Why do probabilistic clinical models fail to transport between sites.

The rising popularity of artificial intelligence in healthcare is highlighting the problem that a computational model achieving super-human clinical performance at its training sites may perform substantially worse at new sites. In this perspective, we argue that we should typically expect this failure to transport, and we present common sources for it, divided into those under the control of the experimenter and those inherent to the clinical data-generating process. Of the inherent sources we look a little deeper into site-specific clinical practices that can affect the data distribution, and propose a potential solution intended to isolate the imprint of those practices on the data from the patterns of disease cause and effect that are the usual target of probabilistic clinical models.

Mitigating pathogenesis for target discovery and disease subtyping.

Treatments ideally mitigate pathogenesis, or the detrimental effects of the root causes of disease. However, existing definitions of treatment effect fail to account for pathogenic mechanism. We therefore introduce the Treated Root causal Effects (TRE) metric which measures the ability of a treatment to modify root causal effects. We leverage TREs to automatically identify treatment targets and cluster patients who respond similarly to treatment. The proposed algorithm learns a partially linear causal model to extract the root causal effects of each variable and then estimates TREs for target discovery and downstream subtyping. We maintain interpretability even without assuming an invertible structural equation model. Experiments across a range of datasets corroborate the generality of the proposed approach.

Counterfactual formulation of patient-specific root causes of disease.

OBJECTIVE: Root causes of disease intuitively correspond to root vertices of a causal model that increase the likelihood of a diagnosis. This description of a root cause nevertheless lacks the rigorous mathematical formulation needed for the development of computer algorithms designed to automatically detect root causes from data. We seek a definition of patient-specific root causes of disease that models the intuitive procedure routinely utilized by physicians to uncover root causes in the clinic. METHODS: We use structural equation models, interventional counterfactuals and the recently developed mathematical formalization of backtracking counterfactuals to propose a counterfactual formulation of patient-specific root causes of disease matching clinical intuition. RESULTS: We introduce a definition of patient-specific root causes of disease that climbs to the third rung of Pearl's Ladder of Causation and matches clinical intuition given factual patient data and a working causal model. We then show how to assign a root causal contribution score to each variable using Shapley values from explainable artificial intelligence. CONCLUSION: The proposed counterfactual formulation of patient-specific root causes of disease accounts for noisy labels, adapts to disease prevalence and admits fast computation without the need for counterfactual simulation.

Discovering Root Causal Genes with High Throughput Perturbations.

Root causal gene expression levels - or root causal genes for short - correspond to the initial changes to gene expression that generate patient symptoms as a downstream effect. Identifying root causal genes is critical towards developing treatments that modify disease near its onset, but no existing algorithms attempt to identify root causal genes from data. RNA-sequencing (RNA-seq) data introduces challenges such as measurement error, high dimensionality and non-linearity that compromise accurate estimation of root causal effects even with state-of-the-art approaches. We therefore instead leverage Perturb-seq, or high throughput perturbations with single cell RNA-seq readout, to learn the causal order between the genes. We then transfer the causal order to bulk RNA-seq and identify root causal genes specific to a given patient for the first time using a novel statistic. Experiments demonstrate large improvements in performance. Applications to macular degeneration and multiple sclerosis also reveal root causal genes that lie on known pathogenic pathways, delineate patient subgroups and implicate a newly defined omnigenic root causal model.

Increasing countries' financial resilience through global catastrophe risk pooling.

Extreme weather events can severely impact national economies, leading the recovery of low- to middle-income countries to become reliant on foreign financial aid. Foreign aid is, however, slow and uncertain. Therefore, the Sendai Framework and the Paris Agreement advocate for more resilient financial instruments like sovereign catastrophe risk pools. Existing pools, however, might not fully exploit their financial resilience potential because they were not designed to maximize risk diversification and because they pool risk only regionally. Here we introduce a method that forms pools by maximizing risk diversification and apply it to assess the benefits of global pooling compared to regional pooling. We find that global pooling always provides a higher risk diversification, it better distributes countries' risk shares in the pool's risk and it increases the number of countries profiting from risk pooling. Optimal global pooling could provide a diversification increase to existing pools of up to 65 %.

