The clinical effectiveness of using a predictive algorithm to guide antidepressant treatment in primary care (PReDicT): an open-label, randomised controlled trial

Using the 'Leapfrog' Design as a Simple Form of Adaptive Platform Trial to Develop, Test, and Implement Treatment Personalization Methods in Routine Practice

The route for the development, evaluation and dissemination of personalized psychological therapies is complex and challenging. In particular, the large sample sizes needed to provide adequately powered trials of newly-developed personalization approaches means that the traditional treatment development route is extremely inefficient. This paper outlines the promise of adaptive platform trials (APT) embedded within routine practice as a method to streamline development and testing of personalized psychological therapies, and close the gap to implementation in real-world settings. It focuses in particular on a recently-developed simplified APT design, the 'leapfrog' trial, illustrating via simulation how such a trial may proceed and the advantages it can bring, for example in terms of reduced sample sizes. Finally it discusses models of how such trials could be implemented in routine practice, including potential challenges and caveats, alongside a longer-term perspective on the development of personalized psychological treatments.

When a test is more than just a test: Findings from patient interviews and survey in the trial of a technology to measure antidepressant medication response (the PReDicT Trial)

BACKGROUND

A RCT of a novel intervention to detect antidepressant medication response (the PReDicT Test) took place in five European countries, accompanied by a nested study of its acceptability and implementation presented here. The RCT results indicated no effect of the intervention on depression at 8 weeks (primary outcome), although effects on anxiety at 8 weeks and functioning at 24 weeks were found.

METHODS

The nested study used mixed methods. The aim was to explore patient experiences of the Test including acceptability and implementation, to inform its use within care. A bespoke survey was completed by trial participants in five countries (n = 778) at week 8. Semi-structured interviews were carried out in two countries soon after week 8 (UK n = 22, Germany n = 20). Quantitative data was analysed descriptively; for qualitative data, thematic analysis was carried out using a framework approach. Results of the two datasets were interrogated together.

OUTCOMES

Survey results showed the intervention was well received, with a majority of participants indicating they would use it again, and it gave them helpful extra information; a small minority indicated the Test made them feel worse. Qualitative data showed the Test had unexpected properties, including: instigating a process of reflection, giving participants feedback on progress and new understanding about their illness, and making participants feel supported and more engaged in treatment.

INTERPRETATION

The qualitative and quantitative results are generally consistent. The Test's unexpected properties may explain why the RCT showed little effect, as properties were experienced across both trial arms. Beyond the RCT, the qualitative data sheds light on measurement reactivity, i.e., how measurements of depression can impact patients.

Depression follow-up monitoring with the PHQ-9: an open cluster-randomised controlled trial

BACKGROUND

Outcome monitoring of depression treatment is recommended but there is a lack of evidence on patient benefit in primary care.

AIM

To test monitoring depression using the Patient Health Questionnaire (PHQ-9) with patient feedback.

DESIGN AND SETTING

An open cluster-randomised controlled trial was undertaken in 141 group practices.

METHOD

Adults with new depressive episodes were recruited through record searches and opportunistically. The exclusion criteria were as follows: dementia; psychosis; substance misuse; and suicide risk. The PHQ-9 was administered soon after diagnosis, and 10-35 days later. The primary outcome was the Beck Depression Inventory (BDI-II) score at 12 weeks. The secondary outcomes were as follows: BDI-II at 26 weeks; Work and Social Adjustment Scale (WSAS) and EuroQol EQ-5D-5L quality of life at 12 and 26 weeks; antidepressant treatment; mental health and social service contacts; adverse events, and Medical Interview Satisfaction Scale (MISS) over 26 weeks.

RESULTS

In total, 302 patients were recruited to the intervention arm and 227 to the controls. At 12 weeks, 254 (84.1%) and 199 (87.7%) were followed-up, respectively. Only 40.9% of patients in the intervention had a GP follow-up PHQ-9 recorded. There was no significant difference in BDI-II score at 12 weeks (mean difference -0.46; 95% confidence interval [CI] = -2.16 to 1.26; adjusted for baseline depression, baseline anxiety, sociodemographic factors, and clustering by practice). EQ-5D-5L quality-of-life scores were higher in the intervention arm at 26 weeks (adjusted mean difference 0.053; 95% CI = 0.013 to 0.093. A clinically significant difference in depression at 26 weeks could not be ruled out. No significant differences were found in social functioning, adverse events, or satisfaction. In a per-protocol analysis, antidepressant use and mental health contacts were significantly greater in patients in the intervention arm with a recorded follow-up PHQ-9 (P = 0.025 and P = 0.010, respectively).

CONCLUSION

No evidence was found of improved depression outcome at 12 weeks from monitoring. The findings of possible benefits over 26 weeks warrant replication, investigating possible mechanisms, preferably with automated delivery of monitoring and more instructive feedback.

Acute neural effects of the mood stabiliser lamotrigine on emotional processing in healthy volunteers: a randomised control trial

Lamotrigine is an effective mood stabiliser, largely used for the management and prevention of depression in bipolar disorder. The neuropsychological mechanisms by which lamotrigine acts to relieve symptoms as well as its neural effects on emotional processing remain unclear. The primary objective of this current study was to investigate the impact of an acute dose of lamotrigine on the neural response to a well-characterised fMRI task probing implicit emotional processing relevant to negative bias. 31 healthy participants were administered either a single dose of lamotrigine (300 mg, n = 14) or placebo (n = 17) in a randomized, double-blind design. Inside the 3 T MRI scanner, participants completed a covert emotional faces gender discrimination task. Brain activations showing significant group differences were identified using voxel-wise general linear model (GLM) nonparametric permutation testing, with threshold free cluster enhancement (TFCE) and a family wise error (FWE)-corrected cluster significance threshold of p < 0.05. Participants receiving lamotrigine were more accurate at identifying the gender of fearful (but not happy or angry) faces. A network of regions associated with emotional processing, including amygdala, insula, and the anterior cingulate cortex (ACC), was significantly less activated in the lamotrigine group compared to the placebo group across emotional facial expressions. A single dose of lamotrigine reduced activation in limbic areas in response to faces with both positive and negative expressions, suggesting a valence-independent effect. However, at a behavioural level lamotrigine appeared to reduce the distracting effect of fear on face discrimination. Such effects may be relevant to the mood stabilisation effects of lamotrigine.

