Guide 

Make the decision claim by claim

Write the proposed claim before reviewing papers. “Improves a measured facial wrinkle outcome in the studied population” is a different proposition from “lifts the face,” “treats acne” or “works for all skin types.” A study cannot answer a question it did not test.

Create three columns: what the study actually used, what the salon device actually delivers, and what remains unexplained. Add page numbers and document versions. A marketing label belongs in the second column only as a label; measured output needs its own evidence.

1. Identify the intervention. Confirm that it is an LED intervention relevant to the proposed use. Keep ultraviolet (UV) phototherapy, intense pulsed light (IPL), lasers, photosensitizer-based photodynamic therapy (PDT), bright-light boxes and infrared heating separate.
2. Compare the delivered exposure. Check the actual operating mode, spectrum, optical output at the treatment position and coverage, rather than color names alone.
3. Compare the clinical question. Check people, body site, condition or appearance concern, comparator, regimen and endpoint.
4. Assign a bounded conclusion. Identify what the paper supports, which differences need technical or clinical assessment, and which claims remain unsupported.

These are editorial review categories, not a regulatory equivalence test. For United States health advertising, the Federal Trade Commission (FTC) says evidence must be relevant to the specific product and benefit, including the research population and real-world use. A valid study can still be irrelevant to a particular claim. [6]

An unknown is an unresolved comparison, not an acceptable match. Missing output data does not prove that a device fails; it prevents a confident transfer of the study’s conclusion.

Compare the optical intervention, not just the wavelength

Photobiomodulation (PBM) describes biological modulation by light. It does not identify one standardized device or exposure. The FDA’s January 2023 PBM guidance addresses specified class II medical devices and remains a draft, marked “Not for implementation.” It is useful context for evidence review, not a worldwide salon operating standard. [1]

The worksheet below translates reporting considerations into questions for a dossier. Jenkins and Carroll’s reporting paper identifies beam and treatment parameters needed for repeatable research; its abstract was checked here, not the full paper. Hamblin’s later editorial likewise emphasizes reporting the intervention’s source and parameters. Neither document establishes a cosmetic treatment prescription. [2] [9]

Ask how the output was measured

Request the test report tied to the actual model and mode: instrument, calibration basis, measurement conditions, uncertainty and test date. The National Institute of Standards and Technology (NIST) describes spectral irradiance calibration in terms of traceability and uncertainty. Its publication is metrology context, not evidence that a salon device works or a do-it-yourself certification method. [3]

A 2025 methodological analysis of visible LED dermatology research highlights inconsistent reporting and dose-validation gaps. It used a convenience sample and a modeled comparison; it was not a systematic review of all LED efficacy. Its observations support careful scrutiny of source data, not a conclusion that every LED device is ineffective. [5]

If either side reports only a single energy-density value, retain the rest of the worksheet. Equal joules per square centimeter do not, by themselves, resolve different spectra, delivery geometry or regimens. Do not change time, distance or intensity to make the numbers resemble a paper; that would introduce a new intervention requiring its own justification.

Match the people, treatment area and measured outcome

Extract eligibility and exclusions, rather than reading only the sample size. Record age range, sex, reported skin phototypes, baseline concern and severity, treatment site and relevant health restrictions. Leave unreported characteristics blank and label them unknown. A paper on one facial area does not automatically cover the neck, scalp or body.

A real example: positive findings did not apply to every endpoint

Bragato and colleagues’ 2025 randomized, sham-controlled facial study enrolled 95 women aged 45–60 and used a specified red LED mask. In the original abstract, the Wrinkle Assessment Scale showed no significant differences among groups. Separate photographic measurements found reductions in selected wrinkle lengths, and participant satisfaction favored active exposure. [4]

That is a mixed endpoint pattern. It does not justify collapsing the findings into “proven facial rejuvenation” for a different salon device. A clinical scale, a selected photographic measurement and satisfaction answer different questions. This example is based on the original abstract only; the complete methods, exclusions and analysis could not be reviewed here.

