Healthy, sub-healthy, diseased: why nutrition and CHC research needs a wider recruiting map
Nutrition and consumer-health studies routinely span three health states, healthy, sub-healthy and diseased, and each needs a different channel. Reaching only one narrows the science.
Insight · · 4 min
In prescription drug development, the participant question is largely settled: you recruit patients who meet a diagnostic threshold, through investigator sites that see that diagnosis. Nutrition, functional food, FSMP and consumer health do not have that luxury. The same product, a protein blend, a botanical, a probiotic, a self-care OTC, may need to demonstrate benefit in healthy consumers, in sub-healthy populations (subclinical, at-risk, borderline, self-perceived) and, for FSMP or adjuvant claims, in diagnosed patients. Three health states. One study programme. And three very different ways to reach the participant.
The drug-trial default fails here because the sub-healthy segment is the largest and least visible. 'Sub-healthy' is not a soft category, it is a defined space between wellness and pathology that regulators, EFSA and payers increasingly recognise. It includes borderline dyslipidaemia, low-grade fatigue, functional digestive discomfort, perimenopausal symptoms, sarcopenic-obese phenotypes, IBS-lite, mild sleep or mood complaints, prediabetic states, glycaemic variability in otherwise healthy adults. These people are not patients, they do not sit in a specialist's clinic. They are also not fully healthy, a claim of 'wellness' is not what the product is really doing. They are precisely the population most functional foods and CHC products are designed for, and precisely the population traditional GCP recruitment cannot see.
Endpoints shift with the state. In healthy participants, the meaningful signals are usually habit, tolerability, preference, hydration or micronutrient status, sleep quality, perceived energy, endpoints that only stand up if captured in the real conditions of use. In sub-healthy populations, endpoints move toward biomarkers with a defensible physiological link (lipid subfractions, HbA1c drift, hs-CRP, sIgA, faecal calprotectin, questionnaire-validated scales), because the effect size is small and the signal must be defended against noise. In diseased participants, FSMP, medical foods, adjuvant CHC, the endpoint bar approaches clinical: nutritional status, functional recovery, disease-specific quality of life, sometimes hard clinical outcomes. A protocol that pretends the three states are one thing produces evidence that is neither regulatory-grade nor commercially useful.
The recruiting map has to match. Six channels, chosen per protocol per state:
1. GCP hospital sites, for diseased participants under specialist care, FSMP under supervision, safety in vulnerable populations. Necessary but slow, and blind to the other two states.
2. Community & online pharmacies, the highest-yield channel for consumer-facing products, and often the best channel for sub-healthy adults: they buy there, a pharmacist can screen and consent, and second-stage checks protect data quality. This is where Evidilya activates a large share of nutrition and CHC studies.
3. Territorial HCPs, general practitioners, dietitians, nutritionists, community paediatricians. Ideal for sub-healthy identification (borderline lab values, at-risk lifestyles) and for early-life or maternal nutrition, where the trust anchor is not a hospital but a familiar clinician.
4. Patient associations & patient-like communities, for named conditions and for organised sub-healthy communities (menopause, IBS, migraine, sleep). They enable ethical, informed recruitment at speed and with retention that anonymous panels cannot match.
5. Digital communities & PHYGITAL-CROWD, the fastest way to reach healthy and self-perceived-sub-healthy consumers at multi-country scale, screened and routed digitally into the study workflow. This is the channel that closes weeks of enrolment risk, not months.
6. At-home / D-T-P, the connective tissue. Once consented, most participants across the three states can be followed at home: eConsent, eDiary, telehealth check-ins, wearables, ePRO. This is what turns a wider recruiting map into an operable study rather than an operational nightmare.
The regulatory context has already moved. EFSA scientific opinions on health claims are explicit that the target population must match the claim: a claim aimed at people with borderline blood pressure is not substantiated by a study in normotensive adults, and vice versa. FSMP evidence expectations sit in Regulation (EU) 609/2013 and its delegated acts, and require populations that reflect the intended medical purpose. Cosmetovigilance and food-safety surveillance obligations extend across all three states once a product is on the market. The 2024–2026 shift toward decentralised and D-T-P research (FDA final DCT guidance, ICH E6(R3) Annex 2) removes the last operational excuse for not reaching them.
Operationally, we run this on one backbone. PHYGITAL-CROWD activates the channel mix per protocol. ECS® orchestrates consent, provenance, audit and data capture so a participant recruited from a pharmacy, an association or a digital community sits in the same governed study as one recruited from a hospital. AIRA helps design endpoints appropriate to the health state, not one generic instrument stretched across three populations. And our people-first model means clinicians, pharmacists and community leads are part of the study team, not intermediaries handed a script.
The question sponsors should now ask their CRO is not 'can you recruit fast?', it is 'can you recruit the three health states this claim actually needs, through the channels that reach each of them, in a single defensible ecosystem?' If the answer is one channel, the science is being narrowed before the protocol is even locked.
Written by the Evidilya scientific team. For interviews, references or a full publication list, use the contact page.
Talk to our team.
Bring us the evidence question you're wrestling with, we'll tell you the shortest defensible path.