach a judgment by starting from an initial internal judgment ('anchored position') based on available external information ('anchoring information') and adjusting it until they are satisfied. We studied the effects of the AAA heuristic during diagnostic decision-making in mammography. We provided practicing radiologists (N = 27 across two studies) a random number that we told them was the estimate of a previous radiologist of the probability that a mammogram they were about to see was positive for breast cancer. We then showed them the actual mammogram. We found that the radiologists' own estimates of cancer in the mammogram reflected the random information they were provided and ignored the actual evidence in the mammogram. However, when the heuristic information was not provided, the same radiologists detected breast cancer in the same set of mammograms highly accurately, indicating that the effect was solely attributable to the availability of heuristic information. Thus, the effects of the AAA heuristic can sometimes be so strong as to override the actual clinical evidence in diagnostic tasks.

Improving Diagnostic Efficiency with Frequency Double-Trees and Frequency Nets in Bayesian Reasoning

Background. Medical students often have problems with Bayesian reasoning situations. Representing statistical information as natural frequencies (instead of probabilities) and visualizing them (e.g., with double-trees or net diagrams) leads to higher accuracy in solving these tasks. However, double-trees and net diagrams (which already contain the correct solution of the task, so that the solution could be read of the diagrams) have not yet been studied in medical education. This study examined the influence of information format (probabilities v. frequencies) and visualization (double-tree v. net diagram) on the accuracy and speed of Bayesian judgments. Methods. A total of 142 medical students at different university medical schools (Munich, Kiel, Goettingen, Erlangen, Nuremberg, Berlin, Regensburg) in Germany predicted posterior probabilities in 4 different medical Bayesian reasoning tasks, resulting in a 3-factorial 2 × 2 × 4 design. The diagnostic efficiency for the different versions was represented as the median time divided by the percentage of correct inferences. Results. Frequency visualizations led to a significantly higher accuracy and faster judgments than did probability visualizations. Participants solved 80% of the tasks correctly in the frequency double-tree and the frequency net diagram. Visualizations with probabilities also led to relatively high performance rates: 73% in the probability double-tree and 70% in the probability net diagram. The median time for a correct inference was fastest with the frequency double tree (2:08 min) followed by the frequency net diagram and the probability double-tree (both 2:26 min) and probability net diagram (2:33 min). The type of visualization did not result in a significant difference. Discussion. Frequency double-trees and frequency net diagrams help answer Bayesian tasks more accurately and also more quickly than the respective probability visualizations. Surprisingly, the effect of information format (probabilities v. frequencies) on performance was higher in previous studies: medical students seem also quite capable of identifying the correct solution to the Bayesian task, among other probabilities in the probability visualizations.

Abuse-Deterrent Opioids: A Survey of Physician Beliefs, Behaviors, and Psychology

Objective

Evaluate beliefs and behaviors pertaining to abuse-deterrent opioids (ADFs).

Design

Survey in 2019 by invitation to all licensed physicians.

Setting

Commonwealth of Kentucky.

Participants

374 physicians.

Methods

Descriptive statistics, and hypothesis test that early adopter prescribers would have greater endorsement of opioid risk management.

Results

Of all prescribers, 55% believed all opioid analgesics should have ADF requirements (15% were unsure); 74% supported mandating insurance coverage. Only one-third considered whether an opioid was ADF when prescribing, motivated by patient family diversion (94%) and societal supply reduction (88%). About half believed ADFs were equally effective in preventing abuse by intact swallowing, injection, chewing, snorting, smoking routes. Only 4% of OxyContin prescribers chose it primarily because of ADF properties. Instead, the most common reason (33%) was being started by another prescriber. A quarter of physicians chose not to prescribe ADFs because of heroin switching potential. Early adopters strongly believed ADFs were effective in reducing abuse (PR 3.2; 95% CI 1.5, 6.6) compared to mainstream physicians. Early-adopter risk-management practices more often included tools increasing agency and measurement: urine drug screens (PR 2.0; 1.3, 3.1), risk screening (PR 1.3; 0.94, 1.9). While nearly all respondents (96%) felt that opioid abuse was a problem in the community, only 57% believed it was a problem among patients in their practice. Attribution theory revealed an externalization of opioid abuse problems that deflected blame from patients on to family members.

Conclusions

The primary motivator for prescribing ADFs was preventing diversion by family members, not patient-level abuse concerns.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40122-021-00343-z.

Provider perspectives on communication and dismissal policies with HPV vaccine hesitant parents

Parental vaccine hesitancy is a growing concern. Less is known about provider or practice characteristics that encounter HPV-specific vaccine-hesitant parents, the providers' confidence in responding to HPV vaccine concerns, and the attitudes and use of vaccine dismissal policies (i.e., removing patients from the practice). North Texas providers completed an online survey. Dependent variables assessed: (1) percentage of HPV vaccine-hesitant parents encountered in practice defined as substantive, or high (≥11%, or among more than one out of ten adolescent patient encounters) versus low (≤10%) levels; (2) confidence in responding to 11 HPV vaccine concerns; (3) attitudes and use of vaccine dismissal policies. Chi-square and Fisher's exact tests were conducted. Among 156 providers, 29% reported high HPV vaccine hesitancy (≥11% of patient population). Overall, providers reported being "very confident" in addressing vaccine concerns (mean: 3.37 out of 4, SD: 0.57). Mean confidence scores were significantly higher for white (vs. non-white) providers and for pediatricians (vs. family practitioners). Providers were least confident in responding to parents' religious/personal beliefs (69%). Some providers (25%) agreed with policies that dismissed vaccine-hesitant parents after repeated counseling attempts. More providers used dismissal policies for childhood (19%) than adolescent (10%) immunizations. Provider communication training should include parental religious/personal beliefs to effectively address HPV vaccine hesitancy. Other regions should examine their HPV-specific vaccine hesitancy levels to understand how the use of dismissal policies might vary between adolescent and childhood immunizations.

Decision Making: Healthy Heuristics and Betraying Biases

Critical care settings are unpredictable, dynamic environments where clinicians face high decision density in suboptimal conditions (stress, time constraints, competing priorities). Experts have described two systems of human decision making: one fast and intuitive; the other slow and methodical. Heuristics, or mental shortcuts, a key feature of intuitive reasoning, are often accurate, applied instinctively, and essential for efficient diagnostic decision making. Heuristics are also prone to failures, or cognitive biases, which can lead to diagnostic errors. A variety of strategies have been proposed to mitigate biases; however, current understanding of such interventions to optimize diagnostic safety is still incomplete.

