The New Challenges

This new technology and ability to access data presents as many challenges as opportunities. One new challenge is that scientists now can generate an extraordinary amount of data about just one person.

“There’s the genome, the epigenome, the microbiome, the exposome and many more to come,” Samet said. “Think about the billions of data appoints that we are generating about single individuals with the idea that we’re going to answer questions about their health risk and disease. Somewhere, we’re going to have to put this all together. Personalized medicine has to reduce all that data into something that a person and their health-care providers can use.”

Questions that scientists are trying to tackle using this information have become more challenging, as well. While, during the 1950s, epidemiologists were looking to determine whether tobacco had an impact on mortality, the questions now must deal with thousands of variables, not only one or a few. Today’s epidemiologists are tracking how genes interact with different environments, looking at life-course epidemiology and trying to generate the healthiest humans possible, and studying environmental carcinogenesis in an attempt to determine what in the environment can cause cancer.

“These are all incredibly challenging,” Samet said. “For example, the sample size for a study on gene/environment interaction needs to be in the tens of thousands or hundreds of thousands. That’s not even acknowledging some of the flaws in the data. These are very big end questions.”

The Future of Epidemiology

“If this isn’t a paradigm shift, I’m not sure what is,” Samet declared. But where that shift will lead is still up in the air.

The future of epidemiology may lie in doing more studies through health systems or using electronic health records. It will likely require an increase in data sharing and access, as well as the creation of networks of scientists.

The latter issue poses a problem in academic circles. “I recently met with all of my faculty, and it was quite interesting talking with all the people that work in the genomics area. There are shared triplets of first authorship on papers and quadruplets of senior authors,” Samet said. “This poses career issues. How do you gain credit for contributions? In academia, we talk about team science, but that hasn’t drilled down yet to promotions and acknowledgements.”

As for data sharing and tracking patients, that will become easier as more health systems adopt EHRs. For the last decade, Samet has worked with a colleague in Korea who has access to a population of 1.3 million Korean civil servants and teachers who answered lifestyle questionnaires as long ago as 20 years.

“We’re able to publish papers simply tracking outcomes and exposures just through record linkages. We can define incident diabetes based on blood sugar or prescription medications,” he said.

The creation of more biobanks may lead to further opportunities for epidemiologists. VA, for example, recently undertook an effort to collect health data and genetic samples from one million veterans nationwide, and has promised to make the data available to all federal researchers.

Asked if making all of this data available will lead to stakeholder analysis and reevaluating studies for the purpose of discrediting them or inducing confusion, Samet said it’s a danger, but not one that has manifested yet.

“For every analysis, there could be an equal and opposite analysis done by the right data person. I’m not sure I’ve seen that happen yet, but I think it’s a real concern. I think what we’re going to need to think through is that, with all this data available, what is the role of the epidemiologist in using it and collecting it? Who are the epidemiologists of the future?”

Back to July Articles