?url_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Adc&rft.relation=http%3A%2F%2Feprints.imtlucca.it%2F4079%2F&rft.title=Improving+the+Prediction+of+Clinical+Success+Using+Machine+Learning&rft.creator=Munos%2C+Bernard&rft.creator=Niederreiter%2C+Jan&rft.creator=Riccaboni%2C+Massimo&rft.subject=HA+Statistics&rft.subject=RM+Therapeutics.+Pharmacology&rft.description=In+pharmaceutical+research%2C+assessing+drug+candidates%E2%80%99+odds+of+success+as+they+move+through+clinical%0D%0Aresearch+often+relies+on+crude+methods+based+on+historical+data.+However%2C+the+rapid+progress+of%0D%0Amachine+learning+offers+a+new+tool+to+identify+the+more+promising+projects.+To+evaluate+its+usefulness%2C%0D%0Awe+trained+and+validated+several+machine+learning+algorithms+on+a+large+database+of+projects.+Using%0D%0Avarious+project+descriptors+as+input+data+we+were+able+to+predict+the+clinical+success+and+failure+rates%0D%0Aof+projects+with+an+average+balanced+accuracy+of+83%25+to+89%25%2C+which+compares+favorably+with+the+56%25%0D%0Ato+70%25+balanced+accuracy+of+the+method+based+on+historical+data.+We+also+identified+the+variables+that%0D%0Acontributed+most+to+trial+success+and+used+the+algorithm+to+predict+the+success+(or+failure)+of+assets%0D%0Acurrently+in+the+industry+pipeline.+We+conclude+by+discussing+how+pharmaceutical+companies+can+use%0D%0Asuch+model+to+improve+the+quantity+and+quality+of+their+new+drugs%2C+and+how+the+broad+adoption+of%0D%0Athis+technology+could+reduce+the+industry%E2%80%99s+risk+profile+with+important+consequences+for+industry%0D%0Astructure%2C+R%26D+investment%2C+and+the+cost+of+innovation.&rft.date=2020-10&rft.type=Working+Paper&rft.type=NonPeerReviewed&rft.format=application%2Fpdf&rft.language=en&rft.rights=cc_by_nc&rft.identifier=http%3A%2F%2Feprints.imtlucca.it%2F4079%2F1%2FWP_3_2020.pdf&rft.identifier=++Munos%2C+Bernard+and+Niederreiter%2C+Jan+and+Riccaboni%2C+Massimo++Improving+the+Prediction+of+Clinical+Success+Using+Machine+Learning.++EIC+working+paper+series++%233%2F2020++++++++++