内容来源：https://news.microsoft.com/source/emea/features/before-the-ambulance-arrives-how-ai-is-helping-the-nhs-rethink-patient-care/

内容总结：

人工智能助力英国国家医疗服务体系重塑患者护理：救护车到达之前，准备早已开始

编者按： 当英国民众拨打999呼叫救护车时，他们最关心的是救援何时到达。然而，在急救人员接触到患者之前，大量的规划、培训、分析与协调工作早已展开。英国国家医疗服务体系（NHS）正借助人工智能工具，从行政管理环节入手，为临床医护减负增效，进而重塑患者护理模式。

背景：救护车到达之前的“看不见的工作”

约克郡救护车服务NHS信托基金首席临床信息官詹姆斯·古尔丁表示：“我们通常只关注救护车到达现场的那一刻，但在急救人员接触到患者之前，有大量工作已经完成。”该信托基金是全英200多个NHS信托机构之一。这些前期工作涵盖服务规划、人员招聘与培训、人口健康数据分析，以及了解英语不流利患者的就医体验等。古尔丁说：“为确保急救人员到达患者面前时能够轻松做出正确判断，我们在背后做了大量工作。”

NHS全面推进生成式AI工具应用

在涉及3万名员工的早期试点之后，英格兰NHS于2026年6月宣布，计划向多达50.5万名临床医护人员和支持人员开放微软Copilot。试点参与者报告称，平均每个工作日节省43分钟，相当于每人每年节省约五周时间。

英格兰NHS全国首席临床信息官亚历克·普莱斯-福布斯博士指出，NHS在许多方面仍是一个“模拟系统”。他说：“我们必须抓住眼前的机会，真正重新构想NHS——减少医护人员的认知负担和行政负担，确保他们获得更好的工作体验，从而有能力提供最高质量的护理。”人工智能和Copilot等数字工具可通过减轻行政负担、释放临床能力，帮助NHS实现更高效、更互联、更以人为本的护理模式。

一线实践：从信息管理到“公平竞争”

约克郡救护车服务区域运营经理汤姆·史密斯从多个角度见证了这项工作。他负责管理运营经理和一线团队、规划人员配置和专业资源，并作为战术指挥官亲自应对重大事件。2025至2026财年，该服务部门接到约125万通999报警电话。他说：“什么情况都有，每天都不一样。”

史密斯需要阅读和撰写大量报告。他表示，Copilot并不替他做决策，而是将信息整理成他可以凭借运营经验进行审阅和质询的格式。“它只是把所有信息以有用的格式呈现给我，易于理解。它帮我完成了许多以前非常耗时的行政任务，我只需审核它完成的内容。”

对患有阅读障碍的史密斯而言，这种支持远不止节省时间。他说，写一封邮件时，如果过度纠结措辞或拼写，耗时会更长。Copilot帮助他搭建草稿结构、总结长篇报告并追问后续问题，这帮他“拉平了起跑线”。

严格的防护措施与信任框架

普莱斯-福布斯表示，NHS对Copilot的使用设定了明确而严格的保障措施：仅用于行政任务，不涉及临床工作。NHS正通过临床安全、信息治理、网络安全、互操作性和实际评估等标准支持数字化转型，并将透明度和人工监督视为维持员工和公众信任的关键。

与许多国家的医疗体系一样，NHS服务的是老龄化人口，许多患者患有多种慢性疾病。普莱斯-福布斯说：“这意味着我们有机会进行重新设计，因为我们无法继续沿用旧的模拟式医疗服务模式。这是全球性的挑战。我们现在需要围绕人、技术和数据重新设计护理，而人工智能将是创建这些新服务模式的唯一途径。”

他强调：“如果做得好，人工智能不会取代NHS的人文核心，而是会保护它——强化决策、恢复护理时间，帮助创建一个更安全、更人性化、更可持续的医疗服务体系。世界已经改变，我们必须为NHS打造一扇数字化的大门。”

普莱斯-福布斯还表示，信任是成功的关键。“患者和员工必须知道何时使用了人工智能、它如何影响决策、何处适用人工监督，并对重大决策有明确的责任归属。”

