Testing new drugs is a slow and expensive process. AI has the potential to disrupt clinical trials — from patient recruitment to adherence monitoring and data collection — and Covid-19 has catalyzed its adoption.
In the past year, nearly 5,000 clinical trials were launched to test life-saving treatments and vaccines for the novel coronavirus.
Covid-19 clinical trial enrollment is 80% higher than average. However, this is less impressive when considering that for many diseases, such as cancer, less than 10% of eligible patients enroll in a trial.
get the full report
Patients often only enroll in a drug trial when existing forms of treatments have already failed. On top of that, not all diagnosed patients are eligible for trial participation — determining eligibility alone can be a herculean task.
For those that are eligible, participating in a trial is often a cost- and time-intensive endeavor. The process is inefficient for other stakeholders too: drug trials average nearly a decade, costing over $1B on average.
The $52B clinical trials market needs a makeover.
Startups and big tech are actively developing clinical trial solutions, from IoT for remote monitoring, to machine learning for electronic health record (EHR) processing, to AI-based cybersecurity for data protection.
Below, we map out a patient’s journey through a typical clinical trial process, and explore use cases for emerging technologies like AI at each step.
Note: We specifically focus on drug-based trials, although technologies discussed in the final section of the report are applicable to a wider range of clinical studies.
TABLE OF CONTENTS
- Why faster clinical trials are critical for pharma companies
- The current state of clinical trials
- How AI could change every stage of clinical trials
- Finding a clinical trial
- Challenges with enrollment
- Medication adherence
- How big tech is disrupting clinical trials
- Google’s healthcare data platform
- Apple’s and Facebook’s moves toward clinical research and trials
- Why AI alone isn’t the magic bullet
- How Covid-19 has affected clinical trial tech adoption
- Study design
- Virtual trials
AI in Healthcare
Healthcare AI startups raised over $6B in 2020. Check out companies involved in healthcare AI in the AI in Healthcare collection.
Track The AI Healthcare SpaceWhy faster clinical trials are critical for pharma companies
Bringing a drug to market is a long and arduous process.
Studies estimate that the clinical trial process — where new drugs are tested on patients before the FDA approves them — lasts 9 years and costs $1.3B on average.
Clinical trials are conducted in multiple phases, with cost and complexity increasing from Phase I to Phase III.
Despite the time and capital invested in trials, only 1 in 10 drugs that enter Phase I of a clinical trial will be approved by the FDA.
Clinical trials fail for a variety of reasons, including failure to recruit enough trial participants, mid-trial patient drop out, side effects, and inconsistent data.
Naturally, trials that fail at a later stage prove more costly.
Switzerland-based Novartis, for instance, attributed a 15% drop in its Q1’17 net income to a failed Phase III drug intended to treat heart failure.
The cost of failure is more pronounced for biopharmaceutical startups. With limited cash, a startup’s lead candidate failing during a clinical trial often means the company will not survive. This is because companies rarely IPO until at least one promising drug is in late-stage clinical trials.
Recently, the SPAC trend has enabled companies to improve survival probability by accessing public capital at an earlier, higher-risk stage.
The high costs associated with clinical trials also have downstream effects on costs for patients. This is because biopharma companies bundle R&D costs of failed trials into the pricing of approved drugs to remain profitable.
The current state of clinical trials
After commercially available treatments have failed, patients must navigate a complicated process to find, enroll, and participate in a clinical trial.
In the infographic below, we map out a typical patient journey.
Furthermore, many clinical studies use rudimentary data collection and verification methods — which often put the onus on the patient — such as: sending patient medical records via fax, manually counting leftover pills in bottles, and relying on patients’ diary entries to determine medication adherence.
This process is ripe for disruption.
How AI could change every stage of clinical trials
Artificial intelligence-powered technology has the potential to change every stage of the clinical trials process, from finding a trial to enrollment to medication adherence.
Finding a clinical trial
Matching the right trial with the right patient is a time-consuming and challenging process for both the clinical study team and the patient.
“In fact, only 3 percent of cancer patients today are enrolled in clinical trials.” — WhiteHouse.gov, May 2018
Roughly 80% of clinical trials fail to meet enrollment timelines, and around one-third of Phase III clinical studies are terminated because of enrollment difficulties.
