DGHI Research Design & Analysis Core

Formerly known as the DGHI Biostatistics Core

Strengthening global health research through rigorous design and analysis.

Mission:  The DGHI Research Design & Analysis Core uses rigorous design and analytic methods to strengthen global health research and education through meaningful collaboration with its faculty and students.


  • Our mission and work extends to both faculty research and graduate student education and training.
  • The Core has seven members, including 1 PhD biostatistician, 3 PhD epidemiologists, 3 master’s level biostatisticians, and one affiliate member who is a PhD candidate in health policy.
  • Leadership is provided by Core Director, Liz Turner, PhD, and Associate Core Director Joe Egger, PhD. Oversight is provided by John Bartlett, MD, DGHI Associate Director for Research.
  • We collaborate with DGHI faculty on grant proposal development and writing. DGHI faculty members are able to draw on our extensive experience in global health research projects by naming the Core and its members on grant proposals.
  • We focus on quantitative methods and have a deep appreciation for qualitative and mixed methods research.
  • We maintain strong links to Duke’s Department of Biostatistics & Bioinformatics  and its own Biostatistics Core, which is the home department of our four biostatisticians.
  • We provide referrals to Duke Library services for DGHI, The Edge, SSRI’s Connection Bar, Duke DTMI Biostatistics Core and other Duke and external organizations that provide complementary services.
  • The Core collaborates and works closely with DGHI’s Evidence Lab to perform rigorous evaluations of global health importance.

Faculty Request FormStudent Request Form

Founded as the DGHI Biostatistics Core in 2012, our name was changed in 2015 to better reflect the importance of our role in meeting DGHI’s research and educational mission.



Collaboration with DGHI Faculty

The Research Design & Analysis Core collaborates with DGHI faculty on the design and analysis of population health research studies. Core members Joe Egger, John Gallis, Larry Park, Alyssa Platt, Ryan Simmons and Liz Turner are available to meet with any DGHI faculty member regarding ongoing and new research proposals. DGHI affiliates, research scholars, postdoctoral scholars/associates, and visiting scholars may request Core support as resources permit. For new grant proposals, our four PhD team members (Joe Egger, Asia Maselko, Larry Park and Liz Turner) are available to be named as co-investigators on new grant proposals and our MS team members are available to further strengthen those new grant proposals by providing high quality in-house quantitative support through funded grant support. Types of research activities include, but are not limited to:

  • Grant preparation, including the possibility of a statistician or epidemiologist named as a co-investigator. Please note: Requests for grant support should be submitted within 5 days of notification to DGHI finance of intention to submit a grant (i.e., at least 90 days prior to the grant’s due date).
  • Grant preparation, including the possibility of listing a Master's level biostatistician on proposal
  • Conceptualization and study design, (e.g., methods, analysis plan)
  • Data collection and sampling (e.g. power/sample size calculations)
  • Reliability/validity analysis (e.g. informal or formal bias analysis)
  • Statistical analysis
  • Manuscript preparation
  • General statistical advice
  • General epidemiological advice

Support is typically offered by meeting face-to-face in Durham, NC. When that is not possible, email or web-based (e.g., Skype, Google) consultation is available.

  • An initial consultation, subsidized by DGHI, is offered free of charge for appropriate requests. If a significant time investment is required, the faculty or affiliate will need to provide funding. Depending on the time commitment, some additional support can be offered free of charge. For instance, assistance with preparation of a grant application may be provided; however, the grant application should include a budget request for continued PhD-level and MS-level biostatistical or epidemiological support after the grant is funded.

