Le Bao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Le Bao has a total of 115 co-authors. Here are the top co-authors by number of shared publications:
Xiaoyue Niu
Department of Statistics, The Pennsylvania State University , Shortlidge Rd , State College, PA 16802
9 shared publicationsIan Laga
Department of Mathematical Sciences, Montana State University, Bozeman, MT
5 shared publicationsTim Brown
Research Program, East-West Center, 1601 East-West Road , Honolulu, HI 96848
5 shared publications- Xiaoyue Niu
Department of Statistics, Pennsylvania State University
4 shared publications - Amrita Rao
Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, United States
3 shared publications Amy Zhang
Affiliation Department of Statistics, Pennsylvania State University, University Park, PA, United States of America
3 shared publications
Latest Publications
- Open Access
- 3 September 2026
- Published by Wiley in Statistics in Medicine
- Vol. 45 (20-22)
A Bayesian Multi‐State Data Integration Approach for Estimating County‐Level Prevalence of Opioid Misuse in the United States
Drug overdose deaths, including from opioids, remain a significant public health threat to the United States (US). To abate the harms of opioid misuse, understanding its prevalence at the local level is crucial for stakeholders in communities to develop response strategies that effectively use limited resources. Although there exist several state‐specific studies that provide county‐level prevalence estimates, such estimates are not widely available across the country, as the datasets used in these studies are not always readily available in other states, which, therefore, has limited the wider applications of existing models. To fill this gap, we propose a Bayesian multi‐state data integration approach that fully utilizes publicly available data sources to estimate county‐level opioid misuse prevalence for all counties in the US. The hierarchical structure jointly models opioid misuse prevalence and overdose death outcomes, leverages existing county‐level prevalence estimates in limited states and state‐level estimates from national surveys, and accounts for heterogeneity across counties and states with counties' covariates and mixed effects. Furthermore, our parsimonious and generalizable modeling framework employs horseshoe+ prior to flexibly shrink coefficients and prevent overfitting, ensuring adaptability as new county‐level prevalence data in additional states become available. Using real‐world data, our model shows high estimation accuracy through cross‐validation and provides nationwide county‐level estimates of opioid misuse for the first time.
- 3 September 2026
- Published by Wiley in Statistics in Medicine
- Vol. 45 (20-22)
Open AccessA Bayesian Multi‐State Data Integration Approach for Estimating County‐Level Prevalence of Opioid Misuse in the United States
Drug overdose deaths, including from opioids, remain a significant public health threat to the United States (US). To abate the harms of opioid misuse, understanding its prevalence at the local level is crucial for stakeholders in communities to develop response strategies that effectively use limited resources. Although there exist several state‐specific studies that provide county‐level prevalence estimates, such estimates are not widely available across the country, as the datasets used in these studies are not always readily available in other states, which, therefore, has limited the wider applications of existing models. To fill this gap, we propose a Bayesian multi‐state data integration approach that fully utilizes publicly available data sources to estimate county‐level opioid misuse prevalence for all counties in the US. The hierarchical structure jointly models opioid misuse prevalence and overdose death outcomes, leverages existing county‐level prevalence estimates in limited states and state‐level estimates from national surveys, and accounts for heterogeneity across counties and states with counties' covariates and mixed effects. Furthermore, our parsimonious and generalizable modeling framework employs horseshoe+ prior to flexibly shrink coefficients and prevent overfitting, ensuring adaptability as new county‐level prevalence data in additional states become available. Using real‐world data, our model shows high estimation accuracy through cross‐validation and provides nationwide county‐level estimates of opioid misuse for the first time.
- 13 August 2026
- Published by Taylor & Francis in Journal of Computational and Graphical Statistics
- p. 1-10
Simultaneous Feature- and Sample-Splitting ADMM for Penalized Rank Regression
1 IntroductionRank regression is an effective tool to deal with heavy-tailed error contamination and outliers due to its robustness and efficiency properties (Hettmansperger and McKean, 1998). It p...
- 13 August 2026
- Published by Taylor & Francis in Journal of Computational and Graphical Statistics
- p. 1-10
Simultaneous Feature- and Sample-Splitting ADMM for Penalized Rank Regression
1 IntroductionRank regression is an effective tool to deal with heavy-tailed error contamination and outliers due to its robustness and efficiency properties (Hettmansperger and McKean, 1998). It p...
