Overview
Our team is developing and applying the methods of causal inference and network science to address some of the most pressing public health and educational challenges, such as opioid use disorder treatment among people who use drugs, HIV prevention among men who have sex with men, and improving learning in statistics education. We recognize in these contexts that interventions may not only benefit the treated patients or exposed students and participants, but also their partners and contacts. Participants in intervention studies are often part of underlying networks even if networks are not explicitly measured or a focus of the study. Importantly, spillover likely occurs in these networks and is largely ignored in the evaluation of HIV interventions and substance use disorder treatment, often underestimating the full intervention impact. Measuring spillover will address the high burden of HIV in marginalized populations because it can expand the reach of HIV intervention benefits to individuals not connected to traditional research and clinical settings while strengthening communities of individuals impacted by HIV. Our research team is currently supported by two awards from the National Institutes of Health (NIH: R01DA058994, R01MH134715) and was also supported by the Avenir Award from the NIH National Institute on Drug Abuse (DP2DA046856).
Acknowledgments and Funding
Funding Agency: National Institute on Drug Abuse
Title: Network-based study design, statistical, and modeling solutions for HIV among populations that use illicit substances: Informing interventions and policy in real-world settings using existing data (R01 Year 1)
NIH NIDA Grant # R01DA058994
PI: Buchanan
Project Period: 09/15/24-7/31/29
Funding Agency: National Institute of Mental Health
Title: Developing Causal Inference Methods to Evaluate and Leverage Spillover Effects through Social Interactions for Designing Improved HIV Prevention Interventions
NIH NIMH Grant# R01MH134715
PI: Forastiere
Project Period: 07/01/23-06/30/28
Funding Agency: National Institute on Drug Abuse
Title: Diversity Postdoctoral Fellowship for Parent Project Causal Inference Methods for HIV Prevention Studies Among Networks of People Who Use Drugs
NIH NIDA Grant # DP2 DA046856
PI: Buchanan
Project Period: 07/01/19 – 05/31/21
Funding Agency: National Institute on Drug Abuse
The Avenir Award Program for Research on Substance Use Disorders and HIV/AIDS (DP2)
Title: Causal Inference Methods for HIV Prevention Studies Among Networks of People Who Use Drugs
NIH NIDA Grant # DP2 DA046856
PI: Buchanan
Project Period: 07/01/18 – 05/31/23
Funding Agency: University of Rhode Island Proposal Development Grant
Title: Causal Inference Methods for HIV Prevention Studies of Multifaceted Interventions Among Networks of People Who Inject Drugs
PIs: Buchanan and Katenka
Project Period: 07/01/18 – 06/30/19
Funding Agency: Rhode Island Foundation Medical Research Grant
Title: New Methods to Evaluate the Effects of Opioids in Provider-Based Networks
PI: Buchanan
Project Period: 04/01/17 – 09/30/18
Funding Agency: Brown University Advance CTR Pilot Project
Title: New Methods to Evaluate the Effects of Opioids in Provider-Based Networks
NIGMS Grant # U54GM115677
PIs: Buchanan and Bratberg
Project Period: 05/01/18 – 04/30/19
Team
We are the Networks and Causal Inference for Public Health and Education Research team at the University of Rhode Island. Our research involves statistical methods development combining causal inference and network analysis approaches to address high priority questions in public health and education.
