A supplement brand may want to understand how customers use a product in everyday life. An ingredient manufacturer may need research that helps prospective customers evaluate an ingredient.
Both need evidence. But the right research approach depends on the question being asked.
Real-world evidence can contribute to an evidence program, but its usefulness depends on the data, study design and analysis. Collecting information outside a clinic does not, by itself, establish that a product works.
This guide explains the terminology, where real-world research can fit, and the decisions to make before collecting data. It focuses on dietary supplements and the U.S. claims context.
What is real-world evidence?
FDA describes real-world data as health and healthcare information routinely collected from sources such as electronic health records, medical claims, registries and certain digital health technologies. Real-world evidence is the clinical evidence generated by analyzing those data to understand a medical product’s use, potential benefits or risks. FDA: Real-World Evidence
For supplement research, these definitions provide a useful starting point. However, FDA’s medical-product RWE programs should not be presented as a general approval pathway for supplement claims.
It helps to separate four concepts:
| Concept | What it describes |
|---|---|
| Real-world data | The source and circumstances of data collection |
| Real-world evidence | Evidence produced through analysis of those data |
| Observational study | Research in which investigators do not assign the intervention being studied |
| Decentralized study | Research that conducts some or all study activities away from a traditional research site |
These concepts can overlap. A randomized trial can use remote assessments, and real-world data can contribute to different study designs.
An at-home study is not automatically an observational study or an RWE study. Describe how participants were assigned, how measurements were collected and what was analyzed.
Start with the decision the evidence needs to inform
Before choosing a device, survey or research partner, write down the decision you want to make.
For example:
- Do people understand how to use the product?
- Can participants complete the proposed measurement schedule?
- Which outcomes deserve further investigation?
- Does the product improve a defined outcome compared with a suitable control?
- How closely does existing ingredient research apply to a proposed formulation?
These questions require different approaches. A study designed to assess participation feasibility should not be presented as a definitive test of efficacy.
NCCIH specifically cautions against using small feasibility pilots to establish efficacy or safety, or using their unstable treatment-effect estimates to size a later efficacy trial. NCCIH: Pilot Studies, Common Uses and Misuses
A practical starting brief is:
We need to understand [specific question] in [defined population], using [specified product or ingredient], over [relevant period], so we can decide [next step].
Where can real-world research fit?
Consider the following as planning options, not a required sequence or a guarantee of sufficient evidence.
| Decision | Research discussion to have | Interpretation boundary |
|---|---|---|
| Understand product use | Examine usage patterns and participant experiences | Reported use does not establish effectiveness |
| Test research procedures | Assess recruitment, measurement completion and participant burden | Feasibility does not establish benefit |
| Explore an association | Examine relationships between exposure and outcomes | Other differences may explain the relationship |
| Evaluate a product effect | Discuss an appropriate controlled design and analysis | Conclusions depend on the comparison and study quality |
| Understand applicability | Compare the studied product and population with the intended use | Findings do not automatically transfer to another formulation or audience |
The goal is to match the method to the decision. A larger dataset cannot compensate for a question the study was not designed to answer.
Choose measurements that answer the question
Depending on the protocol, research may combine wearable measurements, participant questionnaires, laboratory assessments and other tools.
Each should have a defined purpose.
Wearables may provide repeated measurements relevant to activity, sleep or other physiological outcomes. Device choice and measurement suitability need to fit the intended assessment.
Questionnaires can capture participants’ experiences and functioning. Choose an instrument suited to the outcome, population and assessment period.
Laboratory assessments may help address a specific biological question. Their inclusion should follow from the protocol, rather than from a desire to collect more measurements.
These sources are complementary. Agreement is useful to examine, but one measure does not automatically validate another.
FDA’s digital health technology guidance is a reference for evaluating remote measurement tools in medical-product investigations. Its applicability to a particular supplement study should be assessed rather than assumed. FDA digital health technology guidance
For device-specific planning considerations, see our guide to wearables in decentralized clinical trials.
Plan for data quality before collection begins
A useful research brief should ask the team to document:
- Which measurements are required and when.
- How product use and relevant changes in behavior will be recorded.
- What makes a recording usable.
- How missing observations and withdrawals will be reported.
- Which analyses are planned and which will be exploratory.
- Who is responsible for participant safety, consent and data access.
Missing measurements should remain identifiable. A missed questionnaire or absent wearable recording is not a zero result.
Ask the research team to explain its missing-data assumptions and how it will assess whether those assumptions affect the conclusions. Avoid adopting a universal rule for filling gaps across every outcome.
Our clinical research methodology and evidence page explains how measurement choices, missing data and study limitations fit together.
Understand what the findings can support
An observed improvement does not necessarily mean the product caused it.
For example, participants might change their sleep routine during a study. Expectations, other supplements, natural variation or differences between groups could also affect the findings.
For health benefit advertising, FTC guidance generally expects randomized, controlled human clinical testing. The label “real-world” does not lower the substantiation standard. The research must support the specific message consumers receive. FTC Health Products Compliance Guidance
For dietary supplement claims, FDA recommends evaluating the relevance, quality and totality of the evidence. A study on a different dose, formulation or population may not support the intended use. FDA dietary supplement substantiation guidance
For a detailed explanation of claim wording and evidence review, read our structure/function claims guide.
A hypothetical example: planning research on a sleep supplement
Suppose a brand wants to investigate perceived sleep quality in adults who report occasional poor sleep.
A sensible planning discussion would separate three questions:
- Can the procedures work? Can eligible participants follow the schedule and complete the required assessments?
- What outcome matters? Is the main question about perceived sleep quality, sleep duration or another clearly defined outcome?
- What comparison is needed? How will the study distinguish a possible product effect from other explanations?
A questionnaire might assess perceived sleep quality, while wearable measurements provide additional information about selected sleep estimates. The team would need to align the assessment periods and explain how disagreements between measures would be interpreted.
If participants report better sleep, that finding should not automatically become a claim about treating insomnia, improving cognition or accelerating recovery. Each proposed message needs its own evidentiary assessment.
This is a planning illustration, not a recommended protocol or evidence that any product works.
Build an evidence package people can evaluate
A useful output is more than a favorable chart.
Ask for a report that makes the research question, methods, participant population, measurements, analysis and limitations understandable. It should distinguish planned analyses from exploratory findings and report unfavorable or inconclusive results alongside favorable ones.
For ingredient manufacturers, also document exactly what material was studied. Customers need enough information to assess its relevance to their own products.
Our clinical studies for ingredient manufacturers buyer’s guide explains how to compare research partners, proposals, budgets and deliverables.
Frequently asked questions
Can wearable data alone prove a supplement works?
No. A device provides measurements. Whether those measurements support a product benefit depends on their suitability and the study’s design, comparison, execution and analysis.
Is real-world research always faster or cheaper?
No universal comparison is useful. Recruitment, follow-up, measurements, data access and analysis all affect scope. Compare proposals that address the same research question and deliverables.
Can ingredient findings be used for a finished product?
Potentially, but relevance must be assessed. Differences in dose, formulation, other ingredients and intended users can matter. A shared ingredient name is not enough. FDA substantiation guidance
Does a large participant count make a study conclusive?
No. Participant count does not resolve biased recruitment, unsuitable measurements, confounding or an inadequate comparison. Ask how the sample size supports the study’s actual objective.
Plan your next evidence decision
Start with the question, the product specifications and the evidence already available. Identify what remains uncertain before deciding what to collect next.
Explore the Reputable Health platform and our research methodology to prepare for a study-planning discussion.
The aim is an evidence program that makes each decision clearer and communicates the findings within their limits.