- Open Access
- 8 July 2026
- Published by Oxford University Press (OUP) in Journal of the Royal Statistical Society Series A: Statistics in Society
Estimating global HIV prevalence in key populations: a cross-population temporal hierarchical modelling approach
Key populations at high risk of HIV infection are critical for understanding and monitoring HIV epidemics, but global estimation is hampered by sparse, uneven data. We analyse data from 199 countries for female sex workers (FSW), men who have sex with men (MSM), and people who inject drugs (PWID) over 2011–2021, and introduce a cross-population temporal hierarchical model that borrows strength across countries, years, and populations. The model combines region- and population-specific means with country random effects, temporal dependence, and cross-population correlations in a Gaussian Markov random-field formulation on the log-prevalence scale. In fivefold cross-validation, the approach outperforms a regional-median baseline and reduced variants (65% reduction in cross-validation mean squared error) with a well-calibrated posterior predictive coverage (93%). We map the 2021 prevalence and quantify the change between 2011 and 2021 using posterior prevalence ratios to identify countries with substantial increases or decreases. The framework yields globally comparable and uncertainty-quantified country-by-year prevalence estimates, enhancing evidence for resource allocation and targeted interventions for marginalized populations where routine data are limited.
- 8 July 2026
- Published by Oxford University Press (OUP) in Journal of the Royal Statistical Society Series A: Statistics in Society
Open AccessEstimating global HIV prevalence in key populations: a cross-population temporal hierarchical modelling approach
Key populations at high risk of HIV infection are critical for understanding and monitoring HIV epidemics, but global estimation is hampered by sparse, uneven data. We analyse data from 199 countries for female sex workers (FSW), men who have sex with men (MSM), and people who inject drugs (PWID) over 2011–2021, and introduce a cross-population temporal hierarchical model that borrows strength across countries, years, and populations. The model combines region- and population-specific means with country random effects, temporal dependence, and cross-population correlations in a Gaussian Markov random-field formulation on the log-prevalence scale. In fivefold cross-validation, the approach outperforms a regional-median baseline and reduced variants (65% reduction in cross-validation mean squared error) with a well-calibrated posterior predictive coverage (93%). We map the 2021 prevalence and quantify the change between 2011 and 2021 using posterior prevalence ratios to identify countries with substantial increases or decreases. The framework yields globally comparable and uncertainty-quantified country-by-year prevalence estimates, enhancing evidence for resource allocation and targeted interventions for marginalized populations where routine data are limited.
- Open Access
Estimating Global HIV Prevalence in Key Populations: A Cross-Population Hierarchical Modeling Approach
Key populations at high risk of HIV infection are critical for understanding and monitoring HIV epidemics, but global estimation is hampered by sparse, uneven data. We analyze data from 199 countries for female sex workers (FSW), men who have sex with men (MSM), and people who inject drugs (PWID) over 2011-2021, and introduce a cross-population hierarchical model that borrows strength across countries, years, and populations. The model combines region- and population-specific means with country random effects, temporal dependence, and cross-population correlations in a Gaussian Markov random-field formulation on the log-prevalence scale. In 5-fold cross-validation, the approach outperforms a regional-median baseline and reduced variants (65 percent reduction in cross-validated MSE) with well-calibrated posterior predictive coverage (93 percent). We map the 2021 prevalence and quantify the change between 2011 and 2021 using posterior prevalence ratios to identify countries with substantial increases or decreases. The framework yields globally comparable and uncertainty-quantified country-by-year prevalence estimates, enhancing evidence for resource allocation and targeted interventions for marginalized populations where routine data are limited.
Open AccessEstimating Global HIV Prevalence in Key Populations: A Cross-Population Hierarchical Modeling Approach
Key populations at high risk of HIV infection are critical for understanding and monitoring HIV epidemics, but global estimation is hampered by sparse, uneven data. We analyze data from 199 countries for female sex workers (FSW), men who have sex with men (MSM), and people who inject drugs (PWID) over 2011-2021, and introduce a cross-population hierarchical model that borrows strength across countries, years, and populations. The model combines region- and population-specific means with country random effects, temporal dependence, and cross-population correlations in a Gaussian Markov random-field formulation on the log-prevalence scale. In 5-fold cross-validation, the approach outperforms a regional-median baseline and reduced variants (65 percent reduction in cross-validated MSE) with well-calibrated posterior predictive coverage (93 percent). We map the 2021 prevalence and quantify the change between 2011 and 2021 using posterior prevalence ratios to identify countries with substantial increases or decreases. The framework yields globally comparable and uncertainty-quantified country-by-year prevalence estimates, enhancing evidence for resource allocation and targeted interventions for marginalized populations where routine data are limited.