Cross-country risk quantification of extreme wildfires in Mediterranean Europe.

We estimate the country-level risk of extreme wildfires defined by burned area (BA) for Mediterranean Europe and carry out a cross-country comparison. To this end, we avail of the European Forest Fire Information System (EFFIS) geospatial data from 2006 to 2019 to perform an extreme value analysis. More specifically, we apply a point process characterization of wildfire extremes using maximum likelihood estimation. By modeling covariates, we also evaluate potential trends and correlations with commonly known factors that drive or affect wildfire occurrence, such as the Fire Weather Index as a proxy for meteorological conditions, population density, land cover type, and seasonality. We find that the highest risk of extreme wildfires is in Portugal (PT), followed by Greece (GR), Spain (ES), and Italy (IT) with a 10-year BA return level of 50'338 ha, 33'242 ha, 25'165 ha, and 8'966 ha, respectively. Coupling our results with existing estimates of the monetary impact of large wildfires suggests expected losses of 162-439 million € (PT), 81-219 million € (ES), 41-290 million € (GR), and 18-78 million € (IT) for such 10-year return period events. SUMMARY: We model the risk of extreme wildfires for Italy, Greece, Portugal, and Spain in form of burned area return levels, compare them, and estimate expected losses.

Factors Associated With Being Overweight or Obese in Suriname.

Objective: To identify the socio-demographic risk factors that are associated with adult Body Mass Index. Methods: We apply probit and ordinal probit models to a sample of 3,803 adults aged 20 and above from the 2016/17 round of the Suriname Survey of Living Conditions. Results: Women, the elderly, and couples who are either married and/or living together are more likely to be obese or overweight. This is also true for individuals who have chronic illnesses. We also find that individuals who engage in a sport or in other forms of exercise, even if modest, have lower odds of being overweight or obese. Interestingly, our findings indicate that individuals who benefit from government social safety net programs are less likely to be associated with being overweight or obese. Conclusion: Obesity could become a serious public health issue if not addressed appropriately. Policymakers should promptly develop a national strategy to help health care systems cope with the outcomes of obesity and to tackle the risk factors that have the greatest impacts on individual Body Mass Index.

The role of weather conditions in COVID-19 transmission: A study of a global panel of 1236 regions.

It is believed that weather conditions such as temperature and humidity have effects on COVID-19 transmission. However, these effects are not clear due to the limited observations and difficulties in separating impact of social distancing. COVID-19 data and social-economic features of 1236 regions in the world (1112 regions at the provincial level and 124 countries with the small land area) were collected. Large-scale satellite data was combined with these data with a regression analysis model to explore the effects of temperature and relative humidity on COVID-19 spreading, as well as the possible transmission risk by seasonal cycles. The result shows that temperature and relative humidity are negatively correlated with COVID-19 transmission throughout the world. Government intervention (e.g. lockdown policies) and lower population movement contributed to decrease the new daily case ratio. Weather conditions are not the decisive factor in COVID-19 transmission, in that government intervention as well as public awareness, could contribute to the mitigation of the spreading of the virus. So, it deserves a dynamic government policy to mitigate COVID-19 transmission in winter.

Can we mitigate the effect of natural disasters on child health? Evidence from the Indian Ocean tsunami in Indonesia.

