Search

Funder: MRC Mental Health Data Pathfinder grant (MC_PC_17213), Funder: MRC Mental Health Data Pathfinder grant (MC PC 17213), Machine learning (ML), one aspect of artificial intelligence (AI), involves computer algorithms that train themselves. They have been widely applied in the healthcare domain. However, many trained ML algorithms operate as black boxes, producing a prediction from input data without a clear explanation of their workings. Non-transparent predictions are of limited utility in many clinical domains, where decisions must be justifiable. Here, we apply class-contrastive counterfactual reasoning to ML to demonstrate how specific changes in inputs lead to different predictions of mortality in people with severe mental illness (SMI), a major public health challenge. We produce predictions accompanied by visual and textual explanations as to how the prediction would have differed given specific changes to the input. We apply it to routinely collected data from a mental health secondary care provider in patients with schizophrenia. Using a data structuring framework informed by clinical knowledge, we captured information on physical health, mental health, and social predisposing factors. We then trained an ML algorithm and other statistical learning techniques to predict the risk of death. The ML algorithm predicted mortality with an area under receiver operating characteristic curve (AUROC) of 0.80 ( 95\% confidence intervals [0.78, 0.82] ). We used class-contrastive analysis to produce explanations for the model predictions. We outline the scenarios in which class-contrastive analysis is likely to be successful in producing explanations for model predictions. Our aim is not to advocate for a particular model but show an application of the class-contrastive analysis technique to electronic healthcare record data for a disease of public health significance. In patients with schizophrenia, our work suggests that use or prescription of medications like antidepressants was associated with lower risk of death. Abuse of alcohol/drugs and a diagnosis of delirium were associated with higher risk of death. Our ML models highlight the role of co-morbidities in determining mortality in patients with schizophrenia and the need to manage them. We hope that some of these bio-social factors can be targeted therapeutically by either patient-level or service-level interventions. Our approach combines clinical knowledge, health data, and statistical learning, to make predictions interpretable to clinicians using class-contrastive reasoning. This is a step towards interpretable AI in the management of patients with schizophrenia and potentially other diseases.

The European Mental Health Action Plan (EMHAP) 2013-2020 promoted community-based mental health services. One potential success indicator is the provision of antidepressant medication to those needing it.Public data from two surveys (Health Survey for England, UK; Survey of Health, Ageing and Retirement in Europe) covered 19 European countries across EMHAP phases one (2011-2015) and two (2015-2018). People screening positive for depressive symptoms by self-report were included. The primary outcome was antidepressant use: using country-specific weighted regression models, we estimated temporal trends and subgroup disparities in antidepressant receipt, with secondary analysis by country-level measures including healthcare expenditure.Across 37,250 participants, after controlling for age, sex, wealth, and physical disability, antidepressant use (amongst those screening positive) increased significantly in 14/19 countries, with the smallest increase being in Slovenia (adjusted OR[AOR] for trend=1.68[1.20-2.36]) and the highest increase being in Germany (AOR for trend=10.07[7.54-13.46]) and Austria (AOR for trend=10.07[7.32-13.74]). The overall proportion using antidepressants was positively associated with national health expenditure (coefficient=5.43[1.62-9.25]), but not with gross national income per capita or the number of psychiatrists, general practitioners, or psychiatric hospital beds. In 15/19 countries, antidepressants were used less by ≥65-year-olds than 50-64-year-olds, with the smallest differential reported in Luxembourg (AOR=0.70[0.49, 0.98]) and the highest in Germany (AOR=0.28[0.21, 0.37]); this disparity widened in 12/15 countries. Men used antidepressants less than women in 8/19 countries, across phases. In 13/19 countries, people with physical disability were more likely to receive antidepressants, with the smallest gap in Italy (AOR=1.42[1.12-1.80]) and the largest in Israel (AOR=2.34[1.46-3.74]); this disparity narrowed in 5/13 countries. Disparity by wealth was found in 8/19 countries, but its temporal trend varied.Usage of antidepressants by those with depressive symptoms has increased, with wide variation between countries and subgroups. Disparities across age, sex, and disability should prompt further research.Medical Research Council (grants MC_PC_17213 and MR/W014386/1), UK National Institute for Health Research (NIHR) Applied Research Collaboration (ARC) East of England, NIHR Cambridge Biomedical Research Centre (BRC-1215-20014).