Online Request Form for Faculty/Affiliates

Support for DGHI MSc-GH Students

In addition to faculty input provided through MSc-GH course curricula and faculty mentorship, the DGHI Research Design & Analysis Core also offers statistical and epidemiological advice to students related to their MSc-thesis projects. Liz Turner and Larry Park provide support to MSc-GH students through their teaching of core methods courses. Liz Turner leads the fall course Biostatistics and Epidemiology I, whilst Larry Park leads the accompanying data analysis lab and leads the follow-on spring course Biostatistics and Epidemiology II. Core members Joe Egger and Joanna (Asia) Maselko are available to meet with students, and provide support in the following areas:

  • Conceptualization and study design (e.g. methods, analysis plan, power/sample size calculations)
  • Human Subjects (IRB) review
  • Questionnaire design
  • Data entry forms and/or solutions
  • General epidemiological or statistical consultation
  • Questions about data structure or data cleaning
  • Questions about preliminary exploratory analyses
  • Statistical analysis methods/solutions
  • Referral services for students, including Duke Libraries, ORS/IRB, and others.

To set-up a meeting with Asia or Joe, first fill out a student request and they will get in touch with you. Please plan ahead and submit a form as early as possible, as requests are taken in the order in which they are received.

Online Request Form for Students

Undergraduate Global Health students may seek support for quantitative questions from the Academic Resource Center and SSRI’s Connection Bar.


Our Team

The DGHI Research Design & Analysis Core is made up of Joe Egger, PhD (Associate Director), John Gallis, SCM, Joanna (Asia) Maselko, ScD, Larry Park, PhD, Alyssa Platt, MA, Ryan Simmons, MB, and Liz Turner, PhD (Director) and affiliate member Charles Muiruri, MPH.

Liz Turner, PhD, Director

Liz has served as the Core Director since its creation in March 2012 when she joined Duke’s Global Health Institute and Duke’s Department of Biostatistics & Bioinformatics. With a PhD in statistics from McGill University, Canada, followed by four years working as a collaborative biostatistician in the Department of Medical Statistics, Faculty of Epidemiology and Population Health of the London School of Hygiene and Tropical Medicine (LSHTM), Liz has extensive experience working in both epidemiological studies and randomized trials across a range of substantive areas in developed world and resource poor settings. Thanks to her participation in multi-disciplinary projects, she has a great appreciation for the importance of good study design and data collection and is well aware that no fancy statistical analyses can save researchers from the scourge of bad data. Through those experiences and her teaching in different settings, including the UK, Canada, France and Tanzania, she is aware that statisticians and their collaborators sometimes “speak a different language”. As a result, her approach is very much one of translation, pragmatism and collaboration.

Her current substantive interests include malaria, disability and disease burden with an emphasis on eye diseases, cardiovascular disease and mental health, together with child health and education.

Recent research collaborations
  • The International Centre for Eye Health, LSHTM
    Estimation of cataract incidence in order to plan for the surgical volume required to prevent blindness in line with the VISION 2020 goals. Data from three large national prevalence surveys of blindness in Bangladesh, Nigeria and Pakistan were used to estimate incidence and predict future disease burden.

    Ifakara Health Institute, Tanzania
    Research projects in public health entomology in Tanzania including experimental field trials to evaluate registered spatial repellents as potential replacements for DDT against malaria.

    Kenyan Medical Research Institute, Nairobi, & Harvard Graduate School of Education
    Statistician to the Health and Literacy Intervention Trial, a factorial cluster randomized trial to evaluate the impact of malaria control and enhanced literacy instruction on health and educational achievement of school children in Kenya. More than 5000 children from 101 schools on the south-east coast were enrolled.

    National Heart and Lung Institute, Imperial College, London, UK
    Research in preventive cardiology including a national survey to evaluate lifestyle and risk factor management in coronary patients and people at high risk of developing cardiovascular disease, secondary analyses of a European cluster randomized trial of a lifestyle intervention to improve cardiovascular (CV) health and a case-control study of the effects of exposure to violence on CV disease.

    European Huntington’s Disease Network
    Use of longitudinal registry data from more than 2000 people with a range of measures such as motor function, cognition and depression, to develop recommendations for how best to design a forthcoming trial.


