Children think differently from adults when reasoning about resources acquired from parents

Although sharing is often considered a virtuous behavior, individuals rarely share all their extra resources with those less fortunate. The current research investigated conditions under which children believe that someone who has more resources deserves to keep them rather than address an inequality. Specifically, we contrasted resources acquired via merit, windfall, and parental allocations. Across two studies, we showed 5- and 6-year-old children (n = 59), 8- and 9-year-old children (n = 120), and adults (n = 163) three scenarios in which one person acquired more resources than the other due to luck, due to merit, or because that person's parents gave him or her more, and we asked whether that person should share these resources or keep all of them. Results suggested that adults differentiated both the family resource and the merit conditions from the windfall allocation, believing that an agent deserves to keep the extra resources more when they are acquired through one's family or due to merit. However, children did not differentiate family resources from windfall, although they were more likely to believe that individuals deserve to keep their extra resources when they were acquired through merit. The type of the resource (i.e., money vs. balls) did not affect participants' sharing decisions. Overall, these findings suggest that over development the resources acquired from one's family come to be seen as more deserved than windfall resources.

Migrate demographic group for fair Graph Neural Networks

Graph Neural networks (GNNs) have been applied in many scenarios due to the superior performance of graph learning. However, fairness is always ignored when designing GNNs. As a consequence, biased information in training data can easily affect vanilla GNNs, causing biased results toward particular demographic groups (divided by sensitive attributes, such as race and age). There have been efforts to address the fairness issue. However, existing fair techniques generally divide the demographic groups by raw sensitive attributes and assume that are fixed. The biased information correlated with raw sensitive attributes will run through the training process regardless of the implemented fair techniques. It is urgent to resolve this problem for training fair GNNs. To tackle this problem, we propose a brand new framework, FairMigration, which is able to migrate the demographic groups dynamically, instead of keeping that fixed with raw sensitive attributes. FairMigration is composed of two training stages. In the first stage, the GNNs are initially optimized by personalized self-supervised learning, and the demographic groups are adjusted dynamically. In the second stage, the new demographic groups are frozen and supervised learning is carried out under the constraints of new demographic groups and adversarial training. Extensive experiments reveal that FairMigration achieves a high trade-off between model performance and fairness.

Supporting Fair and Efficient Emergency Medical Services in a Large Heterogeneous Region

Emergency Medical Services (EMS) are crucial in delivering timely and effective medical care to patients in need. However, the complex and dynamic nature of operations poses challenges for decision-making processes at strategic, tactical, and operational levels. This paper proposes an action-driven strategy for EMS management, employing a multi-objective optimizer and a simulator to evaluate potential outcomes of decisions. The approach combines historical data with dynamic simulations and multi-objective optimization techniques to inform decision-makers and improve the overall performance of the system. The research focuses on the Friuli Venezia Giulia region in north-eastern Italy. The region encompasses various landscapes and demographic situations that challenge fairness and equity in service access. Similar challenges are faced in other regions with comparable characteristics. The Decision Support System developed in this work accurately models the real-world system and provides valuable feedback and suggestions to EMS professionals, enabling them to make informed decisions and enhance the efficiency and fairness of the system.

A roadmap to artificial intelligence (AI): Methods for designing and building AI ready data to promote fairness

OBJECTIVES

We evaluated methods for preparing electronic health record data to reduce bias before applying artificial intelligence (AI).

METHODS

We created methods for transforming raw data into a data framework for applying machine learning and natural language processing techniques for predicting falls and fractures. Strategies such as inclusion and reporting for multiple races, mixed data sources such as outpatient, inpatient, structured codes, and unstructured notes, and addressing missingness were applied to raw data to promote a reduction in bias. The raw data was carefully curated using validated definitions to create data variables such as age, race, gender, and healthcare utilization. For the formation of these variables, clinical, statistical, and data expertise were used. The research team included a variety of experts with diverse professional and demographic backgrounds to include diverse perspectives.

RESULTS

For the prediction of falls, information extracted from radiology reports was converted to a matrix for applying machine learning. The processing of the data resulted in an input of 5,377,673 reports to the machine learning algorithm, out of which 45,304 were flagged as positive and 5,332,369 as negative for falls. Processed data resulted in lower missingness and a better representation of race and diagnosis codes. For fractures, specialized algorithms extracted snippets of text around keywork "femoral" from dual x-ray absorptiometry (DXA) scans to identify femoral neck T-scores that are important for predicting fracture risk. The natural language processing algorithms yielded 98% accuracy and 2% error rate The methods to prepare data for input to artificial intelligence processes are reproducible and can be applied to other studies.

CONCLUSION

The life cycle of data from raw to analytic form includes data governance, cleaning, management, and analysis. When applying artificial intelligence methods, input data must be prepared optimally to reduce algorithmic bias, as biased output is harmful. Building AI-ready data frameworks that improve efficiency can contribute to transparency and reproducibility. The roadmap for the application of AI involves applying specialized techniques to input data, some of which are suggested here. This study highlights data curation aspects to be considered when preparing data for the application of artificial intelligence to reduce bias.

Proposer's moral identity modulates fairness processing in the ultimatum game: Evidence from behavior and brain potentials

Economic decision-making is pivotal to both human private interests and the national economy. People pursue fairness in economic decision-making, but a proposer's moral identity can influence fairness processing. Previous ERP studies have revealed that moral identity has an effect on fairness considerations in the Ultimatum Game (UG), but the findings are inconsistent. To address the issue, we revised the moral-related sentences and used the ERP technique to measure the corresponding neural mechanism. We have observed that the fairness effect in UG can be mirrored in both MFN and P300 changes, whereas the moral identity effect on fairness in UG can be reflected by MFN but not P300 changes. These findings indicate that the moral identity of the proposer can modulate fairness processing in UG. The current study opens new avenues for clarifying the temporal course of the relationship between the proposer's moral identity and fairness in economic decision-making, which is beneficial for understanding the influencing mechanism of fairness processing and fair allocations in complex social contexts.

Fairness-aware recommendation with meta learning

Fairness has become a critical value online, and the latest studies consider it in many problems. In recommender systems, fairness is important since the visibility of items is controlled by systems. Previous fairness-aware recommender systems assume that sufficient relationship data between users and items are available. However, it is common that new users and items are frequently introduced, and they have no relationship data yet. In this paper, we study recommendation methods to enhance fairness in a cold-start state. Fairness is more significant when the preference of a user or the popularity of an item is unknown. We propose a meta-learning-based cold-start recommendation framework called FaRM to alleviate the unfairness of recommendations. The proposed framework consists of three steps. We first propose a fairness-aware meta-path generation method to eliminate bias in sensitive attributes. In addition, we construct fairness-aware user representations through the meta-path aggregation approach. Then, we propose a novel fairness objective function and introduce a joint learning method to minimize the trade-off between relevancy and fairness. In extensive experiments with various cold-start scenarios, it is shown that FaRM is significantly superior in fairness performance while preserving relevance accuracy over previous work.

How do decision making and fairness mediate the relationship between involuntary hospitalisation and perceived coercion among psychiatric inpatients?

BACKGROUND

Coercion perceived by psychiatric inpatients is not exclusively determined by formal measures such as involuntary admissions, seclusion or restraint, but is also associated with patients' characteristics and professionals' attitude.

AIMS

This study examined how inpatients' involvement in the decision making process, the respect of their decision making preference, and their feeling of having been treated fairly mediate the relationship between involuntary hospitalisation and perceived coercion both at admission and during hospital stay.

METHODS

Mediation analysis were performed in order to study the relationship between involuntary hospitalisation and perceived coercion among 230 patients, voluntarily and involuntarily admitted in six psychiatric hospitals.

