Input data and information sources
The system analyses individual patient records. Vital signs, laboratory values and diagnoses are health data.
Clinical decision support (such as early sepsis detection)
Also known as: Clinical decision support and early warning systems
The system continuously analyses vital signs, laboratory values and patient records and alerts the care team to a possible deterioration.
Understand the use case
What defines this use case?
An early warning system calculates a risk score from the patient record and triggers an alert above a threshold. US hospitals use such systems for early sepsis detection. At Michigan Medicine, however, a widely used model performed substantially worse than its manufacturer's figures.
This page describes a typical system design. Actual data flows and permissions depend on your implementation.
- System design
- Analysis and processing
- Ability to act
- Action permissions depend on integration
- Fields of use
- healthcare
- Also known as
- early warning system, sepsis prediction, clinical risk prediction
How it works, step by step
System design and data flows
Input data
Vital signs, laboratory values, diagnoses and entries arrive continuously from monitoring, laboratories and patient records.
Some models also analyse free text from clinical letters or nursing reports. The system analyses individual patient records.
Processing
A model calculates each patient's risk, for example of sepsis.
The model processes the specified clinical information and may assess developments or anomalies. This produces alerts or prioritised notifications for the care team. A model warning, clinical assessment and treatment actually initiated are separate steps.
Results
An alert appears for the care team when the score exceeds a threshold.
The result describes the defined model task. A match, category or forecast acquires its professional significance only in conjunction with the underlying data and intended use.
Use and impact
The team reviews the alert and decides on examination and treatment; without an alert, this prompt is absent.
Integration determines whether the result is simply displayed, adopted after approval or written directly into a business application. These variants differ in their data flows and permissions.
Explore the system design: Select a component to see its task, incoming and outgoing data, and role in the use case. The guided walkthrough explains each phase. Then use Yes, Uncertain and No to represent your specific system variant.
6 review first11 clarify7 baseline threats
- Present
- Depends on your system
- Absent
- Review first
- To clarify
How to use the model
The diagram shows the typical architecture. Select a component to see what happens there and adapt its features to your system. Enable “Analysis guidance” to see which threats need reviewing for this architecture. The page's checklist continues to describe the typical case.
Changes from the typical case
None yet. Each change will appear here.
Where most threats arise
- 3 review first, 1 clarify
- 4 clarify
- 1 review first, 2 clarify
- 2 clarify
In this system design: 6 to review first, 11 to clarify, 7 baseline threats.
Architecture as a list
External access: Depends on your system
To clarify
Input: Part of every AI system
Review first
To clarify
Baseline for every AI system
AI model: Part of every AI system
Review first
- Model Drift & Degradation Review priority
Provider service: Depends on your system
To clarify
Training and feedback: Present
Review first
To clarify
Result: Part of every AI system
Baseline for every AI system
Business application or equipment: Depends on your system
To clarify
person: Part of every AI system
Review first
- Overreliance Review priority
Affected person: Present
Unapproved AI services: Depends on your system
To clarify
Supply chain and technical foundations: 5 baseline threats applying to every AI system
Baseline for every AI system
How it differs from similar use cases
AI-assisted medical image analysis
Analyses individual X-rays, CT, MRI or mammography images
AI support for emergency call handling
Monitors emergency calls and provides indications such as possible cardiac arrest
Data and information
Processing and operating environment
The Epic Sepsis Model belongs to the vendor's hospital information system; other models originate in research.
Storage and further use
Some hospitals develop or fine-tune the model using their own treatment data; others adopt their software provider's model. Confirmed and dismissed alerts may contribute to model development.
Results and data sharing
Depending on integration, the system only displays alerts or creates tasks or proposed orders in the hospital information system. Hosting the model with the provider sends health data outside the hospital.
Using a provider does not automatically determine where processing takes place. Storage location, subcontractors, deletion periods and use for training must be checked for the specific deployment.
Assess risks and determine measures
From understanding the system to analysing risks
- Define the system boundary
Record data, participants, interfaces and permissions. Resolve uncertain characteristics in the model.
- Investigate threats
Relate review candidates to actual data flows: what could happen, and who would be affected?
- Assess risk
Assess impacts and likelihood in the deployment context. The number of review candidates is not a risk score.
