Optimizing precision medicine for second-step depression treatment: a machine learning approach.

BACKGROUND: Less than a third of patients with depression achieve successful remission with standard first-step antidepressant monotherapy. The process for determining appropriate second-step care is often based on clinical intuition and involves a protracted course of trial and error, resulting in substantial patient burden and unnecessary delay in the provision of optimal treatment. To address this problem, we adopt an ensemble machine learning approach to improve prediction accuracy of remission in response to second-step treatments. METHOD: Data were derived from the Level 2 stage of the STAR*D dataset, which included 1439 patients who were randomized into one of seven different second-step treatment strategies after failing to achieve remission during first-step antidepressant treatment. Ensemble machine learning models, comprising several individual algorithms, were evaluated using nested cross-validation on 155 predictor variables including clinical and demographic measures. RESULTS: The ensemble machine learning algorithms exhibited differential classification performance in predicting remission status across the seven second-step treatments. For the full set of predictors, AUC values ranged from 0.51 to 0.82 depending on the second-step treatment type. Predicting remission was most successful for cognitive therapy (AUC = 0.82) and least successful for other medication and combined treatment options (AUCs = 0.51-0.66). CONCLUSION: Ensemble machine learning has potential to predict second-step treatment. In this study, predictive performance varied by type of treatment, with greater accuracy in predicting remission in response to behavioral treatments than to pharmacotherapy interventions. Future directions include considering more informative predictor modalities to enhance prediction of second-step treatment response.

Rising early warning signals in affect associated with future changes in depression: a dynamical systems approach.

BACKGROUND: Predicting future states of psychopathology such as depressive episodes has been a hallmark initiative in mental health research. Dynamical systems theory has proposed that rises in certain 'early warning signals' (EWSs) in time-series data (e.g. auto-correlation, temporal variance, network connectivity) may precede impending changes in disorder severity. The current study investigates whether rises in these EWSs over time are associated with future changes in disorder severity among a group of patients with major depressive disorder (MDD). METHODS: Thirty-one patients with MDD completed the study, which consisted of daily smartphone-delivered surveys over 8 weeks. Daily positive and negative affect were collected for the time-series analyses. A rolling window approach was used to determine whether rises in auto-correlation of total affect, temporal standard deviation of total affect, and overall network connectivity in individual affect items were predictive of increases in depression symptoms. RESULTS: Results suggested that rises in auto-correlation were significantly associated with worsening in depression symptoms (r = 0.41, p = 0.02). Results indicated that neither rises in temporal standard deviation (r = -0.23, p = 0.23) nor in network connectivity (r = -0.12, p = 0.59) were associated with changes in depression symptoms. CONCLUSIONS: This study more rigorously examines whether rises in EWSs were associated with future depression symptoms in a larger group of patients with MDD. Results indicated that rises in auto-correlation were the only EWS that was associated with worsening future changes in depression.

Use of Staging Models for Treatment-Resistant Depression Is Not Helpful in Predicting Nonresponse to Acute Intravenous Ketamine Treatment.

BACKGROUND: Some staging models for treatment-resistant depression (TRD) have been developed in the attempt to predict treatment outcome, in particular with electroconvulsive therapy. However, these models have not been tested in predicting clinical outcome of ketamine treatment. We assessed the relationship between patients' classification with different TRD staging models and subsequent nonresponse to acute intravenous ketamine treatment. METHODS: A sample of 120 patients with TRD who received acute ketamine treatment from October 2018 to November 2020 were included. Intravenous ketamine was administered twice weekly for 3 weeks as acute treatment. Generalized linear models were fitted to examine if staging classification at baseline could predict percent change in the 16-item Quick Inventory of Depressive Symptomatology-Self Report (QIDS-SR16) scale. Potential confounders such as age, sex, and primary diagnosis were included in the models. Other generalized linear models were also fitted with the Bonferroni correction to investigate if other clinical variables of potential relevance could predict percent change in the QIDS-SR16. RESULTS: No TRD staging model proved accurate in predicting depressive improvement after acute ketamine treatment. Clinical variables such as age (F = 6.68, P = 0.01) and history of neuromodulation therapy (F = 5.12, P = 0.03) were negatively associated with subsequent percent improvement in the QIDS-SR16 with acute ketamine treatment. CONCLUSIONS: The efficacy of acute intravenous ketamine treatment was similar in subjects with higher and lower level of treatment resistance, using definitions based on different TRD staging models. Further exploration of ketamine treatment predictors such as age and neuromodulation therapy is warranted.

