Accelerating Toward New Paradigms: Next Generation Clinical Research and Evidence Based Medicine

Issues with randomized controlled trials (RCTs), the traditional ‘gold standard’ of clinical research

Clinical trials of today represent a culmination of centuries of advances in medical research, and are a vital tool that allow the safety and efficacy of drugs and therapies to be evaluated in a scientifically rigorous manner that also prioritizes the safety of patients. While clinical research has certainly come a long way, continual advances are necessary (as in any field) in order to adapt to changing societal patterns and needs, and to maintain its utility and relevance as new discoveries emerge and new possibilities arise.

In recent decades, randomized controlled trials (RCTs) have been considered the ‘gold standard’ for generating evidence in clinical research evaluating the efficacy and safety of medical interventions. In RCTs, participants are randomly assigned to receive either an experimental intervention or a control treatment, allowing researchers to draw conclusions about causation and treatment efficacy. While RCTs have provided, and continue to provide valuable insights into healthcare interventions, they are burdened by some significant limitations.

One major limitation of RCTs is the strict inclusion and exclusion criteria used to select study participants, which can limit the generalizability of the study’s results to the broader population. Conducting RCTs tends to be time-consuming and expensive, making it challenging to study rare diseases or perform studies on smaller target populations. Ethical concerns also arise when assigning participants randomly to different treatment groups, as randomization can lead to withholding potentially beneficial treatments from participants in control groups who may be in need of the treatment. Finally, rigidity in fixed-design clinical trials means that they can’t be adapted during the course of the study, ignoring potentially important developments based on emerging data that could suggest the need to change aspects such as the treatment protocol, dosages, or treatment assignments.

Improving upon RCTs: What does the future of clinical research look like?

Accelerating advances in technology, medicine, and clinical research protocols are opening up possibilities for overcoming some of the limitations associated with RCTs. These advancements are shaping the future of clinical research and expanding our understanding of various conditions. Below, we go over some of the major technological advances in healthcare over the past 20 years, to provide context for the discussion of next generation medical research that follows.

What are the latest innovations in clinical trials?

In recent years, advancements across various areas within healthcare have opened new possibilities for clinical research:

Digital healthcare technologies and digital endpoints

Digital healthcare technologies have become valuable tools in clinical trials, with wearable devices, mobile apps, and remote monitoring enabling the collection of data and measurement of digital endpoints in real-time, without requiring patients to visit a trial site. These include wearable devices that track patient data such as heart rate, blood pressure, activity levels, blood glucose levels, sleep patterns, and more. Collecting real-time data through these technologies allows for more accurate monitoring of patients' responses to treatments.

Personalized medicine through genome sequencing

Personalized medicine has become increasingly feasible due to advancements in genome sequencing techniques, such as CRISPR-Cas9. By analyzing an individual's genetic makeup, researchers can identify specific biomarkers or genetic variations that may affect responses to certain medications or disease progression. This knowledge opens up opportunities for tailored treatments based on a person's unique genetic profile. Genetic information can also help identify subgroups of patients who are more likely to respond positively or negatively to specific treatments, which can be leveraged in study design.

Biomarkers

The identification of new biomarkers plays a critical role in medical research by providing measurable indicators of biological processes or responses to therapies. Biomarkers allow researchers to monitor disease progression more objectively and assess treatment effectiveness on a molecular level. They also allow for targeted therapies and personalized treatments tailored to individual patient characteristics, and can be used to identify patient subgroups that are most likely to benefit from particular interventions.

Remote and decentralized clinical trials

Remote and decentralized clinical trials leverage various technologies to enable participation by individuals who may face geographical barriers or have limited mobility. Decentralized or hybrid trials involve some or all of the following features: virtual visits/telehealth, remote monitoring, eConsent, electronic data capture (EDC), wearable devices, direct-to-patient distribution models, and ePRO. Remote trials increase accessibility for a wider range of participants while reducing the burden on patients by eliminating the need for frequent in-person visits, supporting improved recruitment rates and enhanced participant diversity.

Patient centricity

The concept of patient centricity is based on recognizing the importance of involving patients as active participants in the research process. Incorporating patient perspectives in study design, recruitment, and outcome measures not only enhances patient engagement but also ensures that research outcomes are aligned with their needs and preferences.

