Risk Modeling Research Articles

Advancing credit risk modelling with Machine Learning: A comprehensive review of the state-of-the-art

Ensuring financial stability necessitates responsible credit granting, so lending institutions maintain sufficient regulatory capital to withstand losses from defaults. Classification methods like Logistic Regression (LR) have long been the standard for estimating default likelihood, but they struggle with increasing operational demands and regulatory standards. Advances in Machine Learning (ML) now allow models to handle larger datasets with greater predictive power. Algorithms such as Decision Trees (DT), Support Vector Machines (SVM), and Neural Networks (NN), though not readily transferable to practical applications, have the potential to enhance the credit risk modelling process. While much of the literature on ML in credit risk focuses on sophisticated, experimental algorithms not yet ready for commercial use, few address the steps and nuances of model development. Our article fills this gap by systematically reviewing the literature on ML in consumer credit risk modelling, highlighting the current state of scientific knowledge. We analyze the frequent steps involved and demonstrate how ML's strengths can be leveraged throughout them. From pre-processing steps like data cleaning, transformation, and reduction, to variable selection and parameter optimization, our findings indicate that ML is becoming increasingly important in financial risk management. However, standardized modelling procedures are needed before complex ML classifiers can be commercially deployed. Meanwhile, although relatively interpretable ML models built around DTs and Random Forests (RF) do showcase increased performance in popular credit risk datasets, it is clear that the potential to benefit from ML spans further, with complex models enabling increased performance by aiding simpler, explainable classification algorithms.

Environmental Risk Assessment and Modeling of TPH and PAHS in OBRIKOM 15 Access Road Mini Manifold Spill

From the anthropogenesis operations of petroleum hydrocarbon related activities in the Niger Delta region of Nigeria, the possibilities of environmental crisis and associated hazards is highly expected. BETX and PAHs as major constituents of petroleum hydrocarbon and their impacts on the environment had caused ecological imbalance with contamination/pollution resulting from crude oil spillages. Random sampling method was adopted and soil samples were collected using amber wide-mouth and vial glass bottles with the aid of soil auger at 0-4m depths and analysed using GC-FID, GC-MSD and APHA analytical techniques. The TPH and PAHs decreased with depth. The TPH measured values were up to 1,636.85mg/kg and PAHs as high as 5,702.21mg/kg in soil with below detectable level (BDL) of BTEX. CF values were between 0.059 to 0.327 and the DC value was 2.511 respectively, made Obrikom spill site soil lowly contaminated while the l-geo calculation values of PAHs was between 3.615 to 6.571 indicating strongly contaminated. The hypothesis testing calculation shows that the calculated value is less than the table value and therefore the hypothesis is upheld which states that there is no relationship between TPH and PAHs concentrations in Obrikom soil impacted site. The contour maps show decreasing values of TPH and PAHs from the point of spill and depth while variogram models revealed a continuous behavior having similar trends of identical exponential variograms.

Forest fire risk modeling in Mediterranean forests using GIS and AHP method: case of the high Rif forest massif (Morocco)

Forest fire risk modeling in Mediterranean forests using GIS and AHP method: case of the high Rif forest massif (Morocco)

An augmented illness-death model for semi-competing risks with clinically immediate terminal events.

Preeclampsia is a pregnancy-associated condition posing risks of both fetal and maternal mortality and morbidity that can only resolve following delivery and removal of the placenta. Because in its typical form preeclampsia can arise before delivery, but not after, these two events exemplify the time-to-event setting of "semi-competing risks" in which a non-terminal event of interest is subject to the occurrence of a terminal event of interest. The semi-competing risks framework presents a valuable opportunity to simultaneously address two clinically meaningful risk modeling tasks: (i) characterizing risk of developing preeclampsia, and (ii) characterizing time to delivery after onset of preeclampsia. However, some people with preeclampsia deliver immediately upon diagnosis, while others are admitted and monitored for an extended period before giving birth, resulting in two distinct trajectories following the non-terminal event, which we call "clinically immediate" and "non-immediate" terminal events. Though such phenomena arise in many clinical contexts, to-date there have not been methods developed to acknowledge the complex dependencies between such outcomes, nor leverage these phenomena to gain new insight into individualized risk. We address this gap by proposing a novel augmented frailty-based illness-death model with a binary submodel to distinguish risk of immediate terminal event following the non-terminal event. The model admits direct dependence of the terminal event on the non-terminal event through flexible regression specification, as well as indirect dependence via a shared frailty term linking each submodel. We develop an efficient Bayesian sampler for estimation and corresponding model fit metrics, and derive formulae for dynamic risk prediction. In an extended example using pregnancy outcome data from an electronic health record, we demonstrate the proposed model's direct applicability to address a broad range of clinical questions.