How family dementia caregivers perceive benefits of a 4-week Mentalizing Imagery Therapy program: a pilot study.

Background: Family caregivers of dementia patients experience high levels of interpersonal stress that often results in elevated anxiety, and depression, and negative impacts on interpersonal relationships. Changes in behaviors and the structure of relationships with the care recipient (CR) and others in the social milieu challenge the caregivers' ability to mentalize, or understand the links between mental states and behaviors. This study investigates the experiences and perceived benefits of family dementia caregivers who underwent Mentalizing Imagery Therapy (MIT), a treatment aiming to improve balanced self-other mentalizing and reduce psychological symptoms. Methods: Purposeful sampling was used to select 11 family dementia caregivers who underwent a 4-week pilot trial of MIT. Semi-structured interviews were completed post-intervention to identify subjective benefits, putative psychological mediators and perceived active components. Results: Caregivers reported improvements in well-being, mood, anxiety, and sleep, and a majority stated MIT helped with forming and maintaining healthier relationships. Some participants noted benefits extending to how they reacted to their social environment and perceived themselves more objectively from others' perspectives. Specific elements of MIT, including self-compassion, self-care, and the ability to reflect on emotionally arousing challenges, might have mediated these improvements. Conclusion: Family dementia caregivers perceived salutary benefits of MIT on multiple domains of well-being. The self reports suggest MIT holds promise for improving well-being, reducing non-mentalizing patterns of thought, and facilitating improvements in balanced mentalization within the caregivers' relationships.

Sex Differences in the Association between Heavy Drinking and Behavioral Distress Tolerance and Emotional Reactivity Among Non-Depressed College Students.

BACKGROUND: Heavy episodic drinking (HED) is a common behavior among college students that is associated with severe negative consequences. Negative reinforcement processes have been applied to elucidate mechanisms underlying relationships between consumption of alcohol and the desire to alleviate negative feelings. Distress tolerance (DT) and emotional reactivity are two mechanisms that are consistent with the self-medication model that may contribute to HED. The current study investigated relationships between DT, emotional reactivity, defined as frustration reactivity and irritability reactivity, and HED in a non-depressed college population. Given differential patterns of consumption and motivation for drinking between males and females, sex differences were also examined. SHORT SUMMARY: The study examined two constructs consistent with negative reinforcement processes, behavioral distress tolerance (DT) and emotional reactivity (frustration reactivity and irritability reactivity), to explain heavy episodic drinking (HED) among non-depressed college students. Behavioral DT and frustration reactivity independently predicted HED. Higher HED was associated with higher frustration reactivity and lower behavioral DT in women, but nor in men. METHODS: One-hundred-ten college students without depressive symptoms completed alcohol use measures and the Paced Auditory Serial Attention Task (PASAT-C) to assess behavioral DT and emotional reactivity. RESULTS: DT and frustration reactivity independently predicted HED. The association between DT and HED was moderated by sex such that higher levels of DT predicted higher HED among females, but not among males. Higher frustration reactivity scores were associated with a greater number of HED. CONCLUSIONS: Results provide supporting evidence that DT and emotional reactivity are distinct factors, and that they predict HED independently. Results underscore the importance of examining sex differences when evaluating the association between HED and negative reinforcement processes in this population.

Prevalence of suicidal behaviour following traumatic brain injury: Longitudinal follow-up data from the NIDRR Traumatic Brain Injury Model Systems.

