image: Niyousha Hosseinichimeh.
Credit: Photo by Peter Means for Virginia Tech.
One in eight new moms will experience sadness, hopelessness, and feelings of overwhelm in the first year after having a baby, a statistic that is even higher for those with newborns receiving emergency services.
Beyond the “baby blues,” perinatal mood and anxiety disorders - often referred to under the umbrella term “postpartum depression” - are serious illnesses that can have significant and lasting impact on the patient, infant, and family.
Assistant Professor Niyousha Hosseinichimeh in Virginia Tech's College of Engineering has partnered with physician researchers at Children’s National Hospital to introduce screenings where there aren’t currently resources for adult populations yet ample opportunities to connect with new parents in need: the pediatric emergency department and the neonatal intensive care unit (NICU).
With a four-year, $1.8 million grant from the Agency for Healthcare Research and Quality, Hosseinichimeh and an interdisciplinary team of medical professionals will launch the Perinatal Mood and Anxiety Learning Laboratory at Children’s National Hospital. They hope to improve mental health screening for parents through referral and treatment, combining medical expertise with systems engineering to advance health care.
“In this collaboration with Children’s National Hospital, we are looking at the parents of children who are being seen in the NICU and emergency department. In a pediatric setting, there aren’t services or providers who can treat the parents or adults of the children receiving treatment, but data shows that these caregivers are more likely to need mental health services,” Hosseinichimeh said. “Right now, the screening rate is low, and we want to understand how we can improve the screening and referral process to provide better mental health care for adults in these settings.”
Objectives
- Optimize screening, referral, and treatment for postpartum depression in the NICU and the pediatric emergency cepartment
- Design and develop a novel software dashboard - a system dynamics simulation model to test solutions in a virtual environment - for real-time tracking of the screening, referral, and treatment stages for eligible mothers
- Implement new solutions and assess safety threats related to missed screening, referral, or treatment in current and future health care systems
Principal investigators
- Hosseinichimeh, assistant professor, Grado Department of Industrial and Systems Engineering
- Lamia Soghier, neonatologist and associate division chief for NICU operations, Children’s National Hospital
- Lenore Jarvis, pediatric emergency medicine physician, director of advocacy and health policy for the Division of Emergency Medicine, Children’s National Hospital
Collaborators
- Virginia Tech’s College of Engineering
- Perinatal Mood and Anxiety Disorder Team at Children’s National Hospital
- Center for Translational Research and the Sheikh Zayed Institute for Surgical Innovations
- Child Health Advocacy Institute
- MedStar Patient Safety Institute
Engineering medical solutions for postpartum depression
Hosseinichimeh and Soghier describe their respective roles in this project, how engineers can partner with medical researchers to find innovative solutions, and why interdisciplinary collaboration is crucial to the future of health care.
How did this collaboration come about and what do each of you bring to this project?
Soghier: I give the medical direction and the expertise in relation to what families experience that can lead to mood and anxiety disorders when their child is born and then hospitalized in the NICU. I have developed a team of screeners, social workers, and psychologists who work with parents in the NICU alongside Lenore Jarvis. who is a pediatric emergency department physician. Like myself, Lenore discovered that many of the families we serve are stressed, depressed, and have mental health concerns after the birth of their child.
We individually bring our knowledge of the system in which we work and our expertise on the ground in the hospital. Niyousha brings her experience in health care systems engineering. From there, we work with a large team of software developers and clinicians to develop solutions that will make our hospital better at recognizing parents with perinatal mood and anxiety disorders, easily identify parents in distress, and give families the help they need.
Why does the medical field need engineers in this area and how is your discipline uniquely equipped to solve that problem?
Hosseinichimeh: My job as a systems engineer is to analyze the current workflow, identify barriers, elicit system requirements from stakeholders, map the system, and develop a system dynamics simulation model to test solutions in a virtual environment. I'm not a physician and I'm not a psychologist, but I facilitate the extraction of knowledge and then integrate that knowledge into a simulation model.
How does a simulation model come about?
Hosseinichimeh: I use group model building techniques to elicit information and knowledge from experts and then build simulation models. After building confidence in the simulation model, we test interventions in a virtual environment before implementing them in a clinical setting. When we test interventions in a virtual environment, we can identify potential barriers, problems, and unintended consequences. Once we implement the selected solutions, we measure how they’re working and use that data to adjust the model. Our goal is to use the predictive model to prioritize screening caregivers at higher risk.
What kinds of questions will you be exploring and what will the data look like?
Hosseinichimeh: A doctor diagnoses individual patients and enhances their health, while a systems engineer diagnoses the entire health care system, analyzing its individual components and their interactions to develop effective solutions for improving patient outcomes, enhancing operational efficiency, and ensuring the delivery of high quality care. We ask questions like:
- Who are the main stakeholders of this problem?
- Who has the interest, the resources, and the power to affect change?
- What are key variables that affect the outcome and processes of this system?
- How do these variables affect each other?
The answers to those questions become our qualitative data. We then take the de-identified quantitative data from the health care providers - like the number of people who were approached for a mental health screening, the number of people who were actually screened, which percentage had positive screens, and so on - to calibrate the model.
After building confidence in the model through various tests, the model is transferred to a website so health care providers can simulate the model without needing our software. We will run what-if analyses, which help the health care providers and managers to make decisions in order to improve their system.
What are some of the “what-if’s” that could arise in screening for postpartum depression?
Hosseinichimeh: In the case of parents and adult caregivers, we might need to know if, for example, we switch to remote mental health screening - instead of health professionals providing the screening - how many resources will we need? If a caregiver is suicidal, they need to be connected to care immediately. Will the number of patients who qualify for care increase and how will we meet those needs with the psychologists, psychiatrists, social workers, and others who provide mental health support? What will the cost be?
Ten years down the line I would also like to see how new techniques, new methods, and new technologies, like ChatGPT or other large language models, can be applied to our approaches and improve them so that patients receive a higher quality of care.
What does the public need to know about postpartum depression?
Soghier: After multi-month admission to our NICU, 45 percent of parents screen positive for depression. I can’t think of any other disorder or disease that screens positive at 45 percent. Similarly, mothers of infants in our emergency department screened positive at 27 percent. This can’t be ignored.