Enhanced visualization of mobile chest X-ray images in the intensive care setting using software scatter correction.

BACKGROUND: Mobile chest X-ray (CXR) scans are performed within intensive treatment units (ITU) without anti-scatter grids for confirming tube and line hardware placement. Assessment is therefore challenging due to degraded subject contrast resulting from scatter. PURPOSE: To evaluate the efficacy of a software scatter correction method (commercially named Trueview) for enhanced hardware visualization and diagnostic quality in the ITU setting. MATERIAL AND METHODS: A total of 30 CXR scans were processed using Trueview and compared with standard original equipment manufacturer (OEM) images via observer scoring study involving two radiology and four ITU doctors to compare visualization of tubes and lines. Results were analyzed to determine observer preference and likelihood of diagnostic quality. RESULTS: Reviewers were more likely to score Trueview higher than OEM for mediastinal structures, bones, retrocardiac region, tube visibility, and tube safety (P < 0.01). Visual grading characteristic analysis suggested a clinical preference for Trueview compared with OEM for mediastinal structures (area under the visual grading characteristic curve [AUCVGC] = 0.60, 95% confidence interval [CI] = 0.55-0.65), bones (AUCVGC = 0.61, 95% CI = 0.55-0.66), retrocardiac region (AUCVGC = 0.64, 95% CI = 0.59-0.69), tube visibility (AUCVGC = 0.65, 95% CI = 0.60-0.70), and tube safety (AUCVGC = 0.68, 95% CI = 0.64-0.73). Reviewers were indifferent to visualization of the lung fields (AUCVGC = 0.49, 95% CI = 0.44-0.55). Registrars (3/6 reviewers) were indifferent to the mediastinal structure regions (AUCVGC = 0.54, 95% CI = 0.47-0.62). CONCLUSION: Reviewers were more confident in identifying the placement and safety of tubes and lines when reviewing Trueview images than they were when reviewing OEM.

A review of biomedical datasets relating to drug discovery: a knowledge graph perspective.

Drug discovery and development is a complex and costly process. Machine learning approaches are being investigated to help improve the effectiveness and speed of multiple stages of the drug discovery pipeline. Of these, those that use Knowledge Graphs (KG) have promise in many tasks, including drug repurposing, drug toxicity prediction and target gene-disease prioritization. In a drug discovery KG, crucial elements including genes, diseases and drugs are represented as entities, while relationships between them indicate an interaction. However, to construct high-quality KGs, suitable data are required. In this review, we detail publicly available sources suitable for use in constructing drug discovery focused KGs. We aim to help guide machine learning and KG practitioners who are interested in applying new techniques to the drug discovery field, but who may be unfamiliar with the relevant data sources. The datasets are selected via strict criteria, categorized according to the primary type of information contained within and are considered based upon what information could be extracted to build a KG. We then present a comparative analysis of existing public drug discovery KGs and an evaluation of selected motivating case studies from the literature. Additionally, we raise numerous and unique challenges and issues associated with the domain and its datasets, while also highlighting key future research directions. We hope this review will motivate KGs use in solving key and emerging questions in the drug discovery domain.

A Knowledge Graph-Enhanced Tensor Factorisation Model for Discovering Drug Targets.

The drug discovery and development process is a long and expensive one, costing over 1 billion USD on average per drug and taking 10-15 years. To reduce the high levels of attrition throughout the process, there has been a growing interest in applying machine learning methodologies to various stages of drug discovery and development in the recent decade, especially at the earliest stage - identification of druggable disease genes. In this paper, we have developed a new tensor factorisation model to predict potential drug targets (genes or proteins) for treating diseases. We created a three-dimensional data tensor consisting of 1,048 gene targets, 860 diseases and 230,011 evidence attributes and clinical outcomes connecting them, using data extracted from the Open Targets and PharmaProjects databases. We enriched the data with gene target representations learned from a drug discovery-oriented knowledge graph and applied our proposed method to predict the clinical outcomes for unseen gene target and disease pairs. We designed three evaluation strategies to measure the prediction performance and benchmarked several commonly used machine learning classifiers together with Bayesian matrix and tensor factorisation methods. The result shows that incorporating knowledge graph embeddings significantly improves the prediction accuracy and that training tensor factorisation alongside a dense neural network outperforms all other baselines. In summary, our framework combines two actively studied machine learning approaches to disease target identification, namely tensor factorisation and knowledge graph representation learning, which could be a promising avenue for further exploration in data-driven drug discovery.

