For Me or for My Relatives? Approximating Self-Protection and Local Altruistic Motivations Underlying Preferences for Public Health Policies Using Risk Perception Metrics.

OBJECTIVES: Research efforts evaluating the role of altruistic motivations behind health policy support are usually based on direct preference elicitation procedures, which may be biased. We propose an indirect measurement approach to approximate self-protection-related and altruistic motivations underlying preferences for public health policies. METHODS: Our new approach relies on associations between on the one hand decision makers' perceived health risk for themselves and for close relatives and on the other hand their observed preferences for health policies that reduce such risks. The approach allows to make a rough distinction between health-related self-protection and local altruistic motives behind preferences for health policies. We illustrate our approach using data obtained from a discrete choice experiment in the context of policies to relax coronavirus-related lockdown measures in The Netherlands. RESULTS: Our results show that the approach is able to uncover that (1) people who think they have a high chance of experiencing health risks from a COVID-19 infection are more willing to accept a societal or personal sacrifice, (2) people with a higher health risk perception for their relatives have a higher willingness to accept sacrifices than people with a higher health risk perception for themselves, and (3) people who perceive that they have a high risk of dying of COVID-19 have a higher willingness to accept sacrifices than those anticipating less severe consequences of COVID-19. CONCLUSIONS: Our method offers a useful proxy metric to distinguish health-related self-protection and local altruism as drivers of citizens' responses to healthcare policies.

Hiding opinions by minimizing disclosed information: an obfuscation-based opinion dynamics model.

In the field of opinion dynamics, the hiding of opinions is routinely modeled as staying silent. However, staying silent is not always feasible. In situations where opinions are indirectly expressed by one's observable actions, people may however try to hide their opinions via a more complex and intelligent strategy called obfuscation, which minimizes the information disclosed to others. This study proposes a formal opinion dynamics model to study the hitherto unexplored effect of obfuscation on public opinion formation based on the recently developed Action-Opinion Inference Model. For illustration purposes, we use our model to simulate two cases with different levels of complexity, highlighting that the effect of obfuscation largely depends on the subtle relations between actions and opinions.

Using natural language processing to explore heterogeneity in moral terminology in palliative care consultations.

BACKGROUND: High quality serious illness communication requires good understanding of patients' values and beliefs for their treatment at end of life. Natural Language Processing (NLP) offers a reliable and scalable method for measuring and analyzing value- and belief-related features of conversations in the natural clinical setting. We use a validated NLP corpus and a series of statistical analyses to capture and explain conversation features that characterize the complex domain of moral values and beliefs. The objective of this study was to examine the frequency, distribution and clustering of morality lexicon expressed by patients during palliative care consultation using the Moral Foundations NLP Dictionary. METHODS: We used text data from 231 audio-recorded and transcribed inpatient PC consultations and data from baseline and follow-up patient questionnaires at two large academic medical centers in the United States. With these data, we identified different moral expressions in patients using text mining techniques. We used latent class analysis to explore if there were qualitatively different underlying patterns in the PC patient population. We used Poisson regressions to analyze if individual patient characteristics, EOL preferences, religion and spiritual beliefs were associated with use of moral terminology. RESULTS: We found two latent classes: a class in which patients did not use many expressions of morality in their PC consultations and one in which patients did. Age, race (white), education, spiritual needs, and whether a patient was affiliated with Christianity or another religion were all associated with membership of the first class. Gender, financial security and preference for longevity-focused over comfort focused treatment near EOL did not affect class membership. CONCLUSIONS: This study is among the first to use text data from a real-world situation to extract information regarding individual foundations of morality. It is the first to test empirically if individual moral expressions are associated with individual characteristics, attitudes and emotions.

Self-interest, positional concerns and distributional considerations in healthcare preferences.