The 2004 Indian Ocean tsunami was an international natural disaster unlike any seen before, killing 166,561 people in Aceh province, Indonesia. It prompted an unprecedented humanitarian response and was a catalyst in ending almost 30 years of civil conflict in Aceh. Since the tsunami was followed by a multitude of events, we first conduct a systematic review to identify those events in Indonesia. We then use a synthetic control method to estimate the combination of those effects on child mortality indicators in Aceh for the 13 years that followed the disaster using data from 258,918 children born between 1990 and 2017. The results show a significant increase in under-5 mortality only the year after the tsunami and no effect in the medium term. However, younger and older children were affected differently in the medium term. In fact, we show a decrease in child mortality among children aged 1-4 years. In contrast, we observe an increase in mortality among children under-1 in 2009 and 2010. Overall, the resilience of Aceh province points to the importance of coordinated international disaster responses after natural disasters.

Air Pollution Exposure and Covid-19 in Dutch Municipalities.

In light of the existing preliminary evidence of a link between Covid-19 and poor air quality, which is largely based upon correlations, we estimate the relationship between long term air pollution exposure and Covid-19 in 355 municipalities in the Netherlands. Using detailed data we find compelling evidence of a positive relationship between air pollution, and particularly P M 2.5 concentrations, and Covid-19 cases, hospital admissions and deaths. This relationship persists even after controlling for a wide range of explanatory variables. Our results indicate that, other things being equal, a municipality with 1 µg/m3 more P M 2.5 concentrations will have 9.4 more Covid-19 cases, 3.0 more hospital admissions, and 2.3 more deaths. This relationship between Covid-19 and air pollution withstands a number of sensitivity and robustness exercises including instrumenting pollution to mitigate potential endogeneity in the measurement of pollution and modelling spatial spillovers using spatial econometric techniques.

The impact of decriminalization on marijuana and alcohol consumption in Jamaica.

We examine whether marijuana decriminalization in Jamaica, a country that historically has had relatively widespread use of the drug, has led to an increase in its use, the frequency of use and the money spent on it. To this end, we use a national drug survey dataset with extensive information on people's use of, attitudes towards, access to marijuana. Our econometric analysis shows that awareness of the legislation has a positive correlation with the use of the substance. Worryingly, decriminalization positively correlates with the likelihood of first time and general use for youths. There is also some evidence that the legislation results in a substitution away from alcohol towards marijuana consumption for youths. From a policy perspective, a marijuana monitoring system can be implemented to follow the consumption patterns of youths. This should involve establishing school-level programmes that monitor students, and where potential drug users are identified, school officials should intervene to curb students' drug appetite before an escalated use of marijuana.

Fast Causal Inference with Non-Random Missingness by Test-Wise Deletion.

Many real datasets contain values missing not at random (MNAR). In this scenario, investigators often perform list-wise deletion, or delete samples with any missing values, before applying causal discovery algorithms. List-wise deletion is a sound and general strategy when paired with algorithms such as FCI and RFCI, but the deletion procedure also eliminates otherwise good samples that contain only a few missing values. In this report, we show that we can more efficiently utilize the observed values with test-wise deletion while still maintaining algorithmic soundness. Here, test-wise deletion refers to the process of list-wise deleting samples only among the variables required for each conditional independence (CI) test used in constraint-based searches. Test-wise deletion therefore often saves more samples than list-wise deletion for each CI test, especially when we have a sparse underlying graph. Our theoretical results show that test-wise deletion is sound under the justifiable assumption that none of the missingness mechanisms causally affect each other in the underlying causal graph. We also find that FCI and RFCI with test-wise deletion outperform their list-wise deletion and imputation counterparts on average when MNAR holds in both synthetic and real data.

Markov Boundary Discovery with Ridge Regularized Linear Models.

Ridge regularized linear models (RRLMs), such as ridge regression and the SVM, are a popular group of methods that are used in conjunction with coefficient hypothesis testing to discover explanatory variables with a significant multivariate association to a response. However, many investigators are reluctant to draw causal interpretations of the selected variables due to the incomplete knowledge of the capabilities of RRLMs in causal inference. Under reasonable assumptions, we show that a modified form of RRLMs can get "very close" to identifying a subset of the Markov boundary by providing a worst-case bound on the space of possible solutions. The results hold for any convex loss, even when the underlying functional relationship is nonlinear, and the solution is not unique. Our approach combines ideas in Markov boundary and sufficient dimension reduction theory. Experimental results show that the modified RRLMs are competitive against state-of-the-art algorithms in discovering part of the Markov boundary from gene expression data.

