Remote Patient Monitoring: How RPM Delivers Stronger Clinical Trial Evidence

The emergence of sensors and wearables is allowing clinical trial sponsors to move beyond point-in-time to continuous patient data capture. But, this large volume of data also poses a problem for study teams: How can these massive quantities of data be effectively monitored to keep patients compliant and safe?  

The answer is artificial intelligence.  

This article takes a look at the benefits of remote patient monitoring, as well as how artificial intelligence is evolving the RPM landscape to streamline data management and analysis.

But, before we dive in, let’s take a look at the origins of remote patient monitoring.  
‍

What is Remote Patient Monitoring?

Remote patient monitoring is the process of remotely collecting and monitoring data using digital tools outside of a clinical environment—while patients are living their lives in the “real world.”  

Most people trace the origins of RPM programs back to clinical care, where it has been used for more than 30 years to track physical systems, chronic conditions, or post-hospitalization rehabilitation. But, in fact, the history of RPM goes back even further. Remote patient monitoring was most famously introduced in 1961 to monitor astronaut Alan Shepard using a primitive EKG, respiration sensor, and thermometer. A decade later, the medical industry began incorporating RPM to telecommunicate medical assessments and consultations by sending x-rays, CT/MRI scans, and lab data over telephone wires.  

The arrival of the internet transformed remote patient monitoring, making it possible to provide data to healthcare providers regarding the health statuses of their patients—particularly those with chronic conditions—when outside the clinic or hospital. When the COVID-19 pandemic struck, the use of remote monitoring accelerated significantly—and continues to do so today. In fact, Insider Intelligence estimates that 70.6 million U.S. patients, or 26.2 percent of the population, will use RPM tools by 2025.  

RPM use in clinical trials has also accelerated rapidly with the advent of decentralized clinical trials, which reduce the frequency of in-person clinic visits to monitor patient progress and well-being.  
‍

A note about the relationship between RPM and telehealth

RPM is not synonymous with telehealth. The latter involves conducting medical visits, administering treatment, or determining diagnoses using telecommunication technologies. The former involves the collection, storing, and monitoring of relevant data and its transmission to medical professionals to gain insights. Telehealth and remote patient monitoring are often incorporated into decentralized clinical trial designs.
‍

‍RPM as a Solution for Common Clinical Trial Challenges  

Patients are the heart of clinical trials. But, patients are human and make mistakes. They might forget to perform study tasks, make an error when inputting data, or downplay a side effect of the intervention. These common challenges of clinical trial patient reporting can degrade data integrity.  

Remote patient monitoring clinical trials can help to overcome these issues. Let’s take a look at how.  
‍

More Complete and Real-World Data via Sensors and Wearables  

Collecting error-free and complete data can be challenging when reporting is left largely to the patients. And, even when patients are meticulous in their reporting, an additional hurdle to real-world accuracy is the punctual nature of data capture, which often represents only a snapshot view of the patient’s life, rather than a holistic picture. Temperatures, blood pressure, and glucose levels can go up and down frequently and sometimes radically; if recorded too punctually, these real-world variations may not show up, thus providing only a partial depiction of the participant’s health.

Enter sensors, wearables, and digital biomarkers. These digital devices can capture data from patients in real-world settings—at home or as they go about their daily lives—providing sponsors with more data collected over a more frequent or even continuous time period. RPM devices like blood pressure monitors or pulse oximeters may not require the patient to manually input any data, because it can be transmitted via DCT integrations.

Remote data collection is most beneficial when accompanied by remote patient monitoring, as discussed below.  
‍

AE Detection  

Patients may experience certain symptoms between clinic visits and choose to ignore them, either because contacting a clinician is too much of a hassle or because the patient doesn’t believe the symptom to be sufficiently intense or relevant for reporting. Severe adverse events (SAEs), which may require urgent intervention, can be missed if patients don’t understand the severity or if they aren’t able to easily access their provider. Sensors and wearables, which capture spikes in temperature or blood pressure, can significantly improve patient safety if detected by a study team who is carefully monitoring this data.  

A trial design that incorporates remote patient monitoring often provides patients with virtual or remote access to the study team via text, email, and phone, and this easy access can encourage patients to report symptoms in near real time, rather than waiting for a site visit.  
‍

Protocol Adherence  

Data quality in clinical trials depends on patient engagement—participants’ completion of their study tasks. But, as many as 40 percent of patients become nonadherent to protocols after 150 days in a clinical trial. In some cases, this can be dangerous for patients; for example, taking too much or too little medication may cause temporary bouts of toxicity due to double-dosing or lack of efficacy due to skipped dosing. Compliance lapses, of course, also introduce data variability.  

When patients are monitored in real time, these compliance lapses can be detected quickly, enabling study teams to reach out to patients and encourage them to remain on track.
‍

AI is Advancing Remote Patient Monitoring in Clinical Trials  

Artificial intelligence (AI) and machine learning (ML) have made inroads into many sectors, increasing the efficiency of computing and data analysis. The clinical trials industry is no exception.  

The introduction of RPM tools such as sensors, wearables, and digital devices can help to power a study. However, the volume of data these devices produce can also overwhelm study teams. Having more data than researchers can reasonably manage or analyze may actually be detrimental to a study. That’s where artificial intelligence and machine learning come into play.
 

AI Streamlines Study Team Monitoring  

Study teams typically monitor patients remotely through the use of a decentralized clinical trial platform. Data captured from patients flows into these platforms, where it can be viewed on dashboards in real time. Artificial intelligence can help study teams by sifting through all of the captured data and surfacing information that needs further attention. This includes alerting the study team to the following data issues.  

Anomaly detection

Outliers—data points that are too extreme and deviate from the mean—can skew statistical results if they aren’t identified and verified. The platform is programmed to detect these outliers to determine whether they are indeed anomalies.

Score discordance

Patients can be monitored by a remote team even when eClinRO is included in the study design. One common issue with clinician ratings is that they can lack consistency, even for a single rater. Artificial intelligence can be trained to monitor these scores and alert the study team when rater variability is detected. The study team can then ask the clinician to re-evaluate.  

AE/SAE alerts

AE/SAE detection is one of the most important features of a monitoring platform, because it keeps patients safe. The system is programmed to immediately alert the study team if a pre-programmed threshold is met or if the patient indicates a change in health status.  

Non-compliance notifications

In cases of noncompliance, the technology can be programmed within the workflow to send patients automatic reminders about pending task completion. This frees up study teams—unless compliance continues to lapse.  

Trend recognition

AI pattern recognition functionalities can help to identify specific points in the protocol where patient compliance starts to drop off or where patients are commonly making data input errors, allowing trial designers to quickly spot and fix study design issues without allowing too much time to elapse.

While artificial intelligence can improve the workflow for study teams and enable them to track high volumes of patient data, it would be unwise or even dangerous at present to rely solely on AI to evaluate data. Instead, a combination of AI functionalities and human review can significantly speed the process of analyzing data and drawing important insights.  
‍

How Can ObvioHealth Help with Remote Patient Monitoring?

At ObvioHealth, we design and run decentralized clinical trials, including study designs that incorporate sensors, wearables, and remote patient monitoring. Our solution can be integrated with over 300 consumer wearables and most medical-grade FDA- or CE-cleared sensors. ObvioHealth studies can be monitored either by trained site teams or our virtual COACH (Clinical Oversight And Coordination Hub) team.  

Our ObvioGo^® platform has everything needed to run an end-to-end DCT, including patient onboarding, eConsent, data collection, processing, and analysis.  

Learn more about how the ObvioGo platform can help you collect stronger evidence by booking a demo today.