Medical Device Clinical Trials: How Patients Get Enrolled

About 11% of activated trial sites never enroll a single patient, according to a Tufts CSDD analysis of nearly 16,000 sites in drug trials [1], and a device sponsor with an approved IDE faces the same risk of paying to open sites that never enroll. Device studies recruit mainly through the clinicians and sites that already treat the condition, so the recruitment plan is a referral plan first and a media plan second. It is judged by qualified referrals and cost per randomization rather than by clicks. Refero does not run recruitment campaigns. We vet and match the clinical research marketing agencies that do, so what follows is written for the buyer, from what those vendors need and see.

How Medical Device Clinical Trials Find and Enroll Participants

Medical device clinical trials recruit mainly through the clinicians and sites that already treat the condition, then widen with registry listings, patient groups and direct outreach. Call this the referral-first trial.

Device studies reach patients through four channels, and each channel has a different owner.

Table: the four channels device studies use to find participants, who runs each, and when each fits.

The site list caps what every channel can reach, because a device procedure needs trained operators and equipped sites. Even a wearable glucose sensor that patients use at home was tested in a published pivotal accuracy study at 12 US investigational sites with repeated in-clinic visits, and 316 participants were analyzed [2]. A sponsor's reach is therefore the catchment of its sites, and budget spent outside it buys inquiries no site can act on.

An illustrative implantable cardiac device shows how referral-first recruitment works. Its candidates are already under a specialist's care, so a broad advertising campaign would mostly reach people without the indication, while the cardiologists who manage the right patients can name them from their own clinic lists. Recruitment therefore runs through those referring physicians to an implanting site, and advertising's role is limited to patients no cardiologist has yet considered.

Distance matters less when the device is used at home. In a small decentralized study of continuous glucose monitoring in type 2 diabetes, 23 of 26 participants were recruited through online advertising and 3 through a pharmacy [3], though as a feasibility result from one small study it sets no benchmark for pivotal enrollment. FDA's final guidance on decentralized elements says investigational devices intended for home use may be used without the investigator's direct oversight when that oversight is not needed to mitigate potential serious risks [4], so home-use devices can plausibly enroll people who live far from a site; the evidence on decentralized trial enrollment covers that case in depth.

Every channel feeds the same funnel, and a sponsor cannot compare channels until the stages are defined:

1. Referral or inquiry: a clinician or patient raises a hand.
2. Prescreen: a site checks the record against the main eligibility criteria.
3. Consent and screening visit: the patient agrees to take part and tests confirm eligibility.
4. Enrollment or randomization: the patient joins the study.

Sponsors track qualified referrals, screen-fail rate and cost per randomization (recruitment spend divided by participants randomized) because a click or a lead says nothing about whether a patient clears those stages. CASRAI's recruitment guidance points the same way: it treats screen-failure rate as a standard recruitment KPI and reports recruitment activity separately from enrollment [5]. The general clinical trial recruitment benchmarks give the baseline timelines against which a device study's funnel can be read.

Medical Device Clinical Trials at a Glance

Device studies start with a handful of patients and grow to hundreds, so the referral network has to scale at every stage.

• FDA describes an early feasibility study as a limited investigation, usually run in fewer than 10 initial subjects [6]
• Across 28 high-risk therapeutic devices with PMA approval in 2010 to 2011, the median was 65 patients in nonpivotal premarket studies and 241 in pivotal studies [7]
• For 27 priority-review device applications approved from 2006 to August 2013, pivotal trials had a median enrollment of 297 and a median duration of 3 years [8]
• As of April 2024, FDA's early feasibility program had approved over 430 study IDEs covering over 4,000 participants [9]

Does Your Device Need a Clinical Study, and How Many Patients?

The device's risk class and regulatory pathway decide whether it needs a clinical study: a 510(k) clearance rests on substantial equivalence to a legally marketed predicate device, supported by performance data that can be clinical or non-clinical bench data, so a new study is not always required [10]. A significant-risk device study needs an FDA-approved IDE before it enrolls anyone, while a non-significant-risk study can proceed under IRB approval without a separate IDE application [11]. A device with no study needs no enrollment plan, and a device with one needs the plan sized to its stage.

Each stage carries a different recruitment consequence. An early feasibility study is a limited investigation, usually in fewer than 10 initial subjects, so it needs a few specialist sites and hand-picked patients [6]. A pivotal study is run in a statistically justified number of subjects, with a median enrollment of 297 in the priority-review device applications above [8], so the referral network must scale across many sites [6]. Post-market studies and EU MDR clinical investigations under Regulation (EU) 2017/745 [12] change the site list again.

This section is not legal advice; confirm the pathway and study classification with regulatory counsel and the IRB of record.