Should I stay or should I go? An exploration of the decision-making behavior of acute cardiac patients during the COVID-19 pandemic

Background

During the SARS-COV-2 (COVID-19) pandemic efforts to reduce virus transmission resulted in non-emergency patients being deterred from seeking help. The number of patients presenting with acute cardiac conditions reduced, significantly

Objectives

To explore the decision-making process, and influential factors in that process, of patients and their family during an acute cardiac event.

Methods

A qualitative research design was employed using purposive sampling of patients who experienced an acute cardiac event during the social containment mandates. Semi-structured interviews were conducted, with thematic analysis of interview transcripts.

Results

Twenty-five participants were recruited from three UK hospitals. Themes identified were reliance on informal support network, lack of awareness of cardiac symptoms leading to delayed help-seeking, and an indirect COVID-19 effect (e.g. avoiding treatment).

Conclusions

These results highlight the need for informed public health messages, targeting patients and their support networks, that allow those in need of treatment to access care.

Lay, professional, and artificial intelligence perspectives on risky medical decisions and COVID-19: How does the number of lives matter in clinical trials framed as gains versus losses?

Outcomes of clinical trials need to be communicated effectively to make decisions that save lives. We investigated whether framing can bias these decisions and if risk preferences shift depending on the number of patients. Hypothetical information about two medicines used in clinical trials having a sure or a risky outcome was presented in either a gain frame (people would be saved) or a loss frame (people would die). The number of patients who signed up for the clinical trials was manipulated in both frames in all the experiments. Using an unnamed disease, lay participants (experiment 1) and would-be medical professionals (experiment 2) were asked to choose which medicine they would have administered. For COVID-19, lay participants were asked which medicine should medical professionals (experiment 3), artificially intelligent software (experiment 4), and they themselves (experiment 5) favour to be administered. Broadly consistent with prospect theory, people were more risk-seeking in the loss frames than the gain frames. However, risk-aversion in gain frames was sensitive to the number of lives with risk-neutrality at low magnitudes and risk-aversion at high magnitudes. In the loss frame, participants were mostly risk-seeking. This pattern was consistent across laypersons and medical professionals, further extended to preferences for choices that medical professionals and artificial intelligence programmes should make in the context of COVID-19. These results underscore how medical decisions can be impacted by the number of lives at stake while revealing inconsistent risk preferences for clinical trials during a real pandemic.

An Iterative and Collaborative End-to-End Methodology Applied to Digital Mental Health

Artificial intelligence (AI) algorithms together with advances in data storage have recently made it possible to better characterize, predict, prevent, and treat a range of psychiatric illnesses. Amid the rapidly growing number of biological devices and the exponential accumulation of data in the mental health sector, the upcoming years are facing a need to homogenize research and development processes in academia as well as in the private sector and to centralize data into federalizing platforms. This has become even more important in light of the current global pandemic. Here, we propose an end-to-end methodology that optimizes and homogenizes digital research processes. Each step of the process is elaborated from project conception to knowledge extraction, with a focus on data analysis. The methodology is based on iterative processes, thus allowing an adaptation to the rate at which digital technologies evolve. The methodology also advocates for interdisciplinary (from mathematics to psychology) and intersectoral (from academia to the industry) collaborations to merge the gap between fundamental and applied research. We also pinpoint the ethical challenges and technical and human biases (from data recorded to the end user) associated with digital mental health. In conclusion, our work provides guidelines for upcoming digital mental health studies, which will accompany the translation of fundamental mental health research to digital technologies.

Impact of Bias in Data Collection of COVID-19 Cases

BACKGROUND/OBJECTIVE

 Prognostic factor research (PFR), prevalence of symptoms and likelihood ratio (LR) play an important role in identifying prescribing indications of useful homeopathic remedies. It involves meticulous unbiased collection and analysis of data collected during clinical practice. This paper is an attempt to identify causes of bias and suggests ways to mitigate them for improving the accuracy in prescribing for better clinical outcomes and execution of randomized controlled studies.

METHODS

 A prospective, open label, observational study was performed from April 2020 to December 2020 at two COVID Health Centers. A custom-made Excel spreadsheet containing 71 fields covering a spectrum of COVID-19 symptoms was shared with doctors for regular reporting. Cases suitable for PFR were selected. LR was calculated for commonly occurring symptoms. Outlier values with LR ≥5 were identified and variance of LRs was calculated.

RESULTS

 Out of 1,889 treated cases of confirmed COVID-19, 1,445 cases were selected for pre-specified reasons. Nine medicines, Arsenicum album, Bryonia alba, Gelsemium sempervirens, Pulsatilla nigricans, Hepar sulphuricus, Magnesia muriaticum, Phosphorus, Nux vomica and Belladonna, were most frequently prescribed. Outlier values and large variance for Hepar sulphuricus and Magnesia muriaticum were noticed as indication of bias. Confirmation bias leading to lowering of symptom threshold, keynote prescribing, and deficiency in checking of all symptoms in each case were identified as the most important sources of bias.

CONCLUSION

 Careful identification of biases and remedial steps such as training of doctors, regular monitoring of data, checking of all pre-defined symptoms, and multicenter data collection are important steps to mitigate biases.

Exploring bias in restraint use: Four strategies to mitigate bias in care of the agitated patient in the emergency department

Agitation is a routine and increasingly common presentation to the emergency department (ED). In the wake of a national examination into racism and police use of force, this article aims to extend that reflection into emergency medicine in the management of patients presenting with acute agitation. Through an overview of ethicolegal considerations in restraint use and current literature on implicit bias in medicine, this article provides a discussion on how bias may impact care of the agitated patient. Concrete strategies are offered at an individual, institutional, and health system level to help mitigate bias and improve care.

Creating a healthcare variant CYNEFIN framework to improve leadership and urgent decision-making in times of crisis

Purpose

The complex and occasionally chaotic nature of health care has been previously described in the literature, as has the broadening recognition that different management approaches are required for different types of problems rather than a “one size fits all” approach. The CYNEFIN framework from Snowden outlines a consistent cognitive approach that offers the leader and leadership team an ability to urgently apply the correct actions to a given situation. This paper proposes a variant CYNEFIN approach for healthcare.

Design/methodology/approach

Consistent and accurate decision-making within health care is the hallmark of an effective and pragmatic leader and leadership team. An awareness of how one’s cognitive biases and heuristics may adversely impact on this cognitive process is paramount, as is an understanding of the calibration between fast and slow thinking.