RESULTS

32.2% of the participants were involuntarily hospitalised. Taken individually, stronger participants' involvement in decision making process, better respect for their decision making preference and higher level of perceived fairness partially mediated the relationship between involuntary hospitalisation and perceived coercion by reducing the level of the latter both at admission and during the hospitalisation. In multiple mediator models, only involvement and respect played an important role at admission. During the hospitalisation, perceived fairness was the most relevant mediator, followed by involvement in decision making.

CONCLUSIONS

During psychiatric hospitalisation patients' involvement in decision making, respect of their decision making preference and perceived fairness determined the relationship between involuntary hospitalisation and perceived coercion, but not in the same way at admission and during the stay. Involving patients in decision making and treating them fairly may be more relevant than taking account of their decision making preference in order to reduce perceived coercion.

A survey of recent methods for addressing AI fairness and bias in biomedicine

OBJECTIVES

Artificial intelligence (AI) systems have the potential to revolutionize clinical practices, including improving diagnostic accuracy and surgical decision-making, while also reducing costs and manpower. However, it is important to recognize that these systems may perpetuate social inequities or demonstrate biases, such as those based on race or gender. Such biases can occur before, during, or after the development of AI models, making it critical to understand and address potential biases to enable the accurate and reliable application of AI models in clinical settings. To mitigate bias concerns during model development, we surveyed recent publications on different debiasing methods in the fields of biomedical natural language processing (NLP) or computer vision (CV). Then we discussed the methods, such as data perturbation and adversarial learning, that have been applied in the biomedical domain to address bias.

METHODS

We performed our literature search on PubMed, ACM digital library, and IEEE Xplore of relevant articles published between January 2018 and December 2023 using multiple combinations of keywords. We then filtered the result of 10,041 articles automatically with loose constraints, and manually inspected the abstracts of the remaining 890 articles to identify the 55 articles included in this review. Additional articles in the references are also included in this review. We discuss each method and compare its strengths and weaknesses. Finally, we review other potential methods from the general domain that could be applied to biomedicine to address bias and improve fairness.

RESULTS

The bias of AIs in biomedicine can originate from multiple sources such as insufficient data, sampling bias and the use of health-irrelevant features or race-adjusted algorithms. Existing debiasing methods that focus on algorithms can be categorized into distributional or algorithmic. Distributional methods include data augmentation, data perturbation, data reweighting methods, and federated learning. Algorithmic approaches include unsupervised representation learning, adversarial learning, disentangled representation learning, loss-based methods and causality-based methods.

Drop the shortcuts: image augmentation improves fairness and decreases AI detection of race and other demographics from medical images

BACKGROUND

It has been shown that AI models can learn race on medical images, leading to algorithmic bias. Our aim in this study was to enhance the fairness of medical image models by eliminating bias related to race, age, and sex. We hypothesise models may be learning demographics via shortcut learning and combat this using image augmentation.

METHODS

This study included 44,953 patients who identified as Asian, Black, or White (mean age, 60.68 years ±18.21; 23,499 women) for a total of 194,359 chest X-rays (CXRs) from MIMIC-CXR database. The included CheXpert images comprised 45,095 patients (mean age 63.10 years ±18.14; 20,437 women) for a total of 134,300 CXRs were used for external validation. We also collected 1195 3D brain magnetic resonance imaging (MRI) data from the ADNI database, which included 273 participants with an average age of 76.97 years ±14.22, and 142 females. DL models were trained on either non-augmented or augmented images and assessed using disparity metrics. The features learned by the models were analysed using task transfer experiments and model visualisation techniques.

FINDINGS

In the detection of radiological findings, training a model using augmented CXR images was shown to reduce disparities in error rate among racial groups (-5.45%), age groups (-13.94%), and sex (-22.22%). For AD detection, the model trained with augmented MRI images was shown 53.11% and 31.01% reduction of disparities in error rate among age and sex groups, respectively. Image augmentation led to a reduction in the model's ability to identify demographic attributes and resulted in the model trained for clinical purposes incorporating fewer demographic features.

INTERPRETATION

The model trained using the augmented images was less likely to be influenced by demographic information in detecting image labels. These results demonstrate that the proposed augmentation scheme could enhance the fairness of interpretations by DL models when dealing with data from patients with different demographic backgrounds.

FUNDING

National Science and Technology Council (Taiwan), National Institutes of Health.

Pandemic lockdowns: who feels coerced and why? - a study on perceived coercion, perceived pressures and procedural justice during the UK COVID-19 lockdowns

BACKGROUND

This study examined perceptions of coercion, pressures and procedural injustice and how such perceptions influenced psychological well-being in those who experienced a UK COVID-19 lockdown, with a view to preparing for the possibility of future lockdowns.

METHODS

40 individuals categorised as perceiving the lockdown(s) as either highly or lowly coercive took part in one of six asynchronous virtual focus groups (AVFGs).

RESULTS

Using thematic analysis, the following key themes were identified in participants' discussions: (1) Choice, control and freedom; (2) threats; (3) fairness; (4) circumstantial factors; and (5) psychological factors.

CONCLUSIONS

As the first qualitative study to investigate the psychological construct of perceived coercion in relation to COVID-19 lockdowns, its findings suggest that the extent to which individuals perceived pandemic-related lockdowns as coercive may have been linked to their acceptance of restrictions. Preparing for future pandemics should include consideration of perceptions of coercion and efforts to combat this, particularly in relation to differences in equity, in addition to clarity of public health messaging and public engagement.

Context Modulates Perceived Fairness in Altruistic Punishment: Neural Signatures from ERPs and EEG Oscillations

Social norms and altruistic punitive behaviours are both based on the integration of information from multiple contexts. Individual behavioural performance can be altered by loss and gain contexts, which produce different mental states and subjective perceptions. In this study, we used event-related potential and time-frequency techniques to examine performance on a third-party punishment task and to explore the neural mechanisms underlying context-dependent differences in punishment decisions. The results indicated that individuals were more likely to reject unfairness in the context of loss (vs. gain) and to increase punishment as unfairness increased. In contrast, fairness appeared to cause an early increase in cognitive control signal enhancement, as indicated by the P2 amplitude and theta oscillations, and a later increase in emotional and motivational salience during decision-making in gain vs. loss contexts, as indicated by the medial frontal negativity and beta oscillations. In summary, individuals were more willing to sanction violations of social norms in the loss context than in the gain context and rejecting unfair losses induced more equity-related cognitive conflict than accepting unfair gains, highlighting the importance of context (i.e., gain vs. loss) in equity-related social decision-making processes.

COVID-19 vaccine refusal as unfair free-riding

Contributions to COVID-19 vaccination programmes promise valuable collective goods. They can support public and individual health by creating herd immunity and taking the pressure off overwhelmed public health services; support freedom of movement by enabling governments to remove restrictive lockdown policies; and improve economic and social well-being by allowing businesses, schools, and other essential public services to re-open. The vaccinated can contribute to the production of these goods. The unvaccinated, who benefit from, but who do not contribute to these goods can be morally criticised as free-riders. In this paper defends the claim that in the case of COVID-19, the unvaccinated are unfair free-riders. I defend the claim against two objections. First, that they are not unfair free-riders because they lack the subjective attitudes and intentions of free-riders; second, that although the unvaccinated may be free-riders, their free-riding is not unfair.

Age-based restrictions on reproductive care: discerning the arbitrary from the necessary

Policies that determine whether someone is allowed access to reproductive healthcare or not vary widely among countries, especially in their age requirements. This raises the suspicion of arbitrariness, especially because often no underlying justification is provided. In this article, we pose the question-under which circumstances is it morally acceptable to use age for policy and legislation in the first place? We start from the notion that everyone has a conditional positive right to fertility treatment. Subsequently, we set off to formulate a framework that helps to determine who should be excluded from treatment nonetheless. The framework's three core elements are: choosing and ethically justifying exclusion criteria (target), determining the actual limit between in- and exclusion (cut-off), and selecting variables that help to predict the exclusion criteria via correlation (as they are not directly measurable) (proxy). This framework allows us to show that referring to age in policy and legislation is only ethically justifiable if there is a sufficiently strong correlation with a non-directly measurable exclusion criterion. Moreover, since age is only one of many predicting variables, it should therefore not be ascribed any special status. Finally, our framework may be used as an argumentative scheme to critically assess the ethical legitimacy of policies that regulate access to (fertility) treatments in general.