Efficiently allocating scarce healthcare resources requires nuanced understanding of individual and collective interests as well as relative concerns, which may overlap or conflict. This paper is the first to empirically investigate whether and to what extent self-interest (SI), positional concerns (PC) and distributional considerations (DC) simultaneously explain individual decision making related to access to healthcare services. Our investigation is based on a stated choice experiment conducted in two countries with different healthcare systems, the United States (US) and the United Kingdom (UK). The choice experiment is on allocation of medical treatment waiting times for a hypothetical disease. We carry out the investigation under two different perspectives: (i) in a socially inclusive personal perspective decision makers were asked to choose between waiting time distributions for themselves and (ii) in a social perspective decision makers were asked to make similar choices for a close relative or friend of opposite gender. The results obtained by estimating a variety of advanced choice models indicate that DC, SI and PC, in this order of importance, are significant drivers of choice behaviour in our empirical context. These findings are consistent regardless of the choice perspective and the country where decision makers live. Comparing the results from different choice perspectives, we find that US respondents who chose for their close relative or friend attach significantly larger weight to their close relative's or friend's waiting times as well as to the overall distribution of waiting times than US respondents who chose for themselves. Looking at differences between countries, our results show that UK respondents who made choices for themselves placed significantly larger weight on SI and DC than US respondents, while US respondents, in turn, displayed relatively stronger but not significantly different positional concerns than UK respondents. In addition, we observe that UK respondents who chose for their close relative or friend put a larger weight on DC than their US counterparts. We conclude that the methodological (data collection and analysis) approach allows for disentangling the relative importance of the three motivations and discusses the potential implications of these findings for healthcare decision making.

Moral rhetoric in discrete choice models: a Natural Language Processing approach.

This paper proposes a new method to combine choice- and text data to infer moral motivations from people's actions. To do this, we rely on moral rhetoric, in other words, extracting moral values from verbal expressions with Natural Language Processing techniques. We use moral rhetoric based on a well-established moral, psychological theory called Moral Foundations Theory. We use moral rhetoric as input in Discrete Choice Models to gain insights into moral behaviour based on people's words and actions. We test our method in a case study of voting and party defection in the European Parliament. Our results indicate that moral rhetoric have significant explanatory power in modelling voting behaviour. We interpret the results in the light of political science literature and propose ways for future investigations.

Towards machine learning for moral choice analysis in health economics: A literature review and research agenda.

BACKGROUND: Discrete choice models (DCMs) for moral choice analysis will likely lead to erroneous model outcomes and misguided policy recommendations, as only some characteristics of moral decision-making are considered. Machine learning (ML) is recently gaining interest in the field of discrete choice modelling. This paper explores the potential of combining DCMs and ML to study moral decision-making more accurately and better inform policy decisions in healthcare. METHODS: An interdisciplinary literature search across four databases - PubMed, Scopus, Web of Science, and Arxiv - was conducted to gather papers. Based on the Preferred Reporting Items for Systematic and Meta-analyses (PRISMA) guideline, studies were screened for eligibility on inclusion criteria and extracted attributes from eligible papers. Of the 6285 articles, we included 277 studies. RESULTS: DCMs have shortcomings in studying moral decision-making. Whilst the DCMs' mathematical elegance and behavioural appeal hold clear interpretations, the models do not account for the 'moral' cost and benefit in an individual's utility calculation. The literature showed that ML obtains higher predictive power, model flexibility, and ability to handle large and unstructured datasets. Combining the strengths of ML methods with DCMs has the potential for studying moral decision-making. CONCLUSIONS: By providing a research agenda, this paper highlights that ML has clear potential to i) find and deepen the utility specification of DCMs, and ii) enrich the insights extracted from DCMs by considering the intrapersonal determinants of moral decision-making.

Surgery or comfort care for neonates with surgical necrotizing enterocolitis: Lessons learned from behavioral artificial intelligence technology.