AI and ML in clinical trials

Both artificial intelligence (AI) and machine learning (ML) have shown great potential in optimizing clinical trial processes. Practical use cases for these technologies have recently exploded across many fields, and in the context of clinical research, they have been used to assist with drug design and candidate molecule selection, patient screening and selection, identifying patterns from large datasets (“big data”) to predict treatment responses or adverse events, and streamlining data analysis for more efficient decision-making. AI and ML also hold promise in optimizing the design of adaptive clinical trials, analyzing real-world data (RWD) to gather new data to inform clinical studies, and accelerating data processing steps.

It is theorized that the overall drug development timeline (from drug design through to approval) could be shortened from the current 12–15 years to less than 4 years.[1]

Adaptive design clinical trial

Adaptive clinical trials allow for mid-trial modifications to the trial or treatment protocol, based on interim analyses or accumulating data, while maintaining scientific integrity. Adaptive clinical trial design tends to be a more intensive process, but this is one area in which AI algorithms have been applied successfully by rapidly testing various potential adaptive designs. The adaptive approach enables flexibility in research protocols, allowing researchers to re-assign participants to a different treatment arm, change dosing or administration protocols, etc., without compromising statistical rigor. One of the major benefits is in ethical concerns, as patients can be re-assigned to a treatment arm if incoming data suggest that the treatment is having a beneficial therapeutic effect.

Next generation clinical trials and evidence-based medicine

Evidence based medicine refers to a comprehensive strategy for making patient care decisions that are informed by three key considerations:[2]

1. The experience and expertise of the clinician
2. The best scientific evidence that is available on the condition at hand
3. The values of the patient

Evidence-based medicine (EBM) may thus draw upon the results of clinical trials, but routine care (whether evidence-based or not) also serves as a source of real-world data, which is increasingly analyzed in observational studies as data mining and analysis capabilities continually improve. The advances in technology and medical research mentioned above have led to the creation of “next-generation clinical trials,” which involve advanced protocol design and are poised to generate even more robust evidence than RCTs, thus supporting the shift toward evidence-based medicine. Some of the main designs considered to represent next generation medical research are described below.[3]

Master protocols: Advanced and adaptive clinical trial designs

One innovation in clinical trials is the design of master protocols. Master protocols provide a framework for studying multiple treatments, biomarkers, or disease subtypes, in separate substudies under a single overall protocol. As of today, the field of oncology has been both the leading source and host of such clinical trial designs. Different types of master protocols have been developed, and 4 of the main master protocol designs are described below.

Umbrella study

In an umbrella study, researchers evaluate multiple targeted interventions for a condition, with substudies investigating different targeted therapies for different subtypes or genetic markers of the disease. Participants are enrolled into various arms based on their specific disease subtype or molecular profile. For example, patients with non-small cell lung cancer (NSCLC) could be enrolled under the umbrella study, and different targeted therapies could be investigated for each of the various subtypes of NSCLC the participants have. The main advantage of umbrella trials is the optimized ratio of potential benefit to risk to patients, but they can require large sample sizes and long durations, and may not be suitable for studying very rare molecular subtypes.

Bucket trial (Basket trial)

In a basket trial, patients are enrolled based on having a specific genetic mutation or biomarker, although it presents as different types of cancer (or another condition, although they have been primarily applied in cancer research). In this way, patients with cancers of different origin are grouped together for evaluation of an experimental treatment targeting an alteration or biomarker that is common to all of those cancers, allowing researchers to directly investigate the genetic/molecular target and its response to a given treatment. A primary feature of bucket trials is the possibility of identifying a therapy that is effective for various types of cancer, or uncovering a new indication for a drug that is already approved for the treatment of another type of cancer. One major limitation is the lack of a control arm or comparison arm.