Findings

The authors propose a variant CYNEFIN approach for health care of “act-probe-sense-respond” to resolve complex and time-critical emergency scenarios, using the differing contexts of a cardiac arrest and an evolving crisis management problem as examples. The variant serves as a pragmatic sense-making framework for the health-care leader and leadership team that can be adopted for many time-critical crisis situations.

Originality/value

The variant serves as a pragmatic sense-making framework for the health-care leader that can be adopted for many crisis situations.

Creation of clinical algorithms for decision-making in oncology: an example with dose prescription in radiation oncology

In oncology, decision-making in individual situations is often very complex. To deal with such complexity, people tend to reduce it by relying on their initial intuition. The downside of this intuitive, subjective way of decision-making is that it is prone to cognitive and emotional biases such as overestimating the quality of its judgements or being influenced by one’s current mood. Hence, clinical predictions based on intuition often turn out to be wrong and to be outperformed by statistical predictions. Structuring and objectivizing oncological decision-making may thus overcome some of these issues and have advantages such as avoidance of unwarranted clinical practice variance or error-prevention. Even for uncertain situations with limited medical evidence available or controversies about the best treatment option, structured decision-making approaches like clinical algorithms could outperform intuitive decision-making. However, the idea of such algorithms is not to prescribe the clinician which decision to make nor to abolish medical judgement, but to support physicians in making decisions in a systematic and structured manner. An example for a use-case scenario where such an approach may be feasible is the selection of treatment dose in radiation oncology. In this paper, we will describe how a clinical algorithm for selection of a fractionation scheme for palliative irradiation of bone metastases can be created. We explain which steps in the creation process of a clinical algorithm for supporting decision-making need to be  performed and which challenges and limitations have to be considered.

Enabling a learning healthcare system with automated computer protocols that produce replicable and personalized clinician actions

Clinical decision-making is based on knowledge, expertise, and authority, with clinicians approving almost every intervention-the starting point for delivery of "All the right care, but only the right care," an unachieved healthcare quality improvement goal. Unaided clinicians suffer from human cognitive limitations and biases when decisions are based only on their training, expertise, and experience. Electronic health records (EHRs) could improve healthcare with robust decision-support tools that reduce unwarranted variation of clinician decisions and actions. Current EHRs, focused on results review, documentation, and accounting, are awkward, time-consuming, and contribute to clinician stress and burnout. Decision-support tools could reduce clinician burden and enable replicable clinician decisions and actions that personalize patient care. Most current clinical decision-support tools or aids lack detail and neither reduce burden nor enable replicable actions. Clinicians must provide subjective interpretation and missing logic, thus introducing personal biases and mindless, unwarranted, variation from evidence-based practice. Replicability occurs when different clinicians, with the same patient information and context, come to the same decision and action. We propose a feasible subset of therapeutic decision-support tools based on credible clinical outcome evidence: computer protocols leading to replicable clinician actions (eActions). eActions enable different clinicians to make consistent decisions and actions when faced with the same patient input data. eActions embrace good everyday decision-making informed by evidence, experience, EHR data, and individual patient status. eActions can reduce unwarranted variation, increase quality of clinical care and research, reduce EHR noise, and could enable a learning healthcare system.

Exome sequencing study in a clinical research setting finds general acceptance of study returning secondary genomic findings with little decisional conflict

The most appropriate strategies for managing secondary genomic findings (SF) in clinical research are being developed and evaluated. We surveyed patients at the National Institute of Allergy and Infectious Diseases (NIAID) to evaluate decisional conflict regarding enrolling in a study that returns SF. Responses were collected using a cross-sectional survey after informed consent but before return of SF. Sixty-six adults of 116 eligible participants responded. No participant explicitly declined because they did not want to possibly receive a SF. Sixty-five of 66 (98%) participants thought it was appropriate to return SFs in research; one participant was unsure. Decisional conflict regarding enrolling in a study returning SF was low overall with 68% of participants reporting a score of less than 10 on a 100-point decisional conflict scale, implying that they felt informed, clear on what they wanted, and supported. Lower genetic literacy was weakly associated with higher decisional conflict (Spearman's rho = -0.297, p = .015). Six participants reported confusion related to the choices about SFs. Our data suggest that participants in our study feel it is appropriate to receive SF and have little decisional conflict about potentially receiving such information; however, some participants may need further education and counseling.

Efficiency over thoroughness in laboratory testing decision making in primary care: findings from a realist review

Background

Existing research demonstrates significant variation in test-ordering practice, and growth in the use of laboratory tests in primary care. Reviews of interventions designed to change test-ordering practice report heterogeneity in design and effectiveness. Improving understanding of clinicians’ decision making in relation to laboratory testing is an important means of understanding practice patterns and developing theory-informed interventions.

Aim

To develop explanations for the underlying causes of patterns of variation and increasing use of laboratory tests in primary care, and make recommendations for future research and intervention design.

Design & setting

Realist review of secondary data from primary care.

Method

Diverse evidence, including data from qualitative and quantitative studies, was gathered via systematic and iterative searching processes. Data were synthesised according to realist principles to develop explanations accounting for clinicians’ decision making in relation to laboratory tests.

Results

A total of 145 documents contributed data to the synthesis. Laboratory test ordering can fulfil many roles in primary care. Decisions about tests are incorporated into practice heuristics and tests are deployed as a tool to manage patient interactions. Ordering tests may be easier than not ordering tests in existing systems. Alongside high workloads and limited time to devote to decision making, there is a common perception that laboratory tests are relatively inconsequential interventions. Clinicians prioritise efficiency over thoroughness in decision making about laboratory tests.

Conclusion

Interventions to change test-ordering practice can be understood as aiming to preserve efficiency or encourage thoroughness in decision making. Intervention designs and evaluations should consider how testing decisions are made in real-world clinical practice.

A qualitative exploration of the factors influencing the decision to transfuse elective platelets in cancer care

OBJECTIVES

To explore factors contributing to decisions to transfuse platelets in cancer care outside guidelines using case study methods.

METHODS

Two case studies were examined, using instrumental case study methodology, to qualitatively explore factors that influence the decision to transfuse platelets. Interviews (n=10) were conducted around cases in haematology and critical care. In-depth review of documentary evidence was undertaken and propositions were developed to provide rigour during the investigation. Thematic analysis and triangulation of documents was undertaken to find specific factors, and propositions analysed, as per instrumental case study methods.