Implications of topological imbalance for representation learning on biomedical knowledge graphs.

Adoption of recently developed methods from machine learning has given rise to creation of drug-discovery knowledge graphs (KGs) that utilize the interconnected nature of the domain. Graph-based modelling of the data, combined with KG embedding (KGE) methods, are promising as they provide a more intuitive representation and are suitable for inference tasks such as predicting missing links. One common application is to produce ranked lists of genes for a given disease, where the rank is based on the perceived likelihood of association between the gene and the disease. It is thus critical that these predictions are not only pertinent but also biologically meaningful. However, KGs can be biased either directly due to the underlying data sources that are integrated or due to modelling choices in the construction of the graph, one consequence of which is that certain entities can get topologically overrepresented. We demonstrate the effect of these inherent structural imbalances, resulting in densely connected entities being highly ranked no matter the context. We provide support for this observation across different datasets, models as well as predictive tasks. Further, we present various graph perturbation experiments which yield more support to the observation that KGE models can be more influenced by the frequency of entities rather than any biological information encoded within the relations. Our results highlight the importance of data modelling choices, and emphasizes the need for practitioners to be mindful of these issues when interpreting model outputs and during KG composition.

Neuronal Graphs: A Graph Theory Primer for Microscopic, Functional Networks of Neurons Recorded by Calcium Imaging.

Connected networks are a fundamental structure of neurobiology. Understanding these networks will help us elucidate the neural mechanisms of computation. Mathematically speaking these networks are "graphs"-structures containing objects that are connected. In neuroscience, the objects could be regions of the brain, e.g., fMRI data, or be individual neurons, e.g., calcium imaging with fluorescence microscopy. The formal study of graphs, graph theory, can provide neuroscientists with a large bank of algorithms for exploring networks. Graph theory has already been applied in a variety of ways to fMRI data but, more recently, has begun to be applied at the scales of neurons, e.g., from functional calcium imaging. In this primer we explain the basics of graph theory and relate them to features of microscopic functional networks of neurons from calcium imaging-neuronal graphs. We explore recent examples of graph theory applied to calcium imaging and we highlight some areas where researchers new to the field could go awry.

ANGPTL6 Genetic Variants Are an Underlying Cause of Familial Intracranial Aneurysms.

PURPOSE: To understand the role of the angiopoietin-like 6 gene (ANGPTL6) in intracranial aneurysms (IAs), we investigated its role in a large cohort of familial IAs. METHODS: Individuals with family history of IA were recruited to the Genetic and Observational Subarachnoid Haemorrhage (GOSH) study. The ANGPTL6 gene was sequenced using Sanger sequencing. Identified genetic variants were compared to a control population. RESULTS: We found 6 rare ANGPTL6 genetic variants in 9/275 individuals with a family history of IA (3.3%) (5 missense mutations and 1 nonsense mutation leading to a premature stop codon), none present in controls. One of these had been previously reported: c.392A>T (p.Glu131Val) on exon 2; another was very close: c.332G>A (p.Arg111His). Two further genetic variants lie within the fibrinogen-like domain of the ANGPTL6 gene, which may influence function or level of the ANGPTL6 protein. The last 2 missense mutations lie within the coiled-coil domain of the ANGPTL6 protein. All genetic variants were well conserved across species. CONCLUSION: ANGPTL6 genetic variants are an important cause of IA. Defective or lack of ANGPTL6 protein is therefore an important factor in blood vessel proliferation leading to IA; dysfunction of this protein is likely to cause abnormal proliferation or weakness of vessel walls. With these data, not only do we emphasize the importance of screening familial IA cases for ANGPTL6 and other genes involved in IA, but also highlight the ANGPTL6 pathway as a potential therapeutic target. CLASSIFICATION OF EVIDENCE: This is a Class III study showing some specificity of presence of the ANGPTL6 gene variant as a marker of familial intracranial aneurysms in a small subset of individuals with familial aneurysms.

Assessment of the Subarachnoid Hemorrhage International Trialists (SAHIT) Models for Dichotomized Long-Term Functional Outcome Prediction After Aneurysmal Subarachnoid Hemorrhage in a United Kingdom Multicenter Cohort Study.

