How many patients qualify for your trial? How to count eligible patients
The number of patients who qualify for a trial is almost always far smaller than the number who have the disease. The reliable way to know it is to count patients against each criterion of the protocol in real, patient-level data, rather than estimate it from totals or site surveys.
- Qualifying patients are those who meet every inclusion and exclusion criterion, not just the diagnosis.
- Each criterion removes its own share, so the qualifying pool is usually a small part of the disease total.
- Prevalence estimates and site surveys tend to overstate it.
- Counting criterion by criterion in patient-level data gives the real number.
- Count before the protocol is final, and again before the bid.
The problemWhy the diagnosis total misleads
Prevalence tells you how many people have a condition. A protocol asks for much more than the condition: an age range, a disease stage, a prior treatment, a lab value inside a threshold, no other serious illness, no medicine that interferes. Each criterion removes its own share of patients, and together they often remove most of them. A site that treats thousands of people with the diagnosis may treat only a handful who qualify.
The optionsThree ways to estimate eligible patients
| Approach | How it works | Where it falls short |
|---|---|---|
| Prevalence and literature | Start from how common the disease is and apply assumed percentages for each criterion | Assumptions stack up, and the result says nothing about which sites treat the patients |
| Site questionnaires | Ask candidate sites how many patients they could enroll | Self-reported, often optimistic, and slow to come back |
| Counting in patient-level data | Check each criterion against de-identified records, patient by patient | Needs patient-level data that follows patients over time, and some criteria are not recorded in routine care |
Only the third gives a count. The first two give estimates, and estimates are where enrollment shortfalls start. See the cost of guessing.
The detailWhich criteria can be counted
Most criteria map to something routine care records: diagnoses, procedures, medicines, lab results and age. Some do not, such as a patient’s willingness to join a trial or the result of a test that is rarely ordered. A good answer says plainly which criteria were counted and which still need checking at the site, so nobody mistakes a partial count for a full one.
The next stepFrom patients to sites
A qualifying count matters most when it is tied to the institutions and investigators who treat those patients. That turns a national number into a site list, and lets you weigh each investigator’s record on similar trials alongside the patients they can reach. See site selection on real patient data.
The timingWhen to count
- Before the protocol is final. A sponsor can still change a criterion that removes too many patients. See the Protocol Stress Test.
- At the RFP. A CRO can size the site plan on counted patients rather than the sponsor’s assumptions.
- Before the bid defense. The sponsor will ask how you know the sites will enroll. See bid defense.
- When a trial is behind. Counting shows where the missing patients are treated before rescue sites are chosen.
Counting eligible patients, answered
How do you estimate how many patients are eligible for a clinical trial?
The reliable way is to count rather than estimate: check each of the protocol’s inclusion and exclusion criteria against de-identified, patient-level real-world data and count the patients who meet all of them. Estimates built from disease prevalence or site questionnaires are usually far too high, because they never test the criteria against real patients.
Why do eligibility criteria shrink the patient pool so much?
Each criterion removes its own share of patients: an age range, a disease stage, a required prior treatment, a lab threshold, another condition or a current medicine. Applied together they often remove most of the people who have the diagnosis, which is why the qualifying pool is usually far smaller than the disease total.
What is a screen failure rate?
The share of patients screened for a trial who turn out not to meet its criteria. A high screen failure rate usually means the site was chosen on disease totals rather than on patients counted against the protocol.
Can real-world data count eligible patients?
Yes, for criteria that routine care records: diagnoses, procedures, medicines, lab results and age. Some criteria are not recorded, such as a patient’s willingness to join a trial, so a good answer says which criteria were counted and which still need checking at the site.
How long does it take to count qualifying patients for a protocol?
With patient-level data already in place it takes days, not weeks. Quantum Research Cloud, a full-service data company, returns counted qualifying patients for a protocol, the sites that treat them and the investigators behind them in days.