High-resolution 3D ambient noise tomography around the Meishan-Chiayi active fault system of western Taiwan

The Meishan-Chiayi area of western Taiwan has a large probability of producing a major earthquake in the near future. Historically, one of the largest and most damaging of Taiwan’s earthquakes occurred there. It is, therefore, important to have a well-constrained upper crustal 3-D shear-wave velocity model that can be used to accurately determine ground motion predictions and fault geometry models used in seismic hazard and risk modelling. In this study, we carried out an ambient noise tomography experiment using 100 seismometers deployed with a ∼2 km spacing on a 20 by 20 km grid. The reliable periods of phase velocity from Rayleigh waves are 0.6 to 6.8 s, providing a well-resolved Vs structure from the surface to a depth of around 4 km. The velocity model displays a prominent, roughly northeast-striking change in Vs that follows the projected surface trace of the blind Chiayi thrust. The uplift of relatively higher Vs rocks in its hanging wall, together with a negative to positive change in dVs suggests that it dips gently eastward across the study area. A northward thickening of the lower Vs crust, together with a high negative dVs in the north of the study area is related to an increased thickness of foreland basin rocks across the Meishan fault. The Vs and dVs models provide reasonable evidence that the Meishan fault can be traced at a high angle from its surface rupture to the base of the model at 4 km depth. It cuts the Chiayi thrust.

A semi-quantitative multi-hazard risk assessment framework for European coastal urban areas

Coastal cities are increasingly exposed to a wide range of hydro-meteorological and climate-related hazards and their negative consequences. Developing disaster risk assessments to support risk management activities in these areas is therefore a critically important but challenging task, due to the diversity of processes involved and associated data requirements. Semi-quantitative, i.e. index-based assessments can be instrumental in addressing these issues: by trading off some detail in the representation of risk for simplicity in the underlying data and methods, it becomes possible to assess, compare, and easily communicate risk for various sectors, hazards, and locations. In this context, the present article presents a new semi-quantitative, GIS-based multi-hazard risk assessment framework for coastal urban areas. The framework uses various large-scale datasets and state-of-the-art risk modelling concepts to provide risk scores for six perils in European coastal cities: fluvial flooding, coastal flooding, extreme precipitation, landslide, heat wave, and coastal erosion. An application is presented for the risk assessment of ten case-study areas, illustrating the potential of the proposed framework for identifying the most pressing risks for each area and supporting the prioritization of more detailed quantitative risk assessments for specific cities, sectors and/or hazards, based on which risk mitigation measures can be defined.

A comprehensive approach for assessing social flood vulnerability and social flood risk: The case of Denmark

Flood risk assessment approaches have traditionally been dominated by measures of economic damage. However, the importance of understanding the social impacts of flooding are increasingly being acknowledged. Social vulnerability indices have been constructed in diverse geographical contexts to understand relative susceptibility of different social groups to flood hazards. However, integrated assessments of social vulnerability, exposure, and hazard information are lacking. Here, we construct a national social vulnerability index (SVI) for Denmark and combine this with direct and indirect social exposure data and coastal flood hazard data to construct a national social flood risk index (SFRI). Results show the spatial distribution of social flood vulnerability and social flood risk in Denmark. Our findings illustrate that including social data in flood risk assessment could significantly change our understanding of flood risk on a national scale. Methodologically, our work introduces a comprehensive flood risk modelling approach that explicitly considers the social impacts of flooding in all model components. The application of this model in Denmark reveals that the social impacts of flooding extend far beyond flooded areas, thus highlighting the importance of explicitly considering direct and indirect social exposure in addition to social vulnerability in flood risk assessment. By introducing a comprehensive, socially specific approach to flood risk assessment that is usable within existing risk management frameworks such as the EU Floods Directive, our work aims to mainstream social wellbeing, resilience, and justice as central considerations in decision making on flood risk management.