Platform study

Platform trials, sometimes known as multi-arm, multi-stage (MAMS) platform trials, are designed to assess multiple different interventions (‘multi-arm’) against a single control arm.[4] The ‘multi-stage’ aspect refers to their adaptive nature, wherein new interventions can be added or removed over time based on emerging data and participant responses.[5] This allows for efficient testing of multiple treatments simultaneously, and has the benefit of allowing researchers to remove certain interventions in favor of others that are proving to be more therapeutically effective. Another advantage is the use of the single control arm, which minimizes the number of patients assigned to the control while maximizing the number who are assigned to a potentially beneficial treatment. A primary drawback is the complex nature of these studies, which can be expensive and lengthy or even indefinite in nature (without a defined end date), and which require a sponsor organization and sites to host them perpetually.

Master observational trial (MOT trial)

Master observational trials (MOTs) are prospective observational trials that represent the synergy of master interventional trial designs (such as the 3 types listed above) with the real-world data (RWD) generated by observational studies.[6] Being prospective studies, MOT trials enroll patients, but revolve around long-term follow-up and the collection of real-world data to assess the effectiveness and safety of different treatments used in routine clinical practice. Such trials help generate real-world evidence, going beyond certain limitations of traditional randomized controlled trials by studying diverse patient populations under real-world scenarios.

N-of-1 studies: Individualized medicine

N of 1 meaning sample size (n) of 1, refers to studies conducted on one single individual to gather data about the safety and/or efficacy of various interventions. N-of-1 studies are currently limited to rare and potentially fatal conditions, partly due to necessity (since sample sizes aren’t large enough for other trial designs) and partly due to ethical questions related to subjecting an individual to numerous treatments.[7] N-of-1 trials involve the intensive monitoring and assessment of an individual patient's response to various treatment regimens; by systematically comparing outcomes during different phases, such studies aim to find the most effective treatment for that specific patient. A drawback is the potentially significant lack of generalizability of the findings of such studies to other patients.

Real-world evidence (RWE) and real-world data (RWD) studies

Real-world data (RWD) refers to information collected from sources such as electronic health records, insurance claims databases, wearable devices, social media platforms, and other healthcare sources outside the controlled environment of clinical trials. RWD can supplement traditional clinical trial data by providing insights into how treatments perform in routine clinical practice and in diverse patient populations. Real-world evidence (RWE), on the other hand, refers to clinical evidence of treatment efficacy, side effects, and risks of a certain medical product – these conclusions are based on analyses of RWD. RWE is often used directly for post-marketing surveillance. As the amount of healthcare data generated by diverse sources becomes increasingly enormous, it represents a treasure trove of potential insights that can be used to inform healthcare decisions and policy. The analysis of such diverse data sources is another avenue in next generation clinical research, and its potential is opened up significantly thanks to artificial intelligence and machine learning tools, which we discuss next.

Artificial intelligence (AI) and machine learning (ML) tools in clinical trials

As discussed in the previous section, the integration of AI and ML technologies into clinical trials has supported accelerated study design, recruitment, data analysis, decision-making, and other trial operations. As the applications of AI, ML, deep neural networks (DNN, and the internet of medical things (IoMT) in clinical research continue to become more powerful and refined, they are expected to open up many new possibilities in research directions. Rapid analysis of enormous real-world data sets, high-throughput screening for new drug targets and drug candidate molecules, more accurate and early diagnosis through highly sensitive analysis of medical images, and improved trial workflows are but some of the advantages, the potential of which are yet to be fully realized.

Conclusion

While randomized controlled trials remain a vital and scientifically rigorous method in clinical research, advancements in technology and new approaches are revolutionizing the field. Next generation medical research will feature new and powerful combinations of digital healthcare technologies, real-world data, personalized medicine through genome sequencing, remote and decentralized trials, artificial intelligence and machine learning algorithms, deep neural networks, advanced analytics, adaptive trial design methodologies such as master protocols, increased patient-centricity, and emphasis on biomarkers. Taken together, these advances offer exciting and promising opportunities for enhancing patient care through evidence-based medicine.

By addressing limitations associated with traditional RCTs through innovative approaches driven by technology-enabled higher-resolution insights into disease mechanisms, at both the individual and group levels, clinical research efforts will continue to yield enhanced treatments tailored to patients' unique needs while simultaneously expanding our understanding of diseases across diverse populations. The future of clinical research seems promising, for researchers and patients alike.