Subjective experience of the environment determines serotoninergic antidepressant treatment outcome in male mice

BACKGROUND

The effects of the selective serotonin reuptake inhibitors (SSRIs), the first-line antidepressant treatment, have been proposed to be affected, at least in part, by the living environment. Since the quality of the environment depends not only on its objective features, but also on the subjective experience, we hypothesized that the latter plays a key role in determining SSRI treatment outcome.

METHODS

We chronically administered the SSRI fluoxetine to two groups of adult CD-1 male mice that reportedly show distinct subjective experiences of the environment measured as consistent and significantly different responses to the same emotional and social stimuli. These distinct socioemotional profiles were generated by rearing mice either in standard laboratory conditions (SN) or in a communal nest (CN) where three dams breed together their offspring, sharing caregiving behavior.

RESULTS

At adulthood, CN mice displayed higher levels of agonistic and anxiety-like behaviors than SN mice, indicating that they experience the environment as more socially challenging and potentially dangerous. We then administered fluoxetine, which increased offensive and anxious response in SN, while producing opposite effects in CN mice. BDNF regulation was modified by the treatment accordingly.

LIMITATIONS

Subjective experience in mice was assessed as behavioral response to the environment.

CONCLUSIONS

These results show that the subjective experience of the environment determines fluoxetine outcome. In a translational perspective, our findings suggest considering not only the objective quality, but also the subjective appraisal, of the patient's living environment for developing effective personalized therapeutic approaches in psychiatry.

Acceptability and utility of digital well-being and mental health support for university students: A pilot study

AIM

To assess the acceptability and explore the utility of a novel digital platform designed as a student-facing well-being and mental health support.

METHODS

An adapted version of i-spero® was piloted as a student-facing well-being support and as part of routine university-based mental health care. In both pathways, student participants completed baseline demographics and brief validated measures of well-being and mental health. Weekly measures of anxiety (GAD-7) and depression (PHQ-9) and a Week 8 Experience Survey were also scheduled. Integrated mixed methods analysis was used to assess acceptability and explore the utility of these platforms.

RESULTS

Students in the well-being (n = 120) and care pathways (n = 121) were mostly female and between 19 and 22 years of age. Baseline screen positive rates for anxiety and depression were high in both the well-being (68%) and care pathways (80%). There was a substantial drop in adherence over Week 1 (50% well-being; 40% care) followed by minor attrition up to Week 8. Anxiety and depressive symptom levels improved from baseline in students who dropped out after Week 1 (p ≤ .06). The student experience was that i-spero® improved their emotional self-awareness, understanding of progress in care, and knowledge about when to seek help. Most students agreed (>75%) that i-spero® should form part of regular university student wellness support.

CONCLUSIONS

Digital well-being and mental health support seems acceptable to university students; however, engagement and persistence are areas for further development. Such digital tools could make a positive contribution to an evidence-based stepped approach to student well-being and mental health support.

Patient-reported outcome measures for monitoring primary care patients with depression: the PROMDEP cluster RCT and economic evaluation

Background

Guidelines on the management of depression recommend that practitioners use patient-reported outcome measures for the follow-up monitoring of symptoms, but there is a lack of evidence of benefit in terms of patient outcomes.

Objective

To test using the Patient Health Questionnaire-9 questionnaire as a patient-reported outcome measure for monitoring depression, training practitioners in interpreting scores and giving patients feedback.

Design

Parallel-group, cluster-randomised superiority trial; 1 : 1 allocation to intervention and control.

Setting

UK primary care (141 group general practices in England and Wales).

Inclusion criteria

Patients aged ≥ 18 years with a new episode of depressive disorder or symptoms, recruited mainly through medical record searches, plus opportunistically in consultations.

Exclusions

Current depression treatment, dementia, psychosis, substance misuse and risk of suicide.

Intervention

Administration of the Patient Health Questionnaire-9 questionnaire with patient feedback soon after diagnosis, and at follow-up 10-35 days later, compared with usual care.

Primary outcome

Beck Depression Inventory, 2nd edition, symptom scores at 12 weeks.

Secondary outcomes

Beck Depression Inventory, 2nd edition, scores at 26 weeks; antidepressant drug treatment and mental health service contacts; social functioning (Work and Social Adjustment Scale) and quality of life (EuroQol 5-Dimension, five-level) at 12 and 26 weeks; service use over 26 weeks to calculate NHS costs; patient satisfaction at 26 weeks (Medical Informant Satisfaction Scale); and adverse events.

Sample size

The original target sample of 676 patients recruited was reduced to 554 due to finding a significant correlation between baseline and follow-up values for the primary outcome measure.

Randomisation

Remote computerised randomisation with minimisation by recruiting university, small/large practice and urban/rural location.

Blinding

Blinding of participants was impossible given the open cluster design, but self-report outcome measures prevented observer bias. Analysis was blind to allocation.

Analysis

Linear mixed models were used, adjusted for baseline depression, baseline anxiety, sociodemographic factors, and clustering including practice as random effect. Quality of life and costs were analysed over 26 weeks.

