Unsupervised Machine Learning to Identify Depressive Subtypes / 대한의료정보학회지

Objectives@#This study evaluated an unsupervised machine learning method, latent Dirichlet allocation (LDA), as a method for identifying subtypes of depression within symptom data. @*Methods@#Data from 18,314 depressed patients were used to create LDA models. The outcomes included future emergency presentations, crisis events, and behavioral problems. One model was chosen for further analysis based upon its potential as a clinically meaningful construct. The associations between patient groups created with the final LDA model and outcomes were tested. These steps were repeated with a commonly-used latent variable model to provide additional context to the LDA results. @*Results@#Five subtypes were identified using the final LDA model. Prior to the outcome analysis, the subtypes were labeled based upon the symptom distributions they produced: psychotic, severe, mild, agitated, and anergic-apathetic. The patient groups largely aligned with the outcome data. For example, the psychotic and severe subgroups were more likely to have emergency presentations (odds ratio [OR] = 1.29; 95% confidence interval [CI], 1.17–1.43 and OR = 1.16; 95% CI, 1.05–1.29, respectively), whereas these outcomes were less likely in the mild subgroup (OR = 0.86; 95% CI, 0.78–0.94). We found that the LDA subtypes were characterized by clusters of unique symptoms. This contrasted with the latent variable model subtypes, which were largely stratified by severity. @*Conclusions@#This study suggests that LDA can surface clinically meaningful, qualitative subtypes. Future work could be incorporated into studies concerning the biological bases of depression, thereby contributing to the development of new psychiatric therapeutics.

Outcomes of COVID-19 infection in patients treated with Clozapine.

BackgroundClozapine, an antipsychotic, is associated with increased susceptibility to infection with COVID-19, compared to other antipsychotics. AimsTo investigate associations between clozapine treatment and increased risk of adverse outcomes of COVID-19, namely COVID-related hospitalisation and intensive care treatment, and death, among patients taking antipsychotics with schizophrenia-spectrum disorders. MethodUsing data from South London and Maudsley NHS Foundation Trust (SLAM) clinical records, via the Clinical Record Interactive Search system, we identified 157 individuals who had an ICD-10 diagnosis of schizophrenia-spectrum disorders, were taking antipsychotics at the time of the COVID-19 pandemic in the UK, and had a laboratory-confirmed COVID-19 infection. The following health outcomes were measured: COVID-related hospitalisation, COVID-related intensive care treatment death. We tested associations between clozapine treatment and each outcome using logistic regression models, adjusting for gender, age, ethnicity, neighbourhood deprivation, obesity, smoking status, diabetes, asthma, bronchitis and hypertension using propensity scores. ResultsIn the 157 individuals who developed COVID while on antipsychotics, there were 44 COVID-related hospitalisations, 13 COVID-related intensive care treatments and 13 deaths of any cause during the follow-up period. In the unadjusted analysis, there was no significant association between clozapine and any of the outcomes and there remained no associations following adjusting for the confounding variables. ConclusionsIn our sample of patients with COVID-19 and schizophrenia-spectrum disorders, we found no evidence that clozapine treatment puts patients at increased risk of hospitalisation, intensive care treatment or death, compared to any other antipsychotic treatment. However, further research should be considered in larger samples to confirm this. Conflict of interestRDH has received research funding from Roche, Pfizer, Janssen, and Lundbeck. DFF has received research funding from Janssen and Lundbeck. JHM has received research funding from Lundbeck. JTT has received research funding from Bristol-Meyers-Squibb. RS declares research support in the last 36 months from Janssen, GSK and Takeda. Ethics statementThe research was conducted under ethical approval reference 18/SC/0372 from Oxfordshire Research Ethics Committee C.

Self-harm presentations to Emergency Departments and Place of Safety during the first wave of the UK COVID-19 pandemic: South London and Maudsley data on service use from February to June 2020.

The lockdown and social distancing policy imposed due to the COVID-19 pandemic has had a substantial impact on both mental health service delivery, and the ways in which people are accessing these services. Previous reports from the South London and Maudsley NHS Trust (SLaM; a large mental health service provider for around 1.2m residents in South London) have highlighted increased use of virtual contacts by mental health teams, with dropping numbers of face-to-face contacts over the first wave of the pandemic. There has been concern that the impact of the COVID-19 pandemic would lead to higher mental health emergencies, particularly instances of self-harm. However, with people advised to stay at home during the first wave lockdown, it is as yet unclear whether this impacted mental health service presentations. Taking advantage of SLaMs Clinical Records Interactive Search (CRIS) data resource with daily updates of information from its electronic mental health records, this paper describes overall presentations to Emergency Department (ED) mental health liaison teams, and those with self-harm. The paper focussed on three periods: i) a pre-lockdown period 1st February to 15th March, ii) a lockdown period 16th March to 10th May and iii) a post-lockdown period 11th May to 28th June. In summary, all attendances to EDs for mental health support decreased during the lockdown period, including those with self-harm. All types of self-harm decreased during lockdown, with self-poisoning remaining the most common. Attendances to EDs for mental health support increased post-lockdown, although were only just approaching pre-lockdown levels by the end of June 2020.

Impact of the COVID-19 pandemic on remote mental healthcare and prescribing in psychiatry

ObjectivesThe recent COVID-19 pandemic has disrupted mental healthcare delivery, with many services shifting from in- person to remote patient contact. We investigated the impact of the pandemic on the use of remote consultation and on the prescribing of psychiatric medications. Design and settingThe Clinical Record Interactive Search tool (CRIS) was used to examine de-identified electronic health records (EHRs) of people receiving mental healthcare from the South London and Maudsley (SLaM) NHS Foundation Trust. Data from the period before and after the onset of the pandemic were analysed using linear regression, and visualised using locally estimated scatterplot smoothing (LOESS). ParticipantsAll patients receiving care from SLaM between 7th January 2019 and 20th September 2020 (around 37,500 patients per week). Outcome measuresO_LIThe number of clinical contacts (in-person, remote or non-attended) with mental healthcare professionals per week C_LIO_LIPrescribing of antipsychotic and mood stabiliser medications per week C_LI ResultsFollowing the onset of the pandemic, the frequency of in-person contacts was significantly reduced compared to that in the previous year ({beta} coefficient: -5829.6 contacts, 95% CI -6919.5 to -4739.6, p<0.001), while the frequency of remote contacts significantly increased ({beta} coefficient: 3338.5 contacts, 95% CI 3074.4 to 3602.7, p<0.001). Rates of remote consultation were lower in older adults than in working age adults, children and adolescents. Despite this change in the type of patient contact, antipsychotic and mood stabiliser prescribing remained at similar levels. ConclusionsThe COVID-19 pandemic has been associated with a marked increase in remote consultation, particularly among younger patients. However, there was no evidence that this has led to changes in psychiatric prescribing. Nevertheless, further work is needed to ensure that older patients are able to access mental healthcare remotely.