Comparing components for seismic risk modelling using data from the 2019 Le Teil (France) earthquake

Abstract. Probabilistic seismic hazard and risk models are essential to improving our awareness of seismic risk, to its management, and to increasing our resilience against earthquake disasters. These models consist of a series of components, which may be evaluated and validated individually, although evaluating and validating these types of models as a whole is challenging due to the lack of recognized procedures. Estimations made with other models, as well as observations of damage from past earthquakes, lend themselves to evaluating the components used to estimate the severity of damage to buildings. Here, we are using a dataset based on emergency post-seismic assessments made after the Le Teil 2019 earthquake, third-party estimations of macroseismic intensity for this seismic event, shake maps, and scenario damage calculations to compare estimations under different modelling assumptions. First we select a rupture model using estimations of ground motion intensity measures and macroseismic intensity. Subsequently, we use scenario damage calculations based on different exposure models, including the aggregated exposure model in the 2020 European Seismic Risk Model (ESRM20), as well as different site models. Moreover, a building-by-building exposure model is used in scenario calculations, which individually models the buildings in the dataset. Lastly, we compare the results of a semi-empirical approach to the estimations made with the scenario calculations. The post-seismic assessments are converted to EMS-98 (Grünthal, 1998) damage grades and then used to estimate the damage for the entirety of the building stock in Le Teil. In general, the scenario calculations estimate lower probabilities for damage grades 3–4 than the estimations made using the emergency post-seismic assessments. An exposure and fragility model assembled herein leads to probabilities for damage grades 3–5 with small differences from the probabilities based on the ESRM20 exposure and fragility model, while the semi-empirical approach leads to lower probabilities. The comparisons in this paper also help us learn lessons on how to improve future testing. An improvement would be the use of damage observations collected directly on the EMS-98 scale or on the damage scale in ESRM20. Advances in testing may also be made by employing methods that inform us about the damage at the scale of a city, such as remote sensing or data-driven learning methods fed by a large number of low-cost seismological instruments spread over the building stock.

Health map for HealthGap: Defining a geographical catchment to examine cardiovascular risk in Victoria, Australia

The HealthGap study aimed to understand cardiovascular risk among Indigenous Australians in Victoria using linked administrative data. A key challenge was differing spatial coverages of sources: state-level data for risk factors but cardiovascular outcomes for three hospitals. Catchments were defined based on hospital postcodes to estimate denominator populations for risk modelling: first- and second-order neighbours, and spatial distribution of outcomes (‘spatial event distribution’). Catchment coverage was assessed through proportions of patients presenting to study hospitals from catchment postcodes. The spatial event distribution performed best, capturing 82% events overall (first-order:40%; second-order:64%) and 65% Indigenous (27% and 45%). No approach excluded proximal non-study hospitals. Spatial event distributions could help define denominator populations when geographic information on outcome data is available but may not avoid potential misclassification.

Evaluating erosion risk models in a Scottish catchment using organic carbon fingerprinting

PurposeIdentification of hotspots of accelerated erosion of soil and organic carbon (OC) is critical to the targeting of soil conservation and sediment management measures. The erosion risk map (ERM) developed by Lilly and Baggaley (Soil erosion risk map of Scotland, 2018) for Scotland estimates erosion risk for the specific soil conditions in the region. However, the ERM provides no soil erosion rates. Erosion rates can be estimated by empirical models such as the Revised Universal Soil Loss Equation (RUSLE). Yet, RUSLE was not developed specifically for the soil conditions in Scotland. Therefore, we evaluated the performance of these two erosion models to determine whether RUSLE erosion rate estimates could be used to quantify the amount of soil eroded from high-risk areas identified in the ERM.MethodsThe study was conducted in the catchment of Loch Davan, Aberdeenshire, Scotland. Organic carbon loss models were constructed to compare land use specific OC yields based on RUSLE and ERM using OC fingerprinting as a benchmark. The estimated soil erosion rates in this study were also compared with recently published estimates in Scotland (Rickson et al. in Developing a method to estimate the costs of soil erosion in high-risk Scottish catchments, 2019).ResultsThe region-specific ERM most closely approximated the relative land use OC yields in streambed sediment however, the results of RUSLE were very similar, suggesting that, in this catchment, RUSLE erosion rate estimates could be used to quantify the amount of soil eroded from the high-risk areas identified by ERM. The RUSLE estimates of soil erosion for this catchment were comparable to the soil erosion rates per land use estimated by Rickson et al. (Developing a method to estimate the costs of soil erosion in high-risk Scottish catchments, 2019) in Scottish soils except in the case of pasture/grassland likely due to the pastures in this catchment being grass ley where periods of surface vegetation cover/root network absence are likely to have generated higher rates of erosion.ConclusionSelection of suitable erosion risk models can be improved by the combined use of two sediment origin techniques—erosion risk modelling and OC sediment fingerprinting. These methods could, ultimately, support the development of targeted sediment management strategies to maintain healthy soils within the EU and beyond.