Qualitative interviews

Practitioner and patient interviews were conducted to reflect on trial processes and use of the Patient Health Questionnaire-9 using the Normalization Process Theory framework.

Results

Three hundred and two patients were recruited in intervention arm practices and 227 patients were recruited in control practices. Primary outcome data were collected for 252 (83.4%) and 195 (85.9%), respectively. No significant difference in Beck Depression Inventory, 2nd edition, score was found at 12 weeks (adjusted mean difference -0.46, 95% confidence interval -2.16 to 1.26). Nor were significant differences found in Beck Depression Inventory, 2nd Edition, score at 26 weeks, social functioning, patient satisfaction or adverse events. EuroQol-5 Dimensions, five-level version, quality-of-life scores favoured the intervention arm at 26 weeks (adjusted mean difference 0.053, 95% confidence interval 0.013 to 0.093). However, quality-adjusted life-years over 26 weeks were not significantly greater (difference 0.0013, 95% confidence interval -0.0157 to 0.0182). Costs were lower in the intervention arm but, again, not significantly (-£163, 95% confidence interval -£349 to £28). Cost-effectiveness and cost-utility analyses, therefore, suggested that the intervention was dominant over usual care, but with considerable uncertainty around the point estimates. Patients valued using the Patient Health Questionnaire-9 to compare scores at baseline and follow-up, whereas practitioner views were more mixed, with some considering it too time-consuming.

Conclusions

We found no evidence of improved depression management or outcome at 12 weeks from using the Patient Health Questionnaire-9, but patients' quality of life was better at 26 weeks, perhaps because feedback of Patient Health Questionnaire-9 scores increased their awareness of improvement in their depression and reduced their anxiety. Further research in primary care should evaluate patient-reported outcome measures including anxiety symptoms, administered remotely, with algorithms delivering clear recommendations for changes in treatment.

Study registration

This study is registered as IRAS250225 and ISRCTN17299295.

Funding

This award was funded by the National Institute for Health and Care Research (NIHR) Health Technology Assessment programme (NIHR award ref: 17/42/02) and is published in full in Health Technology Assessment; Vol. 28, No. 17. See the NIHR Funding and Awards website for further award information.

Digital technology and mental health during the COVID-19 pandemic: a narrative review with a focus on depression, anxiety, stress, and trauma

The sudden appearance and devastating effects of the COVID-19 pandemic resulted in the need for multiple adaptive changes in societies, business operations and healthcare systems across the world. This review describes the development and increased use of digital technologies such as chat bots, electronic diaries, online questionnaires and even video gameplay to maintain effective treatment standards for individuals with mental health conditions such as depression, anxiety and post-traumatic stress syndrome. We describe how these approaches have been applied to help meet the challenges of the pandemic in delivering mental healthcare solutions. The main focus of this narrative review is on describing how these digital platforms have been used in diagnostics, patient monitoring and as a treatment option for the general public, as well as for frontline medical staff suffering with mental health issues.

Prediction of depression treatment outcome from multimodal data: a CAN-BIND-1 report

BACKGROUND

Prediction of treatment outcomes is a key step in improving the treatment of major depressive disorder (MDD). The Canadian Biomarker Integration Network in Depression (CAN-BIND) aims to predict antidepressant treatment outcomes through analyses of clinical assessment, neuroimaging, and blood biomarkers.

METHODS

In the CAN-BIND-1 dataset of 192 adults with MDD and outcomes of treatment with escitalopram, we applied machine learning models in a nested cross-validation framework. Across 210 analyses, we examined combinations of predictive variables from three modalities, measured at baseline and after 2 weeks of treatment, and five machine learning methods with and without feature selection. To optimize the predictors-to-observations ratio, we followed a tiered approach with 134 and 1152 variables in tier 1 and tier 2 respectively.

RESULTS

A combination of baseline tier 1 clinical, neuroimaging, and molecular variables predicted response with a mean balanced accuracy of 0.57 (best model mean 0.62) compared to 0.54 (best model mean 0.61) in single modality models. Adding week 2 predictors improved the prediction of response to a mean balanced accuracy of 0.59 (best model mean 0.66). Adding tier 2 features did not improve prediction.

CONCLUSIONS

A combination of clinical, neuroimaging, and molecular data improves the prediction of treatment outcomes over single modality measurement. The addition of measurements from the early stages of treatment adds precision. Present results are limited by lack of external validation. To achieve clinically meaningful prediction, the multimodal measurement should be scaled up to larger samples and the robustness of prediction tested in an external validation dataset.

Identifying Transdiagnostic Mechanisms in Mental Health Using Computational Factor Modeling

Most psychiatric disorders do not occur in isolation, and most psychiatric symptom dimensions are not uniquely expressed within a single diagnostic category. Current treatments fail to work for around 25% to 40% of individuals, perhaps due at least in part to an overreliance on diagnostic categories in treatment development and allocation. In this review, we describe ongoing efforts in the field to surmount these challenges and precisely characterize psychiatric symptom dimensions using large-scale studies of unselected samples via remote, online, and "citizen science" efforts that take a dimensional, mechanistic approach. We discuss the importance that efforts to identify meaningful psychiatric dimensions be coupled with careful computational modeling to formally specify, test, and potentially falsify candidate mechanisms that underlie transdiagnostic symptom dimensions. We refer to this approach, i.e., where symptom dimensions are identified and validated against computationally well-defined neurocognitive processes, as computational factor modeling. We describe in detail some recent applications of this method to understand transdiagnostic cognitive processes that include model-based planning, metacognition, appetitive processing, and uncertainty estimation. In this context, we highlight how computational factor modeling has been used to identify specific associations between cognition and symptom dimensions and reveal previously obscured relationships, how findings generalize to smaller in-person clinical and nonclinical samples, and how the method is being adapted and optimized beyond its original instantiation. Crucially, we discuss next steps for this area of research, highlighting the value of more direct investigations of treatment response that bridge the gap between basic research and the clinic.

