Artificial intelligence is rapidly transforming from a high-cost luxury to a commercially viable solution. Today, exponential advancement in computing power and telecommunications allows us to share extremely large amounts of data. This, coupled with the need to streamline processes and increase efficiency has placed AI as an ubiquitous tool, not only for commercial purposes but also for personal consumption. As AI becomes more and more popular, many are seeking how to apply artificial intelligence in Clinical Trial Enrollment.
Research has shown that an archetypical drug takes 8 to 12 years to become ready for market consumption and production. Research costs around $1.5 Billion to $2 Billion. Further, a staggering amount of clinical trials end up failing, making this investment a potential sunk cost.
Companies funding clinical trials have to consider the costs, time, and regulations needed to send a new drug or treatment to the market. The possibility exists that a good medical treatment will not make to the market because of high prices and inadequate systems of clinical trial. One of the reasons many clinical trials fail is a poor recruitment / enrollment process. In fact, studies have shown that poor recruitment along with the lack of an effective monitoring system are the main contributors to clinical trials failing to meet deadlines.
Where Does Artificial Intelligence Come Into This?
Artificial intelligence tools can help reduce the high failure rate of clinical trials. Researchers have lauded the benefits of AI , and its potential in helping increase the efficiency of clinical trial recruitment. Additionally, research at AI trendsetting companies like IBM have shown that AI can help by streamlining the enrollment and recruitment process through the automatic selection of candidates based on clinical and personal data.
Artificial Intelligence in Clinical Trial Enrollment
Researchers have noted that over 86 percent of clinical trials fail to meet enrollment deadlines, citing the recruitment process as the main cause of trial delays and over-budgeting issues. Furthermore, manual recruitment also fails to effectively assess the individual conditions and requirements for patients with the necessary complexity.
Often, a patient may suffer from the ailment aiming to be treated, but may not be at the stage required for effective data gathering in a clinical trial phase. Other cases include patients not belonging to the specific sub-phenotype that is targeted by the drug in development. Moreover, many viable candidates are unaware of the clinical trials in progress in their area. A common cause of inconclusive trails is the lack of ethnic and genetic diversity to ensure that the drug is ubiquitous. This issue is known as population heterogeneity.
AI tools can address the afrorementioned concerns. Enrollment can be completed in a timely manner, with a set of patients that are the most suitable for testing the viability of a drug in a clinical trial. AI tools such as Machine Learning and Natural Language Processing are already used in similar applications in other industries, hence they can be retrofitted to improve the complex tasks associated with clinical trial enrollment. Electronic Phenotyping, for example, can address the issue of population heterogeneity. When done right, this will ensure that the subset is diverse so as to avoid inconclusive and potentially dangerous clinical trial results. This use-case of Artificial Intelligence in Clinical Trial Enrollment could simnifically improve the industry margin for recruitment.
Artificial Intelligence’s Advancements in Clinical Trials
IBM’s Watson for Clinical Trial Matching
IBM’s commercial grade AI tool, Watson, is applying aforementioned Natural Language Processing aspect of AI. Using Natural Language Processing to analyze data such as patient records, trial inclusion and exclusion criteria and demographic information, IBM’s Watson for Clinical Trial Matching tool is able to review a potential list of trials for a patient in question while ensuring that the clinical trial is able to meet its enrollment goal. Currently, oncologists have access to IBM’s Watson for Clinical Trial Matching for this very purpose. This is just one example of many applications of artificial intelligence in clinical trial enrollment.
IBM’ research states significant obstacles in the face of a clinical trial success. Their research quotes the 80 percent failure rate in US clinical trials with regards to meeting their recruitment timelines. IBM also notes that 25 percent of sites enroll a lower number of patients than what is required to show conclusive results. Additionally, 10 percent of sites actually fail to enroll a single patient. Additionally, only 15 percent of patients are aware of clinical trial options. Further, clinical trials are only able to recruit 3 percent of cancer patients. IBM’s Watson for Clinical Trial Matching helps improve efficiency for both prospective patients and the clinical trial sites. IBM states that technology feasibility studies have shown a 78 percent reduction in patient screening time. Additionally, a staggering 94 percent of omission of non-matching patients after the employment of their Watson tool was reported.
Another neat application of artificial intelligence in clinical trial enrollments is that of a new entrant in this field is StudyProtocol.io. This is a cloud based collection of microservices and software solutions, designed to help improve the efficiency of the sub-steps in the clinical trial recruitment process. The suite works as a whole to manage the entirety of the clinical trial recruitment process.
This fairly new solution is already making headways in recruitment campaigns. Using Artificial Intelligence and iteration, StudyProtocol.io is capable of refining audiences of interested volunteers to present relevant clinical trials.
The Future of Artificial Intelligence in Clinical Trial Enrollment
The highly lucrative AI research field is only beginning to realize its true potential. As researchers have access to improved hardware capabilities year-after-year, the applications of AI keep increasing and improving.
In the case of clinical trials, the potential for AI to transform the process is almost endless. Companies like IBM and StudyProtocol are already using AI to help streamline the process. Some of the AI features in use include:
- Natural Language
- Machine learning
As the role of AI increases in clinical trials, more patients will have access to the right trial, helping medical researchers bring life-saving drugs to market faster than before while simultaneously maximizing the safety of the patients and the effectiveness of their medical process.