The emergence of young children's tolerance for inequality: With age, children stop showing numerically sensitive fairness

One persistent and pernicious feature of outstanding social inequality is that even relatively extreme forms of inequality can be justified with reference to merit-based considerations. One key feature of fairness with respect to resource allocation is that it is numerically sensitive; greater (more extreme) inequalities are generally seen as less fair than less extreme ones. This work sought to document the emergence of numerically sensitive fairness in children aged 4 to 8 years. A total of 81 4- to 8-year-olds completed a series of within-participants fairness judgment trials in which they observed two characters receive either equitable or inequitable shares of resources-ranging from 50/50 (completely fair) to 0/100 (completely unfair)-in two contexts: one in which the two characters were described as working the same amount (equality context) and one in which one character was described as working harder than the other character (merit context). Children of all ages showed numerically sensitive fairness in the equality context. However, whereas younger children continued to show numerically sensitive fairness in the merit context, older children approved even relatively extreme inequalities when one person was described as working harder. This effect emerged with age, suggesting a double-edged sword to acquiring beliefs in merit-based fairness; as children get older, they may begin to accept even relatively extreme forms of inequality when presented in a merit context. Results are discussed with respect to the acquisition of meritocracy as a normative belief of fairness.

Disrupting the Dorsolateral Prefrontal Cortex Attenuates the Difference in Decision-Making for Altruistic Punishment Between the Gain and Loss Contexts

Altruistic punishment is a primary response to social norms violations; its neural mechanism has also attracted extensive research attention. In the present studies, we applied a low-frequency repetitive transcranial magnetic stimulation (rTMS) to the bilateral dorsolateral prefrontal cortex (DLPFC) while participants engaged in a modified Ultimatum Game (Study 1) and third-party punishment game (Study 2) to explore how the bilateral DLPFC disruption affects people's perception of violation of fairness norms and altruistic punishment decision in the gain and loss contexts. Typically, punishers intervene more often against and show more social outrage towards Dictators/Proposers who unfairly distribute losses than those who unfairly share gains. We found that disrupting the function of the left DLPFC in the second-party punishment and the bilateral DLPFC in the third-party punishment with rTMS effectively obliterated this difference, making participants punish unfairly shared gains as often as they usually would punish unfairly shared losses. In the altruistic punishment of maintaining the social fairness norms, the inhibition of the right DLPFC function will affect the deviation of individual information integration ability; the inhibition of the left DLPFC function will affect the assessment of the degree of violation of fairness norms and weaken impulse control, leading to attenuate the moderating effect of gain and loss contexts on altruistic punishment. Our findings emphasize that DLPFC is closely related to altruistic punishment and provide causal neuroscientific evidence.

A transformer-based deep learning approach for fairly predicting post-liver transplant risk factors

Liver transplantation is a life-saving procedure for patients with end-stage liver disease. There are two main challenges in liver transplant: finding the best matching patient for a donor and ensuring transplant equity among different subpopulations. The current MELD scoring system evaluates a patient's mortality risk if not receiving an organ within 90 days. However, the donor-patient matching should also consider post-transplant risk factors, such as cardiovascular disease, chronic rejection, etc., which are all common complications after transplant. Accurate prediction of these risk scores remains a significant challenge. In this study, we used predictive models to solve the above challenges. Specifically, we proposed a deep learning model to predict multiple risk factors after a liver transplant. By formulating it as a multi-task learning problem, the proposed deep neural network was trained to simultaneously predict the five post-transplant risks and achieve equal good performance by exploiting task-balancing techniques. We also proposed a novel fairness-achieving algorithm to ensure prediction fairness across different subpopulations. We used electronic health records of 160,360 liver transplant patients, including demographic information, clinical variables, and laboratory values, collected from the liver transplant records of the United States from 1987 to 2018. The model's performance was evaluated using various performance metrics such as AUROC and AUPRC. Our experiment results highlighted the success of our multi-task model in achieving task balance while maintaining accuracy. The model significantly reduced the task discrepancy by 39 %. Further application of the fairness-achieving algorithm substantially reduced fairness disparity among all sensitive attributes (gender, age group, and race/ethnicity) in each risk factor. It underlined the potency of integrating fairness considerations into the task-balancing framework, ensuring robust and fair predictions across multiple tasks and diverse demographic groups.

Fairness of artificial intelligence in healthcare: review and recommendations

In this review, we address the issue of fairness in the clinical integration of artificial intelligence (AI) in the medical field. As the clinical adoption of deep learning algorithms, a subfield of AI, progresses, concerns have arisen regarding the impact of AI biases and discrimination on patient health. This review aims to provide a comprehensive overview of concerns associated with AI fairness; discuss strategies to mitigate AI biases; and emphasize the need for cooperation among physicians, AI researchers, AI developers, policymakers, and patients to ensure equitable AI integration. First, we define and introduce the concept of fairness in AI applications in healthcare and radiology, emphasizing the benefits and challenges of incorporating AI into clinical practice. Next, we delve into concerns regarding fairness in healthcare, addressing the various causes of biases in AI and potential concerns such as misdiagnosis, unequal access to treatment, and ethical considerations. We then outline strategies for addressing fairness, such as the importance of diverse and representative data and algorithm audits. Additionally, we discuss ethical and legal considerations such as data privacy, responsibility, accountability, transparency, and explainability in AI. Finally, we present the Fairness of Artificial Intelligence Recommendations in healthcare (FAIR) statement to offer best practices. Through these efforts, we aim to provide a foundation for discussing the responsible and equitable implementation and deployment of AI in healthcare.

A Fair Individualized Polysocial Risk Score for Identifying Increased Social Risk in Type 2 Diabetes

Background

Racial and ethnic minority groups and individuals facing social disadvantages, which often stem from their social determinants of health (SDoH), bear a disproportionate burden of type 2 diabetes (T2D) and its complications. It is crucial to implement effective social risk management strategies at the point of care.

Objective

To develop an electronic health records (EHR)-based machine learning (ML) analytical pipeline to address unmet social needs associated with hospitalization risk in patients with T2D.

Methods

We identified real-world patients with T2D from the EHR data from University of Florida (UF) Health Integrated Data Repository (IDR), incorporating both contextual SDoH (e.g., neighborhood deprivation) and individual-level SDoH (e.g., housing instability). The 2015-2020 data were used for training and validation and 2021-2022 data for independent testing. We developed a machine learning analytic pipeline, namely individualized polysocial risk score (iPsRS), to identify high social risk associated with hospitalizations in T2D patients, along with explainable AI (XAI) and fairness optimization.

Results

The study cohort included 10,192 real-world patients with T2D, with a mean age of 59 years and 58% female. Of the cohort, 50% were non-Hispanic White, 39% were non-Hispanic Black, 6% were Hispanic, and 5% were other races/ethnicities. Our iPsRS, including both contextual and individual-level SDoH as input factors, achieved a C statistic of 0.72 in predicting 1-year hospitalization after fairness optimization across racial and ethnic groups. The iPsRS showed excellent utility for capturing individuals at high hospitalization risk because of SDoH, that is, the actual 1-year hospitalization rate in the top 5% of iPsRS was 28.1%, ~13 times as high as the bottom decile (2.2% for 1-year hospitalization rate).

Conclusion

Our ML pipeline iPsRS can fairly and accurately screen for patients who have increased social risk leading to hospitalization in real word patients with T2D.

