Precision medicine development
Our customer is a municipal US hospital that is regularly organizing and hosting clinical trials.
The typical drug dosing protocol is developed during randomized clinical trials – an expensive and time-consuming process that also is limited from the patient variety point of view. Moreover, due to the highly individual drug susceptibility for different people, the protocol often does not work as expected. Some patients react differently to the baseline; others do not react at all. Personalized drug prescriptions are becoming a new standard that is maximizing the effectiveness of medicine and reducing patient risks.
Algorithmically, a combination of classic statistical learning and deep learning approaches was used to stratify patients into drug susceptibility cohorts based on their phenotypes and provide them with strata-specific recommendations for optimal treatment. Over that, the cloud data platform was developed for the aggregation of clinical data for patients under observation.
- Standard protocol with doses and limits was used to calculate a drug dosage for each patient
- Statistical stratification unlocked much shorter active compound expenditure – 24% less than baseline and personalized drug prescriptions