- Open Access
Bridging Data Gaps in Oncology: Large Language Models and Collaborative Filtering for Cancer Treatment Recommendations
Background: Patients with rare cancers face substantial challenges due to limited evidence-based treatment options, resulting from sparse clinical trials. Advances in large language models (LLMs) and recommendation algorithms offer new opportunities to utilize all clinical trial information to improve clinical decisions. Methods: We used LLM to systematically extract and standardize more than 100,000 cancer trials from ClinicalTrials.gov . Each trial was annotated using a customized scoring system reflecting cancer-treatment interactions based on clinical outcomes and trial attributes. Using this structured data set, we implemented three state-of-the-art collaborative filtering algorithms to recommend potentially effective treatments across different cancer types. Results: The LLM-driven data extraction process successfully generated a comprehensive and rigorously curated database from fragmented clinical trial information, covering 78 cancer types and 5,315 distinct interventions. Recommendation models demonstrated high predictive accuracy (cross-validated RMSE: 0.49–0.62) and identified clinically meaningful new treatments for melanoma, independently validated by oncology experts. Conclusions: Our study establishes a proof of concept demonstrating that the combination of LLMs with sophisticated recommendation algorithms can systematically identify novel and clinically plausible cancer treatments. This integrated approach may accelerate the identification of effective therapies for rare cancers, ultimately improving patient outcomes by generating evidence-based treatment recommendations where traditional data sources remain limited.
Open AccessBridging Data Gaps in Oncology: Large Language Models and Collaborative Filtering for Cancer Treatment Recommendations
Background: Patients with rare cancers face substantial challenges due to limited evidence-based treatment options, resulting from sparse clinical trials. Advances in large language models (LLMs) and recommendation algorithms offer new opportunities to utilize all clinical trial information to improve clinical decisions. Methods: We used LLM to systematically extract and standardize more than 100,000 cancer trials from ClinicalTrials.gov . Each trial was annotated using a customized scoring system reflecting cancer-treatment interactions based on clinical outcomes and trial attributes. Using this structured data set, we implemented three state-of-the-art collaborative filtering algorithms to recommend potentially effective treatments across different cancer types. Results: The LLM-driven data extraction process successfully generated a comprehensive and rigorously curated database from fragmented clinical trial information, covering 78 cancer types and 5,315 distinct interventions. Recommendation models demonstrated high predictive accuracy (cross-validated RMSE: 0.49–0.62) and identified clinically meaningful new treatments for melanoma, independently validated by oncology experts. Conclusions: Our study establishes a proof of concept demonstrating that the combination of LLMs with sophisticated recommendation algorithms can systematically identify novel and clinically plausible cancer treatments. This integrated approach may accelerate the identification of effective therapies for rare cancers, ultimately improving patient outcomes by generating evidence-based treatment recommendations where traditional data sources remain limited.
- Open Access
- 14 November 2024
- Published by Taylor & Francis in Journal of Computational and Graphical Statistics
- Vol. 34 (2) , 488-497
Approximate Cross-Validated Mean Estimates for Bayesian Hierarchical Regression Models
We introduce a novel procedure for obtaining cross-validated predictive estimates for Bayesian hierarchical regression models (BHRMs). BHRMs are popular for modeling complex dependence structures (e.g., Gaussian processes and Gaussian Markov random fields) but can be computationally expensive to run. Cross-validation (CV) is, therefore, not a common practice to evaluate the predictive performance of BHRMs. Our method circumvents the need to rerun computationally costly estimation methods for each cross-validation fold and makes CV more feasible for large BHRMs. We shift the CV problem from probability-based sampling to a familiar and straightforward optimization problem by conditioning on the variance-covariance parameters. Our approximation applies to leave-one-out CV and leave-one-cluster-out CV, the latter of which is more appropriate for models with complex dependencies. In many cases, this produces estimates equivalent to full CV. We provide theoretical results, demonstrate the efficacy of our method on publicly available data and in simulations, and compare the model performance with several competing methods for CV approximation. Code and other supplementary materials available online.
