"What do you think of precision medicine?" my interviewer asked, as I stared blankly and nervously, not knowing what to answer. I’d never heard that term before. That was in 2015, at my first interview for a postdoctoral research position in the field of lung cancer. Squirming, I asked, "What do you mean by precision medicine?" "Oh, it’s a personalized approach to treatment and prevention based on a patient’s lifestyle, environment, and genetics," he answered. It clicked. "I think it’s a great approach. I always thought treatment and prevention plans should be cosmeticized for the individual, but won’t we need a lot of data for that to work? And won’t these types of plans be very expensive?" Later that day I learned that the Precision Medicine Initiative (PMI) is aiming to do just that, collect data to help scientists and clinicians create a more targeted approach toward diagnostics and treatment! But how is it going to accomplish this and can we help make medicine more personalized?
Introduced by President Obama at the State of the Union in 2015, the PMI aims, in part, to create "All of Us", a research program that encompasses a cohort of more than one million volunteers who will share genomic data, biological samples, and diet/lifestyle information, all of which will be linked to their electronic health records, over time. Management and analysis of this cluster of “big data” will advance our understanding of disease, foster drug discovery, and ensure proper delivery of necessary therapies and interventions. The program started its beta testing phase in May, 2017, and as of January, 2018, 18,000 volunteers have enrolled. With very few restrictions, any adult living in the United States can join the program and donate their data. There is currently a waiting list to join but that will be lifted once the program fully launches nationwide in Spring, 2018.
The More the Merrier
The idea of collecting medical information from a large cohort of individuals in order to gain new insights into targeted treatment and prevention strategies is not new. In fact, the use of PARP inhibitors in BRCA cancers, which I explored in my previous post, Silver Linings for BRCA Mutation Carriers, is one of the great examples of the successes of precision medicine in clinical practice. What makes All of Us so momentous and unique, however, is its scale and reach. Collecting medical data from one million participants has never been done before and will greatly improve the robustness of the findings and predictive power of the study. The more data we are able to collect the more accurate our models and analysis of the data will be. It is not only the amount of data that matters, however. At the moment, more than 80% of data collected worldwide pertains to individuals of European decent. All of Us will be able to change that because of its richly diverse setting; the United States is one of the most ethnically diverse countries in the world. The power of diversity will allow researchers to interpret the data better and develop accurate tailored medical plans for everyone. In essence, what fuels “All of Us” is, in fact, just that— all of us. We, as diverse individuals and communities, can make sure we are represented in this endeavor by being engaged participants whose medical data contributions will supply the biomedical research community with targeted information that will transform clinical practice and delivery of healthcare.
The Timing Is Right
Another reason why All of Us is so powerful is the recent advances in biomedical data acquisition, especially in genomics. In 2001, it cost around $100M to sequence the genome of a single person. Now it costs less than $1,500! This price is predicted to continue to decrease even more in the coming years and will make collecting and sequencing a huge cluster of genomic data much more affordable. Technologies for biomedical imaging and analysis, such as CAT scans and X-Rays, have also improved. Healthcare providers have established standards for fast biomedical image acquisition and sharing. Other forms of data, such as health records, are mostly electronic today. As of 2015, Electronic Health Records have been implemented for 80% of doctors and hospitals which makes sharing those data very easy, fast, and standardized.
This surge in the ability to generate and share huge amounts of data will require a proper and secure “big data” pipeline. Technology companies, such as Google and Amazon, have already taken advantage of the improvements in computing power and capacity to develop tools and models that streamline data acquisition, accessibility, management, processing, storage, analysis, and interpretation. The PMI will implement similar approaches in a biomedical research and clinical setting to analyze and interpret the acquired vast and varied types of datasets. The PMI cannot develop such complicated data science approaches on its own, however. Recently, a panel of biopharma experts and government officials at the World Economic Forum participated in a panel discussion entitled, The Promise of Precision Medicine, which emphasized the need for a collaborative effort among all research, healthcare, and technology sectors to develop an efficient pipeline. Harvard University, for example, has already setup the Harvard Data Science Initiative to help and train biomedical scientists in their big data analysis.
These different types of datasets from All of Us will also help develop and drive artificial intelligence models and algorithms that can, for example, perform biomedical image analysis autonomously. The more data, the better machines can learn how to process and analyze it. In fact, advances in machine learning will be essential for fast and accurate analysis of the big data derived from All of Us. Otherwise, it would not be humanly possible for even experts in the fields of genomics and biomedical imaging to sift through and analyze the large amounts of complicated and interrelated data. At the end of the day, the great value of acquiring big data only comes if we can extract accurate information and knowledge from it. We can collect all the puzzle pieces we want but if we cannot link them together we won’t be able to discover who, what, how, and when we need to target for treatment or prevention of a specific condition.
So They’ll Have All This Data. Then What?
Donating our data will help empower us as a society to take control of our health and eventually transform our visit to the doctor. But how will our data be used, exactly, and will our privacies be protected? As part of a blog post series on PMI, I will explore these questions and concerns in my next two posts. My upcoming one will discuss the types of translational and clinical studies that will be driven by the data from All of Us. I will delve into how data scientists, biomedical scientists, and clinicians will collaborate to study disease mechanisms and discover novel therapeutic targets and biomarkers. Finally, I will explore how All of Us will help apply precision medicine to cancer, especially hereditary cancers. In my third and the last installment of the series, I will discuss some of the ethical and technical concerns facing this historic endeavor such as data privacy and security and responsible sharing of personal information. In the meantime, show yourself and your community some love this Valentine’s Day and look into All of Us.