Evaluation of a Model to Target High-risk Psychiatric Inpatients for an Intensive Postdischarge Suicide Prevention Intervention.

Importance: The months after psychiatric hospital discharge are a time of high risk for suicide. Intensive postdischarge case management, although potentially effective in suicide prevention, is likely to be cost-effective only if targeted at high-risk patients. A previously developed machine learning (ML) model showed that postdischarge suicides can be predicted from electronic health records and geospatial data, but it is unknown if prediction could be improved by adding additional information. Objective: To determine whether model prediction could be improved by adding information extracted from clinical notes and public records. Design, Setting, and Participants: Models were trained to predict suicides in the 12 months after Veterans Health Administration (VHA) short-term (less than 365 days) psychiatric hospitalizations between the beginning of 2010 and September 1, 2012 (299 050 hospitalizations, with 916 hospitalizations followed within 12 months by suicides) and tested in the hospitalizations from September 2, 2012, to December 31, 2013 (149 738 hospitalizations, with 393 hospitalizations followed within 12 months by suicides). Validation focused on net benefit across a range of plausible decision thresholds. Predictor importance was assessed with Shapley additive explanations (SHAP) values. Data were analyzed from January to August 2022. Main Outcomes and Measures: Suicides were defined by the National Death Index. Base model predictors included VHA electronic health records and patient residential data. The expanded predictors came from natural language processing (NLP) of clinical notes and a social determinants of health (SDOH) public records database. Results: The model included 448 788 unique hospitalizations. Net benefit over risk horizons between 3 and 12 months was generally highest for the model that included both NLP and SDOH predictors (area under the receiver operating characteristic curve range, 0.747-0.780; area under the precision recall curve relative to the suicide rate range, 3.87-5.75). NLP and SDOH predictors also had the highest predictor class-level SHAP values (proportional SHAP = 64.0% and 49.3%, respectively), although the single highest positive variable-level SHAP value was for a count of medications classified by the US Food and Drug Administration as increasing suicide risk prescribed the year before hospitalization (proportional SHAP = 15.0%). Conclusions and Relevance: In this study, clinical notes and public records were found to improve ML model prediction of suicide after psychiatric hospitalization. The model had positive net benefit over 3-month to 12-month risk horizons for plausible decision thresholds. Although caution is needed in inferring causality based on predictor importance, several key predictors have potential intervention implications that should be investigated in future studies.

The Evolution of a Large Biobank at Mass General Brigham.

The Mass General Brigham Biobank (formerly Partners HealthCare Biobank) is a large repository of biospecimens and data linked to extensive electronic health record data and survey data. Its objective is to support and enable translational research focused on genomic, environmental, biomarker and family history associations with disease phenotypes. The Biobank has enrolled more than 135,000 participants, generated genomic data on more than 65,000 of its participants, distributed approximately 153,000 biospecimens, and served close to 450 institutional studies with biospecimens or data. Although the Biobank has been successful, based on some measures of output, this has required substantial institutional investment. In addition, several challenges are ongoing, including: (1) developing a sustainable cost model that doesn't rely as heavily on institutional funding; (2) integrating Biobank operations into clinical workflows; and (3) building a research resource that is diverse and promotes equity in research. Here, we describe the evolution of the Biobank and highlight key lessons learned that may inform other efforts to build biobanking efforts in health system contexts.

Effects of social support on depression risk during the COVID-19 pandemic: What support types and for whom?