Multimodal predictions of treatment outcome in major depression: A comparison of data-driven predictors with importance ratings by clinicians

BACKGROUND

Reliable prediction models of treatment outcome in Major Depressive Disorder (MDD) are currently lacking in clinical practice. Data-driven outcome definitions, combining data from multiple modalities and incorporating clinician expertise might improve predictions.

METHODS

We used unsupervised machine learning to identify treatment outcome classes in 1060 MDD inpatients. Subsequently, classification models were created on clinical and biological baseline information to predict treatment outcome classes and compared to the performance of two widely used classical outcome definitions. We also related the findings to results from an online survey that assessed which information clinicians use for outcome prognosis.

RESULTS

Three and four outcome classes were identified by unsupervised learning. However, data-driven outcome classes did not result in more accurate prediction models. The best prediction model was targeting treatment response in its standard definition and reached accuracies of 63.9 % in the test sample, and 59.5 % and 56.9 % in the validation samples. Top predictors included sociodemographic and clinical characteristics, while biological parameters did not improve prediction accuracies. Treatment history, personality factors, prior course of the disorder, and patient attitude towards treatment were ranked as most important indicators by clinicians.

LIMITATIONS

Missing data limited the power to identify biological predictors of treatment outcome from certain modalities.

CONCLUSIONS

So far, the inclusion of available biological measures in addition to psychometric and clinical information did not improve predictive value of the models, which was overall low. Optimized biomarkers, stratified predictions and the inclusion of clinical expertise may improve future prediction models.

Development of a model to predict combined antidepressant medication and psychotherapy treatment response for depression among veterans

BACKGROUND

Although research shows that more depressed patients respond to combined antidepressants (ADM) and psychotherapy than either alone, many patients do not respond even to combined treatment. A reliable prediction model for this could help treatment decision-making. We attempted to create such a model using machine learning methods among patients in the US Veterans Health Administration (VHA).

METHODS

A 2018-2020 national sample of VHA patients beginning combined depression treatment completed self-report assessments at baseline and 3 months (n = 658). A learning model was developed using baseline self-report, administrative, and geospatial data to predict 3-month treatment response defined by reductions in the Quick Inventory of Depression Symptomatology Self-Report and/or in the Sheehan Disability Scale. The model was developed in a 70 % training sample and tested in the remaining 30 % test sample.

RESULTS

30.0 % of patients responded to treatment. The prediction model had a test sample AUC-ROC of 0.657. A strong gradient was found in probability of treatment response from 52.7 % in the highest predicted quintile to 14.4 % in the lowest predicted quintile. The most important predictors were episode characteristics (symptoms, comorbidities, history), personality/psychological resilience, recent stressors, and treatment characteristics.

LIMITATIONS

Restrictions in sample definition, a low recruitment rate, and reliance on patient self-report rather than clinician assessments to determine treatment response limited the generalizability of results.

CONCLUSIONS

A machine learning model could help depressed patients and providers predict likely response to combined ADM-psychotherapy. Parallel information about potential harms and costs of alternative treatments would be needed, though, to inform optimal treatment selection.

Deep phenotyping towards precision psychiatry of first-episode depression - the Brain Drugs-Depression cohort

BACKGROUND

Major Depressive Disorder (MDD) is a heterogenous brain disorder, with potentially multiple psychosocial and biological disease mechanisms. This is also a plausible explanation for why patients do not respond equally well to treatment with first- or second-line antidepressants, i.e., one-third to one-half of patients do not remit in response to first- or second-line treatment. To map MDD heterogeneity and markers of treatment response to enable a precision medicine approach, we will acquire several possible predictive markers across several domains, e.g., psychosocial, biochemical, and neuroimaging.

METHODS

All patients are examined before receiving a standardised treatment package for adults aged 18-65 with first-episode depression in six public outpatient clinics in the Capital Region of Denmark. From this population, we will recruit a cohort of 800 patients for whom we will acquire clinical, cognitive, psychometric, and biological data. A subgroup (subcohort I, n = 600) will additionally provide neuroimaging data, i.e., Magnetic Resonance Imaging, and Electroencephalogram, and a subgroup of patients from subcohort I unmedicated at inclusion (subcohort II, n = 60) will also undergo a brain Positron Emission Tomography with the [¹¹C]-UCB-J tracer binding to the presynaptic glycoprotein-SV2A. Subcohort allocation is based on eligibility and willingness to participate. The treatment package typically lasts six months. Depression severity is assessed with the Quick Inventory of Depressive Symptomatology (QIDS) at baseline, and 6, 12 and 18 months after treatment initiation. The primary outcome is remission (QIDS ≤ 5) and clinical improvement (≥ 50% reduction in QIDS) after 6 months. Secondary endpoints include remission at 12 and 18 months and %-change in QIDS, 10-item Symptom Checklist, 5-item WHO Well-Being Index, and modified Disability Scale from baseline through follow-up. We also assess psychotherapy and medication side-effects. We will use machine learning to determine a combination of characteristics that best predict treatment outcomes and statistical models to investigate the association between individual measures and clinical outcomes. We will assess associations between patient characteristics, treatment choices, and clinical outcomes using path analysis, enabling us to estimate the effect of treatment choices and timing on the clinical outcome.

DISCUSSION

The BrainDrugs-Depression study is a real-world deep-phenotyping clinical cohort study of first-episode MDD patients.

TRIAL REGISTRATION

Registered at clinicaltrials.gov November 15th, 2022 (NCT05616559).

