Clinical Research Methodology

How wearable data, questionnaires, and laboratory measurements contribute to clinical research.

Reputable Health helps wellness brands and ingredient manufacturers plan clinical research around a defined question. Depending on the protocol, research can combine wearable data, participant questionnaires, and laboratory measurements. Each contributes a different perspective; none establishes a product’s effects on its own.

Clinical research methodology is the plan for designing a study, collecting measurements, analyzing results, and explaining their limitations. This page explains the methodological considerations behind that approach. The study-specific protocol, analysis plan, and final report are the sources for how an individual study is conducted and what its findings support.

Explore Reputable’s clinical research services →

Start with the question, then choose the measurements

A study starts with a question about a particular product, population, and outcome. An endpoint is the specific outcome used to evaluate that question. The protocol defines the study design, eligibility criteria, measurement schedule, and planned comparison. The analysis plan describes how those measurements will be evaluated.

Not every study needs every type of data. The choice depends on the question, the suitability of the measurement, participant burden, and the agreed study scope.

Three complementary ways to measure outcomes

Wearables: measurements from daily life

Wearables can provide repeated measurements or estimates of features such as activity, heart rate, or sleep. What is available depends on the device and the data it makes accessible. A device-derived estimate is not automatically equivalent to a clinical assessment, and a proprietary wellness score should not be treated as a diagnosis.

Device selection needs to account for the intended outcome, relevant validation evidence, the participant population, and practical use. Device and software versions, recording schedules, and usable-data criteria matter when interpreting results.

For a practical overview, read our guide to wearables in decentralized clinical trials. See also the FDA guidance on remote digital measurements.

Questionnaires: the participant’s experience

Questionnaires and diaries capture experiences that a sensor may not measure directly, such as perceived sleep quality, symptoms, or daily functioning. The instrument, recall period, completion schedule, and scoring method need to fit the research question. Where appropriate, a validated instrument can support consistent measurement; its suitability still depends on the population and intended use.

Self-reported outcomes can be affected by recall, expectations, and incomplete responses. They add context rather than serving as an automatic confirmation of wearable findings.

Laboratory measurements: protocol-specific biological information

Laboratory work may be included when a research question calls for a particular biological measurement. It is not part of every study. Sample timing, preparation requirements, collection and handling, assay methods, and units all affect interpretation.

A change in a laboratory marker does not necessarily demonstrate a meaningful health benefit. The result must be interpreted in relation to the study question, design, and other evidence.

How the measurements fit together

The protocol and analysis plan define how records from the different sources are linked to the same participant and study period, which outcomes are primary, and which are secondary or exploratory. Keeping the measurement windows clear helps avoid comparing a questionnaire about one period with sensor or laboratory data from another.

For example, a sleep study might pair device-derived sleep estimates with a sleep questionnaire. Laboratory measurements would only be added if they address a relevant question. This is a planning example, not a description of a completed Reputable study or a promised outcome.

The sources may agree, disagree, or show no clear change. Those differences belong in the interpretation. They should not be combined into a single success score unless that score and its use are justified in the analysis plan.

For a worked planning illustration, see the sleep study-planning example in our wearables guide.

Missing data: a gap is not a zero

Missing data can arise from device non-wear, charging or syncing issues, unanswered questionnaires, missed sample collection, or withdrawal. An absent measurement must be distinguished from an observed value of zero.

The approach to missing data is study-specific. The protocol and analysis plan should address:

What counts as usable data: the recording, questionnaire, or sample criteria for each outcome.
How gaps are documented: the amount and timing of missing information, and known reasons where available.
Who contributes to each analysis: the number enrolled, the number with usable measurements, and reasons for exclusions.
How gaps are analyzed: whether an analysis uses available observations, estimates missing values, or uses another justified method, with its assumptions stated.
How conclusions are checked: whether appropriate sensitivity analyses show that different reasonable assumptions would change the findings.

There is no universal wear-time or completion threshold for every study. If participants with missing data differ from those with complete data, an analysis of completers alone may give a misleading picture. A final report should make that uncertainty visible.

For technical background on analysis assumptions and sensitivity analysis, see ICH E9(R1), published by the FDA. This is a methodological reference, not a claim that every Reputable study is a regulatory submission.

Participation metrics need a definition

Retention, questionnaire completion, usable wearable data, and adherence to product use measure different things. Any reported participation rate needs to identify the activity measured, numerator, denominator, time period, and applicable exclusions. These measures should be reported separately unless a combined measure has a documented definition.

What a study can—and cannot—show

A change from baseline does not, by itself, show that a product caused the change. A comparison group, randomization, and blinding can help address different sources of bias when appropriate to the research question and design. They should only be described as features of a study when that study actually used them.

Interpretation also depends on sample size, study duration, participant selection, measurement limitations, missing data, and other factors that could explain the findings. Results from one product, dose, population, or setting may not apply to another.

A statistically significant finding is not automatically a meaningful benefit. Effect size and uncertainty matter. Exploratory findings should be identified as exploratory, and inconclusive findings should be reported alongside positive results. A small study that does not detect a difference does not establish that no effect exists.

For help interpreting a study report, read How to Read Clinical Trial Data.

Evidence behind a public claim

A study result is most useful when a reader can see what supports it. For a numerical claim, the supporting report should identify the product and population studied, design, measurement, comparison, analysis period, number analyzed, result, and limitations. Funding and relevant conflicts of interest also provide important context.

Testimonials, pilot findings, observational studies, and randomized trials provide different kinds of information. Ethics approval concerns oversight of the research; it does not establish that a product works. Publication and peer review should only be claimed when they have occurred.

Common questions about clinical research methodology

Does every study combine wearables, questionnaires, and lab tests?

No. The research question and protocol determine which measurements are needed. Laboratory work is included only when relevant and within the agreed scope.

How is missing study data handled?

The study-specific analysis plan defines usable-data criteria, how gaps are analyzed, and the assumptions behind that approach. Missing observations are not automatically counted as zero, and their effect on conclusions should be reported.

Does a change during a study prove that a product works?

No. Interpretation depends on the comparison, study design, measurement quality, and uncertainty. A change from baseline alone does not establish causation.

Plan research around your evidence needs

Bring your product, intended participants, and research question to a discussion with Reputable Health. The agreed study scope determines which measurements, services, and deliverables are included. Laboratory work and publication support are not automatically included in every engagement.

Explore clinical research services or discuss your study with our team.

Clinical Research Methodology and Evidence