Background: Critical decision making in surgical necrotizing enterocolitis (NEC) is highly complex and hard to capture in decision rules due to case-specificity and high mortality risk. In this choice experiment, we aimed to identify the implicit weight of decision factors towards future decision support, and to assess potential differences between specialties or centers. Methods: Thirty-five hypothetical surgical NEC scenarios with different factor levels were evaluated by neonatal care experts of all Dutch neonatal care centers in an online environment, where a recommendation for surgery or comfort care was requested. We conducted choice analysis by constructing a binary logistic regression model according to behavioral artificial intelligence technology (BAIT). Results: Out of 109 invited neonatal care experts, 62 (57%) participated, including 45 neonatologists, 16 pediatric surgeons and one neonatology physician assistant. Cerebral ultrasound (Relative importance = 20%, OR = 4.06, 95% CI = 3.39-4.86) was the most important factor in the decision surgery versus comfort care in surgical NEC, nationwide and for all specialties and centers. Pediatric surgeons more often recommended surgery compared to neonatologists (62% vs. 57%, p = 0.03). For all centers, cerebral ultrasound, congenital comorbidity, hemodynamics and parental preferences were significant decision factors (p < 0.05). Sex (p = 0.14), growth since birth (p = 0.25), and estimated parental capacities (p = 0.06) had no significance in nationwide nor subgroup analyses. Conclusion: We demonstrated how BAIT can analyze the implicit weight of factors in the complex and critical decision for surgery or comfort care for (surgical) NEC. The findings reflect Dutch expertise, but the technique can be expanded internationally. After validation, our choice model/BAIT may function as decision aid.

A simple satisficing model.

Economic theory is built on the assumption that people are omniscient utility maximizers. In reality, this is unlikely to be true and often people lack information about all alternatives that are available to them; either because the information is unavailable or that the cost of searching for and evaluating that information is high. In this paper, we develop a simple and tractable model that captures satisficing behavior. We show that the model can retrieve consistent parameters under a large range of experimental conditions. We test our model on synthetic data and present an empirical application. We discuss the implications of our results for the use of satisficing choice models in explaining choice.

Why Are General Moral Values Poor Predictors of Concrete Moral Behavior in Everyday Life? A Conceptual Analysis and Empirical Study.

Within moral psychology, theories focusing on the conceptualization and empirical measurement of people's morality in terms of general moral values -such as Moral Foundation Theory- (implicitly) assume general moral values to be relevant concepts for the explanation and prediction of behavior in everyday life. However, a solid theoretical and empirical foundation for this idea remains work in progress. In this study we explore this relationship between general moral values and daily life behavior through a conceptual analysis and an empirical study. Our conceptual analysis of the moral value-moral behavior relationship suggests that the effect of a generally endorsed moral value on moral behavior is highly context dependent. It requires the manifestation of several phases of moral decision-making, each influenced by many contextual factors. We expect that this renders the empirical relationship between generic moral values and people's concrete moral behavior indeterminate. Subsequently, we empirically investigate this relationship in three different studies. We relate two different measures of general moral values -the Moral Foundation Questionnaire and the Morality As Cooperation Questionnaire- to a broad set of self-reported morally relevant daily life behaviors (including adherence to COVID-19 measures and participation in voluntary work). Our empirical results are in line with the expectations derived from our conceptual analysis: the considered general moral values are poor predictors of the selected daily life behaviors. Furthermore, moral values that were tailored to the specific context of the behavior showed to be somewhat stronger predictors. Together with the insights derived from our conceptual analysis, this indicates the relevance of the contextual nature of moral decision-making as a possible explanation for the poor predictive value of general moral values. Our findings suggest that the investigation of morality's influence on behavior by expressing and measuring it in terms of general moral values may need revision.

A healthy debate: Exploring the views of medical doctors on the ethics of artificial intelligence.