Background: Rates of depression have increased worldwide during the COVID-19 pandemic. One known protective factor for depression is social support, but more work is needed to quantify the extent to which social support could reduce depression risk during a global crisis, and specifically to identify which types of support are most helpful, and who might benefit most. Methods: Data were obtained from participants in the All of Us Research Program who responded to the COVID-19 Participant Experience (COPE) survey administered monthly from May 2020 to July 2020 (N=69,066, 66% female). Social support was assessed using 10 items measuring emotional/informational support (e.g., someone to confide in or talk to about yourself or your problems), positive social interaction support (e.g., someone to do things with to help you get your mind off things), and tangible support (e.g., someone to help with daily chores if sick). Elevated depression symptoms were defined based on having a moderate-to-severe (≥10) score on the Patient Health Questionnaire (PHQ-9). Mixed-effects logistic regression models were used to test associations across time between overall social support and its subtypes with depression, adjusting for age, sex, race, ethnicity, and socioeconomic factors. We then assessed interactions between social support and potential effect modifiers: age, sex, pre-pandemic mood disorder, and pandemic-related stressors (e.g., financial insecurity). Results: Approximately 16% of the sample experienced elevated depressive symptoms. Overall social support was associated with significantly reduced odds of depression (adjusted odds ratio, aOR [95% CI]=0.44 [0.42-0.45]). Among subtypes, emotional/informational support (aOR=0.42 [0.41-0.43]) and positive social interactions (aOR=0.43 [0.41-0.44]) showed the largest protective associations with depression, followed by tangible support (aOR=0.63 [0.61-0.65]). Sex, age, and pandemic-related financial stressors were statistically significant modifiers of the association between social support and depression. Conclusions: Individuals reporting higher levels of social support were at reduced risk of depression during the early COVID-19 pandemic. The perceived availability of emotional support and positive social interactions, more so than tangible support, was key. Individuals more vulnerable to depression (e.g., women, younger individuals, and those experiencing financial stressors) may particularly benefit from enhanced social support, supporting a precision prevention approach.

Prediction of Suicide Attempts Using Clinician Assessment, Patient Self-report, and Electronic Health Records.

Importance: Half of the people who die by suicide make a health care visit within 1 month of their death. However, clinicians lack the tools to identify these patients. Objective: To predict suicide attempts within 1 and 6 months of presentation at an emergency department (ED) for psychiatric problems. Design, Setting, and Participants: This prognostic study assessed the 1-month and 6-month risk of suicide attempts among 1818 patients presenting to an ED between February 4, 2015, and March 13, 2017, with psychiatric problems. Data analysis was performed from May 1, 2020, to November 19, 2021. Main Outcomes and Measures: Suicide attempts 1 and 6 months after presentation to the ED were defined by combining data from electronic health records (EHRs) with patient 1-month (n = 1102) and 6-month (n = 1220) follow-up surveys. Ensemble machine learning was used to develop predictive models and a risk score for suicide. Results: A total of 1818 patients participated in this study (1016 men [55.9%]; median age, 33 years [IQR, 24-46 years]; 266 Hispanic patients [14.6%]; 1221 non-Hispanic White patients [67.2%], 142 non-Hispanic Black patients [7.8%], 64 non-Hispanic Asian patients [3.5%], and 125 non-Hispanic patients of other race and ethnicity [6.9%]). A total of 137 of 1102 patients (12.9%; weighted prevalence) attempted suicide within 1 month, and a total of 268 of 1220 patients (22.0%; weighted prevalence) attempted suicide within 6 months. Clinicians' assessment alone was little better than chance at predicting suicide attempts, with externally validated area under the receiver operating characteristic curve (AUC) of 0.67 for the 1-month model and 0.60 for the 6-month model. Prediction accuracy was slightly higher for models based on EHR data (1-month model: AUC, 0.71; 6 month model: AUC, 0.65) and was best using patient self-reports (1-month model: AUC, 0.76; 6-month model: AUC, 0.77), especially when patient self-reports were combined with EHR and/or clinician data (1-month model: AUC, 0.77; and 6 month model: AUC, 0.79). A model that used only 20 patient self-report questions and an EHR-based risk score performed similarly well (1-month model: AUC, 0.77; 6 month model: AUC, 0.78). In the best 1-month model, 30.7% (positive predicted value) of the patients classified as having highest risk (top 25% of the sample) made a suicide attempt within 1 month of their ED visit, accounting for 64.8% (sensitivity) of all 1-month attempts. In the best 6-month model, 46.0% (positive predicted value) of the patients classified at highest risk made a suicide attempt within 6 months of their ED visit, accounting for 50.2% (sensitivity) of all 6-month attempts. Conclusions and Relevance: This prognostic study suggests that the ability to identify patients at high risk of suicide attempt after an ED visit for psychiatric problems improved using a combination of patient self-reports and EHR data.