Team Members
Current team members
Director, Data Science and Associate Professor
Dept. of Computer Science and Statistics
Graduate Chair, Statistics and Associate Professor
Dept. of Computer Science and Statistics
Past team members
Post Doctoral Fellow
Graduate Research Assistant
Collaborators
Assistant Professor of Biostatistics on the Teaching Scholar Track
Brown University
Founding Director of the People, Place & Health Collective (PPHC), and Professor of Epidemiology
Brown University School of Public Health
Vice Chair, Department of Health, Behavior, and Society Professor
John Hopkins Bloomberg School of Public Health
Susan Dwight Bliss Professor of Biostatistics; Professor, Department of Statistics and Data Science; Founding Director, Center for Methods in Implementation and Prevention Science (CMIPS); Director, Interdisciplinary Research Methods Core, Center for Interdisciplinary Research on AIDS (CIRA); Assistant Director, Global Oncology, Yale Cancer Center; Affiliated Faculty, Yale Institute for Global Health
Yale School of Public Health
Assistant Professor in Epidemiology
Boston University School of Public Health
Professor and Head of Biostatistics, Bioinformatics and Epidemiology Program Vaccine and Infectious Disease Division, and Professor Public Health Sciences Division
Fred Hutch Cancer Center
Associate Professor in Biostatistics
John Hopkins Bloomberg School of Public Health
Associate Professor of Biostatistics, Associate Professor of Ecology and Evolutionary Biology, Associate Professor of Management, and Associate Professor of Statistics and Data Science; Co-director, Public Health Modeling Concentration
Yale School of Public Health
Assistant Professor of Epidemiology and Public Health at the Medical School
Medical School, University of Cyprus, Nicosia, Cyprus
Researcher
Medical School, University of Cyprus, Nicosia, Cyprus
Professor of Statistics
Department of Mathematics & Statistics, University of Cyprus
Assistant Professor of Biostatistics
Yale Institute for Global Health
Assistant Professor
Université du Québec à Montréal (UQUAM)
Associate Professor Community Health and Social Sciences
City University of New York School of Public Health
Research Scientist in Biostatistics; Associate Director, Interdisciplinary Research Methods Core, Center for Interdisciplinary Research on AIDS; Affiliated Faculty, Yale Institute for Global Health
Center for Methods in Implementation and Prevention Science, Yale School of Public Health
Senior Theoretician and Associate Director of the Infectious Disease Epidemiology and Theory Core at the Center for Drug Use and HIV Research (CDUHR), and Research Professor, Department of Population Health
Center for Opioid Epidemiology and Policy in the Department of Population Health at NYU Grossman School of Medicine
Professor
College of Pharmacy, URI
Manning Assistant Professor of Biostatistics
Brown University
News
- 2023 Presentation at IBS
Congratulations to Dr. Youjin Lee for presenting about the paper, “Finding influential subjects in a network using a causal framework,” at the International Biometric Society (IBS) Journal Club in April 2023!
- 2023 Pathogens Paper!
Congratulations to Dr. Ashley Buchanan et. al. whose paper Methods for Assessing Spillover in Network-Based Studies of HIV/AIDS Prevention among People Who Use Drugs was accepted in the journal Pathogens and published in February 2023.
- 2023 NESS Presentation
Congratulations to Dr. Natallia Katenka for her presentation “Power and sample size calculations for evaluating spillover effects in networks with non-randomized interventions” at the New England Statistical Symposium (NESS), Boston, MA in June 2023!
- 2023 Johns Hopkins University Presentation
Congratulations to Dr. Natallia Katenka for her presentation “Power and sample size calculations for evaluating spillover effects in networks with non-randomized interventions” at the Department of Biostatistics, Johns Hopkins University, Baltimore, MD in February 2023!
- 2023 CUNY Presentation on RDS
In January 2023, Gabrielle Lemire presented “Regression Methods for Respondent Driven Samples (RDS): An Application to Young Adults Using Presciption Opioids Non-Medically in NYC” for the Institute for Implementation Science in Population Health (ISPH) at City University of New York (CUNY). You can find more information here.
- 2023 American Causal Inference Conference (ACIC) Poster Presentation
Congratulations to Gabrielle Lemire for presenting the poster “Simulating Potential Outcomes in the Presence of Interference” at ACIC in May 2023.
- 2023 “Causal Inference in Networks” Virtual Spring Workshop
Congratulations to Dr. Ashley Buchanan, Ke Zhang, and Gabrielle Lemire who presented during the Lightnight Talks, and ran a code demo, at “Causal Inference in Networks: Applications to Public Health” Spring Virtual Workshop in March 2023. Thank you to the other speakers who also presented: Drs. Laura Forastiere, Samrachana Adhikari, and Elizabeth Ogburn. We had over 100 individuals register and attendees were from all over the globe!
