Justin Morgan Leach
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Justin Morgan Leach has a total of 44 co-authors. Here are the top co-authors by number of shared publications:
Gary Cutter
Department of Biostatistics University of Alabama at Birmingham Birmingham Alabama USA
3 shared publicationsInmaculada Aban
Department of Biostatistics University of Alabama at Birmingham Birmingham Alabama USA
3 shared publications- Nengjun Yi
Department of Biostatistics, University of Alabama at Birmingham, Birmingham, AL 35294, USA
3 shared publications - Inmaculada B. Aban
University of Alabama at Birmingham
2 shared publications - Jonghwa Oh
Department of Environmental Health Sciences, University of Alabama at Birmingham, Birmingham, AL, 35294, USA.
2 shared publications - Nathan Chen
Environmental Health Sciences, The University of Alabama at Birmingham, Birmingham, Alabama, USA
2 shared publications
Latest Publications
- 15 July 2026
- Published by SAGE Publications in Statistical Methods in Medical Research
- Vol. 35 (9) , 1882-1896
Bayesian mediation analysis for zero-inflated mediator
Mediation analysis is a powerful tool for exploring the causal relationships between exposures and outcomes that are mediated by intermediate variables. In this paper, we propose a flexible Bayesian mediation analysis framework to accommodate zero-inflated mediators, compatible with a wide range of outcome distributions. This novel technique employs Bayesian models for both the mediator and the outcome, utilizing Markov Chain Monte Carlo algorithms for parameter estimation. While addressing the challenges posed by an excess of zeros, we further decompose the mediation effects into components influenced by either the probability of zero or the mean of the non-zero distribution in the mediator. An associated R package mediationBayes ( https://github.com/jhcuibst/mediationBayes.git ) has been developed to facilitate the application of this framework. Through comprehensive simulation studies, we demonstrate that our method outperforms alternatives in terms of the accuracy of point estimates, coverage probabilities, and the precision of the mediation effects decomposition. We further illustrate the practical applicability of our method by conducting an analysis on the REasons for Geographic And Racial Differences in Stroke Study to investigate the mediating influence of smoking pack-years on the association between educational levels and incident hypertension, where mediation effects are quantified on a risk ratio scale for binary outcomes.
- 15 July 2026
- Published by SAGE Publications in Statistical Methods in Medical Research
- Vol. 35 (9) , 1882-1896
Bayesian mediation analysis for zero-inflated mediator
Mediation analysis is a powerful tool for exploring the causal relationships between exposures and outcomes that are mediated by intermediate variables. In this paper, we propose a flexible Bayesian mediation analysis framework to accommodate zero-inflated mediators, compatible with a wide range of outcome distributions. This novel technique employs Bayesian models for both the mediator and the outcome, utilizing Markov Chain Monte Carlo algorithms for parameter estimation. While addressing the challenges posed by an excess of zeros, we further decompose the mediation effects into components influenced by either the probability of zero or the mean of the non-zero distribution in the mediator. An associated R package mediationBayes ( https://github.com/jhcuibst/mediationBayes.git ) has been developed to facilitate the application of this framework. Through comprehensive simulation studies, we demonstrate that our method outperforms alternatives in terms of the accuracy of point estimates, coverage probabilities, and the precision of the mediation effects decomposition. We further illustrate the practical applicability of our method by conducting an analysis on the REasons for Geographic And Racial Differences in Stroke Study to investigate the mediating influence of smoking pack-years on the association between educational levels and incident hypertension, where mediation effects are quantified on a risk ratio scale for binary outcomes.