Fair-split distribution of multi-dose vaccines with prioritized age groups and dynamic demand: The case study of COVID-19

The emergence of the SARS-CoV-2 virus and new viral variations with higher transmission and mortality rates have highlighted the urgency to accelerate vaccination to mitigate the morbidity and mortality of the COVID-19 pandemic. For this purpose, this paper formulates a new multi-vaccine, multi-depot location-inventory-routing problem for vaccine distribution. The proposed model addresses a wide variety of vaccination concerns: prioritizing age groups, fair distribution, multi-dose injection, dynamic demand, etc. To solve large-size instances of the model, we employ a Benders decomposition algorithm with a number of acceleration techniques. To monitor the dynamic demand of vaccines, we propose a new adjusted susceptible-infectious-recovered (SIR) epidemiological model, where infected individuals are tested and quarantined. The solution to the optimal control problem dynamically allocates the vaccine demand to reach the endemic equilibrium point. Finally, to illustrate the applicability and performance of the proposed model and solution approach, the paper reports extensive numerical experiments on a real case study of the vaccination campaign in France. The computational results show that the proposed Benders decomposition algorithm is 12 times faster, and its solutions are, on average, 16% better in terms of quality than the Gurobi solver under a limited CPU time. In terms of vaccination strategies, our results suggest that delaying the recommended time interval between doses of injection by a factor of 1.5 reduces the unmet demand up to 50%. Furthermore, we observed that the mortality is a convex function of fairness and an appropriate level of fairness should be adapted through the vaccination.

Locating helicopter ambulance bases in Iceland: efficient and fair solutions

BACKGROUND

Fixed-wing air ambulances play an important role in healthcare in rural Iceland. More frequent use of helicopter ambulances has been suggested to shorten response times and increase equity in access to advanced emergency care. In finding optimal base locations, the objective is often efficiency-maximizing the number of individuals who can be reached within a given time. This approach benefits people in densely populated areas more than people living in remote areas and the solution is not necessarily fair. This study aimed to find efficient and fair helicopter ambulance base locations in Iceland.

METHODS

We used high-resolution population and incident location data to estimate the service demand for helicopter ambulances, with possible base locations limited to twenty-one airports and landing strips around the country. Base locations were estimated using both the maximal covering location problem (MCLP) optimization model, which aimed for maximal coverage of demand, and the fringe sensitive location problem (FSLP) model, which also considered uncovered demand (i.e., beyond the response time threshold). We explored the percentage of the population and incidents covered by one to three helicopter bases within 45-, 60-, and 75-min response time thresholds, conditioned or not, on the single existing base located at Reykjavík Airport. This resulted in a total of eighteen combinations of conditions for each model. The models were implemented in R and solved using Gurobi.

RESULTS

Model solutions for base locations differed between the demand datasets for two out of eighteen combinations, both with the lowest service standard. Base locations differed between the MCLP and FSLP models for one combination involving a single base, and for two combinations involving two bases. Three bases covered all or almost all demand with longer response time thresholds, and the models differed in four of six combinations. The two helicopter ambulance bases can possibly obtain 97% coverage within 60 min, with bases in Húsafell and Grímsstaðir. Bases at Reykjavík Airport and Akureyri would cover 94.2%, whereas bases at Reykjavík Airport and Egilsstaðir would cover 88.5% of demand.

CONCLUSION

An efficient and fair solution would be to locate bases at Reykjavík Airport and in Akureyri or Egilsstaðir.

Cardiometabolic risk estimation using exposome data and machine learning

BACKGROUND

The human exposome encompasses all exposures that individuals encounter throughout their lifetime. It is now widely acknowledged that health outcomes are influenced not only by genetic factors but also by the interactions between these factors and various exposures. Consequently, the exposome has emerged as a significant contributor to the overall risk of developing major diseases, such as cardiovascular disease (CVD) and diabetes. Therefore, personalized early risk assessment based on exposome attributes might be a promising tool for identifying high-risk individuals and improving disease prevention.

OBJECTIVE

Develop and evaluate a novel and fair machine learning (ML) model for CVD and type 2 diabetes (T2D) risk prediction based on a set of readily available exposome factors. We evaluated our model using internal and external validation groups from a multi-center cohort. To be considered fair, the model was required to demonstrate consistent performance across different sub-groups of the cohort.

METHODS

From the UK Biobank, we identified 5,348 and 1,534 participants who within 13 years from the baseline visit were diagnosed with CVD and T2D, respectively. An equal number of participants who did not develop these pathologies were randomly selected as the control group. 109 readily available exposure variables from six different categories (physical measures, environmental, lifestyle, mental health events, sociodemographics, and early-life factors) from the participant's baseline visit were considered. We adopted the XGBoost ensemble model to predict individuals at risk of developing the diseases. The model's performance was compared to that of an integrative ML model which is based on a set of biological, clinical, physical, and sociodemographic variables, and, additionally for CVD, to the Framingham risk score. Moreover, we assessed the proposed model for potential bias related to sex, ethnicity, and age. Lastly, we interpreted the model's results using SHAP, a state-of-the-art explainability method.

RESULTS

The proposed ML model presents a comparable performance to the integrative ML model despite using solely exposome information, achieving a ROC-AUC of 0.78±0.01 and 0.77±0.01 for CVD and T2D, respectively. Additionally, for CVD risk prediction, the exposome-based model presents an improved performance over the traditional Framingham risk score. No bias in terms of key sensitive variables was identified.

CONCLUSIONS

We identified exposome factors that play an important role in identifying patients at risk of CVD and T2D, such as naps during the day, age completed full-time education, past tobacco smoking, frequency of tiredness/unenthusiasm, and current work status. Overall, this work demonstrates the potential of exposome-based machine learning as a fair CVD and T2D risk assessment tool.

Conceptualizing fairness in the secondary use of health data for research: A scoping review

With the introduction of the European Health Data Space (EHDS), the secondary use of health data for research purposes is attracting more attention. Secondary health data processing promises to address novel research questions, inform the design of future research and improve healthcare delivery generally. To comply with the existing data protection regulations, the secondary data use must be fair, among other things. However, there is no clear understanding of what fairness means in the context of secondary use of health data for scientific research purposes. In response, we conducted a scoping review of argument-based literature to explore how fairness in the secondary use of health data has been conceptualized. A total of 35 publications were included in the final synthesis after abstract and full-text screening. Using an inductive approach and a thematic analysis, our review has revealed that balancing individual and public interests, reducing power asymmetries, setting conditions for commercial involvement, and implementing benefit sharing are essential to guarantee fair secondary use research. The findings of this review can inform current and future research practices and policy development to adequately address concerns about fairness in the secondary use of health data.

Availability of information needed to evaluate algorithmic fairness - A systematic review of publicly accessible critical care databases

BACKGROUND

Machine learning (ML) may improve clinical decision-making in critical care settings, but intrinsic biases in datasets can introduce bias into predictive models. This study aims to determine if publicly available critical care datasets provide relevant information to identify historically marginalized populations.

METHOD

We conducted a review to identify the manuscripts that report the training/validation of ML algorithms using publicly accessible critical care electronic medical record (EMR) datasets. The datasets were reviewed to determine if the following 12 variables were available: age, sex, gender identity, race and/or ethnicity, self-identification as an indigenous person, payor, primary language, religion, place of residence, education, occupation, and income.

RESULTS

7 publicly available databases were identified. Medical Information Mart for Intensive Care (MIMIC) reports information on 7 of the 12 variables of interest, Sistema de Informação de Vigilância Epidemiológica da Gripe (SIVEP-Gripe) on 7, COVID-19 Mexican Open Repository on 4, and eICU on 4. Other datasets report information on 2 or fewer variables. All 7 databases included information about sex and age. Four databases (57%) included information about whether a patient identified as native or indigenous. Only 3 (43%) included data about race and/or ethnicity. Two databases (29%) included information about residence, and one (14%) included information about payor, language, and religion. One database (14%) included information about education and patient occupation. No databases included information on gender identity and income.

