According to the most recent research carried out by Grand View Research, the global market for clinical trials is increasing alongside the demand for clinical trials in developing countries. Several reasons contribute to the industry's popularity, ranging from the increasing incidence of chronic diseases to the need for personalized and orphan drugs.
By now, the introduction of a new medication on the market takes about 10 - 15 years and requires USD 1.5 -2 bn of investment. A significant share of these financial and time resources is spent precisely during the clinical trial phase. This is one of the decisive factors that drive the implementation of modern technologies, such as Artificial Intelligence, into the industry. Let's take a closer look at how AI can facilitate the clinical trial design.
What are clinical trials and their place in the drug development process
The drug development process is a bedrock of the healthcare system as it's a direct provider of therapeutic benefit during treatment. It falls on 4 equally important stages: drug discovery, pre-clinical, clinical trials, and regulatory approval. In recent years, the clinical trial design captures the attention of pharma, biotech, and device companies that want to improve the overall process.
A clinical trial is a type of research conducted in humans with the aim to assess how the drug affects the subjects being tested, whether it has side effects, and whether it has properties not previously described. It is an essential stage because it allows investigators to measure the safety of the drug and its impact on the health of subjects. The deliverables of a clinical trial directly affect the possibility that the drug will be released and distributed.
Since the design of a clinical trial is a comprehensive and long-lasted process, it is unacceptable to make a mistake at this stage. Any misstep will result in the loss of a significant amount of money and time. However, due to the need to meet numerous conditions, the pharmaceutical industry continues to face many challenges. These challenges and the opportunities to overcome them with Artificial Intelligence technologies we will consider further.
Challenges during clinical trials design
The flow of clinical trial design was changing to fit the United States Food and Drug Administration (FDA) requirements, thereby speed up drug delivery.
Firstly, it was a conventional (linear) process followed by a strict protocol. Research advocates formulated a scientific hypothesis to accept or reject it with no modifications alongside. Even though the approach had a strong scientific foundation, the tests weren't successful enough. Lack of flexibility in interpreting the outcomes and difficulties of administering effective medicine made enthusiasts look for new ways of fostering clinical trials.
Due to the stagnation of innovations in the market, the pharmaceutical industry has created a new model - adaptive clinical trial design. The FDA justifies the need for such a model in contemporary research by allowing adjustments to be made at different stages of the trial, based on data received from trial participants. Depending on different factors, scientists modify treatment assignments and process data faster. All of these should improve the overall drug development process.
This is a theoretical scenario, but what do we have in practice? We've selected pitfalls of current methodologies that inhibit clinical trials design:
- errors prone - most clinical trials fail in the designing stage due to non-optimal assessment, and a human element. By following the procedure, patients should take medications regularly and keep a diary. But they often forget to take a pill or at least record it.
If talking about medical centers, their recording system also can't keep documentation in proper order. Current technology doesn't meet specifications. Such a situation gives rise to abuses.
- high costs, overblown budgets - The average cost of one patient in a US-based clinical trial was $36,500 (£28,736) according to PhRMA findings. R&D costs rise exponentially since clinical trials include several phases as trial design, protocol design, site selection, and trial execution; moreover, they could be prolonged. Experts cite inadequate planning in response to inflated costs.
- regulatory issues - adaptive design implies the use of the unblinded interim results as a part of the statistical method. They threaten not only the relevance of the study, but also European, and U.S. regulators restrict it via the Adaptive Charter. The unblinded interim results should preserve statistical rigor anyway. Hence not every unit could enforce the rules, while the institutions provide updates every year.
The recent meeting of FDA & EMA Global Regulators revealed new details in COVID-19 vaccine development. How much time should it take the clinical trial process when people's lives are at stake?
Impact of AI on clinical trials design
With the evolution of new technologies, for about the last 6 years, scientists
began applying Artificial Intelligence for Clinical Trial Design to refine clinical testings. It will take another decade before it becomes clear what contribution AI has made to the pharmaceutical industry. However, it is already possible to evaluate how this technology can be used to deal with the challenges mentioned above.