- Open Access
- 30 April 2026
- Published by Springer Nature in International Archives of Occupational and Environmental Health
- Vol. 99 (4) , 22
Measurement of muscle passive stiffness in vibration-exposed groundskeepers
Objective: Currently, there is no standardized method to quantify musculoskeletal health of hand-arm vibration syndrome (HAVS). This study evaluated passive stiffness (PS) of the abductor pollicis brevis (APB) muscle among groundskeepers and examined the potential relationship of the PS with lifetime hand-arm vibration (HAV) exposure of the workers. Methods: PS of APB muscle was measured using a myotonometer among 17 groundskeepers and 10 office workers for 3 days. Study participants’ HAV exposure, demographic information, social history, and health-related information were pulled from our previous studies conducted in parallel with the present study. Linear mixed models were employed to estimate the association between lifetime HAV exposure dose and PS. Results: The average PS for right and left hands was 509.7 N/m (range 402.2–752.9 N/m) and 501.7 N/m (range 380.4–660.7 N/m), respectively, in the exposure group and 422.0 N/m (range 365.0–517.2 N/m) and 430.3 N/m (range 359.2–591.9 N/m), respectively, in the reference group. A significant positive association between lifetime HAV exposure and PS was observed in the multivariable linear mixed model after adjusting for age, body mass index, and race/ethnicity (right hand: β = 6.972 and p-value = 0.0128; left hand: β = 9.039 and p-value = 0.0108). Conclusions: A significant association between lifetime HAV exposure and PS was found, supporting PS measurement using myotonometry as a promising health indicator for muscular disorders induced by HAV.
- 30 April 2026
- Published by Springer Nature in International Archives of Occupational and Environmental Health
- Vol. 99 (4) , 22
Open AccessMeasurement of muscle passive stiffness in vibration-exposed groundskeepers
Objective: Currently, there is no standardized method to quantify musculoskeletal health of hand-arm vibration syndrome (HAVS). This study evaluated passive stiffness (PS) of the abductor pollicis brevis (APB) muscle among groundskeepers and examined the potential relationship of the PS with lifetime hand-arm vibration (HAV) exposure of the workers. Methods: PS of APB muscle was measured using a myotonometer among 17 groundskeepers and 10 office workers for 3 days. Study participants’ HAV exposure, demographic information, social history, and health-related information were pulled from our previous studies conducted in parallel with the present study. Linear mixed models were employed to estimate the association between lifetime HAV exposure dose and PS. Results: The average PS for right and left hands was 509.7 N/m (range 402.2–752.9 N/m) and 501.7 N/m (range 380.4–660.7 N/m), respectively, in the exposure group and 422.0 N/m (range 365.0–517.2 N/m) and 430.3 N/m (range 359.2–591.9 N/m), respectively, in the reference group. A significant positive association between lifetime HAV exposure and PS was observed in the multivariable linear mixed model after adjusting for age, body mass index, and race/ethnicity (right hand: β = 6.972 and p-value = 0.0128; left hand: β = 9.039 and p-value = 0.0108). Conclusions: A significant association between lifetime HAV exposure and PS was found, supporting PS measurement using myotonometry as a promising health indicator for muscular disorders induced by HAV.
- Open Access
- 16 April 2025
- Published by Wiley in Annals of Clinical and Translational Neurology
- Vol. 12 (5) , 888-897
Handling rescue therapy in myasthenia gravis clinical trials: why it matters and why you should care
Myasthenia gravis (MG) clinical trials typically allow rescue therapy during follow‐up in the event of marked worsening of MG symptoms. Failure to appropriately address rescue therapy in defining treatment effects and planning statistical analyses may yield biased estimates, increase false positive rates, or decrease statistical power – all of which increase the risk that inaccurate information influences patient care. In alignment with recent International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) guidelines, we review strategies based on the estimand framework that produce rigorously defined treatment effects in MG trials where rescue therapy may be administered. We also discuss the interpretation, clinical relevance, and pitfalls of each strategy in the context of MG trials, suggesting circumstances in which a strategy would or would not be appropriate. Finally, we outline available statistical methods for estimating treatment effects based on each strategy. As primary endpoints and statistical analyses for trials must be defined prior to conducting the study, it is necessary to consider how to address rescue therapy during study planning.
- 16 April 2025
- Published by Wiley in Annals of Clinical and Translational Neurology
- Vol. 12 (5) , 888-897
Open AccessHandling rescue therapy in myasthenia gravis clinical trials: why it matters and why you should care
Myasthenia gravis (MG) clinical trials typically allow rescue therapy during follow‐up in the event of marked worsening of MG symptoms. Failure to appropriately address rescue therapy in defining treatment effects and planning statistical analyses may yield biased estimates, increase false positive rates, or decrease statistical power – all of which increase the risk that inaccurate information influences patient care. In alignment with recent International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) guidelines, we review strategies based on the estimand framework that produce rigorously defined treatment effects in MG trials where rescue therapy may be administered. We also discuss the interpretation, clinical relevance, and pitfalls of each strategy in the context of MG trials, suggesting circumstances in which a strategy would or would not be appropriate. Finally, we outline available statistical methods for estimating treatment effects based on each strategy. As primary endpoints and statistical analyses for trials must be defined prior to conducting the study, it is necessary to consider how to address rescue therapy during study planning.