- 14 November 2024
- Published by Taylor & Francis in Journal of Computational and Graphical Statistics
- Vol. 34 (2) , 488-497
Open AccessApproximate Cross-Validated Mean Estimates for Bayesian Hierarchical Regression Models
We introduce a novel procedure for obtaining cross-validated predictive estimates for Bayesian hierarchical regression models (BHRMs). BHRMs are popular for modeling complex dependence structures (e.g., Gaussian processes and Gaussian Markov random fields) but can be computationally expensive to run. Cross-validation (CV) is, therefore, not a common practice to evaluate the predictive performance of BHRMs. Our method circumvents the need to rerun computationally costly estimation methods for each cross-validation fold and makes CV more feasible for large BHRMs. We shift the CV problem from probability-based sampling to a familiar and straightforward optimization problem by conditioning on the variance-covariance parameters. Our approximation applies to leave-one-out CV and leave-one-cluster-out CV, the latter of which is more appropriate for models with complex dependencies. In many cases, this produces estimates equivalent to full CV. We provide theoretical results, demonstrate the efficacy of our method on publicly available data and in simulations, and compare the model performance with several competing methods for CV approximation. Code and other supplementary materials available online.
- 3 October 2024
- Published by Oxford University Press (OUP) in Biometrics
- Vol. 80 (4)
A likelihood approach to incorporating self-report data in HIV recency classification
Estimating new HIV infections is significant yet challenging due to the difficulty in distinguishing between recent and long-term infections. We demonstrate that HIV recency status (recent versus long-term) could be determined from self-report testing history and biomarkers, which are increasingly available in bio-behavioral surveys. HIV recency status is partially observed, given the self-report testing history. For example, people who tested positive for HIV over 1 year ago should have a long-term infection. Based on the nationally representative samples collected by the Population-based HIV Impact Assessment (PHIA) Project, we propose a likelihood-based probabilistic model for HIV recency classification. The model incorporates individuals with known recency status based on testing histories and individuals whose recency status could not be determined and integrates the mechanism of how HIV recency status depends on biomarkers and the mechanism of how HIV recency status, together with the self-report time of the most recent HIV test, impacts the test results. We compare our method to logistic regression and the binary classification tree (current practice) on Malawi PHIA data, as well as on simulated data. Our model obtains more efficient and less biased parameter estimates and is relatively robust to potential reporting error and model misspecification.
- 3 October 2024
- Published by Oxford University Press (OUP) in Biometrics
- Vol. 80 (4)
A likelihood approach to incorporating self-report data in HIV recency classification
Estimating new HIV infections is significant yet challenging due to the difficulty in distinguishing between recent and long-term infections. We demonstrate that HIV recency status (recent versus long-term) could be determined from self-report testing history and biomarkers, which are increasingly available in bio-behavioral surveys. HIV recency status is partially observed, given the self-report testing history. For example, people who tested positive for HIV over 1 year ago should have a long-term infection. Based on the nationally representative samples collected by the Population-based HIV Impact Assessment (PHIA) Project, we propose a likelihood-based probabilistic model for HIV recency classification. The model incorporates individuals with known recency status based on testing histories and individuals whose recency status could not be determined and integrates the mechanism of how HIV recency status depends on biomarkers and the mechanism of how HIV recency status, together with the self-report time of the most recent HIV test, impacts the test results. We compare our method to logistic regression and the binary classification tree (current practice) on Malawi PHIA data, as well as on simulated data. Our model obtains more efficient and less biased parameter estimates and is relatively robust to potential reporting error and model misspecification.
- Open Access
- 29 February 2024
- Published by Oxford University Press (OUP) in Biostatistics
- Vol. 25 (4) , 1049-1061
Dynamic models augmented by hierarchical data: an application of estimating HIV epidemics at sub-national level
Summary: Dynamic models have been successfully used in producing estimates of HIV epidemics at the national level due to their epidemiological nature and their ability to estimate prevalence, incidence, and mortality rates simultaneously. Recently, HIV interventions and policies have required more information at sub-national levels to support local planning, decision-making and resource allocation. Unfortunately, many areas lack sufficient data for deriving stable and reliable results, and this is a critical technical barrier to more stratified estimates. One solution is to borrow information from other areas within the same country. However, directly assuming hierarchical structures within the HIV dynamic models is complicated and computationally time-consuming. In this article, we propose a simple and innovative way to incorporate hierarchical information into the dynamical systems by using auxiliary data. The proposed method efficiently uses information from multiple areas within each country without increasing the computational burden. As a result, the new model improves predictive ability and uncertainty assessment.