RESULTS

Both cases emphasised how patient complexity, and individual response to platelet administration, was an influencing factor. Other themes included uncertainty of clinical situations coupled with uncertainty about platelet availability. Other factors worthy of further investigation include the concept of professional safety and the trustworthiness of platelet thresholds against platelet monitoring for bleeding episodes.

CONCLUSION

The findings indicated several factors influence decisions to transfuse, including clinical context, and individual ability to trust guidelines and assume any perceived risks.

Indwelling catheters increase altered mental status and urinary tract infection risk: A retrospective Cohort Study

Background

Although indwelling urinary catheters (IUCs) are used intraoperatively and may cause complications (e.g., delirium), only few robust studies have investigated the association between intraoperative IUC use and complications. We hypothesized that IUC use might increase the postoperative incidence of altered mental status and/or urinary catheter infection.

Materials and methods

In this retrospective single-center cohort study, we analyzed the data of adult patients undergoing surgery at our facility between January 2013 and December 2018. The primary endpoint was altered mental status and/or incidence of urinary catheter infections. The patients were divided into IUC and control groups. A multivariable logistic regression model was used to identify the predictors of postoperative complications, and a multivariable Cox proportional hazards regression model was used to analyze hospital discharge in unmatched and inverse propensity-weighted patients.

Results

Of the 14,284 patients that were reviewed, we analyzed 5112 patients (control group, 44.0%; IUC group, 56.0%). Almost all procedures comprised less invasive surgeries. The prevalence of postoperative altered mental status and postoperative urinary catheter infection were 3.56% and 0.04%, respectively. After inverse propensity weighting, all baseline characteristics were similar between the two groups. However, patients with IUCs had a higher risk of postoperative complications (adjusted odds ratio, 1.97; 95% confidence interval [CI], 1.50-2.59) and prolonged hospital stays (hazard ratio, 0.84; 95% CI, 0.80-0.89).

Conclusion

In patients undergoing less invasive surgery, IUCs may be associated with a relatively high risk of altered mental status or urinary catheter infection. These data may facilitate preoperative discussions regarding the perioperative use of IUCs.

Attending Surgeons Demonstrate Greater Correlations of Skill Assessment and Anticipation of Adverse Events Than Residents During Laparoscopic Cholecystectomy

BACKGROUND

Surgical training includes the development of technical and nontechnical skills. While technical skills are more easily quantified, nontechnical skills such as situation awareness (SA) are more difficult to measure and quantify. This study investigated the relationships between different SA elements and expertise.

METHODS

Twenty attending and resident surgeons rated their anticipation of an impending adverse event while watching 20 videos of laparoscopic cholecystectomies with and without adverse events. After watching each video, they assessed surgeon skills and self-assessed their anticipation ratings. All participants answered a general confidence questionnaire before and after the study.

RESULTS

Videos with adverse events led to significantly higher anticipation of adverse events (P < 0.001), lower surgeon skill rating (P < 0.001), and higher self-assessment in their anticipation ratings (P < 0.001) across both participant groups. General confidence was significantly lower for residents than that for attending surgeons (P < 0.001). Compared with the residents, attendings exhibited stronger and more stable correlations between measurements of SA. When viewing videos with adverse events, attendings showed significantly higher correlation between anticipation of an impending adverse event and skill assessment of the surgeon (P = 0.005).

CONCLUSIONS

This study investigated how different elements of SA and their relationships were influenced by experience. The results indicated that attendings had stronger and more stable correlations between SA elements than residents, demonstrating how measurement correlations could be meaningful and sensitive indicators of expertise and autonomy readiness.

Parental preferences for vesicoureteral reflux treatment: Profile case best-worst scaling

INTRODUCTION

Vesicoureteral reflux is a common pediatric urologic condition that often has several reasonable treatment options depending on condition severity. In order to choose the best treatment for their child, parents are expected to make judgements that weigh attributes such as treatment cost, effectiveness, and complication rate. Prior research has shown that factors such as treating hospital and surgeon also influence patient treatment choice.

OBJECTIVES

This study evaluates parental preferences for reflux treatment using profile case best-worst scaling, an emerging technique in both urologic and health care preference estimation. The study also uses latent class analysis (LCA) to identify parental sub-classes with different preferences.

STUDY DESIGN

Data were collected from a community sample of parents via a multimedia best-worst scaling survey instrument published to Amazon's Mechanical Turk online community. After extensive review of the literature, reflux attributes and attribute levels were selected to correspond with available treatments. The profile case best-worst scaling exercise elicited preferences for granular attributes of reflux treatments. Data were analyzed using multinomial logistic regression and class analysis to distinguish preference heterogeneity. Probability scaled values (PSVs) reflected the order of desirability of the attributes. Attribute preference importance was rescaled into dollar units for comparison as well.

RESULTS

We analyzed data for 248 respondents. The highest treatment effectiveness was more desirable than all other leveled treatment attributes (PSV 17.8, all p < 0.01) (Table). Low complication rate and doctor recommendation were amongst the other most desirable treatment attributes (PSV 11.3 and 9.0, respectively). Latent class analysis identified a class with more extreme preferences, for whom doctor recommendation and avoiding hospitalization were particularly desirable.

DISCUSSION

In this community-based sample, high treatment effectiveness and low complication rate were the most desirable treatment attributes to parents, though parents likely have heterogenous treatment preference structures. Shared parent-physician decision-making that incorporates parental preferences will likely allow more effective, targeted decision-making in the future.

Cognitions and behaviours of general practitioners in France regarding HPV vaccination: A theory-based systematic review

Human papillomavirus (HPV) vaccination is safe and efficacious to prevent persistent HPV infection, precancerous anogenital lesions and cervical cancer. However, in countries where vaccination programmes are implemented outside of schools, such as France, reaching high HPV vaccination coverage of the target population is challenging. Many studies have been performed in France to assess cognitions of general practitioners (GPs) regarding HPV vaccination. However, the evidence is not consistent about which cognitions are key. To provide a comprehensive overview, we performed a systematic review of studies conducted in France on GPs' cognitions regarding HPV vaccination and used the reasoned action approach to extract and synthesize data. The systematic search was performed up to July 2020 in Medline via PubMed, PsycINFO, PsycARTICLES, Embase, CINAHL Plus, Web of Science, Pascal and Francis databases. Grey literature was searched for in the French Public Health Database, Cairn. Info, yahoo.fr, and Google Scholar. Twenty-five scientific publications were selected based on eligibility criteria and assessed for quality. Our qualitative synthesis highlights that although 73% of GPs report recommending HPV vaccination, up to 50% would not recommend it because of concerns, including changes in patients' health behaviours and doubts about safety and/or efficacy. GPs' injunctive norms, i.e. trust in institutional information, were shown to be associated positively with GPs' willingness to recommend HPV vaccination. Parents' fears, girls' age, and potential connection with sexuality do not seem to affect GPs' recommendations. These results will inform the development of a professional educational intervention targeting GPs in France.

