In this post, I will discuss how broad availability of medical and patient data as well as new data analytics approaches will accelerate the Digital Transformation of the Life Sciences industry.
How AI and Big Data will improve healthcare for everyone
Artificial Intelligence (AI) and Big Data are expected to play a key role in scientific research and development. The amount of data available for clinical studies in all phases is growing exponentially, driven by OMICS-, Image- and Biomarker Data. Analyzing this vast amount of data requires new approaches leveraging predictive analytics and machine learning.
Preventive medicine and the recent trend to fight the diseases of aging such as Alzheimer’s at its root cause – aging itself – will eventually turn everybody into a patient.
With enough data available and advanced analytics capabilities such as AI at our disposal the industry is better equipped than ever to understand diseases and tailor individualized treatments for each patient. We will identify patterns and triggers related to aging and diseases of old age and how to overcome them. Eventually this will not only lead to longer lives but also better quality of life at old age.
What is the main challenge between today and this promising future of longer and healthier lives?
Our technical capabilities are evolving fast and computing power has finally arrived at a level where machine learning is a viable approach. Who would have thought of the capabilities it delivers today, such as semi-autonomous cars, ten years ago? It is difficult to imagine the possibilities in ten years from now. Clearly technology is not the limiting factor anymore.
Data is the limiting factor!
Taking the example of self-driving cars, what is making Tesla so successful? Since the early days they have included sensors and cameras in their cars and have connected them to their servers over the internet. This allowed them to collect vast amounts of data about the roads, situations and drivers through the ten-thousands of cars they have on the road. With this data, they can train and improve their AI system in a much more effective way than companies who only have a few self-driving cars on the road. Data is the fuel that gets the AI engine going.
What is the situation with data in the life sciences and healthcare industry?
Pharmaceutical companies have collected large amounts of data over the past decades through clinical trials. Since about two decades the data is captured electronically through Clinical Trial Management Systems. Does this mean the data is available and can be used for Machine Learning and Big Data? The reality is that studies have been conducted for their primary purpose of developing a specific drug and receive its market authorization. While the data could of course be used for further research and development (secondary purposes), the reality is that the reuse of data is very limited today. Some of the main challenges for reusing data include:
- No index or catalog of data available – It is not clear what data is available
- Lack of metadata – Not knowing the source and quality of data limits its reusability
- Data quality – Data not fit for purpose without manual curation
- Fragmented systems – Data not accessible without manual efforts
- Different data standards – The data cannot be reused due to incompatibilities
- Data protection laws – Legal constraints limit the reuse, manual data anonymization
Big pharmaceuticals have identified some of these issues and are working on better data sharing and reuse. Large programs have been initiated and it will take many years for them to show substantial outcomes.
On the healthcare provider side, the situation does not look better. Projects to store patient data in one single place to make it accessible where needed have failed in many cases. Data protection, governance, standardization and security concerns are amongst the key issues. We are far away from a world where one could share its relevant patient records with any doctor or hospital in the world. Think of a scenario where somebody with cardio vascular problems is traveling abroad and suddenly suffers from severe heart rhythm disturbances. Without his patient record the local clinic would have to run tests which could be avoided. Precious time is lost and resources are wasted. In certain cases, having such data could even be lifesaving!
Besides the above-mentioned issues, trust is a key issue. People are hesitant to share their medical data. They want to be sure their data is not being used against them in the future. That the data will benefit themselves and society, not some corporations or governments. Fair use and control over the data is a critical success factor.
What could be the future?
Think of a world where all individuals would truly own and have access to their medical data. Where they could share this data with whatever entity they find trustworthy while remaining anonymous. Where they could manage access to their data and revoke it at any time.
We would have a global data repository for all health-related data. It would be the fuel for AI approaches to fight diseases and provide us with recommendations and treatments on how to live longer and stay healthier. It would accelerate scientific advancements and the development of new treatments significantly!
It is time to level the playing field!
In such a world, innovative startups can get access to large amounts of medical data at low or no cost if they can present promising approaches. It would disrupt the system of drug development as we know it today for the benefit of everyone. It would accelerate the Digital Transformation of the Life Sciences industry!
What’s next?
In my next blog post, I will talk about how we can overcome the main obstacles such as security, data governance and trust and how to shape such a future.