Our technology, unlike other emerging AI solutions, has been developed and tested over 20 years. We are unique in offering the industry’s most validated and accurate results.
We don’t just predict if drugs will work, we predict how well, with which other drugs, at what dose and disease stage, in which populations.
Beginning in 2014, we conducted in silico trials:
Millions of patients predictions.
Thousands of clinical response parameters per patient per day.
Different doses, different genetic backgrounds.
Following our predictions, doctors held clinical trials.
9 prospective clinical trials.
29,111 human patients.
Hundreds of daily response parameters per patient.
Our predictions of patient response parameters, per patient, per day were within 99.94% of those observed in the 29,111 patients.
Any predictions provided outside of confidentiality agreements were published in open access journals.
Our 2019 work predicting TB drug cure rates guided the selection of a single clinical trial dosage to achieve cure, confirmed in human trials 2 years later and published in 2024. This is one of several predictions validated in subsequent human trials.
Phase Advance made predictions in lung infections, metabolic syndrome, and liver disease, with hundreds of thousands of in silico patients in each prediction, and thousands of data points and responses generated per patient. Phase Advance’s forecast of those data points at time 1 (t1) was compared to the data points per patient observed in clinical trials later (29,111 patients in total) at time 2 (t2) of either dose or clinical response rates.
Our pooled weighted accuracy was 99.94% (95% credible intervals: 99.52-100%).
