Policies and frameworks

This plan is adopted by the Association Board, which is responsible for its implementation. The Ethics, Law & Governance working group supports its follow-up.

Gender equality is part of the regular work of the Board and the working groups. Expertise is drawn from our Fellows and network, including research on gender aspects of AI in medicine.

We collect and report, once a year, anonymised and aggregated sex- and gender-disaggregated data on the Board, Fellows, Convenors, working group members and volunteers, and on authorship of hub publications.

We provide information on gender equality and unconscious bias to the Board, Convenors and working group members.

1. Organisational culture and work–life balance: flexible, remote and asynchronous participation, so that contributing is possible alongside work, studies and care responsibilities.
2. Gender balance in leadership and decision-making: we aim for gender balance on the Board, among Fellows and among Convenors.
3. Equality in selection and progression: transparent and fair criteria for Fellows, Convenors, volunteers and co-authorship (ICMJE).
4. The gender dimension in research and in the platform: sex and gender are considered in study design, analysis and AI development, and the platform records biological sex for clinical purposes alongside optional gender identity and pronouns.
5. Zero tolerance of harassment: zero tolerance of gender-based violence, sexual harassment and discrimination, with a confidential way to raise concerns through our contact form.

• Platform data: personal health records created and controlled by patients.
• Research data: data collected in approved studies, for example surveys, interviews and analyses.
• Publications, software and documentation.

Health data is structured with international classifications: ICD-11 for diagnoses, ICHI for procedures and ATC for medicines, with exchange through HL7 FHIR and the International Patient Summary as explicit goals.

Data is stored within the EU (Frankfurt, Germany), encrypted in transit, with access controls and logging. Platform data is never used for research without the patient's explicit consent.

Health data is processed in line with the GDPR. Research using personal data requires explicit consent or another valid legal basis, a data protection impact assessment, and approval from the Swedish Ethical Review Authority (or the equivalent body in other countries).

We collect only the data needed for each purpose. Research data is pseudonymised or anonymised wherever possible.

Research outputs follow the FAIR principles (Findable, Accessible, Interoperable, Reusable). Publications are open access wherever possible. Code will be released under an open licence. Personal data is never published.

Retention periods are set for each study and agreed with the ethics approval. Patients can export and permanently delete their platform data at any time.

The Research Director is responsible for data management. Each study has a named data steward. Questions can be sent through our data protection form.

The GlobeDoc platform currently uses no AI models that process health data. The only automated service is machine translation of fixed interface text, which never includes health information.

AI will be developed and researched at GlobeDoc Research Hub. All AI work will follow:

• The EU Artificial Intelligence Act. AI intended for medical purposes is expected to be classified as high-risk, with requirements for risk management, data governance, technical documentation, transparency and human oversight.
• The WHO guidance on the ethics and governance of artificial intelligence for health.

1. Human oversight: AI supports, and never replaces, the judgement of patients and clinicians.
2. Transparency: users will always know when AI is used and what it does.
3. Consent: patient data is never used to train or test AI without explicit consent.
4. Fairness: AI is tested for bias across sex, gender, age, ethnicity and geography.
5. Privacy by design: we pursue privacy-preserving methods such as federated learning and homomorphic encryption.
6. Documentation and accountability: every AI model is documented, versioned and evaluated before use.

GlobeDoc is a research platform for patient co-owned health records. It is not a certified medical device and is not intended for diagnosis or treatment.

Our explicit goal is for the GlobeDoc platform, and any future clinical functions, to meet the requirements of the EU Medical Device Regulation (MDR) and to obtain CE marking where required, and to meet the requirements for electronic health record systems under the European Health Data Space (EHDS).

1. Define the intended purpose of each function and its classification under the MDR.
2. Establish a quality management system in line with ISO 13485.
3. Software lifecycle in line with IEC 62304, risk management in line with ISO 14971, and usability in line with IEC 62366-1.
4. Information security certification in line with ISO/IEC 27001.
5. Clinical evaluation and assessment by a notified body for functions classified as medical devices.
6. Alignment with the EHDS requirements for interoperability and security of electronic health record systems.

The roadmap is led by the Technology & Security and the Ethics, Law & Governance working groups.

• Digital-first collaboration: remote meetings by default, and travel only where it adds clear value.
• Responsible hosting: we assess the energy efficiency and renewable energy use of our hosting providers, with preference for sustainable EU-based options.
• Efficient software and data minimisation: we store only the data needed and design the platform to be lightweight, which also supports low-bandwidth settings.
• Responsible equipment: long use of hardware and responsible recycling.