Impaired Recognition and Regulation of Disgust Is Associated with Distinct but Partially Overlapping Patterns of Decreased Gray Matter Volume in the Ventroanterior Insula.

BACKGROUND: The ventroanterior insula is implicated in the experience, expression, and recognition of disgust; however, whether this brain region is required for recognizing disgust or regulating disgusting behaviors remains unknown. METHODS: We examined the brain correlates of the presence of disgusting behavior and impaired recognition of disgust using voxel-based morphometry in a sample of 305 patients with heterogeneous patterns of neurodegeneration. Permutation-based analyses were used to determine regions of decreased gray matter volume at a significance level p <= .05 corrected for family-wise error across the whole brain and within the insula. RESULTS: Patients with behavioral variant frontotemporal dementia and semantic variant primary progressive aphasia were most likely to exhibit disgusting behaviors and were, on average, the most impaired at recognizing disgust in others. Imaging analysis revealed that patients who exhibited disgusting behaviors had significantly less gray matter volume bilaterally in the ventral anterior insula. A region of interest analysis restricted to behavioral variant frontotemporal dementia and semantic variant primary progressive aphasia patients alone confirmed this result. Moreover, impaired recognition of disgust was associated with decreased gray matter volume in the bilateral ventroanterior and ventral middle regions of the insula. There was an area of overlap in the bilateral anterior insula where decreased gray matter volume was associated with both the presence of disgusting behavior and impairments in recognizing disgust. CONCLUSIONS: These findings suggest that regulating disgusting behaviors and recognizing disgust in others involve two partially overlapping neural systems within the insula. Moreover, the ventral anterior insula is required for both processes.

Predicting the risk of psychosis onset: advances and prospects.

AIM: To conduct a systematic review of the methods and performance characteristics of models developed for predicting the onset of psychosis. METHODS: We performed a comprehensive literature search restricted to English articles and identified using PubMed, Medline and PsychINFO, as well as the reference lists of published studies and reviews. Inclusion criteria included the selection of more than one variable to predict psychosis or schizophrenia onset, and selection of individuals at familial risk or clinical high risk. Eighteen studies met these criteria, and we compared these studies based on the subjects selected, predictor variables used and the choice of statistical or machine learning methods. RESULTS: Quality of life and life functioning as well as structural brain imaging emerged as the most promising predictors of psychosis onset, particularly when they were coupled with appropriate dimensionality reduction methods and predictive model algorithms like the support vector machine (SVM). Balanced accuracy ranged from 100% to 78% in four studies using the SVM, and 67% to 81% in 14 studies using general linear models. CONCLUSIONS: Performance of the predictive models improves with quality of life measures, life functioning measures, structural brain imaging data, as well as with the use of methods like SVM. Despite these advances, the overall performance of psychosis predictive models is still modest. In the future, performance can potentially be improved by including genetic variant and new functional imaging data in addition to the predictors that are used currently.

BDNF serum concentrations show no relationship with diagnostic group or medication status in neurodegenerative disease.