Returning actionable genomic results in a research biobank: Analytic validity, clinical implementation, and resource utilization.

Over 100 million research participants around the world have had research array-based genotyping (GT) or genome sequencing (GS), but only a small fraction of these have been offered return of actionable genomic findings (gRoR). Between 2017 and 2021, we analyzed genomic results from 36,417 participants in the Mass General Brigham Biobank and offered to confirm and return pathogenic and likely pathogenic variants (PLPVs) in 59 genes. Variant verification prior to participant recontact revealed that GT falsely identified PLPVs in 44.9% of samples, and GT failed to identify 72.0% of PLPVs detected in a subset of samples that were also sequenced. GT and GS detected verified PLPVs in 1% and 2.5% of the cohort, respectively. Of 256 participants who were alerted that they carried actionable PLPVs, 37.5% actively or passively declined further disclosure. 76.3% of those carrying PLPVs were unaware that they were carrying the variant, and over half of those met published professional criteria for genetic testing but had never been tested. This gRoR protocol cost approximately $129,000 USD per year in laboratory testing and research staff support, representing $14 per participant whose DNA was analyzed or $3,224 per participant in whom a PLPV was confirmed and disclosed. These data provide logistical details around gRoR that could help other investigators planning to return genomic results.

Accuracy Requirements for Cost-effective Suicide Risk Prediction Among Primary Care Patients in the US.

Importance: Several statistical models for predicting suicide risk have been developed, but how accurate such models must be to warrant implementation in clinical practice is not known. Objective: To identify threshold values of sensitivity, specificity, and positive predictive value that a suicide risk prediction method must attain to cost-effectively target a suicide risk reduction intervention to high-risk individuals. Design, Setting, and Participants: This economic evaluation incorporated published data on suicide epidemiology, the health care and societal costs of suicide, and the costs and efficacy of suicide risk reduction interventions into a novel decision analytic model. The model projected suicide-related health economic outcomes over a lifetime horizon among a population of US adults with a primary care physician. Data analysis was performed from September 19, 2019, to July 5, 2020. Interventions: Two possible interventions were delivered to individuals at high predicted risk: active contact and follow-up (ACF; relative risk of suicide attempt, 0.83; annual health care cost, $96) and cognitive behavioral therapy (CBT; relative risk of suicide attempt, 0.47; annual health care cost, $1088). Main Outcomes and Measures: Fatal and nonfatal suicide attempts, quality-adjusted life-years (QALYs), health care sector costs and societal costs (in 2016 US dollars), and incremental cost-effectiveness ratios (ICERs) (with ICERs ≤$150 000 per QALY designated cost-effective). Results: With a specificity of 95% and a sensitivity of 25%, primary care-based suicide risk prediction could reduce suicide death rates by 0.5 per 100 000 person-years (if used to target ACF) or 1.6 per 100 000 person-years (if used to target CBT) from a baseline of 15.3 per 100 000 person-years. To be cost-effective from a health care sector perspective at a specificity of 95%, a risk prediction method would need to have a sensitivity of 17.0% or greater (95% CI, 7.4%-37.3%) if used to target ACF and 35.7% or greater (95% CI, 23.1%-60.3%) if used to target CBT. To achieve cost-effectiveness, ACF required positive predictive values of 0.8% for predicting suicide attempt and 0.07% for predicting suicide death; CBT required values of 1.7% for suicide attempt and 0.2% for suicide death. Conclusions and Relevance: These findings suggest that with sufficient accuracy, statistical suicide risk prediction models can provide good health economic value in the US. Several existing suicide risk prediction models exceed the accuracy thresholds identified in this analysis and thus may warrant pilot implementation in US health care systems.

Greater Neurobiological Resilience to Chronic Socioeconomic or Environmental Stressors Associates With Lower Risk for Cardiovascular Disease Events.