BACKGROUND: Long-term outcome after subarachnoid hemorrhage, beyond the first few months, is difficult to predict, but has critical relevance to patients, their families, and carers. OBJECTIVE: To assess the performance of the Subarachnoid Hemorrhage International Trialists (SAHIT) prediction models, which were initially designed to predict short-term (90 d) outcome, as predictors of long-term (2 yr) functional outcome after aneurysmal subarachnoid hemorrhage (aSAH). METHODS: We included 1545 patients with angiographically-proven aSAH from the Genetic and Observational Subarachnoid Haemorrhage (GOSH) study recruited at 22 hospitals between 2011 and 2014. We collected data on age, WNFS grade on admission, history of hypertension, Fisher grade, aneurysm size and location, as well as treatment modality. Functional outcome was measured by the Glasgow Outcome Scale (GOS) with GOS 1 to 3 corresponding to unfavorable and 4 to 5 to favorable functional outcome, according to the SAHIT models. The SAHIT models were assessed for long-term outcome prediction by estimating measures of calibration (calibration slope) and discrimination (area under the receiver-operating characteristic curve [AUC]) in relation to poor clinical outcome. RESULTS: Follow-up was standardized to 2 yr using imputation methods. All 3 SAHIT models demonstrated acceptable predictive performance for long-term functional outcome. The estimated AUC was 0.71 (95% CI: 0.65-0.76), 0.73 (95% CI: 0.68-0.77), and 0.74 (95% CI: 0.69-0.79) for the core, neuroimaging, and full models, respectively; the calibration slopes were 0.86, 0.84, and 0.89, indicating good calibration. CONCLUSION: The SAHIT prediction models, incorporating simple factors available on hospital admission, show good predictive performance for long-term functional outcome after aSAH.

An Intentional Aconite Overdose: A Case Report.

BACKGROUND: Aconite is one of the most toxic known herbs, widely used for centuries as an essential Chinese medicine, but also for deliberate poisoning throughout history. Clinically indicated in herbal medicine for a range of ailments from headaches to muscle spasm, unfortunately the narrow therapeutic window may lead to a range of toxic presentations. The mechanism of action of the pharmacologically active compounds in Aconite relate to the activation of voltage gated sodium channels within a range of tissue including myocardial, neuronal and smooth muscle leading to persistent cellular activity. CASE PRESENTATION: We report on a rare case of a fifty year old male with intentional aconite overdose presenting with refractory cardiovascular instability from persistent life threatening arrhythmias, respiratory failure and seizure activity. CONCLUSION: An overview of Aconite, its history, pharmacological effects, treatment of overdose and outcomes is presented.

Haptoglobin genotype and outcome after aneurysmal subarachnoid haemorrhage.

OBJECTIVE: After aneurysmal subarachnoid haemorrhage (aSAH), extracellular haemoglobin (Hb) in the subarachnoid space is bound by haptoglobin, neutralising Hb toxicity and helping its clearance. Two exons in the HP gene (encoding haptoglobin) exhibit copy number variation (CNV), giving rise to HP1 and HP2 alleles, which influence haptoglobin expression level and possibly haptoglobin function. We hypothesised that the HP CNV associates with long-term outcome beyond the first year after aSAH. METHODS: The HP CNV was typed using quantitative PCR in 1299 aSAH survivors in the Genetics and Observational Subarachnoid Haemorrhage (GOSH) Study, a retrospective multicentre cohort study with a median follow-up of 18 months. To investigate mediation of the HP CNV effect by haptoglobin expression level, as opposed to functional differences, we used rs2000999, a single nucleotide polymorphism associated with haptoglobin expression independent of the HP CNV. Outcome was assessed using modified Rankin and Glasgow Outcome Scores. SAH volume was dichotomised on the Fisher grade. Haemoglobin-haptoglobin complexes were measured in cerebrospinal fluid (CSF) of 44 patients with aSAH and related to the HP CNV. RESULTS: The HP2 allele associated with a favourable long-term outcome after high-volume but not low-volume aSAH (multivariable logistic regression). However rs2000999 did not predict outcome. The HP2 allele associated with lower CSF haemoglobin-haptoglobin complex levels. The CSF Hb concentration after high-volume and low-volume aSAH was, respectively, higher and lower than the Hb-binding capacity of CSF haptoglobin. CONCLUSION: The HP2 allele carries a favourable long-term prognosis after high-volume aSAH. Haptoglobin and the Hb clearance pathway are therapeutic targets after aSAH.

Biomarker-guided antibiotic stewardship in suspected ventilator-associated pneumonia (VAPrapid2): a randomised controlled trial and process evaluation.