Antidepressant Advisor (ADeSS): a decision support system for antidepressant treatment for depression in UK primary care - a feasibility study

OBJECTIVES

To develop and probe the first computerised decision-support tool to provide antidepressant treatment guidance to general practitioners (GPs) in UK primary care.

DESIGN

A parallel group, cluster-randomised controlled feasibility trial, where individual participants were blind to treatment allocation.

SETTING

South London NHS GP practices.

PARTICIPANTS

Ten practices and eighteen patients with treatment-resistant current major depressive disorder.

INTERVENTIONS

Practices were randomised to two treatment arms: (a) treatment-as-usual, (b) computerised decision support tool.

RESULTS

Ten GP practices participated in the trial, which was within our target range (8-20). However, practice and patient recruitment were slower than anticipated and only 18 of 86 intended patients were recruited. This was due to fewer than expected patients being eligible for the study, as well as disruption resulting from the COVID-19 pandemic. Only one patient was lost to follow-up. There were no serious or medically important adverse events during the trial. GPs in the decision tool arm indicated moderate support for the tool. A minority of patients fully engaged with the mobile app-based tracking of symptoms, medication adherence and side effects.

CONCLUSIONS

Overall, feasibility was not shown in the current study and the following modifications would be needed to attempt to overcome the limitations found: (a) inclusion of patients who have only tried one Selective Serotonin Reuptake Inhibitor, rather than two, to improve recruitment and pragmatic relevance of the study; (b) approaching community pharmacists to implement tool recommendations rather than GPs; (c) further funding to directly interface between the decision support tool and self-reported symptom app; (d) increasing the geographic reach by not requiring detailed diagnostic assessments and replacing this with supported remote self-report.

TRIAL REGISTRATION NUMBER

NCT03628027.

Novel and emerging treatments for major depression

Depression is common, costly, debilitating, and associated with increased risk of suicide. It is one of the leading global public health problems. Although existing available pharmacological treatments can be effective, their onset of action can take up to 6 weeks, side-effects are common, and recovery can require treatment with multiple different agents. Although psychosocial interventions might also be recommended, more effective treatments than those currently available are needed for people with moderate or severe depression. In the past 10 years, treatment trials have developed and tested many new targeted interventions. In this Review, we assess novel and emerging biological treatments for major depressive disorder, evaluate their putative brain and body mechanisms, and highlight how close each might be to clinical use.

Predicting treatment outcome in depression: an introduction into current concepts and challenges

Improving response and remission rates in major depressive disorder (MDD) remains an important challenge. Matching patients to the treatment they will most likely respond to should be the ultimate goal. Even though numerous studies have investigated patient-specific indicators of treatment efficacy, no (bio)markers or empirical tests for use in clinical practice have resulted as of now. Therefore, clinical decisions regarding the treatment of MDD still have to be made on the basis of questionnaire- or interview-based assessments and general guidelines without the support of a (laboratory) test. We conducted a narrative review of current approaches to characterize and predict outcome to pharmacological treatments in MDD. We particularly focused on findings from newer computational studies using machine learning and on the resulting implementation into clinical decision support systems. The main issues seem to rest upon the unavailability of robust predictive variables and the lacking application of empirical findings and predictive models in clinical practice. We outline several challenges that need to be tackled on different stages of the translational process, from current concepts and definitions to generalizable prediction models and their successful implementation into digital support systems. By bridging the addressed gaps in translational psychiatric research, advances in data quantity and new technologies may enable the next steps toward precision psychiatry.

Psychological Distress Impact of Coronavirus Disease (COVID-19) Outbreak on Three Continents: A Systematic Review and Meta-analysis

BACKGROUND

The dire state of coronavirus disease (COVID-19) outbreak has had a substantial psychological impact on society.

METHODS

A systematic search was performed through Medline, PubMed, Embase, Scopus, and Web of Science, to investigate the impact of the COVID-19 pandemic on the psychological health of individuals in various countries. Subgroup analyses considered gender and classification of countries into three continents of America, Europe, and Asia. Only studies that used the COVID-19 Peritraumatic Distress Index (CPDI) questionnaire as a tool to assess mental distress were included in this meta-analysis. Heterogeneity among studies was assessed by I² statistic, and the random-effects model was utilized to obtain the pooled prevalence.

RESULTS

This pooled analysis included a large data sample of 21 studies consisting of 94,414 participants. The pooled prevalence of the psychological distress during the time of COVID-19 pandemic by CPDI for the continent of Asia was 43% (34.6% mild-to-moderate and 8.4% severe) which was greater than that for Europe (35%; 30% mild-to-moderate and 5% severe) but lower than that for America (64.3%; 45.8% mild to moderate and 18.5% severe). In addition, the prevalence of psychological distress according to CPDI was higher in females (48%; 40% mild to moderate, 13% severe) compared with males (59%; 36% mild to moderate and 5% severe).

CONCLUSIONS

Our findings suggest that psychological distress in the Americas is a larger problem than in Asia and European continents. Females appear to be more vulnerable and may therefore require further attention in terms of preventive and management strategies. Implementation of both digital and molecular biomarkers is encouraged to increase objectivity and accuracy of assessing the dynamic changes in mental health in the current and future pandemics.

Feasibility, acceptability and costs of nurse-led Alpha-Stim cranial electrostimulation to treat anxiety and depression in university students

Background

Only a relatively low proportion of university students seek help for anxiety and depression disorders, partly because they dislike current drug and psychological treatment options and would prefer home-based care. The aim of this study is to determine the feasibility, acceptability and cost utility of Alpha-Stim cranial electrostimulation (CES) delivered through a nurse led primary care clinic as a daily treatment for anxiety and depression symptoms by the student at home in contrast to usual primary care.

