About

Alan Hubbard is a Professor and Chair of BioStatistics at the School of Public Health. His research interests encompass causal inference, statistical issues in infectious disease, bioinformatics, data adaptive target parameters, and prediction and variable importance in trauma. He serves as a Principal Investigator and is actively involved in advancing methodologies within these areas. Hubbard's work focuses on developing and applying statistical techniques to address complex biological and health-related questions, contributing to the fields of computational biology and public health.

Research topics

• Medicine
• Artificial Intelligence
• Environmental health
• Machine Learning
• Internal medicine
• Computer Science
• Operations management
• Mathematics
• Emergency medicine
• Virology
• Algorithm
• Statistics
• Socioeconomics
• Engineering
• Economic growth
• Biology
• Economics

Selected publications

• Effect of a theory-driven health education intervention on personal protective equipment use among commercial motorcycle riders in Cameroon: A quasi-experimental study

  medRxiv · 2026-04-09

  articleOpen access

  Abstract Background Commercial motorcycle riders are among the most vulnerable road users in low- and middle-income countries and contribute substantially to the burden of road traffic injuries. The use of personal protective equipment (PPE), including helmets and protective clothing, reduces injury severity; however, uptake remains suboptimal. This study evaluated the effectiveness of a theory-driven health education intervention in improving knowledge, attitudes, and use of PPE among commercial motorcycle riders in Cameroon. Methods A quasi-experimental, non-randomized controlled before-and-after study was conducted in Limbe (intervention) and Tiko (control) Health Districts between August 4, 2024, and April 6, 2025. Participants were recruited from a cohort of commercial motorcycle riders and followed over an eight-month intervention period. The intervention, guided by the Health Belief Model and developed using the Intervention Mapping framework, combined face-to-face sensitization sessions with mobile phone–based educational messaging adapted to participants’ literacy levels and communication preferences. Data were collected at baseline and endline using structured questionnaires and direct observation checklists. Intervention effects were estimated using difference-in-differences analysis with generalized estimating equations, adjusting for socio-demographic factors. Results A total of 313 riders were enrolled at baseline (183 intervention, 130 control), with 249 retained at endline (149 intervention, 100 control). The intervention was associated with significant improvements in PPE knowledge (β = 2.91; 95% CI: 2.14–3.68; p < 0.001) and attitudes (β = 5.76; 95% CI: 4.32–7.21; p < 0.001) compared with the control group. No statistically significant effect was observed for PPE practice scores (β = 0.21; 95% CI: −0.09–0.52; p = 0.171). Among individual PPE items, helmet use increased significantly in the intervention group relative to the control group (AOR = 2.38; 95% CI: 1.19–9.45; p = 0.036), while no significant effects were observed for gloves, trousers, eyeglasses, or closed-toe shoes. Conclusion The theory-driven health education intervention significantly improved knowledge and attitudes toward PPE and increased helmet use among commercial motorcycle riders but did not lead to broader improvements in the uptake of other protective equipment. These findings highlight the need for complementary structural and policy interventions to address persistent barriers to PPE use in similar low-resource settings. Trial registration ClinicalTrials.gov Identifier: NCT07087444 (registered July 28, 2025, retrospectively)

  PublisherDOI

• Impact of a health education intervention on knowledge, attitudes, and use of visibility materials among commercial motorcycle riders: a quasi-experimental study in Cameroon

  Research Square · 2026-03-27 · 1 citations

  preprintOpen access
  PublisherOA PDFDOI

• Determinants of knowledge and attitudes toward personal protective equipment use among commercial motorcycle riders in a conflict-affected region of Cameroon: implications for road safety in Sub-Saharan Africa