For context, there are over 22,000 clinical studies in the US that are currently recruiting patients. The heatmap below shows the location of 1,000+ active breast cancer studies.
Source: ClinicalTrials.gov
Patients may occasionally get trial recommendations from their doctors if the physician is aware of an ongoing trial. Otherwise, the onus of scouring through ClinicalTrials.gov — a comprehensive federal database of past and ongoing clinical trials — often falls on the patient.
Natural language processing (NLP) can help extract and analyze relevant information from a patient’s EHR records, compare with eligibility criteria for ongoing trials, and recommend matching studies.
In fact, extracting information from medical records — including EHRs and lab images — is one of the most sought-after applications of artificial intelligence in healthcare.
However, solutions accessing patient data face a number of challenges, including unstructured healthcare data and disparate data sources that don’t communicate with each other.
The EHR interoperability challenge
Despite a $27B federally funded incentive program to encourage hospitals and providers to adopt EHRs, there is no standard format or centralized repository of patient medical data.
In fact, it’s still difficult for patients to access their own records from all the health institutions they’ve visited.
Under the Health Insurance Portability and Accountability Act (HIPAA), data sharing is allowed with patient consent. This creates opportunities for AI startups to analyze medical data and suggest eligible patients within minutes — a process that would otherwise take months.
However, issues with securely sharing health information between institutions and software systems — or interoperability — persist. The Covid-19 pandemic has underscored this issue and has driven investor attention to the EHR ecosystem. News mentions for EHRs have also skyrocketed.
Different hospitals and providers treating the same patient may not use the same EHR software to enter data. In many clinical trials, researchers still fax requests for patient records to hospitals, who often send the data back as PDFs or images (including pictures of handwritten notes).
This poses a challenge for AI technology. As one study by researchers from MIT, Harvard, Johns Hopkins, and NYU highlights:
“Standard natural language processing tasks such as sentiment analysis and word sense disambiguation are difficult in clinical notes, which are misspelled, acronym-laden, and copy-paste heavy.”
Health AI company Flatiron Health explains this further in a patent filing: “Structured data can also become unstructured due to transmission methods. For example, a spreadsheet that is faxed or turned into a read-only document (such as PDF) loses much of its structure.”
This dated, manual system makes it difficult for clinical trial researchers to collect accurate data needed to determine a patient’s eligibility.
Startups are approaching the patient recruitment problem from multiple angles.
Deep 6 AI uses NLP to extract clinical data — such as symptoms, diagnoses, and treatments — from patient records. Its software can even identify patients with conditions not explicitly mentioned in EHR data, improving the match rate between patients and clinical trials. Deep 6 AI was valued at over $140M in its latest fundraise.
Clinical trial marketplaces, such as the one SubjectWell offers, are another approach. SubjectWell’s platform allows researchers to access pre-screened patients.
Source: SubjectWell
A smaller group of companies is attempting to work around interoperability hurdles with a direct-to-consumer approach. For example, Clara Health offers a patient-friendly solution to find new treatment options. Its solution not only matches patients to trials but offers ongoing support throughout the process, improving retention as well. The company is backed by Founders Fund and Khosla Ventures.
Established players in other areas of healthcare are also entering the clinical trials recruitment space. 23andMe is now offering recommendations about what studies might be a good fit for its 12M+ customers based on their genetics.
Other emerging consumer-focused recruitment solutions include:
- Disease-specific social networking platforms like Be the Partner
- Open-source frameworks like Google Health Studies, which researchers can use to create their own apps for conducting research studies
Acquisitions as a strategy to get patient data
Flatiron Health tackled the interoperability problem by acquiring an oncology-focused electronic medical record (EMR) company Altos Solutions in 2014.
At the time, Flatiron was selling its cloud analytics platform to healthcare and life science companies, and Altos’ EMR was being used by oncology institutions like Florida Cancer Specialists. The deal gave Flatiron direct access to raw patient data, instead of relying solely on access to third-party EMRs.
Roche subsequently acquired Flatiron for $2B+ in 2018 to gain access to its real-world evidence — insights generated from EHRs, claims, and wearable sensors. Roche plans to use this data to improve its cancer treatment pipeline development.
Meanwhile, biopharma companies continue to partner to gain access to patient data. For example, Janssen Pharmaceuticals partnered with Tempus in November 2020 to gain access to its clinical and molecular patient database and improve patient identification.