• Name the measure. Is it a validated clinical scale, an instrument reading, standardized image analysis or self-reported satisfaction?
• Check the comparison. Did change differ from the control group, or was it only a before-and-after change in treated participants?
• Check the claim’s size. A statistically significant result does not, by itself, establish a noticeable or meaningful benefit for a client.
• Keep outcomes separate. A collagen-related laboratory marker is not a measured face-lifting outcome; an acne endpoint is not a scar-removal endpoint.

These questions are a reading framework, not a reanalysis of the trial. Mechanism papers and animal or cell experiments can inform hypotheses, but cannot supply missing human outcomes for the proposed service.

Read the complete regimen and the follow-up

The intervention includes more than one session. Extract exposure duration, treatment intervals, number of sessions, anatomical positioning, preparation, accompanying treatments, adherence and any maintenance phase. Then record when the outcome was measured, whether treatment was still continuing and how adverse events were collected.

Keep two separate timeline entries: the treatment period and the observation period. Improvement assessed during a course cannot establish how long it lasts after stopping. A brief observation period cannot settle long-term safety. Do not create a maintenance schedule from an untested gap in follow-up.

Inspect the control as carefully as the active arm. Was there a credible sham? Were participants and outcome assessors blinded? Were skincare or other procedures balanced? Were dropouts and missed sessions reported? A protocol describing planned work is not a report of completed results.

Read funding and device-supply disclosures, too. Sponsorship does not automatically invalidate a study, but it is a reason to examine design, reporting and independent replication. Keep adverse-event findings attached to the actual participants, exposures and monitoring period; “none reported” is not a universal safety guarantee.

A supplier’s bibliography needs a documented link to the device

For each cited paper, ask the supplier to identify the claimed benefit, the studied device, the current model and mode, the matching data and every consequential difference. An unexplained reference list leaves the central question unanswered: why should this paper’s result apply to this device and use?

A useful submission includes the complete paper and supplements, current instructions for use (IFU), optical test documentation and a written assessment by appropriately qualified people. If there is no trial on the exact model, the supplier should explain the basis for using another intervention’s evidence and whether further clinical evidence is needed. A specification comparison alone cannot establish equal patient outcomes.

In United States advertising, third-party literature can contribute to implied health claims, and disclaimers do not cure misleading claims. Review the overall message of a training slide or service description, not just whether a reference appears beneath it. [6]

Keep regulatory identity on a separate track

FDA establishment registration and device listing do not denote approval or clearance. Verify any United States authorization claim against the relevant record and the actual device identity. A registration certificate cannot close a clinical evidence gap. [7]

This article does not determine any device’s legal classification, market access or a practitioner’s permitted scope. Those questions require the applicable jurisdiction’s records and rules; a research match cannot answer them.

Frequently asked questions

Can a home-use study ever be relevant to a salon service?

Yes, as evidence to examine. The home setting does not automatically make a paper irrelevant, but the salon label does not establish a match either. Compare the intervention, operator, use conditions, participants and endpoint. Relevance still does not authorize the proposed professional service.

What if the supplier says the current model is an upgrade?

Ask for the documented differences from the studied version. New emitters, changed optics, software modes or positioning can leave the previous comparison unresolved. “Upgrade” is not an evidence category; the assessment should explain what changed and how that affects the particular claim.

Does a systematic review remove the need to inspect individual trials?

No. It helps organize the broader literature, but you still need to know whether the included interventions and outcomes match the proposed use. The 2025 methodological analysis illustrates why incomplete and heterogeneous LED reporting makes comparisons difficult. [5]

What should the file say while full methods are unavailable?

Record “original abstract verified; full methods not reviewed,” followed by the exact missing fields. Keep the paper as a lead or a bounded example. Do not fill gaps from a supplier slide, guess missing values or describe an incomplete review as a completed equivalence assessment.

Keep the conclusion as narrow as the evidence

A useful review ends with a traceable statement: the paper supports this measured outcome under these studied conditions; these device differences remain unresolved; these additional documents or assessments are needed. That record gives a salon a defensible way to evaluate evidence without borrowing another intervention’s results.