Healthcare, Big Data, and Emory

From our phones to wearable devices, modern technology allows us to both collect and analyze data like never before. In healthcare, reams of data are collected from a number of sources; often there is too much information for patterns and trends to stand out to the human eye. Up until recent advances in data analysis the use and benefits of such “big data” was difficult and the potential benefits to patients and research remained obscured. Although the relationship between healthcare and big data is still in its early phases, researchers at Emory and all over the globe are using previously developed knowledge of health science and pairing it with computer science to extend the abilities of those working in modern healthcare.

Big data describes data sets that are so vast that computational analysis is required to derive meaning from them. Because public health is built around finding trends and links between an abundance of factors, this technology is a useful tool. It gathers information and uncovers connections with extreme efficiency, although the computers are not working alone, but in concert with humans. The National Institutes of Health (NIH) All of Us research program is teaming up with research institutions like Emory all over the country to gather data on nearby populations.

Michael Zwick, PhD, assistant dean of research at Emory School of Medicine and assistant vice president of research at Woodruff Health Sciences Center, is principally overseeing Emory’s role in the All of Us Research Program with the help of Arshed Quyyumi, MD, Greg Martin, MD, James Lah, MD, PhD, Andrew Post, MD, PhD, and Alvaro Alonso, MD, PhD. They are making an effort to include as much of the population as possible to increase the accuracy of the research, ensuring underserved communities that are otherwise rarely considered are included. Big data, with the help of some of the top minds at Emory and other institutions, will help make sense of the social and health experiences of millions of people in an effort to improve lives.

While big data can be applied to populations, collecting and efficiently interpreting large amounts of data can also be helpful when applied to the individual. When it comes to our health, our bodies have a multitude of measurable responses to ailments and illnesses. Some are telltale signs and some are subtle, interconnected, and difficult to identify. Emory researchers like Timothy Buchman, MD, PhD director of the Center for Critical Care (ECCC) and Shamim Nemati, PhD assistant professor the Department of Biomedical Engineering are using technology to help monitor patients’ statuses while in critical condition, taking some of the weight off the shoulders of staff and leading to stronger accuracy of diagnoses and predictions. Buchman founded the ECCC and recently initiated Emory’s electronic ICU (eICU), a patient data monitoring system that has a specific taskforce in charge of overseeing the process and work in conjunction with the other ICU staff. He is currently working on a streaming analytics system that looks for patterns in real-time patient data and alerts medical personnel of crises like heart failure, sepsis, or pneumonia.

Nemati’s work is in much the same field, but with a focus on sepsis. Sepsis occurs when the body overcompensates for infection, causing rampant inflammation. It is a severe, life-threatening complication that causes more deaths in the United States than breast cancer, prostate cancer, and AIDS combined. It is vital to catch it as early as possible to lessen the potential for death. Nemati and his colleagues are working on monitoring algorithms that can predict the onset of sepsis very early using measurements from more than 65 different indicators. By interpreting these measures in concert with one another, computers can predict sepsis up to 12 hours ahead of time, better than any previously developed sepsis prediction technology. The computers are “trained” on years of Emory ICU admission data to be as accurate as possible.

The application of big data concepts extends beyond physical ailments. Ying Guo, PhD, director of the Center for Biomedical Imaging Statistics at Emory, is looking to create analytical tools that search for and find biosignatures in brain scans that identify abnormalities such as mental illnesses, addiction, and cognitive issues. The data being collected from brain scanning technologies (fMRIs, sMRIs, DW-MRIs) is getting continually more sophisticated with vast amounts of data contained within a single scan, far too much for an individual to fully interpret. Guo and her collaborators at the Emory School of Medicine are developing algorithms that filter through the immense amount of data from scans in search of signs and trends that point to cognitive dysfunctions. Guo’s work aims to combine information from various types of scans to achieve more holistic understanding of the cognitive functioning of patients. This multi-modality imaging is already proving helpful to other researchers. Tanja Jovanovic, PhD, of Grady Hospital has used Guo’s technology to validate her previous findings concerning the brains of PTSD patients. Guo is also extending her work to characterize data sets collected to study subtypes of depression and therefore can be used to develop and train Guo’s algorithms.

The interconnection between technology and healthcare is only just beginning. The work being done by programs like the All of Us research and other big data projects by Emory researchers are exciting fusions between computer science and health science. These will open pathways for the researchers, scientists, and healthcare providers of the future, and will continue to lead to safer and more capable healthcare practices.

Works Cited