Portable 12 V power supplies are used extensively for communications and power tool applications. These devices demand fast response times of the power supply. Fuel cells are generally best suited to continuous power applications and require an initial warm-up period, although they offer the prospect of increased operational duration over a battery for a given weight of portable system. This paper investigates the combination of specific energy performance from the fuel cell system with the specific power and response time of the battery. Two separate hybrid systems have been developed and tested; a planar, 20-cell, polymer electrolyte membrane fuel cell (PEMFC) stack together with either a lead–acid or nickel/cadmium battery; and a conventional 20-cell, bipolar, PEMFC stack. Both systems have been tested under pulse-load conditions at temperatures between −20°C and +40°C, and for comparison, the individual components have undergone similar tests. The hybrid systems have successfully operated continuously for several weeks under load profiles that the fuel cell alone could not sustain.

Background Lithium carbonate and valproate semisodium are both recommended as monotherapy for prevention of relapse in bipolar disorder, but are not individually fully effective in many patients. If combination therapy with both agents is better than monotherapy, many relapses and consequent disability could be avoided. We aimed to establish whether lithium plus valproate was better than monotherapy with either drug alone for relapse prevention in bipolar I disorder. Methods 330 patients aged 16 years and older with bipolar I disorder from 41 sites in the UK, France, USA, and Italy were randomly allocated to open-label lithium monotherapy (plasma concentration 0·4–1·0 mmol/L, n=110), valproate monotherapy (750–1250 mg, n=110), or both agents in combination (n=110), after an active run-in of 4–8 weeks on the combination. Randomisation was by computer program, and investigators and participants were informed of treatment allocation. All outcome events were considered by the trial management team, who were masked to treatment assignment. Participants were followed up for up to 24 months. The primary outcome was initiation of new intervention for an emergent mood episode, which was compared between groups by Cox regression. Analysis was by intention to treat. This study is registered, number ISRCTN 55261332. Findings 59 (54%) of 110 people in the combination therapy group, 65 (59%) of 110 in the lithium group, and 76 (69%) of 110 in the valproate group had a primary outcome event during follow-up. Hazard ratios for the primary outcome were 0·59 (95% CI 0·42–0·83, p=0·0023) for combination therapy versus valproate, 0·82 (0·58–1·17, p=0·27) for combination therapy versus lithium, and 0·71 (0·51–1·00, p=0·0472) for lithium versus valproate. 16 participants had serious adverse events after randomisation: seven receiving valproate monotherapy (three deaths); five lithium monotherapy (two deaths); and four combination therapy (one death). Interpretation For people with bipolar I disorder, for whom long-term therapy is clinically indicated, both combination therapy with lithium plus valproate and lithium monotherapy are more likely to prevent relapse than is valproate monotherapy. This benefit seems to be irrespective of baseline severity of illness and is maintained for up to 2 years. BALANCE could neither reliably confirm nor refute a benefit of combination therapy compared with lithium monotherapy. Funding Stanley Medical Research Institute; Sanofi-Aventis.

The analysis of the data reported by Tuma ([2000][1]) is seriously flawed. In this report there are no primary outcome data for 26 (48%) of the elderly cohort and 8 (14%) of the younger adults. The eight elderly people developing dementia at the 4.5 years outcome point are included in the analysis

OBJECTIVE: To describe the development and validation of the Clinical Global Impression-Schizophrenia (CGI-SCH) scale, designed to assess positive, negative, depressive and cognitive symptoms in schizophrenia. METHOD: The CGI-SCH scale was adapted from the CGI scale. Concurrent validity and sensitivity to change were assessed by comparison with the Positive and Negative Symptom Severity (PANSS) and Global Assessment of Functioning (GAF) scales. To evaluate inter-rater reliability, all patients were assessed by two clinicians. RESULTS: Symptoms were assessed in 114 patients. Correlation coefficients between the CGI-SCH and the GAF and PANSS scores were high (most above 0.75), and were highest for positive and negative symptoms. Reliability was substantial (intraclass correlation coefficient, ICC > 0.70) in all but one dimension (depressive dimension, ICC = 0.64). CONCLUSION: The CGI-SCH scale is a valid, reliable instrument to evaluate severity and treatment response in schizophrenia. Given its simplicity, brevity and clinical face validity, the scale is appropriate for use in observational studies and routine clinical practice. Acta Psychiatr Scand Suppl