The Covid-19 pandemic has placed unprecedented pressure on healthcare systems and workers around the world. Such pressures may impact on working conditions, psychological wellbeing and perception of safety. In spite of this, no study has assessed the relationship between safety attitudes and psychological outcomes. Moreover, only limited studies have examined the relationship between personal characteristics and psychological outcomes during Covid-19. From 22nd March 2020 to 18th June 2020, healthcare workers from the United Kingdom, Poland, and Singapore were invited to participate using a self-administered questionnaire comprising the Safety Attitudes Questionnaire (SAQ), Oldenburg Burnout Inventory (OLBI) and Hospital Anxiety and Depression Scale (HADS) to evaluate safety culture, burnout and anxiety/depression. Multivariate logistic regression was used to determine predictors of burnout, anxiety and depression. Of 3,537 healthcare workers who participated in the study, 2,364 (67%) screened positive for burnout, 701 (20%) for anxiety, and 389 (11%) for depression. Significant predictors of burnout included patient-facing roles: doctor (OR 2.10; 95% CI 1.49–2.95), nurse (OR 1.38; 95% CI 1.04–1.84), and ‘other clinical’ (OR 2.02; 95% CI 1.45–2.82); being redeployed (OR 1.27; 95% CI 1.02–1.58), bottom quartile SAQ score (OR 2.43; 95% CI 1.98–2.99), anxiety (OR 4.87; 95% CI 3.92–6.06) and depression (OR 4.06; 95% CI 3.04–5.42). Significant factors inversely correlated with burnout included being tested for SARS-CoV-2 (OR 0.64; 95% CI 0.51–0.82) and top quartile SAQ score (OR 0.30; 95% CI 0.22–0.40). Significant factors associated with anxiety and depression, included burnout, gender, safety attitudes and job role. Our findings demonstrate a significant burden of burnout, anxiety, and depression amongst healthcare workers. A strong association was seen between SARS-CoV-2 testing, safety attitudes, gender, job role, redeployment and psychological state. These findings highlight the importance of targeted support services for at risk groups and proactive SARS-CoV-2 testing of healthcare workers.

Background A significant proportion of individuals experience lingering and debilitating symptoms following acute COVID-19 infection. The National Institute for Health and Care Excellence (NICE) have coined the persistent cluster of symptoms as post-COVID syndrome. This has been further sub-categorised into acute post-COVID syndrome for symptoms persisting three weeks beyond initial infection and chronic post-COVID syndrome for symptoms persisting beyond twelve weeks. The aim of this review was to detail the prevalence of clinical features and identify potential predictors for acute and chronic post-COVID syndrome. Methods A systematic literature search, with no language restrictions, was performed to identify studies detailing characteristics and outcomes related to survivorship of post-COVID syndrome. The last search was performed on 6 March 2021 and all pre-dating published articles included. A means of proportion meta-analysis was performed to quantify characteristics of acute and chronic post-COVID syndrome. Study quality was assessed with a specific risk of bias tool. PROSPERO Registration: CRD42020222855 Findings A total of 43 studies met the eligibility criteria; of which, 38 allowed for meta-analysis. Fatigue and dyspnoea were the most prevalent symptoms in acute post-COVID (0·37 and 0·35) and fatigue and sleep disturbance in chronic post-COVID syndrome (0·48 and 0·44), respectively. The available evidence is generally of poor quality, with considerable risk of bias, and are of observational design. Interpretation In conclusion, this review highlights that flaws in data capture and interpretation, noted in the uncertainty within our meta-analysis, affect the applicability of current knowledge. Policy makers and researchers must focus on understanding the impact of this condition on individuals and society with appropriate funding initiatives and global collaborative research.

Deep learning (DL) has the potential to transform medical diagnostics. However, the diagnostic accuracy of DL is uncertain. Our aim was to evaluate the diagnostic accuracy of DL algorithms to identify pathology in medical imaging. Searches were conducted in Medline and EMBASE up to January 2020. We identified 11,921 studies, of which 503 were included in the systematic review. Eighty-two studies in ophthalmology, 82 in breast disease and 115 in respiratory disease were included for meta-analysis. Two hundred twenty-four studies in other specialities were included for qualitative review. Peer-reviewed studies that reported on the diagnostic accuracy of DL algorithms to identify pathology using medical imaging were included. Primary outcomes were measures of diagnostic accuracy, study design and reporting standards in the literature. Estimates were pooled using random-effects meta-analysis. In ophthalmology, AUC’s ranged between 0.933 and 1 for diagnosing diabetic retinopathy, age-related macular degeneration and glaucoma on retinal fundus photographs and optical coherence tomography. In respiratory imaging, AUC’s ranged between 0.864 and 0.937 for diagnosing lung nodules or lung cancer on chest X-ray or CT scan. For breast imaging, AUC’s ranged between 0.868 and 0.909 for diagnosing breast cancer on mammogram, ultrasound, MRI and digital breast tomosynthesis. Heterogeneity was high between studies and extensive variation in methodology, terminology and outcome measures was noted. This can lead to an overestimation of the diagnostic accuracy of DL algorithms on medical imaging. There is an immediate need for the development of artificial intelligence-specific EQUATOR guidelines, particularly STARD, in order to provide guidance around key issues in this field.

