Navid Ghaffarzadegan

Navid Ghaffarzadegan: Modeling Complex Problems

for Real World Impact

As far as complex problems go, COVID-19 is currently among those at the top of the list. With the coronavirus pandemic continuing through a fourth surge in the U.S., Americans are weary of changing public health guidance, wearing masks, tensions over vaccinations and continued uncertainty about the future. This messy, unpredictable global issue is just the kind that Navid Ghaffarzadegan, an associate professor in the Grado Department of Industrial Systems and Engineering and an Institute for Society, Culture and Environment Scholar, likes to tackle. He is passionate about understanding complex social, managerial and policy problems and how his research can inform important policy decisions.  

“I am interested in studying complex societal issues, problems that matter and that are challenging to deal with, such as healthcare problems, higher education challenges, or barriers for scientific advancement,” said Ghaffarzadegan who also directs a research team of undergraduate and graduate students at the Social Dynamics and Analytics (SoDA) lab on the Virginia Tech campus in the greater Washington, D.C. metropolitan area.

To ensure he has a broad and deep understanding of complex issues, he uses systems thinking and a variety of simulation models and data analytic techniques, such as system dynamics models, statistics, machine learning and network analysis, as the basis for his research.

Using multiple methods, including qualitative methods, “enriches the models and helps me better understand the results of the model and the issue at hand,” explained Ghaffarzadegan. These models can then help him identify potential outcomes and gain valuable insights that may have real world policy implications.

“Complex systems include many interacting variables, webs of interconnections and feedback loops,” said Ghaffarzadegan. “If you take a reductionist approach, and focus on a single cause, or assume a linear chain of events, you may miss things which may then have unintended, often negative, consequences.”

It is these same techniques that Ghaffarzadegan used for the past eighteen months to help Virginia Tech leaders better understand the potential impact of rising COVID-19 cases on the university’s operations. He developed a simulation model that showed how the virus could spread on campus depending on certain scenarios and mitigation strategies.

“Working closely with university administrators helped me gain a better understanding of the university’s operations and its complex problems,” Ghaffarzadegan said.

He explained that many insights about a system are in the stakeholders' mental models. Without intensive communication and discussion, “modelers may fail to deeply understand the problems they are dealing with. It is also important for modelers and stakeholders to build trust which will be essential in a successful implementation of the modeling outcomes.”

Ghaffarzadegan points out that “such two-way communication should continue throughout any project. This is in contrast with the dominant mode of modeling, and research in general, where scientists are disjoined from stakeholders, which affects the quality of their work and the potential to make an impact.”

Ghaffarzadegan has made an interactive COVID-19 model available for other universities to use on the SoDA website. A calibrated version of the model, which uses Virginia Tech data, can be used for projections and policy analysis. His COVID-19 research has been published in journals such as BioScience, Lancet Planetary Health, System Dynamics Review and PLOS One and been featured in popular news media such as the New York Times, The Washington Post and the BBC.

Ghaffarzadegan’s previous research has focused on complex health issues such as post-traumatic stress disorder (PTSD) and HIV/AIDS. For example, he has developed simulation models that estimate PTSD prevalence among veterans and military personnel.

“We used the models to look at different policy alternatives such as more screening, or better recruitment procedures on PTSD prevalence, as well as provided an analysis of healthcare costs for those in the military and the VA for each of the policies,” said Ghaffarzadegan. He explained that the models can be used by policy makers as they develop programs to prevent or reduce PTSD.

Ghaffarzadegan is also interested in better understanding scientific advancement and workforce development. In one project, for which ISCE provided some support, Ghaffarzadegan examined how research topics predict research success.

“Is it better to do research on a popular topic or a novel one?” Ghaffarzadegan asked. He reviewed over 200,000 HIV-AIDS-related publications over three decades. “We constructed a dynamic topic network from the abstracts of these publications through a topic-modeling approach. The network depicts the evolution of HIV-AIDS scientific knowledge over time.”

The results of his study indicated that many high impact studies include novel topics or combine topics of varying maturity, such as a well-studied approach applied to a new topic. The policy implication of these findings is that faculty advisors and mentors can help guide graduate students to choose research topics that are of interest to them, but that may also increase their chances of publication success, which may ultimately improve their long-term career prospects.  

Ghaffarzadegan’s work on workforce development has also focused on the number of PhD students that universities produce and their long-term career success. Ghaffarzadegan has used a systems approach to examine the impact of increased research funding by federal funding agencies on job opportunities.

For example, from 1998-2003, the National Institutes of Health doubled the funding available for research, which increased PhD graduation rates, and in turn, created a need for new positions; when the funding dried up, however, there was not enough funding available to support all the new hires. Thus, this increase in funding actually had a negative long-term impact on the careers of many of the researchers coming into the market. His research suggests that it would be optimal for federal agencies to incrementally increase research funding rather than make large increases all at once.

In addition to his research, Ghaffarzadegan regularly teaches a course on System Dynamics which provides students with a foundation in how to build complex models, such as the one he developed for COVID-19. He was also one of three instructors for a new policy course for the Policy Destination Area that provided non-policy majors with a grounding in policy skills for complex problems and decision-making.

To learn more about Ghaffarzadegan’s research, visit the Social Dynamics and Analytics Lab website.