Not surprisingly most people adopt a blank look when I explain we’re using a Bayesian Artificial Neural Network to predict hypotensive episodes.
When I try to explain they can get the idea of how drops in blood pressure can damage organs. The example of air leaking from a tyre does the trick.
But the BANN is a totally different matter. The concept of a computer identifying patterns which it then uses for prediction when humans have no idea of what it’s seeing is the stuff of SciFi novels to me, and I’ve seen it working.
But we aren’t alone in thinking this stuff has real value it can offer clinicians.
Promedas is providing similar technology for a wide range of diagnostic applications including:
* Infectious diseases, tropical medicine and sexually transmitted diseases
* Genetic disorders
* Clinical immunology
* Gastrointestinal disease including Gall bladder and biliary system
* Haematological disease
* Malignant disease
* Rheumatology and bone disease
* Renal disease
* Water, electrolytes and acid-base balance disorders
* Cardiovascular disease
* Respiratory disease
* Intensive care medicine
* Adverse drug reactions and poisoning
* Endocrine disease
* Diabetes mellitus and other disorders of metabolism
* Disorders of the eye, the ear, the nose, the throat
* Neurological disease
* Psychological medicine
* Skin disease
Promedas is the biggest and fastest probabilistic medical diagnostic network in the World based on medical expert knowledge, acquired from the literature by our medical specialists. More than 3.500 diagnoses and 47.000 network connections provide a diagnosis in seconds.
The output of Promedas
Promedas provides for each patient involved:
* A rapid differential diagnostic list of possible diseases in the field of internal medicine
* Proposals for additional clinical tests
* Warnings for interactions between drugs used by the patient
* A list of drugs that may cause side effects, taking into account the age, gender, liver and renal function.
* Drugs that may be responsible for toxicity experience by the patient
Additional output can be delivered such as a disease description and links to the literature (or in The Netherlands, the so called “Diagnose Behandel Combinatie”)
Related articles by Zemanta
- Artificial intelligence helps diagnose cardiac infections (scienceblog.com)