Bringing drugs to the market is a long and arduous process. The study estimates that the clinical trial process—testing new drugs on patients before approval—lasts nine years at an average cost of $1.3B.
The Covid-19 pandemic has triggered the adoption of technologies that can greatly increase the efficiency and cost of the traditional clinical trial process. To keep people healthy and safe in the future, we will increasingly rely on fast-moving and efficient experiments.
Artificial Intelligence (AI) will play an important role in the transformation experimentation process, ultimately making us healthier and better able to prevent disease. The healthcare industry is leading the way in adopting AI, and is experimenting with applications ranging from machine learning-assisted diagnosis to extracting information from electronic health records.
However, the adoption of artificial intelligence in actual clinical trials is still in its early stages. Compared with other health care fields, there are fewer start-up companies that directly target customers in the clinical trial field. In many aspects of clinical trials, the need for digitization precedes the need for artificial intelligence.
Many clinical studies use basic data collection and verification methods-this usually places the responsibility on the patient, such as sending patient medical records by fax, manually calculating the remaining pills in the bottle, and relying on patient diary entries to determine medication compliance. Needless to say, this process has matured.
Time to change
We cannot simply accept that testing new drugs will continue to be a slow and expensive process. Artificial intelligence has the potential to disrupt current clinical trial methods — from patient recruitment to compliance monitoring and data collection — now is the time to seize these opportunities.
Patients usually participate in drug trials only when existing forms of treatment have failed. In addition, not all confirmed patients are eligible to participate in the trial, because determining eligibility alone can be a daunting task.
For those eligible, participating in trials is often expensive and time-consuming. This process is also inefficient for other stakeholders: As mentioned above, drug trials take nearly ten years on average, and the average cost exceeds $1B.
Current trials still lack the analytical power, flexibility, and speed required to develop complex new therapies for a small and often heterogeneous patient population.
In addition, sub-optimal patient selection, recruitment, and retention, as well as the difficulty of effective management and monitoring of patients, have led to high trial failure rates and increased R&D costs.
The use of AI-enabled digital health technologies and patient support platforms can revolutionize clinical trials and achieve greater success in attracting, attracting, and retaining loyal patients throughout the study period and after the study is terminated.
In short, the application of artificial intelligence can reduce the cycle time of clinical trials while increasing productivity and the cost of clinical development results.
The combination of artificial intelligence algorithms and effective digital infrastructure can clean, aggregate, encode, store, and manage continuous clinical trial data streams.
Artificial intelligence technology has the potential to change every stage of the clinical trial process, from search for trials to registration to drug compliance.
The adoption of artificial intelligence has the potential to change clinical trials in four key areas:
- Clinical trial design: Biopharmaceutical companies are adopting a series of strategies to innovate test designs. More and more scientific research data, such as current and past clinical trials, patient support plans, and post-marketing monitoring, inject vitality into the trial design. Artificial intelligence technology has the unparalleled potential to collect, organize, and analyze data (including failed data) generated by more and more clinical trials, and can extract meaningful information patterns to help design.
- Patient enrichment, recruitment and registration: Matching the right trial with the right patient is a time-consuming and challenging process for clinical research teams and patients. In fact, only 3% of cancer patients participate in clinical trials today. By mining, analyzing and interpreting multiple data sources, including electronic health records (EHR), medical imaging, and “omics” data, digital transformation supported by artificial intelligence can improve patient selection and increase the efficiency of clinical trials.
- Patient monitoring, medication compliance and retention: Artificial intelligence algorithms can help monitor and manage patients through automated data collection, digital standard clinical evaluation, and data sharing across systems. The combination of artificial intelligence algorithms and wearable technology can achieve continuous patient monitoring and real-time insight into the safety and effectiveness of treatment, while predicting the risk of dropout, thereby increasing participation and retention.
- Use operational data to drive clinical trial analysis that supports artificial intelligence: Trials will generate large amounts of operational data, but functional data islands and disparate systems may prevent the company from fully understanding its clinical trial portfolio at multiple global sites. Integrating all data (regardless of source) into a shared analysis platform, supported by open data standards, can facilitate collaboration and integration, and provide insights across important indicators. Combined with a self-learning system, it aims to improve predictions and prescriptions over time, coupled with data visualization tools, can proactively provide users with reliable analysis insights.
Looking to the future
Artificial intelligence has been used to change the clinical trial process and experience, but there are also some challenges. In many clinical trials, researchers still fax requests for patient records to hospitals, and hospitals usually send the data back in the form of PDFs or images (including pictures of handwritten notes).
Due to these transmission methods, structured data may also become unstructured. For example, an electronic form sent by fax or converted to a read-only document (such as PDF) loses most of its structure. This outdated manual system makes it difficult for clinical trial researchers to collect the accurate data needed to determine patient eligibility.
The AI solution uses natural language processing (NLP) to extract clinical data from patient records, such as symptoms, diagnosis, and treatment. Its software can even identify patients not explicitly mentioned in the EHR data, thereby improving the matching rate between patients and clinical trials.
Non-compliance is another challenge that may adversely affect the health of patients. If a study must recruit new patients, it will incur costs and interfere with the accuracy of the results of the study. Generally, treatment effects require a compliance rate of 80% or higher. However, in the United States, as many as 50% of prescription drugs are taken incorrectly. In response, clinical research sponsors are investing in emerging technologies to minimize non-compliance.
Some start-up companies are providing visual confirmation of drug management. For example, some platforms use interactive medical assistants (IMA) to identify patients at risk of non-compliance based on visual data collection. The patient uses his mobile phone to take a video of himself swallowing the pill, allowing the platform to confirm that the correct person has taken the correct pill.
Embrace change
Due to interoperable data, open and secure platforms, consumer-driven care, and fundamental changes in the way we manage health, the healthcare and life sciences industries are on the verge of massive disruption.
Together, large technology companies and start-ups have laid the foundation for faster and more effective clinical trials in the future. The ultimate goal is to promote innovation and better healthcare through the use of artificial intelligence throughout the clinical trial process.
The next decade will continue to empower patients as individuals rather than cohorts characterized by medical diagnosis or a series of symptoms. Those participating in clinical trials will look forward to a more comprehensive and immersive experience, providing opportunities for continuous learning while receiving quality care. The personalized health report, data visualization, and portable raw data provided at the end of the trial will provide participants with opportunities to carry forward their experience and strengthen ownership of their health and care.
For healthcare, 2021 is still a complicated and uncertain year. But despite the challenges, dedicated leaders around the world are using the power of healthcare IT to create more efficient and effective clinical trials so that patients and clinicians can truly change the lives of the future.