- 2022 Uni of Pittsburgh Presentation
In February 2022, Dr. Natallia Katenka presented “Estimating Causal Effects of Non-Randomized HIV Prevention Interventions with Spillover in Network-based Studies among People who Inject Drugs” for the Department of Statistics at the University of Pittsburgh, PN.
- 2022 Student Presentations at URI Conference
Congratulations to graduate students Ke Zhang, Ryan Humphreys and Zhejia Dong who presented at the research conference “Advancing Data Analysis in Interdisciplinary Studies: Deciphering the World’s Black Boxes” on October 21, 2022 at the University of Rhode Island (URI) in collaboration with the HAW Hamburg. An additional thanks to Dr Katenka for co-leading and co-facilitating the conference.
- 2022 Presentation at CMIPS and MSU
In December 2022, Gabrielle Lemire presented “Simulating Potential Outcomes in the Presence of Interference” at Center for Methods in Implementation and Prevention Science (CMIPS) at the Yale School of Public Health and for the Student Chapter of the American Statistical Association (ASA) at Montana State University (MSU). Congrats!
- 2022 Presentation at 9th Annual ASA Workshop
In October 2022, Dr. Natallia Katenka presented “Causal Inference in Networks with Applications in Public Health” at the 9th Annual ASA New Jersey Chapter and Bayer Statistics and Data Insights workshop.
- 2022 Paper Accepted
Congratulations to Dr. Buchanan et al. whose paper “Spillover effects of pre-exposure prophylaxis delivery for HIV prevention: Evaluating the importance of effect modification using an agent-based model” has been accepted in the journal of Epidemiology and Infection.
Buchanan AL, Park, CJ, Bessey S, Goedel WC, Murray EJ, Friedman SR, Halloran ME, Katenka NV, Marshall BDL. Spillover effects of pre-exposure prophylaxis delivery for HIV prevention: Evaluating the importance of effect modification using an agent-based model. Epidemiology and Infection. In Press, 2022.
- 2022 JSM Sessions and Presentations
Dr. Youjin Lee organized the invited session “Advances in Social Network Analysis for Public Health Solutions” at 2022 Joint Statistical Meetings (JSM) titled “Advances in Social Network Analysis for Public Health Solutions,” with Gabrielle Lemire as discussant and Dr. Natallia Katenka as an invited speaker. Dr. Katenka’s talk is titled “New Approaches to Model Multi-Level Social Networks for Evaluating the Impact of Demographics on Health Disparities of Epidemic Trajectories in the NYC Population.”
Gabrielle Lemire also presented “Bias Correction for Sampled Genetic Network Data” as a part of the Bayesian Methods and Social Statistics Speed Session.
- 2022 Invited Speaker at BU Seminar
Dr. Ashley Buchanan was an invited speaker at the Causal Inference Seminar Boston University in April 2022 to speak on “Estimating Causal Effects of Non-Randomized HIV Prevention Interventions with Spillover in Network-based Studies among People who Inject Drugs”.
- 2022 International Symposium on Mathematical Methods Applied to the Sciences Conference Presentation
In February 2022, Dr. Anarina Murillo delivered a presentation titled “Novel Application of a Multistate Model to Evaluate the Opioid Use Disorder Care Cascade in the Rhode Island All-Payer Claims Database” at the International Symposium on Mathematical Methods Applied to the Sciences, SIMMAC XXIII.
https://eventos.cimpa.ucr.ac.cr/index.php/simmac/XXIIISIMMAC
- 2022 ENAR Spring Conference Presentations on Eliminating Health Disparities
In March 2022, Dr. Aroke organized an invited session that was accepted through a competitive process at the Eastern North American Region International Biometric Society 2022 Spring Meeting. The title of the session was “Novel Data Science Approaches to Eliminate Health Disparities”. The speakers included Drs. Natallia Katenka, Anarina Murillo, and Hilary Aroke.