Challenges with enrollment
Unfortunately, enrollment challenges do not end when a patient chooses a clinical trial.
To confirm eligibility, the patient must complete a preliminary phone screen and then undergo examination by a participating site in person or virtually.
Every trial includes inclusion and exclusion criteria that each patient must meet in order to participate. These terms are often riddled with medical jargon, as can be seen in the below screenshot of eligibility criteria from a Phase II breast cancer trial.
According to ClinicalTrials.gov, this study began in November 2017 and is expected to end in May 2025, highlighting the time and cost involved in each phase of the trial. Please click to enlarge.
In the above example, patients must go through evaluations, like laboratory and imaging tests, to make sure they meet all the inclusion and exclusion criteria.
Depending on their availability and how far they live from a trial site, some patients may be able to complete these procedures in less than a week. But for others with children, inflexible jobs, or long commutes, the process could take multiple visits.
Telehealth services could help streamline this process. For example, Deaconess Health System partnered with TytoCare in 2020 to integrate the startup’s connected exam tools with Deaconess Clinic LIVE, the health system’s proprietary virtual care platform.
Source: TytoCare
Tyto’s device and accompanying smartphone app allow quarantined patients to carry out their own exams by capturing data from the heart, lungs, throat, ears, skin, and abdomen. They can then share their results in real time with remote physicians.
Platforms like TytoCare may prove to be critical for determining eligibility and conducting virtual clinical trials.
If eligible, the patient signs a consent form agreeing to the terms of the clinical trial. This includes awareness of potential side effects, willingness to provide biological samples, and covering any expenses not included within the study budget.
Solutions using AI to extract information from patient medical records can help simplify the enrollment process by automatically verifying some of the inclusion and exclusion criteria.
Medication adherence
Once patients enroll in a study, they receive the experimental study drug (or placebo).
Patients go home with the first course of the medication (for example, a 30-day pill bottle with instructions on dosage) and a diary to fill out daily. Many clinical studies still use paper diaries instead of electronic systems.
Patients are asked to note when they took the study drug, what other medications were taken on those days, and any adverse reactions (including headache, stomach ache, or muscle aches).
This process is plagued with inefficiencies:
- Reliance on patient memory. When a patient returns to the clinic for check-ins, the study investigator checks their pill bottle to make sure there are no pills left and reviews the patient’s diaries for any blanks or inconsistencies. If there is missing information in the diary entries, the investigator relies on the patient’s memory of events.
- Outdated recording system. Paper documents, which may be misplaced or missing key information, are an outdated and unreliable way to record key data points for a trial.
- Risk of drop out. Frequent travel to a clinical study and research site for regular check-ins is a strain on patients’ time and money, particularly for patients traveling from out-of-state. This heightens the risk of drop out.
- Additional payments. Although out-of-pocket costs are included in the consent document that the patient signs, many patients do not grasp the magnitude of those fees. For example, additional MRI and lab tests during follow-up visits may not be included in the trial — and health insurers may not cover such tests, since they’re for research purposes and not out of medical necessity.
Non-adherence can have adverse effects on a patient’s health, incur costs if a study has to recruit new patients, and interfere with the accuracy of study outcomes.
Generally, adherence rates of 80% or more are required for therapeutic efficacy. However, up to 50% of medications prescribed in the US are taken incorrectly. In response, clinical study sponsors are investing in emerging technology to minimize non-adherence.
Visual, auditory, and digital phenotyping for medication adherence assessment
Some startups are providing visual confirmation of medication administration.
Platforms like AiCure use an interactive medical assistant (IMA) to identify patients at risk of non-adherence based on visual data collection. Patients use their phones to take a video of themselves swallowing a pill, and AiCure confirms that the right person took the right pill.
The company raised a $24.5M Series C led by Baird Capital in late 2019 and partnered with Science 37 in April 2020 to enable virtual clinical trials.
Source: AiCure
Other emerging technologies include digital phenotyping and speech analytics for medication adherence assessment.
ARCH Venture Partners-backed mental health startup Mindstrong uses digital phenotyping technology to measure mood based on how users interact with their mobile device.
Aural Analytics, which uses speech detection to identify subtle changes in users’ brain health, recently partnered with Mass General Hospital for an amyotrophic lateral sclerosis (ALS) trial.
