How Fastclinic Intelligence is used responsibly
In short
Fastclinic Intelligence is the part of our software that uses machine learning to support clinical and operational work. It is designed to support professionals, not to replace them. This policy sets out the principles we follow when we build and run it, and what we expect from the organisations that use it.
This summary helps you find your way. The full text below is what applies.
Section 1. What this policy covers
Fastclinic Limited (“Fastclinic”) develops and operates software that may use machine learning and related modelling techniques to support clinical workflows, operational efficiency, analytics and the experience of using our products. We call these capabilities Fastclinic Intelligence, and each of them a Fastclinic Intelligence feature. This Fastclinic Intelligence Policy describes the principles and practices we apply when we design, deploy and maintain them across our enterprise healthcare platform.
This policy applies to the Fastclinic Intelligence features we provide as part of the Services. Customers remain responsible for lawful use, clinical governance and configuration within their organisations. This policy adds to your agreement, your data processing terms and the regulations that apply. It does not replace them.
Section 2. The principles we hold ourselves to
We commit to these principles:
- Fairness
- We seek to identify and reduce unjust bias that could disadvantage individuals or groups because of protected characteristics, or characteristics that are clinically irrelevant, as far as is technically feasible and appropriate to the clinical context.
- Transparency
- We provide documentation of the intended use, limitations and known risks of Fastclinic Intelligence features, suited to enterprise customers and, where relevant, disclosures to end users that we coordinate with the customer organisation.
- Accountability
- We keep internal roles and controls for managing the model lifecycle, change management, incident review and escalation.
- Safety and reliability
- We use testing, monitoring and guardrails in proportion to each feature’s risk, with closer scrutiny for outputs that could influence clinical or public health decisions.
Section 3. Qualified people stay in charge
Fastclinic Intelligence features are designed to support decisions. They are not autonomous substitutes for licensed healthcare professionals or for roles that are legally accountable in an organisation. Where outputs may affect patient care, we expect customers to enforce policies that require review by a qualified person before action is taken, except where automation has been explicitly approved by the appropriate clinical governance and the law.
We give users ways to override, correct or disregard model suggestions. We log certain interactions where needed for auditability and product improvement, subject to the NDPA 2023 and our contractual commitments on data minimisation.
Section 4. Testing for bias and fairness
Before a high-risk Fastclinic Intelligence feature becomes generally available, we carry out proportionate testing. This may include statistical evaluation across relevant subgroups where data is available and statistically meaningful, stress tests with simulated or red-team scenarios, and review by subject-matter advisers. We document known limitations and remaining risks.
Once a feature is deployed, we monitor for performance drift, data skew and unusual error patterns. We encourage customers to report suspected systematic bias through our support channels, so we can investigate it together with the right technical and clinical people.
Section 5. Protecting patient safety
Our development processes follow healthcare safety thinking. That includes classifying failure modes by severity, keeping rollback plans for model or configuration changes, and working with customers during incidents that may affect care.
Fastclinic Intelligence features are not validated as medical devices unless they are explicitly certified or registered as such in the relevant jurisdiction, and customers must not represent them otherwise. Regulatory classifications vary from place to place, and customers are responsible for compliance where they deploy them.
Section 6. How Fastclinic Intelligence uses data
Training, fine-tuning, evaluating and running Fastclinic Intelligence features rely on datasets that are subject to data minimisation, purpose limitation and lawful basis requirements under the NDPA 2023. We use customer production data to improve models only where the contract permits it and, where appropriate, with de-identification or aggregation, or in separate environments.
We keep records of where datasets came from, how they may be used, and which subprocessors are involved in the infrastructure behind Fastclinic Intelligence. Enterprise customers can see these records under confidentiality as part of due diligence.
Section 7. Showing how a result was reached
We provide documentation that describes, at a suitable technical depth, the intended use case, the inputs and outputs, performance metrics from validation, known failure modes, versioning and how updates are made. For some models, where their design supports it, we offer feature importance, attention visualisations, written rationales or citations to source records. These techniques have limits and may not fully capture how a complex model behaves.
Section 8. Third-party and open models
Where we integrate third-party foundation models or APIs, we carry out due diligence on the vendor, impose contractual data protection and security requirements consistent with the NDPA 2023, and limit data flows to what the subscribed feature needs. We monitor vendor notices for material changes that affect privacy or safety.
Section 9. How this fits the Nigeria Data Protection Act 2023
Processing of personal data in connection with Fastclinic Intelligence features complies with the Nigeria Data Protection Act 2023, including its principles of lawfulness, fairness, transparency, storage limitation, integrity and confidentiality. Data subjects can use their rights through the appropriate controller (often the healthcare organisation that is our customer), or directly with us where we act as controller, at contact@getfastclinic.com.
Section 10. Keeping this policy up to date
We review this policy, our risk assessments and industry standards from time to time. When changes to how we govern Fastclinic Intelligence materially affect customers, we will tell them through release notes, contractual notice channels or revised documentation. For questions about this policy, contact contact@getfastclinic.com.