Predictive risk modeling for child maltreatment detection and enhanced decision-making: Evidence from Danish administrative data.

Child maltreatment is a widespread problem with significant costs for both victims and society. In this retrospective cohort study, we develop predictive risk models using Danish administrative data to predict removal decisions among referred children and assess the effectiveness of caseworkers in identifying children at risk of maltreatment. The study analyzes 195,639 referrals involving 102,309 children Danish Child Protection Services received from April 2016 to December 2017. We implement four machine learning models of increasing complexity, incorporating extensive background information on each child and their family. Our best-performing model exhibits robust predictive power, with an AUC-ROC score exceeding 87%, indicating its ability to consistently rank referred children based on their likelihood of being removed. Additionally, we find strong positive correlations between the model's predictions and various adverse child outcomes, such as crime, physical and mental health issues, and school absenteeism. Furthermore, we demonstrate that predictive risk models can enhance caseworkers' decision-making processes by reducing classification errors and identifying at-risk children at an earlier stage, enabling timely interventions and potentially improving outcomes for vulnerable children.

Fluvial flood inundation and socio-economic impact model based on open data

Abstract. Fluvial floods are destructive hazards that affect millions of people worldwide each year. Forecasting flood events and their potential impacts therefore is crucial for disaster preparation and mitigation. Modeling flood inundation based on extreme value analysis of river discharges is an alternative to physical models of flood dynamics, which are computationally expensive. We present the implementation of a globally applicable, open-source fluvial flood model within a state-of-the-art risk modeling framework. It uses openly available data to rapidly compute flood inundation footprints of historic and forecasted events for the estimation of associated impacts. For the example of Pakistan, we use this flood model to compute flood depths and extents and employ it to estimate population displacement due to floods. Comparing flood extents to satellite data reveals that incorporating estimated flood protection standards does not necessarily improve the flood footprint computed by the model. We further show that, after calibrating the vulnerability of the impact model to a single event, the estimated displacement caused by past floods is in good agreement with disaster reports. Finally, we demonstrate that this calibrated model is suited for probabilistic impact-based forecasting.

Self-Explainable Graph Neural Network for Alzheimer Disease and Related Dementias Risk Prediction: Algorithm Development and Validation Study.

Alzheimer disease and related dementias (ADRD) rank as the sixth leading cause of death in the United States, underlining the importance of accurate ADRD risk prediction. While recent advancements in ADRD risk prediction have primarily relied on imaging analysis, not all patients undergo medical imaging before an ADRD diagnosis. Merging machine learning with claims data can reveal additional risk factors and uncover interconnections among diverse medical codes. The study aims to use graph neural networks (GNNs) with claim data for ADRD risk prediction. Addressing the lack of human-interpretable reasons behind these predictions, we introduce an innovative, self-explainable method to evaluate relationship importance and its influence on ADRD risk prediction. We used a variationally regularized encoder-decoder GNN (variational GNN [VGNN]) integrated with our proposed relation importance method for estimating ADRD likelihood. This self-explainable method can provide a feature-important explanation in the context of ADRD risk prediction, leveraging relational information within a graph. Three scenarios with 1-year, 2-year, and 3-year prediction windows were created to assess the model's efficiency, respectively. Random forest (RF) and light gradient boost machine (LGBM) were used as baselines. By using this method, we further clarify the key relationships for ADRD risk prediction. In scenario 1, the VGNN model showed area under the receiver operating characteristic (AUROC) scores of 0.7272 and 0.7480 for the small subset and the matched cohort data set. It outperforms RF and LGBM by 10.6% and 9.1%, respectively, on average. In scenario 2, it achieved AUROC scores of 0.7125 and 0.7281, surpassing the other models by 10.5% and 8.9%, respectively. Similarly, in scenario 3, AUROC scores of 0.7001 and 0.7187 were obtained, exceeding 10.1% and 8.5% than the baseline models, respectively. These results clearly demonstrate the significant superiority of the graph-based approach over the tree-based models (RF and LGBM) in predicting ADRD. Furthermore, the integration of the VGNN model and our relation importance interpretation could provide valuable insight into paired factors that may contribute to or delay ADRD progression. Using our innovative self-explainable method with claims data enhances ADRD risk prediction and provides insights into the impact of interconnected medical code relationships. This methodology not only enables ADRD risk modeling but also shows potential for other image analysis predictions using claims data.