AI and IoT for remote patient monitoring
To enable remote patient monitoring for clinical trials, some startups are developing their own monitoring devices and sensors, then adding a layer of machine learning to interpret the data. Others are only developing the AI software and integrating with third-party at-home monitoring devices.
Connected devices can facilitate real-time study medication adherence. For example, optimize.health (previously Pillsy) launched a smart medication bottle with a corresponding mobile application that provides reminders, educational content, dose tracking, and patient-reported data capabilities for providers. The company, which is now positioned as a remote patient monitoring platform, raised a $15M Series A in August 2020.
Other solutions focus on capturing physiological data. AliveCor’s wearable electrocardiogram (EKG) device applies machine learning to real-time data to detect abnormal heart rhythms, such as atrial fibrillation (AF). In April 2020, the company partnered with Medable to enable decentralized cardiology-focused clinical trials. AliveCor most recently raised a $65M Series E round in November.
Source: AliveCor
Meanwhile, Sequoia Capital-backed Biofourmis uses wearable medical devices to track users’ vitals on its AI platform and provide predictive insights into their health.
The startup partnered with the University of Hong Kong to capture the temperature, heart rate, and oxygen levels of patients infected with Covid-19 to detect subtle changes in health and help accelerate virus research.
Source: Biofourmis
Biofourmis’ platform is used by payers, providers, and pharmaceutical companies to inform decisions on treatment efficacy and to offer personalized care. In April 2020, the startup acquired Gaido Health from Takeda Pharmaceutical to expand into remote monitoring services for oncology.
AI and wearables offer real-time, continuous monitoring of physiological and behavioral changes in patients, potentially reducing the cost, frequency, and difficulty of on-site check-ups.
get the full report
How big tech is disrupting clinical trials
Startups are not the only companies working on tech for clinical trials. Big tech companies are leveraging their mobile devices to build platforms that span across the clinical trial process.
Since 2015, Apple has been building a clinical study ecosystem around the iPhone and Apple Watch, both of which enable real-time health data collection. Its open-source frameworks — ResearchKit and CareKit — help clinical trials recruit patients and monitor their health remotely.
Recently, however, Google has been more active in the space. The company is building a clinical research ecosystem through its Google Health Studies Android application and developing products through its life science subsidiary, Verily Life Sciences.
Through these products, Google can provide medical researchers with streams of patient health data that were not easily accessible until now.
Google’s healthcare data platform
Verily launched Project Baseline in 2017 to fuel medical research by mapping human health. By mid-2019, Novartis, Sanofi, Otsuka, and Pfizer had partnered with Verily to use its tools for more efficient clinical trials. The initiative has also partnered with Stanford Medicine, the Duke University School of Medicine, and the American Heart Association.
Some of Verily’s successes to date include an FDA-approved EKG watch and an FDA-approved irregular pulse monitor.
In December 2020, Google launched a new Android app — Google Health Studies — that streamlines study participation for consumers and provides transparency around how their data is being used for health research.
Source: Google
Already, Boston Children’s Hospital and Harvard Medical School are partnering with Google to enroll Android users in a 100,000-person study of acute respiratory illnesses. The study will utilize survey responses and mobility data to analyze transmission dynamics of pathogens such as Covid-19.
Disrupting EHR data sharing
Google recently launched its Healthcare Interoperability Readiness Program to help healthcare organizations understand the current status of their data and create a path to standardize and integrate across systems.
Source: Google
In April 2020, Google opened up its Cloud Healthcare API to health systems and quickly signed on top medical centers such as Mayo Clinic. These actions follow Google’s 2018 pledge to support healthcare interoperability and data-sharing standards (also signed by Amazon, IBM, Microsoft, and Salesforce).
Google is also working with EHR vendors, including Meditech, to take their systems and data to the cloud. One potential end goal of these partnerships could be a two-way data flow, where EHR vendors are incentivized to integrate patient-generated data into Google software.
What does this data mean for clinical trials?
The widespread adoption of mobile devices has placed Google at the center of the healthcare data ecosystem, offering previously unavailable real-time data while also gathering difficult-to-consolidate EHR information.
The possibilities are seemingly endless when it comes to using AI and machine learning for early diagnosis, driving decisions in drug design, enrolling the right pool of patients for studies, and remotely monitoring patients’ progress throughout the study.