Dr. Aroke presented work titled “Novel Application of a Multistate Model to Investigate Disparities in Access to Treatment for Opioid Use Disorder in Rhode Island”.
Dr. Natallia Katenka presented “New Approaches to Model Multi-level Social Networks for Evaluating the Impact of Demographics on Health Disparities of Epidemic Trajectories in the NYC Population” at the Eastern North American Region International Biometric Society 2022 Spring Meeting in March 2022.
https://www.enar.org/meetings/spring2022/
- 2022 BU Workshop Presentation and Tutorial
Dr. Murray and Dr. Buchanan presented at a workshop on simulation modeling for causal inference “Incorporating Causal Inference into Simulation Models.” The workshop included presentations on the estimation of causal effects in network models. Dr. Buchanan also developed and delivered a tutorial on spillover in agent based models (ABM) as part of this workshop, titled “Adding complexity: How to think about causal effects in systems models with interference or transmission”.
This workshop was hosted in conjunction with the Focused Research Program group at the Rafik B. Hariri Institute for Computing and Computational Science and Engineering at Boston University in April 2022.
- 2021 Welcome!
In 2021, we welcomed a new collaborator to our team, Dr. Youjin Lee from Brown University, an expert in network dependence. We also expanded collaborations across institutions to include Boston University (with site PI Dr. Eleanor Murray, an expert in causal agent-based models), and University of Cyprus (with site PI Dr. Georgios Nikolopoulos).
- 2021 NESS Session and Presentations
In August 2021, Dr. Buchanan organized an invited session at the New England Statistical Society (NESS) symposium titled “Novel statistical approaches for public health and education studies with clustering and network features”. This session highlighted key work from the URI Avenir team and collaborators.
Dr. Youjin Lee presented “Partially Pooled Propensity Score Models for Average Treatment Effect Estimation with Multilevel Data”. Dr. Hilary Aroke presented “Novel Application of a Multistate Model to Investigate the Opioid Use Disorder Care Cascade among Medicaid Enrollees in Rhode Island”. Dr. TingFang Lee presented “Toward Evaluation of Disseminated Effects of Non-Randomized HIV Prevention Interventions Among Observed Networks of People who Inject Drugs”. Dr. Ashley Buchanan presented “Spillover of Pre-Exposure Prophylaxis for HIV Prevention: Evaluating the Importance of Effect Modification using an Agent-Based Model”.
- 2021 JSM Sessions and Presentations
Dr. TingFang Lee (Postdoctoral Fellow) organized a topic contributed session that was accepted through a competitive process at the American Statistical Association Joint Statistical Meetings (JSM) in August 2021. This symposium was titled “Evaluating Causal Effects Using Incomplete Data with Interference in Public Health Research” and brought together leading researchers in this area, including Drs. Costanza Tortù, Imke Mayer, TingFang Lee, and Junhan Fang. Dr. Samrachana Adhikari from the New York School of Medicine served as the Discussant.
Dr. Lee presented work on evaluating the impact of missing data when assessing dissemination titled “Evaluating Disseminated Effects of Incomplete Data in Network-Based Studies for Public Health”.
Hilary Aroke also presented “Novel Application of a Multistate Model to Investigate the Opioid Use Disorder Care Cascase in Rhode Island” at JSM 2021. This presentation was as a part of the session “Innovative Methods for Predicting Public Health Outcomes and Informing Policy” which was organized by Dr. Ashley Buchanan.
- 2021 Invited Speaker at NYU Seminar
Dr. Ashley Buchanan was an invited speaker at the Biostatistics Seminar at New York University in October 2021 on “Estimating Causal Effects of Non-Randomized HIV Prevention Interventions with Spillover in Network-based Studies among People who Inject Drugs”.