从数据中寻找信号

约克郡救护车服务首席数字信息官萨曼莎·罗宾逊负责帮助信息在一线团队、地方运营部门和更广泛的NHS之间流动。她说，Copilot“赋予我们串联线索的能力，否则我们需要花好几个小时翻遍数据报告、寻找共同主题才能发现”。

罗宾逊表示，NHS的信息流向四面八方。她可能需要将救护车服务各处的运营信息整理成给NHS高层领导的简报，有时又需要将国家指导方针转化为面向不同团队的清晰实用的指令。Copilot可以帮助生成初稿，再由罗宾逊和同事审核并针对目标受众进行调整。“我们有大量数据和信息需要传达给员工、一线急救人员和呼叫中心，Copilot帮助将这些材料转化为真正易于消化、人人都能理解的信息。”

寻找最合适的护理路径

这些信息在员工考虑如何将患者与最合适的护理对接时至关重要。并非每位被救护车救治的患者都需要送往急诊科。根据病情不同，有些人可以在家中安全接受治疗或转介至其他服务。

约克郡救护车服务正在推出一项与社区团队合作的计划，包括一个护理转移中心，为需要非紧急治疗的患者安排居家护理和转诊路径。目标是帮助患者在最适合的环境中接受照护，同时减少急诊科不必要的压力。史密斯说：“我们正努力将送往急诊科的转运率降至50%。”目前已有进展——Copilot帮助该服务部门整理数据、发现改进机会，转运率已从60%改善至56%。他说：“Copilot帮助我真正理解数据，能够将数据分享到不同论坛，并作为参考工具——比如‘如果我们这样做，你认为影响会是什么？’”团队成员因此拥有了更好的信息来支持患者路径决策。史密斯说：“最重要的是让患者获得最佳结果，这才是我们真正努力推动的。”

中文翻译：

救护车到达之前：AI如何助力英国国民医疗服务体系重新思考患者护理

当英国民众拨打999呼叫救护车时，他们的注意力理所当然地集中在救援何时到达上。而在此之前很久就已进行的规划、培训、分析和协调工作，则不那么为人所见。

“当救护车到达现场时，我们都会关注它，但在救护车 crew 载着患者到达之前，有大量工作已经完成，”约克郡救护车服务国民医疗服务体系信托基金首席临床信息官詹姆斯·古尔丁表示。该信托基金是英国国民医疗服务体系200多个信托基金（即组织单位）之一。

这些准备工作包括规划服务、招募和培训员工、分析人口健康数据，以及了解那些英语不流利的患者的体验。

“为了让急救人员到达患者面前时能够轻松地为患者做出正确决策，我们做了大量的工作，”古尔丁说。

约克郡救护车服务首席临床信息官詹姆斯·古尔丁。照片由乔安妮·克劳福德为微软拍摄。

在先前涉及3万名员工的试验之后，英国国民医疗服务体系英格兰分部于2026年6月宣布，计划向多达50.5万名临床医生和支持人员提供Microsoft Copilot。试验参与者报告称，平均每个工作日节省43分钟——相当于每人每年约五周的时间。

英国国民医疗服务体系英格兰分部国家首席临床信息官亚历克·普莱斯-福布斯博士表示，国民医疗服务体系在许多方面仍是一个模拟系统。“至关重要的是，我们要利用眼前的机会真正重新构想国民医疗服务体系，”他说，“减少医疗专业人员的认知负担和行政负担，确保他们获得更好的体验，从而能够提供最高质量的护理。”

AI和Microsoft Copilot等数字工具可以通过减轻行政负担、释放临床能力、实现更高效、更互联、更以人为本的护理模式来帮助维持国民医疗服务体系的运转。

帮助国民医疗服务体系团队管理信息

汤姆·史密斯从多个角度看待这项工作。作为约克郡救护车服务的区域运营经理，他负责监督运营经理和一线团队，规划人员配置和专家资源，并作为战术指挥官亲自应对重大事件。该服务在2025-26财年接到了约125万个999紧急呼叫。

“什么事都有，”他说，“没有一天是相同的。”

史密斯还需要阅读和制作大量报告。他说，Microsoft Copilot不会替他做决定。相反，它将信息整理成一种他可以运用运营专业知识进行审阅和质询的格式。

“它所做的就是以有用的、易于理解的格式给我提供所有信息，”史密斯说，“它为我完成了大量以前非常耗时的行政任务。然后我只需审阅它完成的工作。”