Drug development life cycle improvement
The most significant shortcomings of a designing phase that can cause a failure of a trial are ineffective mechanisms for selecting participants for the research and absence of technical equipment to conduct it. To overcome these cornerstones, experts are slowly starting to use Machine Learning and Deep Learning, which allow them to find identical patterns among a large data set - in text, speech, or images. Natural Language Processing helps to understand and proceed with written and spoken speech while Human - Machine Interfaces further allows exchanging information between the machine and the person.
These AI capabilities will be very helpful in selecting volunteers prior to trial and in monitoring compliance more closely for accurate deliverables evaluation. The main objective of Artificial Intelligence technology for clinical trials is to boost the chance that the FDA will approve the compounds and thus accelerate the process of bringing a new drug to the market.
It is no longer a secret that process automation leads to cost reduction. The discovery of new drugs is very time-consuming and costly at each stage, especially when it comes to the design of a clinical trial.
Speaking of the design of a trial, there is a need for experts to analyze a large amount of information, ranging from data on previous related researches to regulatory information. Machine Learning, as a part of Artificial Intelligence technology, can analyze multiple times more information and make a report much faster than any human using a manual method. Such data processing has great potential to reduce costs by improving data quality, increasing patient retention, and measuring drug efficacy better than ever.
Timely medication intake
When developing a clinical trial protocol, it is necessary to eliminate the human factor that could prevent a valid outcome from being obtained. To ensure each trial participant takes the required medication in the right quantity and at the agreed time, an AI-based solution was invented.
This software allows users to videotape the process of drug intake. In turn, based on the computer vision algorithm, the solution analyzes the video and identifies the person taking the drug and the drug itself. The solution engineers claim that it can even measure people's facial expressions to track their response to treatment, to guide the development of therapies. However, the objective is not the guidance, but a first crude protocol written by AI.
Practical use: How startups reimagined the process of designing clinical trials
There are plenty of solutions that already facilitate clinical testing design via AI technologies for drug development and clinical trials.
Startup company Trials.ai launched an AI tool for crafting trial protocols. Considering the complexity of the clinical trial procedure, San Diego's company decided to optimize the protocols. With NLP (Natural Language Processing), collecting and analyzing data is accelerated in comparison with human capabilities. The technology proceeds trial-related documents allowing them to obtain more precise outcomes and reach maximum patient participation.
"When protocols are right, drug development is faster and cheaper," founder of Trials.ai Craig Lipset says. For now, the company reported 98% of patient adherence.
Winterlights Labs is a Toronto-based company that specializes in cognitive neuroscience. They mainly deal with dementia and psychiatric illnesses hard to predict. AI algorithms substantially contribute to interpreting the data by analyzing short snippets of speech (voice patterns, amplitude, the emotional impact of the patient).
AiCure is another startup that utilizes analytics to work closely with patients. The solution allows identifying whether the patient takes medication or not via the camera. The computer-vision algorithm not only controls the treatment process but also analyzes behavior patterns to adjust the therapy.
AiCure platform shows the AI technology has desired results, as 90% of people with schizophrenia undergo treatment as expected.
We hope that the clinical trial process will be replenished with AI-powered solutions very soon. Since the rooms for innovations are more than enough, this sphere is hugely demanding more comprehensive tools that could cover all clinical trial phases at once.
The clinical trial design is a crucial part of medication development, as it affects human well-being. Innovations in the market, like AI power, foster their conduction, cut costs, and promise better patient adherence.
Nowadays, in the COVID-19 splash, humanity needs a safe and effective vaccine as never before. Moreover, clinical trials should run simultaneously to provide us a miracle drug. So we can't rely on old-fashioned techniques anymore.
It is Artificial Intelligence that can help society to fight the epidemic quickly and effectively. Many companies are now implementing AI technologies to accelerate the process of vaccine invention. And they are doing it pretty successfully! Another COVID-19 vaccine was recently released and tested positively, and its model was developed exactly using AI. Will it help humanity? We'll see!
Your submission is received and we will contact you soon