- 2021 Guest Appearance on Causal Inference Podcast
Dr Ashley Buchanan appeared as a guest on season 3 of the podcast Casual Inference. She discusses the importance of correctly estimating causal effects in dependent data structures, such as networks. Fun fact, this is how our Research Associate, Gabrielle Lemire, heard of the Avenir team!
You can find the episode here, or take a listen below:
- 2021 Congratulations!
Congratulations to Dr. Tingfang Lee on completing her post-doctoral work as part of the Avenir team! In her role, she built novel methods and statistical models to assess the impact of interventions and risk factors on individuals embedded in networks, specifically those pertaining to HIV and illicit drug use. Congratulations to Dr. Hilary Aroke on completing his post-doctoral work as a part of the Avenir team! Aroke was awarded the prestigious National Institute on Drug Abuse (NIDA) Diversity Post Doctoral Fellowship. During the past 2-3 years he has conducted epidemiologic and network science research using a variety of data sources. His current interest is in substance abuse research. Congratulations to Dr. Tianyu Sun on completing his PhD. His thesis title is “Evaluation of Medications for Opioid Use Disorder on Overdose and Healthcare Utilization in the US.”
- 2020 ENAR Conference Presentation
Dr. Buchanan presented “Toward Evaluation of Disseminated Effects of Non-Randomized HIV Prevention Interventions Among Observed Networks of PWID” at the Eastern North American Region International Biometric Society 2020 conference. She presents a novel method for estimation of disseminated effects among members of a network as well as approaches to bootstrap estimation of the variance of the novel estimator.
- 2020 Congratulations!
Congratulations to our recent graduates! Ben Skov graduated with a Master of Science in Pharmaceutical Science. Ben’s thesis is titled Evaluation of the Relationship between Centrality and Individual-Level Characteristics among PWID. Joseph Puleo graduated with a Master of Science in Statistics (2020). Joseph’s thesis is titled Evaluating the Effects of Medications for Opioid Use Disorder on HIV Risk Behaviors Among a Network of People Who Inject Drugs. Joseph’s current job is as a Biostatistician I at the Center for Biostatistics in AIDS Research, Harvard T.H. Chan School of Public Health. Valerie Ryan graduated with a Doctor of Philosophy in Behavior Science, Master of Science in Statistics, and Master of Arts in Behavior Science (2020). Valerie’s thesis title is Graph Models with Missing Data Techniques for Applications in Drug Use Networks. Valerie is now a Behavior Scientist at the National Center for Health Statistics.
- 2019 Speaker at UW Biostatistics Seminar
Dr. Ashley Buchanan presented at the Biostatistics Seminar at the University of Washington (UW) in 2019 on some of our research regarding the effect of injection partner beliefs on health seeking behavior.
- 2019 JSM Conference Presentation
Dr. Ashley Buchanan presented at Joint Statistical Meetings (JSM) 2019 as a part of the “Novel Statistical Methods for Network-Based Studies Among People Who Use Drugs” session. Her talk was titled “Toward Evaluation of Dissemination of HIV Prevention Interventions Among Networks of People Who Inject Drugs” and covers research on the application of cluster based disseminated effects methods to networked data. It included results from an analysis on the effects of individual attitudes towards HIV risk on engagement with HIV care among people who inject drugs (PWID) from the Social Factors and HIV Risk study.
Research
- Assessing Individual and Disseminated Effects in Network-Randomized Studies
This paper describes disseminated effects in network-randomized clinical trials and applies these methods to a study of injection drug users at risk for HIV.
- Causal Inference in Networks
This presentation is an introduction to the concepts used in our research. It briefly presents that potential outcomes framework used in causal inference, an introduction to network theory, and the concept of interference as it applies to health outcomes.
- Disseminated Effects in Agent Based Models: A Potential Outcomes Framework and Application to Inform Pre-Exposure Prophylaxis Coverage Levels for HIV Prevention
We developed an agent-based model using a novel trial emulation approach to quantify disseminated effects of PrEP use among men who have sex with men in Atlanta, GA. Components (subsets of agents connected through partnerships in a sexual network, but not sharing sexual partnerships any other agents) were first randomized to an intervention coverage level or control, then within intervention components, eligible agents were randomized to PrEP.