  Frontiers in Public Health · 2026-03-18 · 2 citations

  articleOpen access

  Background: Road traffic injuries are a leading cause of death globally, disproportionately affecting motorcyclists in low- and middle-income countries (LMICs). Personal protective equipment (PPE) such as helmets and visibility materials substantially reduce injury severity, yet their consistent use remains suboptimal in many African settings. In conflict-affected areas, where displacement and informal transport systems are common, safety behaviors may be further compromised. This study examined the determinants of knowledge and attitudes toward PPE use among commercial motorcycle riders in the Limbe and Tiko Health Districts of Cameroon to inform regional road safety interventions. Methods: A community-based cross-sectional study was conducted among 499 commercial motorcycle riders aged ≥18 years between May and August 2024. Participants were selected using a multistage sampling approach, and data were collected by trained research assistants through interviewer-administered structured questionnaires at randomly selected motorcycle pick-up points. The questionnaire captured sociodemographic characteristics, riding experience, and PPE knowledge (13 items) and attitudes (9 Likert-scale items). Knowledge and attitude scores were categorized using a 60% threshold. Multivariable logistic regression was applied to identify factors independently associated with good knowledge and positive attitudes. Results: The mean age of participants was 32.2 ± 7.6 years; all were male. Overall, 66.1% had good knowledge and 45.9% had positive attitudes toward PPE use. License ownership (AOR = 1.7; 95% CI: 1.1-2.6) and prior PPE training (AOR = 2.0; 95% CI: 1.4-3.0) were positively associated with knowledge, while internal displacement reduced the odds (AOR = 0.6; 95% CI: 0.4-0.9). Predictors of positive attitudes included license ownership (AOR = 1.8; 95% CI: 1.1-2.8), PPE training (AOR = 1.7; 95% CI: 1.1-2.6), and good knowledge (AOR = 10.4; 95% CI: 6.3-17.3). Internal displacement again reduced the likelihood of positive attitudes (AOR = 0.6; 95% CI: 0.4-0.9). Conclusion: Although knowledge of PPE was relatively high, attitudes remained inadequate, particularly among internally displaced riders. Strengthening motorcycle licensing systems, integrating PPE training into road safety programs, and addressing displacement-related vulnerabilities could improve safety behaviors. Findings from this conflict-affected region of Cameroon provide valuable insights for designing community-based interventions to enhance PPE use and reduce motorcycle-related injuries across Sub-Saharan Africa.

  PublisherOA PDFDOI

• Bridging binarization: causal inference with dichotomized continuous exposures

  Journal of Causal Inference · 2026-01-01

  preprintOpen access

  The average treatment effect (ATE) is a common parameter estimated in causal inference literature, but it is only defined for binary exposures. Thus, despite concerns raised by some researchers, many studies seeking to estimate the causal effect of a continuous exposure create a new binary exposure variable by dichotomizing the continuous values into two categories. In this paper, we affirm binarization as a statistically valid method for answering causal questions about continuous exposures by showing the equivalence between the binarized ATE and the difference in the average outcomes of two specific modified treatment policies. These policies impose cut-offs corresponding to the binarized exposure variable and assume preservation of relative self-selection. Relative self-selection is the ratio of the probability density of an individual having an exposure equal to one value of the continuous exposure variable versus another. The policies assume that, for any two values of the exposure variable with non-zero probability density after the cut-off, this ratio will remain unchanged. Through this equivalence, we clarify the assumptions underlying binarization and discuss how to properly interpret the resulting estimator. Additionally, we introduce a new target parameter that can be computed after binarization that considers the observed world as a benchmark. We argue that this parameter addresses more relevant causal questions than the traditional binarized ATE parameter. We present a simulation study to illustrate the implications of these assumptions when analyzing data and to demonstrate how to correctly implement estimators of the parameters discussed. Finally, we present an application of this method to evaluate the effect of a law in the state of California which seeks to limit exposures to oil and gas wells on birth outcomes to further illustrate the underlying assumptions.

