The Future of Clinical Trials

December 21, 2020

BY NATHALIE SMYTH, DANIEL BRETTLER, MICHELE FIELDS, AND CAMILLE HEPSWORTH

The landscape of clinical trials is evolving at an unprecedented pace, being both propelled and incentivized by the global race to advance the COVID-19 vaccines. Advancements in artificial intelligence (AI) and machine learning (ML) for data collection and interpretation, and a move towards virtual and adaptive trials and use of Real World Data (RWD), have paved the way for the development of novel data collection techniques and analysis.
Alongside such advancements is the ever-present threat of cyber-attacks and malicious hacking of private medical data as well as new potential liability risks associated with novel trial techniques and reliance on AI for decision-making. Increased flexibility and diversity of trial designs also come with the need for clarity of protocol and patient consent wording in addition to a good understanding of local coverage and liability issues in developing countries. Last, but not least, the expected roll-out of a worldwide vaccine at an unprecedented rate will necessitate careful consideration of policy wording encompassing all risks pertaining to the entire life cycle of medical products, including their development, manufacture, storage, transportation and beyond.

In this article, Conner Strong & Buckelew and Kennedys Law LLP explore current and potential future trends and liabilities in clinical trials alongside insurance considerations for key players in the life sciences market, including:

• AI and ML for Data Collection and Interpretation
• Novel Trial Designs
• Fast Tracking and Supply of Vaccines
• A Move Towards Diversity in Clinical Trials
• Looking forward

AI and ML for Data Collection and Interpretation
AI and ML tools are already being used to review thousands of study protocols and trial results and to provide automated advice during development of trial protocols, with the ultimate goal being the first machine-drafted protocol.

While there is currently limited use of the decision-making capabilities of AI to interpret clinical data, real advancements are being made in the use of AI and ML-based platforms and apps to assist with various clinical trial procedures. These include enrollment of trial subjects, confirmation of medication ingestion, facilitation of electronic clinical outcome assessment (eCOA) technology towards a more patient-centered approach to trial design and administration, and integration of data into a centralized platform to expedite analysis.

As the use of this technology develops, it is imperative that trial sponsors have adequate insurance in place to cover the potential new exposures they may bring. This includes consideration of whether and to what extent additional coverage is required in respect of potential cyber-attacks leading to unauthorized access to patient data and other GDPR breaches. Manufacturers and trial sponsors should also ensure that existing levels of coverage for professional, general and directors and officers (D&O) liabilities remain adequate, particularly when their duties and responsibilities may be linked to reliance upon AI and ML for decision-making.

Novel Trial Designs
Numerous manufacturers and trial sponsors are already spearheading several novel trial techniques, utilizing virtual and adaptive trial design and Real World Data (RWD)1.

Virtual clinical trials

While virtual trials have been feasible for a number of years, the impact of COVID-19 has pushed trial sponsors to quickly adjust their existing practices in favor of remote monitoring and data collection through the use of apps, electronic monitoring devices and online social engagement platforms.

For instance, in March 2020, J&J launched its first fully virtual trial (the Heartline trial) which explores whether its Heart Healthy app, paired with the Apple Watch’s Irregular Rhythm Notification (IRN) and electrocardiogram (ECG) apps, can help reduce the risk of stroke. The new virtual model means that people can participate remotely throughout the study without having to travel to research sites.

Similarly to AI and ML, virtual monitoring and use of apps and monitoring devices create a new area of potential exposure related to breach of private information, malicious hacking to disrupt or steal information, and potential bodily injury exposure if the monitoring device fails. As traditional insurance for clinical trials may not cover such exposures, detailed reviews of existing liability and cyber privacy insurance wordings should be carried out.

Move towards adaptive clinical trials

Clinical trials have traditionally taken the form of Randomized Controlled Trials (RCTs)2. However, there has been a recent move towards a more bespoke or ‘adaptive’ trial structure, which allows for prospective modifications to trial design based on accumulated trial data. For example, if one treatment is seen to perform better than others, the adaptive design would allow patients to be allocated to the better performing treatment during the trial period.

A move to more bespoke trial designs can provide huge benefits for both the patient and the manufacturer in terms of efficiency, ethicality, cost and time savings, improvement of patient outcomes and potentially the use of fewer participants.

There is also significant potential for a reduced timescale to obtain regulatory approval. This is demonstrated by the emergence of Real Time Oncology Review (RTOR) in 2018, which allows regulators to review data prior to the full application for regulatory approval with a view to reducing approval time from 6-10 months to under 6 months without compromising the safety and efficacy of the treatment. While this pilot program only considers applications for cancer drugs, if successful, it could extend to real-time review of drug applications for other disease groups.

