Improving Access through School Based Care
The Division of Child and Adolescent Psychiatry’s school-based services growth in their partnerships with local school districts is surging in recent years. School-based services began seven years ago in the Oak Hills Local School District on the West side of Cincinnati with a handful of clinicians and over the years experienced tremendous growth. During FY19 alone, we expanded the number of social workers and clinical counselors providing clinical mental health counseling from 16 to 28 across 37 schools in the Cincinnati area. Embedding therapists within the schools provides care on site for ease of access to care. This greatly improves access for patients who may previously have run into barriers that kept them from either receiving treatment at one of our outpatient locations. In FY18 our school-based therapy program saw 586 patients for 6,769 visits. In FY19, therapists saw 1,037 patients for a total of 15,256 throughout the school year and summer.
Psychiatry also provides medication management services at our schools provided by Advance Practice Registered Nurses (APRNs). We increased our reach in schools for these services in FY19, as well. In FY18, our APRNs saw 64 patients for a total of 575 visits in the schools. In FY19, our APRNs increased the number of schools that received services and saw 131 patients for 850 visits. Patients that receive these services must also receive school-based therapy with our embedded therapists.
The partnership between the schools and Cincinnati Children's is strong, and the stakeholders within the school communities find these services to be a valuable asset in their ability to support students coping with a variety of challenges. Patients seen by school-based clinicians are often in treatment for longer periods of time, which allows the opportunity for continuity of care and maintenance of growth. Many of our clinicians also provide ancillary knowledge and expertise for school faculty and staff. Our clinicians receive training in a variety of backgrounds and modalities and commonly provide professional development, which increases school staff knowledge about how to best support students who may be coping with mental health diagnoses, trauma backgrounds, or other challenging circumstances. We are continuing to increase our reach for school-based services in FY20 to provide care to children in need.
Evaluating School Violence Risk through Natural Language Processing
Our research team previously established a standardized risk assessment program to evaluate students for risk of school violence and began the development of a natural language processing (NLP) and machine learning technologies to automate this process. Our research aimed to further develop and evaluate these technologies. An interview process of the 219 participants used our two innovative risk assessment scales to determine their risk of violence based on clinical judgment by the Cincinnati Children's forensic psychiatry team. Extraction of different types of linguistic features from the interview content occurred by leveraging NLP technologies. Applying machine learning classifiers helped to predict risk of school violence for individual subjects. Implementation of a two-stage feature selection identified violence-related predictors. Assessment of positive predictive value (PPV), sensitivity (SEN), negative predictive value (NPV), specificity (SPEC), and area under the ROC curve (AUC) validate the performance on the psychiatrist-generated reference standard of risk levels . Participants were between the ages of 10 to 18 years old, enrolled in school (excluding homeschool and online school), and were not in state custody. We recruited an equal number of females and males with no exclusion of race, ethnicity or socioeconomic standings. Our study population represented a large spectrum of severities of behavioral concerns or behavioral changes. We included subjects without behavioral changes to stimulate a real-life school population. Compared to subjects’ demographics and socioeconomic information, use of linguistic features significantly improved classifiers’ predictive performance (P<0.01). The best-performing classifier in 2019 with forest algorithm features achieved 66.67%/95.24%/92%/53.49%/91.75% (PPV/SEN/NPV/SPEC/AUC) on the test data. Natural language processing and machine learning classifiers show strong promise in detecting risk of perpetrating school violence. By continuing to expand the number of subjects and continuing to develop these computerized algorithms, we will eventually create an end-to-end automated risk assessment program.
Improving Care for Patients with Neurodevelopmental Needs –Expanding the Care Continuum
Youth with neurodevelopmental disorders, intellectual disability, and autism spectrum disorder (ASD) often have complex psychiatric symptoms along with frequent medical comorbidities. Difficulty accessing specialty care and the difficulty of accommodating the needs of these patients in outpatient clinics leads to more emergency department visits, more inpatient psychiatric admissions, and likely to overall suboptimal care. Cincinnati Children's created a neurobehavioral psychiatry continuum of care to address these specialized needs. It consists of a multidisciplinary team focusing on coordinated outpatient psychiatric care and inpatient psychiatry crisis management. Assignment of a nurse care manager to high acuity patients helps coordinates care throughout the continuum. Psychiatric services within the continuum of care significantly expanded over the last four years, with five psychiatrists, four psychiatric nurse practitioners, and a trainee clinic for child psychiatry fellows seeing patients within a dedicated psychiatry clinic space. The continuum also significantly expanded the number and types of outpatient services offered to patients within the continuum including psychology, neurology, genetics, adaptive equipment evaluations, speech therapy, and occupational therapy. In FY19 the neurobehavioral psychiatry inpatient unit expanded into a new 10-bed specialized unit designed to meet the needs of this patient population. In addition, the development of a direct admission waitlist system helps prevent ER visits for high-acuity patients needing hospitalization on the specialized inpatient unit. This continuum of care model is an effective way to manage youth with ASD and/or intellectual disabilities across disciplines and care settings. Enrollment in the nurse care management program associates with a significant decrease in annual ER visits, with an estimated annual cost savings of $950 per patient per year. In addition, use of the direct admission waitlist system for the inpatient neurobehavioral psychiatry unit effectively helped to avoid 170 ER visits while saving an estimated $374,000 in health care utilization costs in FY19.