Beliefs and Risk Perceptions About COVID-19: Evidence From Two Successive French Representative Surveys During Lockdown

BACKGROUND

The outbreak of COVID-19 has been a major interrupting event, challenging how societies and individuals deal with risk. An essential determinant of the virus' spread is a series of individual decisions, such as wearing face masks in public space. Those decisions depend on trade-offs between costs (or benefits) and risks, and beliefs are key to explain these.

METHODS

We elicit beliefs about the COVID-19 pandemic during lockdown in France by means of surveys asking French citizens about their belief of the infection fatality ratio (IFR) for COVID-19, own risk to catch the disease, risk as perceived by others, and expected prevalence rate. Those self-assessments were measured twice during lockdown: about 2 weeks after lockdown started and about 2 weeks before lockdown ended. We also measured the quality of these beliefs with respect to available evidence at the time of the surveys, allowing us to assess the calibration of beliefs based on risk-related socio-demographics. Finally, comparing own risk to expected prevalence rates in the two successive surveys provides a dynamic view of comparative optimism with respect to the disease.

RESULTS

The risk perceptions are rather high in absolute terms and they increased between the two surveys. We found no evidence for an impact of personal experience with COVID-19 on beliefs and lower risk perceptions of the IFR when someone in the respondent's family has been diagnosed with a disease. Answers to survey 1 confirmed this pattern with a clear indication that respondents were optimistic about their chances to catch COVID-19. However, in survey 2, respondents revealed comparative pessimism.

CONCLUSION

The results show that respondents overestimated the probabilities to catch or die from COVID-19, which is not unusual and does not necessarily reflect a strong deviation from rational behavior. While a rational model explains why the own risk to catch COVID-19 rose between the two surveys, it does not explain why the subjective assessment of the IFR remained stable. The comparative pessimism in survey 2 was likely due to a concomitant increase in the respondents' perceived chances to catch the disease and a decreased expected prevalence rate.

Nudging clinicians: A systematic scoping review of the literature

BACKGROUND

While the quality of medical care delivered by physicians can be very good, it can also be inconsistent and feature behaviours that are entrenched despite updated information and evidence. The "nudge" paradigm for behaviour change is being used to bring clinical practice in line with desired standards. The premise is that behaviour can be voluntarily shifted by making particular choices instinctively appealing. We reviewed studies that are explicit about their use of nudge theory in influencing clinician behaviour.

METHODS

Databases were searched from April 2008 (the publication date of the book that introduced nudge theory to a wider audience) to November 2018, inclusive. The search strategy and narrative review of results addressed the Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews.

RESULTS

22 studies were identified. Randomized trials or pre-post comparisons were generally used in community-based settings; single-site pre-post studies were favoured in hospitals. The studies employed eight intervention types: active choice; patient chart redesign; default and default alerts; partitioning of prescription menus; audit and feedback; commitment messages; peer comparisons; and redirection of workflow. Three core cognitive factors underpinned the eight interventions: bias towards prominent choices (salience); predisposition to social norms; and bias towards time or cost savings.

CONCLUSIONS

Published studies that are explicit about their use of nudge theory are few in number and diverse in their settings, targets, and results. Default and chart re-design interventions reported the most substantial improvements in adherence to evidence and guideline-based practice. Studies that are explicit in their use of nudge theory address the widespread failure of clinical practice studies to identify theoretical frameworks for interventions. However, few studies identified in our review engaged in research to understand the contextual and site-specific barriers to a desired behaviour before designing a nudge intervention.

Integrating infection and sepsis management through holistic early warning systems and heuristic approaches: a concept proposal

This is a first attempt to integrate the three pillars of infection management: the infection prevention and control (IPC), and surveillance (IPCS), antimicrobial stewardship (AMS), and rapid identification and management of sepsis (RIMS). The new 'Sepsis-3' definition extrapolates the diagnosis of sepsis from our previously slightly naïve concept of a stepwise evolving pattern. In doing so, however, we have placed the transition from infection toward sepsis in the domain of uncertainty and time-dependency. This now demands that clinical judgment be used in the risk stratification of patients with infection, and that pragmatic local solutions be used to prompt clinicians to evaluate formally for sepsis. We feel it is necessary to stimulate the development of a new generation of concepts and models aiming at embracing uncertainty. We see the opportunity for a heuristic approach focusing on the relevant clinical predictors at hand allowing to navigate the uncertainty of infection diagnosis under time constraints. The diverse and situated clinical approaches eventually emerging need to focus on the understanding of infection as the unbalanced interactions of host, pathogen, and environment. In order extend such approach throughout the patient journey we propose a holistic early warning system underpinned by the risk-based categories of hazards and vulnerabilities iteratively fostered by the information gathered by the infection prevention control and surveillance, clinical microbiology, and clinical chemistry services.

Data Science-Based Analysis of Patient Subgroup Structures Suggest Effects of Rhinitis on All Chemosensory Perceptions in the Upper Airways

Viral rhinitis contributes significantly to olfactory dysfunction, but it is unclear how many patients have other chemosensory symptoms in addition to olfactory loss. This was addressed in the present reanalysis of data previously published in Pellegrino R, Walliczek-Dworschak U, Winter G, Hull D, Hummel T. 2017. Investigation of chemosensitivity during and after an acute cold. Int Forum Allergy Rhinol. 7(2):185-191, using unsupervised and supervised machine-learning methods. Fifty-eight patients with acute rhinitis and 59 healthy controls were assessed for orthonasal and retronasal olfactory function, taste, and intranasal trigeminal sensitivity. Unsupervised analysis showed that during rhinitis, clinical scores of olfactory function, expressed as threshold, discrimination, identification (TDI) values, were trimodally distributed. Two minor modes were separated from the main mode at TDI = 30.5, which corresponds to the established limit of hyposmia. This trimodal distribution was not observed after the rhinitis subsided. Olfactory function was not significantly impaired in 40% of all rhinitis patients, whereas it was transiently impaired in 59%. For this group, supervised machine-learning algorithms could be trained with information on retronasal olfactory function, gustatory function, and trigeminal sensitivity to assign patients to subgroups based on orthonasal olfactory function with a balanced classification accuracy of 64-65%. The ability to recognize patients with olfactory loss based on retronasal olfactory function as well as gustatory function and trigeminal sensitivity suggests in turn that these modalities are affected by rhinitis. However, the only modest accuracy at which this information allowed to reproduce the olfactory diagnosis indicated they are involved in the symptomatology of rhinitis to a lesser extent compared with the orthonasal olfactory function.

