Sponsors who come to us with a healthy-volunteer collection often open the conversation by acknowledging the difficulty. We know healthy is the hardest cohort. They have read it in trade publications. They have heard it from prior CROs who tried and stalled. They have braced themselves for long timelines and high screen-failure rates. They are wrong about the difficulty, and the wrongness is informative — it is a signal that the prior CRO had the wrong design, not that the cohort was inherently impossible.
Healthy is not the hardest cohort we recruit. It is the most miscategorized. Across more than 300 IVD studies since 2011, including a recent rescue engagement that closed at 1,240 healthy subjects across fifteen sub-cohorts, the studies that struggled with healthy collection were the studies whose protocols had been written for sick people and then asked to recruit healthy ones. The studies that ran cleanly were the studies whose authors had reframed the work around what healthy collection actually is.
What healthy collection actually is, is a reference-range study with a positively defined population, recruited around the subject's convenience because the subject has no clinical reason to come in. Reframed this way, the work is one of the cleanest study types we produce. The framework below is the reframe.
Define the population positively, not as absence of disease.
"Healthy" is not a positive criterion. Demographic boundaries, lifestyle attributes, and claims-data adjacencies are.
The first principle. Most healthy-volunteer protocols define the cohort negatively: no diagnosed cardiovascular disease, no diabetes, no thyroid disease, no current medications, no recent surgery. The negative definition is reasonable on its face — the protocol is trying to specify the absence of confounding conditions. The negative definition is also operationally unworkable, because it produces an eligibility criterion the site cannot screen for in advance. A site cannot identify "patients without thyroid disease" from their EHR efficiently; they can only identify patients with thyroid disease and exclude them, one at a time, after the patient has already walked in. The screen-failure rate ends up high not because healthy is rare but because the screening method is wrong.
The corrective is positive definition. Adult women between 40 and 60, on no chronic medications, with a routine annual exam scheduled in the next 60 days, willing to consent to one additional 8.5 mL blood draw at that visit. The positive definition is recruitable. The site can identify candidates from their existing visit calendar. The patient already has a reason to be there. The negative criteria — no thyroid disease, no recent surgery — become exclusion checks at consent, not pre-screening filters. The funnel inverts, and what looked like a hard recruitment problem becomes a straightforward overlay on a visit that was already happening.
On the recent multi-cohort rescue, sub-cohorts defined positively (women in specific cycle phases, pediatric subjects at well-child visits, age-banded postmenopausal women) recruited at roughly twice the rate per active site of cohorts defined primarily by exclusion. The differentiator was not the eligibility envelope's tightness; it was the screening direction. Positive definitions screen forward. Negative definitions screen backward, and the backward screen is where the time goes.
For every healthy-cohort criterion, ask whether the site can identify candidates against it from their existing visit calendar. If the answer is no, the criterion is exclusion, not inclusion, and it should not appear in the eligibility-screening pass. Inclusion criteria specify who comes in. Exclusion criteria specify who does not consent.
It is a reference range study with a defined population.
The reference range is the deliverable. Treating the study as range-generation, not absence-of-disease confirmation, changes how it is designed and how it is judged at submission.
The second principle is about purpose. Most healthy-volunteer protocols are written as if the study's job is to confirm that the subjects are healthy and then collect samples from them. This is the wrong job description. The study's job is to characterize the analyte's distribution in a defined population that the IFU's intended-use claim will reference. Confirmation of health is a screening step in service of the deliverable; the deliverable is the reference range.
This reframe changes what the protocol prioritizes. If the deliverable is a reference range, the study's design questions become statistical: what is the population whose distribution we are trying to estimate, what sample size produces an interval narrow enough for the claim, what stratifications matter for clinical interpretation. The eligibility criteria become the population definition the range will reference. The cohort design becomes a sub-cohort plan that the FDA reviewer can map to the IFU's intended-use language. None of this is what protocols written under the absence-of-disease framing produce.
The 1,240-subject rescue we recently completed had been originally scoped as a 240-subject "healthy collection" by a prior CRO. The scope grew to 1,240 across 15 sub-cohorts not because the original number was insufficient on average, but because the IFU's intended-use claims required reference ranges in specific sub-populations — pediatric Tanner stages, pregnancy trimesters, age-banded postmenopausal women — that the original protocol had collapsed into a single "healthy adults" bucket. Reframing the study as reference-range generation across those sub-populations was the unlock.