- 6 March 2025
- Published by Elsevier in Social Science & Medicine
- Vol. 372, 117922
Health insurance, race, and receipt of a postpartum visit among patients giving birth in a referral hospital in the US South
- 6 March 2025
- Published by Elsevier in Social Science & Medicine
- Vol. 372, 117922
Health insurance, race, and receipt of a postpartum visit among patients giving birth in a referral hospital in the US South
- Open Access
- 1 February 2025
- Published by BMJ in Occupational and Environmental Medicine
- Vol. 82 (2) , 83-89
Finger blood flow (FBF) measurement among vibration-exposed groundskeepers: a pilot study in the southeastern USA
Objectives: This study assessed finger blood flow (FBF) among groundskeepers using laser Doppler flowmetry (LDF) and evaluated the association of the FBF with hand-arm vibration (HAV) exposure dose.Methods: Baseline FBF measured before a work shift (FBFbaseline) and daily changes in FBF before and after a work shift (ΔFBFdaily) were measured among 17 groundskeepers and 10 office workers using LDF (PeriFlux 6000, Perimed, Järfälla, Sweden) for 3 days. Study participants’ health-related information was obtained through questionnaires, while HAV exposure and demographic information were pulled from our previous study conducted in parallel with the present study. Linear mixed models were employed to estimate the association between HAV exposure dose and FBF.Results: The average FBFbaselinefor right and left hands was 241.5 and 239.9 perfusion units (PUs), respectively, among the exposure group and 305.6 and 307.3 PU, respectively, among the reference group. The average ΔFBFdailyfor right and left hands was 44.2 and 25.4 PU, respectively, among the exposure group and −35.2 and −33.2 PU, respectively, among the reference group. A significant negative association between lifetime HAV exposure and FBFbaselinewas observed in the linear mixed model after adjusting for age, body mass index, race/ethnicity and hypertension (right hand: β=−0.0006 and p=0.0055; left hand: β=−0.0009 and p=0.0068). Inconsistent significances were observed between lifetime HAV exposure and ΔFBFdailyand between daily HAV exposure and ΔFBFdaily.Conclusions: A significant negative association between lifetime HAV exposure and baseline FBF among groundskeepers was observed, supporting FBF measurement using LDF as a promising health indicator for vascular disorders induced by HAV.
- 1 February 2025
- Published by BMJ in Occupational and Environmental Medicine
- Vol. 82 (2) , 83-89
Open AccessFinger blood flow (FBF) measurement among vibration-exposed groundskeepers: a pilot study in the southeastern USA
Objectives: This study assessed finger blood flow (FBF) among groundskeepers using laser Doppler flowmetry (LDF) and evaluated the association of the FBF with hand-arm vibration (HAV) exposure dose.Methods: Baseline FBF measured before a work shift (FBFbaseline) and daily changes in FBF before and after a work shift (ΔFBFdaily) were measured among 17 groundskeepers and 10 office workers using LDF (PeriFlux 6000, Perimed, Järfälla, Sweden) for 3 days. Study participants’ health-related information was obtained through questionnaires, while HAV exposure and demographic information were pulled from our previous study conducted in parallel with the present study. Linear mixed models were employed to estimate the association between HAV exposure dose and FBF.Results: The average FBFbaselinefor right and left hands was 241.5 and 239.9 perfusion units (PUs), respectively, among the exposure group and 305.6 and 307.3 PU, respectively, among the reference group. The average ΔFBFdailyfor right and left hands was 44.2 and 25.4 PU, respectively, among the exposure group and −35.2 and −33.2 PU, respectively, among the reference group. A significant negative association between lifetime HAV exposure and FBFbaselinewas observed in the linear mixed model after adjusting for age, body mass index, race/ethnicity and hypertension (right hand: β=−0.0006 and p=0.0055; left hand: β=−0.0009 and p=0.0068). Inconsistent significances were observed between lifetime HAV exposure and ΔFBFdailyand between daily HAV exposure and ΔFBFdaily.Conclusions: A significant negative association between lifetime HAV exposure and baseline FBF among groundskeepers was observed, supporting FBF measurement using LDF as a promising health indicator for vascular disorders induced by HAV.