  PublisherOA PDFDOI

• Causal Inference via Electronic Health Records in the National Clinical Cohort Collaborative: Challenges and Solutions in Long COVID Research

  medRxiv · 2025-06-08 · 1 citations

  preprintOpen accessSenior author

  Abstract Observational analyses of electronic health record (EHR) data using databases such as the National Clinical Cohort Collaborative include unique challenges for researchers seeking causal inferences, particularly when evaluating subjectively-defined outcomes like Long COVID. We explore several challenges and describe potential solutions. 1. Lack of true negatives: Many diagnoses and conditions either have a positive indicator or a missing status, requiring investigators to carefully consider which patients are likely negative for this condition. 2. Differential monitoring: EHR data include nonrandom missingness driven by patients engaging with the healthcare system at different rates, which is often related to both the exposure and outcome of interest. 3. Bias: EHR data sources face many biases, but are particularly vulnerable to informative missingness, differential monitoring, and model misspecification. 4. Large sample size: High precision (i.e., narrow confidence intervals) paired with potential bias leads to a high risk of incorrectly rejecting the null hypothesis. 5. Defining index time: It is important that investigators deliberately define index time (i.e., t 0 , baseline) to ensure that they only adjust for baseline confounders and do not adjust for (or condition on) factors that are affected by the exposure of interest (i.e., colliders or mediators). 6. Parameter selection: Investigators should only select parameters that are supported by the data distribution. This manuscript provides an overview of these challenges and solutions, using both simulated data and real-world data, with the outcome of Long COVID as the running example.

  PublisherOA PDFDOI

• Knowledge and attitudes of commercial motorcyclists on personal protective equipment in the crisis-affected Limbe and Tiko Health Districts of Cameroon toward injury prevention: A formative research study for a health education intervention

  medRxiv · 2025-05-07 · 1 citations

  preprintOpen access

  Abstract Background In Cameroon, commercial motorcycle riders are essential for urban transportation; however, they face considerable health risks from road traffic accidents and workplace hazards. Despite the critical role of personal protective equipment (PPE) in reducing injury risks, riders often possess limited knowledge and attitudes towards PPE. This study aimed to assess the knowledge and attitudes of motorcycle riders in the Limbe and Tiko Health Districts. Methods A community-based cross-sectional study was conducted with 499 commercial motorcycle riders aged 18 and older in these districts from the 15 th of May 2024 to the 17 th of August 2024. Participants were selected through consecutive sampling at motorcycle pick-up locations after obtaining ethical approval from the University of Buea, with severely ill individuals excluded from the study. Trained research assistants administered structured questionnaires to gather data on socio-demographics, riding habits, and riders’ knowledge and attitudes regarding PPE use. Data analysis was performed using descriptive statistics with SPSS version 25, and Bloom’s Criteria was applied to classify participants’ knowledge as good or poor. Results The average age of the 499 riders was 32.2±7.6 years, all of whom were male, with 48.5% aged between 21-30 years. Only 32.1% held a valid motorcycle license, and 37.1% were internally displaced due to the ongoing socio-political crisis in the two English-speaking regions of Cameroon. The findings showed that 30.7% of riders had good knowledge of PPE, 26.1% displayed positive attitudes, while only 13.2% practiced good PPE usage. This study highlights significant deficiencies in knowledge, attitudes, and practices related to PPE among motorcycle riders in Limbe and Tiko Health Districts, underscoring the necessity for targeted health education interventions to enhance their understanding and usage of PPE, ultimately improving safety and reducing injuries among riders. What is known in the Topic Commercial motorcycle riders are essential for urban transportation but face significant health risks from road traffic accidents and workplace hazards. Personal protective equipment (PPE) is critical in reducing injury risks among motorcycle riders. There is a general lack of awareness and understanding regarding the importance of PPE among motorcycle riders. What this study adds This study provides empirical data on the knowledge, attitudes, and practices regarding PPE usage among commercial motorcycle riders specifically in the Limbe and Tiko Health Districts of Cameroon. It highlights the low percentage of riders with good knowledge (30.7%), positive attitudes (26.1%), and proper practices (13.2%) related to PPE usage, indicating significant deficiencies in these areas. The findings emphasize the urgent need for targeted health education interventions to improve understanding and usage of PPE among motorcycle riders, aiming to enhance safety and reduce injuries in this population.

