University of New Mexico Health Sciences researcher Christophe Lambert is adept at using computer algorithms to detect patterns of behavior hidden in vast troves of digital health care data.
Lambert, division chief of Translational Informatics in the Department of Internal Medicine, leads a team that has been awarded a four-year $2.9 million grant from the National Institute of Mental Health to develop and apply new methods for identifying undiagnosed cases of post-traumatic stress disorder (PTSD) in medical records and comparing treatment outcomes for this condition.
Many people silently live with PTSD, and it is only discovered by the time a person has accumulated multiple co-occurring physical and mental health conditions that drive them to get care, Lambert said.
PTSD can engender sleep problems, cardiovascular disease and depression. You can have hyperarousal to triggers that remind you of that trauma, and you can be in a fight-or-flight mode even when there is no danger
“PTSD can engender sleep problems, cardiovascular disease and depression,” he said. “You can have hyperarousal to triggers that remind you of that trauma, and you can be in a fight-or-flight mode even when there is no danger. Stress hormones are released into your system, which can create physical and emotional problems, including overarousal, alternating with shutdown and depression. PTSD often comes with multiple other mental health conditions, making treatment challenging, and often unsuccessful.”
The study will use machine learning techniques to analyze patient records for more than five million veterans and for more than 40 million people in the general population who are commercially insured or covered by Medicare.
Lambert’s past research has shown that computer analysis can identify patterns of symptoms that reveal underlying behaviors, like self-harm, that may not be coded into the electronic health records (EHR) healthcare providers use to help guide their patient encounters. For example, one study found that when providers treated males who presented with injuries they were less likely to code those injuries as having evidence of self-harm.
One of the study’s goals is to identify disparities in the diagnosis and recording of PTSD and other mental health conditions in different demographic groups, to inform interventions that can improve health equity. In the future, when a certain pattern of symptoms is noted, the software might prompt the clinician to explore whether PTSD might be an appropriate diagnosis, leading to appropriate treatment.
Another goal is to improve patient outcomes by assessing the quality of various treatment regimens. Only two drugs have been approved in the U.S. for PTSD treatment, Lambert said, and they are often used in conjunction with psychosocial methods.
“For half of people, or more, those interventions leave them still with PTSD,” he said. “They're trying all kinds of off-label and combination therapies to try to help folks, and there's not a great body of evidence. Part of our grant is to add to that body of evidence and sort out which treatments are safer and more effective, including with and without psychosocial interventions.”
This analysis should help improve the quality of clinical psychiatric decision-making and guide improved care for those suffering from PTSD and/or those who are at high risk for self-harm, including suicide attempts, Lambert said.
The grant includes collaborators from across the UNM campus and outside the university.
They include Mauricio Tohen, MD, chair of the Department of Psychiatry & Behavioral Sciences, D.J. Perkins, PhD, director of the Center for Global Health, Yiliang Zhu, PhD, chief of the Division of Epidemiology, Biostatistics & Preventive Medicine, David van der Goes, PhD, associate professor in the UNM Department of Economics, Gerardo Villarreal, MD, a UNM professor and psychiatrist at Raymond G. Murphy VA Medical Center, and three partners at Vanderbilt University.
Lambert credited Douglas Ziedonis, MD, MPH, executive vice president for UNM Health Sciences and CEO of the UNM Health System, for his strong institutional support of research partnerships between UNM and VA health system that helped make this research possible.