Artificial Intelligence (AI) is moving towards the health space. It is generally acknowledged that, while there is great promise in the implementation of AI technologies in healthcare, it also raises important ethical issues. In this study we surveyed medical doctors based in The Netherlands, Portugal, and the U.S. from a diverse mix of medical specializations about the ethics surrounding Health AI. Four main perspectives have emerged from the data representing different views about this matter. The first perspective (AI is a helpful tool: Let physicians do what they were trained for) highlights the efficiency associated with automation, which will allow doctors to have the time to focus on expanding their medical knowledge and skills. The second perspective (Rules & Regulations are crucial: Private companies only think about money) shows strong distrust in private tech companies and emphasizes the need for regulatory oversight. The third perspective (Ethics is enough: Private companies can be trusted) puts more trust in private tech companies and maintains that ethics is sufficient to ground these corporations. And finally the fourth perspective (Explainable AI tools: Learning is necessary and inevitable) emphasizes the importance of explainability of AI tools in order to ensure that doctors are engaged in the technological progress. Each perspective provides valuable and often contrasting insights about ethical issues that should be operationalized and accounted for in the design and development of AI Health.

BAIT: A New Medical Decision Support Technology Based on Discrete Choice Theory.

We present a novel way to codify medical expertise and to make it available to support medical decision making. Our approach is based on econometric techniques (known as conjoint analysis or discrete choice theory) developed to analyze and forecast consumer or patient behavior; we reconceptualize these techniques and put them to use to generate an explainable, tractable decision support system for medical experts. The approach works as follows: using choice experiments containing systematically composed hypothetical choice scenarios, we collect a set of expert decisions. Then we use those decisions to estimate the weights that experts implicitly assign to various decision factors. The resulting choice model is able to generate a probabilistic assessment for real-life decision situations, in combination with an explanation of which factors led to the assessment. The approach has several advantages, but also potential limitations, compared to rule-based methods and machine learning techniques. We illustrate the choice model approach to support medical decision making by applying it in the context of the difficult choice to proceed to surgery v. comfort care for a critically ill neonate.

Diabolical dilemmas of COVID-19: An empirical study into Dutch society's trade-offs between health impacts and other effects of the lockdown.

We report and interpret preferences of a sample of the Dutch adult population for different strategies to end the so-called 'intelligent lockdown' which their government had put in place in response to the COVID-19 pandemic. Using a discrete choice experiment, we invited participants to make a series of choices between policy scenarios aimed at relaxing the lockdown, which were specified not in terms of their nature (e.g. whether or not to allow schools to re-open) but in terms of their effects along seven dimensions. These included health-related effects, but also impacts on the economy, education, and personal income. From the observed choices, we were able to infer the implicit trade-offs made by the Dutch between these policy effects. For example, we find that the average citizen, in order to avoid one fatality directly or indirectly related to COVID-19, is willing to accept a lasting lag in the educational performance of 18 children, or a lasting (>3 years) and substantial (>15%) reduction in net income of 77 households. We explore heterogeneity across individuals in terms of these trade-offs by means of latent class analysis. Our results suggest that most citizens are willing to trade-off health-related and other effects of the lockdown, implying a consequentialist ethical perspective. Somewhat surprisingly, we find that the elderly, known to be at relatively high risk of being affected by the virus, are relatively reluctant to sacrifice economic pain and educational disadvantages for the younger generation, to avoid fatalities. We also identify a so-called taboo trade-off aversion amongst a substantial share of our sample, being an aversion to accept morally problematic policies that simultaneously imply higher fatality numbers and lower taxes. We explain various ways in which our results can be of value to policy makers in the context of the COVID-19 and future pandemics.

From welcome culture to welcome limits? Uncovering preference changes over time for sheltering refugees in Germany.