Joe Egger, PhD, Associate Director

Joe joined Duke’s Global Health Institute in June 2013. Joe is an epidemiologist and collaborates with, and provides epidemiological and statistical support to DGHI students, faculty and affiliates. Prior to joining Duke, Joe was the lead epidemiologist for SciMetrika, a public health consulting firm based in Research Triangle Park, NC. Joe has also worked in academic and government settings on a wide array of public health research topics, including influenza, HIV/AIDS, dengue, tobacco, obesity/nutrition, and cancers. Joe has a master’s degree in medical geography from the University of Washington, and a PhD in infectious disease epidemiology from the London School of Hygiene & Tropical Medicine.

Joe’s research interests include vector-borne disease ecology, disease transmission dynamics, and more recently, the use of electronic health record data in epidemiological studies and disease surveillance.

Recent research collaborations
  • Epidemiologic biases in post-marketing safety surveillance
    Co-investigator on an FDA-funded project to study the impact of bias parameters (e.g., misclassification, confounding, selection bias) on epidemiologic studies. The study is adapting existing probabilistic bias analysis techniques for use in post-licensure medical product safety surveillance.

    Traumatic Brain Injury
    Co-investigator with CDC’s National Center for Injury Prevention and Control to study traumatic brain injury in American youth. Designing a study to calculate the sensitivity of relevant ICD-9 CM codes to accurately detect TBI in multiple clinical settings using electronic health record data. Results will be used to improve the sensitivity of CDC’s traumatic brain injury surveillance system. A secondary objective of the project is the study predictors of concussion in high school sports using a nationally-representative sample of US high schools. 

    Health-Related Quality of Life
    Researching the relationship between health-related quality of life and the adoption and effectiveness of interventions to prevent chronic diseases. Specifically, the study uses data from the Behavioral Risk Factor Surveillance System and National Health Interview Survey to model whether HRQOL moderates the association between an intervention to prevent chronic diseases (e.g., lung cancer from tobacco use) and success in that intervention (e.g., tobacco cessation program).

    Expanding Healthcare Data Access and Use
    Co-investigator on a two-year study with CDC’s Office of Surveillance, Epidemiology, and Laboratory Services (OSELS) to systematically evaluate the use of several large healthcare data sources (e.g., medical claims, electronic health records) to augment public health surveillance activities at CDC. A component of the evaluation was to determine the ability of each data source to monitor the impacts of Health Reform, as enacted under PPACA and ARRA/HITECH, on the US healthcare system.

    Co-investigator on a multi-year longitudinal study of the effects of changes in fitness and body mass index on academic outcomes in New York City middle school students. By using data from a large and heterogeneous population the study also assessed the impact of socio-economic status, race/ethnicity and gender on the fitness-academic relationship.

    Seroprevalence of pandemic influenza viruses, New York.
    Data from the 2004 NYC Health and Nutrition Examination Survey, a representative sample of New York City adults, were used to determine the seroprevalence of preexisting hemagglutinin inhibition (HAI) antibody titers to influenza strains with pandemic potential. Serum samples were tested for antibodies to A(H1N1)pdm09 virus and the 1918, 1957, and 1968 pandemic viruses. Statistical modeling was used to assess the relationship between antibody titers to previously circulating viral strains and titers to A(H1N1)pdm09 virus. The results revealed that immunologic background during the emergence of A(H1N1)pdm09 virus in NYC beginning in late April 2009 help explain why fewer cases of A(H1N1)pdm09 infection were clinically detected among older persons than younger persons, supporting the conclusion that the difference was a result of, at least in part, antibodies elicited by prior H1N1 subtype infection in older persons.

    Non-pharmaceutical interventions on rates of influenza in schools
    Studied the effect of reactive school dismissal in New York City during the 2009 influenza pandemic on rates of influenza transmission in those schools, compared to schools that did not enact school dismissal policies.  Results indicated that if non-pharmaceutical interventions, such as school dismissal, are employed after cases are detected through passive surveillance, the effect on reducing the burden of influenza will be minimal.