BACKGROUND: Chronic exposure to socioeconomic or environmental stressors associates with greater stress-related neurobiological activity (ie, higher amygdalar activity [AmygA]) and higher risk of major adverse cardiovascular events (MACE). However, among individuals exposed to such stressors, it is unknown whether neurobiological resilience (NBResilience, defined as lower AmygA despite stress exposure) lowers MACE risk. We tested the hypotheses that NBResilience protects against MACE, and that it does so through decreased bone marrow activity and arterial inflammation. METHODS: Individuals underwent 18F-fluorodeoxyglucose positron emission tomography/computed tomography; AmygA, bone marrow activity, and arterial inflammation were quantified. Chronic socioeconomic and environmental stressors known to associate with AmygA and MACE (ie, transportation noise exposure, neighborhood median household income, and crime rate) were quantified. Heightened stress exposure was defined as exposure to at least one chronic stressor (ie, the highest tertile of noise exposure or crime or lowest tertile of income). MACE within 5 years of imaging was adjudicated. Relationships were evaluated using linear and Cox regression, Kaplan-Meier survival, and mediation analyses. RESULTS: Of 254 individuals studied (median age [interquartile range]: 57 years [46-67], 36.7% male), 166 were exposed to at least one chronic stressor. Among stress-exposed individuals, 12 experienced MACE over a median follow-up of 3.75 years. Among this group, higher AmygA (ie, lower resilience) associated with higher bone marrow activity (standardized ß [95% CI]: 0.192 [0.030-0.353], P=0.020), arterial inflammation (0.203 [0.055-0.351], P=0.007), and MACE risk (standardized hazard ratio [95% CI]: 1.927 [1.370-2.711], P=0.001). The effect of NBResilience on MACE risk was significantly mediated by lower arterial inflammation (P<0.05). CONCLUSIONS: Among individuals who are chronically exposed to socioeconomic or environmental stressors, NBResilience (AmygA <1 SD above the mean) associates with a >50% reduction in MACE risk, potentially via reduced arterial inflammation. These data raise the possibility that enhancing NBResilience may decrease the burden of cardiovascular disease.

Using Administrative Data to Predict Suicide After Psychiatric Hospitalization in the Veterans Health Administration System.

There is a very high suicide rate in the year after psychiatric hospital discharge. Intensive postdischarge case management programs can address this problem but are not cost-effective for all patients. This issue can be addressed by developing a risk model to predict which inpatients might need such a program. We developed such a model for the 391,018 short-term psychiatric hospital admissions of US veterans in Veterans Health Administration (VHA) hospitals 2010-2013. Records were linked with the National Death Index to determine suicide within 12 months of hospital discharge (n=771). The Super Learner ensemble machine learning method was used to predict these suicides for time horizon between 1 week and 12 months after discharge in a 70% training sample. Accuracy was validated in the remaining 30% holdout sample. Predictors included VHA administrative variables and small area geocode data linked to patient home addresses. The models had AUC=.79-.82 for time horizons between 1 week and 6 months and AUC=.74 for 12 months. An analysis of operating characteristics showed that 22.4%-32.2% of patients who died by suicide would have been reached if intensive case management was provided to the 5% of patients with highest predicted suicide risk. Positive predictive value (PPV) at this higher threshold ranged from 1.2% over 12 months to 3.8% per case manager year over 1 week. Focusing on the low end of the risk spectrum, the 40% of patients classified as having lowest risk account for 0%-9.7% of suicides across time horizons. Variable importance analysis shows that 51.1% of model performance is due to psychopathological risk factors accounted, 26.2% to social determinants of health, 14.8% to prior history of suicidal behaviors, and 6.6% to physical disorders. The paper closes with a discussion of next steps in refining the model and prospects for developing a parallel precision treatment model.

The "All of Us" Research Program.

Knowledge gained from observational cohort studies has dramatically advanced the prevention and treatment of diseases. Many of these cohorts, however, are small, lack diversity, or do not provide comprehensive phenotype data. The All of Us Research Program plans to enroll a diverse group of at least 1 million persons in the United States in order to accelerate biomedical research and improve health. The program aims to make the research results accessible to participants, and it is developing new approaches to generate, access, and make data broadly available to approved researchers. All of Us opened for enrollment in May 2018 and currently enrolls participants 18 years of age or older from a network of more than 340 recruitment sites. Elements of the program protocol include health questionnaires, electronic health records (EHRs), physical measurements, the use of digital health technology, and the collection and analysis of biospecimens. As of July 2019, more than 175,000 participants had contributed biospecimens. More than 80% of these participants are from groups that have been historically underrepresented in biomedical research. EHR data on more than 112,000 participants from 34 sites have been collected. The All of Us data repository should permit researchers to take into account individual differences in lifestyle, socioeconomic factors, environment, and biologic characteristics in order to advance precision diagnosis, prevention, and treatment.

