Faculty Bibliography

Neuroimaging-based subtyping is increasingly used to explain heterogeneity in psychiatric disorders. However, the clinical utility of these subtyping efforts remains unclear, and replication has been challenging. Here we examined how the choice of neuroimaging measures influences the derivation of neuro-subtypes and the consequences for clinical delineation. On a clinically heterogeneous dataset (total n = 566) that included controls (n = 268) and cases (n = 298) of psychiatric conditions, including individuals diagnosed with post-traumatic stress disorder (PTSD), traumatic brain injury (TBI), and comorbidity of both (PTSD&TBI), we identified neuro-subtypes among the cases using either structural, resting-state, or task-based measures. The neuro-subtypes for each modality had high internal validity but did not significantly differ in their clinical and cognitive profiles. We further show that the choice of neuroimaging measures for subtyping substantially impacts the identification of neuro-subtypes, leading to low concordance across subtyping solutions. Similar variability in neuro-subtyping was found in an independent dataset (n = 1642) comprised of major depression disorder (MDD, n = 848) and controls (n = 794). Our results suggest that the highly anticipated relationships between neuro-subtypes and clinical features may be difficult to discover.

We introduce a simple variant of a Purely Random Forest, an Absolute Random Forest (ARF) for clustering. At every node splits of units are determined by a randomly chosen feature and a random threshold drawn from a uniform distribution whose support, the range of the selected feature in the root node, does not change. This enables closed-form estimators of parameters, such as pairwise proximities, to be obtained without having to grow a forest. The probabilistic structure corresponding to an ARF is called a Treeless Absolute Random Forest (TARF). With high probability, the algorithm will split units whose feature vectors are far apart and keep together units whose feature vectors are similar. Thus, the underlying structure of the data drives the growth of the tree. The expected value of pairwise proximities is obtained for three pathway functions. One, a completely common pathway function, is an indicator of whether a pair of units follow the same path from the root to the leaf node. The properties of TARF-based proximity estimators for clustering and classification are compared to other methods in eight real-world data sets and in simulations. Results show substantial performance and computing efficiencies of particular value for large data sets.

Post-traumatic stress disorder (PTSD) is a mental disorder diagnosed by clinical interviews, self-report measures and neuropsychological testing. Traumatic brain injury (TBI) can have neuropsychiatric symptoms similar to PTSD. Diagnosing PTSD and TBI is challenging and more so for providers lacking specialized training facing time pressures in primary care and other general medical settings. Diagnosis relies heavily on patient self-report and patients frequently under-report or over-report their symptoms due to stigma or seeking compensation. We aimed to create objective diagnostic screening tests utilizing Clinical Laboratory Improvement Amendments (CLIA) blood tests available in most clinical settings. CLIA blood test results were ascertained in 475 male veterans with and without PTSD and TBI following warzone exposure in Iraq or Afghanistan. Using random forest (RF) methods, four classification models were derived to predict PTSD and TBI status. CLIA features were selected utilizing a stepwise forward variable selection RF procedure. The AUC, accuracy, sensitivity, and specificity were 0.730, 0.706, 0.659, and 0.715, respectively for differentiating PTSD and healthy controls (HC), 0.704, 0.677, 0.671, and 0.681 for TBI vs. HC, 0.739, 0.742, 0.635, and 0.766 for PTSD comorbid with TBI vs HC, and 0.726, 0.723, 0.636, and 0.747 for PTSD vs. TBI. Comorbid alcohol abuse, major depressive disorder, and BMI are not confounders in these RF models. Markers of glucose metabolism and inflammation are among the most significant CLIA features in our models. Routine CLIA blood tests have the potential for discriminating PTSD and TBI cases from healthy controls and from each other. These findings hold promise for the development of accessible and low-cost biomarker tests as screening measures for PTSD and TBI in primary care and specialty settings.

