APBN New Site

APBN Developing Site

The Hyper-Intelligent Clinical Trial

Ross Rothmeier shares how AI is being applied in the recruitment process of clinical trials.

by Ross Rothmeier

The success of clinical trials will soon no longer be enshrined in two sacred numbers – progression-free survival or overall survival rates. Measurement of treatments now extends beyond historical gold standard endpoints and provides a third dimension that is made possible with the burgeoning age of intelligence.

New treatments and trial designs derived from massive amounts of health data being continuously monitored and measured will reveal far more insights than could previously be imagined.1 Soon targeted interventions will be ushered in by intelligent clinical trials.

This new form of intelligence clusters the human population beyond ethnicity or geographical location, and deep dives into the genetic make-up of mankind. The ability to do this will eventually pave the way to cutting edge personalised medicine. Treatment efficacy and safety will be so precise that historical measures of success will seem primitive.

With the union of great minds and machine algorithms, will this be the answer to incurable rare diseases or put an end to chronic diseases like cancer and diabetes? Will machines one day replace doctors in medical diagnosis and treatment recommendations?


Artificial Intelligence (AI) Is Here to Stay

The impacts of artificial intelligence (AI) can be felt across all industries and it is no longer an up and coming fad, but a reality that is here to stay. It is an established, albeit sometimes confusing part of the digital transformation of healthcare.4

While the late renowned physicist Stephen Hawking warned that AI “may be the best, or the worst thing, ever to happen to humanity”, he lauded its creation saying that it is “crucial to the future of our civilisation and our species”.3

AI can be defined by the ability of a computer or machine to simulate human cognition, such as visual perception, speech and language recognition, and decision-making.3

It serves to augment and assist medical practitioners to make sound clinical judgements, help detect trends, and even predict outcomes.

An example of a significant breakthrough in technology is the development of intelligent platforms for life sciences where clinical researchers can now navigate through highly complex multi-national clinical trials with ease and on one screen.


Bullseye for Healthcare Leaders

The healthcare industry has historically generated copious amounts of data and the volume has only increased in recent years with the digital transformation of the industry.

This growth has led to a culmination of data that holds the promise of supporting a wider range of medical and healthcare functions, including clinical decision support, disease surveillance, and population health management amongst others.1

AI’s potential is now being realised in the form of equipping and empowering practitioners with deeper medical insights, based on data, that fuel rapid medical advancement and enhance care for patients.

Specifically, the five major benefits that AI has brought to healthcare are:

    • Speed: High speed data analytics has enabled more data to be processed within a shorter time. Today, the unraveling of the intricate world of human biology takes only milliseconds with data being processed at the speed of light.4 With shortened time spent on data review and verification, drug developments are also accelerated and the road to commercialisation will be a much smoother and faster journey.
    • Flexibility: As sponsors strive to realise greater return on investment (ROI), clinical trials increasingly face changing requirements and priorities. AI-enabled platforms can now be programmed to handle complex requirements and enable quick and easy study change implementation, allowing for greater flexibility in today’s adaptive environment. Now, prototypes can be established in hours, setup can be completed in weeks and mid-study changes such as removing a treatment arm can be made without a vendor change order, thus eliminating unnecessary downtime.5
    • Cost-effectiveness: With AI introduced into healthcare, the impact of cost-savings has been felt across the clinical trial continuum. Besides overall study time reduction and faster route to commercialisation, one can already experience cost savings at the initial stages of trial planning. Today, AI technology allows trialists to utilise synthetic comparisons,6 or data-driven computer-generated control arms, to predict outcomes based on precedent. This serves as a powerful tool to reduce patient recruitment numbers by comparing test group with a control group created through algorithm mapping. In addition, AI-enabled platforms ensure that trial processes can be streamlined, resources are optimised to reduce wastage, and predictive modelling capabilities are leveraged to re-forecast budgets based on actual site performance.
    • Data quality: It is reported that up to one in six new molecular entities (NMEs) fail first cycle approval due in part to data integrity issues. With AI platforms, they can now analyse the data for a select clinical trial to identify trends that can impact data quality or integrity. Trends include but are not limited to: data anomalies, procedural differences between sites (such as dosing pattern differences), patients or sites discrepancies, and differences in patient- or investigator-reported outcomes. With the quality of data preserved, AI systems can quickly alert the trial team to take immediate actions to obviate potential problems throughout the trial period.7
    • Accuracy and reliability: The accuracy and reliability of historical data captured with AI technology allows us to set study criteria, clinical benchmarks, and even facilitate study site selection based on technical data and not just pure observations. With machine learning, we can now eliminate human bias, personal theories or instinctive deductions.3


Unlocking the Future Potential of AI

AI is on the cusp of breakthrough discoveries involving new diagnostic tools and treatments for various chronic terminal illnesses.3 Today we are watching and participating in a revolution of medicine and healthcare – from a focus on populations to individuals, with an understanding that no single human is genetically identical.

One example is the identification of predictive biomarkers and gene signatures for a rare disease known as idiopathic multi-centric Castleman Disease (iMCD). Leveraging on Medidata’s Rave Omics machine learning system, the Castleman Disease Collaborative Network embarked on a large-scale proteomics study to co-analyze Omics data and clinical trial data. This collaborative study eventually led to the discovery of six proteomically unique disease subtypes and disease states. Trace evidence also identified additional proteomic predictors of anti-leukin-6 treatment responses and revealed etiological insights into the poorly understood disease. This was made possible because Omics data analyzes a wide array of genes and proteins to understand the phenotypic and genotypic responses of complex diseases. For such rare diseases that often lacks sample size, resources, and treatment options, this strategy proves highly essential in overcoming these challenges. Besides, the discovery of these biomarkers has provided powerful insights into treatment response and potential new drug targets, highlighting the value of precision medicine.8

While we revere the power of AI as we get a glimpse of what the future might entail for us and machines, one thing is still certain – innovation is still driven by the human mind, which is still the most powerful tool in clinical research. It is doubtful AI will soon replace humans despite all of the hype about what it can do, but it is rapidly moving us toward days when there are far more treatment options that we know will make us healthier, based on the combined natural and artificial intelligence that is bringing them to us. [APBN]


  1. Harvard Business Review, May 2, 2019 “The Future of Drug Trials is Better Data and Continuous Monitoring”
  2. AI In Healthcare, Machine Learning 101: Simplifying It One Term at a Time
  3. www.clinicalinformaticsnews.com/2017/09/29/the-intelligent-trial-ai-comes-to-clinical-trials.aspx
  4. https://ec.europa.eu/programmes/horizon2020/en/news/data-processing-speed-light
  5. Medidata Rave RTSM Agile Randomization and Trial Supply Management. Medidata fact sheet, 2019.
  6. D. Berry, et. al., “Creating a synthetic control arm from previous clinical trials: Application to establishing early end points as indicators of overall survival in acute myeloid leukemia (AML).” Journal of American Oncology, May 30, 2017.
  7. Medidata Trial Assurance: Protect Your Potential Blockbuster from Avoidable Failure. Medidata fact sheet, 2017.
  8. www.medidata.com/en/press-releases/medidata-rave-omics-collaboration-uncovers-novel-insights-rare-disease

About the Author

Ross Rothmeier is vice president of technology solutions & innovation labs at Medidata Solutions.