- 29 February 2024
- Published by Oxford University Press (OUP) in Biostatistics
- Vol. 25 (4) , 1049-1061
Open AccessDynamic models augmented by hierarchical data: an application of estimating HIV epidemics at sub-national level
Summary: Dynamic models have been successfully used in producing estimates of HIV epidemics at the national level due to their epidemiological nature and their ability to estimate prevalence, incidence, and mortality rates simultaneously. Recently, HIV interventions and policies have required more information at sub-national levels to support local planning, decision-making and resource allocation. Unfortunately, many areas lack sufficient data for deriving stable and reliable results, and this is a critical technical barrier to more stratified estimates. One solution is to borrow information from other areas within the same country. However, directly assuming hierarchical structures within the HIV dynamic models is complicated and computationally time-consuming. In this article, we propose a simple and innovative way to incorporate hierarchical information into the dynamical systems by using auxiliary data. The proposed method efficiently uses information from multiple areas within each country without increasing the computational burden. As a result, the new model improves predictive ability and uncertainty assessment.
- 7 September 2023
- Published by Oxford University Press (OUP) in Journal of the Royal Statistical Society Series A: Statistics in Society
- Vol. 187 (2) , 321-337
Multivariate spatial modelling for predicting missing HIV prevalence rates among key populations
Ending the HIV/AIDS pandemic is among the sustainable development goals for the next decade. To overcome the problem caused by the imbalances between the need for care and the limited resources, we shall improve our understanding of the local HIV epidemics, especially for key populations at high risk of HIV infection. However, HIV prevalence rates for key populations have been difficult to estimate because their HIV surveillance data are very scarce. This paper develops a multivariate spatial model for predicting unknown HIV prevalence rates among key populations. The proposed multivariate conditional auto-regressive model efficiently pools information from neighbouring locations and correlated populations. As the real data analysis illustrates, it provides more accurate predictions than independently fitting the sub-epidemic for each key population. Furthermore, we investigate how different pieces of surveillance data contribute to the prediction and offer practical suggestions for epidemic data collection.
- 7 September 2023
- Published by Oxford University Press (OUP) in Journal of the Royal Statistical Society Series A: Statistics in Society
- Vol. 187 (2) , 321-337
Multivariate spatial modelling for predicting missing HIV prevalence rates among key populations
Ending the HIV/AIDS pandemic is among the sustainable development goals for the next decade. To overcome the problem caused by the imbalances between the need for care and the limited resources, we shall improve our understanding of the local HIV epidemics, especially for key populations at high risk of HIV infection. However, HIV prevalence rates for key populations have been difficult to estimate because their HIV surveillance data are very scarce. This paper develops a multivariate spatial model for predicting unknown HIV prevalence rates among key populations. The proposed multivariate conditional auto-regressive model efficiently pools information from neighbouring locations and correlated populations. As the real data analysis illustrates, it provides more accurate predictions than independently fitting the sub-epidemic for each key population. Furthermore, we investigate how different pieces of surveillance data contribute to the prediction and offer practical suggestions for epidemic data collection.
- Open Access
- 10 July 2023
- Published by Wiley in Environmetrics
- Vol. 34 (6)
Air pollution estimation under air stagnation—A case study of Beijing
Air pollution continues to be a major environmental concern in China. The wind‐driven transmission poses difficulties in understanding the air pollution patterns at the local level. The main objective of this study is to offer a straightforward approach for investigating the temporal trends and meteorological effects on the air pollutant concentrations during the generation process without being confounded by the complex wind‐driven transmission effect. We focus on the hourly data of the three most common air pollutants: PM2.5, NO, and CO under air stagnation in Beijing, China, during 2014–2017. We find that the local pollution levels under air stagnation in Beijing have decreased over the years; winter is the severest month of the year; Sunday is the clearest day of the week. Our model also interpolates the air pollutant concentrations at sites without monitoring stations and provides a map of air pollution concentrations under air stagnation. The results could be used to identify locations where air pollutants easily accumulate.
- 10 July 2023
- Published by Wiley in Environmetrics
- Vol. 34 (6)
Open AccessAir pollution estimation under air stagnation—A case study of Beijing
Air pollution continues to be a major environmental concern in China. The wind‐driven transmission poses difficulties in understanding the air pollution patterns at the local level. The main objective of this study is to offer a straightforward approach for investigating the temporal trends and meteorological effects on the air pollutant concentrations during the generation process without being confounded by the complex wind‐driven transmission effect. We focus on the hourly data of the three most common air pollutants: PM2.5, NO, and CO under air stagnation in Beijing, China, during 2014–2017. We find that the local pollution levels under air stagnation in Beijing have decreased over the years; winter is the severest month of the year; Sunday is the clearest day of the week. Our model also interpolates the air pollutant concentrations at sites without monitoring stations and provides a map of air pollution concentrations under air stagnation. The results could be used to identify locations where air pollutants easily accumulate.