A lower screen-failure rate looks like better screening.
And sometimes that’s exactly what it means.
But the number leaves out a decision that can change the entire result: who was allowed to enter formal screening in the first place.
Before a candidate can formally fail screening, someone has to advance that candidate from pre-screening.
Those who advance and enter formal screening become part of the population used to calculate the rate. Those who don’t advance remain outside the calculation, regardless of how many candidates were considered before that decision.
The calculation captures what happened to the people who made it through. It doesn’t carry forward the full story of how they were selected.
And that’s how a perfectly accurate percentage can create a misleading impression of performance.
If advancement becomes more selective, candidates judged more likely to fail may never reach formal screening. The remaining candidates can produce a lower failure rate without any improvement in outcomes for a comparable group.
A falling rate could mean fewer failures during formal screening or a change in how candidates were selected for formal screening.
The number itself can’t tell you which.
The denominator was selected before It was counted
There’s a difference between a candidate who doesn’t advance and one who fails formal screening.
The first is a pre-screening decision. The second is a recorded formal screening outcome.
A candidate stopped before screening hasn’t formally failed anything. No formal outcome exists to record.
And that distinction matters because the screen-failure rate measures failures among candidates who were actually screened. It doesn’t measure potential failures among everyone considered beforehand.
Now consider two reporting periods.
Both periods use the same screening definitions and failure-rate calculation, but differ in the share of candidates advanced from pre-screening.
The second period can report a lower failure rate without improved outcomes for comparable candidates. The calculation stayed consistent. The population being measured didn’t necessarily stay comparable.
That’s the problem with treating numerical consistency as proof of operational improvement.
And there’s another part of the story the percentage doesn’t tell you.
Volume.
A lower failure rate can accompany fewer candidates entering formal screening. In that case, the lower rate may reflect more selective advancement into formal screening without indicating that a greater share of the original candidate pool meets screening requirements.
At one hospital in the Light Touch Trial, automated and manual pre-screening narrowed a pool of more than 13,000 patients to 647 pre-qualified candidates. All 48 candidates who attended formal screening there met the eligibility criteria. That’s a zero eligibility-based failure rate among those screened, but it doesn’t establish what would have happened to the candidates who never reached that stage. [1]
Even unchanged formal screening volume doesn’t settle the matter. If the number considered during pre-screening has changed, advancing the same number of candidates may represent a different selection pattern.
The number who advanced tells you little without knowing how many candidates were considered.
Pre-screening records show who was selected, not who would have failed
This is where records from the earlier stage become useful.
Start with the number of candidates considered before formal screening and the number allowed to advance. Together, those figures tell you what proportion of the initial pool was allowed to advance toward formal screening.
The number not advanced counts the candidates whose potential formal screening outcomes never became part of the reported rate.
Records of advancement decisions can help explain how that population was selected. [2]
These figures don’t make the formal failure rate less useful. They make it possible to understand what the rate actually represents.
But there’s a limit to what those records can establish.
The candidates who didn’t advance were never formally screened. Their formal screening outcomes remain unknown.
Someone may have assessed a candidate as likely to fail. That doesn’t mean the candidate formally failed screening. And treating every excluded candidate as a failure successfully avoided would be no more defensible.
Pre-screening records establish which candidates advanced and which didn’t. They can’t establish what would have happened if the excluded candidates had entered formal screening.
That distinction grounds the analysis in observed decisions, not hypothetical outcomes.
Don’t give formal screening credit for what selection may have changed
The useful comparison starts by looking at advancement decisions alongside the formal failure rate.
Suppose the failure rate declines while fewer candidates advance from a similar pre-screening pool. That gives leadership a reason to examine whether more selective advancement explains the better-looking result.
Now suppose advancement patterns remain stable while the proportion failing formal screening declines. That gives stronger support to the possibility that outcomes within formal screening have improved.
Neither pattern delivers a definitive answer.
Advancing the same proportion of candidates doesn’t guarantee that the candidates selected are comparable. And a change in advancement decisions doesn’t rule out a simultaneous change in formal screening outcomes.
Both stages can influence the result.
The trouble starts when leadership treats the downstream percentage as a complete measure of screening performance.
It isn’t.
The reported rate tells you the share of recorded failures among candidates who entered formal screening. That’s a legitimate measurement, and nothing about earlier selection decisions makes the arithmetic invalid.
But the rate doesn’t show how much of the change came from formal screening and how much may reflect pre-screening decisions.
Without examining advancement decisions, leadership can give formal screening credit for an improvement that may have occurred before the measured population was even defined.
A lower rate is still a lower rate.
The percentage alone doesn’t establish whether it represents better screening across a comparable population.
Sources:
[1] Springer Nature (Eye). Leveraging structured EMR data for efficient patient prescreening: a practical approach to reducing screen-failure rates in Light Touch Trial. Published journal comment documenting a trial-site implementation, 4 February 2026. Journal article
[2] Wolters Kluwer Health (Sexually Transmitted Diseases). Impact of Eligibility Criteria on Participant Enrollment for a Randomized Clinical Trial of Gonorrhea Treatment. Peer-reviewed clinical-trial implementation study, June 2017. PubMed-indexed research article