We call attention to an important, but overlooked finding in research reported by Longoni, Bonezzi and Morewedge (2019). Longoni et al. claim that people always prefer a human to an artificially intelligent (AI) medical provider. We show that this was only the case when the historical performance of the human and AI providers was equal.  When the AI is known to outperform the human, their data showed a clear preference for the automated provider. We provide additional statistical analyses of their data to support this claim.

Background: The Domain-Specific Risk Taking scale (DOSPERT) is a widely used instrument that measures perceived risk and benefit and attitude toward risk for activities in several domains, but does not include medical risks.

Objective: To develop a medical risk domain subscale for DOSPERT.

Methods: Sixteen candidate risk items were developed through expert discussion. We conducted cognitive telephone interviews, an online survey, and a random-digit dialing (RDD) telephone survey to reduce and refine the scale, explore its factor structure, and obtain estimates of reliability.

Participants: Eight patients recruited from UIC medical center waiting rooms participated in 45-60 minute cognitive interviews. Thirty Amazon Mechanical Turk workers completed the online survey. One hundred Chicago-area residents completed the RDD telephone survey.

Results: On the basis of cognitive interviews, we eliminated five items due to poor variance or participant misunderstanding. The online survey suggested that two additional items were negatively correlated with the scale, and we considered them candidates for removal. Factor analysis of the responses in the RDD telephone survey and non-statistical factors led us to recommend a final set of 6 items to represent the medical risk domain. The final set of items included blood donation, kidney donation, daily medication use for allergies, knee replacement surgery, general anesthesia in dentistry, and clinical trial participation. The interitem reliability (Cronbach’s α) of the final set of 6 items ranged from 0.57-0.59 depending on the response task. Older respondents gave lower overall ratings of expected benefit from the activities.

Conclusion: We refined a set of items to measure risk and benefit perceptions for medical activities. Our next step will be to add these items to the complete DOSPERT scale, confirm the scale’s psychometric properties, determine whether medical risks constitute a psychologically distinct domain from other risky activities, and characterize individual differences in medical risk attitudes.

Research endeavors to determine the effectiveness of patient decision aids (PtDAs) have yielded mixed results. The conflicting evaluations are largely due to the different metrics used to assess the validity of judgments made using PtDAs. The different approaches can be characterized by Hammond’s (1996) two frameworks for evaluating judgments: correspondence and coherence. This paper reviews the literature on the effectiveness of PtDAs and recasts this argument as a renewed debate between these two meta-theories of judgment. Evaluation by correspondence criteria involves measuring the impact of patient decision aids on metrics for which there are objective, external, and empirically justifiable values. Evaluation on coherence criteria involves assessing the degree to which decisions follow the logical implications of internal, possibly subjective, value systems/preferences. Coherence can exist absent of correspondence and vice versa. Therefore, many of the seemingly conflicting results regarding the effectiveness of PtDAs can be reconciled by considering that the two meta-theories contribute unique perspectives. We argue that one approach cannot substitute for the other, and researchers should not deny the value of either approach. Furthermore, we suggest that future research evaluating PtDAs include both correspondence and coherence criteria.

In preventative health decisions, such as the decision to undergo an invasive screening test or treatment, people may be deterred from selecting the test because its perceived disutility relative to not testing is greater than the utility associated with prevention of possible disease. The prospect theory editing operation, by which a decision maker’s reference point is determined, can have important effects on the disutility of the test. On the basis of the prospect theory value function, this paper develops two approaches to reducing disutility by directing the decision maker’s attention to either (actual) past or (expected) future losses that result in shifted reference points. After providing a graphical description of the approaches and a mathematical proof of the direction of their effect on judgment, we briefly illustrate the potential value of these approaches with examples from qualitative research on prostate cancer treatment decisions.

