An emergency arrives – the 2025 scenario
Blog: Capgemini CTO Blog
I’ve just spent another day hearing how digital is transforming the ways of working in banking and utilities. On my journey home, I start thinking about what would happen if some of this technology were applied to the emergency department where I was a consultant emergency physician. It would probably go something like this …
My smartphone vibrates, a message tells me that a 48-year-old, medium-risk chest pain patient is arriving in cubicle 15 in 10 minutes. Data has been transmitted from the ambulance and analyzed to determine the outline needs of the patient. It has determined the initial presentational pathway and the resources needed and has allocated a cubicle in ER.
When I arrive at the cubicle, the plasma screen is displaying additional patient details but also has the pre-arrival checklist of people and resources, and the clinical checklist for the predicted pathway. As staff arrive, they are checked-in automatically with their ID badges RFID and allocated preparatory tasks. If a team member is not present just before the ambulance arrives, another clinician can be messaged. EDwin (the virtual assistant) knows who is available and who is undertaking low-priority tasks. One minute before arrival, the final checklist is displayed on the screen ensuring that staff are in place and understand their roles and that the appropriate equipment is ready. EDwin checks in with me (the team leader) so that he can recognize my voice against the background.
The ambulance paramedic arrives and gives his handover. EDwin is listening and spots keywords (key information from the ambulance electronic record and the verbal handover) that are being displayed on the screen in a standardized format that is familiar to the ED staff. I agree that this is a medium-risk chest pain patient and confirm this to EDwin. He changes the screen to the initial care checklist, which also shows who should undertake each task. As the tasks are completed, staff either tap the on-screen tick boxes or tell EDwin that a particular task is complete.
Knowing the results of the initial vital signs and the examination, EDwin suggests a continuing clinical plan, which I then validate. I have used the troponin patch, which detects troponin levels in the patient’s sweat, and the cubicle webcam measures the color intensity and tells me the troponin level. I have also asked to see previous records on screen to help determine a course of action. I have allocated a clinician and a nurse to undertake the next stage of care. The system has suggested staff based on their existing workloads as well as skill- and productivity levels. In the background, the system monitors when results become available and knows when to inform the clinician. Any delays are automatically highlighted and often resolved by messaging the appropriate person or another department’s virtual assistant.
EDwin assists in delivering treatments. He is asked for a drug and told the indication. He calculates the appropriate dose (helped by the weighing device in the trolley) and then the pharmacy robot delivers a preloaded syringe with instructions on delivery route and method. His error rate is significantly lower than that based on human calculations.
As additional patient information is gathered and it becomes apparent that the patient would be more appropriately cared for in a different area or by a different team, EDwin suggests this to the team leader. If the team leaders agrees, EDwin organizes and transfers the information. The receiving team receives a message with a link to the data and an ETA to a specific space in their unit. The portering team and transfer team are allocated the task, knowing their present location and their availability. If necessary, other tasks may be delayed – for example, meal distribution can be delayed to ensure timely patient transfer, however catering and the wards are aware in advance because of the predictions in the system.
From before the time the patient arrived, the system has calculated the likelihood of the patient going to another department or needing an admission. It has already looked for a space. It may have saved a bed for a specific patient or reserved a bed for one of the four patients with a 25% chance of admission. It will provide a name later.
The system has been undertaking an analysis of many aspects of departmental running over the previous few months to achieve this level of efficiency and accuracy. The clinical protocols were all derived from traditional clinical guidelines, however EDwin has used artificial intelligence to learn from every case. Any changes are analyzed and learning is achieved. The system may ask the clinician the rationale for certain decisions so that it can learn. It may have been updated because of new evidence detected in literature and made suggestions to the clinical team at the governance meeting. National guidelines will have been automatically incorporated and localized; the first time any clinician has an appropriate case after a change, a learning screen will appear. When any case is completed, the clinician can ask EDwin for a two-minute teaching session during which he will present a variety of learning opportunities, including case reflection, new evidence, related clinical challenges, or colleague feedback. These are automatically added to the individual’s portfolio and improve education by grasping the educational moment.
When the system first started, there were concerns that efficiency and staff utilization targets would get in the way of empathetic care, but the system soon learned that it needed to include time for the compassion and personal aspects of care. It also allowed patients to speak to EDwin, to find out what was happening, what they were waiting for, and the likely outcome (diagnosis, care plan, admission). However, it could also highlight to the health professionals the need for more information, better explanation, and communication.
This technology is either already available or will soon be, so perhaps this future could be a lot closer than 2025. More challenging are the governance and ethics questions behind these digital innovations. How will patients respond to a new, electronic voice in wards offering up diagnoses and advice on a constant basis? How will staff feel if their every movement in the hospital is tracked? What happens when privacy rights cause a reduction in the quality of treatment for an individual? Will there need to be new, more sophisticated patient permission processes? And how might these concerns, and others, delay or prevent such advances?
Professor Matthew Cooke is Chief Clinical officer at Capgemini UK. He was an emergency physician in the UK and the National Clinical Director for Urgent and Emergency Care in England. He is also Professor of Clinical Systems Design at Warwick Medical School, UK.