From cure to prediction: the algorithmic transformation of healthcare

“A few years ago, we came togeth­er as part of a group of research­ers from the Man­age­ment Research Centre at Ecole Poly­tech­nique work­ing in the health­care sec­tor,” explains Étienne Min­vi­elle, CNRS research dir­ect­or (IP Par­is). “One of the object­ives was to think about how to bring tech­no­lo­gic­al innov­a­tions into dia­logue with the needs of the health­care sys­tem.” This meet­ing launched a series of sem­inars on algorithmic pre­ven­tion. “Two years ago, I per­son­ally didn’t know much about this top­ic,” he admits. “To tell the truth, I didn’t really see what could be said about it. How­ever, after hav­ing led these sem­inars, I real­ise how import­ant this sub­ject is for improv­ing prevention.”

Because, although ini­tially little was known about this sub­ject, even among pro­fes­sion­als in the field, these dis­cus­sions have high­lighted the fact that algorithmic pre­ven­tion affects almost all areas of medi­cine (onco­logy, geri­at­rics, psy­chi­atry, neur­o­logy, etc.).

From digit­al twins to the pre­ven­tion of epi­dem­ics, age­ing well and aug­men­ted psy­chi­atry, these sem­inars show that algorithmic pre­ven­tion is not lim­ited to a spe­cif­ic field. It paves the way for a sys­tem­ic trans­form­a­tion of medi­cine, link­ing tech­no­lo­gic­al innov­a­tions to soci­et­al issues. Health pre­ven­tion can now take two forms: con­ven­tion­al pre­ven­tion, which is aimed at a large group of the pop­u­la­tion, and so-called algorithmic pre­ven­tion, which is more per­son­al­ised. “Algorithmic pre­ven­tion dif­fers from con­ven­tion­al pre­ven­tion in that it is per­son­al­ised and accom­pan­ied by dynam­ic mon­it­or­ing,” says Étienne Min­vi­elle. “This involves the col­lec­tion of sub­stan­tial data on genet­ic, but also socio-eco­nom­ic and beha­vi­our­al, factors.”

___________________________________________________________________________________________

Read also: How digit­al tech­no­logy will per­son­al­ise healthcare

___________________________________________________________________________________________

Once this data has been col­lec­ted, coupled with a more in-depth know­ledge of the devel­op­ment of the dis­ease, it becomes pos­sible to make bet­ter pre­dic­tions. “For example, the Inter­cep­tion pro­gramme at Gust­ave Roussy is based on the obser­va­tion that 40% of severe forms of can­cer could have been iden­ti­fied earli­er by these algorithmic pro­cesses,” he explains. “Tests are thus car­ried out to identi­fy genet­ic poly­morph­isms, i.e. com­bin­a­tions of genet­ic muta­tions, in people iden­ti­fied as being at risk. Coupled with an ana­lys­is of envir­on­ment­al risk factors, they make it pos­sible to pre­dict the risk of can­cer occur­ring and to “inter­cept” it even before it can devel­op, thanks to per­son­al­ised monitoring.” 

40% of can­cers could there­fore be avoided, as Fabrice Bar­lesi, CEO of Gust­ave Roussy, is well aware: “Once we know this, we can­not fail to recog­nise the import­ance of pre­ven­tion. But we must also ask ourselves why pre­ven­tion is not work­ing today – smoking is a case in point. Moreover, our screen­ing pro­grammes, i.e. early detec­tion of the dis­ease, are also strug­gling. To rem­edy this, we will need to be able to identi­fy the people at highest risk with a view to inter­cept­ing this disease.”

A sim­il­ar pat­tern is found in the pre­ven­tion estab­lished by the ICOPE pro­gramme in its quest for healthy age­ing, in the pre­ven­tion of cog­nit­ive decline (such as with Alzheimer’s dis­ease), and in oth­er con­di­tions (car­di­ology, men­tal health, well-being).

How­ever, examples of the applic­a­tion of this type of pre­ven­tion high­light its depend­ence on our sci­entif­ic, tech­no­lo­gic­al and organ­isa­tion­al advances. “Today, we can see that innov­a­tion is a major lever for con­trib­ut­ing to an effect­ive pre­ven­tion policy,” says Lise Alter, former dir­ect­or gen­er­al of the Health Innov­a­tion Agency. “And, between an age­ing pop­u­la­tion, which means an increase in the pre­val­ence of chron­ic dis­eases, the diverse and var­ied chal­lenges of cli­mate change, and the lim­it­ing factor of human resources in the health­care sec­tor, we are going to have to face major chal­lenges that will require the trans­form­a­tion of our health­care system.”

And it is these major chal­lenges that make the prom­ises of algorithmic pre­ven­tion so appeal­ing. “When we talk about “trans­form­a­tions” it means “changes” in our fin­an­cing mod­els, which are mainly based on cur­at­ive rather than pre­vent­ive care. Eval­u­ation and also demon­stra­tion of value – requir­ing applic­a­tion on a pop­u­la­tion scale suf­fi­cient to have a power of demon­stra­tion.” Above all, the demon­stra­tion of effect­ive­ness must not stop at the clin­ic­al aspect but must also focus on the impact of such a change on the organ­isa­tion of care or on the qual­ity of life of health­care per­son­nel. “These are there­fore much broad­er con­sid­er­a­tions than the simple clin­ic­al impact on the patient, even if this remains a fun­da­ment­al point.”

For his part, Nic­olas Rev­el, Dir­ect­or Gen­er­al of AP-HP, Assist­ance Pub­lique – Hôpitaux de Par­is, is clear: “I am con­vinced that we are going to have to turn pre­ven­tion, which is a great idea, into a real­ity. This will require the remov­al of a few obstacles, both eco­nom­ic and fin­an­cial. And, indeed, at a time when we are seek­ing to reduce expendit­ure, innov­a­tion will be the key to con­vin­cing decision-makers to invest and to suc­cess­fully imple­ment­ing it.” One of the lines of attack could also be to demon­strate effect­ive­ness, for primary as well as sec­ond­ary and ter­tiary pre­ven­tion. “Although primary pre­ven­tion requires long-term invest­ment, it cre­ates bene­fits that have an impact on sec­ond­ary and ter­tiary pre­ven­tion. This would enable us to bring the health­care sys­tem closer to the patient.”

This trans­form­a­tion can­not hap­pen overnight. As Lise Alter rightly points out: “Before any changes can be made, suf­fi­cient evid­ence is needed to provide these ele­ments of objectivity.”

Pablo Andres