汤姆·史密斯，约克郡救护车服务区域运营经理。照片由埃斯梅·迈为微软拍摄。

对于有阅读障碍的史密斯来说，这种支持不仅仅是节省时间。他说，当他开始过度思考措辞或拼写时，写一封电子邮件可能会花费更长时间。能够构建草稿、总结长篇报告并提出后续问题的能力，有助于为他“创造公平的竞争环境”。

他的经历正值英国国民医疗服务体系英格兰分部在英格兰全境扩大生成式AI工具的访问权限之际。

重新构想国民医疗服务体系的运作方式

普莱斯-福布斯表示，国民医疗服务体系对其使用Microsoft Copilot制定了明确而严格的保障措施：它仅用于行政任务，不用于临床工作。国民医疗服务体系正在通过临床安全、信息治理、网络安全、互操作性和实际评估方面的标准来支持其数字化转型，并将透明度和人工监督视为维持员工和公众信心的关键。

与许多国家的医疗系统一样，国民医疗服务体系服务于老龄化人口，许多患者患有多种慢性疾病。

普莱斯-福布斯说：“这意味着我们有机会重新设计，因为我们无法继续沿用旧的模拟式医疗服务提供模式。这是全球面临的挑战。因此，我们现在需要围绕人和技术、数字和数据重新设计护理，而AI将是创建这些新服务护理模式的唯一途径。”

“人工智能可以变革国民医疗服务体系——改善护理、将时间还给员工、创造更大价值——但前提是我们将雄心转化为大规模的可信行动，”他说。“包括Microsoft 365 Copilot在内的经过验证的工具，必须迅速超越孤立的试点，进入日常 workflows，只要证据表明它们能改善结果、生产力或员工体验。”

“如果做得好，AI不会取代国民医疗服务体系的人性核心；它会保护它——加强决策、恢复护理时间，帮助创建一个更安全、更个性化、更可持续的医疗服务体系。世界已经改变。我们必须为国民医疗服务体系打造一扇数字化前门。”

亚历克·普莱斯-福布斯博士，英国国民医疗服务体系英格兰分部国家首席临床信息官。照片由乔安妮·克劳福德为微软拍摄。

他说，这种转型超越了技术层面；它要求重新设计工作方式，使AI能够帮助临床医生放大其专业知识，同时减轻他们不必要的负担。

这将解放“临床医生，使他们能够专注于判断、验证和人际连接。”

他还表示，信任是成功的关键。

“患者和员工必须知道何时使用了AI、它如何影响决策以及在哪里适用人工监督，并对重大选择有明确无误的问责机制，”他说。

普莱斯-福布斯说，现在的挑战是利用数字能力重新设计路径、改善流程、支持临床决策并将能力释放回护理中。这意味着要快速推进，但要有纪律。

“如果做得好，AI不会取代国民医疗服务体系的人性核心；它会保护它——加强决策、恢复护理时间，帮助创建一个更安全、更个性化、更可持续的医疗服务体系。世界已经改变。我们必须为国民医疗服务体系打造一扇数字化前门。”

在噪声中寻找信号

Copilot“赋予我们串联信息点的能力，如果没有它，我们自己需要花费数小时费力浏览数据报告并试图寻找共同主题才能发现这些联系。”

萨曼莎·罗宾逊，约克郡救护车服务首席数字信息官。照片由乔安妮·克劳福德为微软拍摄。

约克郡救护车服务首席数字信息官萨曼莎·罗宾逊负责帮助信息在一线团队、地方运营部门和更广泛的国民医疗服务体系之间流动。

Copilot“赋予我们串联信息点的能力，如果没有它，我们自己需要花费数小时费力浏览数据报告并试图寻找共同主题才能发现这些联系，”她说。

罗宾逊说，国民医疗服务体系的信息向各个方向流动。她可能需要将整个救护车服务的运营信息转化为面向国民医疗服务体系高级领导人的简报。在其他时候，她必须将国家指导方针转化为面向不同团队的清晰、实用的指令。Copilot可以帮助生成初稿，然后罗宾逊和她的同事审阅并根据目标受众进行调整。