- Provide evidence of measures
Select appropriate controls, assign responsibilities and define effectiveness checks. Document residual risks and approvals.
Key threats in the application context
- 3 particularly important
- 3 typical
- 11 depending on the system
- 7 additional threats for every AI system
A total of 24 threats from the threat catalogue may be relevant to this use case.
Also review: whether the outputs are substantively correct
Check probability reliability, possible test-data leakage into training and the consequences of an error in the business process.
Also review: whether the system disadvantages groups
Check whether selection, ranking or assessment disadvantages people with protected characteristics, such as sex, age or origin.
Also review: Safety of people and equipment
Check independent safety functions, intended operating limits and the safety evidence your sector requires.
Priority review points
A model developed elsewhere encounters different patients, workflows and documentation practices and performs worse. At Michigan Medicine, measured performance fell well short of the manufacturer's claims.
A team relying on sepsis alerts may quickly consider an unflagged patient stable. Treatment then begins too late.
Sensitive Information Disclosure
Development and operation require complete treatment histories from many patients. Unauthorised access during training or at the provider exposes diagnoses and laboratory values.
3 further review candidates in the typical design
Other relevant threats
Targeted Poisoning / Label Poisoning
Falsified training labels can deliberately shift model outputs.
Listed because of this characteristic
- Learns from examples with correct answers
Targeted queries can recover some personal or protected training data from the model.
Listed because of these characteristics
- Processes personal data
- Processes confidential data
Exfiltration from ML Application
Protected data can leak through tools, network connections, storage or insiders.
Listed because of this characteristic
- Processes confidential data
Up to 11 further threats may apply, depending on your system design. Nine threats need to be reviewed for every AI system. Two of these are particularly important here and are therefore listed above.
System features and open questions
Each Yes adds further threats to your checklist, up to 11 in total. If you cannot yet answer a question, include its threats in your review.
- Added by this Yes
- Also requires a Yes to another question
How your system is built
- Learns from your own data2 threats, plus 1 only in combination with other characteristics
- Processes text, images and speech together2 threats
- Changes data or triggers processes1 threat
- Uses another provider's model1 threat
- Learns in operation without renewed approval1 threat
- Uses operational feedback1 threat
How you operate the system
- Available to people outside the organisation5 threats
- Accepts queries directly through an interface4 threats
- Accessible over a network2 threats, plus 1 only in combination with other characteristics
- Incurs costs with each query1 threat
- Shares computing resources with other customers or applications1 threat
- Used alongside unapproved AI services1 threat
How your system is built
Can the system change things itself, such as storing data, sending messages, initiating transactions or granting permissions?
Depending on integration, the system only displays alerts or creates tasks or proposed orders in the hospital information system.
If yes, add
Does your organisation train or fine-tune the model itself using its own data?
Some hospitals develop or fine-tune the model using their own treatment data; others adopt their software provider's model.
If yes, add
Does the model or its base model come from another provider, even if you host it?
The Epic Sepsis Model belongs to the vendor's hospital information system; other models originate in research.
If yes, add
Does the system keep learning in operation and change its behaviour without renewed approval?
If yes, add
Does operational feedback or output later feed into analysis or training?
Confirmed and dismissed alerts may contribute to model development.
If yes, add
Does the system process several input types together, such as text with images or speech?
Some models also analyse free text from clinical letters or nursing reports.
If yes, add
How you operate the system
These questions arise for almost every AI system.
Can people or programs query the model directly, for example through an interface?
If yes, add
Is the system accessible over a network or dependent on network services?
If yes, add
Do you pay per query or unit of computing time, so costs increase with use?
If yes, add
Does the system share servers or computing resources with other customers or applications?
If yes, add
Can people outside your organisation use the system, including without login?
If yes, add
Do staff also use AI services for this task that your organisation has not approved?
If yes, add
Controls and implementation
Measures for the priority review areas are expanded. Each linked measure explains how to implement it.
Review priorityModel Drift & Degradation3 measures
Review priorityOverreliance3 measures
Review prioritySensitive Information Disclosure3 measures
Targeted Poisoning / Label Poisoning3 measures
Privacy Attacks3 measures
Exfiltration from ML Application3 measures
Controls for the remaining threats are available on their pages in the threat catalogue.