As adaptive trials become more prominent, insurers and trial sponsors should also be aware of the potential exposure created by the more flexible design and the corresponding protocol and informed consent language. In particular, the underwriting of clinical liability insurance places a significant emphasis on the patients’ ability to understand the protocol. Therefore, wording must be clear and comprehensive, yet sufficiently simplified, for a non-health care professional or scientist to understand.

A growing reliance on Real World Data (RWD)

RWD is used at each stage of the drug development process to analyze the potential benefits or risks of a product or treatment, which, in turn, enables pharmaceutical companies to minimize risk and invest in only the most promising treatments.

While RWD is not a replacement for RCTs, it can supplement and augment analysis of data on the safety and efficacy of treatments over a longer period of time for a larger section of the patient population with a particular disease. Indeed, only last year, the FDA approved Pfizer’s Ibrance, which was the first drug where analysis was largely based on RWD from clinical registries.

The use of RWD (and emergence of technologies to capture it) could readily be the answer to the rising risks and costs of clinical trials, particularly in the study of rare diseases, or where there are difficulties with recruitment or funding of a full-scale RCT.

However, the implementation of new design methods based on RWD, especially those adopted mid-trial, may bring with it new additional focus on shareholder disclosure and potential litigation from a D&O’s liability perspective. For instance, if the sponsor company’s shares experience a drop in price due to negative clinical trial results, shareholders and their counsel will closely scrutinize all public statements made by the company before, during and after the clinical trial in order to assess their validity and accuracy, particularly regarding the reasons for any design changes made during the trial period. It is therefore important for insureds to ensure their coverage adapts sufficiently.

Fast Tracking and Supply of Vaccines
AI and ML will undoubtedly play an important role in the fast-tracking of future vaccines and other treatments through the regulatory process and onto the market. Indeed, AI-powered tools have already been utilized to expedite the search for potential COVID-19 treatments and vaccines by combing through thousands of studies for relevant information. The results allow researchers and policymakers to readily identify the most promising studies and to make faster and better decisions about prioritizing time and funding.

In particular, AI and ML could play an important role in Phase Three testing, which typically involves the evaluation of the risks and benefits of a medication over a large number of participants. Specifically, apps can be utilized by trial subjects to report side effects, with the data being analyzed by a centralized data system.

Again, it is important to properly insure against the risks of privacy and cyber breaches associated with the use of apps and e-technology to collect and share sensitive information, in addition to closely examining the possible implications of over-reliance on AI data for decision making.

Further, and in anticipation of the large-scale roll-out of regulatory-approved COVID-19 vaccines, manufacturers need to carefully consider whether and to what extent they bear the risk of loss of vaccines that have been stockpiled into the supply chain and the valuation of such risk. This is an exposure, which is often overlooked, although supply chain insurance for property perils and potential spoilage arising out of a change in controlled environment or condemnation is available.

A Move Towards Diversity in Clinical Trials
There has been a recent trend towards the relocation of clinical trials to less developed countries. While this has obvious benefits to manufacturers, sponsors and their insurers in terms of lower litigation risk, the importance of conducting trials in less developed countries should not be underestimated.

Africa’s virtual absence from the clinical trials map (impeded by limited infrastructure, low visibility of existing sites and unpredictable regulatory timelines) is problematic. Many potential trial subjects have had no previous exposure to pharmaceutical drugs and a number of diseases (particularly tropical) are endemic to the continent.

Africa boasts an unrivaled amount of genetic diversity, which, if not well-represented in trials, will not feed through to study findings and their generalized application to large populations. The promotion of such trials in less developed countries could therefore be the key to unlocking the potential for more diversified clinical trials going forward.

As sponsors consider studies in less developed countries, it is important to be aware that local insurance infrastructure and regulations therein are still developing. While many insurance regulations require that a local admitted insurance policy is purchased by the sponsor, these will not be as ‘well-tested’ as those in developed countries and sponsors will need to ensure that they strictly conform to local requirements and customs.

Even where local admitted coverage is not required, there may still be very strict requirements for sponsors to pay extensive damages to patients, their families and dependents, regardless of negligence. It is therefore important for brokers and insurers to have an in-depth understanding of what coverage is available in these emergent jurisdictions and under what circumstances it may be triggered, particularly as the policy may not cover certain heads of loss.

Looking forward
There is most certainly a place for new technologies in the clinical trial process, which will require detailed collaboration and an openness between manufacturers, trial sponsors and regulators to produce adequate guidance and a framework for the possibilities of AI and ML to be fully realized.

It is also hoped that the increased flexibility and diversity in trial designs and use of RWD, alongside lessons to be learned from the development of a COVID-19 vaccine, will lead to more efficient trials and development of medicinal products, with reduced risk, timescale and costs for all involved.

1 RWD is the collection of information pertaining to a patient’s health status and medical history from a variety of sources such as electronic medical records, insurance claims, disease or product registries, and mobile health apps.
2 A Randomized Controlled Trial (RCT) is a type of study in which treatment is compared in two or more randomized groups of people against a control group.