Method

Feasibility and acceptability of a nurse led clinic offering Alpha-Stim CES in terms of the take up and completion of the six-week course of Alpha-Stim CES. Change in score on the GAD-7 and PHQ-9 as measures of anxiety and depression symptoms at baseline and at 8 weeks following a course of Alpha-Stim CES. Similar evaluation in a non-randomised control group attending a family doctor over the same period. Cost-utility analysis of the nurse led Alpha-Stim CES and family doctor pathways with participants failing to improve following further NICE Guideline clinical care (facilitated self-help and cognitive behaviour therapy).

Results

Of 47 students (mean age 22.1, years, 79% female opting for Alpha-Stim CES at the nurse-led clinic 46 (97.9%) completed a 6-week daily course. Forty-seven (47) students comprised a comparison group receiving usual family doctor care. Both Alpha-Stim CES and usual family doctor care were associated with large effect size reductions in GAD-7 and PHQ-9 scores from baseline to 8 weeks. There were no adverse effects and only one participant showed a clinically important deterioration in the Alpha-Stim group. In the cost utility analysis, Alpha-Stim CES was a cheaper option than usual family doctor care under all deterministic or probabilistic assumptions.

Conclusion

Nurse delivered Alpha-Stim CES may be a feasible, acceptable and cheaper way of providing greater choice and home-based care for some university students seeking help from primary care with new presentations of anxiety and depression.

Toward Population Health: Using a Learning Behavioral Health System and Measurement-Based Care to Improve Access, Care, Outcomes, and Disparities

Achieving population behavioral health is urgently needed. The mental health system struggles with enormous challenges of providing access to mental health services, improving quality and equitability of care, and ensuring good health outcomes across subpopulations. Little data exists about increasing access within highly constrained resources, staging/sequencing treatment along care pathways, or personalizing treatments. The conceptual model of the learning healthcare system offers a potential paradigm shift for addressing these challenges. In this article we present an overview of how the three constructs of population health, learning health systems, and measurement-based care are inter-related, and we provide an example of how one academic, community-based, safety net health system is approaching integrating these paradigms into its service delivery system. Implementation outcomes will be described in a subsequent publication. We close by discussing how ultimately, to meaningfully improve population behavioral health, a learning healthcare system could expand into a learning health community in order to target critical points of prevention and intervention.

Randomized Controlled Trials of Artificial Intelligence in Clinical Practice: Systematic Review

BACKGROUND

The number of artificial intelligence (AI) studies in medicine has exponentially increased recently. However, there is no clear quantification of the clinical benefits of implementing AI-assisted tools in patient care.

OBJECTIVE

This study aims to systematically review all published randomized controlled trials (RCTs) of AI-assisted tools to characterize their performance in clinical practice.

METHODS

CINAHL, Cochrane Central, Embase, MEDLINE, and PubMed were searched to identify relevant RCTs published up to July 2021 and comparing the performance of AI-assisted tools with conventional clinical management without AI assistance. We evaluated the primary end points of each study to determine their clinical relevance. This systematic review was conducted following the updated PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 guidelines.

RESULTS

Among the 11,839 articles retrieved, only 39 (0.33%) RCTs were included. These RCTs were conducted in an approximately equal distribution from North America, Europe, and Asia. AI-assisted tools were implemented in 13 different clinical specialties. Most RCTs were published in the field of gastroenterology, with 15 studies on AI-assisted endoscopy. Most RCTs studied biosignal-based AI-assisted tools, and a minority of RCTs studied AI-assisted tools drawn from clinical data. In 77% (30/39) of the RCTs, AI-assisted interventions outperformed usual clinical care, and clinically relevant outcomes improved with AI-assisted intervention in 70% (21/30) of the studies. Small sample size and single-center design limited the generalizability of these studies.

CONCLUSIONS

There is growing evidence supporting the implementation of AI-assisted tools in daily clinical practice; however, the number of available RCTs is limited and heterogeneous. More RCTs of AI-assisted tools integrated into clinical practice are needed to advance the role of AI in medicine.

TRIAL REGISTRATION

PROSPERO CRD42021286539; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=286539.

Relationship between change in social evaluation learning and mood in early antidepressant treatment: A prospective cohort study in primary care

BACKGROUND

Antidepressants are proposed to work by increasing sensitivity to positive versus negative information. Increasing positive affective learning within social contexts may help remediate negative self-schema. We investigated the association between change in biased learning of social evaluations about the self and others, and mood during early antidepressant treatment.

METHOD

Prospective cohort assessing patients recruited from primary care in South West England at four timepoints over the first 8 weeks of antidepressant treatment (n = 29). At each timepoint, participants completed self-report measures of depression (Beck Depression Inventory II (BDI-II) and Patient Health Questionnaire 9 (PHQ-9)), anxiety (Generalised Anxiety Disorder Questionnaire 7 (GAD-7)), and a computerised task measuring learning of social evaluations about the self, a friend and a stranger.

RESULTS

We did not find evidence that learning about the self was associated with a reduction in PHQ-9 (b = 0.08, 95% CI: -0.05, 0.20, p = 0.239) or BDI-II scores (b = 0.10, 95% CI: -0.18, 0.38, p = 0.469). We found some weak evidence that increased positive learning about the friend was associated with a reduction in BDI-II scores (b = 0.30, 95% CI: -0.02, 0.62, p = 0.069). However, exploratory analyses indicated stronger evidence that increased positive learning about the self (b = 0.18, 95% CI: 0.07, 0.28, p = 0.002) and a friend (b = 0.22, 95% CI: 0.10, 0.35, p = 0.001) was associated with reductions in anxiety.