Many controversies in medical science can be framed as tension between a coherence approach (which seeks logic and explanation) and a correspondence approach (which emphasizes empirical correctness). In many instances, a coherence-based theory leads to an understanding of disease that is not supported by empirical evidence. Physicians and patients alike tend to favor the coherence approach even in the face of strong, contradictory correspondence evidence. Examples include the management of atrial fibrillation, treatment of acute bronchitis, and the use of Vitamin E to prevent heart disease. Despite the frequent occurrence of controversy stemming from coherence-correspondence conflicts, medical professionals are generally unaware of these terms and the philosophical traditions that underlie them. Learning about the coherence-correspondence distinction and using the best of both approaches could not only help reconcile controversy but also lead to striking advances in medical science.

This paper asks whether moral preferences in eight medical dilemmas change as a function of how preferences are expressed, and how people choose when they are faced with two equally attractive help projects. In two large-scale studies, participants first read dilemmas where they “matched” two suggested helping projects (which varied on a single attribute) so that they became equally attractive. They did this by filling in a missing number (e.g., how many male patients must Project M save in order to be equally attractive as Project F which can save 100 female patients). Later, the same participants were asked to choose between the two equally attractive projects. We found robust evidence that people do not choose randomly, but instead tend to choose projects that help female (vs. male), children (vs. adult), innocent (vs. non-innocent), ingroup (vs. outgroup) and existing (vs. future) patients, and imply no (vs. some) risk of a harmful side-effect, even when these projects have been matched as equally attractive as, and save fewer patients than the contrasting project. We also found that some moral preferences are hidden when expressed with matching but apparent when expressed with forced choice. For example, 88–95% of the participants expressed that female and male patients are equally valuable when doing the matching task, but over 80% of them helped female patients in the choice task.

We clarify two points made in our commentary (Pezzo & Beckstead, 2020, this issue) on a recent paper by Longoni, Bonezzi, and Morewedge (2019). In both Experiments 1 and 4 from their paper, it is not possible to determine whether accuracy can compensate for algorithm aversion. Experiments 3A-C, however, do show a strong effect of accuracy such that AI that is superior to a human provider is embraced by patients. Many papers, including Longoni et al. tend to minimize the role of this compensatory process, apparently because it seems obvious to the authors (Longoni, Bonezzi, Morewedge, 2020, this issue). Such minimization, however, can lead to (mis)citations in which research that clearly demonstrates a compensatory role of AI accuracy is cited as non-compensatory.

When survey respondents rate the quality of life (QoL) associated with a health condition, they must not only evaluate the health condition itself, but must also interpret the meaning of the rating scale in order to assign a specific value. The way that respondents approach this task depends on subjective interpretations, resulting in inconsistent results across populations and tasks. In particular, patients and non-patients often give very different ratings to health conditions, a discrepancy that raises questions about the objectivity of either groups’ evaluations. In this study, we found that the perspective of the raters (i.e., their own current health relative to the health conditions they rated) influences the way they distinguish between different health states that vary in severity. Consistent with prospect theory, a mild and a severe lung disease scenario were rated quite differently by lung disease patients whose own health falls between the two scenarios, whereas healthy non-patients, whose own health was better than both scenarios, rated the two scenarios as much more similar. In addition, we found that the context of the rating task influences the way participants distinguish between mild and severe scenarios. Both patients and non-patients gave less distinct ratings to the two scenarios when each were presented in isolation than when they were presented alongside other scenarios that provided contextual information about the possible range of severity for lung disease. These results raise continuing concerns about the reliability and validity of subjective QoL ratings, as these ratings are highly sensitive to differences between respondent groups and the particulars of the rating task.

Decision makers have been found to bias their interpretation of incoming information to support an emerging judgment (predecisional information distortion). This is a robust finding in human judgment, and was recently also established and measured in physicians’ diagnostic judgments (Kostopoulou et al. 2012). The two studies reported here extend this work by addressing the constituent modes of distortion in physicians. Specifically, we studied whether and to what extent physicians distort information to strengthen their leading diagnosis and/or to weaken a competing diagnosis. We used the “stepwise evolution of preference” method with three clinical scenarios, and measured distortion on separate rating scales, one for each of the two competing diagnoses per scenario.