Europe recently experienced a large influx of refugees, spurring much public debate about the admission and integration of refugees and migrants into society. Previous research based on cross-sectional data found that European citizens generally favour asylum seekers with high employability, severe vulnerabilities, and Christians over Muslims. These preferences and attitudes were found to be homogeneous across countries and socio-demographic groups. Here, we do not study the general acceptance of asylum seekers, but the acceptance of refugee and migrant homes in citizens' vicinity and how it changes over time. Based on a repeated stated choice experiment on preferences for refugee and migrant homes, we show that the initially promoted "welcome culture" towards refugees in Germany was not reflected in the views of a majority of a sample of German citizens who rather disapproved refugee homes in their vicinity. Their preferences have not changed between November 2015, the peak of "welcome culture," and November 2016, after political debates, media reporting and public discourse had shifted towards limiting admission of immigrants. A minority of one fifth of the sample population, who were initially rather approving of refugee and migrant homes being established in their vicinity, were more likely to change their preferences towards a rather disapproving position in 2016. Experience of contact with refugees and migrants, higher education, and general pro-immigration attitudes explain acceptance of refugee and migrant homes as well as preference stability over time. Country of origin and religion of refugees and migrants are considered less important than decent housing conditions and whether refugee and migrants arrive as families or single persons. In this respect our results highlight the importance of humanitarian aspects of sheltering and integration of refugees and other migrants into society.

A Large-Scale Analysis of Impact Factor Biased Journal Self-Citations.

Based on three decades of citation data from across scientific fields of science, we study trends in impact factor biased self-citations of scholarly journals, using a purpose-built and easy to use citation based measure. Our measure is given by the ratio between i) the relative share of journal self-citations to papers published in the last two years, and ii) the relative share of journal self-citations to papers published in preceding years. A ratio higher than one suggests that a journal's impact factor is disproportionally affected (inflated) by self-citations. Using recently reported survey data, we show that there is a relation between high values of our proposed measure and coercive journal self-citation malpractices. We use our measure to perform a large-scale analysis of impact factor biased journal self-citations. Our main empirical result is, that the share of journals for which our measure has a (very) high value has remained stable between the 1980s and the early 2000s, but has since risen strongly in all fields of science. This time span corresponds well with the growing obsession with the impact factor as a journal evaluation measure over the last decade. Taken together, this suggests a trend of increasingly pervasive journal self-citation malpractices, with all due unwanted consequences such as inflated perceived importance of journals and biased journal rankings.

Random regret-based discrete-choice modelling: an application to healthcare.

BACKGROUND: A new modelling approach for analysing data from discrete-choice experiments (DCEs) has been recently developed in transport economics based on the notion of regret minimization-driven choice behaviour. This so-called Random Regret Minimization (RRM) approach forms an alternative to the dominant Random Utility Maximization (RUM) approach. The RRM approach is able to model semi-compensatory choice behaviour and compromise effects, while being as parsimonious and formally tractable as the RUM approach. OBJECTIVES: Our objectives were to introduce the RRM modelling approach to healthcare-related decisions, and to investigate its usefulness in this domain. METHODS: Using data from DCEs aimed at determining valuations of attributes of osteoporosis drug treatments and human papillomavirus (HPV) vaccinations, we empirically compared RRM models, RUM models and Hybrid RUM-RRM models in terms of goodness of fit, parameter ratios and predicted choice probabilities. RESULTS: In terms of model fit, the RRM model did not outperform the RUM model significantly in the case of the osteoporosis DCE data (p = 0.21), whereas in the case of the HPV DCE data, the Hybrid RUM-RRM model outperformed the RUM model (p < 0.05). Differences in predicted choice probabilities between RUM models and (Hybrid RUM-) RRM models were small. Derived parameter ratios did not differ significantly between model types, but trade-offs between attributes implied by the two models can vary substantially. CONCLUSION: Differences in model fit between RUM, RRM and Hybrid RUM-RRM were found to be small. Although our study did not show significant differences in parameter ratios, the RRM and Hybrid RUM-RRM models did feature considerable differences in terms of the trade-offs implied by these ratios. In combination, our results suggest that RRM and Hybrid RUM-RRM modelling approach hold the potential of offering new and policy-relevant insights for health researchers and policy makers.