Asia Maselko, ScD

Joanna (Asia) Maselko joined DGHI and the Dept. of Psychiatry and Behavioral Sciences in August 2009. Asia is a social epidemiologist whose research uses a developmental and epidemiological approach to study the mechanisms through which social and cultural factors influence mental health over the lifecourse. One specific area of interest is trying to identify how the family and social context shapes the intergenerational transmission of psychopathology risk. Towards this end, Dr. Maselko is currently evaluating the long term cognitive and socio-emotional outcomes among school-aged children whose mothers participated in a community based perinatal depression intervention in Pakistan. Dr. Maselko also leads a project that examines the family and social context determinants of well-being and cognitive decline among elderly in Sri Lanka. Another area of research focuses on how religious engagement facilitates risk or resilience in the context of socioeconomic adversities. With an emphasis of on disentangling causal processes, Maselko's previous research has described the potential contribution of selection effects in the observed religious engagement-depression association.

Recent research collaborations
  • Family Composition and Risk of Mental Health Problems in the Offspring
    This pilot project looks at assessing social support and the stimulating home environment in the South Asian setting. The role of family, social support, behaviors of mothers and other caregivers, and the home environment of infants are explored to gain insight into factors that may affect mothers and infants' well-being during the perinatal period.

    Youth in Transition: Developmental Pathways to Risk and Resilience in South India
    A project to develop hypotheses for the association among the school environment, life course models, and neuropsychological functioning. This study addresses the public health and education policy issue of the relationship between the educational environment and mental health employing three methodological perspectives: anthropology, pediatric neuropsychology, and social epidemiology. The project will use ethnographic methods to develop a quantitative assessment of life-course models of protective and risk factors for adolescent well-being.

    Perinatal Depression Treatment and Child Development
    A clustered randomized trial where lay Lady Health Workers (LHWs) were trained to deliver a cognitive behavior therapy (CBT) intervention to mothers with major depression. Research will examine whether participating in a successful perinatal depression intervention (Thinking Healthy Programme, THP) leads to improved developmental outcomes in offspring. The THP was delivered in Pakistan in 2006; in 2012 the children will be 5-6 years old and entering kindergarten, thus allowing for an assessment of a wide range of cognitive, socio-emotional, motor and physical developmental outcomes.

    The Sri Lanka Healthy Minds Study
    The goal of this study is to lay the foundation for the Sri Lanka Healthy Minds Study, a population-based study where individual, caregiver, and other family level factors are being examined using a lifecourse perspective to better understand depression risk and its influence on cognitive impairment.

    Religious Engagement and Psychopathology over the Lifecourse
    This study examines the relationship between religious engagement and psychopathology over the life-course by taking advantage of multiple waves of data on both religious engagement and psychopathology that cover several decades. The study uses survival analysis techniques, analyzing data from three follow-up studies of the New England Cohorts of the National Collaborate Perinatal Project (NCPP).


Larry Park, PhD

Larry Park, Ph.D. is affiliated with the Duke Department of Medicine and the Durham VA Medical Center. Dr. Park joined the Duke Global Health Institute in 2013 as adjunct faculty and the DGHI Biostatistics Core in July 2014. He is the statistical analysis lab instructor for Quantitative Methods I and course instructor for Quantitative Methods II. Dr Park holds a Ph.D. in epidemiology from UNC Chapel Hill and Master’s degrees in both computer science and electrical engineering from the Johns Hopkins School of Engineering. He has spent many years in basic science (protein chemistry, molecular biology, immunology and viral evolution).

His current research interests and collaborations include respiratory infectious disease, HIV and HCV infection and their effects on chronic disease, cardiovascular disease and endocarditis, healthcare cost and utilization, hospital infection control and prevention, the application of advanced statistical methods in epidemiology, analysis of longitudinal data, survival analysis, and methods for causal inference in observational epidemiology.

Recent research collaborations
  • Institute for Genomic Sciences and Policy – Duke University
    A study examining in vivo peripheral blood gene expression signatures in response to respiratory infection to differentiate among viral, bacterial and fungal etiology and further to distinguish among specific pathogens. The aims are to identify pre-symptomatic infection to reduce inappropriate antibiotic use and initiate appropriate treatment earlier. This study uses Bayesian factor analysis and other pattern recognition and machine learning methods in developing expression and symptom signatures to detect infection and differentiate among the infectious agents.