If in Doubt Don't Act Out! Exploring Behaviours in Clinical Decision Making by General Surgeons Towards Surgical Procedures

BACKGROUND

Clinical decision-making (CDM) plays an integral role to surgeons work and has ramifications for patient outcomes and experience. The factors influencing a surgeons decision-making and the utility of cognitive decisional short cuts used in CDM known as 'heuristics' remains unknown. The aim of this paper is to explore how general surgeons make decisions in high-stake biliary tract clinical scenarios.

METHODS

This was a cross sectional survey comprising of two sections-a 'demographics section' and a 'clinical vignettes section'. Participants were recruited by an email distributed by the Royal College of Surgeons in Ireland. Non-parametric testing examined relationships and content analysis was applied for clinical reasoning.

RESULTS

73 participants or 37.6% of the overall population completed the survey. 71.4% of these were male. Most (50%) were higher trainees with moderate levels of overall reflective practice in decision-making. A majority of participants chose conservatively in high-stake biliary tract clinical cases with disease factors (43.5%) weighted highest, followed by personal factors (41.1%) and patient factors (15.4%) in clinical reasoning. The presence of a 'hook' associated with commonly used heuristics did not significantly change decision-making behaviour.

CONCLUSION

In high-stake scenarios, surgeons make conservative clinical decisions, predominantly dominated by disease and personal justifications. The utility of heuristics in lower-stake scenarios should be explored regarding clinical decision-making rationale and outcomes. Practitioners should consider use of patient factors in high-stake decisions to enable shared decision-making when appropriate which can reduce post-decisional regret and support the vision of patient-centred care.

Bias in Medicine: Lessons Learned and Mitigation Strategies

Cognitive bias consists of systematic errors in thinking due to human processing limitations or inappropriate mental models. Cognitive bias occurs when intuitive thinking is used to reach conclusions about information rather than analytic (mindful) thinking. Scientific progress is delayed when bias influences the dissemination of new scientific knowledge, as it has with the role of human leucocyte antigen antibodies and antibody-mediated rejection in cardiac transplantation. Mitigating strategies can be successful but involve concerted action by investigators, peer reviewers, and editors to consider how we think as well as what we think.

Availability Bias Causes Misdiagnoses by Physicians: Direct Evidence from a Randomized Controlled Trial

Objective Empirical evidence on the availability bias associated with diagnostic errors is still insufficient. We investigated whether or not recent experience with clinical problems can lead physicians to make diagnostic errors due to availability bias and whether or not reflection counteracts this bias. Methods Forty-six internal medicine residents were randomly divided into a control group (CG) and experimental group (EG). Among the eight clinical cases used in this study, three experimental cases were similar to the disease of dengue fever (DF) but exhibited different diagnoses, one was actually DF, and the other four filler cases were not associated with DF. First, only the EG received information on DF, while the CG knew nothing about this study. Then, six hours later, all participants were asked to diagnose eight clinical cases via nonanalytic reasoning. Finally, four cases were diagnosed again via reflective reasoning. Results In stage 2, the average score of the CG in the diagnosis of experimental cases was significantly higher than that of the filler cases (0.80 vs. 0.59, p<0.01), but the EG's average score in the two types of cases was not significantly different (0.66 vs. 0.64, p=0.756). The EG and CG had significantly different scores for each experimental case, while no difference was observed in the filler cases. The proportion of diseases incorrectly diagnosed as DF among experimental cases ranged from 71% to 100% in the EG. There were no significant differences between the mean diagnostic accuracy scores obtained by nonanalytic reasoning and those obtained by the reflective reasoning in any cases. Conclusion Availability bias led to diagnostic errors. Misdiagnoses cannot always be repaired solely by adopting a reflective approach.

Why We Fail to Rescue During Critical Events

Many factors come together probabilistically to affect clinician response to critical events in the operating room; no 2 critical events are alike. These factors involve 4 primary domains: (1) the event itself, (2) the individual anesthetist(s), (3) the operating room team, and (4) the resources available and environments in which the event occurs. Appreciating these factors, anticipating how they create vulnerabilities for error and poor response, and actively addressing those vulnerabilities (before events occur as well as during) will help clinicians manage critical event response more effectively and avoid errors.

Patterns of Service Provision in Older People's Mental Health Care in Australia

Australia has a population of around 4 million people aged 65 years and over, many of whom are at risk of developing cognitive decline, mental illness, and/or psychological problems associated with physical illnesses. The aim of this study was to describe the pattern of specialised mental healthcare provision (availability, placement capacity, balance of care and diversity) for this age group in urban and rural health districts in Australia. The Description and Evaluation of Services and DirectoriEs for Long Term Care (DESDE-LTC) tool was used in nine urban and two rural health districts of the thirty-one Primary Health Networks across Australia. For the most part service provision was limited to hospital and outpatient care across all study areas. The latter was mainly restricted to health-related outpatient care, and there was a relative lack of social outpatient care. While both acute and non-acute hospital care were available in urban areas, in rural areas hospital care was limited to acute care. Limited access to comprehensive mental health care, and the uniformity in provision across areas in spite of differences in demographic, socioeconomic and health characteristics raises issues of equity in regard to psychogeriatric care in this country. Comparing patterns of mental health service provision across the age span using the same classification method allows for a better understanding of care provision and gap analysis for evidence-informed policy.

Gender Bias in Collaborative Medical Decision Making: Emergent Evidence

This initial, exploratory study on gender bias in collaborative medical decision making examined the degree to which physicians' reliance on a team member's patient care advice differs as a function of the gender of the advice giver. In 2018, 283 anesthesiologists read a brief, online clinical vignette and were randomly assigned to receive treatment advice from 1 of 8 possible sources (physician or nurse, man or woman, experienced or inexperienced). They then indicated their treatment decision, as well as the degree to which they relied upon the advice given.The results revealed 2 patterns consistent with gender bias in participants' advice taking. First, when treatment advice was delivered by an inexperienced physician, participants reported replying significantly more on the advice of a man versus a woman, F(1,61) = 4.24, P = .04. Second, participants' reliance on the advice of the woman physician was a function of her experience, F(1,62) = 6.96, P = .01, whereas reliance on the advice of the man physician was not, F(1,60) = 0.21, P = .65.These findings suggest women physicians, relative to men, may encounter additional hurdles to performing their jobs, especially at early stages in their careers. These hurdles are rooted in psychological biases of others, rather than objective features of cases or treatment settings. Cultural stereotypes may shape physicians' information use and decision-making processes (and hinder collaboration), even in contexts that appear to have little to do with social category membership. The authors recommend institutions adopt policies and practices encouraging equal attention to advice, regardless of the source, to help ensure advice taking is a function of information quality rather than the attributes of the advice giver. Such policies and practices may help surface and implement diverse expert perspectives in collaborative medical decision making, promoting better and more effective patient care.