Assessment of Bidirectional Relationships Between Physical Activity and Depression Among Adults: A 2-Sample Mendelian Randomization Study.

Importance: Increasing evidence shows that physical activity is associated with reduced risk for depression, pointing to a potential modifiable target for prevention. However, the causality and direction of this association are not clear; physical activity may protect against depression, and/or depression may result in decreased physical activity. Objective: To examine bidirectional relationships between physical activity and depression using a genetically informed method for assessing potential causal inference. Design, Setting, and Participants: This 2-sample mendelian randomization (MR) used independent top genetic variants associated with 2 physical activity phenotypes-self-reported (n = 377 234) and objective accelerometer-based (n = 91 084)-and with major depressive disorder (MDD) (n = 143 265) as genetic instruments from the largest available, nonoverlapping genome-wide association studies (GWAS). GWAS were previously conducted in diverse observational cohorts, including the UK Biobank (for physical activity) and participating studies in the Psychiatric Genomics Consortium (for MDD) among adults of European ancestry. Mendelian randomization estimates from each genetic instrument were combined using inverse variance weighted meta-analysis, with alternate methods (eg, weighted median, MR Egger, MR-Pleiotropy Residual Sum and Outlier [PRESSO]) and multiple sensitivity analyses to assess horizontal pleiotropy and remove outliers. Data were analyzed from May 10 through July 31, 2018. Main Outcomes and Measures: MDD and physical activity. Results: GWAS summary data were available for a combined sample size of 611 583 adult participants. Mendelian randomization evidence suggested a protective relationship between accelerometer-based activity and MDD (odds ratio [OR], 0.74 for MDD per 1-SD increase in mean acceleration; 95% CI, 0.59-0.92; P = .006). In contrast, there was no statistically significant relationship between MDD and accelerometer-based activity (ß = -0.08 in mean acceleration per MDD vs control status; 95% CI, -0.47 to 0.32; P = .70). Furthermore, there was no significant relationship between self-reported activity and MDD (OR, 1.28 for MDD per 1-SD increase in metabolic-equivalent minutes of reported moderate-to-vigorous activity; 95% CI, 0.57-3.37; P = .48), or between MDD and self-reported activity (ß = 0.02 per MDD in standardized metabolic-equivalent minutes of reported moderate-to-vigorous activity per MDD vs control status; 95% CI, -0.008 to 0.05; P = .15). Conclusions and Relevance: Using genetic instruments identified from large-scale GWAS, robust evidence supports a protective relationship between objectively assessed-but not self-reported-physical activity and the risk for MDD. Findings point to the importance of objective measurement of physical activity in epidemiologic studies of mental health and support the hypothesis that enhancing physical activity may be an effective prevention strategy for depression.

Socioeconomic disadvantage and neural development from infancy through early childhood.

BACKGROUND: Early social experiences are believed to shape neurodevelopment, with potentially lifelong consequences. Yet minimal evidence exists regarding the role of the social environment on children's neural functioning, a core domain of neurodevelopment. METHODS: We analysed data from 36 443 participants in the United States Collaborative Perinatal Project, a socioeconomically diverse pregnancy cohort conducted between 1959 and 1974. Study outcomes included: physician (neurologist or paediatrician)-rated neurological abnormality neonatally and thereafter at 4 months and 1 and 7 years; indicators of neurological hard signs and soft signs; and indicators of autonomic nervous system function. RESULTS: Children born to socioeconomically disadvantaged parents were more likely to exhibit neurological abnormalities at 4 months [odds ratio (OR) = 1.20; 95% confidence interval (CI) = 1.06, 1.37], 1 year (OR = 1.35; CI = 1.17, 1.56), and 7 years (OR = 1.67; CI = 1.48, 1.89), and more likely to exhibit neurological hard signs (OR = 1.39; CI = 1.10, 1.76), soft signs (OR = 1.26; CI = 1.09, 1.45) and autonomic nervous system dysfunctions at 7 years. Pregnancy and delivery complications, themselves associated with socioeconomic disadvantage, did not account for the higher risks of neurological abnormalities among disadvantaged children. CONCLUSIONS: Parental socioeconomic disadvantage was, independently from pregnancy and delivery complications, associated with abnormal child neural development during the first 7 years of life. These findings reinforce the importance of the early environment for neurodevelopment generally, and expand knowledge regarding the domains of neurodevelopment affected by environmental conditions. Further work is needed to determine the mechanisms linking socioeconomic disadvantage with children's neural functioning, the timing of such mechanisms and their potential reversibility.