How Device Trial Recruitment Differs From Drug Trials

Device recruitment differs from drug recruitment in three ways: device studies draw smaller, procedure-bound populations through specialists, use designs that change between studies, and rely on controls other than a placebo pill.

The popular claim that device trials need fewer participants holds only as a tendency in published samples. The 188 novel drugs FDA approved from 2005 to 2012 rested on pivotal trials with a median of 760 patients enrolled per indication [13], while pivotal studies for the high-risk devices approved in 2010 to 2011 had a median of 241 patients [7]. These figures come from two studies with different samples, so they do not form a ratio, yet the device figure shows a recruiter working a smaller pool per study, where each referring clinician carries more weight.

Iterative design adds a second difference, because FDA expects the device and the protocol to change during an early feasibility study and allows more changes under 5-day notification than in other study types [6]. Revising the device between the feasibility and pivotal studies can change eligibility and the operators needed. An illustrative redesigned delivery system may require operators trained on the new version, and that requirement shrinks the usable site list until the operators are trained, so the referral network built for the first study has to be rebuilt.

Control design is the third difference, and it changes what patients are told and will accept. In a comparison of FDA ophthalmic approvals from 2002 to 2012, only 16% of device pivotal trials were randomized and 12% masked, against 91% of drug trials for both [14]. A patient offered a device study may face a sham procedure, an active comparator or a single-arm design instead of a placebo pill, so the recruitment message must explain the design plainly.

Where Enrollment in Device Studies Stalls

Enrollment in device studies most often stalls at the site rather than for lack of interested patients. This position draws on general trial data, because published figures are not device-specific and vary by source.

The Tufts CSDD finding that about 11% of activated sites enroll no one [1] shows that opening a site does not guarantee it can enroll. Device studies concentrate that risk, because the trained-operator and equipment requirements keep the site list short, especially in early feasibility studies [6], so one idle site removes a large share of a sponsor's reach.

Three device-specific drivers sit behind that risk: operators trained on the investigational device, referral pathways that run through physicians outside the trial site, and equipment or imaging that only some sites have. Patient mix is a fourth. In the eight US trials of FDA-approved hybrid closed-loop and continuous glucose monitoring devices in type 1 diabetes that reported race, 84.5% of the 1,354 participants were non-Hispanic White [15], and at one academic center baseline glucose-monitor use was 7.9% for Black adults against 30.3% for non-Black adults [16]. A plan that depends on patients who already use the device therefore draws on a skewed pool, and the diversity evidence for trials treats that skew as a recruitment constraint.

What Patients Ask Before Joining a Medical Device Clinical Study

Patients weighing a device study are deciding about an implant or procedure they cannot stop like a pill, so a recruitment plan that cannot say what the device is and answer their questions quickly loses people at consent.

The first answer the plan must give is what the device is. FDA's guidance on recruiting study subjects treats recruitment advertising as part of the consent process, requires the IRB to review it, and says materials must not claim or imply that the investigational device is safe or effective, or equivalent or superior to other treatments [17]. Ad platforms apply their own clinical trial advertising policies as a separate clearance, so IRB approval does not guarantee an ad will run [18].

The second answer is speed: when a named person at the site or a call center can answer a patient's questions during the first contact, the inquiry does not go cold before prescreen. For home-use devices, 85% of participants in the small decentralized glucose-monitoring study preferred an online consent conversation [3], and that preference suggests remote consent is worth offering. Remote contact still takes time: in a fully virtual glucose-monitoring study, visit time for participants aged 65 and older was 41 minutes longer than for those under 40 [19], so a plan that relies on remote onboarding should budget longer first contacts for older patients.

How to Choose and Brief Recruitment Help for a Device Trial

A sponsor should decide which recruitment work stays with sites and the CRO, set funnel metrics before launch, and check any vendor's device experience and patient handoff.

Decide Early Which Recruitment Work Stays With Sites or the CRO

If sites fall below projected referrals at the first enrollment checkpoint, a specialist recruitment agency is worth adding. Sites and the CRO keep the referral networks, EHR searches and prescreen, because those depend on clinical relationships and records an outside party cannot reach. A specialist agency fits outreach and inquiry handling for patients no referring clinician has identified, and its work must hand patients to a site that has the capacity to prescreen them.

Set Funnel Metrics Before Launch, by Source

If a vendor proposal is priced or reported per lead, agree on the metrics before the contract: qualified referrals, screen-fail rate and cost per randomization, each broken out by source. These metrics measure patients who reach enrollment, so they expose a cheap channel that sends unqualified referrals.

Check Device Experience and the Patient Handoff With Five Questions

Put these five questions to every vendor before shortlisting:

1. Which device studies have you recruited for, at what stage and with what kind of sites?
2. How do referrals reach the sites, and what route do referring clinicians have besides the patient calling in?
3. Who owns prescreen, and what is the response time to a patient inquiry?
4. Do you report by funnel stage and source, including screen failures?
5. How do your materials state that the device is investigational, and who submits them to the IRB?

