Let’s start with the main purpose of clinical trials. Briefly speaking, clinical trials help to define and evaluate new treatment procedures. Patients for clinical trials are selected based on their treatment and medical history, and other factors such as severity of symptoms and health conditions. Most of these tasks are done manually. However, machine learning (ML) and AI can simplify data management during clinical studies. Continue reading our blog post to learn more about the role of ML in pharma trials.
Medical research industry in clinical studies: What the process looks like
How do you identify a patient who suits the clinical trials best? Usually, medical studies have a research team that consists of doctors, nurses, social workers, and other health care professionals. Before starting a clinical trial, they design a protocol that includes the following information:
- The reason for conducting the clinical trial and its main purpose
- Criteria for choosing the participants in the clinical study and the rules of patient enrollment in clinical trials
- The number of patients needed
- The length of the clinical trial
- A description of the drugs or processes that will be tested
- Data that will be gathered from participants
Here are the main stages of a clinical trial:
- The first stage covers testing a treatment on a small number of people to check its safety. Here, the researchers identify the possible side effects and safe dosage. This stage involves a small number of people (usually under 15). Note that patients who take part in clinical trials must sign a special document giving informed consent. It describes the benefits, risks, instructions, and process of the clinical study in great detail.
- In the second stage, a research team investigates the impact of the medicine or process on a particular illness. This stage takes around 3-4 months and involves 20 to 80 people who have no underlying health conditions.
- The third stage compares the new treatment with existing ones. It involves people who have conditions that the new medicine or process is meant to treat. During this stage, specialists gather more information about the side effects and different dosages
- The fourth stage happens after the FDA approves the medication. The purpose of this stage is to examine the long-term efficiency of the medication and its benefits.
How to choose patients for clinical studies to make them more effective
The selection of patients is a complicated process that includes a lot of factors. Researchers must thoroughly study the health conditions and medical history of a patient and analyze the medical records of the available subjects.
The global clinical trials market was $44.3 billion in 2020 and is expected to grow 5.7% annually from 2021 to 2028.
Each trial stage requires a distinct severity of symptoms and has special parameters the patients should fit. The important thing to remember is that all trial participant information is confidential and not to be shared with third parties. In the same way, trial patients and participants must keep trial information confidential.
Many therapeutic areas use individual patient markers that are associated with differential treatment responses. These include both baseline characteristics and short-term changes that come after treatment. These predictive markers help create more targeted studies and reduce the number of subjects to recruit.
As trials may include some risks, many trials require participants to undergo additional procedures, tests, and assessments based on the study protocol. A potential participant should also discuss any health issues with members of the research team and with their usual health care provider.
Patients get the following benefits from taking part in a clinical trial:
- They can get a treatment before it’s available on the market
- Medical trials accelerate the development of treatments that can benefit patients
- The cost of tests and visits related to the trial are paid for
There are special clinical trial recruitment companies that assist in choosing subjects for clinical trials. These organizations usually conduct patient research and create campaigns to reach more patients using Facebook ads, search ads, digital promotion through partnerships, patient databases, and EHR matching..
Instead of using a large team of researchers, AI can identify patients for clinical and observational studies based on clinical attributes. One example of this process is Watson for clinical trial matching. AI uses special identifiers to find a patient. How to identify a patient? Usually, special determiners are used for this purpose. What are patient identifiers? A patient identifier is a piece of information directly associated with clinical patients that reliably shows the individual to be the person for whom the service or treatment is intended. These identifiers include:
- Assigned identification number (e.g., medical record number)
- Date of birth
- Phone number
- Social security number
Machine learning clinical trials: Examples and benefits
Machine learning clinical trials are increasingly automated and, as a result, include fewer mistakes and failures. According to studies, the amount of data collected in clinical trials grows by 40% each year. If trials use more and more variables and millions of data points, it becomes more and more difficult to find the relationship between them.
Machine learning can predict the outcome of trials better as it helps manage resources more efficiently, leads to faster drug approval, and eliminates most of the manual work by using neural networks patterns.
“Everyone is affected by the risk of a drug failing in its clinical trial process. With more accurate measures of the risk of drug and device development, we hope to encourage greater investment at this unique inflection point in biomedicine.”
— Andrew Lo, the study’s senior author and director of the MIT Laboratory for Financial Engineering and investigator at the MIT Computer Science and Artificial Intelligence Laboratory.
Every clinical trial follows a protocol that describes all the details of the study and possible outcomes. If any problems occur during the trial, medical professionals need to adjust the protocol, which may take months and lead to major expenses.
“When designing a trial, researchers lean on information from numerous sources, including comparative studies, clinical data, and regulatory information. AI-powered software can not only process all of that information faster but also collate more data than a person could read.”
— Eric Topol, director of the Scripps Research Translational Institute in La Jolla, California
Trends in clinical trials that influence the market
Now let’s review the main trends that affect any global clinical trial in 2021.
- Virtual trials
Researchers use technology such as wearables that provide real-time data and spend less time on data entry and more time on the studies themselves. Wearables produce more exact data that allows for more accurate analysis and, ultimately, better results.
- Need for human connection in medicine
Taking part in a clinical trial often requires a human connection and the need to talk to a medical professional about sensitive medical topics. While many companies use automated follow-up emails or text messages that will remind patients about next steps and outcomes of a trial, it’s also important to maintain the human connection and support patients face-to-face.
- Post-trial monitoring
Clinical trial participants should be able to continue the investigational treatment they received in the trial. The market for post-trial monitoring of this kind is growing as post-trial responsibilities increase year by year.
“Our software development services are based on the latest trends in the medical industry. For example, we recently developed Triage, a custom software solution for healthcare equipped with an AI-powered voice attendant. The patient is asked to answer a series of questions with the help of a voice prompter.”
— Vlad Medvedovsky, Founder and CEO at Proxet (ex - Rails Reactor), a software development solutions company
Custom clinic solutions for clinical trials
Let’s take a look at a cutting-edge custom solution for clinical participant selection. Computer science and engineering researchers at the University of South Florida have developed a system that can identify appropriate trials for eligible cancer patients and designed a web-based interface that enables a clinician to enter new pharma trials without the help of a programmer.
Their system includes heuristics for ordering medical tests. Medical testing is a significant part of the cost of clinical trials, and the implemented heuristics reduce this cost. As the next step, the researchers plan to deploy the system at Moffitt and evaluate its effectiveness in helping clinicians.
The system uses a “patient-centered” approach, choosing trials tailored for each patient based on information collected from their intake appointment with a doctor. A physicians typically delegates the data entry to a nurse, who enters the related information after a patient’s visit.
If you’re looking for a custom solution for your clinic, we’re ready to help you at Proxet. Proxet is an AI medical company with experience in developing ML solutions for medical institutions worldwide. Furthermore, we offer digital transformation solutions to enterprise and mid-size clients, cutting-edge engineering solutions and expertise to startups, and team augmentation services to companies of all sizes
Dive deep into the technicalities of embedding models, vector databases, and optimization strategies to revolutionize information retrieval
Explore the correlation between latency and token generation in Large Language Models. Learn how prompt size impacts response time.