
A screener question can match the protocol word for word, pass institutional review board (IRB) review and still lose eligible patients. The problem shows up when the question asks for something the patient doesn’t know, such as a subtype the physician never named, a test result the patient never saw or the exact date a medication dose last changed.
When a patient can’t answer a question, the screener can lose that patient in two ways. The patient answers “no” and drops out, or guesses “yes” and fails screening at the clinical research site weeks later. Neither outcome appears in a recruitment report as a screener problem.
Antidote’s feasibility tests are short online surveys that put a study’s screening questions in front of people from the target patient population before recruitment begins. In one test for a heart failure with preserved ejection fraction (HFpEF) study, each screening step asked patients to recall something more specific.
Interest stayed high among the patients who got through: 87.5% were very interested in taking part. The respondents were motivated. The questions asked for clinical details many patients don’t carry with them.
Across Antidote’s feasibility tests, protocols that relied on patients recalling these details triggered early self-disqualification, even among motivated respondents.
The lung cancer result points to a second issue. When a screener tells patients to “check with your doctor” before answering, fewer than half of the respondents in that test said they were comfortable asking. Patients who aren’t comfortable asking may leave the screener.
Sponsors test screening questions by running the draft screener with people from the target patient population before recruitment begins. The results show which questions patients can answer on their own and which need confirmation by phone, through medical records or at the clinical research site. Five steps make that test useful.
Review the draft protocol and mark each criterion that asks for a subtype, a test value, a prior test or a date. These are the questions most likely to lose eligible patients.
“Has a doctor told you your heart pumps normally but is stiff?” is easier to answer than a subtype name. Ask for height and weight instead of body mass index (BMI), and offer date ranges instead of exact dates.
A patient who isn’t sure should move forward for confirmation by phone or through medical records, rather than be screened out by the form.
A short feasibility test shows where patients drop off, which questions draw the most “I’m not sure” answers and which criteria screen out the most respondents.
Keep criteria that patients answer reliably in the online screener. Move uncertain ones to phone confirmation, and leave test values and imaging to the clinical site.
For many criteria, yes: age, a confirmed diagnosis, height and weight. Accuracy drops for subtype diagnoses, numeric thresholds, prior test results and exact treatment dates. Those usually need confirmation by phone, through medical records or at the clinical site.
For criteria that depend on recall, an “I’m not sure” option keeps eligible patients from screening themselves out. The answer should send the patient to a confirmation step, not end the screener.
Patient-facing screening materials are often reviewed with the recruitment materials. Your compliance team should confirm what your IRB requires before you change screener wording.
Antidote runs short feasibility tests that put protocol-based screening questions and draft patient-facing messages in front of relevant patient populations before patient recruitment begins. Each test returns a criterion-level breakdown of responses that shows sponsors:
From those results, Antidote recommends which criteria to keep in the online screener, which to move to contact center clarification and which to leave for clinical confirmation at the site.