Understanding the role of GPs' gut feelings in diagnosing cancer in primary care: a systematic review and meta-analysis of existing evidence

BACKGROUND

Growing evidence for the role of GPs' gut feelings in cancer diagnosis raises questions about their origin and role in clinical practice.

AIM

To explore the origins of GPs' gut feelings for cancer, their use, and their diagnostic utility.

DESIGN AND SETTING

Systematic review and meta-analysis of international research on GPs' gut feelings in primary care.

METHOD

Six databases were searched from inception to July 2019, and internet searches were conducted. A segregated method was used to analyse, then combine, quantitative and qualitative findings.

RESULTS

Twelve articles and four online resources were included that described varied conceptualisations of gut feelings. Gut feelings were often initially associated with patients being unwell, rather than with a suspicion of cancer, and were commonly experienced in response to symptoms and non-verbal cues. The pooled odds of a cancer diagnosis were four times higher when gut feelings were recorded (OR 4.24, 95% confidence interval = 2.26 to 7.94); they became more predictive of cancer as clinical experience and familiarity with the patient increased. Despite being included in some clinical guidelines, GPs had varying experiences of acting on gut feelings as some specialists questioned their diagnostic value. Consequently, some GPs ignored or omitted gut feelings from referral letters, or chose investigations that did not require specialist approval.

CONCLUSION

GPs' gut feelings for cancer were conceptualised as a rapid summing up of multiple verbal and non-verbal patient cues in the context of the GPs' clinical knowledge and experience. Triggers of gut feelings not included in referral guidance deserve further investigation as predictors of cancer. Non-verbal cues that trigger gut feelings appear to be reliant on continuity of care and clinical experience; they tend to remain poorly recorded and are, therefore, inaccessible to researchers.

What is optimal in optimal inference?

Inferring hidden structure from noisy observations is a problem addressed by Bayesian statistical learning, which aims to identify optimal models of the process that generated the observations given assumptions that constrain the space of potential solutions. Animals and machines face similar "model-selection" problems to infer latent properties and predict future states of the world. Here we review recent attempts to explain how intelligent agents address these challenges and how their solutions relate to Bayesian principles. We focus on how constraints on available information and resources affect inference and propose a general framework that uses benefit(accuracy) and accuracy(cost) curves to assess optimality under these constraints.

Diffusion-weighted MRI in transient global amnesia and its diagnostic implications

OBJECTIVE

To analyze how the evidence of hippocampal diffusion-weighted imaging (DWI) lesions may support the clinical diagnosis of transient global amnesia (TGA).

METHODS

In this retrospective observational study, 390 consecutive patients with isolated TGA were analyzed, who were evaluated at our institution between July 1999 and August 2018. The size, location, and number of lesions and time-dependent lesion detectability were examined. The incidence of DWI lesions was reviewed with regard to different levels of clinical diagnostic certainty upon presentation to the emergency department.

RESULTS

Hippocampal DWI lesions were detected in 272 (70.6%) patients with TGA, with a mean of 1.05 ± 0.98 (range 0-6) and a mean lesion size of 4.01 ± 1.22 mm (range 1.7-8.6 mm). In the subgroups of lower diagnostic certainty (amnesia witnessed by layperson or self-reported amnestic gap), DWI was helpful in supporting the diagnosis of TGA in 76 (69.1%) patients. In 187 patients with information about the exact onset, DWI lesions were analyzed in relation to latency between onset and MRI. Lesions could be detected at all time points and up to 6 days after symptom onset in individual patients; the highest rate of DWI-positive MRI (93%) was in the 12-24 hours time window.

CONCLUSION

MRI findings can support the diagnosis of TGA and may be particularly valuable in situations of low clinical certainty. DWI-ideally performed with a minimum delay of 20 hours after onset-should therefore be considered a useful adjunct to the diagnosis of TGA.

The Use of Common Knowledge in Fuzzy Logic Approach for Vineyard Site Selection

A multitude of factors considered necessary for an informed choice of the location of the vineyard can be overwhelming for the decision-maker. Is there still a place for knowledge valuable from the perspective of an experienced winegrower in the era of precise measurements? The informative use of so-called common knowledge is possible owing to fuzzy-based techniques, which allow for the representation of intuitive notions in terms of quantitative measures. The work uses tools based on fuzzy logic to cover the scope of common knowledge within the decision-making process. Owing to its flexibility and ability to deal with imprecise input data while maintaining the simple construction, the fuzzy logic solution filled the gap between GIS data and wine grower’s experience. Based on the data from the thematic literature, a set of rules was created to interpret the relationships between popular site selection criteria. The dynamics and manner of interaction between variables were determined using adequate membership functions. Pre-processing using GIS with remote sensing data was considered as a preliminary stage for the analysis. By using the graphical interface, the system operation facilitates the work of a potential user. The obtained results indicated the possibility of an alternative approach to classical analyses by replacing or extending the meaning of some variables using information based on feelings and perceptions. Research constitutes a premise for the further development of expert systems using widely understood domain knowledge.

Respected as a Client, Cared for as a Patient: Evidence of Heuristic Decision-Making from Yelp Reviews of Obstetrician-Gynecologists

Previous literature has demonstrated a gap between what health care consumers say they want to know about their physicians and the publicly available information the health care industry provides. This systematic analysis of Yelp reviews from the 25 most populous U.S. metropolitan areas is predicated on the assumptions that patients who post online physician reviews include information they would find useful when choosing a physician, and that this information represents an ecologically valid sample for making inferences regarding patients' decision making process. Obstetrician gynecologists are twice as likely to be reviewed as other physicians, but this is, to our knowledge, the first study examining online reviews of obstetrician-gynecologists specifically. This study contributes to the literature on medical decision making, demonstrating that the physician choice decision is made using a heuristic tallying model, in which only two, nearly equally weighted parameters are meaningful: patient (which, within the reviews, incorporates physician interpersonal manner and physician knowledge and skills) and office management. Yelp reviews of obstetrician-gynecologists follow patterns previously established in scholarship: ratings are bimodal, and approximately two-thirds are positive. In the absence of objective, user-friendly physician-level information, patients will turn to review sites like Yelp when choosing physicians. Rather than resisting this trend, physicians and hospital systems would be better served by working with review sites, or creating their own, to emphasize aspects of the experience patients are qualified to evaluate: the physician's interpersonal manner and office management.