  PublisherOA PDFDOI

• Experimental and ethnographic evidence on household-based lockboxes and savings nudges from rural Tanzania

  World Development Perspectives · 2025-08-07

  article
  PublisherDOI

• Deconstructing the (un)affordability of clean cooking fuels through a randomized trial in rural Tanzania

  Nature Energy · 2025-06-03 · 5 citations

  articleSenior author
  PublisherDOI

• POS1271 ENHANCING PATIENT CARE AND CLINICAL CAPACITY IN INFLAMMATORY ARTHRITIS WITH AN INTEGRATED ELECTRONIC PATIENT-REPORTED OUTCOME SYSTEM

  Annals of the Rheumatic Diseases · 2025-06-01

  articleOpen accessSenior author

  <h2>Abstract</h2><h3>Background:</h3> The management of inflammatory arthritis (IA), including Rheumatoid Arthritis (RA), Axial Spondyloarthritis (axSpA), and Psoriatic Arthritis (PsA), requires accurate disease activity monitoring and timely interventions to achieve optimal outcomes. Traditional paper-based patient-reported outcomes (PROMs) often lacked completeness, hindered workflow efficiency, and limited their utility in routine practice. To address these challenges, the Rheumatology Department at Royal Berkshire NHS Foundation Trust implemented a fully integrated electronic patient-reported outcomes (ePROs) system within its digital care pathways. This system, designed for scalability and personalisation, enhances disease management while optimising clinic capacity. <h3>Objectives:</h3> This study aimed to evaluate the effectiveness of a digitally integrated ePRO system by increasing PROM completion rates for comprehensive disease monitoring, reducing routine follow-ups through patient-initiated follow-ups (PIFU), maintaining a low new-to-follow-up ratio to optimize capacity and enhancing clinician-patient efficiency to prioritise high-need cases. <h3>Methods:</h3> Between January 2022 and August 2024, ePROs were implemented via a patient engagement platform integrated with the hospital electronic patient record (EPR). PROMs, such as BASDAI, RAPID-3, PSAID-12 and HAQ-DI were sent to patients pre-visit, enabling remote completion and longitudinal monitoring of disease activity. A dynamic dashboard visualised trends and stratified patients by severity, facilitating timely interventions. Metrics such as completion rates, follow-up intervals, clinician time saved, and patient satisfaction were analysed. Comparisons were made with traditional workflows to highlight improvements. <h3>Results:</h3> Electronic PROs were sent out to 2151 RA, 910 PsA and 722 axSpA patients during the period of the study. Patient Engagement: Completion rates exceeded 60%, with condition-specific rates of 62.8% (AS), 60.5% (RA), and 62.5% (PsA). These rates significantly exceeded pre-digital completion rates (<30%). Clinic Efficiency: The total clinic time saved over 36 months was 1,188 hours, or 396 hours annually. New-to-Follow-Up Ratio: the service achieved a ratio of 2.5, outperforming the national average of 4.2. User Feedback: Over 90% of clinicians rated the system as highly effective for integrating patient data into decision-making. Patients reported greater satisfaction, citing reduced travel and improved convenience. <h3>Conclusion:</h3> This study demonstrates the transformative potential of integrating ePROs into IA care pathways. By enhancing disease monitoring, reducing in-person visits, and empowering patients as active participants in their care, the system significantly improved outcomes and clinic efficiency. The ePRO completion rates were above the average reported in literature (50-55%) [1]. The new-to-follow up ratio was lower than the national average with some hospitals reporting ratios as high as 10. The introduction of PIFU and virtual follow-ups saved clinical time. This allowed for reallocating resources to high-need patients, enabling timely reviews for flares or complex cases. This supports the drive to see the right patient at the right time [2]. Key challenges, such as digital literacy and data security, were mitigated through targeted training and robust infrastructure. Future plans include scaling this model to other conditions, such as SLE, and integrating predictive analytics to detect flares earlier, further optimising personalised care. <h3>REFERENCES:</h3> [1] Arumalla et al., Arthritis Rheumatol. 2023; 75:1892-1903. [2] Chan et al., Nat Rev Rheumatol. 2023 Nov;19(11):680-681. Figure 1The traditional versus digital pathway used with the implementation of an integrated electronic patient reported outcome (ePRO) platform. <h3>Acknowledgements:</h3> <b>NIL</b>. <h3>Disclosure of Interests:</h3> Antoni Chan Speakers bureau for UCB, Abbvie, Novartis, Janssen, Amgen, Kathryn Rigler: None declared, Nneoma Zabbey: None declared, Mustansar Hussain: None declared, Andrew Hubbard: None declared. © The Authors 2025. This abstract is an open access article published in Annals of Rheumatic Diseases under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Neither EULAR nor the publisher make any representation as to the accuracy of the content. The authors are solely responsible for the content in their abstract including accuracy of the facts, statements, results, conclusion, citing resources etc.