- 23 April 2024
- Published by SAGE Publications in Statistical Methods in Medical Research
- Vol. 33 (6) , 1043-1054
Bayesian compositional models for ordinal response
Ordinal response is commonly found in medicine, biology, and other fields. In many situations, the predictors for this ordinal response are compositional, which means that the sum of predictors for each sample is fixed. Examples of compositional data include the relative abundance of species in microbiome data and the relative frequency of nutrition concentrations. Moreover, the predictors that are strongly correlated tend to have similar influence on the response outcome. Conventional cumulative logistic regression models for ordinal responses ignore the fixed-sum constraint on predictors and their associated interrelationships, and thus are not appropriate for analyzing compositional predictors.To solve this problem, we proposed Bayesian Compositional Models for Ordinal Response to analyze the relationship between compositional data and an ordinal response with a structured regularized horseshoe prior for the compositional coefficients and a soft sum-to-zero restriction on coefficients through the prior distribution. The method was implemented with R package rstan using efficient Hamiltonian Monte Carlo algorithm. We performed simulations to compare the proposed approach and existing methods for ordinal responses. Results revealed that our proposed method outperformed the existing methods in terms of parameter estimation and prediction. We also applied the proposed method to a microbiome study HMP2Data, to find microorganisms linked to ordinal inflammatory bowel disease levels. To make this work reproducible, the code and data used in this paper are available at https://github.com/Li-Zhang28/BCO .
- 23 April 2024
- Published by SAGE Publications in Statistical Methods in Medical Research
- Vol. 33 (6) , 1043-1054
Bayesian compositional models for ordinal response
Ordinal response is commonly found in medicine, biology, and other fields. In many situations, the predictors for this ordinal response are compositional, which means that the sum of predictors for each sample is fixed. Examples of compositional data include the relative abundance of species in microbiome data and the relative frequency of nutrition concentrations. Moreover, the predictors that are strongly correlated tend to have similar influence on the response outcome. Conventional cumulative logistic regression models for ordinal responses ignore the fixed-sum constraint on predictors and their associated interrelationships, and thus are not appropriate for analyzing compositional predictors.To solve this problem, we proposed Bayesian Compositional Models for Ordinal Response to analyze the relationship between compositional data and an ordinal response with a structured regularized horseshoe prior for the compositional coefficients and a soft sum-to-zero restriction on coefficients through the prior distribution. The method was implemented with R package rstan using efficient Hamiltonian Monte Carlo algorithm. We performed simulations to compare the proposed approach and existing methods for ordinal responses. Results revealed that our proposed method outperformed the existing methods in terms of parameter estimation and prediction. We also applied the proposed method to a microbiome study HMP2Data, to find microorganisms linked to ordinal inflammatory bowel disease levels. To make this work reproducible, the code and data used in this paper are available at https://github.com/Li-Zhang28/BCO .
- 14 September 2023
- Published by Taylor & Francis in Journal of Applied Statistics
- Vol. 51 (11) , 2039-2061
The spike-and-slab lasso and scalable algorithm to accommodate multinomial outcomes in variable selection problems
Spike-and-slab prior distributions are used to impose variable selection in Bayesian regression-style problems with many possible predictors. These priors are a mixture of two zero-centered distributions with differing variances, resulting in different shrinkage levels on parameter estimates based on whether they are relevant to the outcome. The spike-and-slab lasso assigns mixtures of double exponential distributions as priors for the parameters. This framework was initially developed for linear models, later developed for generalized linear models, and shown to perform well in scenarios requiring sparse solutions. Standard formulations of generalized linear models cannot immediately accommodate categorical outcomes with > 2 categories, i.e. multinomial outcomes, and require modifications to model specification and parameter estimation. Such modifications are relatively straightforward in a Classical setting but require additional theoretical and computational considerations in Bayesian settings, which can depend on the choice of prior distributions for the parameters of interest. While previous developments of the spike-and-slab lasso focused on continuous, count, and/or binary outcomes, we generalize the spike-and-slab lasso to accommodate multinomial outcomes, developing both the theoretical basis for the model and an expectation-maximization algorithm to fit the model. To our knowledge, this is the first generalization of the spike-and-slab lasso to allow for multinomial outcomes.