CONCLUSIONS

Change in social evaluation learning was associated with a reduction in anxiety but not depression. Antidepressants may treat anxiety symptoms by remediating negative affective biases towards socially threatening information directed towards the self and close others. However, our findings are based on exploratory analyses within a small sample without a control group and are therefore at risk of type 1 errors and order effects. Further research with larger samples is required.

Real-World Implementation of Precision Psychiatry: A Systematic Review of Barriers and Facilitators

Background: Despite significant research progress surrounding precision medicine in psychiatry, there has been little tangible impact upon real-world clinical care. Objective: To identify barriers and facilitators affecting the real-world implementation of precision psychiatry. Method: A PRISMA-compliant systematic literature search of primary research studies, conducted in the Web of Science, Cochrane Central Register of Controlled Trials, PsycINFO and OpenGrey databases. We included a qualitative data synthesis structured according to the ‘Consolidated Framework for Implementation Research’ (CFIR) key constructs. Results: Of 93,886 records screened, 28 studies were suitable for inclusion. The included studies reported 38 barriers and facilitators attributed to the CFIR constructs. Commonly reported barriers included: potential psychological harm to the service user (n = 11), cost and time investments (n = 9), potential economic and occupational harm to the service user (n = 8), poor accuracy and utility of the model (n = 8), and poor perceived competence in precision medicine amongst staff (n = 7). The most highly reported facilitator was the availability of adequate competence and skills training for staff (n = 7). Conclusions: Psychiatry faces widespread challenges in the implementation of precision medicine methods. Innovative solutions are required at the level of the individual and the wider system to fulfil the translational gap and impact real-world care.

Temporal dynamics of affect in the brain: Evidence from human imaging and animal models

Emotions are time-varying internal states that promote survival in the face of dynamic environments and shifting homeostatic needs. Research in non-human organisms has recently afforded specific insights into the neural mechanisms that support the emergence, persistence, and decay of affective states. Concurrently, a separate affective neuroscience literature has begun to dissect the neural bases of affective dynamics in humans. However, the circuit-level mechanisms identified in animals lack a clear mapping to the human neuroscience literature. As a result, critical questions pertaining to the neural bases of affective dynamics in humans remain unanswered. To address these shortcomings, the present review integrates findings from humans and non-human organisms to highlight the neural mechanisms that govern the temporal features of emotional states. Using the theory of affective chronometry as an organizing framework, we describe the specific neural mechanisms and modulatory factors that arbitrate the rise-time, intensity, and duration of emotional states.

Validation of the P1vital® Faces Set for Use as Stimuli in Tests of Facial Emotion Recognition

BACKGROUND

Negative bias in facial emotion recognition is a well-established concept in mental disorders such as depression. However, existing face sets of emotion recognition tests may be of limited use in international research, which could benefit from more contemporary and diverse alternatives. Here, we developed and provide initial validation for the P1vital® Affective Faces set (PAFs) as a contemporary alternative to the widely-used Pictures of Facial Affect (PoFA).

METHODS

The PAFs was constructed of 133 color photographs of facial expressions of ethnically-diverse trained actors and compared with the PoFA, comprised of 110 black and white photographs of facial expressions of generally Caucasian actors. Sixty-one recruits were asked to classify faces from both sets over six emotions (happy, sad, fear, anger, disgust, surprise) varying in intensity in 10% increments from 0 to 100%.

RESULTS

Participants were significantly more accurate in identifying correct emotions viewing faces from the PAFs. In both sets, participants identified happy faces more accurately than fearful faces, were least likely to misclassify facial expressions as happy and most likely to misclassify all emotions at low intensity as neutral. Accuracy in identifying facial expressions improved with increasing emotion intensity for both sets, reaching peaks at 60 and 80% intensity for the PAFs and PoFA, respectively. The study was limited by small sizes and age-range of participants and ethnic diversity of actors.

CONCLUSIONS

The PAFs successfully depicted a range of emotional expressions with improved performance over the PoFA and may be used as a contemporary set in facial expression recognition tests.

Machine learning in the prediction of depression treatment outcomes: a systematic review and meta-analysis

BACKGROUND

Multiple treatments are effective for major depressive disorder (MDD), but the outcomes of each treatment vary broadly among individuals. Accurate prediction of outcomes is needed to help select a treatment that is likely to work for a given person. We aim to examine the performance of machine learning methods in delivering replicable predictions of treatment outcomes.

METHODS

Of 7732 non-duplicate records identified through literature search, we retained 59 eligible reports and extracted data on sample, treatment, predictors, machine learning method, and treatment outcome prediction. A minimum sample size of 100 and an adequate validation method were used to identify adequate-quality studies. The effects of study features on prediction accuracy were tested with mixed-effects models. Fifty-four of the studies provided accuracy estimates or other estimates that allowed calculation of balanced accuracy of predicting outcomes of treatment.

RESULTS

Eight adequate-quality studies reported a mean accuracy of 0.63 [95% confidence interval (CI) 0.56-0.71], which was significantly lower than a mean accuracy of 0.75 (95% CI 0.72-0.78) in the other 46 studies. Among the adequate-quality studies, accuracies were higher when predicting treatment resistance (0.69) and lower when predicting remission (0.60) or response (0.56). The choice of machine learning method, feature selection, and the ratio of features to individuals were not associated with reported accuracy.

CONCLUSIONS

The negative relationship between study quality and prediction accuracy, combined with a lack of independent replication, invites caution when evaluating the potential of machine learning applications for personalizing the treatment of depression.

Effectiveness of common antidepressants: a post market release study

BACKGROUND

This study summarizes the experiences of patients, who have multiple comorbidities, with 15 mono-treated antidepressants.

