When people take prescribed medication from health professionals, they usually assume that the prescriptions are correct and that these should help them feel better.
However, medication errors are in fact relatively common in general practice and in hospitals. Both errors in the act of writing (prescription errors) and prescribing faults due to incorrect medical decisions can result in harm to patients. According to the Pharmaceutical Journal, prescribing error rates of 8.9 errors per 100 medicine orders have been observed in hospitals and in 4.9% of all prescription items in general practice.
Posos, a startup in the Microsoft AI Factory at STATION F, is building a solution to prevent these common but preventable errors. Their algorithm allows health professionals to pinpoint the most relevant information in medical literature and to choose the most appropriate treatment for their patients. Unlike existing language-based tools, Posos adopts a semantic approach, valuing context and meaning of words over research of keywords.
The Posos team currently counts about 12 people – pharmacists and natural language processing engineers – and aims to double in size within a year. To develop a highly performing tool suitable for all therapeutic areas, they need to process an enormous volume of medical literature in French, in English, and in many other languages. They have recently closed a funding round of €2M to bolster their growth. Posos aims to provide their solution to public and private actors in the health sector in France and abroad with a focus on the USA.