Digital Neuralhealth - Research Program
Project: Readmission Risk Assessment of Mental Health Patients
In collaboration with the Psychiatry Department and Systems Quality Group of Houston Methodist, this research project aims to apply machine learning approach to identify psych patients who are at high-risk and lists personalized post-acute care intervention that is best suited for individual patients. Evidence suggests that mental health conditions and symptoms can increase physical health readmission rates. Depression is an independent risk factor for many comorbid conditions. Little is known about predicting psychiatric readmissions and identifying modifiable factors that may reduce early readmissions. The study would also help hospitals, insurers, and homecare agencies improve the quality of care for patients while reducing 30-day readmission rates. Readmission risk assessment tool will determine if interventions designed to treat mental health symptoms can effectively reduce readmission risk. While our focus is on the psychiatric conditions, we will also focus on adult patients hospitalized for any physical health condition to identify interventions that could be adapted for study in our populations of interest. Readmission Risk Assessment tool will be administered any time before discharge and stratify the patients into two groups of Low Risk and High Risk of readmission.
Project: Hospital Outcomes and Readmission Risk Assessment for Hospitalization in Geriatric Neurodegenerative Patients
More than 5 million Americans are living with Alzheimer’s disease, and at least 1 million Americans live with Parkinson's disease. The risk of being affected by a neurodegenerative disease increases dramatically with age. Geriatric patients will use up more healthcare costs and geriatric patients with neurodegenerative conditions stay in the hospitals more often than other people of the same age group because of their neurological deficits and functional decline. However, hospitalization may be unavoidable due to many reasons, such as trauma, infection, stroke, surgery, or exacerbation of systemic illnesses. Therefore, it is important to understand all the factors that are significant to lessen the hospital ordeal and to achieve a good outcome post hospitalization. In collaboration with the Department of Neurology and the Neuro ICU, the objective of this project is to develop machine learning models that can predict: (1) by 1st or 2nd day of admission, the outcomes of hospitalization of geriatric patients with neurodegenerative diseases, notably dementia, Alzheimer’s, and Parkinson’s, and (2) the risk of readmission of these patients in order to devise effective follow-up interventions for modifiable factors that could improve patients’ outcomes and reduce readmission risks. We are applying machine learning to analyze ten-year’s worth of dementia patient data at Houston Methodist and narrowing down risk factors (both modifiable and non-modifiable) vital in determining patient outcomes. Accurate prediction of readmissions in this patient segment could also help mitigate rising health costs.
Project: Personalized Digital Music Therapeutics for Mental Health Patients
In collaboration with the Houston Methodist Center for Performing Arts & Medicine, this project aims to study the effects of music therapy (MT) and other musical interventions on different aspects of neurological disorders, including mood, anxiety, depressive syndromes, and quality of life on neurological patients. Psychological distress, depressive syndromes, and mood disorders represent typical comorbid conditions in neurological disorders that can lead to acute and chronic stress responses. Cognitive neuroscience research over the past decade has created increasingly fascinating insights into how the human brain processes music, while also enhancing the understanding of brain processes related to many mental health disorders. Music therapy is considered a complementary treatment that can be utilized in conjunction with psychiatry and psychology to treat and manage mental health needs. People with moderate symptoms of anxiety and depression may benefit from self-directed protocols for mood symptom management, particularly when paired with therapist-directed intervention and guidance. This project implements individual music therapy for mood symptom management through a blended model incorporating music therapist-directed and self-directed treatment components. MOCHA digital therapeutics, which has previously been tested successfully for cancer patients and survivors, encourages independent accountability and self-reinforcement in maintaining wellness plus a means of ongoing communication with healthcare providers, which may help to mitigate neurological disorders. A workflow of the patient-side app of music therapy delivery via MOCHA digital therapeutics is shown above.