BACKGROUND: Ventilator-associated pneumonia is the most common intensive care unit (ICU)-acquired infection, yet accurate diagnosis remains difficult, leading to overuse of antibiotics. Low concentrations of IL-1ß and IL-8 in bronchoalveolar lavage fluid have been validated as effective markers for exclusion of ventilator-associated pneumonia. The VAPrapid2 trial aimed to determine whether measurement of bronchoalveolar lavage fluid IL-1ß and IL-8 could effectively and safely improve antibiotic stewardship in patients with clinically suspected ventilator-associated pneumonia. METHODS: VAPrapid2 was a multicentre, randomised controlled trial in patients admitted to 24 ICUs from 17 National Health Service hospital trusts across England, Scotland, and Northern Ireland. Patients were screened for eligibility and included if they were 18 years or older, intubated and mechanically ventilated for at least 48 h, and had suspected ventilator-associated pneumonia. Patients were randomly assigned (1:1) to biomarker-guided recommendation on antibiotics (intervention group) or routine use of antibiotics (control group) using a web-based randomisation service hosted by Newcastle Clinical Trials Unit. Patients were randomised using randomly permuted blocks of size four and six and stratified by site, with allocation concealment. Clinicians were masked to patient assignment for an initial period until biomarker results were reported. Bronchoalveolar lavage was done in all patients, with concentrations of IL-1ß and IL-8 rapidly determined in bronchoalveolar lavage fluid from patients randomised to the biomarker-based antibiotic recommendation group. If concentrations were below a previously validated cutoff, clinicians were advised that ventilator-associated pneumonia was unlikely and to consider discontinuing antibiotics. Patients in the routine use of antibiotics group received antibiotics according to usual practice at sites. Microbiology was done on bronchoalveolar lavage fluid from all patients and ventilator-associated pneumonia was confirmed by at least 104 colony forming units per mL of bronchoalveolar lavage fluid. The primary outcome was the distribution of antibiotic-free days in the 7 days following bronchoalveolar lavage. Data were analysed on an intention-to-treat basis, with an additional per-protocol analysis that excluded patients randomly assigned to the intervention group who defaulted to routine use of antibiotics because of failure to return an adequate biomarker result. An embedded process evaluation assessed factors influencing trial adoption, recruitment, and decision making. This study is registered with ISRCTN, ISRCTN65937227, and ClinicalTrials.gov, NCT01972425. FINDINGS: Between Nov 6, 2013, and Sept 13, 2016, 360 patients were screened for inclusion in the study. 146 patients were ineligible, leaving 214 who were recruited to the study. Four patients were excluded before randomisation, meaning that 210 patients were randomly assigned to biomarker-guided recommendation on antibiotics (n=104) or routine use of antibiotics (n=106). One patient in the biomarker-guided recommendation group was withdrawn by the clinical team before bronchoscopy and so was excluded from the intention-to-treat analysis. We found no significant difference in the primary outcome of the distribution of antibiotic-free days in the 7 days following bronchoalveolar lavage in the intention-to-treat analysis (p=0·58). Bronchoalveolar lavage was associated with a small and transient increase in oxygen requirements. Established prescribing practices, reluctance for bronchoalveolar lavage, and dependence on a chain of trial-related procedures emerged as factors that impaired trial processes. INTERPRETATION: Antibiotic use remains high in patients with suspected ventilator-associated pneumonia. Antibiotic stewardship was not improved by a rapid, highly sensitive rule-out test. Prescribing culture, rather than poor test performance, might explain this absence of effect. FUNDING: UK Department of Health and the Wellcome Trust.

Histamine and Delirium: Current Opinion.

Delirium is a very common, but refractory clinical state, notably present in intensive care and in the growing aging community. It is characterized by fluctuating disturbances in a number of key behavioral features, namely cognition, mood, attention, arousal, and self-awareness. Histamine is arguably the most pleotropic neurotransmitter in the human brain, and this review provides a rationale, and proposes that this neuroactive amine plays a role in modulating the characteristic features of delirium. While centrally permeable H1 and H2 histamine receptor antagonists have pro-delirium potential, we propose that centrally permeable H3 histamine receptor antagonists may provide an exciting new strategy to combat delirium. The Histamine H4 receptor may also have an indirect inflammatory neuroglial role which requires further exploration.

Association of functional MMP-2 gene variant with intracranial aneurysms: case-control genetic association study and meta-analysis.