“我们有大量的数据和信息需要传达给我们的员工、一线 crew 和联络中心，内容涉及国民医疗服务体系的运作情况，”罗宾逊说。Copilot帮助将这些材料转化为“真正易于消化的信息，以每个人都能理解的方式。”

找到合适的护理

当员工考虑如何将患者与最合适的护理对接时，这些信息至关重要。并非每位由救护车诊治的患者都需要被送往急诊科。根据病情不同，有些人可以在家中安全接受治疗或被转介到其他服务。

约克郡救护车服务正在推出一项与社区团队合作的计划，包括一个护理转移中心，为需要非紧急类型治疗的患者安排居家护理和转介路径。目标是在最合适的环境中支持患者，同时减少急诊科可避免的压力。

“我们正努力将送往急诊科的转运率降至50%，”史密斯说。进展正在取得；Copilot帮助约克郡救护车服务整理了数据并了解了机会所在，这支持了转运率从60%改善到如今的56%。

“Copilot在这方面帮助我的是真正理解数据，能够将数据分享到不同论坛，并将其作为一个大的参谋平台，”他说。“比如，‘如果我们这样做，你认为影响会是什么？’”

因此，团队成员拥有了更好的信息来为患者路径决策提供依据。

“最重要的是获得最佳患者结果，”史密斯说，“这才是我们真正努力推动的。”