OBJECTIVE: This study utilized the Traumatic Brain Injury Model Systems (TBIMS) National Database to examine the prevalence of depression and suicidal behaviour in a large cohort of patients who sustained moderate-to-severe TBI. METHOD: Participants presented to a TBIMS acute care hospital within 72 hours of injury and received acute care and comprehensive rehabilitation in a TBIMS designated brain injury inpatient rehabilitation programme. Depression and suicidal ideation were measured with the Patient Health Questionnaire (PHQ-9). Self-reported suicide attempts during the past year were recorded at each follow-up examination, at 1, 2, 3, 10, 15 and 20 years post-injury. RESULTS: Throughout the 20 years of follow-up, rates of depression ranged from 24.8-28.1%, suicidal ideation ranged from 7.0-10.1% and suicide attempts (past year) ranged from 0.8-1.7%. Participants who endorsed depression and/or suicidal behaviour at year 1 demonstrated consistently elevated rates of depression and suicidal behaviour 5 years after TBI. CONCLUSION: Compared to the general population, individuals with TBI are at greater risk for depression and suicidal behaviour many years after TBI. The significant psychiatric symptoms evidenced by individuals with TBI highlight the need for routine screening and mental health treatment in this population.

College students with depressive symptoms with and without fatigue: Differences in functioning, suicidality, anxiety, and depressive severity.

BACKGROUND: We examined whether fatigue was associated with greater symptomatic burden and functional impairment in college students with depressive symptoms. METHODS: Using data from the self-report Beck Depression Inventory (BDI), we stratified a group of 287 students endorsing significant symptoms of depression (BDI score ≥ 13) into 3 levels: no fatigue, mild fatigue, or moderate/severe fatigue. We then compared the 3 levels of fatigue across a battery of psychiatric and functional outcome measures. RESULTS: Approximately 87% of students endorsed at least mild fatigue. Students with moderate/severe fatigue had significantly greater depressive symptom severity compared with those with mild or no fatigue and scored higher on a suicide risk measure than those with mild fatigue. Students with severe fatigue evidenced greater frequency and intensity of anxiety than those with mild or no fatigue. Reported cognitive and functional impairment increased significantly as fatigue worsened. CONCLUSIONS: Depressed college students with symptoms of fatigue demonstrated functional impairment and symptomatic burden that worsened with increasing levels of fatigue. Assessing and treating symptoms of fatigue appears warranted within this population.

Treatment Outcome in Depressed Latinos Predicted by Concomitant Psychosislike Symptoms.

We compared treatment response (≥50 decrease in Nine-Item Patient Health Questionnaire total score) among 24 Latinos with major depressive disorder, presenting with and without specific psychosislike symptoms: A, hearing noises or house sounds, B, hearing voices calling one's name, C, seeing fleeting visions such as shadows, and D, symptoms more likely to be truly psychotic (e.g., poorly defined and short-lasting voices [other than B], fleeting paranoid ideation, or fleeting ideas of reference). 18 subjects (75%) endorsed symptoms of cluster A, 12 (50%) of cluster B, 10 (31%) of cluster C, and 12 (50%) of cluster D. Only subjects who reported symptoms from the D cluster exhibited significantly unfavorable depressive outcomes (compared to those with absence of D symptoms). The authors propose a phenomenological differentiation between benign psychosislike symptoms (clusters A-C) and the expression of the psychotic continuum (cluster D) in depressed Latinos.

College Students: Mental Health Problems and Treatment Considerations.

Attending college can be a stressful time for many students. In addition to coping with academic pressure, some students have to deal with the stressful tasks of separation and individuation from their family of origin while some may have to attend to numerous work and family responsibilities. In this context, many college students experience the first onset of mental health and substance use problems or an exacerbation of their symptoms. Given the uniqueness of college students, there is a need to outline critical issues to consider when working with this population. In this commentary, first, the prevalence of psychiatric and substance use problems in college students and the significance of assessing age of onset of current psychopathology are described. Then, the concerning persistent nature of mental health problems among college students and its implications are summarized. Finally, important aspects of treatment to consider when treating college students with mental health problems are outlined, such as the importance of including parents in the treatment, communicating with other providers, and employing of technology to increase adherence. It is concluded that, by becoming familiar with the unique problems characteristic of the developmental stage and environment college students are in, practitioners will be able to better serve them.

Reliability and validity of the Symptoms of Depression Questionnaire (SDQ).

