Supriya Kumar's home page

Supriya Kumar, PhD, MPH

Post-doctoral Associate
Graduate School of Public Health, University of Pittsburgh
Contact information

Curriculum Vitae
Google Scholar Citations
Teaching

Research

I transitioned from a productive career in laboratory-based Biological Sciences research into behavioral health research looking for a more socially relevant career in scientific research. I now study the impact of social factors, in addition to cognitive factors, on health behaviors. I am specifically focused on how the policy and social environment, and individual social position impact health behaviors. In addition, I am interested in the translation of research and evaluation to inform the development of policy. I am thoroughly enjoying my new career in public health, and have multiple published manuscripts. My goal as a public health researcher is to effect the implementation of health-enabling policies by researching how the social and policy environments interact to impact health behaviors.

Research Theme: Health Disparities
The overarching theme in my research is health disparities. I studied racial disparities in vaccine uptake and the voluntary decision to stay home from work if ill during the 2009 H1N1 pandemic. I use the social ecological model and models of the social determinants of health extensively to inform my studies. I have reviewed the national pandemic plan in India from a health disparities perspective. I am currently using computational modeling to assess the impact of a universal paid sick days policy on socioeconomic disparities in influenza incidence among working adults.

Research Methods
I use survey analysis, Geographic Information Systems (GIS), computational modeling, as well as qualitative (focus group) methods and employ a participatory research approach whenever appropriate.

Survey Research
I have used a nationally representative survey of the US population to examine public attitudes and abilities to carry out recommended behaviors during the 2009 H1N1 pandemic. During this research, I led a study of social, policy, and individual-level determinants of acceptance of the 2009 H1N1 vaccine. I used behavioral theory extensively to inform this research, resulting in a manuscript in Health Education and Behavior. I also studied the impact of access to workplace resources on self-reported influenza incidence during the 2009 H1N1 pandemic, leading to a manuscript in the American Journal of Public Health. Using structural equation modeling, my collaborators and I showed that Hispanics, who reported work-related barriers to social distancing, including lack of access to sick leave, had higher odds of self-reported ILI incidence compared to Whites.

This work on the H1N1 pandemic afforded me the opportunity to collaborate with a multi-disciplinary team of researchers at the University of Pittsburgh, University of Georgia, and the University of Tennessee. I enjoyed working with a diverse group, and was proud to find that I could collaborate effectively and efficiently on multiple studies simultaneously. The group, including statisticians and behavioral researchers, also provided a wonderful opportunity to learn new ideas and ways to analyze and interpret data.

I am currently using large national and regional surveys to analyze the determinants of health behaviors. I am using the 2009 Behavioral Risk Factor Surveillance System survey in Allegheny County, PA, to examine the geographic and demographic correlates of influenza vaccine uptake, and the 2010 National Health Interview Survey and the 2009 Medical Expenditure Panel Survey to examine the relationship between employees' decision to stay home from work if ill and their access to paid sick days in the US.

GIS in public health research
I use GIS extensively to conduct descriptive as well as quantitative analyses of access to health services and other resources. I have collaborated with geographers at Penn State University to examine multi-level correlates of childhood vaccine uptake rates in Niger. We modeled the travel network between villages and health centers in Niger and used Demographic and Health Surveys with characteristics of vaccine acceptors to show that living within a 1 hour walk to a health center resulted in a higher probability of vaccine uptake than living >1 hour from a health center. This study was recently published in the International Journal of Health Geographics. I am currently examining influenza vaccine uptake behavior of Allegheny County residents in conjunction with travel times to pharmacies that provide vaccines.

Computational Models
To explore how determinants of behavior interact over space and time, and to evaluate the impact of potential policies, I have recently collaborated with computational modelers. Agent-based models provide a powerful method to understand the causes of health behaviors and health disparities. In fact, the Institute of Medicine has recommended that modeling be used to assess the benefits and costs of policy options. I am joint-Principal Investigator on a newly submitted R21 proposal to the NIH aimed at studying the health and economic impacts of universal access to paid sick days among employees and their children. In this line of research I have had the opportunity to collaborate with a trans-disciplinary team of researchers including computer scientists, epidemiologists, infectious disease experts, health policy experts, and health behavior researchers.

Community-based Participatory Research
As stated above, I am joint-Principal Investigator on a newly submitted R21 proposal to the NIH in which we are taking a stakeholder-engaged, participatory approach to modeling the impact of universal access to paid sick days on influenza incidence among employees and their children. Our participatory approach will directly tackle the perception of models as "black boxes" by involving end-users of models (the public, policy-makers, and employers, recognizing and integrating the unique expertise that each brings) as partners in the process of modeling itself. Our aim is to increase transparency of our model to end-users, to gather feedback on face-validity of our model and results (in line with recent recommended good-practices for model validation), and to ultimately promote the utility of model findings in decision-making.

In the past, I initiated and led a community-based participatory research study examining the impact of neighborhood grocery store quality on dietary and shopping behaviors in Pittsburgh. This research, employing mixed methods including GIS, focus groups, and a survey, informed us that though a number of high-quality supermarkets exist in or bordering neighborhoods in which 60% or greater of the population is African American, the community only shops at two supermarkets. These "Black-identified" stores were perceived to have lower quality produce and food. The community continued to shop in these two stores for reasons including its dependence on Jitneys, which were easier to access at these stores. My focus groups led to community-generated ideas for interventions to empower the African American community to monitor and manage the food resources in their neighborhood.