INTRODUCTION: Abnormalities in Matrix Metalloproteinase (MMP) genes, which are important in extracellular matrix (ECM) maintenance and therefore arterial wall integrity are a plausible underlying mechanism of intracranial aneurysm (IA) formation, growth and subsequent rupture. We investigated whether the rs243865 C > T SNP (single nucleotide polymorphism) within the MMP-2 gene (which influences gene transcription) is associated with IA compared to matched controls. MATERIALS AND METHODS: We conducted a case-control genetic association study, adjusted for known IA risk factors (smoking and hypertension), in a UK Caucasian population of 1409 patients with intracranial aneurysms (IA), and 1290 matched controls, to determine the association of the rs243865 C > T functional MMP-2 gene SNP with IA (overall, and classified as ruptured and unruptured). We also undertook a meta-analysis of two previous studies examining this SNP. RESULTS: The rs243865 T allele was associated with IA presence in univariate (OR 1.18 [95% CI 1.04-1.33], p = .01) and in multi-variable analyses adjusted for smoking and hypertension status (OR 1.16 [95% CI 1.01-1.35], p = .042). Subgroup analysis demonstrated an association of the rs243865 SNP with ruptured IA (OR 1.18 [95% CI 1.03-1.34] p = .017), but, not unruptured IA (OR 1.17 [95% CI 0.97-1.42], p = .11). CONCLUSIONS: Our study demonstrated an association between the functional MMP-2 rs243865 variant and IAs. Our findings suggest a genetic role for altered extracellular matrix integrity in the pathogenesis of IA development and rupture.

Intensive versus standard physical rehabilitation therapy in the critically ill (EPICC): a multicentre, parallel-group, randomised controlled trial.

BACKGROUND: Early physical rehabilitation in the intensive care unit (ICU) has been shown to improve short-term clinical outcomes but long-term benefit has not been proven and the optimum intensity of rehabilitation is not known. METHODS: We conducted a randomised, parallel-group, allocation-concealed, assessor-blinded, controlled trial in patients who had received at least 48 hours of invasive or non-invasive ventilation. Participants were randomised in a 1:1 ratio, stratified by admitting ICU, admission type and level of independence. The intervention group had a target of 90 min physical rehabilitation per day, the control group a target of 30 min per day (both Monday to Friday). The primary outcome was the Physical Component Summary (PCS) measure of SF-36 at 6 months. RESULTS: We recruited 308 participants over 34 months: 150 assigned to the intervention and 158 to the control group. The intervention group received a median (IQR) of 161 (67-273) min of physical rehabilitation on ICU compared with 86 (31-139) min in the control group. At 6 months, 62 participants in the intervention group and 54 participants in the control group contributed primary outcome data. In the intervention group, 43 had died, 11 had withdrawn and 34 were lost to follow-up, while in the control group, 56 had died, 5 had withdrawn and 43 were lost to follow-up. There was no difference in the primary outcome at 6 months, mean (SD) PCS 37 (12.2) in the intervention group and 37 (11.3) in the control group. CONCLUSIONS: In this study, ICU-based physical rehabilitation did not appear to improve physical outcomes at 6 months compared with standard physical rehabilitation. TRIAL REGISTRATION NUMBER: ISRCTN 20436833.

Characteristics of Unruptured Compared to Ruptured Intracranial Aneurysms: A Multicenter Case-Control Study.

BACKGROUND: Only a minority of intracranial aneurysms rupture to cause subarachnoid hemorrhage. OBJECTIVE: To test the hypothesis that unruptured aneurysms have different characteristics and risk factor profiles compared to ruptured aneurysms. METHODS: We recruited patients with unruptured aneurysms or aneurysmal subarachnoid hemorrhages at 22 UK hospitals between 2011 and 2014. Demographic, clinical, and imaging data were collected using standardized case report forms. We compared risk factors using multivariable logistic regression. RESULTS: A total of 2334 patients (1729 with aneurysmal subarachnoid hemorrhage, 605 with unruptured aneurysms) were included (mean age 54.22 yr). In multivariable analyses, the following variables were independently associated with rupture status: black ethnicity (odds ratio [OR] 2.42; 95% confidence interval [CI] 1.29-4.56, compared to white) and aneurysm location (anterior cerebral artery/anterior communicating artery [OR 3.21; 95% CI 2.34-4.40], posterior communicating artery [OR 3.92; 95% CI 2.67-5.74], or posterior circulation [OR 3.12; 95% CI 2.08-4.70], compared to middle cerebral artery). The following variables were inversely associated with rupture status: antihypertensive medication (OR 0.65; 95% CI 0.49-0.84), hypercholesterolemia (0.64 OR; 95% CI 0.48-0.85), aspirin use (OR 0.28; 95% CI 0.20-0.40), internal carotid artery location (OR 0.53; 95% CI 0.38-0.75), and aneurysm size (per mm increase; OR 0.76; 95% CI 0.69-0.84). CONCLUSION: We show substantial differences in patient and aneurysm characteristics between ruptured and unruptured aneurysms. These findings support the hypothesis that different pathological mechanisms are involved in the formation of ruptured aneurysms and incidentally detected unruptured aneurysms. The potential protective effect of aspirin might justify randomized prevention trials in patients with unruptured aneurysms.