Current measures for major depressive disorder focus primarily on the assessment of depressive symptoms, while often omitting other common features. However, the presence of comorbid features in the anxiety spectrum influences outcome and may effect treatment. More comprehensive measures of depression are needed that include the assessment of symptoms in the anxiety-depression spectrum. This study examines the reliability and validity of the Symptoms of Depression Questionnaire (SDQ), which assesses irritability, anger attacks, and anxiety symptoms together with the commonly considered symptoms of depression. Analysis of the factor structure of the SDQ identified 5 subscales, including one in the anxiety-depression spectrum, with adequate internal consistency and concurrent validity. The SDQ may be a valuable new tool to better characterize depression and identify and administer more targeted interventions.

Relationship between sleep disturbance and depression, anxiety, and functioning in college students.

BACKGROUND: Sleep disturbance (SD) has complex associations with depression, both preceding and following the onset and recurrence of depression. We hypothesized that students with depressive symptoms with SD would demonstrate a greater burden of comorbid psychiatric symptoms and functional impairment compared to students with depressive symptoms without SD. METHODS: During a mental health screening, 287 undergraduate students endorsed symptoms of depression (Beck Depression Inventory [BDI] ≥ 13) and filled out the following self-report measures: demographic questionnaire, BDI, Anxiety Symptom Questionnaire-intensity and frequency (ASQ), Beck Hopelessness Scale (BHS), Beck Anxiety Inventory (BAI), Quality of Life Enjoyment and Satisfaction Questionnaire (QLESQ), and the Massachusetts General Hospital Cognitive and Physical Functioning Questionnaire (CPFQ). SD was measured using the BDI sleep item #16 dichotomized (score 0: no SD; or score > 0: some SD). RESULTS: Students with depressive symptoms and SD (n = 220), compared to those without SD (n = 67), endorsed significantly more intense and frequent anxiety and poorer cognitive and physical functioning. Students with depressive symptoms with and without SD did not significantly differ in depressive severity, hopelessness, or quality of life. CONCLUSIONS: College students with depressive symptoms with SD may experience a greater burden of comorbid anxiety symptoms and hyperarousal, and may have impairments in functioning, compared to students with depressive symptoms without SD. These findings require replication.

Factors that distinguish college students with depressive symptoms with and without suicidal thoughts.

BACKGROUND: Suicide among college students is a significant public health concern. Although suicidality is linked to depression, not all depressed college students experience suicidal ideation (SI). The primary aim of this study was to determine potential factors that may distinguish college students with depressive symptoms with and without SI. METHODS: A total of 287 undergraduate college students with substantial depressive symptoms (Beck Depression Inventory [BDI] total score >13) with and without SI were compared across psychiatric and functional outcome variables. Independent sample t tests were conducted for each outcome variable using the suicide item of the BDI as a dichotomous (ie, zero vs nonzero score) grouping variable. RESULTS: Relative to students with substantial depressive symptoms without SI, those with SI were more symptomatic overall, having significantly higher levels of depressive symptoms, hopelessness, and anxiety. However, contrary to our expectations, nonsuicidal and suicidal students did not differ on measures of everyday functioning (ie, cognitive and physical functioning and grade point average). CONCLUSIONS: Our findings suggest that SI among college students is associated with increased subjective distress but may not adversely impact physical or cognitive functioning or academic performance.

Correlates of irritability in college students with depressive symptoms.

Depression is a prevalent psychiatric disorder associated with significant personal and societal burden. There is accumulating evidence for the presence of a subtype of depression characterized by the presence of irritability that is associated with increased morbidity, risk for suicidal ideation, and functional impairments in adults. Little is known about the features of depressive symptoms with and without irritability among young adults in college. The primary aim of this study was to characterize the presentation of college students with depressive symptoms and irritability. Two-hundred eighty-seven undergraduate college students with depressive symptoms with and without irritability were compared across several psychiatric and functional outcome variables. Independent samples t-tests or logistic regressions were conducted for each outcome variable using the irritability item of the Beck Depression Inventory as a dichotomous grouping variable. Analyses were conducted separately for the men and the women. Both male and female students with depressive symptoms and severe irritability reported a greater severity of depressive symptoms compared with their peers with no or mild irritability. In the women, the presence of irritability was associated with greater symptoms of anxiety, whereas in the men, it was associated with increased likelihood of engaging in risky behaviors, including compulsive use of alcohol, illicit drugs, and prescription drugs. The male and female college students with depressive symptoms with and without irritability did not differ on severity of suicidal ideation, hopelessness, or cognitive functioning. The findings from this study suggest that depressive symptoms and irritability may characterize a subtype of college students who have a greater symptom burden and with the potential need for more aggressive and prompt treatment.

