The future of clinical research: AI personalized, real-world studies
Data is all around us, and applying AI to make sense of this information, is changing the way we live, work and interact with our family, friends, and work colleagues on a daily basis. When it comes to science, just like in every aspect of our life, AI and real-world Big Data will navigate us into a new era of clinical research. A future where we can develop even better treatments sooner, as clinical trial quality is drastically improved, and costs and study times are halved. A world where studies will be ongoing, flexible and personalized to our real-life experience.
When we think of clinical research today – we often think of clinical trials evaluating the safety and efficacy of drugs – involving several hundred or thousands of patients, costing millions of dollars and lasting years. However, Artificial Intelligence (AI), mHealth, Internet of Things (IoMT), big data, and not forgetting social media, means we are entering into a new era of clinical research – the “modern AI data-driven clinical research”.
Where innovative study designs, patient-centric approaches, biomarkers and personalized medicine, enable scientists and researchers to explore new ways to advance clinical treatment and diagnosis. Where AI and the access to previously unavailable genomic, clinical and environmental Big Data enable us to carve out a new clinical research landscape – AI personalized, real-world studies!
Driven by AI threading its way through every part of our life, this new personalized approach to clinical research, will enable us to address the ever-increasing expectations of the modern patient – the development of targeted personalized treatments!
So let me explain more.
Transforming clinical research with AI and real-world data
Traditional research to this day, starts by asking a specific, focussed question, designing a theory or working hypothesis and then testing this in a very specific and controlled environment.
In clinical trials, for example, we ask if this drug is safe? We then test this specific question using a defined protocol on a selected group of patients with defined characteristics and health status. We then analyse the collected data to reveal new insights that help us formulate the answer to our research question. For example, drug X showed a good safety profile in this group of patients. However, the selection of a specific study cohort leaves us with many open questions; for example, how safe is the drug to a different cohort? What is the long-term effect of the drug? Are there any adverse interactions of the drug with other drugs?
The revolution in AI and real-world data now enables us transform this style of traditional clinical trial. We can now drastically improve clinical trial quality by asking and answering additional questions, as well as lower costs and study time through the analysis of existing real-word data.
With the availability of data today, we can also have much greater impact and influence on the complete breadth of clinical research, such as genetic, epidemiological, diagnostic, quality of life and prevention studies! The advent of real-world data enables us to approach these latter studies in a whole new way.
We can now design research, which involves tens of thousands of patients who are almost instantly recruited by technological means like the Apple ResearchKit. We can expand our initial, specific, focussed question with many additional questions that are constantly fine-tuned and optimized as we collect ongoing data. And we can do this all in an uncontrolled, real-life setting at much less cost than ever before!
Yet, this is not all. We can also use the already accumulating genetic, clinical, demographic and social data from previously conducted clinical trials, and from additional available data sources. The challenge now is how do we analyse and make sense of this ever-expanding amount of new information to reveal intelligent findings. This is where AI can help us truly leverage the value of real-world data in clinical research!
So what do I mean?
The dawn of AI real-world studies
We are now living in the data-driven era.
In its simplest form, we have large data-sets at our finger-tips, which previously did not exist. The explosion in wearables, medical devices and IoT means we now have access to Big Data collected outside of conventional controlled trials. This includes datasets from electronic health records, registries, hospital records, health insurance data alongside biobank, genomic and digital phenotyping information.
However, that doesn’t mean we forget everything and start a scratch with clinical research. Instead, it means we take what we have learned and build even better clinical trials with the tools and data we now have available. We “reinvent” them!
This is where AI and “real-world” data come in.
These huge mountains of data allows us to ask many different questions in an ongoing way. We can drill down to discover new research insights – even ones we didn’t even know we were looking for when starting…!
