The Evolution of AI in Clinical Decision Support for Acute Care
Table of Contents Help Others Discover – Click to Share! Facebook Twitter LinkedIn Table of Contents THE ADOPTION GAP: 70% of physicians cited difficulties integrating AI with existing systems as a barrier to adoption, and only 19% reported that AI supported clinical decision-making in their organisation. Introduction: Acute Care Needs More Than Artificial Intelligence 92.5% general and acute bed occupancy in Q4 2025/26. 70% of physicians cited difficulties integrating AI with existing systems as a barrier to adoption. 65% identified poor interoperability, 73% raised concerns about AI errors, and only 19% reported that AI supported clinical decision-making in their organisation. By April 2026, the NHS App was connected to every acute NHS trust in England, with around 41 million registered users and 64% of hospital appointments visible through the app. The numbers, at a glance: Metric Figure General and acute bed occupancy (Q4 2025/26) 92.5% Physicians citing AI integration difficulties as a barrier 70% Physicians identifying poor interoperability 65% Physicians raising concerns about AI errors 73% Physicians reporting AI supported clinical decision-making 19% NHS App registered users (April 2026) ~41m Hospital appointments visible through the NHS App 64% These figures highlight the challenge: hospitals need connected clinical systems, accessible patient information, effective patient portals and appropriate NHS App integration to support safe, coordinated acute care. The NHS 10 Year Health Plan sets an ambition for hospitals to become fully AI-enabled within the plan’s lifetime, alongside a wider shift from analogue to digital care. At Cellma, we align with this direction by supporting a connected approach to healthcare information, clinical workflows and digital access. The future of AI and clinical decision support system technology depends not only on algorithms, but also on reliable data, interoperable systems, clinical safety and technology that fits the way healthcare teams work. Why Acute Care Needs Connected Clinical Decision Support Acute care teams make time-critical decisions every day. When a patient arrives at A&E, clinicians may need to review medical history, current observations, medication, allergies, previous investigations and information from other care settings. They also need to coordinate assessments, referrals, transfers, bed allocation and discharge. When this information is held across systems that do not communicate effectively, staff may have to switch between screens, search for missing details or repeat administrative tasks. This can make handovers harder and delay access to the information needed for clinical review. An AI and clinical decision support system can help clinicians interpret relevant information, identify potential risks and support decisions when appropriately designed and implemented. However, its usefulness depends on the quality, completeness and accessibility of the underlying data. AI cannot reliably compensate for information that is missing, outdated or inaccessible. The Royal College of Physicians’ January 2026 findings reinforce the point. Alongside integration and interoperability barriers, physicians reported concerns about clinical errors, liability, model drift, bias and explainability. Safe implementation therefore requires clinical oversight, suitable governance, staff training and clear accountability, not simply the introduction of another digital tool. The Role of the NHS App and Patient Portals Patient access is another part of the connected care picture. The NHS App gives people access to services such as appointment information, prescriptions, test results and patient records. NHS England reported in April 2026 that around two-thirds of hospital appointments were visible through the app, with patients at every acute NHS trust in England able to access referral and appointment information. A connected patient portal can complement this national access by giving patients an appropriate digital route to information and services offered by their healthcare provider. Depending on implementation, this may include appointment details, selected health information, communications and access to relevant records. For acute care providers, the goal is to reduce avoidable information gaps between patients, clinicians and services. IMPORTANT NOTE: Patient portals and NHS App integration are not interchangeable: integration with the NHS App requires NHS onboarding, NHS Login where applicable, and compliance with the relevant standards for accessibility, usability, clinical safety and data privacy. Secondary-care appointment integration may also involve the Patient Care Aggregator, known as Wayfinder. When planned around clinical and operational requirements, patient-facing digital services can form part of a wider approach to connected care. They can sit alongside the clinical information and workflows that support assessment, communication and care coordination. What NHS Interoperability Means in Practice Interoperability is the ability of systems to exchange information and use it across organisational and software boundaries. For acute care, this means making relevant information available to authorised professionals when and where it is needed. An AI and clinical decision support system also depends on data being exchanged in a consistent, secure and usable format. NHS England identifies standards and services that support this exchange. Depending on the use case, relevant requirements and integrations can include: Standard / Service What It Supports FHIR UK Core and HL7 Standards supporting structured healthcare data exchange and interoperability REST APIs and NHS API standards Approaches for connecting digital services and exchanging data GP Connect Enables authorised staff to access specified GP clinical information across systems, subject to access rules and the relevant implementation Spine services Including the Personal Demographics Service (PDS) and Summary Care Record (SCR), where appropriate to the use case SNOMED CT Structured clinical terminology for recording clinical information ICD-10 and dm+d Standards used for disease classification and medicines and devices terminology, respectively, where applicable NHS number and ODS codes Supporting patient identification and organisation identification in relevant workflows CIS2 and role-based access control (RBAC) Relevant to secure access and identity management in NHS environments NHS App integration standards Requirements for eligible patient-facing services, including privacy, accessibility, usability and clinical safety MESH and other approved messaging or data-exchange routes Where required by the relevant NHS service and integration design IMPORTANT NOTE: Not every system requires every interface. The correct integration scope depends on the clinical use case, the sending and receiving systems, information governance, technical specifications and national service onboarding. NHS England advises that systems should use agreed APIs and FHIR UK Core where appropriate, alongside relevant