by Elizabeth Marshall, Director of Clinical Analytics at Linguamatics, an IQVIA company
April 2020 – The coronavirus pandemic has made its mark on every aspect of our lives. Not only has COVID-19 threatened our physical health, it has also created huge economic challenges, altered our social routines, and forced us to stay at home.
We don’t know how long it will take us to recover. As we assess the widespread impact and seek to find solutions to move forward, we cannot overlook the emotional and psychological toll that this health crisis has taken on all of us. As of April 2020, a Kaiser Family Foundation poll found that 45 percent of all Americans felt the coronavirus was negatively affecting their mental health – a figure that is hardly surprising as we practice social distancing, and even isolation for some – and experience fear and anxiety about our health and financial well-being. The situation is made worse as the endpoint continues to shift from weeks to months to the unknown.
The country’s healthcare system is not well-equipped to proactively identify patients who may be at risk of mental health issues. Often individuals are not diagnosed with major depression, suicidality or other mental health conditions until a significant event occurs, such as a failed suicide attempt that sends a patient to the hospital. Most EHRs were designed to support billing processes and lack structured fields to store vital mental health clues, like feelings of hopelessness, involvement in risky behavior, and increases in substance abuse, especially if there is a lack of social support. Increased isolation is now necessary, which increases the risk for mental health related issues like suicidal behavior. When a concerned clinician does include such observations in a patient’s chart, the details are usually captured within free text fields.
As we move beyond this health crisis, clinicians need to be mindful of the red flags that indicate the potential risk of a mental health disorder- a task that may be difficult if the clues are buried deep within the record. Now is the time for healthcare organizations to consider programs and tools to help identify the warning signs, so we can help patients and proactively avoid a national mental health crisis.
Finding hidden mental health clues
Earlier in my career I worked for the Veteran’s Administration as a research physician, and as a veteran myself, I was dedicated to advancing mental health treatment for my fellow military veterans. At the time, VA clinicians were not well-equipped to identify early signs of possible depression or mental illness. Even today, looking for clues within an EHR’s structured fields, such as the chief complaint or diagnosis, rarely yields deep insights into an individual’s current status. Chief complaints within most EHRs are often outdated or inaccurate and don’t reflect whether a patient’s condition has improved or deteriorated, if they are taking medication as prescribed, etc.
During a typical encounter, a patient might mention having difficulty getting out of bed in the morning, drinking excessively or not wanting to socialize with family and friends, which are just a few behaviors that indicate possible depression. Most EHRs don’t have structured fields to store these observations, and even when they do, a clinician often notes such information in a free-text field. As I worked to identify VA patients who were at risk and might need additional follow-up, I quickly realized that without a solid informatic solution, there was no easy way to find these individuals without time-intensive manual reviews of the unstructured clinical notes for every patient. Without better search tools, at-risk patients would likely fall through the cracks.
Shortly after my time at the VA, I pursued my fellowship training in informatics, hoping to prove that better search tools could improve clinical care for individuals at risk of mental health crises. Between my experience at the VA, where I spent several months reviewing EHRs for thousands of patients from hundreds of clinicians, and my fellowship training, I became convinced that NLP technology could help clinicians search patients’ records more efficiently at the primary point of care, enabling them to identify many more at-risk patients and take appropriate action before a major mental health event occurs.
Avoiding a national mental health crisis
The current pandemic is creating a perfect storm for a national mental health crisis. For many people, especially those who live alone, social distancing can easily lead to substance abuse, feelings of social isolation and loneliness. Watching the escalating number of infected patients makes us anxious for our own safety and the health of our loved ones. And, as jobs disappear, more families are facing severe economic hardships, fueling high levels of anxiety and fears for the future.
Healthcare providers must be on high alert for the red flags that indicate a patient could be at risk for a major mental health event – but this is challenging without the right tools. As I realized at the VA, finding these clues can be much easier with NLP and related informatic tools and workflows.
As we move forward and heal from this health crisis, healthcare leaders worldwide must make sure their organizations have the right infrastructure in place to proactively identify and address not only the physical manifestations of disease, but also the psychological needs of their patient populations.
About the Author
Elizabeth (Liz) Marshall, MD, MBA is the Director of Clinical Analytics at Linguamatics, an IQVIA company, where she is responsible for clinical oversight of all healthcare projects. Prior to joining Linguamatics, Marshall completed her fellowship training in informatics at the Medical University of South Carolina and served as an assistant professor and clinical manager for a study focused on decreasing veteran suicidality. Early in her career she served in the United States Air Force as a logistics team member in computer operations for Operation Enduring Freedom. She later became a research physician dedicated to the development of informatics solutions to improve the effectiveness of mental health treatments for military veterans. In honor of her work advancing the treatment of suicidality and PTSD for veterans while a clinical research health scientist at the Ralph H. Johnson Veterans Administration Medical Center, she was awarded the Research Training Institute Scholar Award from the ICRC-S Injury Control Research Center for Suicide Prevention (2013).