AI Act provisions
-
Art. 4: AI literacy
For every AI system Applies since 2 February 2025.
-
Art. 6 para. 1, Annex I Section A: Safety component of an assessed product
In typical use Review trigger from 2 August 2028.
Applies to every AI system
Art. 4: AI literacy
Applies if staff or appointed representatives operate or use AI systems.
Applies since 2 February 2025.
Obligations and details for Art. 4: AI literacy
- Who bears the obligations
- Provider, Deployer, Product manufacturer, Provider of a general-purpose AI model
Detailed obligations
May apply in typical use
Art. 6 para. 1, Annex I Section A: Safety component of an assessed product
Applies if the system is placed on the market as a medical device under Regulation (EU) 2017/745 and requires notified body conformity assessment, as diagnostic and early warning software usually does under Rule 11.
Applies from 2 August 2028.
Obligations and details for Art. 6 para. 1, Annex I Section A: Safety component of an assessed product
- The general condition is
- Applies if the system is itself a product or a safety component of a product requiring third-party assessment under Annex I legislation, such as machinery, lifts, medical devices or toys.
- Exception
- Systems improving only operation, performance or comfort without a safety role are not safety components under Article 6(1a)–(1c).
- Who bears the obligations
- Provider, Deployer, Product manufacturer
Detailed obligations
Generally does not apply
Annex III no. 5 point d: Emergency calls and emergency care
A ward-based early warning system does not assess emergency calls or prioritise emergency patients. This may differ if an emergency department uses it for initial triage.
Obligations and details for Annex III no. 5 point d: Emergency calls and emergency care
- The general condition is
- Applies if the system evaluates or classifies emergency calls, dispatches responders or prioritises emergency patients.
- Who bears the obligations
- Provider, Deployer, Product manufacturer
Detailed obligations
Classify your system with the quick checker All obligations in the obligations compass
Cross-cutting threats
Review these threats for every AI system, whatever its purpose.
Atypical system features
View 17 further questions
If any of these features does apply to your system, add the listed threats.
Does the system generate its own text, such as answers, drafts or summaries?
The system produces a risk score and alert rather than generating text.
If yes, add
Does the system generate or edit images, sounds, voices or videos?
If yes, add
Does the system create new content from open-ended requests that people read, see or hear directly? Translation, reading aloud and transcription do not count.
If yes, add
Can users give the system natural language tasks that it then performs?
If yes, add
Does the system search its own collection of documents or records for relevant content?
If yes, add
Does the system formulate answers from retrieved documents?
If yes, add
Does the system use a knowledge graph storing concepts and their relationships?
If yes, add
Does the system remember content from earlier conversations or tasks and reuse it later?
If yes, add
Does the system call other programs, interfaces or search services itself, even if only to read?
If yes, add
Does the system determine its next steps itself based on earlier results?
If yes, add
Does the system delegate tasks to other AI agents or incorporate their results?
If yes, add
Is the system connected to tools or data sources through the Model Context Protocol (MCP)?
If yes, add
Does the system process content outsiders may have written or altered, such as emails, uploads or websites?
Data comes from the hospital's monitoring devices, laboratory and patient records.
If yes, add
Do documents from several customers, departments or permission groups share one search store?
If yes, add
Does the system generate code, scripts or database queries for further use?
If yes, add
Does the system execute programs, scripts or commands itself?
If yes, add
Is the system intended to maximise a specified metric, such as revenue, utilisation or clicks?
If yes, add
Sources
- Specialist study
The prospective study examines a machine learning early warning system used in five hospitals.
- Specialist study
Johns Hopkins University research portal, Entry for Adams et al.'s TREWS study, Nature Medicine 2022
The open-access entry summarises monitoring of 590,736 patients and lower mortality where alerts were confirmed promptly.
- Specialist study
The study tests the Epic Sepsis Model on 38,455 admissions and measures an AUC of 0.63, compared with the manufacturer's 0.76–0.83.
Related use cases
Also in the field of Healthcare and emergency services
Updated: 30 September 2026. Review checklist rules version 0.3.2, threat catalogue version 2026.10.24.2. How these pages are structured