When is pharmacogenetic testing for antidepressant response ready for the clinic? A cost-effectiveness analysis based on data from the STAR*D study.

The potential of personalized medicine to transform the treatment of mood disorders has been widely touted in psychiatry, but has not been quantified. We estimated the costs and benefits of a putative pharmacogenetic test for antidepressant response in the treatment of major depressive disorder (MDD) from the societal perspective. Specifically, we performed cost-effectiveness analyses using state-transition probability models incorporating probabilities from the multicenter STAR*D effectiveness study of MDD. Costs and quality-adjusted life years (QALYs) were compared for sequential antidepressant trials, with or without guidance from a pharmacogenetic test for differential response to selective serotonin reuptake inhibitors (SSRIs). Likely SSRI responders received an SSRI, whereas likely nonresponders received the norepinephrine/dopamine reuptake inhibitor bupropion. For a 40-year old with MDD, applying the pharmacogenetic test and using the non-SSRI bupropion for those at higher risk for nonresponse cost $93,520 per additional QALY compared with treating all patients with an SSRI first and switching sequentially in the case of nonremission. Cost per QALY dropped below $50,000 for tests with remission rate ratios as low as 1.5, corresponding to odds ratios approximately 1.8-2.0. Tests for differential antidepressant response could thus become cost effective under certain circumstances. These circumstances, particularly availability of alternative treatment strategies and test effect sizes, can be estimated and should be considered before these tests are broadly applied in clinical settings.

Critical thinking about adverse drug effects: lessons from the psychology of risk and medical decision-making for clinical psychopharmacology.

Systematic biases in decision-making have been well characterized in medical and nonmedical fields but mostly ignored in clinical psychopharmacology. The purpose of this paper is to sensitize clinicians who prescribe psychiatric drugs to the issues of the psychology of risk, especially as they pertain to the risk of side effects. Specifically, the present analysis focuses on heuristic organization and framing effects that create cognitive biases in medical practice. Our purpose is to increase the awareness of how pharmaceutical companies may influence physicians by framing the risk of medication side effects to favor their products.

Pharmacogenetic testing in the clinical management of schizophrenia: a decision-analytic model.

Clinical application of pharmacogenetic testing has been proposed as a means of improving treatment outcomes in psychiatry. The identification of a putative genetic test for better clozapine response in schizophrenia offers an opportunity to evaluate the cost-effectiveness of such testing. The authors performed a cost-effectiveness analysis of a genetic test that may identify individuals with greater likelihood of responding to clozapine treatment. We modeled a target population of schizophrenia patients in an acute psychotic episode, using a lifetime time horizon and societal perspective. Outcome measures included life expectancy, quality-adjusted life expectancy, costs, and incremental cost-effectiveness. Effects of variations in testing parameters were also examined. For a 30-year-old with schizophrenia, applying the pharmacogenetic test and treating those predicted to respond to clozapine with clozapine-first cost US $47,705 per additional quality-adjusted life-year, compared with treating all patients with conventional agents and reserving clozapine for treatment-resistant patients. In 1-way sensitivity analyses, test sensitivity and cost had the greatest impact on the incremental cost-effectiveness. We conclude that pharmacogenetic tests may achieve utility in clinical psychiatry, although their cost-effectiveness depends on several clinical parameters. More consistent reporting of test parameters such as sensitivity and specificity would greatly facilitate assessment of future pharmacogenetic studies.