Combined MI + CBT for Depressive Symptoms and Binge Drinking Among Young Adults: Two Case Studies.

There are high rates of comorbidity between heavy drinking and depressive symptoms among college students, often resulting in severe alcohol-related consequences. No empirically supported treatment exists that concurrently addresses both of these problems in this population. Research with college students has demonstrated that brief motivational interventions (BMIs) reduce heavy drinking and alcohol-related consequences, and that cognitive behavioral therapy for depression (CBT-D) is effective in reducing depressive symptoms. Thus, a program combining BMI and CBT-D appears ideal for college students with co-occurring binge drinking and depressive symptoms. This manuscript presents the rationale and format of a BMI + CBT-D treatment protocol for this population, and provides a case example of a female college student who received the protocol and experienced improvement in depressive symptoms, a reduction in alcohol use and alcohol-related negative consequences, and an increase in readiness to change alcohol consumption. We discuss theoretical and clinical implications of these findings, and suggest directions for future research.

A Randomized Controlled Trial of Community-Delivered Heated Hatha Yoga for Moderate-to-Severe Depression.

Objective: To evaluate feasibility, acceptability, and preliminary efficacy of heated yoga to treat moderate-to-severe depression.Design: An 8-week randomized controlled trial (RCT) of heated yoga versus waitlist control was conducted from March 2017 to August 2019.Methods: Participants in the yoga condition were asked to attend heated yoga classes at 2 community heated yoga studios at least twice weekly. We assessed acceptability and feasibility using exit interview and attendance data, respectively. The primary intervention efficacy outcome variable was change in the Inventory of Depressive Symptomatology-Clinician Rated (IDS-CR) score from baseline to post-intervention (week 8).Results: We randomized 80 participants and included 65 (mean [± SD] age 32.7 [± 11.7] years; 81.5% female) in the analyses (yoga n = 33, waitlist n = 32). The mean IDS-CR score at baseline was 35.6 (± 7.9) for the full sample, 36.9 (± 8.8) for yoga participants, and 34.4 (± 6.7) for waitlist participants. Participants attended an average of 10.3 (± 7.1) total classes over the 8-week intervention period. Yoga participants had a significantly greater pre- to post-intervention reduction in IDS-CR scores than waitlist participants (Cohen d = 1.04, P < .001). More yoga participants (59.3%; n = 16) than waitlist participants (6.3%; n = 2) evidenced larger treatment responses (IDS-CR ≥ 50% decrease in symptoms). Participants rated the heated yoga and its aftereffects positively in exit interviews.Conclusions: Approximately 1 heated yoga session per week (mean of 10.3 classes over 8 weeks) was associated with significantly greater reduction in depression symptoms than a waitlist control. Participants rated heated yoga positively. Taken together, results suggest feasibility, acceptability, and preliminary efficacy for patients with depression and warrant further research using active control conditions.Trial Registration: ClinicalTrials.gov identifier: NCT02607514.

A randomized, controlled, pilot study of acamprosate added to escitalopram in adults with major depressive disorder and alcohol use disorder.