Person- and job-specific factors of intuitive decision-making in clinical practice: results of a sample survey among Hungarian physicians and nurses

Objective: To assess the prevalence of intuitive decision-making (IDM) among health care practitioners (HCPs) and explore its person- and job-specific factors. Design and Outcome Measures: We used on-line survey data from a cross-sectional sample of Hungarian physicians and nurses (N = 460) to assess their reliance on IDM. In a second survey we asked physicians (N = 104) to rate medical specialties on dimensions of 'emergency' (necessity of making instantaneous decisions in unforeseeable situations) and 'complexity' (necessity of considering multiple perceptual and diagnostic aspects of patients' health condition along with diverse treatment options). Results: Altogether 40% of participants reported ever relying on IDM. Using logistic regression analysis, we found the estimated probability of IDM was 0.24 greater for physicians than for nurses, 0.10 greater for females than for males, and 0.11 greater for advanced level HCPs than for novices. Reaching expert level further increased (by 0.31) the probability of IDM for physicians, but not for nurses. Concerning physicians, practicing in a medical specialty of 'high likelihood of emergency' or 'high complexity' increased the probability of IDM by 0.25 and 0.23; the same effects for nurses were 0.20 and 0.07. We found some (inconclusive) evidence for education positively influencing HCPs' propensity for IDM. Additionally, we performed content analysis of participants' free-text answers to explore the psychological background of IDM instances. HCPs educated in the subject of IDM were found more disposed to perform or request further medical investigation, less prone to deviate from medical protocols, apter to reflect on their mental processes, and more inclined to rely on a large scope of information for their decisions. Conclusions: The associations between job- and person-specific factors and HCPs' propensity for IDM may have implications for their training and allocation in the health care system. Education has great potential for enhancing the quality of IDM in clinical practice.

Strategies in Abduction: Generating and Selecting Diagnostic Hypotheses

We distinguish three aspects of medical diagnosis: generating new diagnostic hypotheses, selecting hypotheses for further pursuit, and evaluating their probability in light of the available evidence. Drawing on Peirce's account of abduction, we argue that hypothesis generation is amenable to normative analysis: physicians need to make good decisions about when and how to generate new diagnostic hypothesis as well as when to stop. The intertwining relationship between the generation and selection of diagnostic hypotheses is illustrated through the analysis of a detailed clinical case study. This interaction is not adequately captured by the existing probabilistic, decision-theoretic models of the threshold approach to clinical decision-making. Instead, we propose to conceptualize medical diagnosis in terms of strategic reasoning.

Identifying Acute Low Back Pain Episodes in Primary Care Practice From Clinical Notes: Observational Study

BACKGROUND

Acute and chronic low back pain (LBP) are different conditions with different treatments. However, they are coded in electronic health records with the same International Classification of Diseases, 10th revision (ICD-10) code (M54.5) and can be differentiated only by retrospective chart reviews. This prevents an efficient definition of data-driven guidelines for billing and therapy recommendations, such as return-to-work options.

OBJECTIVE

The objective of this study was to evaluate the feasibility of automatically distinguishing acute LBP episodes by analyzing free-text clinical notes.

METHODS

We used a dataset of 17,409 clinical notes from different primary care practices; of these, 891 documents were manually annotated as acute LBP and 2973 were generally associated with LBP via the recorded ICD-10 code. We compared different supervised and unsupervised strategies for automated identification: keyword search, topic modeling, logistic regression with bag of n-grams and manual features, and deep learning (a convolutional neural network-based architecture [ConvNet]). We trained the supervised models using either manual annotations or ICD-10 codes as positive labels.

RESULTS

ConvNet trained using manual annotations obtained the best results with an area under the receiver operating characteristic curve of 0.98 and an F score of 0.70. ConvNet's results were also robust to reduction of the number of manually annotated documents. In the absence of manual annotations, topic models performed better than methods trained using ICD-10 codes, which were unsatisfactory for identifying LBP acuity.

CONCLUSIONS

This study uses clinical notes to delineate a potential path toward systematic learning of therapeutic strategies, billing guidelines, and management options for acute LBP at the point of care.

Machine-learned identification of psychological subgroups with relation to pain interference in patients after breast cancer treatments

Background

Persistent pain in breast cancer survivors is common. Psychological and sleep-related factors modulate perception, interpretation and coping with pain and may contribute to the clinical phenotype. The present analysis pursued the hypothesis that breast cancer survivors form subgroups, based on psychological and sleep-related parameters that are relevant to the impact of pain on the patients’ life.

Methods

We analysed 337 women treated for breast cancer, in whom psychological and sleep-related parameters as well as parameters related to pain intensity and interference had been acquired. Data were analysed by using supervised and unsupervised machine-learning techniques (i) to detect patient subgroups based on the pattern of psychological or sleep-related parameters, (ii) to interpret the detected cluster structure and (iii) to relate this data structure to pain interference and impact on life.

Results

Artificial intelligence-based detection of data structure, implemented as self-organizing neuronal maps, identified two different clusters of patients. A smaller cluster (11.5% of the patients) had comparatively lower resilience, more depressive symptoms and lower extraversion than the other patients. In these patients, life-satisfaction, mood, and life in general were comparatively more impeded by persistent pain.

Conclusions

The results support the initial hypothesis that psychological and sleep-related parameter patterns are meaningful for subgrouping patients with respect to how persistent pain after breast cancer treatments interferes with their life. This indicates that management of pain should address more complex features than just pain intensity. Artificial intelligence is a useful tool in the identification of subgroups of patients based on psychological factors.

•

Pain interference with daily life is relevant for post-surgery breast cancer patients.

•

11.5% of 337 patients belonged to a subgroup with high life interference of pain.

•

Life interference was not a function of pain intensity.

•

Risk factors were lower resilience, more depressive symptoms and lower extraversion.

•

Life-satisfaction, mood, and life in general were impeded by persistent pain.