- 14 September 2023
- Published by Taylor & Francis in Journal of Applied Statistics
- Vol. 51 (11) , 2039-2061
The spike-and-slab lasso and scalable algorithm to accommodate multinomial outcomes in variable selection problems
Spike-and-slab prior distributions are used to impose variable selection in Bayesian regression-style problems with many possible predictors. These priors are a mixture of two zero-centered distributions with differing variances, resulting in different shrinkage levels on parameter estimates based on whether they are relevant to the outcome. The spike-and-slab lasso assigns mixtures of double exponential distributions as priors for the parameters. This framework was initially developed for linear models, later developed for generalized linear models, and shown to perform well in scenarios requiring sparse solutions. Standard formulations of generalized linear models cannot immediately accommodate categorical outcomes with > 2 categories, i.e. multinomial outcomes, and require modifications to model specification and parameter estimation. Such modifications are relatively straightforward in a Classical setting but require additional theoretical and computational considerations in Bayesian settings, which can depend on the choice of prior distributions for the parameters of interest. While previous developments of the spike-and-slab lasso focused on continuous, count, and/or binary outcomes, we generalize the spike-and-slab lasso to accommodate multinomial outcomes, developing both the theoretical basis for the model and an expectation-maximization algorithm to fit the model. To our knowledge, this is the first generalization of the spike-and-slab lasso to allow for multinomial outcomes.
- Open Access
- 25 June 2022
- Published by Wiley in Muscle & Nerve
- Vol. 66 (4) , 411-420
One‐year follow‐up of disease burden and medication changes in patients with myasthenia gravis: From the MG Patient Registry
Introduction/Aims: We studied the progression of myasthenia gravis (MG) disease burden and medication adjustment among MG Patient Registry participants.Methods: Participants diagnosed with MG (age ≥18 years), registered between July 1, 2013 and July 31, 2018 and completing both 6‐ and 12‐month follow‐up surveys, were included in this investigation. Participants were grouped into high‐burden (Myasthenia Gravis Activity of Daily Living scale [MG‐ADL] score ≥6) and low‐burden (MG‐ADL 20%) at 6 months significantly increased the likelihood of achieving MSE at 12 months (P = .0004).Discussion: In both groups, but more so in the high‐burden group, patients infrequently achieved MSE after 1 year of MG treatment. Baseline low disease burden, improvement at 6 months and no pyridostigmine use were associated with a higher likelihood of MSE at 12 months.
- 25 June 2022
- Published by Wiley in Muscle & Nerve
- Vol. 66 (4) , 411-420
Open AccessOne‐year follow‐up of disease burden and medication changes in patients with myasthenia gravis: From the MG Patient Registry
Introduction/Aims: We studied the progression of myasthenia gravis (MG) disease burden and medication adjustment among MG Patient Registry participants.Methods: Participants diagnosed with MG (age ≥18 years), registered between July 1, 2013 and July 31, 2018 and completing both 6‐ and 12‐month follow‐up surveys, were included in this investigation. Participants were grouped into high‐burden (Myasthenia Gravis Activity of Daily Living scale [MG‐ADL] score ≥6) and low‐burden (MG‐ADL 20%) at 6 months significantly increased the likelihood of achieving MSE at 12 months (P = .0004).Discussion: In both groups, but more so in the high‐burden group, patients infrequently achieved MSE after 1 year of MG treatment. Baseline low disease burden, improvement at 6 months and no pyridostigmine use were associated with a higher likelihood of MSE at 12 months.
- 20 February 2022
- Published by Elsevier in Multiple Sclerosis and Related Disorders
- Vol. 60, 103704
Measuring cognitive function by the SDMT across functional domains: Useful but not sufficient
- 20 February 2022
- Published by Elsevier in Multiple Sclerosis and Related Disorders
- Vol. 60, 103704
Measuring cognitive function by the SDMT across functional domains: Useful but not sufficient