Before drafting any healthy-volunteer protocol, write the sentence the IFU will eventually contain about reference range. The sentence specifies a population. The protocol's job is to characterize the analyte's distribution in that population. Every design decision should be evaluated against whether it improves that characterization.
Recruit people who don't routinely come in.
Healthy people do not have a clinical reason to visit. The recruitment plan has to acknowledge this, and the visit has to be designed for their convenience, not the practice's.
The third principle is the one that distinguishes healthy collection from disease-based collection most fundamentally. A patient with a chronic condition has a clinical reason to come in; the study can be embedded in that visit. A healthy subject does not. The recruitment plan has to either find a clinical visit the subject was already going to have — annual physical, routine screening, well-child visit, prenatal visit — or accept that a study-only visit is required and design that visit around the subject's convenience.
The second path, when it is necessary, is where most healthy protocols fail. They schedule the study visit at the practice's preferred time, in the practice's preferred location, with the practice's preferred draw window. Healthy subjects, who have no clinical reason to be there, decline at high rates and the rates worsen as the study progresses. The corrective is to design the visit around the subject. Saturday morning availability. Locations near the subject's home or workplace. Draw windows that accommodate work schedules. Reasonable reimbursement. The visit is being created for them; the design has to acknowledge that.
On healthy-volunteer cohorts where a clinical visit was available — adult women's annual exams, pediatric well-child visits — recruitment ran on cadence and screen-failure rates were low. On cohorts where the visit had to be created from scratch — adult male healthy controls outside an annual-exam window — the recruitment design had to be inverted: subject-convenience scheduling, expanded draw windows, paid time, and direct-to-patient outreach through channels the subjects already used. The cohorts ran successfully under the inverted design and would have stalled indefinitely under the practice-centric one.
For every healthy cohort, name the visit type the protocol overlays on. If a clinical visit exists in the subject's normal pattern, use it. If not, design the study-only visit around the subject's convenience — not the practice's calendar, not the CRO's logistics, not the sponsor's preferred timeline.
Build the cohort as a gold standard the claim can rest on.
The reference range produced is not a side-deliverable. For most IVD claims that touch a healthy comparator, it is the foundational evidence the entire validation rests on.
The fourth principle integrates the prior three. If the population is positively defined, the study is reference-range generation, and the recruitment is designed around subject convenience, the result is a healthy cohort whose distribution characterization is the reference for every downstream comparison the assay will need to defend. The pivotal trial that comes later compares its disease cohort to this reference. The 510(k) submission cites this reference. The IFU's intended-use claim invokes this reference. The healthy cohort, done right, becomes the gold standard the IVD's regulatory profile rests on.
This implies a level of investment in the healthy cohort that conventional protocol writing under-prioritizes. Sample sizes large enough to support sub-population stratification. Replication where the science requires. Storage at the conditions the downstream assays will actually need. Documentation of collection conditions adequate for any future re-validation. The cohort is being built once. It has to be built well, because everything that follows references it.
Done at this level of care, the healthy collection becomes one of the most leveraged investments in the IVD's submission package. It is also one of the most defensible at FDA review. The agency understands reference-range characterization; what they push back on is healthy collection that was treated as a checkbox rather than a deliverable. The framework above produces the deliverable.
Treat the healthy cohort's reference range as the IVD's foundational evidence, not as a screening artifact. Size the study against the precision the downstream claims require. Design the sub-cohorts to map to the IFU's eventual intended-use language. The investment compounds across the submission lifecycle.
What this framework rules out.
The four principles describe how healthy collection becomes a deliverable rather than a difficulty. They also rule out a few conventions worth naming.
They rule out "healthy is hard" as the explanation for why prior healthy collections stalled. The cohort is not hard. The protocol design is wrong. The reframing is the unlock.
They rule out "healthy" as a primary eligibility criterion. Absence-of-disease does not produce a recruitable population or a defensible reference range. Positive definition does both.
They rule out protocol-overlay recruitment as the only strategy for healthy cohorts. The visit does not exist without the study, so when no clinical visit is available, the study has to create the visit — and the visit has to be designed for the subject's convenience, not the practice's.
They rule out treating the reference range as a side-deliverable. For most IVD claims, the healthy cohort's reference range is the foundational evidence. Investing in it accordingly is the difference between a clean submission and a difficult one.
The framework is not closed. When the study outcome matters, you call RDI. Healthy is not the hardest cohort. It is the most miscategorized. Reframe it, and it becomes one of the cleanest deliverables a CRO can produce.