  PublisherDOI

• Protocol: the International Milk Composition (IMiC) Consortium - a harmonized secondary analysis of human milk from four studies

  Frontiers in Nutrition · 2025-06-10 · 8 citations

  articleOpen access

  Introduction: Human milk (HM) contains a multitude of nutritive and nonnutritive bioactive compounds that support infant growth, immunity and development, yet its complex composition remains poorly understood. Integrating diverse scientific disciplines from nutrition and global health to data science, the International Milk Composition (IMiC) Consortium was established to undertake a comprehensive harmonized analysis of HM from low, middle and high-resource settings to inform novel strategies for supporting maternal-child nutrition and health. Methods and analysis: = 290). Altogether IMiC includes 1,946 HM samples across time-points ranging from birth to 5 months. Using HM-validated assays, we are measuring macronutrients, minerals, B-vitamins, fat-soluble vitamins, HM oligosaccharides, selected bioactive proteins, and untargeted metabolites, proteins, and bacteria. Multi-modal machine learning methods (extreme gradient boosting with late fusion and two-layered cross-validation) will be applied to predict infant growth and identify determinants of HM variation. Feature selection and pathway enrichment analyses will identify key HM components and biological pathways, respectively. While participant data (e.g., maternal characteristics, health, household characteristics) will be harmonized across studies to the extent possible, we will also employ a meta-analytic structure approach where HM effects will be estimated separately within each study, and then meta-analyzed across studies. Ethics and dissemination: IMiC was approved by the human research ethics board at the University of Manitoba. Contributing studies were approved by their respective primary institutions and local study centers, with all participants providing informed consent. Aiming to inform maternal, newborn, and infant nutritional recommendations and interventions, results will be disseminated through Open Access platforms, and data will be available for secondary analysis. Clinical trial registration: ClinicalTrials.gov, identifier, NCT05119166.

  PublisherOA PDFDOI

Recent grants

• In situ destruction of halogenated Superfund contaminants with persulfate-generated radicals

  NIH · $84.5M · 1997–2027

• NIH Grant P20ES018173

  NIH · $1.4M · 2014

Frequent coauthors

• John M. Colford

  Berkeley Public Health Division

  145 shared

• Benjamin F. Arnold

  Global Brain Health Institute

  138 shared

• Brenda Eskenazi

  Center for Environmental Health

  114 shared

• Martyn T. Smith

  99 shared

• Asa Bradman

  University of California, Merced

  93 shared

• Jade Benjamin‐Chung

  Stanford University

  91 shared

• Kim G. Harley

  Center for Environmental Health

  82 shared

• Dana Boyd Barr

  Emory University

  79 shared

Labs

• Center for Computational BiologyPI