Denys Pak

The rapid development of technologies relating to telemedicine has brought with it new opportunities and potential particularly for use in time critical settings such as emergency care. It is also thought that telemedicine may help prevent attendance for minor illness or injury to major hospital emergency departments. We reviewed the evidence for telemedicine based approaches to emergency and acute healthcare settings in comparison to face to face patient care. Searches were performed in MEDLINE, EMBASE, PubMed and the Cochrane Database. In total, seven studies involving 958 patients with an acute or emergency medical presentation were identified. The quality of included trials was assessed using the Critical Appraisal Skills Programme (CASP) tool. Outcome data were pooled under four headings: time from symptom onset to consultation, patient satisfaction, mean duration of consultation and accuracy of diagnosis. During this review no results were found for a specific comparison between patient journey time to main unit and standard treatment intervention compared to telemedicine administered at a satellite clinic or facility. Further evidence is needed regarding the efficacy of telemedicine with regard to unnecessary patient recall and the possible difficulty it presents in clinician agreement rates within diagnostic and patient management decision making. In addition greater focus could be given to the patient and practitioner satisfaction rates as well as further examination of possible time saving in response rates and implementation of appropriate treatment with the use of telemedicine.

Introduction: Many patients have cardiac monitoring on HASUs across the UK in current practice. An appropriate question would be how good are those monitoring systems at AF detection? We installed a Nihon Kohden monitoring system with remote rhythm analysis for AF detection (Apolplex). Every night rhythm is remotely analysed. An email report is generated. By the time of the HASU ward round the analysis is available for review Methods: We retrospectively analysed positive AF reports for 4 months in 2016. Notes were reviewed to confirm that new AF was detected. Demographics and anticoagulation plans were recorded. Results: 30 positive cases were identified via Apoplex report. 2 cases were identified as false positive. Of the remaining 28 cases there were 17 new AF cases, 11 previously known. The average cohort age was 83 y. The patients were monitored for an average of 22.9 hours. 43% were TACS. Out of 17 new AF cases 10 were anticoagulated. Conclusion: This is a novel system with evidence based AF detection. UK Stroke units are using serendipitous AF detection via cardiac monitoring systems – eg. tachycardia alarms or clinician recognition of irregular rhythm. Our system has proved effective in AF detection and easy to integrate into daily HASU care. This retrospective analysis confirms new AF detection in 17 cases allowing early anticoagulation, also avoiding outpatient 24 h tapes. There may be clinical and cost benefit of this system.

Objective Educating clinicians about artificial intelligence (AI) is urgent as the UK General Medical Council places liability with practitioners and the European Union AI Act with employers for appropriate training, but also because AI, like any tool, requires training to use safely. National Health Service England (NHSE) Capability Framework provides guidance, but frontline clinicians’ perspectives are unknown, so we sought to identify their priorities. Methods and analysis Iterative interviews with residents, educators and experts synthesised 10 contextualised AI-related problem statements. We surveyed residents and consultant-educators in the East of England, who rated their confidence and importance. Participants also ranked their preferred learning modality. Results We received 317 responses. Clinicians’ priorities, defined by high importance (I) and low confidence (C), were: ‘understanding liability implications’ (I: 40%; C: 1.82/5), ‘determining appropriate levels of confidence in AI algorithms’ (I: 36.5%; C: 1.98/5) and ‘mitigating security and privacy risks’ (I: 34%; C: 1.68). Confidence was low (mean 20, range 10–50), with no significant difference between educators and residents. Residents preferred integration of training into regional teaching, while consultant–educators favoured webinars. Conclusion Our findings show that clinicians prioritise practical concerns, such as liability and determining confidence in algorithmic outputs. In contrast, critical appraisal and explaining AI to patients were deprioritised, despite their relevance to clinical safety. This study enhances the NHSE Capability Framework by contextualising AI-related capabilities for clinicians as users and identifying priorities with which to develop scalable training.