- Estimating Causal Effects of HIV Prevention Interventions with Interference in Network-based Studies among People Who Inject Drugs
This paper described methods to evaluate disseminated effects in sociometric network studies and applies these methods to network-based study of injection drug users at risk for HIV. The manuscript is submitted to Annals of Applied Statistics.
- Generalizing Evidence from Randomized Trials using Inverse Probability of Sampling Weights
This paper describes the use of inverse probability of sampling weights which are used to adjust the results of clinical trials to generalize to a given population of interest.
- Network Analysis of Collaborations Among Undergraduate Statistics Students
In the project, we investigate the drivers of peer collaborations in two undergraduate statistics courses at the University of Rhode Island.
Shared Resources
- Check out our GitHub Page
Check out our most recent work at https://github.com/uri-ncipher
- Evaluating causal effects in network-based studies in R
This tutorial IPW models includes organizing data for nearest neighborhood IPW estimator modeling. The functions output point estimations and Weld-type 95% confidence intervals under assigned allocation strategies of the average potential outcomes and causal effects.
GitHub:: https://github.com/uri-ncipher/Nearest-Neighbor-estimators
- Inverse probability Sample Weights in R
Below is the R code from the “Generalizing Evidence from Randomized Trials using Inverse Probability of Sampling Weights” Paper published by Dr. Buchanan.
Due to security restrictions the code is in a .txt file which can be copied into R
- Worth the Weight: IPW Cox Models Tutorial Code
In this tutorial, we demonstrate how inverse probability weighted Cox models can be used to account for multiple measured confounders, while concentrating inferences on the treatment or exposure effects of central interest and providing graphical summaries of these effects.
Projects
NIDA R01
We propose to evaluate the spillover of HIV interventions in marginalized populations, specifically young Black men who have sex with men (YBMSM) in Chicago, using both secondary data analyses of existing data and simulation approaches requiring the development of novel study design, inferential, and modeling methods. We will collaborate with a local community advisory board and AIDS Foundation Chicago to expedite the impact of project findings. Importantly, measuring spillover will address the high burden of HIV among YBMSM because it can expand the reach of HIV intervention benefits to the most marginalized individuals not connected to traditional research and clinical settings while strengthening communities of YBMSM.
NIH HPM2
We are developing innovative statistical methods for designing more effective HIV treatment and prevention interventions, along with more effective implementation strategies to deliver them, focusing on two large cluster randomized trials in Botswana and South Africa. The methods research related to package interventions is represented in the attached publication. To achieve global HIV control targets, innovative methodological tools are needed to effectively and efficiently quantify, control for, and leverage spillover effects of HIV prevention and treatment interventions.
URI Avenir
The goal of the Avenir project was to develop causal inference methodology combined in novel ways with network science to solve critical challenges at the nexus of HIV and substance use for network-based studies of HIV treatment and prevention among people who use drugs (PWUD). Participants in intervention studies are often part of underlying networks even if networks are not explicitly measured or a focus of the study. Importantly, spillover likely occurs in these networks and is largely ignored in the evaluation of HIV interventions and OUD treatment, often underestimating the full intervention impact. We developed methodology for both empirical analyses and simulation-based approaches to quantify the spillover effects of network-based interventions. Our methods and findings were disseminated through publications, conference presentations, corresponding statistical programs, a network database resource, training, and an open access tutorial. Finally, this project has supported multiple graduate students, postdoctoral research associates, and a research associate who have continued to internships and employment as (bio)statisticians and epidemiologists.
Stats Teaching
For substantive work in statistics education research, we evaluated peer collaborations among statistics students in two undergraduate courses at the University of Rhode Island. The primary tools include network-based visualization, characterization, and modeling methods including Exponential Random Graph Models (ERGMs). The results from this study can provide data-driven approaches to improve statistics pedagogy to engage students successfully.