CONCLUSION

This review demonstrates that critical care publicly available data used to train AI algorithms do not include enough information to properly look for intrinsic bias and fairness issues towards historically marginalized populations.

Drop edges and adapt: A fairness enforcing fine-tuning for graph neural networks

The rise of graph representation learning as the primary solution for many different network science tasks led to a surge of interest in the fairness of this family of methods. Link prediction, in particular, has a substantial social impact. However, link prediction algorithms tend to increase the segregation in social networks by disfavouring the links between individuals in specific demographic groups. This paper proposes a novel way to enforce fairness on graph neural networks with a fine-tuning strategy. We Drop the unfair Edges and, simultaneously, we Adapt the model's parameters to those modifications, DEA in short. We introduce two covariance-based constraints designed explicitly for the link prediction task. We use these constraints to guide the optimization process responsible for learning the new 'fair' adjacency matrix. One novelty of DEA is that we can use a discrete yet learnable adjacency matrix in our fine-tuning. We demonstrate the effectiveness of our approach on five real-world datasets and show that we can improve both the accuracy and the fairness of the link prediction tasks. In addition, we present an in-depth ablation study demonstrating that our training algorithm for the adjacency matrix can be used to improve link prediction performances during training. Finally, we compute the relevance of each component of our framework to show that the combination of both the constraints and the training of the adjacency matrix leads to optimal performances.

Multidisciplinary considerations of fairness in medical AI: A scoping review

INTRODUCTION

Artificial Intelligence (AI) technology has been developed significantly in recent years. The fairness of medical AI is of great concern due to its direct relation to human life and health. This review aims to analyze the existing research literature on fairness in medical AI from the perspectives of computer science, medical science, and social science (including law and ethics). The objective of the review is to examine the similarities and differences in the understanding of fairness, explore influencing factors, and investigate potential measures to implement fairness in medical AI across English and Chinese literature.

METHODS

This study employed a scoping review methodology and selected the following databases: Web of Science, MEDLINE, Pubmed, OVID, CNKI, WANFANG Data, etc., for the fairness issues in medical AI through February 2023. The search was conducted using various keywords such as "artificial intelligence," "machine learning," "medical," "algorithm," "fairness," "decision-making," and "bias." The collected data were charted, synthesized, and subjected to descriptive and thematic analysis.

RESULTS

After reviewing 468 English papers and 356 Chinese papers, 53 and 42 were included in the final analysis. Our results show the three different disciplines all show significant differences in the research on the core issues. Data is the foundation that affects medical AI fairness in addition to algorithmic bias and human bias. Legal, ethical, and technological measures all promote the implementation of medical AI fairness.

CONCLUSIONS

Our review indicates a consensus regarding the importance of data fairness as the foundation for achieving fairness in medical AI across multidisciplinary perspectives. However, there are substantial discrepancies in core aspects such as the concept, influencing factors, and implementation measures of fairness in medical AI. Consequently, future research should facilitate interdisciplinary discussions to bridge the cognitive gaps between different fields and enhance the practical implementation of fairness in medical AI.

Enhancing Fairness in Disease Prediction by Optimizing Multiple Domain Adversarial Networks

Predictive models in biomedicine need to ensure equitable and reliable outcomes for the populations they are applied to. Unfortunately, biases in medical predictions can lead to unfair treatment and widening disparities, underscoring the need for effective techniques to address these issues. To enhance fairness, we introduce a framework based on a Multiple Domain Adversarial Neural Network (MDANN), which incorporates multiple adversarial components. In an MDANN, an adversarial module is applied to learn a fair pattern by negative gradients back-propagating across multiple sensitive features (i.e., characteristics of individuals that should not be used to discriminate unfairly between individuals when making predictions or decisions.) We leverage loss functions based on the Area Under the Receiver Operating Characteristic Curve (AUC) to address the class imbalance, promoting equitable classification performance for minority groups (e.g., a subset of the population that is underrepresented or disadvantaged.) Moreover, we utilize pre-trained convolutional autoencoders (CAEs) to extract deep representations of data, aiming to enhance prediction accuracy and fairness. Combining these mechanisms, we alleviate biases and disparities to provide reliable and equitable disease prediction. We empirically demonstrate that the MDANN approach leads to better accuracy and fairness in predicting disease progression using brain imaging data for Alzheimer's Disease and Autism populations than state-of-the-art techniques.

Fairness and generalizability of OCT normative databases: a comparative analysis

PURPOSE

In supervised Machine Learning algorithms, labels and reports are important in model development. To provide a normality assessment, the OCT has an in-built normative database that provides a color base scale from the measurement database comparison. This article aims to evaluate and compare normative databases of different OCT machines, analyzing patient demographic, contrast inclusion and exclusion criteria, diversity index, and statistical approach to assess their fairness and generalizability.

METHODS

Data were retrieved from Cirrus, Avanti, Spectralis, and Triton's FDA-approval and equipment manual. The following variables were compared: number of eyes and patients, inclusion and exclusion criteria, statistical approach, sex, race and ethnicity, age, participant country, and diversity index.

RESULTS

Avanti OCT has the largest normative database (640 eyes). In every database, the inclusion and exclusion criteria were similar, including adult patients and excluding pathological eyes. Spectralis has the largest White (79.7%) proportionately representation, Cirrus has the largest Asian (24%), and Triton has the largest Black (22%) patient representation. In all databases, the statistical analysis applied was Regression models. The sex diversity index is similar in all datasets, and comparable to the ten most populous contries. Avanti dataset has the highest diversity index in terms of race, followed by Cirrus, Triton, and Spectralis.

CONCLUSION

In all analyzed databases, the data framework is static, with limited upgrade options and lacking normative databases for new modules. As a result, caution in OCT normality interpretation is warranted. To address these limitations, there is a need for more diverse, representative, and open-access datasets that take into account patient demographics, especially considering the development of supervised Machine Learning algorithms in healthcare.

Differentiated Instruction as a Viable Framework for Meeting the Needs of Diverse Adult Learners in Health Professions Education

Health professions education (HPE) instructors are often challenged with simultaneously teaching adult learners of varying educational levels, needs, and backgrounds. With an increased focus on interprofessional education, instructors may be tasked with teaching extremely diverse audiences during a single educational session. While some aspects of differentiated instruction (DI) have been implemented within HPE contexts, the DI framework appears to be relatively unknown in many fields. Evidence from a range of educational fields outside of HPE supports the use of DI as a framework to enhance fairness, diversity and inclusion while meeting core instructional needs. In this Monograph, we explore DI and offer strategies for implementation amenable to many HPE settings.

SAF: Stakeholders' Agreement on Fairness in the Practice of Machine Learning Development

This paper clarifies why bias cannot be completely mitigated in Machine Learning (ML) and proposes an end-to-end methodology to translate the ethical principle of justice and fairness into the practice of ML development as an ongoing agreement with stakeholders. The pro-ethical iterative process presented in the paper aims to challenge asymmetric power dynamics in the fairness decision making within ML design and support ML development teams to identify, mitigate and monitor bias at each step of ML systems development. The process also provides guidance on how to explain the always imperfect trade-offs in terms of bias to users.

Multi-task learning with dynamic re-weighting to achieve fairness in healthcare predictive modeling

The emphasis on fairness in predictive healthcare modeling has increased in popularity as an approach for overcoming biases in automated decision-making systems. The aim is to guarantee that sensitive characteristics like gender, race, and ethnicity do not influence prediction outputs. Numerous algorithmic strategies have been proposed to reduce bias in prediction results, mitigate prejudice toward minority groups and promote prediction fairness. The goal of these strategies is to ensure that model prediction performance does not exhibit significant disparity among sensitive groups. In this study, we propose a novel fairness-achieving scheme based on multitask learning, which fundamentally differs from conventional fairness-achieving techniques, including altering data distributions and constraint optimization through regularizing fairness metrics or tampering with prediction outcomes. By dividing predictions on different sub-populations into separate tasks, we view the fairness problem as a task-balancing problem. To ensure fairness during the model-training process, we suggest a novel dynamic re-weighting approach. Fairness is achieved by dynamically modifying the gradients of various prediction tasks during neural network back-propagation, and this novel technique applies to a wide range of fairness criteria. We conduct tests on a real-world use case to predict sepsis patients' mortality risk. Our approach satisfies that it can reduce the disparity between subgroups by 98% while only losing less than 4% of prediction accuracy.