BACKGROUND:De-escalation of behavioral emergencies in the inpatient medical setting may involve restrictive clinical interventions that directly challenge patient autonomy. OBJECTIVE:We describe a quality improvement framework used to examine associations between patient characteristics and behavioral emergency de-escalation strategies. This project may inform other Consultation-Liaison Psychiatry teams seeking to promote equity in care. METHODS:We examined behavioral emergency response team (BERT) management at an urban, tertiary-care medical center in the United States over a 3-year period. BERT data from an existing dataset were combined with demographic information from the hospital's electronic medical record. Race and ethnic identities were categorized as Black, Hispanic, Asian, White, and unknown. BERT events were coded based on the most restrictive intervention utilized per unique patient. Cross-tabulations and adjusted odds ratios from multivariate logistic regression were used to identify quality improvement targets in this exploratory project. RESULTS:The sample included N = 902 patients and 1532 BERT events. The most frequent intervention reached was verbal de-escalation (n = 419 patients, 46.45%) and the least frequent was 4-point restraints (n = 29 patients, 3.2%). Half of BERT activations for Asian and a third for Hispanic patients required interpreter services. Anxiety and cognitive disorders and 2 BERT interventions, verbal de-escalation, and intramuscular/intravenous/ medications, were significantly associated with race/ethnic category. The most restrictive intervention for BERTs involving Black and Asian patients were verbal de-escalation (60.1%) and intramuscular/intravenous(53.7%), respectively. These proportions were higher compared with other race/ethnic groups. There was a greater percentage of patients from the unknown (6.3%) and Black (5.9%) race/ethnic groups placed in 4-point restraints compared with other groups (3.2%) that did not reach statistical significance. A logistic regression model predicting 4-point restraints indicated that younger age, multiple BERTs, and violent behavior as a reason for BERT activation, but not race/ethnic group, resulted in significantly higher odds. CONCLUSIONS:This project illustrates that a quality improvement framework utilizing existing clinical data can be used to engage in organizational introspection and identify potential areas of bias in BERT management. Our findings suggest opportunities for further exploration, enhanced education, and programmatic improvements regarding BERT intervention; 4-point restraints; interpreter services; and the influence of race on perception of psychopathology.

The quality of a cluster analysis of unlabeled units depends on the quality of the between units dissimilarity measures. Data dependent dissimilarity is more objective than data independent geometric measures such as Euclidean distance. As suggested by Breiman, many data driven approaches are based on decision tree ensembles, such as a random forest (RF), that produce a proximity matrix that can easily be transformed into a dissimilarity matrix. A RF can be obtained using labels that distinguish units with real data from units with synthetic data. The resulting dissimilarity matrix is input to a clustering program and units are assigned labels corresponding to cluster membership. We introduce a General Iterative Cluster (GIC) algorithm that improves the proximity matrix and clusters of the base RF. The cluster labels are used to grow a new RF yielding an updated proximity matrix which is entered into the clustering program. The process is repeated until convergence. The same procedure can be used with many base procedures such as the Extremely Randomized Tree ensemble. We evaluate the performance of the GIC algorithm using benchmark and simulated data sets. The properties measured by the Silhouette Score are substantially superior to the base clustering algorithm. The GIC package has been released in R: https://cran.r-project.org/web/packages/GIC/index.html.

