E-health: how far will Big Data go to heal us?
In recent years, the use of data has developed strongly in the field of marketing, but its development is disrupting many other sectors such as agriculture, the sport… And health is no exception. It is even one of the sectors where they have the most potential, according to the conclusion of the last November report on life sciences from Ernst & Young (EY). The firm concludes that only healthcare companies that have anticipated its surge will be able to face the new challenges. “Data” has potential in the medicine and pharmacy sectors. “The use of Big Data in health is a revolution, which will impose itself and have an extremely structuring impact on the health system in general”, estimates Dr. Eric Baseilhac, Director of Economic Affairs at Leem, one of the major unions of drug manufacturers.
Moreover, Google is already investing in it. She Flatiron Health finance, an American start-up – founded by two young entrepreneurs from marketing – looking for new factors to reduce Cancer by analyzing the data of several million patients. The firm had already set up the Google Flu Trends project in 2008, to try to predict the spread of epidemics from the data; and work at project Calico Genetic Disease Screening.
In fact, in the field of health, the use of data could make it possible to support medicine on various levels:
- Predictive medicine: that is, to predict a risk of disease in certain patients from the genome. “Often the problem with prevention is that it is too broad. With the data, we will be able to target the right people according to the risk they have of developing a particular disease, ”explains Eric Baseilhac. This predictive medicine is favored by the participatory medicine, i.e. the collection of health data through quantified self. The data then helps to strengthen epidemiology. Chronic pathologies whose parameters are well identified and linked to clear physiological mechanisms are those for which “data” has the most potential. Manuel Gea, vice-president of the Centrale Santé association, gives the example of diabetes. Sanofi created Diabeo to monitor changes in blood glucose levels and share data by smartphone with the doctor.
- Personalized medicine: all the data known thanks to genetics will make it possible to adapt the treatments to
certain types of patients, depending on their genetic profile. This is particularly the case for certain conditions such as breast and colon cancer. The sharing of data collected at national and global level, between different health establishments will gradually lead to an adaptation of the diagnoses and especially of the treatments to difficult and rare cases. “We will be able to better and better adapt the diagnoses and treatments to this type of case, in pediatrics for example,” comments Eric Baseilhac. “It will nevertheless be advisable to always take care qualify, contextualize and relativize correctly the data. Unlike marketing, we are not in the same level of responsibility for health ”, however, warns Manuel Gea, CEO and co-founder of BMSystems, specializing in research into new treatments and prevention solutions.
- Relevant medicine: Decision-making algorithms, proposing a diagnostic outline have been developed. IBM, for example, designed the Watson supercomputer which provides a first diagnosis based on remarks from practitioners, notes from interviews with the patient, family history, analysis results… Several American hospitals have already adopted it. , as the entrepreneur and Digital Champion Gilles Babinet tells us in his latest book “Big Data, thinking about man and the world differently”. Many other diagnostic aid software based on the use of data exist. In France, the Parisian start-up Khresterion has developed such software. It helps caregivers of patients with diabetes and cancer to make decisions using artificial intelligence.
- Pharmacovigilance: large-scale data collection through connected bracelets, for example, will make it possible to deduce data on the side effects of drugs and to better prevent them. By crossing the data recorded by health insurance, we can imagine observing the rate of hospitalization of patients taking a particular drug, imagines Mr. Baseilhac. But for the moment, the CNIL refuses to authorize the sharing of this data to drug manufacturers.
The data is also of interest to pharmacy researchers. In biology, where the hypothesis is at the origin of all reasoning, the use of “data” can open a new field of reflection. Rather than deriving a hypothesis from a first observation of a gene or a protein, researchers are likely to identify a probable trend from a mass of data on subjects of many molecules. This is what the American is working on Ernest Fraenkel, researcher in biology at MIT. His work is not yet unanimous within the scientific community.
Then, during the creation of a drug, the data will make it possible to move to a more predictive model, using mathematical analysis techniques in order to precisely predict the physiological effects of candidate molecules. Finally, the analysis of patient activity on social networks could lead in some cases to selecting the right candidates for clinical trials, according to a study by McKinsey published in early March 2015.
Still, there is still some way to go.
“Even today, data that the medical profession actually uses are extremely limited: a brief summary of the patient’s life, some information on the main diseases he has known, measurement of the pulse, blood pressure, supplemented by expensive medical examinations, if necessary (…) There are also many analyzes that contradict the statements of those concerned (…) However, these data do exist; they were created by the medical profession, but they are often scattered, ”emphasizes Gilles Babinet in his latest book.
In addition, when the use of data will develop, Manuel Gea considers that care should be taken to ensure who collects these health data. He considers the possibility of them falling into the hands of insurers to be a “risk”.
” People are naive about this collection because there has not yet been a problem (…). It will be necessary to be very careful that actors do not seize the data by turning them against the individuals ”he warns.
It would be more inclined to use data for the purposes of prevention and explanation of pathologies to patients. With the arrival of connected health objects, “scientific communication must become more and more serious, doctors must be trained to better explain pathologies to patients and the proper way to use these objects”, Manuel believes. Gea.
The health law currently being discussed in the Assembly will lead parliamentarians to define more precisely the framework for opening up health data.