Bias Analysis in Healthcare Time Series (BAHT) Decision Support Systems from Meta Data

One of the hindrances in the widespread acceptance of deep learning-based decision support systems in healthcare is bias. Bias in its many forms occurs in the datasets used to train and test deep learning models and is amplified when deployed in the real world, leading to challenges such as model drift. Recent advancements in the field of deep learning have led to the deployment of deployable automated healthcare diagnosis decision support systems at hospitals as well as tele-medicine through IoT devices. Research has been focused primarily on the development and improvement of these systems leaving a gap in the analysis of the fairness. The domain of FAccT ML (fairness, accountability, and transparency) accounts for the analysis of these deployable machine learning systems. In this work, we present a framework for bias analysis in healthcare time series (BAHT) signals such as electrocardiogram (ECG) and electroencephalogram (EEG). BAHT provides a graphical interpretive analysis of bias in the training, testing datasets in terms of protected variables, and analysis of bias amplification by the trained supervised learning model for time series healthcare decision support systems. We thoroughly investigate three prominent time series ECG and EEG healthcare datasets used for model training and research. We show the extensive presence of bias in the datasets leads to potentially biased or unfair machine-learning models. Our experiments also demonstrate the amplification of identified bias with an observed maximum of 66.66%. We investigate the effect of model drift due to unanalyzed bias in datasets and algorithms. Bias mitigation though prudent is a nascent area of research. We present experiments and analyze the most prevalently accepted bias mitigation strategies of under-sampling, oversampling, and the use of synthetic data for balancing the dataset through augmentation. It is important that healthcare models, datasets, and bias mitigation strategies should be properly analyzed for a fair unbiased delivery of service.

Exploring views of members of the public and policymakers on the acceptability of population level dietary and active-travel policies: a qualitative study

BACKGROUND

There is limited evidence on what shapes the acceptability of population level dietary and active-travel policies in England. This information would be useful in the decision-making process about which policies should be implemented and how to increase their effectiveness and sustainability. To fill this gap, we explored public and policymakers' views about factors that influence public acceptability of dietary and active-travel policies and how to increase public acceptability for these policies.

METHODS

We conducted online, semi-structured interviews with 20 members of the public and 20 policymakers in England. A purposive sampling frame was used to recruit members of the public via a recruitment agency, based on age, sex, socioeconomic status and ethnicity. Policymakers were recruited from existing contacts within our research collaborations and via snowball sampling. We explored different dietary and active-travel policies that varied in their scope and focus. Interviews were transcribed verbatim and analysed using thematic reflexive analysis with both inductive and deductive coding.

RESULTS

We identified four themes that informed public acceptability of dietary and active-travel policies: (1) perceived policy effectiveness, i.e., policies that included believable mechanisms of action, addressed valued co-benefits and barriers to engage in the behaviour; (2) perceived policy fairness, i.e., policies that provided everyone with an opportunity to benefit (mentioned only by the public), equally considered the needs of various population subgroups and rewarded 'healthy' behaviours rather than only penalising 'unhealthy' behaviours; (3) communication of policies, i.e., policies that were visible and had consistent and positive messages from the media (mentioned only by policymakers) and (4) how to improve policy support, with the main suggestion being an integrated strategy addressing multiple aspects of these behaviours, inclusive policies that consider everyone's needs and use of appropriate channels and messages in policy communication.

CONCLUSIONS

Our findings highlight that members' of the public and policymakers' support for dietary and active-travel policies can be shaped by the perceived effectiveness, fairness and communication of policies and provide suggestions on how to improve policy support. This information can inform the design of acceptable policies but can also be used to help communicate existing and future policies to maximise their adoption and sustainability.

Emotions as social information in unambiguous situations: role of emotions on procedural justice perception

Emotion as Social Information Theory claims that in an ambiguous situation, people rely on others' emotions to make sense of the level of fairness encountered. We tested whether the information provided by emotions about the fairness of a procedure is still a significant factor in explaining individual differences in perception of variance, even in unambiguous situations. We assessed the effects of others' emotions on observers inferred procedural justice during (un)ambiguous situations when people are treated (un)fairly. We collected data using Qualtrics online survey software from 1012 employees across different industry services in the United States. The participants were assigned randomly to one of the 12 experimental conditions (fair, unfair, and unknown x happiness, anger, guilt, and neutral). The results indicated that emotions played a significant role in the psychology of justice judgments under the ambiguous situation, as predicted by the EASI, as well as under unambiguous conditions. The study revealed significant interactions between the procedure and emotion. These findings emphasized the importance of considering how others' emotions influence an observer's perception of justice. The theoretical and practical implications of these findings were also discussed.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12144-023-04640-y.

Fairness in Artificial Intelligence: Regulatory Sanbox Evaluation of Bias Prevention for ECG Classification

As the use of artificial intelligence within healthcare is on the rise, an increased attention has been directed towards ethical considerations. Defining fairness in machine learning is a well explored topic with an extensive literature. However, such definitions often rely on the existence of metrics on the input data and well-defined outcome measurements, while regulatory definitions use general terminology. This work aims to study fairness within AI, particularly bringing regulation and theoretical knowledge closer. The study is done via a regulatory sandbox implemented on a healthcare case, specifically ECG classification.

Narrative review of mathematical and psychological studies of staff scheduling for holidays as applicable to anesthesiologists and nurse anesthetists

We performed a narrative review of articles applicable to anesthesiologists' and nurse anesthetists' choices of who works each statutory holiday for operating room and non-operating room anesthesia. We include search protocols and detailed supplementary annotated comments. Studies showed that holiday staff scheduling is emotional. Working on holidays often is more stressful and undesirable than comparable workdays. Intrinsic motivation may overall, among practitioners, be greater by preferentially scheduling practitioners who choose to work on holidays, for compensation, before mandating that practitioners who would prefer to be off must work on holidays. Granting each practitioner (who so desires) at least one major holiday off can depend on identifying and scheduling other clinicians who want to work holidays for monetary compensation or extra compensatory time off. Scheduling holidays by random priority (i.e., a lottery choosing who gets to pick their holiday[s] first, second, etc.) is inefficient, resulting in fewer practitioners having their preferences satisfied, especially for small departments or divisions (e.g., cardiac anesthesia). No article that we reviewed implemented a random priority mechanism for staff scheduling. The selection of practitioners to take turns in choosing their holidays is perceived to have less fairness than a selection process that collects each participants' preferences. Although holidays often are scheduled separately from regular workdays and weekends, doing so will not increase efficiency or fairness. Holidays can, in practice, be scheduled simultaneously with non-holidays. Models can explicitly include fairness as an objective. For example, fairness can be based on the difference between the maximum and minimum number of holidays for which practitioners of the same division are scheduled. Holidays can be given greater weights than other shifts when estimating fairness. Staff scheduling for holidays, when done simultaneously with regular workdays, nights, and weekends, can also use personalized weights, specifying practitioners' preferences to be satisfied if possible.