METHODS

This is a retrospective, observational, matched case control study. The cohort was organized using claims data available through OptumLabs for depressed patients treated with antidepressants between January 1, 2001 and December 31, 2018. The cohort included patients from all states within United States of America. The analysis focused on 3,678,082 patients with major depression who had 10,221,145 antidepressant treatments. Using the robust, and large predictors of remission, and propensity to prescribe an antidepressant, the study created 16,770 subgroups of patients. The study reports the remission rate for the antidepressants within the subgroups. The overall impact of antidepressant on remission was calculated as the common odds ratio across the strata.

FINDINGS

The study accurately modelled clinicians' prescription patterns (cross-validated Area under the Receiver Operating Curve, AROC, of 82.0%, varied from 77% to 90%) and patients' remission (cross-validated AROC of 72.0%, varied from 69.5% to 78%). In different strata, contrary to published randomized studies, remission rates differed significantly and antidepressants were not equally effective. For example, in age and gender subgroups, the best antidepressant had an average remission rate of 50.78%, 1.5 times higher than the average antidepressant (30.30% remission rate) and 20 times higher than the worst antidepressant. The Breslow-Day chi-square test for homogeneity showed that across strata a homogenous common odds-ratio did not exist (alpha<0.0001). Therefore, the choice of the optimal antidepressant depended on the strata defined by the patient's medical history.

INTERPRETATION

Study findings may not be appropriate for specific patients. To help clinicians assess the transferability of study findings to specific patient, the web site http://hi.gmu.edu/ad assesses the patient's medical history, finds similar cases in our data, and recommends an antidepressant based on the experience of remission in our data. Patients can share this site's recommendations with their clinicians, who can then assess the appropriateness of the recommendations.

FUNDING

This project was funded by the Robert Wood Johnson foundation grant #76786. The development of related web site was supported by grant 247-02-20 from Virginia's Commonwealth Health Research Board.

Revisiting Treatment Options for Depressed Patients with Generalised Anxiety Disorder

Symptoms of anxiety and depression often coexist, and evidence suggests that this has a genetic basis, among other possible causes. However, the current classification of comorbid generalised anxiety disorder (GAD) and depression (anxious depression) in the Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition; DSM-5) does not fully reflect the high prevalence of anxiety symptoms in people with depression and the International Classification of Diseases (10th and 11th revisions) has tended to identify anxious depression with minor disorders seen in primary care. As a result, few dedicated therapeutic trials have been conducted in patients with anxious depression, and specific treatment guidelines and recommendations are lacking. Fortunately, there is considerable therapeutic overlap between anxiety and depression, such that many agents with antidepressant efficacy are also effective for symptoms of GAD. The initial treatment of a patient with depression and symptoms of anxiety should be with an agent that is approved for both major depressive disorder and GAD, such as a selective serotonin reuptake inhibitor. There is an obvious need for greater recognition of anxious depression in order to boost the volume of high-quality clinical data, which should translate over time into better, more specific treatment recommendations and improved outcomes.

An Experimental Medicine Investigation of the Effects of Subacute Pramipexole Treatment on Emotional Information Processing in Healthy Volunteers

Treatment with the dopamine D2/D3 receptor agonist pramipexole has demonstrated promising clinical effects in patients with depression. However, the mechanisms through which pramipexole might alleviate depressive symptoms are currently not well understood. Conventional antidepressant drugs are thought to work by biasing the processing of emotional information in favour of positive relative to negative appraisal. In this study, we used an established experimental medicine assay to explore whether pramipexole treatment might have a similar effect. Employing a double-blind, parallel-group design, 40 healthy volunteers (aged 18 to 43 years, 50% female) were randomly allocated to 12 to 15 days of treatment with either pramipexole (at a peak daily dose of 1.0 mg pramipexole salt) or placebo. After treatment was established, emotional information processing was assessed on the neural level by measuring amygdala activity in response to positive and negative facial emotional expressions, using functional magnetic resonance imaging (MRI). In addition, behavioural measures of emotional information processing were collected at baseline and on drug, using an established computerized task battery, tapping into different cognitive domains. As predicted, pramipexole-treated participants, compared to those receiving placebo, showed decreased neural activity in response to negative (fearful) vs. positive (happy) facial expressions in bilateral amygdala. Contrary to our predictions, however, pramipexole treatment had no significant antidepressant-like effect on behavioural measures of emotional processing. This study provides the first experimental evidence that subacute pramipexole treatment in healthy volunteers modifies neural responses to emotional information in a manner that resembles the effects of conventional antidepressant drugs.

The promise of machine learning in predicting treatment outcomes in psychiatry

For many years, psychiatrists have tried to understand factors involved in response to medications or psychotherapies, in order to personalize their treatment choices. There is now a broad and growing interest in the idea that we can develop models to personalize treatment decisions using new statistical approaches from the field of machine learning and applying them to larger volumes of data. In this pursuit, there has been a paradigm shift away from experimental studies to confirm or refute specific hypotheses towards a focus on the overall explanatory power of a predictive model when tested on new, unseen datasets. In this paper, we review key studies using machine learning to predict treatment outcomes in psychiatry, ranging from medications and psychotherapies to digital interventions and neurobiological treatments. Next, we focus on some new sources of data that are being used for the development of predictive models based on machine learning, such as electronic health records, smartphone and social media data, and on the potential utility of data from genetics, electrophysiology, neuroimaging and cognitive testing. Finally, we discuss how far the field has come towards implementing prediction tools in real-world clinical practice. Relatively few retrospective studies to-date include appropriate external validation procedures, and there are even fewer prospective studies testing the clinical feasibility and effectiveness of predictive models. Applications of machine learning in psychiatry face some of the same ethical challenges posed by these techniques in other areas of medicine or computer science, which we discuss here. In short, machine learning is a nascent but important approach to improve the effectiveness of mental health care, and several prospective clinical studies suggest that it may be working already.