No abstract

Covid-19 has placed an unprecedented demand on healthcare systems worldwide. A positive safety culture is associated with improved patient safety and, in turn, with patient outcomes. To date, no study has evaluated the impact of Covid-19 on safety culture. The Safety Attitudes Questionnaire (SAQ) was used to investigate safety culture at a large UK healthcare trust during Covid-19. Findings were compared with baseline data from 2017. Incident reporting from the year preceding the pandemic was also examined. SAQ scores of doctors and “other clinical staff”, were relatively higher than the nursing group. During Covid-19, on univariate regression analysis, female gender, age 40–49 years, non-White ethnicity, and nursing job role were all associated with lower SAQ scores. Training and support for redeployment were associated with higher SAQ scores. On multivariate analysis, non-disclosed gender (−0.13), non-disclosed ethnicity (−0.11), nursing role (−0.15), and support (0.29) persisted to a level of significance. A significant decrease (p < 0.003) was seen in error reporting after the onset of the Covid-19 pandemic. This is the first study to investigate SAQ during Covid-19. Differences in SAQ scores were observed during Covid-19 between professional groups when compared to baseline. Reductions in incident reporting were also seen. These changes may reflect perception of risk, changes in volume or nature of work. High-quality support for redeployed staff may be associated with improved safety perception during future pandemics.

Objective: We compared the diagnostic performance of a novel point-ofcare duplex ultrasound test (podiatry ankle duplex scan; PAD-scan) against commonly used bedside tests for the detection of PAD in diabetes. Background: PAD is a major risk factor for diabetic foot ulceration and amputation. Its diagnosis is fundamental though challenging. Although a variety of bedside tests are available, there is no agreement as to which is the most useful. PAD-scan may be advantageous over current tests as it allows for vessel visualization and more accurate arterial waveform assessment. However, its accuracy has not been previously evaluated. Methods: From March to October 2019, we recruited 305 patients from 2 diabetic foot clinics. The diagnostic performance of ankle-brachial pressure index, toe-brachial pressure index, transcutaneous pressure of oxygen, pulse palpation, and ankle waveform assessment using PADscan and Doppler devices (audible and visual waveform assessment) were assessed. The reference test was a full lower limb duplex ultrasound. Results: Based on the reference test, 202 (66.2%) patients had evidence of PAD. PAD-scan had a significantly higher sensitivity [95%, confidence interval (CI) 90%-97%) as compared to all other tests. Particularly low sensitivities were seen with pulse palpation (43%, CI 36%-50%) and transcutaneous pressure of oxygen (31%, CI 24%-38%). PAD-scan had a lower specificity (77%, CI 67%-84%) compared to toe-brachial pressure index (86%, CI 78%-93%; P < 0.001), but not statistically different when compared to all other tests. Conclusions: PAD-scan has superior diagnostic utility and is a valid first line investigation.

IntroductionStandards for Reporting of Diagnostic Accuracy Study (STARD) was developed to improve the completeness and transparency of reporting in studies investigating diagnostic test accuracy. However, its current form, STARD 2015 does not address the issues and challenges raised by artificial intelligence (AI)-centred interventions. As such, we propose an AI-specific version of the STARD checklist (STARD-AI), which focuses on the reporting of AI diagnostic test accuracy studies. This paper describes the methods that will be used to develop STARD-AI.Methods and analysisThe development of the STARD-AI checklist can be distilled into six stages. (1) A project organisation phase has been undertaken, during which a Project Team and a Steering Committee were established; (2) An item generation process has been completed following a literature review, a patient and public involvement and engagement exercise and an online scoping survey of international experts; (3) A three-round modified Delphi consensus methodology is underway, which will culminate in a teleconference consensus meeting of experts; (4) Thereafter, the Project Team will draft the initial STARD-AI checklist and the accompanying documents; (5) A piloting phase among expert users will be undertaken to identify items which are either unclear or missing. This process, consisting of surveys and semistructured interviews, will contribute towards the explanation and elaboration document and (6) On finalisation of the manuscripts, the group’s efforts turn towards an organised dissemination and implementation strategy to maximise end-user adoption.Ethics and disseminationEthical approval has been granted by the Joint Research Compliance Office at Imperial College London (reference number: 19IC5679). A dissemination strategy will be aimed towards five groups of stakeholders: (1) academia, (2) policy, (3) guidelines and regulation, (4) industry and (5) public and non-specific stakeholders. We anticipate that dissemination will take place in Q3 of 2021.

Objective: To identify the most prevalent symptoms and those with greatest impact upon health-related quality of life (HRQOL) among esophageal cancer survivors. Background: Long-term symptom burden after esophagectomy, and associations with HRQOL, are poorly understood.Patients and Methods: Between 2010 and 2016, patients from 20 European Centers who underwent esophageal cancer surgery, and were disease-free at least 1 year postoperatively were asked to complete LASER, EORTC-QLQ-C30, and QLQ-OG25 questionnaires. Specific symptom questionnaire items that were associated with poor HRQOL as identified by EORTC QLQ-C30 and QLQ-OG25 were identified by multivariable regression analysis and combined to form a tool. Results: A total of 876 of 1081 invited patients responded to the questionnaire, giving a response rate of 81%. Of these, 66.9% stated in the last 6 months they had symptoms associated with their esophagectomy. Ongoing weight loss was reported by 10.4% of patients, and only 13.8% returned to work with the same activities. Three LASER symptoms were correlated with poor HRQOL on multivariable analysis; pain on scars on chest (odds ratio (OR) 1.27; 95% CI 0.97-1.65), low mood (OR 1.42; 95% CI 1.15-1.77) and reduced energy or activity tolerance (OR 1.37; 95% CI 1.18-1.59). The areas under the curves for the development and validation datasets were 0.81 AE 0.02 and 0.82 AE 0.09 respectively. Conclusion: Two-thirds of patients experience significant symptoms more than 1 year after surgery. The 3 key symptoms associated with poor HRQOL identified in this study should be further validated, and could be used in clinical practice to identify patients who require increased support.