Subgroup analyses and heterogeneity of treatment effects in randomized trials: a primer for the clinician.

To date, most randomized clinical trials in critical care report neutral overall results. However, research as to whether heterogenous responses underlie these results and give opportunity for personalized care is gaining momentum but has yet to inform clinical practice guidance. Thus, we aim to provide an overview of methodological approaches to estimating heterogeneity of treatment effects in randomized trials and conjecture about future paths to application in patient care. Despite their limitations, traditional subgroup analyses are still the most reported approach. More recent methods based on subphenotyping, risk modeling and effect modeling are still uncommonly embedded in primary reports of clinical trials but have provided useful insights in secondary analyses. However, further simulation studies and subsequent guidelines are needed to ascertain the most efficient and robust manner to validate these results for eventual use in practice. There is an increasing interest in approaches that can identify heterogeneity in treatment effects from randomized clinical trials, extending beyond traditional subgroup analyses. While prospective validation in further studies is still needed, these approaches are promising tools for design, interpretation, and implementation of clinical trial results.

Spatiotemporal dynamics of epidemiology diseases: mobility based risk and short-term prediction modeling of COVID-19.

Nowadays, epidemiological modeling is applied to a wide range of diseases, communicable and non-communicable, namely AIDS, Ebola, influenza, Dengue, Malaria, Zika. More recently, in the context of the last pandemic declared by the World Health Organization (WHO), several studies applied these models to SARS-CoV-2. Despite the increasing number of researches using spatial analysis, some constraints persist that prevent more complex modeling such as capturing local epidemiological dynamics or capturing the real patterns and dynamics. For example, the unavailability of: (i) epidemiological information such as the frequency with which it is made available; (ii) sociodemographic and environmental factors (e.g., population density and population mobility) at a finer scale which influence the evolution patterns of infectious diseases; or (iii) the number of cases information that is also very dependent on the degree of testing performed, often with severe territorial disparities and influenced by context factors. Moreover, the delay in case reporting and the lack of quality control in epidemiological information is responsible for biases in the data that lead to many results obtained being subject to the ecological fallacy, making it difficult to identify causal relationships. Other important methodological limitations are the control of spatiotemporal dependence, management of non-linearity, ergodicy, among others, which can impute inconsistencies to the results. In addition to these issues, social contact, is still difficult to quantify in order to be incorporated into modeling processes. This study aims to explore a modeling framework that can overcome some of these modeling methodological limitations to allow more accurate modeling of epidemiological diseases. Based on Geographic Information Systems (GIS) and spatial analysis, our model is developed to identify group of municipalities where population density (vulnerability) has a stronger relationship with incidence (hazard) and commuting movements (exposure). Specifically, our framework shows how to operate a model over data with no clear trend or seasonal pattern which is suitable for a short-term predicting (i.e., forecasting) of cases based on few determinants. Our tested models provide a good alternative for when explanatory data is few and the time component is not available, once they have shown a good fit and good short-term forecast ability.

HERisk and statistical clustering integrated for health risk modelling of PTEs in natural water resources for drinking and sanitary uses

Potentially toxic elements (PTEs) are well-known for exposing living organisms and humans to different levels of risk. The present study aimed to evaluate the extent of exposure to health risks sustained by the inhabitants of different suburbs across Southeastern Nigeria as a result of contaminated water sources. There are existing literatures on human health risk assessment in the study region. However, this is the first study to report a detailed breakdown of the risks faced by nine age groups in the study area. This was achieved by integrating a novel code (HERisk), statistical clustering, and water quality data. Laboratory analysis showed that the levels of PTEs (Fe2+, Ni2+, Cr3+, and Pb2+) in 53.6%, 17.8%, 3.5%, and 46% of the samples were found to be above the recommended limits. Aggregated human health risk scores for ingestion and dermal routes revealed that children aged 1 to 2 years are the most vulnerable to health hazards. According to cumulative carcinogenic health risk values (0.00E+00 to 2.41E-04), samples proximal to regions with significant human activities expose the locals to a high risk of developing cancer via ingestion. Q-mode hierarchical cluster analysis successfully validated the classification schemes used for the interpretation of the HERisk code values, captured the patterns in the dataset, and brought forth new perspectives. The comprehensive approach detailed in this study can be adopted as a framework for ongoing monitoring and assessment of water quality.