Background/Significance: During the COVID-19 pandemic, people with substance use disorders have experienced increased rates of overdose, decreased access to substance use disorder treatment, and increased risk for adverse COVID outcomes (NIDA, 2020). Throughout the pandemic, NYU Langone Health has continued using the Tobacco, Alcohol, and Prescription Substance (TAPS) screening tool for all inpatient admissions in order to identify and provide proactive consultation to hospitalized patients at risk for substance use disorders.
Method(s): We conducted a retrospective review of adult inpatient medical and surgical admissions to NYU Langone Health, using data collected from a pre-defined Epic report based on TAPS documentation. We compared groups pre-COVID-19 pandemic (defined as 9/2018-9/2019) and during COVID-19 pandemic (defined as 3/2020-3/2021) for the following outcomes: (1) nursing compliance rate with TAPS administration, (2) prevalence of patients with substance use disorders as measured by positive TAPS screen, and (3) severity of alcohol use disorder among patients with TAPS positive for alcohol.
Result(s): During the pre-COVID-19 period, 24,057 patients were screened with a compliance rate of 90% and a positivity rate of 6% (N=1673). ICU compliance was 84%. Prevalence of patients at risk for various substance use disorders was as follows: 4.3% (N=1027) alcohol, 1.5% (N=357) cannabis, 0.32% (N=78) heroin, 0.24% (N=57) opiates, 0.15% (N=35) sedatives, 0.48% (N=116) stimulants, and 0.01% (N=3) prescription stimulants. Of positive alcohol screens, 26.7% (274/1027) represented the highest severity of use (Alcohol Score 4). During the COVID-19 period, 17,931 patients were screened with a compliance rate of 82% and positivity rate of 6% (N=1374). ICU compliance was 74%. Prevalence of patients at risk for various substance use disorders was as follows: 4.3% (N=772) alcohol, 1.5% (N=272) cannabis, 0.60% (N=108) heroin, 0.26% (N=46) opiates, 0.20% (N=35) sedatives, 0.69% (N=124) stimulants, and 0.04% (N=7) prescription stimulants. Of positive alcohol screens, 41.2% (318/772) were highest severity. We were unable to meaningfully test for significant given limitations of Epic datasets and variability in unit composition and staffing throughout COVID-19 period.
Discussion(s): There was decreased compliance with TAPS administration during COVID-19 as compared to pre-COVID-19, as well as overall low compliance in ICUs during both time periods. There were similar rates of positive screens for all substance use disorders pre-COVID-19 and during COVID-19, with an increase in positive heroin and other opiate screens during COVID-19. Among patients with positive alcohol screens, there was increased severity of alcohol scores during COVID-19 relative to pre-COVID-19. Conclusion/Implications: These results suggest a change in patterns of substance use during the COVID-19 pandemic, consistent with findings from prior studies of increased opioid overdoses (Slavova 2020, Georgia Department of Public Health 2020) and severity of substance use (NIDA 2020). Poor ICU compliance suggests increased barriers to TAPS administration in patients with critical illness and/or altered mental status, which may lead to decreased identification and treatment of patients at increased risk for substance use disorders. These results may inform clinical practice and future studies regarding utilization of TAPS screen and proactive addiction psychiatry consultation service in acute care settings. References: 1. NIDA. 2020, September 14. Addressing the Unique Challenges of COVID-19 for People in Recovery. Retrieved from https://www.drugabuse.gov/about-nida/noras-blog/2020/09/addressing-u nique-challenges-covid-19-people-in-recovery on 2021, March 15 2. Slavova, S., Rock, P., Bush, H. M., Quesinberry, D., & Walsh, S. L. (2020). Signal of increased opioid overdose during COVID-19 from emergency medical services data. Drug and alcohol dependence, 214, 108176. 3. Georgia Department of Public Health. 2020, June 19. Suspected Drug Overdose Increases in Georgia Amid COVID-19. Retrieved from https://www.drugabuse.gov/sites/default/files/suspected_drug_overdos e_increases_in_georgia_amid_covid-19_1.pdf
Copyright

Background: CL psychiatrists are uniquely positioned to combat structural racism in medicine Currently, there are no published papers examining racial bias in the management of psychiatric emergencies in the general medical hospital. Given the potential for restrictive clinical interventions that directly challenge a patient's autonomy (including intramuscular injections and restraints), our group embarked on a long-term, quality improvement project to detect and address racial bias affecting the clinical management of these psychiatric emergencies.
Method(s): Our institution has a multidisciplinary behavioral code team known as the Behavioral Emergency Response Team (BERT) that responds to behavioral emergencies throughout the medical hospital. Secondary BERT event data occurring from 2017 to 2020 was combined with demographic data from the electronic medical record. Race and ethnic data were collapsed into unique, phenotypic categories. BERT events were coded based on the most restrictive intervention utilized. Descriptive statistics were used to describe the sample and examine whether race / ethnicity correlated with BERT intervention utilized, diagnostic impression, reason for BERT activation, or recurrent BERTs.
Result(s): Our sample included 1532 BERT events representing N = 902 unique patients. The main interaction of BERT intervention by Race / Ethnic category reached statistical significance (p=0.04). Though most BERTs only required verbal de-escalation (n=419, 46.45%), 3% of BERTs (n = 29) escalated to 4-pt restraints (most restrictive intervention). Though reaching level 5 was rare, Black patients had a statistically significant higher likelihood of receiving this intervention compared with White patients (6% v 2%, p=0.027) and compared with all other non-Black patients (6% v 2%, p=0.040). Although the overall comparison for Race/Ethnicity and the diagnostic impression "Psychosis" did not reach significance (p=0.086), targeted analysis showed that Black patients were significantly more likely to have "Psychosis" listed as a contributing factor compared with White patients (p=0.009) and all other non-Black patients (p=0.016). Several other comparisons with Race / Ethnic category reached statistical significance: Age (p=0.048), and need for interpreter yes/no (p<0.001). Closer examination of the interaction of Race/Ethnicity x Need for Interpreter revealed that half of events involving Asian patients (n=22, 53.66%) and a third of events involving Hispanic patients (n=29, 30.53%) required interpreter services.
Discussion(s): This study demonstrates the feasibility of investigating racial bias in behavioral emergency management. The results of this preliminary analysis suggest multiple areas for enhanced education, self-awareness development, and programmatic improvement to target systemic racism, decrease racial bias, and improve patient care. These areas include bias in restraints use, the role of language in behavioral emergencies, and the influence of race on perception of underlying diagnosis.
Copyright

