Device Trial Pre-Screening: The Questions That Cost You Eligible Patients

Some of the eligible patients a device trial needs are lost before anyone at the clinical site or the sponsor ever speaks to them. The patient meets the criteria, reaches a screening question the patient can’t confidently answer and leaves the screener. The loss never shows up in a report. The patient simply never becomes a referral.

Here is how it happens. A patient with the right heart condition reads a screener question: “Have you been told you have a specific subtype of heart failure?”
The patient isn’t sure, answers no and leaves. The clinical trial loses a patient who may have qualified, and the sponsor has no way of knowing why.
Device trials ask patients for more than a diagnosis. Many require a prior procedure, a recent test result, a treatment the patient stopped or a device used within a set window. Some of those details patients know. Many sit in medical records the patient has never read.

For sponsors, the challenge is finding the screening questions that create unnecessary barriers before the clinical trial launches, not after enrollment falls behind. That starts with reading the screener the way a patient reads it.

Paul Egli, Senior Vice President of Sales and Business Development at 83bar, does exactly that when he reviews a new client’s patient pathway:

“I try and go through myself as a patient to see what is that experience… How much knowledge is required of me. How much clinical understanding do I have to have to move forward. And I think a lot of times there’s a miss there.”

What happens when patients can’t confirm a criterion

Antidote, an 83bar company, ran a feasibility test for a heart failure with preserved ejection fraction (HFpEF) study. The study targeted a common umbrella diagnosis, then screened for a specific subtype and a clinical threshold.

Screening step Share of respondents
Reported the umbrella heart failure diagnosis 59.2%
Reported the specific subtype 56.2%
Met a key clinical threshold 44.4%

Of 26 survey submissions, two patients qualified. Interest was not the problem: the patients who reached the end of the screener were highly interested in the clinical trial. The drop came at the steps that asked for more clinical detail, a heart failure subtype and a test-based threshold, which many patients hear once in a physician’s office and may not be able to confirm on a form.

For the sponsor, a criterion that patients can’t confirm hurts patient recruitment efficiency in two directions:

What happens What it costs the sponsor
A qualified patient answers “no” or quits at a question the patient can’t answer The sponsor loses an eligible patient the campaign already paid to reach, and the loss never appears in a report
An unqualified patient guesses “yes” and gets referred The sponsor pays for a referral that fails clinical site screening after using site staff time, an appointment and often imaging or lab work

Both outcomes raise the cost of every randomized patient. The same advertising spend produces fewer eligible patients, and clinical sites spend more of their screening capacity on patients who were never going to qualify. Across a full enrollment period, that gap shows up as slower enrollment and a higher cost per randomized patient.

Sort every criterion into three groups

Before the screener is written, sort each inclusion and exclusion criterion by who can reliably confirm it.

Group Device trial examples Where to confirm it
The patient knows it Age, a diagnosis a physician confirmed, height and weight, willingness to attend clinic visits Online pre-screener
The patient roughly knows it and needs help to confirm it Heart failure subtype, prior therapy and when the patient stopped it, date of last device use, time on a current medication dose Phone call from a clinical contact center
The patient rarely knows it Test values, imaging measurements, documented results from a prior sleep study Medical records or clinical site screening

Three device-related feasibility tests from Antidote, an 83bar company, show how this sorting plays out in screener design. The numbers don’t measure patient memory directly. They show which details a screener can ask for, which details work better as a calculation and which details are worth confirming by phone before a patient is screened out.

Feasibility test Criterion What it shows about question design
Sleep apnea device study Prior sleep study and prior positive airway pressure (PAP) use 86.3% reported a completed sleep study and 51.5% reported stopping PAP use. These questions ask about events the patient lived through, which makes them a good fit for the online screener. The sleep study result and the reason the patient stopped PAP belong in the phone call.
Sleep apnea device study Body mass index (BMI) of 32 or lower The screener asked for height and weight rather than BMI, so the patient never had to know the number. 36.8% were excluded for a BMI above 32, and the screener caught it with a calculation instead of a question the patient might guess at.
Parkinson’s disease study Use of a specific medication delivery device in the past four months Only 4.2% reported recent device use. A four-month lookback depends on a date many patients don’t track, so a date range plus a phone call for borderline answers can keep eligible patients from screening themselves out.

The pattern across all three: patients reliably report events, such as finishing a sleep study or using a device. Results, measurements and exact dates are less reliable. The screener should ask about the event, and the confirmation step should handle the detail.

How can sponsors write patient-friendly device trial screening questions?

  • Ask what the patient experienced, not the clinical term. “Have you had an overnight sleep study, and did a doctor go over the results with you?” is easier to answer than a question built on diagnostic terms.
  • Add an “I’m not sure” answer and route it to a phone call. A patient who isn’t sure should hear from someone who can help, not be screened out by a form.
  • Ask for numbers the patient knows, then calculate the threshold. Height and weight produce BMI. Approximate dates produce a lookback window.
  • Use date ranges for lookback windows. “Within the past three months” or “four to six months ago” is easier to answer accurately than an exact date.
  • Keep the phone script consistent with the screener. The clinical contact center should confirm the same criteria the screener flagged as uncertain. Have the clinical and regulatory team confirm which patient-facing materials and scripts require review or approval before use.

Paul points to why this matters at the very start of the pathway:

“A lot of these patients are suffering, but they don’t know what they actually have when they’re searching.”

Why this shows up in randomization

Every criterion confirmed before referral is one fewer reason for a patient to fail at the clinical site. In one cardiovascular device program, 83bar referred 15,000 patients. More than 90% of randomizations came from those referrals, and enrollment finished on time. For the metrics that track this from first response to randomization, see From Referral to Randomization: The Metric That Predicts Device Trial Enrollment.

Frequently asked questions

  • How can a sponsor tell which criteria are causing unnecessary drop-off? Look at where patients leave the screener and where “I’m not sure” answers cluster. When patients drop at one criterion while showing strong interest elsewhere in the screener, the question is often the barrier, not the patient’s eligibility. Testing the screener with real patient responses before launch surfaces these points early.
  • When should a criterion move from the online screener to a phone call? Move it when the answer depends on a clinical term, a test result or an exact date, or when a large share of patients choose “I’m not sure.” A trained team member on the phone can help the patient find the answer instead of the form screening the patient out.
  • Does moving criteria to a phone call add work for clinical sites? Not when the call happens before referral. The clinical site receives patients already checked against every criterion the patient can verify, so site screening can focus on test values, imaging and other documented measurements that still need medical records.

Find the screening barriers before they cost you enrollment

Every eligible patient lost at a confusing question is a patient the sponsor already paid to reach. 83bar reviews device trial criteria with medtech sponsors before launch, so the online screener asks what patients know, the clinical contact center confirms what patients need help with and clinical sites receive patients ready for screening.