Intelligent inference engine–Infermedica

World-class medical inference engine

At Infermedica, we aim to connect people with the medical services they need. This is possible thanks to our AI‑powered inference engine that precisely analyzes patients' symptoms.

)

HOW IT WORKS

AI-led medical interview

Infermedica’s technology draws conclusions from patient data to construct dynamic medical interviews on the fly. Our AI asks the right questions to gather supplementary health information, better understand patients' state of health, and give appropriate recommendations. The process consists primarily of the following:

  1. Collecting initial evidence

    The patient interview starts with gathering initial symptoms, risk factors, and demographic data. Additional input regarding the symptoms' occurrence or severity makes the engine even more precise.

    • The engine is backed with a comprehensive library of thousands of medical conditions
    • It covers multiple medical specialties, including pediatrics
  2. Intelligent interview

    The inference engine uses initial evidence to construct a dynamic interview based on probabilistic models and reasoning techniques. It was developed by our data scientists, who collaborated closely with medical experts.

    • Algorithms follow the rules of differential diagnosis that physicians use to interview the patient
    • The engine verifies various conditions at once and is more effective than a fixed rule-based system
  3. Drawing conclusions

    Infermedica’s AI knows exactly when to stop asking questions. The inference engine observes correlations between patients’ symptoms and hundreds of diseases. With every answer from the user, it becomes more confident about the most probable conditions.

    • It’s one of the shortest pre-interviews on the market
    • Questions are generated dynamically during the interview
  4. Triage recommendations

    Finally, the inference engine presents the most probable causes of the symptoms, which are paired with the suggested level of care. Combined with the answers of the intelligent interview, recommendations are individual to each situation.

    • Our triage recommendations are based on a 5-level scale (self-care, consultation, consultation 24h, emergency, ambulance)
    • We provide recommendations for the level of care, type of medical specialist, communication channels, and more

)

Our engine’s abilities

Complexity

The inference engine is able to analyze numerous symptoms, conditions, risk factors, and demographics at once.

Adaptability

Small bits of evidence can make a big difference. Our engine adapts as it learns about patients’ complaints.

Instant results

Our engine can investigate multiple conditions in real time and provide recommendations in a matter of minutes.

Dynamic logic

Our AI asks fully dynamic questions—there is no fixed logic or prescribed path.

Blazing speed

It accelerates to reach high performance with an unmatched query response time of ~0.1 s!

Foundations of AI

Advanced modeling and seamless communication

Natural Language Processing

Developed by a team of NLP experts and leveraging the Medical Knowledge Base, our Inference Engine empowers patients to express their concerns using natural language effortlessly.

Probabilistic modeling

Scientists, engineers, and doctors have contributed to our algorithms and probabilistic models. These parts of our engine are based on modified Bayesian networks and machine learning methods.

See how we test engine accuracy in the series of blog articles →

27M+

successful health checkups

1+ million

structured patient cases

96%

conditions accuracy

)

KNOWLEDGE BASE

Medical quality ensured by physicians

Thousands of symptoms, risk factors, conditions, and the precise connections between them are sourced and verified by a team of experienced physicians. They use only proven and reliable sources of medical knowledge and follow scientific principles of clinical validation.

)

3000+

medical concepts

155,000+

hours of physician input

36

medical experts engaged

1800+

symptoms

900+

conditions

340+

risk factors

)

Jakub Jaszczak
Medical Doctor & Product Manager at Infermedica

PLATFORM FEATURES

Our engine’s extensions

The inference engine lies at the core of Infermedica’s platform. It can be complemented by various components and features that strengthen its symptom-assessing and patient-routing abilities. These elements can be selected and implemented according to a businesses’ unique needs.

)

How to use it

Incorporate our top-notch inference engine and its vast medical knowledge into your businesses’ virtual healthcare product. Check out our ready-to-use features or leverage the flexible API. Use it freely to build custom solutions, create multiple interfaces, scale up, or simply integrate it with your existing systems.

)

Discover the power of clinically validated AI