by Peter Song
- 1The growing availability of big and complex data sets means new opportunities for interdisciplinary research in public health.
- 2Biologists once studied individual genes through conventional laboratory experiments . Today, using high-throughput microarray, sequencing, and other technologies, they collaborate with statisticians to measure and analyze gene expressions across the entire genome.
- 3The vast amounts of data generated by new technologies allow scientists to better understand the relationships among genetic variants as well as gene-environment interactions—crucial factors in understanding complex diseases.
- 4Because of the increased capacity of computer storage and cloud-computing facilities, scientists have access to—and are able to process—more variables than ever before.
- 5Previously, scientists looked at data that was often collected at a single point in time. Today they follow people's life-course trajectories and measure data at multiple points, making it easier to track the progression of chronic diseases and the mechanisms of human growth and development.
- 6Instead of examining just one type of data, such as physiological outcomes, researchers now assess and integrate multiple kinds of data, such as imaging and behavioral data and environmental exposures. This kind of rich, subject-level information can help scientists develop personalized interventions to improve quality of life.
- 7Digital health records yield vast amounts of information on conditions like asthma and diabetes and allow researchers to do a better job of tracking both compliance with and the efficacy of medications.
- 8Instead of conducting costly and time-consuming conventional clinical trials with specific groups of people under certain exclusion criteria, scientists can use digital health records to streamline clinical trials—and save both money and time.
- 9By combining data from national disease registries, insurance companies, hospitals, and individual patients, scientists are better able to evaluate the quality of health care—and people in need of treatment can make better choices about facilities and treatments.
- 10Conditions like childhood obesity have very complex pre- and postnatal exposure patterns. Big data enables researchers to measure a wide span of environmental exposures, from pesticides to food intake to neighborhoods. This high-dimensional data can yield critical insights into childhood growth and development and help lead to effective interventions to prevent obesity.
Peter Song, professor of biostatistics, is engaged in several projects involving big data, including studies of renal disease using registry data from the U.S. Renal Data System and from the NIH and EPA–co-funded study Lifecourse Exposures & Diet: Epigenetics, Maturation & Metabolic Syndrome. He hopes this research will provide relevant knowledge and better evidence for policymakers.
Back in the Day ...
.... researchers like Thomas Francis Jr. stored big data in file cabinets—more than a million records of children who participated in the Salk polio vaccine trials. No one had ever conducted a study of that size, and it's not likely to be repeated. As epidemiologist William Foege, the first recipient of U-M's Thomas Francis Jr. Medal in Global Health, noted on the 50th anniversary of the conclusion of the Salk vaccine trials, a century from now people will still marvel "at the audacity of a field trial that kept track of 1.8 million children before the age of computers."
Public health students don't always get the chance to work with real-life data in the classroom, so field experience is critical, says Dana Thomas of the U-M SPH Office of Public Health Practice.
Field work teaches students how data is actually collected and lets them see the faces behind the data. It helps them understand why questions need to be worded in certain ways and why questions need to be asked the same way with each respondent. "Sometimes data comes from focus groups," Thomas notes. "Sometimes it comes from going to the market and having conversations." Wherever it comes from, the context matters because it influences how students interpret the data they gather.
Since 2006, when Thomas's office sent a group of SPH students to the Gulf Coast to help collect data in the wake of Hurricane Katrina, data collection has been a key component of the practice experience at SPH. In partnership with health departments and community organizations, students on the school's interdisciplinary Public Health Action Support Team, or PHAST, have worked on projects both in the U.S. and overseas. It's a win-win arrangement, Thomas says, because students get indispensable hands-on experience, and health departments and community organizations—who often lack the capacity to undertake significant data collection—get the data they need to launch new initiatives, determine programmatic direction, and identify areas of need. Specific PHAST projects include:
- Mississippi Gulf Coast and Delta: Since 2006, SPH students have helped assess post-Katrina needs and quality of life in communities in Mississippi and Louisiana, with the aim of informing community and economic development.
- Texas: Since 2010, PHAST teams have worked in the border region near Brownsville to gather information on food security, electronic medical records, and implementation of the Affordable Care Act.
- Grenada: PHAST students have conducted focus groups on workplace wellness and worked with the Sickle Cell Association of Grenada to assess caregiver needs.
- Kentucky: In collaboration with the CDC's Community Assessment for Public Health Emergency Response, or CASPER, program, a PHAST team conducted a door-to-door community assessment in the wake of tornado strikes in 2012.