If you would rather not run that evaluation alone, Refero screens healthcare marketing and patient recruitment agencies against five published criteria and introduces up to three that fit your brief. It is free for buyers, and there is no obligation to hire anyone we introduce. Tell us what you need.

Frequently Asked Questions

Do medical devices go through clinical trials?

Some do, and risk class and regulatory pathway decide which. A 510(k) clearance rests on substantial equivalence to a legally marketed predicate device, supported by performance data that can be clinical or non-clinical bench data, so a new clinical study is not always required [10]. A significant-risk study also needs an IDE from FDA before it enrolls anyone [11]. Confirm your own pathway with regulatory counsel.

What are the stages of a medical device clinical trial?

Device studies typically move from early feasibility, a limited investigation in a small number of subjects, to a pivotal study run in a statistically justified number of subjects [6]. Post-market follow-up comes after approval or clearance. Each stage usually adds sites and patients, so the referral network built for the first stage has to be extended or rebuilt for the next.

How are patients recruited for medical device trials?

Mostly through referring clinicians and the trial sites' own patient lists and EHR searches, with registry listings, patient groups and direct outreach widening the pool. Vendors help most with outreach and inquiry handling. Sponsors judge the mix by qualified referrals, screen-fail rate and cost per randomization, because clicks and leads do not show who enrolls.

Do medical device trials need fewer participants than drug trials?

Published samples often show smaller device studies, but no rule sets the number. In one comparison of FDA ophthalmic approvals from 2002 to 2012, drug pivotal trials enrolled a median of 809 patients and device pivotal trials a mean of 324 [14]. Sample size follows the statistical design and the intended population.

Are participants paid in medical device clinical trials?

Payment and travel reimbursement vary by study, and the IRB reviews both the amount and how it is described to participants. FDA's guidance says payment should not unduly influence someone to join and should accrue as the study proceeds rather than depend on completion [20]. Ask the IRB of record before fixing a figure.

Sources

• Tufts Center for the Study of Drug Development. Enrollment Performance: Weighing the Facts. Applied Clinical Trials, 2012. Link
• Garg SK, et al. Accuracy and Safety of Dexcom G7 Continuous Glucose Monitoring in Adults with Diabetes. Diabetes Technology & Therapeutics, 2022. Link
• Ali, et al. Exploring Decentralized Glucose and Behaviometric Monitoring of Persons with Type 2 Diabetes in the Setting of a Clinical Trial. Journal of Diabetes Science and Technology, 2023. Link
• US Food and Drug Administration. Conducting Clinical Trials With Decentralized Elements (final guidance), 2024. Link
• CASRAI. Clinical Trial Patient Recruitment guide. Link
• US Food and Drug Administration. IDEs for Early Feasibility Medical Device Clinical Studies, Including Certain First in Human (FIH) Studies (final guidance), 2013. Link
• Rathi VK, et al. Characteristics of clinical studies conducted over the total product life cycle of high-risk therapeutic medical devices receiving FDA premarket approval in 2010 and 2011. JAMA, 2015. Link
• Rising J, Moscovitch B. The Food and Drug Administration's priority review of medical devices: pivotal trial characteristics. PLoS One, 2015. Link
• Medical Device Innovation Consortium. A Decade of Innovation in Medical Device Testing, 2024. Link
• US Food and Drug Administration. Premarket Notification 510(k). Link
• Electronic Code of Federal Regulations. 21 CFR Part 812, Investigational Device Exemptions. Link
• European Union. Regulation (EU) 2017/745 on medical devices. Official Journal of the European Union, 2017. Link
• Downing NS, et al. Clinical trial evidence supporting FDA approval of novel therapeutic agents, 2005-2012. JAMA, 2014. Link
• Hwang, et al. Pivotal clinical trials of novel ophthalmic drugs and medical devices: retrospective observational study, 2002-2012. BMJ Open, 2015. Link
• Inequity in Racial-Ethnic Representation in Randomized Controlled Trials of Diabetes Technologies in Type 1 Diabetes: Critical Need for New Standards. Diabetes Care, 2021. Link
• Racial Disparities in Access and Use of Diabetes Technology Among Adult Patients With Type 1 Diabetes in a U.S. Academic Medical Center. Diabetes Care, 2023. Link
• US Food and Drug Administration. Recruiting Study Subjects (information sheet). Link
• Google Ads Help. Healthcare and medicines policy. Link
• Weinstock, et al. Older Adults Benefit From Virtual Support for Continuous Glucose Monitor Use But Require Longer Visits. Journal of Diabetes Science and Technology, 2026. Link
• US Food and Drug Administration. Payment and Reimbursement to Research Subjects (information sheet). Link