The research needs of an ICM trainee: The RAFT national survey results and initiatives to improve trainee research opportunities.

Research is key to intensive care practice but it is not certain whether trainees get the research experience they want to prepare them for clinical practice as consultants. We distributed a survey via Research & Audit Federation of Trainees to 478 current intensive care medicine trainees to establish their views. We restricted analysis to regions with an 80% or greater response rate (n = 235 responses) to reduce self-selection bias. Our results indicated that the desire for research activities was high but multiple barriers exist. We suggest simple initiatives combined with engagement of trainee research networks and local Clinical Research Networks to improve trainee access to research opportunities. Greater freedom to participate in research activities could be created by broader ARCP clinical governance requirements, flexible interpretation of the FICM curriculum and support of trainees wishing to use their non-clinical time for projects. A resource detailing local research activity and appropriate contacts could help trainees plan suitable activities.

Extra Physiotherapy in Critical Care (EPICC) Trial Protocol: a randomised controlled trial of intensive versus standard physical rehabilitation therapy in the critically ill.

INTRODUCTION: Patients discharged from Critical Care suffer from excessive longer term morbidity and mortality. Physical and mental health measures of quality of life show a marked and immediate fall after admission to Critical Care with some recovery over time. However, physical function is still significantly reduced at 6 months. The National Institute for Health and Care Excellence clinical guideline on rehabilitation after critical illness, identified the need for high-quality randomised controlled trials to determine the most effective rehabilitation strategy for critically ill patients at risk of critical illness-associated physical morbidity. In response to this, we will conduct a randomised controlled trial, comparing physiotherapy aimed at early and intensive patient mobilisation with routine care. We hypothesise that this intervention will improve physical outcomes and the mental health and functional well-being of survivors of critical illness. METHODS AND ANALYSIS: 308 adult patients who have received more than 48 h of non-invasive or invasive ventilation in Critical Care will be recruited to a patient-randomised, parallel group, controlled trial, comparing two intensities of physiotherapy. Participants will be randomised to receive either standard or intensive physiotherapy for the duration of their Critical Care admission. Outcomes will be recorded on Critical Care discharge, at 3 and 6 months following initial recruitment to the study. The primary outcome measure is physical health at 6 months, as measured by the SF-36 Physical Component Summary. Secondary outcomes include assessment of mental health, activities of daily living, delirium and ventilator-free days. We will also include a health economic analysis. ETHICS AND DISSEMINATION: The trial has ethical approval from Newcastle and North Tyneside 2 Research Ethics Committee (11/NE/0206). There is a Trial Oversight Committee including an independent chair. The results of the study will be submitted for publication in peer-reviewed journals and presented at national and international scientific meetings. TRIAL REGISTRATION NUMBER: ISRCTN20436833.

Project Post Intensive Care eXercise (PIX): A qualitative exploration of intensive care unit survivors' perceptions of quality of life post-discharge and experience of exercise rehabilitation.

Patients who survive critical illness often report deterioration in health related quality of life. This has not been shown to improve following post-intensive care unit (ICU) self-directed exercise. The Post Intensive Care eXercise (PIX) study demonstrated improved objectively measured fitness following a supervised exercise programme following critical illness and also suggested beneficial effects on physical and mental health. The qualitative arm of the PIX study reported here utilised focus groups to explore in more detail recovery from critical illness, quality of life following hospital discharge, perceptions of the exercise programme and it's impact on perceived well-being. Sixteen participants (eight of whom underwent the supervised exercise programme) were allocated to four psychologist lead focus groups. Themes identified after hospital discharge centred on social isolation, abandonment, vulnerability and reduced physical activity. However, patients in the exercise group described exercise training as motivating, increasing energy levels and sense of achievement, social interaction and confidence. This study adds to the sparse literature on the patient experience post critical illness. It supports the improvements in physical and mental health suggested with exercise in the PIX study and would support further research in relation to the effects of supervised exercise and rehabilitation programmes post critical illness. It recommends that future comparative outcome studies in this patient population also include interview-based assessment as part of assessment of quality of life and an individual's functional status.