Clinical research is a branch of healthcare which determines the safety and efficacy of medicines, devices, diagnostic products and treatment regimens intended for human use.
Whenever there are any new device diagnostic products that need to be launched in the market or any new condition has to be treated with already existing medication, it needs to be checked for the safety and efficacy at the dose that needs to be administered.
A medicine or devices or diagnostic process undergoes the following phases in clinical research:
In this phase, the drug is tested in non-humans. In this evaluation of the efficacy, toxicity and pharmacokinetics are made.
b) Phase 0:
In this phase, a small number of healthy volunteers such as around 10 people are tested for the pharmacokinetic parameters. The dose for the healthy volunteer is calculated based on the pre-clinical trials.
In most cases, this phase is skipped and phase I is conducted directly.
c) Phase I:
This phase is conducted to check the safety of the drug. This is conducted in healthy volunteers ranging from 20-100 people. It involves testing multiple doses to calculate the apt dosage for the efficacy in patients.
This phase is conducted in around 100 to 300 patients in different parts of the country to involve all types of sampling pools with a dosage based on the phase I trial. This phase of the study is conducted to assess the efficacy and side effects of the devices or drugs.
e) Phase III:
This phase is conducted in a large population of patients from different parts of the country around 300 to 3000 patients in order to study the efficacy, safety and effectiveness of the drug or device. Once the drug passes through this phase, it is eligible for a marketing license.
f) Phase IV or post-marketing study:
This involves the study of how well the drug performs in the market after being launched which is in terms of efficacy and safety.
Clinical research is a hub of huge amounts of data related to the drug performance, efficacy in each patient, adverse events produced in different scenarios in each patient, etc. Thus clinical research leads to huge data of different variables for the analysis using artificial intelligence.
Artificial intelligence has a high scope in the field of clinical research as follows:-
- a) It helps in determining the pattern of how well the drug is performing in the patients.
- b) It helps in determining the pattern of side effects or adverse events that are seen in different patient pools. Hence this can lead to efforts for avoiding such side effects in the patients.
- c) The postmarketing study is a tedious task as it involves a huge number of data which is tough for normal eyes to analyze and predict the defects in terms of efficacy and safety. Whereas, with the improved technology of machine learning, this can be easily analyzed. The drugs that can cause harmful adverse events or side effects at a higher proportion compared to its therapeutic activity can be easily withdrawn at a much earlier stage.
Other than the above mentioned broad scope of AI and ML in clinical research, some of the in-depth spheres of clinical research where AI and ML plays an important role are as follows:-
As AI and ML can help in the prediction of appropriate dosage and design required for the drug to pass the clinical trial phases, the same can be incorporated in the protocol designing which would help the manufacturing companies to reduce cost and provide a better medication or treatment to the patients at a faster rate.
One of the case studies of cognizant of how AI helped in fast-tracking the cancer drug development is as follows:-
One of the major clients of the cognizant that required full range of cancer treatments including acute myeloid leukemia (AML), needed a method for more quickly and accurately processing the massive amounts of data emerging from their own trials, from available research, and from the Cancer Cell Line Encyclopedia (CCLE).
Using a variety of data science tools and techniques, the cognizant team was able to build an automated solution that made the identification of optimal doses for drugs dramatically faster.
Hence, with the full drug development process taking from ten to eighteen years and costing $40,000 to $50,000 per patient, the data science solution could trim up to four years from the process and offers savings of as much as 10% of total costs.
Monitoring by Clinical Research Organization (CRO)
Traditionally monitoring of 100% source data verification was performed in clinical research by the CRO team. As this is a cost consuming and time-consuming process, the new ICH-GCP guidelines have introduced a lean approach to clinical monitoring. This involves monitoring on the basis of risk or Risk-Based Monitoring (RBM).
FDA defines RBM as, “This guidance assists sponsor of clinical investigations in developing risk-based monitoring strategies and plans for investigational studies of medical products, including human drug, biological products, medical devices, and combinations thereof. The overarching goal of this guidance is to enhance human subject protection and the quality of clinical trial data by focusing on sponsor oversight on the most important aspects of study conduct and reporting.”
Data science tools and techniques can help to integrate data from various systems, and effectively analyze and track the issues and risks in a timely manner which might be overseen by humans.
Site selection with the required population pool
Site selection having the population pool as required by the protocol is one of the biggest challenges faced by the CRO. This can be overcome by AI and ML tools that identify and suggest the sites based on the highest recruitment potential and using appropriate recruitment strategies. This involves mapping patient populations and proactively targeting sites with high predicted potential to deliver the most patients.
Identifying and recruitment of patients
Identifying patients and recruitment are one of the crucial issues faced by most of the CRO which leads to crossing the initially accepted study guidelines. This happens mainly because the patient pool is tracked and recruited during the study. Due to medical conditions and other events, the patient might get dropped out before the study completes. This dropout rate can be reduced by AI as it can help in reducing the population heterogeneity during the enrollment phase itself. By analyzing the medical history and the protocol requirements, the data science tools can predict whether the patient would complete the study endpoints.
To ensure drug safety, a huge amount of structured and unstructured data has to be analyzed. Hence, AI and ML technologies could address many of the challenges faced and provide new insights into drug safety.
Artificial intelligence can hence play a vital role in each stage of the phases and help the manufacturers to reduce the cost of clinical research. A better treatment is also possible by analyzing the huge data produced during each stage from the available repositories. This can also help to provide a better design of the study.