    International Collaboration on Endocarditis
    Currently the largest study on endocarditis (comprised of >7000 inpatient stays in 64 hospitals and 38 countries around the world) this project aims to describe the epidemiology, clinical course and outcomes of patients hospitalized with infectious endocarditis. This study has elucidated geographic, demographic and clinical factors associated with outcomes.

    Hospital epidemiology and infection control in the Durham VA Medical Center
    Hospital-acquired infection (HAI) with multi-drug resistant organisms (MDRO) is a significant contributor to morbidity and mortality worldwide, and reducing incidence of HAI has been a focus of the infection control groups in hospitals nationwide. At Durham VAMC, real-time clinical surveillance data is used to identify patients colonized with MDRO to insure that they are put on appropriate isolation protocols. These data also allow researchers to conduct outbreak investigations within the facility. In addition, patients with incident hospital-acquired infections are identified in order to administer appropriate treatment and to monitor the sources of those infections to reduce future incident events.

    Clinical Case Registry – VA Healthcare System
    The VA Clinical Case Registry has compiled a clinical dataset for all patients infected with either HIV or HCV. This database covers the years 1998 through 2010 and includes 450,000 patients. Data are being used to quantify overall healthcare utilization and temporal trends of utilization in this patient population. The data are also being used to study the epidemiology of cardiovascular disease, pulmonary disease and other chronic conditions in this patient population.

    Institute for Immunology and Infectious Diseases – Murdoch University and Royal Perth Hospital, Perth Western Australia
    A major focus of this group is the association between host HLA genotype and mutation/evolution of HIV and hepatitis viruses HBV and HCV. Certain HLA Class I and II alleles and haplotypes are associated with better or poorer survival among those infected with HIV, and specific HLA alleles are strongly associated with escape mutations in HIV. This research uses full-length viral sequence and high-resolution HLA typing to identify the associations between HLA and viral mutation and viral mutation with clinical disease course.


Alyssa Platt, MA

Alyssa joined Duke’s Global Health Institute and the Department of Biostatistics and Bioinformatics in February 2012. Alyssa received a MA in applied economics from Duke University. Previous to joining DGHI and the Department of Biostatistics and Bioinformatics, she worked in the field of health economics in the Department of Economics, at Duke University. Alyssa has a range of experience in the field of health behaviors and health outcomes, including: obesity and physical activity; infectious disease; risk perceptions; alcohol and tobacco use; effects of built environment on health behavior; court policies and deterrence in domestic violence and DWI behavior; and associations of adherence to recommendations for preventive care in health maintenance and disease management.

Her interests include health behavior and access to care, risk behavior, and economic development, and the use of spatial methods to determine environmental correlates of health and disease risk.

Recent research collaborations
  • MESA
    A cross-sectional matched case/control study linking children presenting to the pediatric ward at Webuye District Hospital to children of the same age/gender and village examining a range of data on household members, sleeping spaces, bednet quality and usage, larval sites, agriculture, building structure, and neighborhood bednet usage to determine the best predictors of febrile malaria.

    Innovative Public Private Partnerships
    A two phase semi-longitudinal case control study in the Ndivisi district of Western Kenya. The study will examine rapid diagnostic tests (RDT) and anti-malarial purchasing behavior in response to coupons for free testing and discounted price on Artemisinin-based Combination Therapy (ACT).

    Facility Incentives Study
    A longitudinal cluster randomized study that examines the effectiveness of pay-for-performance (P4P) incentives in improving malaria diagnostic, treatment, and reporting practices by health centers in high transmission and low malaria transmission areas in Western Kenya.

    DSS Febrile Study
    A longitudinal observational study using census data from the Webuye DSS (Western Kenya). This study identifies spatial “hotspots” of self-reported febrile malaria and finds static and dynamic factors predictive of a household’s location in a “hotspot”

    A longitudinal observational study on physical activity level and weight status of children in two Mexico cities. Children are observed prior to after starting preschool, using survey, accelerometry, and GPS methods.