We also no longer need to struggle to recruit a few hundred volunteers, we can involve and analyze hundreds of thousands or even millions of patients from these new databases at the press of a button (okay – it might take a little bit longer!). Moreover, from a statistical point of view, by using more patients and datasets, our findings will be more robust and valid, enabling for more accurate and repeatable results. Research that may have taken years can be done in weeks!
The recent trial by Pfizer with 23andme is a fantastic example of the potential of using Big Data real-world datasets in research. By being able to access much larger datasets (the trial involved over 300,000 people) through 23andMe’s massive genetic database, Pfizer was able to link 15 genome sites to depression using the traditional genome-wide association approach. A great success when considering previous traditional genome-wide association studies had failed to identify genetic regions that correlate to risk of major depression. I believe that by using AI techniques they could go even further to reach even more significant and interesting findings – please take a read of some of my previous blogs on these topics: Personalizing medicine: Using AI to analyze combinations of genetic and environmental factors and Cracking the genetic mysteries in the mental health modern space.
For us at Taliaz, clinical research that leads to practical solutions, which help doctors find out who is most likely to respond to a drug and who is not is the essence of why we exist.
During my research into depression, the #1 worldwide health disability according to the World Health Organization, it became very clear that one antidepressant does not fit all. Though clinical research had provided breakthrough antidepressant medications, tragically 65% of patients do not achieve remission with their first antidepressant prescribed by their doctor. What doctors needed was more scientific knowledge and real-world data, so they could know which medication to prescribe for each individual. Taliaz’s AI algorithmic platform, Predictix, does exactly this and this has huge benefits for personalizing the treatment experience in clinical research.
Turning scripted study experiences into real-life experience
AI and real-world data can also enable us to enhance the complete clinical trial experience.
Whereas before, clinical studies have been delivered in very controlled environments – like a scripted play, AI and real-world data enables us to be flexible and see the full 360-degree patient view to capture the “unscripted” part of the play.
We can look to integrate the chaotic, chance-and-circumstance related treatment reality that patients experience in the ‘real world’ of hospitals, with the scientific and robust, hypothesis-driven drug testing from clinical research.
Peter Arlett, Head of Pharmacovigilance and Epidemiology at the European Medicines Agency (EMA), summarizes it nicely “There’s a realization that real-world data can complement evidence from controlled trials and address questions that trials can’t – particularly the performance of a medicine when used in clinical practice”.
In essence, real-word data gives us the capabilities to deliver ever-evolving clinical studies that using AI can capture and analyze billions of data points to collect “real-life” insights. Most importantly, it has the ability to adapt, learn and personalize study treatments to individual patient experiences – in my opinion the “holy grail” for clinical research.
The future of clinical research: AI evolving studies
AI is transforming our business, personal and life experience, and just like in other fields – clinical trials are not going to be left behind!
AI solutions will be able to recruit eligible clinical trial patients in minutes, find disease-causing biomarkers and gene signatures, read volumes of text in seconds, and discover breakthrough diagnostic tools and treatments for chronic and terminal illnesses such as cancer and Alzheimer’s disease.
We at Taliaz are hard at work exploring ways we can deliver new innovative solutions in this changing field. Our Predictix AI algorithm is already able to analyze genomic, clinical and demographic big data sets to empower doctors to prescribe the right treatment sooner. This technology can also be harnessed to help pharmaceutical companies and other drug development companies design trials with those patients who are most likely to respond to treatment.
Clearly, the AI clinical research journey is just getting started.
I recommend you read this article “The future of clinical trials” to see how some of our great industry pharma leaders see how this space will change.
As you can see from my article, for me personally, I envisage AI will enable the healthcare industry to design clinical trials for the patient and that evolve with the patient. In essence, evolving studies with AI technology entwined throughout every stage personalized to each patient’s trial experience. Where AI algorithms are constantly learning, improving and optimizing treatments not just to genetic profiles, but also incorporating their clinical, emotional and behavioural state.
Exciting stuff and I can’t wait to see how the industry deliver on these possibilities to develop even better treatments and get them to more patients sooner!