OBJECTIVE:To further the precision medicine goal of tailoring medical treatment to individual patient characteristics by providing a method of analysis of the effect of test treatment, T, compared to a reference treatment, R, in participants in a RCT who are likely responders to T. METHODS:Likely responders to T are individuals whose expected response at baseline exceeds a prespecified minimum. A prognostic score, the expected response predicted as a function of baseline covariates, is obtained at trial completion. It is a balancing score that can be used to match likely responders randomized to T with those randomized to R; the result is comparable treatment groups that have a common covariance distribution. Treatments are compared based on observed outcomes in this enriched sample. The approach is illustrated in a RCT comparing two treatments for opioid use disorder. RESULTS:A standard statistical analysis of the opioid use disorder RCT found no treatment difference in the total sample. However, a subset of likely responders to T were identified and in this group, T was statistically superior to R. CONCLUSION/CONCLUSIONS:The causal treatment effect of T relative to R among likely responders may be more important than the effect in the whole target population. The prognostic score function provides quantitative information to support patient specific treatment decisions regarding T furthering the goal of precision medicine.

Active-duty Army personnel can be exposed to traumatic warzone events and are at increased risk for developing post-traumatic stress disorder (PTSD) compared with the general population. PTSD is associated with high individual and societal costs, but identification of predictive markers to determine deployment readiness and risk mitigation strategies is not well understood. This prospective longitudinal naturalistic cohort study-the Fort Campbell Cohort study-examined the value of using a large multidimensional dataset collected from soldiers prior to deployment to Afghanistan for predicting post-deployment PTSD status. The dataset consisted of polygenic, epigenetic, metabolomic, endocrine, inflammatory and routine clinical lab markers, computerized neurocognitive testing, and symptom self-reports. The analysis was computed on active-duty Army personnel (N = 473) of the 101st Airborne at Fort Campbell, Kentucky. Machine-learning models predicted provisional PTSD diagnosis 90-180 days post deployment (random forest: AUC = 0.78, 95% CI = 0.67-0.89, sensitivity = 0.78, specificity = 0.71; SVM: AUC = 0.88, 95% CI = 0.78-0.98, sensitivity = 0.89, specificity = 0.79) and longitudinal PTSD symptom trajectories identified with latent growth mixture modeling (random forest: AUC = 0.85, 95% CI = 0.75-0.96, sensitivity = 0.88, specificity = 0.69; SVM: AUC = 0.87, 95% CI = 0.79-0.96, sensitivity = 0.80, specificity = 0.85). Among the highest-ranked predictive features were pre-deployment sleep quality, anxiety, depression, sustained attention, and cognitive flexibility. Blood-based biomarkers including metabolites, epigenomic, immune, inflammatory, and liver function markers complemented the most important predictors. The clinical prediction of post-deployment symptom trajectories and provisional PTSD diagnosis based on pre-deployment data achieved high discriminatory power. The predictive models may be used to determine deployment readiness and to determine novel pre-deployment interventions to mitigate the risk for deployment-related PTSD.