Hydrological Study and Hydraulic Modeling of Flood Risk in the Watershed of the Oued Lahdar (Upper Inaouene, Morocco)

Hydrological Study and Hydraulic Modeling of Flood Risk in the Watershed of the Oued Lahdar (Upper Inaouene, Morocco)

Spatial modeling of soil erosion risk: a multi-criteria decision-making (MCDM) approach in the paguyaman watershed, gorontalo, Indonesia

Spatial modeling of soil erosion risk: a multi-criteria decision-making (MCDM) approach in the paguyaman watershed, gorontalo, Indonesia

Dynamic Risk Estimation of Adverse Events in Ambulatory LVAD Patients: A MOMENTUM 3 Analysis

BackgroundHemocompatibility-related adverse events affect patients after left ventricular assist device (LVAD) implantation but are hard to predict. ObjectivesDynamic risk modeling with a multistate model can predict risk of gastrointestinal bleeding (GIB), stroke, or death in ambulatory patients. MethodsThis was a secondary analysis of the MOMENTUM 3 (Multicenter Study of MagLev Technology in Patients Undergoing Mechanical Circulatory Support Therapy with HeartMate 3) trial. HeartMate 3 LVAD recipients who survived to hospital discharge and were followed for up to 2 years. A total of 145 variables were included in the multistate model with multivariate logistic regression. Model performance was assessed with the area under the curve in a holdout validation cohort. A risk stratification tool was created by dividing patients into categories of predicted risk using the final model variables and associated OR. ResultsAmong 2,056 LVAD patients, the median age was 59.4 years (20.4% women, 28.6% Black). At 2 years, the incidence of GIB, stroke, and death was 25.6%, 6.0%, and 12.3%, respectively. The multistate model included 39 total variables to predict risk of GIB (16 variables), stroke (10 variables), and death (19 variables). When ambulatory patients were classified according to their risk category, the 30-day observed event rate in the highest risk group for GIB, stroke, or death was 26.9%, 1.8%, and 4.8%, respectively. The multistate model predicted GIB, stroke, and death at any 30-day period with an area under the curve of 0.70, 0.69, and 0.86, respectively. ConclusionsThe multistate model informs 30-day risk in ambulatory LVAD recipients and allows recalculation of risk as new patient-specific data become available. The model allows for accurate risk stratification that predicts impending adverse events and may guide clinical decision making. (MOMENTUM 3 IDE Clinical Study Protocol; NCT02224755)

Serum uric acid to high-density lipoprotein cholesterol ratio is a predictor for all-cause and cardiovascular disease mortality in patients with diabetes: Evidence from NHANES 2005–2018

Background and aimsThe relationship between uric acid to high-density lipoprotein cholesterol ratio (UHR) and mortality in individuals with diabetes remains uncertain. This study aimed to explore the relationship between serum UHR and all-cause and cardiovascular disease (CVD) mortality in adults with diabetes. Methods and resultsA total of 5,665 patients with diabetes were enrolled from the 2005–2018 National Health and Nutrition Examination Survey (NHANES). Mortality data were determined through the National Death Index (NDI) until December 31, 2019. The multivariate hazard ratio (HR) and 95% confidence interval (CI) were examined by Cox proportional risk modeling and threshold effects analysis. Stratified analyses were conducted to identify the populations with high-risk mortality. Among the participants with diabetes, 1,088 all-cause mortality, containing 310 CVD mortality occurred. Following adjustment for multivariate, higher UHR was significantly and nonlinearly associated with increased all-cause mortality (HR 1.02, 95% CI 1.02–1.02) and CVD mortality (HR 1.03, 95% CI 1.03–1.03). Furthermore, a U-shaped relationship between UHR and all-cause and CVD mortality, with a plateau at 12.57% for all-cause mortality and 9.86% for CVD mortality. Below the inflection points, a higher UHR was associated with a 4% reduced risk for all-cause mortality. Conversely, exceeding the inflection points, a 4% higher risk for all-cause and a 3% higher risk for CVD mortality associated with elevated UHR. ConclusionsNonlinearity of UHR with all-cause and CVD mortality was observed in adults with diabetes in the United States, with thresholds identified at 12.57% for all-cause and 9.86% for CVD mortality respectively.