We sought to examine the efficacy and safety of acamprosate augmentation of escitalopram in patients with concurrent major depressive disorder (MDD) and alcohol use disorders. Twenty-three adults (43% female; mean ± SD age, 46 ± 14 years) were enrolled and received 12 weeks of treatment with psychosocial support; escitalopram, 10 to 30 mg/d; and either acamprosate, 2000 mg/d (n = 12), or identical placebo (n = 11). Outcomes included change in clinician ratings of depressive symptoms, MDD response and remission rates, changes in frequency and intensity of alcohol use, retention rates, and adverse events. Twelve subjects (acamprosate, n = 7; placebo, n = 5) completed the study. There was significant mean reduction in ratings of depressive symptoms from baseline in both treatment arms (P < 0.05), with no significant difference between the groups. Those in the acamprosate group had a 50% MDD response rate and a 42% remission rate, whereas those in the placebo arm had a 36% response and remission rate (not significant). Those assigned to acamprosate had significant reduction in number of drinks per week and drinks per month during the trial, whereas those assigned to placebo demonstrated no significant change in any alcohol use parameter, but the between-group difference was not significant. There were no significant associations between change in depressive symptoms and change in alcohol use. Attrition rates did not differ significantly between the 2 arms. Acamprosate added to escitalopram in adults with MDD and alcohol use disorders was associated with reduction in the frequency of alcohol use. The present study was not powered to detect superiority versus placebo. Further study in a larger sample is warranted.

The Quick Inventory of Depressive Symptomatology, clinician rated and self-report: a psychometric assessment in Chinese Americans with major depressive disorder.

This study aimed to investigate the psychometric properties of the Chinese translations of the Quick Inventory of Depressive Symptomatology (QIDS(16)), including the Clinician-Rated (QIDS-C(16)), Self-report (QIDS-SR(16)), and Interactive Voice Response (QIDS-SR-IVR(16)) formats. Thirty depressed Chinese Americans were assessed with Chinese translations of the QIDS-SR(16), QIDS-SR-IVR(16), and QIDS-C(16). Cronbach alpha estimates of internal scale consistency on the QIDS-SR(16), QIDS-SR-IVR(16), and QIDS-C(16) were 0.70, 0.74, and 0.79, respectively. Intercorrelations among the measures were QIDS-SR(16) and QIDS-SR-IVR(16), r = 0.79; QIDS-SR(16) and QIDS-C(16), r = 0.61; and QIDS-SR-IVR(16) and QIDS-C(16), r = 0.69 (all p values < 0.01). The areas under the curve for the receiver operating characteristics of the QIDS-SR(16) and QIDS-SR-IVR(16) were 0.78 (95% confidence interval, 0.61-0.95) and 0.81 (95% confidence interval, 0.65-0.96), respectively. The respective screening sensitivities/specificities were 0.73/0.74 and 0.86/0.58. The Chinese translations of the QIDS(16) have adequate psychometric properties and may be useful tools for depression screening.

Depression and suicidal ideation in college students.

BACKGROUND/AIMS: Suicide is one of the leading causes of death in college students and is often associated with depression. The aim of this study was to assess the rates of suicidal ideation (SI) on college campuses and to identify its correlates. METHODS: On-campus depression screening sessions were conducted at 3 universities (n = 898; 55% female; mean age 20.07 ± 1.85 years). Participants completed the Beck Depression Inventory (BDI; mean ± SD of total score = 6.27 ± 6.31) and other measures. Eighty-four students endorsed a '1' on the BDI suicidality item, suggesting thoughts of suicide. RESULTS: Results showed that students with greater depression severity, higher levels of hopelessness, and poorer quality of life were more likely to endorse SI. CONCLUSION: Factors associated with SI highlighted in this study may aid in the identification of college students at risk for suicide.

Quantifying and addressing the impact of measurement error in network models.

Network psychometric models are often estimated using a single indicator for each node in the network, thus failing to consider potential measurement error. In this study, we investigate the impact of measurement error on cross-sectional network models. First, we conduct a simulation study to evaluate the performance of models based on single indicators as well as models that utilize information from multiple indicators per node, including average scores, factor scores, and latent variables. Our results demonstrate that measurement error impairs the reliability and performance of network models, especially when using single indicators. The reliability and performance of network models improves substantially with increasing sample size and when using methods that combine information from multiple indicators per node. Second, we use empirical data from the STAR*D trial (n = 3,731) to further evaluate the impact of measurement error. In the STAR*D trial, depression symptoms were assessed via three questionnaires, providing multiple indicators per symptom. Consistent with our simulation results, we find that when using sub-samples of this dataset, the discrepancy between the three single-indicator networks (one network per questionnaire) diminishes with increasing sample size. Together, our simulated and empirical findings provide evidence that measurement error can hinder network estimation when working with smaller samples and offers guidance on methods to mitigate measurement error.