英文来源：

Before the ambulance arrives: How AI is helping the NHS rethink patient care
When someone in the UK dials 999 for an ambulance, their focus is understandably on when help will arrive. Less visible is the planning, training, analysis and coordination that happen long before a paramedic reaches the patient.
“We all think about the ambulance when it’s on scene, but there’s so much that happens before that goes on before an ambulance crew even arrives with a patient,” says James Goulding, chief clinical information officer at Yorkshire Ambulance Service (YAS) NHS Trust. YAS is one of 200 plus trusts, or organizational units, of the UK’s National Health Service.
That preparation includes planning services, recruiting and training staff, analyzing population health data and understanding the experiences of patients who do not speak English fluently.
“So much goes into making sure that when that paramedic arrives in front of a patient, that we’ve done all of this work to make it really easy for them to make good decisions for patients,” Goulding says.
James Goulding, chief clinical information officer at Yorkshire Ambulance Service. Photo by Joanne Crawford for Microsoft.
Following an earlier trial involving 30,000 employees, NHS England announced plans in June 2026 to make Microsoft Copilot available to as many as 505,000 clinicians and support staff. Participants in the trial reported saving an average of 43 minutes per working day — which adds up to about five weeks of time per person a year.
Dr. Alec Price-Forbes, national chief clinical information officer for NHS England, says that the NHS is still in many ways an analog system. “It’s critically important that we use the opportunity that we have in front of us to truly reimagine the NHS,” he says, “to reduce cognitive load and administrative burden for healthcare professionals, make sure they get a better experience and then are in the position to give the best quality care.”
AI and digital tools such as Microsoft Copilot can help sustain the NHS by reducing administrative burden, releasing clinical capacity and enabling more productive, connected and person-centered models of care.
Helping NHS teams manage information
Tom Smith sees that work from several angles. As an area operations manager for Yorkshire Ambulance Service, he oversees operational managers and frontline teams, plans staffing and specialist resources, and responds personally to significant incidents as a tactical commander. The service received about 1.25 million calls to its 999 number in the 2025-26 fiscal year.
“It’s everything and anything,” he says. “No day is the same.”
Smith also has to read and produce many reports. Microsoft Copilot does not make decisions for him, he says. Instead, it organizes information into a format he can review and interrogate using his operational expertise.
“What it does is just give me all the information in a helpful format, which is easy to understand,” Smith says. “It does a lot of my admin tasks for me that would previously have been really time consuming. Then I can just review what it’s done.”
Tom Smith, area operations manager for Yorkshire Ambulance Service. Photo by Esme Mai for Microsoft.
For Smith, who is dyslexic, that support extends beyond saving time. Writing an email can take longer when he begins to overthink the wording or spelling, he says. The ability to structure a draft, summarize a long report and ask follow-up questions helps “level the playing field” for him.
His experience comes as NHS England expands access to generative AI tools across the health service in England.
Reimagining how the NHS works
The NHS has put clear, strict safeguards around its use of Microsoft Copilot: it is only for administrative tasks, not clinical work, says Price-Forbes. The NHS is supporting its digital transformation with standards for clinical safety, information governance, cybersecurity, interoperability and real-world evaluation, with transparency and human oversight considered essential to maintaining the confidence of staff and the public.
Like healthcare systems in many countries, the NHS serves an aging population, and many patients have multiple chronic illnesses.
Price-Forbes says, “It means that we’ve got the opportunity to redesign because we just cannot continue with the old analog model of healthcare delivery. And this is a challenge across the globe. So we now need to redesign care around people and technology, digital and data, and AI is going to be the only way to create those new service models of care.”
“Artificial intelligence can transform the NHS — improving care, giving time back to staff and delivering greater value — but only if we turn ambition into trusted action at scale,” he says. “Proven tools, including Microsoft 365 Copilot, must move rapidly beyond isolated pilots and into everyday workflows wherever evidence shows they improve outcomes, productivity or staff experience.”
“Done well, AI will not replace the human heart of the NHS; it will protect it — strengthening decisions, restoring time for care and helping create a safer, more personal and sustainable health service. The world has changed. We’ve got to create a digital front door for the NHS.”
Dr. Alec Price-Forbes, national chief clinical information officer for NHS England. Photo by Joanne Crawford for Microsoft.
He says that this transformation goes beyond technology; it requires that work be redesigned so that AI can help clinicians amplify their expertise while relieving them of unnecessary burdens.
This will free “clinicians to focus on judgment, validation and human connection.”
He also says that trust is essential for success.
“Patients and staff must know when AI is being used, how it shapes decisions and where human oversight applies, with unambiguous accountability for consequential choices,” he says.
The challenge now is to use digital capability to redesign pathways, improve flow, support clinical decision-making and release capacity back into care, Price-Forbes says. That means moving at pace, but with discipline.
“Done well, AI will not replace the human heart of the NHS; it will protect it — strengthening decisions, restoring time for care and helping create a safer, more personal and sustainable health service. The world has changed. We’ve got to create a digital front door for the NHS.”
Finding the signal in the noise
Copilot “gives us that ability to join dots that we wouldn’t see ourselves without a laborious several hours’ worth of trawling through data reports and trying to look at common themes.”
Samantha Robinson, chief digital information officer at Yorkshire Ambulance Service. Photo by Joanne Crawford for Microsoft.
Samantha Robinson, chief digital information officer at Yorkshire Ambulance Service, is responsible for helping information move between frontline teams, local operations and the wider NHS.
Copilot “gives us that ability to join dots that we wouldn’t see ourselves without a laborious several hours’ worth of trawling through data reports and trying to look at common themes,” she says.
The NHS’s information flows in all directions, Robinson says. She may need to turn operational information from across the ambulance service into a briefing for senior NHS leaders. At other times, she must translate national guidance into clear, practical instructions for different teams. Copilot can help produce a first draft, which Robinson and her colleagues review and adapt for the intended audience.
“We have a huge amount of data and information that we need to get out to our staff, frontline crew and contact center about how the NHS is performing,” Robinson says. Copilot helps turn that material into “really digestible information and in a way everyone can understand.”
Finding the right care
That information matters when staff are considering how to connect patients with the most appropriate care. Not every patient attended by an ambulance needs to be taken to an emergency department. Depending on their condition, some people may be treated safely at home or referred to another service.
Yorkshire Ambulance Service is introducing a program with community teams, including a transfer-of-care hub that arranges for home care and referral pathways for patients who need non-urgent types of treatment. The aim is to support patients in the most appropriate setting while reducing avoidable pressure on emergency departments.
“We’re trying to get our conveyance to the emergency department down to 50%,” Smith says. Progress is being made; Copilot has helped Yorkshire Ambulance Service to collate the data and understand where opportunities could be, which has supported an improvement from 60% to 56% conveyance today.
“How Copilot has helped me with that is really understanding the data, being able to share that data to different forums, to use it as a big of a sounding board as well,” he says. “Like, ‘if we did this, what do you think the impact would be?’”
As a result, team members have better information to inform decisions about patient pathways.
“The big bit is getting the best patient outcome,” Smith says. “That’s what we’re really trying to drive.”