Fairness and COVID: Conducting Research During the Crisis

The COVID-19 pandemic forced Principal Investigators (PIs) to make rapid and unprecedented decisions about ongoing research projects and research teams. Confronted with vague or shifting guidance from institutional administrators and public health officials, PIs nonetheless had to decide whether their projects were "essential," who could conduct on-site "essential" research, how to continue research activities by remote means if possible, and how to safely and effectively manage personnel during the crisis. Based on both narrative comments from a federally sponsored survey of over a thousand NIH- and NSF-funded PIs and their personnel, as well as follow-up interviews with over 60 survey participants, this study examines various ways PI and institutional decisions raised issues of procedural and distributive fairness. These fairness issues include the challenge of treating research personnel fairly in light of their disparate personal circumstances and inconsistent enforcement of COVID-19-related directives. Our findings highlight aspects of fairness and equitability that all PIs and research administrators should keep in mind for when future research disruptions occur.

A guide to formulating fairness in an optimization model

Optimization models typically seek to maximize overall benefit or minimize total cost. Yet fairness is an important element of many practical decisions, and it is much less obvious how to express it mathematically. We provide a critical survey of various schemes that have been proposed for formulating ethics-related criteria, including those that integrate efficiency and fairness concerns. The survey covers inequality measures, Rawlsian maximin and leximax criteria, convex combinations of fairness and efficiency, alpha fairness and proportional fairness (also known as the Nash bargaining solution), Kalai-Smorodinsky bargaining, and recently proposed utility-threshold and fairness-threshold schemes for combining utilitarian with maximin or leximax criteria. The paper also examines group parity metrics that are popular in machine learning. We present what appears to be the best practical approach to formulating each criterion in a linear, nonlinear, or mixed integer programming model. We also survey axiomatic and bargaining derivations of fairness criteria from the social choice literature while taking into account interpersonal comparability of utilities. Finally, we cite relevant philosophical and ethical literature where appropriate.

A fair and interpretable network for clinical risk prediction: a regularized multi-view multi-task learning approach

In healthcare domain, complication risk profiling which can be seen as multiple clinical risk prediction tasks is challenging due to the complex interaction between heterogeneous clinical entities. With the availability of real-world data, many deep learning methods are proposed for complication risk profiling. However, the existing methods face three open challenges. First, they leverage clinical data from a single view and then lead to suboptimal models. Second, most existing methods lack an effective mechanism to interpret predictions. Third, models learned from clinical data may have inherent pre-existing biases and exhibit discrimination against certain social groups. We then propose a multi-view multi-task network (MuViTaNet) to tackle these issues. MuViTaNet complements patient representation by using a multi-view encoder to exploit more information. Moreover, it uses a multi-task learning to generate more generalized representations using both labeled and unlabeled datasets. Last, a fairness variant (F-MuViTaNet) is proposed to mitigate the unfairness issues and promote healthcare equity. The experiments show that MuViTaNet outperforms existing methods for cardiac complication profiling. Its architecture also provides an effective mechanism for interpreting the predictions, which helps clinicians discover the underlying mechanism triggering the complication onsets. F-MuViTaNet can also effectively mitigate the unfairness with only negligible impact on accuracy.

Fairness and Tax Morale in Developing Countries

This paper investigates the relationship between individuals' attitudes towards fairness and their views about tax compliance in developing countries. It argues that individuals' attitudes regarding fairness shape their views about paying taxes and their ethical stances regarding tax evasion. Using survey data for 18 major cities in Latin America, we find that individuals who are highly sensitive to fairness are less likely to consider paying taxes as a civic duty and more likely to justify tax evasion. These attitudes toward tax compliance are not inelastic. We also find evidence that individualst argues about reciprocity and merit mediate the effect of fairness on personal views about tax compliance. Finally, this paper shows that the heuristics people use to explain their position in the income distribution make them sensitive to inequality, and it affects their tax morale. These findings help us better understand the concept of reciprocity and provide valuable lessons on the urgent task of expanding fiscal capacity to promote economic growth and inequality in developing countries.

PreCoF: counterfactual explanations for fairness

This paper studies how counterfactual explanations can be used to assess the fairness of a model. Using machine learning for high-stakes decisions is a threat to fairness as these models can amplify bias present in the dataset, and there is no consensus on a universal metric to detect this. The appropriate metric and method to tackle the bias in a dataset will be case-dependent, and it requires insight into the nature of the bias first. We aim to provide this insight by integrating explainable AI (XAI) research with the fairness domain. More specifically, apart from being able to use (Predictive) Counterfactual Explanations to detect explicit bias when the model is directly using the sensitive attribute, we show that it can also be used to detect implicit bias when the model does not use the sensitive attribute directly but does use other correlated attributes leading to a substantial disadvantage for a protected group. We call this metric PreCoF, or Predictive Counterfactual Fairness. Our experimental results show that our metric succeeds in detecting occurrences of implicit bias in the model by assessing which attributes are more present in the explanations of the protected group compared to the unprotected group. These results could help policymakers decide on whether this discrimination is justified or not.

Children with medical complexities: their distinct vulnerability in health systems' Covid-19 response and their claims of justice in the recovery phase

In this paper, we discuss the lack of consideration given to children in the COVID-19 health systems policy response to the pandemic. We do this by focusing on the case of children with complex medical needs. We argue that, in broad terms, health systems policies that were implemented during the pandemic failed adequately to meet our obligations to both children generally and those with complex medical needs by failing to consider those needs and so to give them fair protection against harm and disadvantage. We argue that justice requires that the distinct needs and vulnerabilities of children with medical complexities are explicitly integrated and prioritised in decisions concerning healthcare and operational planning in the recovery phase and beyond.

Peer review practices in academic medicine: how the example of orthopaedic surgery may help shift the paradigm?

PURPOSE

To establish the current peer-reviewed practices in the discipline of orthopaedic surgery and correlate these to the journal's impact factor. Unfortunately, this is not receiving much attention and a critical literature gap in various disciplines; thus, determining the current practices in the discipline of orthopaedic surgery could provide valid insight that may be potentially applicable to other academic medicine disciplines as well.

METHODS

Orthopaedic surgery journals belonging to the Journal Citation Reports were queried, and the following was extracted: impact factor (IF) and blinding practices: single (SBPR), double (DBPR), triple (TBPR), quadruple (QBPR), and open (OPR) blinding review process and possibility of author-suggested reviewer (ASR) and non-preferred reviewer (NPR) options.

RESULTS

Of the 82 journals, four were excluded as they allowed submission by invitation only. In the remaining, blinding was as follows: SBPR nine (11.5%), DBPR 52 (66.7%), TBPR two (2.6%), QBPR zero (0%), and OPR three (3.8%), and in 12 (15.4%), this was unclear. ASR and NPR options were offered by 34 (43.6%) and 27 (34.6%) journals respectively, whereas ASR was mandatory in eight (10.2%). No correlation between IF and any other parameter was found.

CONCLUSION

The rules of the "game" are unclear/not disclosed in a significant number of cases, and the SBPR system, along with the ASR (mandatory sometimes) and NPR, is still extensively used with questionable integrity and fairness. Several recommendations are provided to mitigate potentially compromising practices, along with future directions to address the scarcity of research in this critical aspect of science.

Mining for equitable health: Assessing the impact of missing data in electronic health records

Electronic health records (EHR) are collected as a routine part of healthcare delivery, and have great potential to be utilized to improve patient health outcomes. They contain multiple years of health information to be leveraged for risk prediction, disease detection, and treatment evaluation. However, they do not have a consistent, standardized format across institutions, particularly in the United States, and can present significant analytical challenges- they contain multi-scale data from heterogeneous domains and include both structured and unstructured data. Data for individual patients are collected at irregular time intervals and with varying frequencies. In addition to the analytical challenges, EHR can reflect inequity- patients belonging to different groups will have differing amounts of data in their health records. Many of these issues can contribute to biased data collection. The consequence is that the data for under-served groups may be less informative partly due to more fragmented care, which can be viewed as a type of missing data problem. For EHR data in this complex form, there is currently no framework for introducing realistic missing values. There has also been little to no work in assessing the impact of missing data in EHR. In this work, we first introduce a terminology to define three levels of EHR data and then propose a novel framework for simulating realistic missing data scenarios in EHR to adequately assess their impact on predictive modeling. We incorporate the use of a medical knowledge graph to capture dependencies between medical events to create a more realistic missing data framework. In an intensive care unit setting, we found that missing data have greater negative impact on the performance of disease prediction models in groups that tend to have less access to healthcare, or seek less healthcare. We also found that the impact of missing data on disease prediction models is stronger when using the knowledge graph framework to introduce realistic missing values as opposed to random event removal.