Brain-derived neurotrophic factor (BDNF) is a growth factor implicated in neuronal survival. Studies have reported altered BDNF serum concentrations in patients with Alzheimer's disease (AD). However, these studies have been inconsistent. Few studies have investigated BDNF concentrations across multiple neurodegenerative diseases, and no studies have investigated BDNF concentrations in patients with frontotemporal dementia. To examine BDNF concentrations in different neurodegenerative diseases, we measured serum concentrations of BDNF using enzyme-linked immunoassay in subjects with behavioral-variant frontotemporal dementia (bvFTD, n=20), semantic dementia (SemD, n=16), AD (n=34), and mild cognitive impairment (MCI, n=30), as well as healthy older subjects (HS, n=38). BDNF serum concentrations were compared across diagnoses and correlated with cognitive tests and patterns of brain atrophy using voxelbased morphometry. We found small negative correlations between BDNF serum concentrations and some of the cognitive tests assessing learning, information processing speed and cognitive control in complex situations, however, BDNF did not predict disease group membership despite adequate power. These findings suggest that BDNF serum concentration may not be a reliable diagnostic biomarker to distinguish among neurodegenerative diseases.

Association between skin surface pH, temperature and Staphylococcus pseudintermedius in dogs with immunomodulatory-responsive lymphocytic-plasmacytic pododermatitis.

Secondary bacterial infection is a frequent complication in lesional skin of dogs with immunomodulatory-responsive lymphocytic-plasmacytic pododermatitis (ImR-LPP). However, the influence of skin pH and temperature in determining the composition of the cutaneous microflora at lesional sites has not been investigated. The association between ImR-LPP and pedal skin temperature, pH and Staphylococcus pseudintermedius isolates was thus evaluated. Temperature and pH were measured in 20 dogs with ImR-LPP and in 30 clinically healthy control dogs, and S. pseudintermedius was cultured from interdigital and palmoplantar swabs in both groups and scored semi-quantitatively for bacterial growth. In the ImR-LPP group, mean skin pH was slightly, but significantly, higher at both interdigital and palmoplantar sites. Staphylococcus pseudintermedius was isolated more frequently, and scores for bacterial growth were also significantly higher. However, mean skin temperatures were not significantly different from those in the control group. The isolation of S. pseudintermedius was significantly associated with ImR-LPP, with the single exception of isolates on Columbia blood agar from the palmoplantar region. However, pH and temperature were not significantly associated with the disease, and were not associated with the isolation of S. pseudintermedius at most sites sampled. Staphylococcus pseudintermedius was not isolated from all feet sampled in dogs with ImR-LPP. Taken together, these data would suggest that S. pseudintermedius infection is most likely to be a secondary phenomenon in dogs with ImR-LPP, and that changes in skin pH and temperature are not significant risk factors for this disease.

Cutaneous infiltrates and peripheral blood immune responses in dogs with immunomodulatory-responsive lymphocytic-plasmacytic pododermatitis.

This study characterizes T- and B-lymphocyte responses in the peripheral blood and lesional skin of dogs with immunomodulatory-responsive lymphocytic-plasmacytic pododermatitis (ImR-LPP), a term previously proposed to denote a subpopulation of dogs with idiopathic pododermatitis. T-cell (CD3+, CD4+ and CD8+ ) and B-cell (CD21+) counts were significantly increased in both the epidermis and dermis of lesional ImR-LPP skin compared with that in pedal skin from healthy controls. CD3+ , CD4+, CD8+ and CD21+ cells were commonly observed in perivascular sites in the superficial dermis, periadnexally, beneath the dermal-epidermal (DE) junction and in the epidermis of lesional ImR-LPP skin. The CD8+ /CD3+ T-cell ratio in peripheral blood was significantly increased in the ImR-LPP group (0.42 versus 0.35 in controls). Serum IgA, IgG and IgM concentrations were all significantly elevated in affected dogs. Lymphocyte stimulation indices in ImR-LPP dogs were comparable with control levels except for a lower response to ionomycin (6.0 versus 11.1). Dogs with ImR-LPP had a higher incidence and mean (semi-quantitative) score for IgA, IgG and IgM deposits in the epidermis, and a significantly increased incidence of dermal IgA+, IgG+ and IgM+ mononuclear inflammatory cells. The results indicate that upregulated T- and B-lymphocyte responses may contribute to the pathogenesis of the skin lesions observed in dogs with ImR-LPP.