Many trials have an experimental group (patients who get the study drug) and a control group (patients who get a placebo drug). The purpose of a control group is to establish a baseline to compare to the experimental group’s symptoms.
Patient-generated data — like the data from Project Baseline — could help create digital twins and eliminate the need for a control group, further reducing recruitment bottlenecks.
Source: Unlearn
Apple and Facebook’s moves toward clinical research and trials
For Apple, the strategic importance of clinical trials is less clear. Within 3 years of launching its open-sourced ResearchKit and CareKit software, over 500 doctors and medical researchers used the tools for studies involving over 3M participants.
While Apple has continued to partner with pharmaceutical companies like Johnson & Johnson for clinical studies, the ResearchKit website has not been updated since 2018.
Another tech company that may enter the space is Facebook, which launched its Preventive Health tool in late 2019. Given the depth of personal data that Facebook captures and its self-organizing communities around health issues, this may be the first step toward a clinical trial recruitment solution.
Why AI alone isn’t the magic bullet
The healthcare industry leads in AI adoption, experimenting with applications ranging from machine learning-assisted diagnostics to extracting information from electronic health records.
In particular, using AI software to design new drugs has gained momentum, with pharma giant Merck partnering with startup Atomwise and GlaxoSmithKline partnering with Insilico Medicine, among others.
The entire healthcare AI heatmap and analysis is available to CB Insights clients here.
However, AI adoption in the actual clinical trial process is still in its early stages.
Compared to other areas of healthcare, fewer startups are directly targeting clients in the clinical trials space. And in many aspects of clinical trials, the need for digitization precedes the need for AI.
How Covid-19 has affected clinical trial tech adoption
The Covid-19 pandemic has catalyzed the adoption of technologies that can improve the efficiency and cost of clinical trials.
Study design
Adaptive design — which involves a more flexible approach to conducting a trial — has been a key trend as researchers grapple with Covid-19.
While traditional studies can be rigid about key endpoints and dosing regimens before commencing the next phase, an adaptive design allows researchers to modify such metrics as trials progress.
Regeneron Pharmaceuticals‘ and Sanofi’s trial evaluating an antibody treatment for Covid-19 patients followed an adaptive design. The World Health Organization’s SOLIDARITY trial has also adopted this model by using a randomized, non-double-blind model to emphasize speed. (In double-blind studies, neither the researcher nor the patient knows which treatment the patient is receiving.)
Another trend in Covid-19 study initiatives is the use of open research forums to expedite potential findings and conclusions.
Source: Mendel.ai
For example, AI-based clinical trials company Mendel.ai partnered with DCM Ventures to create a Covid-19 search engine where researchers can pull data to run their respective studies. Owkin also launched its Covid-19 Open AI Consortium (COAI), which promotes collaboration in key research areas such as cardiovascular complications.
As clinical trial designs come under the microscope during this global pandemic, these initiatives could provide learning opportunities for researchers as alternative methods to traditional study design practices are challenged.
Virtual trials
The Covid-19-driven adoption of telemedicine and remote monitoring solutions has spurred interest in virtual or decentralized clinical trials.
The remote model is not compatible with all types of clinical trials, such as those that require frequent diagnostic imaging or other in-person assessments. However, this model presents an opportunity to better accommodate patient participation.
The potential benefits of virtual trials include reduced costs, a wider network of eligible patients, and better patient retention rates.
One company in this space is Science 37, which offers end-to-end clinical trial services using its virtual Metasite model. It leverages a network of investigators, mobile nurses, and study coordinators with the aim of making studies more accessible to patients.
It has raised almost $107M in total disclosed funding from investors such as GV, Amgen Ventures, Novartis Venture Funds, and Sanofi-Genzyme BioVentures.
Source: Science 37
In April 2020, the company announced a partnership with Innovo Research to decrease the time required to begin a Covid-19 clinical trial. The initiative plans to do this by leveraging Innovo’s national network of sites and patients and applying Science 37’s platform.
The novel coronavirus has spurred new discussions around how technology can address key gaps in today’s clinical trial landscape. As reliance on these technologies grows, future trials could be reshaped by these software platforms.
If you aren’t already a client, sign up for a free trial to learn more about our platform.