Revolutionizing Clinical Trials with Artificial Intelligence
A key requirement before any drug can get the approval of the U.S. Food and Drug Administration is for such to have successful and convincing clinical trials. The process is not easy, is fraught with issues, and costs companies millions of dollars to complete.
Clinical trials are carried out in phases to test their efficacy before they are made available to patients. The requirements for these phases increase with each succeeding one. For instance, Phase I trials typically involve fewer participants and so are less expensive than Phase III trials.
The process can take any length of time from about 7 to 15 years, which means it can consume a significant amount of resources. The clinical trial industry is reported to be worth around $65 billion. Each study costs any amount between $161 million and $2 billion, as reported by CB Insights.
Yet, a good number of these trials fail at a significant cost to pharmaceutical companies. The financial effect of failed trials is more evident on the smaller players.
While the need for clinical trials is obvious, some have argued that the current state of things deprives patients of medications that can potentially preserve their lives. It is believed that the application of artificial intelligence (AI) has the potential for transforming the process considerably.
There are diverse ways by which artificial intelligence is thought to be capable of improving the clinical trial process.
There are many patients who may have an interest in taking part in a clinical trial and yet not be aware of the availability of such. They may only find out about one when their doctor suggests it, especially when other options have failed.
On the other side of the divide, there are researchers looking for study participants and cannot get enough. There are more than 18,000 trials currently recruiting subjects in the US, according to CB Insights.
AI can help to solve this problem. It has the potential of reducing the level of time-wasting involved in finding and matching patients with trials, as this entails sifting through a lot of information or records. A piece of software can easily search through medical records for people fitting to particular trials.
It is also possible to increase the level of enrollment in studies through the use of machine learning. This is a challenging aspect due to a lack of awareness and other issues.
A report by Cognizant revealed that about 4 in every 5 clinical trials fail due to the inability to meet enrollment deadlines. Around a third of studies terminated in Phase III failed because of issues pertaining to enrollment.
The use of algorithms can streamline the processing of patient information and identifying ideal candidates for studies. This will be useful for informing qualified patients of clinical trials they might be interested in.
Furthermore, AI may make the process of evaluating treatment results easier. A test trial involving healthy volunteers and using both Orbita.ai and Amazon Voice System displayed the potential usefulness of machine learning in this regard. These tools were used to design a questionnaire and engagement feature that is activated using voice.
The trial showed how AI can be used to evaluate the condition of a patient to see what effects treatment is producing. The majority of the subjects also found the technique preferable to interactions via apps or voice assistants alone.
However, an aspect that needs to be clarified or addressed is that of security. Patients need assurance that their information is secure.