In Study 1, distortion in an experimental group was measured against the responses of a separate control group. In Study 2, distortion in a new experimental group was measured against participants’ own, personal responses provided under control conditions, with the two response conditions separated by a month. The two studies produced consistent results. On average, we found considerable distortion of information to weaken the trailing diagnosis but little distortion to strengthen the leading diagnosis. We also found individual differences in the tendency to engage in either mode of distortion. Given that two recent studies found both modes of distortion in lay preference (Blanchard, Carlson & Meloy, 2014; DeKay, Miller, Schley & Erford, 2014), we suggest that predecisional information distortion is affected by participant and task characteristics. Our findings contribute to the growing research on the different modes of predecisional distortion and their stability to methodological variation.

While the medical ethics literature has well explored the harm to patients, families, and the integrity of the profession in failing to disclose medical errors once they occur, less often addressed are the moral and professional obligations to take all available steps to prevent errors and harm in the first instance. As an expanding body of scholarship further elucidates the causes of medical error, including the considerable extent to which medical errors, particularly in diagnostics, may be attributable to cognitive sources, insufficient progress in systematically evaluating and implementing suggested strategies for improving critical thinking skills and medical judgment is of mounting concern. Continued failure to address pervasive thinking errors in medical decisionmaking imperils patient safety and professionalism, as well as beneficence and nonmaleficence, fairness and justice. We maintain that self-reflective and metacognitive refinement of critical thinking should not be construed as optional but rather should be considered an integral part of medical education, a codified tenet of professionalism, and by extension, a moral and professional duty.

Background: Patient involvement is widely acknowledged to be a valuable component in health technology assessment (HTA) and healthcare decision making. However, quantitative approaches to ascertain patients' preferences for treatment endpoints are not yet established. The objective of this study is to introduce the analytic hierarchy process (AHP) as a preference elicitation method in HTA. Based on a systematic literature review on the use of AHP in health care in 2009, the German Institute for Quality and Efficiency in Health Care (IQWiG) initiated an AHP study related to its HTA work in 2010.

Methods: The AHP study included two AHP workshops, one with twelve patients and one with seven healthcare professionals. In these workshops, both patients and professionals rated their preferences with respect to the importance of different endpoints of antidepressant treatment by a pairwise comparison of individual endpoints. These comparisons were performed and evaluated by the AHP method and relative weights were generated for each endpoint.

Results: The AHP study indicates that AHP is a well-structured technique whose cognitive demands were well handled by patients and professionals. The two groups rated some of the included endpoints of antidepressant treatment differently. For both groups, however, the same six of the eleven endpoints analyzed accounted for more than 80 percent of the total weight.

Conclusions: AHP can be used in HTA to give a quantitative dimension to patients' preferences for treatment endpoints. Preference elicitation could provide important information at various stages of HTA and challenge opinions on the importance of endpoints.

This study investigated cognitive predictors of medical decision-making capacity (MDC) in patients with amnestic mild cognitive impairment (MCI). A total of 56 healthy controls, 60 patients with MCI, and 31 patients with mild Alzheimer's disease (AD) were administered the Capacity to Consent to Treatment Instrument (CCTI) and a neuropsychological test battery. The CCTI assesses MDC across four established treatment consent standards—S1 (expressing choice), S3 (appreciation), S4 (reasoning), and S5 (understanding)—and one experimental standard [S2] (reasonable choice). Scores on neuropsychological measures were correlated with scores on each CCTI standard. Significant bivariate correlates were subsequently entered into stepwise regression analyses to identity group-specific multivariable predictors of MDC across CCTI standards. Different multivariable cognitive models emerged across groups and consent standards. For the MCI group, measures of short-term verbal memory were key predictors of MDC for each of the three clinically relevant standards (S3, S4, and S5). Secondary predictors were measures of executive function. In contrast, in the mild AD group, measures tapping executive function and processing speed were primary predictors of S3, S4, and S5. MDC in patients with MCI is supported primarily by short-term verbal memory. The findings demonstrate the impact of amnestic deficits on MDC in patients with MCI. (JINS, 2008, 14, 297–308.)Disclosure: The capacity outcome measure used in the study is owned by the UAB Research Foundation (UABRF). Dr. Marson and Dr. Harrell receive royalty income through UABRF. Statistical analyses were carried out by Mr. Okonkwo, and statistical consultation was provided by Dr. Raman.