    NC On the Move
    A pre-post case/control evaluation study examining the effects of a grassroots community intervention to promote physical activity on physical activity and BMI of elementary school age children in two small cities in North Carolina.

    A pre-post case/control evaluation study examining the effects of a statewide policy for minimum mandatory requirements for physical activity for children in preschool.

    Promoting Breastfeeding through State Regulation
    A national cross-sectional spatial analysis of state policies promoting breastfeeding in childcare centers and family home child care.

    Implementation Science to Optimize Malaria Vector Control and Disease
    A longitudinal cluster randomized 2x2 factorial trial of malaria prevention interventions in Tanzania using disease management, vector management, and their combination to reduce prevalence of malaria in 24 randomly selected villages.

    Home-based HIV Counseling & Testing Survey
    A cross-sectional analysis of uptake in antenatal care in 6 administrative districts Western Kenya. Individual, household, village correlates are examined to explain spatial autocorrelation in uptake in care.


John Gallis, ScM

John joined Duke’s Global Health Institute in June 2014 and the Department of Biostatistics and Bioinformatics in August 2012. John received a master’s degree (ScM) in biostatistics from the Johns Hopkins School of Public Health. John’s collaborations at Duke have included performing data management and analysis and creating statistical reports for a long-term randomized controlled trial (RCT) of weight loss in young adults in Durham; working with a neurosurgery research group and other statisticians to answer clinical questions of interest using a large national insurance database; and assisting research relating to childcare use and obesity.

John’s research interests include the application of longitudinal and spatial statistical methods to public health and medical observational and clinical data.

Recent research collaborations
  • Cell Phone Intervention for You (CITY) Clinical Trial
    A longitudinal (24-month) RCT of the effectiveness of a cell phone app-based intervention and a personal contact-based intervention in increasing weight loss in young adults (ages 18-35) versus a control group.

    Neurosurgery Research Group
    Studying the relationship between various procedures related to neurosurgery and various health outcomes, such as complications following surgery and healthcare resource utilization, using a large longitudinal insurance database (MarketScan).

    Danish National Birth Cohort
    A study of the relationship between childhood overweight and obesity and childcare use in Denmark using a large national birth cohort.

    Nurture Study
    Ongoing prospective longitudinal data collection on mothers and infants in Durham, NC, the main goal of which is to study the factors influencing childhood overweight and obesity. Data collection is in its early stages, but a current study on the data collected up to this point seeks to examine the relationship between food insecurity (i.e., not having enough food to eat) and various health outcomes, such as obesity and infant feeding styles.


Ryan Simmons, MB

Ryan collaborates with DGHI faculty and affiliates via the DGHI Biostatistics Core. Collaboration efforts include a priori analyses (e.g. sample size and power calculations) for grant applications, assistance in research design and methods, and analysis of results. Past projects include examining access to healthcare among elderly individuals in Laos and analyzing malaria serology data from a cluster randomized trial in Tanzania. Ryan received his master in biostatistics degree from Duke in 2015.




Recent research collaborations
  • Changes in Access to Healthcare Among the Elderly in Laos
    A study using cross-sectional survey data in Laos evaluating changes in access to and utilization of both inpatient and outpatient services by people 60 years and older.

    Implementation Science to Optimize Malaria Vector Control and Disease
    A longitudinal cluster randomized 2x2 factorial trial of malaria prevention interventions in Tanzania using disease management, vector management, and their combination to reduce prevalence of malaria in 24 randomly selected villages.


Charles Muiruri, MPH

Charles is an affiliate member of the Core. He joined Duke’s Global Health Institute in 2007. Prior to joining DGHI, Charles was a program coordinator in the division of Infectious Diseases at Duke. Charles spent his early years at DGHI working on research capacity building at the Kilimanjaro Christian Medical Centre, Moshi Tanzania. Charles has a master’s degree in Public Health from UNC Chapel Hill and he is doctoral candidate in department of Health Policy and Management at UNC Chapel Hill. Charles supports faculty with research designs and methodology particularly in implementation science approaches.

Charles’s research interests include health services, patient-centered outcomes research and evaluation of health care models for multiple chronic conditions.