Evidence-Informed Deliberative Processes for UHC: Progress, Potential and Prudence Comment on "Evidence-Informed Deliberative Processes for Health Benefit Package Design - Part II: A Practical Guide"

In their recent article on evidence-informed deliberative processes (EDPs) for health benefit package decisions, Oortwijn et al examine how the different steps of EDP play out in eight countries with relatively mature institutions for using health technology assessment (HTA). This commentary examines how EDP addresses stakeholder involvement in decision-making for equitable progress towards universal health coverage (UHC). It focuses on the value of inclusiveness, the need to pay attention to trade-offs between desirable features of EDP and the need to broaden the scope of processes examined beyond those specifically tied to producing and using HTAs . It concludes that EDPs have contributed to significant progress for health benefit design decisions worldwide and holds much potential in further application. At the same time, this commentary calls for prudence: investments in EDPs should be efficiently deployed to enhance the pre-existing legislative, institutional and political framework that exist to promote fair and legitimate healthcare decisions.

Equitable modelling of brain imaging by counterfactual augmentation with morphologically constrained 3D deep generative models

We describe CounterSynth, a conditional generative model of diffeomorphic deformations that induce label-driven, biologically plausible changes in volumetric brain images. The model is intended to synthesise counterfactual training data augmentations for downstream discriminative modelling tasks where fidelity is limited by data imbalance, distributional instability, confounding, or underspecification, and exhibits inequitable performance across distinct subpopulations. Focusing on demographic attributes, we evaluate the quality of synthesised counterfactuals with voxel-based morphometry, classification and regression of the conditioning attributes, and the Fréchet inception distance. Examining downstream discriminative performance in the context of engineered demographic imbalance and confounding, we use UK Biobank and OASIS magnetic resonance imaging data to benchmark CounterSynth augmentation against current solutions to these problems. We achieve state-of-the-art improvements, both in overall fidelity and equity. The source code for CounterSynth is available at https://github.com/guilherme-pombo/CounterSynth.

Intention-based evaluations of distributive actions by 4-month-olds

Four-month-olds' ability to consider the intentions of agents performing distributive actions was investigated in four experiments, using the Violation of Expectation paradigm (VoE) (Experiments 1-3) and the Preferential Looking paradigm (Experiment 4). In Experiment 1, infants were presented with two events showing two types of failed attempts to perform a distribution. In an attempt to distribute fairly, the distributor first tried to reach one of the recipients to deliver an apple, he failed, and then attempted to reach the other recipient to deliver a second apple and also failed. In an attempt to distribute unfairly, a different distributor tried unsuccessfully to bring resources always to the same recipient. Infants looked reliably longer at failed fair distribution events, suggesting that they did not just react to the actions outcomes and they attended to agents' intentions. Experiments 2 and 3 assessed alternative explanations based on perceptual factors or affiliative behaviors. In Experiment 4, during the test trials, infants were shown both distributors simultaneously and they preferred to look at the fair rather than at the unfair distributor. Overall, these findings reveal an early ability to take into account distributors' intentions and a preference for watching agents that tried to distribute resources fairly.

Whose public, whose goods? Generations of patients and visions of fairness in Ghanaian health

Since Ghana's independence in 1957, the country has seen an ebb and flow of reforms intended to expand and fund state healthcare, informed by diverse notions of affordability and adequate provision. Cycles of attempted health reforms have emerged from disparate political and economic ideologies, themselves a product of broader global histories and specific national experiences. Based on group interviews with people across most administrative regions of Ghana, this paper examines how the formative historical experiences of different generations gives rise to a multiplicity of understandings of what constitutes a 'fair' distribution of national health resources. It discusses the forms and contents of arguments that people of different ages raised in both rural and urban settings in the course of the study - with particular reference to the operation of Ghana's current National Health Insurance Scheme, and in light of their perceptions of the justice or injustice of present day healthcare in relation to earlier periods.

Evaluating and mitigating bias in machine learning models for cardiovascular disease prediction

OBJECTIVE

The study aims to investigate whether machine learning-based predictive models for cardiovascular disease (CVD) risk assessment show equivalent performance across demographic groups (such as race and gender) and if bias mitigation methods can reduce any bias present in the models. This is important as systematic bias may be introduced when collecting and preprocessing health data, which could affect the performance of the models on certain demographic sub-cohorts. The study is to investigate this using electronic health records data and various machine learning models.

METHODS

The study used large de-identified Electronic Health Records data from Vanderbilt University Medical Center. Machine learning (ML) algorithms including logistic regression, random forest, gradient-boosting trees, and long short-term memory were applied to build multiple predictive models. Model bias and fairness were evaluated using equal opportunity difference (EOD, 0 indicates fairness) and disparate impact (DI, 1 indicates fairness). In our study, we also evaluated the fairness of a non-ML baseline model, the American Heart Association (AHA) Pooled Cohort Risk Equations (PCEs). Moreover, we compared the performance of three different de-biasing methods: removing protected attributes (e.g., race and gender), resampling the imbalanced training dataset by sample size, and resampling by the proportion of people with CVD outcomes.

RESULTS

The study cohort included 109,490 individuals (mean [SD] age 47.4 [14.7] years; 64.5% female; 86.3% White; 13.7% Black). The experimental results suggested that most ML models had smaller EOD and DI than PCEs. For ML models, the mean EOD ranged from -0.001 to 0.018 and the mean DI ranged from 1.037 to 1.094 across race groups. There was a larger EOD and DI across gender groups, with EOD ranging from 0.131 to 0.136 and DI ranging from 1.535 to 1.587. For debiasing methods, removing protected attributes didn't significantly reduced the bias for most ML models. Resampling by sample size also didn't consistently decrease bias. Resampling by case proportion reduced the EOD and DI for gender groups but slightly reduced accuracy in many cases.

CONCLUSIONS

Among the VUMC cohort, both PCEs and ML models were biased against women, suggesting the need to investigate and correct gender disparities in CVD risk prediction. Resampling by proportion reduced the bias for gender groups but not for race groups.

Fair detection of poisoning attacks in federated learning on non-i.i.d. data

Reconciling machine learning with individual privacy is one of the main motivations behind federated learning (FL), a decentralized machine learning technique that aggregates partial models trained by clients on their own private data to obtain a global deep learning model. Even if FL provides stronger privacy guarantees to the participating clients than centralized learning collecting the clients' data in a central server, FL is vulnerable to some attacks whereby malicious clients submit bad updates in order to prevent the model from converging or, more subtly, to introduce artificial bias in the classification (poisoning). Poisoning detection techniques compute statistics on the updates to identify malicious clients. A downside of anti-poisoning techniques is that they might lead to discriminate minority groups whose data are significantly and legitimately different from those of the majority of clients. This would not only be unfair, but would yield poorer models that would fail to capture the knowledge in the training data, especially when data are not independent and identically distributed (non-i.i.d.). In this work, we strive to strike a balance between fighting poisoning and accommodating diversity to help learning fairer and less discriminatory federated learning models. In this way, we forestall the exclusion of diverse clients while still ensuring detection of poisoning attacks. Empirical work on three data sets shows that employing our approach to tell legitimate from malicious updates produces models that are more accurate than those obtained with state-of-the-art poisoning detection techniques. Additionally, we explore the impact of our proposal on the performance of models on non-i.i.d local training data.