Is Big Data the Key to Tailor-Made Health Solutions?



When former US president Bill Clinton announced in June 2000 that the first survey of the entire human genome had been completed, he predicted the achievement would “revolutionize the diagnosis, prevention and treatment of most, if not all, human diseases” by allowing doctors to tailor sophisticated medicines to attack the “genetic roots” of disease.

Eighteen years later, Clinton’s hope that “our children's children will know the term cancer only as a constellation of stars” now seems wildly optimistic, as did dreams of flying cars 30 or even 40 years after the first moon landing.

Stephane Kirkesseli, Deputy Head, Translational Medicine and Early Development, Clinical Pharmacology at Sanofi, agrees that initial expectations of genetic medicine were too high. “Some people were thinking about testing their genetic code, and based on the analysis of their genome, getting truly personalized treatment,” he says. “We are still far from that.”

The concept has evolved to the point of, what Kirkesseli calls, “stratified medicine,” where genetic testing and other techniques can determine that subgroups of patients will respond in certain ways to certain treatments. “We are now able to define subgroups of patients based on biomarkers and as a result we have developed a number of targeted drugs for specific subsets of patients,” he says. 

He uses asthma to illustrate how “stratified medicine” can work for a complex disease. Asthma was once thought to be a “single-entity” disease, but after years of effort, “a number of subsets have been defined based on different markers” like the types of inflammatory cells present in the lungs of patients, resulting in the development of new drugs.  

It’s an important step towards true precision or personalized medicine, but we are not where we expected to be 20 years ago. Part of the reason for this is that the genome is not the simple, interpretable map of the human body it was initially thought to be. 

In the early history of genetics, researchers thought of the genome as a code, like computer code, but as understanding developed over time, precision medicine became not only about examining genomes and parsing them as if they were complex algorithms, but also about combining our knowledge of genetic data with as much other data – medical records, exercise and diet information, exposure to toxins – as we can collect. 

The goal of this holistic approach, sometimes called personalized medicine, is to take advantage of the latest in computer science, connected health-monitoring devices and data analyzing techniques and come up with a tailored plan for a patient or a subset of patients.  The integration of data from real life, randomized clinical trials, electronic medical records and public data sets is opening new ways to provide the right treatment at the right time, sometimes combined with adjustments in diet or behavior. 

Harmonizing data for the big picture

In order for computational techniques like machine learning to become more effective in advancing precision care, different data sets – genomic information, health records, behavioral and environmental data – need to be harmonized so that they can be used together. 

The concept is to put all that data into a big data platform, explains Javier Jimenez, Vice President, Global Head for Real World Evidence and Clinical Outcomes at Sanofi, so that machine learning, visualization, analytical and predictive modeling techniques would help  provide better insights on how a drug might affect different types of patients. This also means “taking into consideration not only the patient’s biology but also the healthcare environment.” 

We can then use the findings based on anonymous data from thousands or millions of patients to determine the probabilities of particular outcomes for individual patients in an effort to provide the right treatment at the right time.  

An increasing number of healthcare companies are using real world data and cloud-based platforms like our data platform, DARWIN “to better understand patients, to put more focus on their needs and to demonstrate the value of products,” says Jimenez, adding that there was still room to improve the quality and harmony of the data, including privacy and security issues. “Privacy is a critical element in the area of real world evidence,” says Jimenez, “for Sanofi that is a priority.” 

Going mobile to better understand patients

The evolution of connected mobile devices is playing a big role in gathering behavioral data that can be used to improve care. 

From highly specialized devices to consumer products like the Fitbit or the Apple watch, the amount of data produced by individuals is increasing exponentially. The new Apple Watch includes an electrocardiogram and a function that can detect an irregular heart rhythm. Both functions were cleared by the U.S. Food and Drug Administration. 

Sanofi works with a Silicon Valley company called Evidation Health, for example, to consolidate this kind of patient-behavior data and link it with clinical data. A system developed by Evidation Health is aimed at detecting potential changes in patient behavior to improve long-term care. Patients with diabetes, for example, tend to put on weight when they start taking insulin, which could be a reason for them to stop taking it. By tracking the patient’s weight, activity and pharmacy purchases, the system can alert doctors about patients who may not stick to their treatment plans.  

“Knowing that may happen and seeing the trend ahead of time, we can be much more proactive in providing education and care to those patients before they fall off the bandwagon,” says Christine Lemke, Co-Founder and President of Evidation Health, in an article published in October 2017.

Dozens of start-ups already offer services based on mobile health monitoring. 

Sanofi’s venture with Verily Life Sciences, an Alphabet company, created a connected device and software system that helps diabetes patients and doctors collaborate on a personalized treatment system.

“There is a tremendous amount of data available in the real world,” says Dr. Bernard Hamelin, Global Head of Medical Evidence Generation for Sanofi. “All this information can help us measure the outcomes of relevance for the treatments that are of interest to us. It also helps us contextualize these outcomes.”

The bigger picture brings a sharper focus to the individual

Some of the early, misplaced hopes about the genome were due to a failure to see the bigger picture, explains Kirkesseli. “We thought that we would be able to link genetic mutations with specific diseases in a better way. But maybe people forgot that most diseases are multifactorial - and so not just linked to one specific gene mutation.”

The effort to find different ways of classifying disease is evolving with precision medicine. “Classification based on clinical signs of symptoms or environmental factors has been known for a long, long time,” he says. “We are evolving towards a more mechanistic classification; this is how a pharmaceutical company can intervene. We need to dissect the mechanisms that drive the disease.  If we know, for example that certain types of cells are directly involved, then we can develop specific treatments based on the pathways we should target to treat that disease.”

So while we are making progress on developing tailor made solutions, it will be some time yet before “our children's children will know the term cancer only as a constellation of stars”.

Seconds Behind the Science

Genome

A genome is an organism's complete set of DNA. Each genome contains all of the hereditary information needed to build and maintain that organism. The term was coined in 1920 by German botanist Hans Winkler who combined the words GENe and chromosOME.  

Sources: https://ghr.nlm.nih.gov/primer/hgp/genome, last accessed December 2018, https://simple.wikipedia.org/wiki/Genome, last accessed December 2018, http://www.genomenewsnetwork.org/resources/glossary/, last accessed December 2018

Biomarkers

Biomarkers, short for biological markers, refer to a measurable indicator of a biological state or condition. Biomarkers are used to help predict, diagnose and monitor a disease. A biomarker can be a traceable substance that is introduced into the body to examine the heart muscle, for example. It can also be a substance, for example the presence of an antibody may indicate an infection. Biomarkers are also used to help identify patients who are more likely to benefit from a specific medicine or intervention. 

Sources: http://www.inchem.org/documents/ehc/ehc/ehc222.htm, last accessed December 2018, retrieved via https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3078627/#R2
https://en.wikipedia.org/wiki/Biomarker, last accessed December 2018