What can Big Data do for complex diseases?
Millions of individuals enter observations on their health actively or passively into their smart phones. Data collected through mobile phone apps on a global scale and in a short amount of time can help researchers to uncover new ways to detect complex diseases significantly earlier.
Massive datasets have proven very helpful and transformative to biomedical research. The entire genomics, proteomics and the other high-throughput obtained -omics heavily utilize faster and better computing power to dig deeper and deeper into our molecular structures. Genomic and proteomic sequencing provides us with both resolution and data enormity better than the electron microscope can output, and it is getting cheaper by the day.
Big Data is more than the sum of its numbers. It is mostly about how we use what we collect. Google has shown in 2014 how powerful algorithms can be used to warn and suppress the epidemic spread at an early stage. Big Data offers us an amazing opportunity to increase our health awareness through fast collection, its analysis, interpretation and actionability.
We need Big Data if we want to make a big change with a global impact. That way we will give a better chance to our health. We can start finding finer and finer patterns of the disease, as it always happens when one starts to study subject in depth and width. Consequently, new symptoms will start emerging, and we can visualize them in charts that medical healthcare professionals use when assessing us. We are making first steps towards this, FitBit is a bright example of a commercially available wearable sensor technology participating in many clinical trial studies.
Where does Big Data come from and who can use it?
For the most of my career, I was an academic (15 years, you might have read in my previous posts), and know how much great knowledge and expertise academia houses. At the same time I know how much academic research groups need a pool of good Big Data. By good I mean: diverse, inclusive, actionable and trustable.
Personally, deep understanding and linking early patterns of diverse complex diseases to us all is the major point of collecting this massive amount of data on a global scale, and all in order to help us manage and maintain our health in the best way possible. Determinants of our health can be bound to our lifestyles, our regional “selection” of a pool of genes that have proven to display certain disease in a specific way; mostly it is both. To understand the origins and the pattern of a disease, and connect it to the Big Data, a lot of diverse knowledge is needed.
This post is mainly about why creating collaborative efforts between research institutions and mobile phone health apps or wearable/implantable sensors collecting an immense amount of data should be a solution for tackling early disease detection.
Stanford Medicine is one of the bright examples of how proximity to the transforming technology hub at Silicon Valley can foster a new movement in the precision health and an collaborative effort between one of the leading health institutions and mHealth companies that collect big data. MyHeart Counts is the most known project they pushed out (at least to me), with an app that allows people to contribute to a study of human heart health while learning about the health of their own hearts. Powerful stuff!
As my case study, I will use female health, and two prevalent conditions: endometriosis and polycystic ovary syndrome (PCOS). The choice is not random at all, I work in a female health company, and equally important I have friends and family members suffering from diagnosed and undiagnosed conditions affecting their reproductive health. The fact is that although called reproductive conditions, many of them affect not only reproductive, but also person’s overall health (such as diabetes type 2, infertility, cardiovascular diseases and cancer).
There is one more reason I picked female reproductive disorders, it is because dealing with anything connected to menstrual cycle is still sometimes considered somewhat uncomfortable. If the health subject is considered uncomfortable, it will surely not get it’s deserved attention, and without that diagnosis will fail. It will come to late, and it might be wrong because not many and not right questions were asked, or the patient thought it is not important to mention. To be able to diagnose, healthcare provider needs to know the symptoms.
Why is menstrual cycle so important? It is a vital sign of health for half of worlds population that experiences it, did or will experience it. Just like the body temperature, pulse and respiration are for all of us. And still, not everyone will recognize or agree with this fact.
What can we learn from menstrual cycle? I hope we can learn how to detect complex diseases, including cancer, earlier.
Let’s talk about endometriosis and PCOS
What is known so far on menstrual cycle-related diseases is not enough. The scientific history of endometriosis and PCOS indicates at that.
How do symptoms look for endometriosis? They revolve around pain: painful periods, painful ovulation, pain during or after sexual intercourse, chronic pelvic pain, as well as heavy and irregular menstrual bleeding, fatigue, and infertility. All of this impacts person’s wellbeing, and reduces the life quality in individual that experiences it. PCOS, being a more complex disease of the two, has a more dispersed symptom pattern, but at the same time, possibly in some ways more specific.
If we look into the published scientific research, which is where majority of our knowledge on health and disease comes from, it becomes apparent that the topic has only gained increased interest in the last 10 years, when more than 40% of the total scientific literature on endometriosis and more than 70% on PCOS is published.
So, finally complex conditions are getting increasingly researched. Diverse angles and approaches are being used, but this might not be enough. There is a lot to do, and a large empty area to cover in order to reach knowledge parity with some other prominent conditions (cardiovascular as a first example). In order to reach the parity and make lives easier and better with an immediate effect we need a fast and a vast data collection. And that’s why academia needs Big Data.
By merging this old-school knowledge from academia and the new ways to collect massive amounts of data in a short span of time by a mobile health apps we can open a way to faster and a more efficient research. We can ask questions and already start getting an idea of how the answer will be shaped in days and weeks instead of months, years and decades.
Academia has probably the biggest knowledge collection in great amount of published work; it has a deep knowledge on health and easily accessible expertise on specific subtopics.
Complementing that, the mobile health apps, are able to collect on a daily basis more data than an average research project generates during its entire run, which is measured in years. Knowing what users wants enables mHealth apps to have easier global reach.
With this hybrid research form, we have a way to test and reproduce some of the older publications that are still used as a reference when creating normality charts. This old research can be expanded by including millions of diverse individuals, so the modern version of old research can be faster, bigger and more inclusive.
Inclusivity here is having more than one dimension, while the first one represents including diverse populations in the research, the second dimension is about providing an immediate knowledge of its health and possible perils to an individual. By that we can have an increasingly decentralized health knowledge, which brings a more democratic process in the healthcare.
Gaps left to fill
Bigger Data, both individualized and global can help fill the today’s “underresearched” gap.
The more health data academic institutions have access to, the better they will be at asking right questions. It will become easier for research groups to plan and conduct research leading to significant improvements in our health.
There are many questions left to answer; and it will be like that for some time, as there are many layers left to peel down. Many controversies, when detecting and diagnosing complex diseases still persist, and need to be cleared as soon as possible. Even the simple questions on how many people are suffering from a certain disease, or are at risk of it? What is, and is there a signature to these diseases that can be applied to certain populations? How is a condition one suffers from today connected to the later ailments, does it predispose people to other illnesses?
Big Data makes it possible to answer the questions within a reasonable time. It will enable us to create distinct pools of individuals sharing similar phenotype. Through that we can project how will persons health change through the next few decades. It will significantly lower the costs of a healthcare that we have today. It is important to mention that there are still a lot of issues and somewhat heated debate on processing of sensitive data in medical research (not everyone is in favour of ‘consent or anonymise approach’).
This actionable information will enable YOU to have a lifelong full control over your health.
References:
http://online.liebertpub.com/doi/pdf/10.1089/big.2013.0027
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1069067/
http://www.nature.com/nrcardio/journal/vaop/ncurrent/full/nrcardio.2016.42.html
http://www.nature.com/ejhg/journal/vaop/ncurrent/full/ejhg2015239a.html
