EWG Skin Deep ratings: why dermatologists ignore them in 2026

The EWG Skin Deep database rates cosmetic ingredients on a one to ten hazard scale, and its ratings have become a reflex reference for consumers picking a moisturiser. Yet very few board-certified dermatologists cite it in 2026, and most cosmetic regulators do not treat it as a scientific source. The gap comes down to method: EWG scores hazard, not risk, and skincare safety is a risk question, not a hazard label.

TL;DR: The EWG Skin Deep score is a hazard-only indicator, not a risk assessment. It ignores concentration, exposure route, product residence time and the difference between systemic and topical exposure. Regulatory bodies such as the European Commission, ANSM, HAS, EMA and the FDA use full risk modelling with concentration limits and post-market surveillance, which is why an ingredient can be red in the app and fully approved in every serious regulatory framework. Dermatologists follow the risk-based frameworks and clinical evidence, not a one to ten hazard label.

What Is EWG Skin Deep and Who Runs It

The Environmental Working Group is a US non-profit advocacy organisation. It launched the Skin Deep cosmetics database in 2004 to give consumers a searchable ingredient tool. Each ingredient is rated from 1 (low hazard) to 10 (high hazard), with a data availability flag. Products are then aggregated to an overall score.

EWG is not a regulator, not a peer-reviewed body and not a research institution. It is an advocacy group that publishes ratings as part of a broader environmental health campaign. That does not automatically disqualify its content, but it does mean the database should be read as advocacy, not as a scientific reference. The distinction matters because the app interface makes the score look like an authoritative safety verdict, which it is not.

The tool has real strengths. It aggregates ingredient information in one place, it flags allergens for people with contact dermatitis, and it can start a useful conversation about transparency. The problems begin when the numerical score is used as a stand-alone verdict on whether a product is safe to use.

How EWG Assigns Hazard Scores

The Skin Deep methodology, published on the EWG website, aggregates data from open databases such as PubChem, the NIH's HazMap, the EU Cosmetic Ingredient Database (CosIng) and various toxicology reports. A score is assigned based on the presence of hazard flags across those sources: potential carcinogenicity, reproductive toxicity, endocrine activity, contact allergy signals, respiratory or environmental data.

The critical point is that most of those flags are hazard identifiers, not risk conclusions. A hazard is an intrinsic property of a substance. A risk is the probability of harm under a given exposure. Toxicology has known this distinction since Paracelsus wrote that the dose makes the poison in the sixteenth century, and every modern regulatory framework, including the European Cosmetic Regulation 1223/2009 that the ANSM enforces in France, is built on it.

EWG does adjust scores by data availability, so an ingredient with sparse data can still be flagged high on precaution. This is called the precautionary principle applied to hazard, and it structurally biases the app toward higher scores when data are scarce, which is common for newer or niche cosmetic actives.

The Structural Problems with the Methodology

Peer-reviewed critiques of the EWG methodology have been published for more than a decade. A 2011 analysis in the Journal of the American Academy of Dermatology noted that EWG scores frequently disagreed with regulatory risk assessments and that the app's aggregation formula amplified low-quality data. More recent commentaries from the American Academy of Dermatology and the European Academy of Dermatology and Venereology have reached similar conclusions.

The most cited structural issues are:

• No exposure modelling. A preservative used at 0.1 percent in a rinse-off cleanser is treated similarly to the same molecule at 1 percent in a leave-on serum, even though real-world exposure differs by orders of magnitude.
• Hazard is not risk. Water has an LD50. Vitamin A is teratogenic at systemic doses that are simply not achievable with a topical cosmetic. Scoring on intrinsic hazard flags without factoring dose makes the scale non-actionable.
• Extrapolation from irrelevant studies. Some scores rely on inhalation or ingestion studies in animals at doses far above cosmetic use conditions. Dermal exposure at cosmetic percentages is a different toxicological scenario.
• Aggregation opacity. The final product score is a weighted average that is not fully documented, so it is difficult to reproduce independently.
• Data recency lag. Ingredients whose risk profile was updated by SCCS or CIR opinions can remain scored on older data for months or years.

For an ingredient-level reading of how percentages actually behave in a formula, our guide to how to read an INCI list walks through concentration order and use-condition logic.

Dose, Concentration and Residence Time

Regulatory science evaluates cosmetics on three axes at minimum: intrinsic properties of the ingredient, its concentration in the finished product, and the conditions of use, which include residence time on the skin, area applied, frequency and target population. The Scientific Committee on Consumer Safety publishes opinions that follow this exact model for every ingredient reviewed at EU level.

EWG's public methodology page acknowledges that concentration matters, but the app does not consistently apply concentration data to the final numeric score displayed to the consumer. That is the practical failure. A well-formulated 0.3 percent salicylic acid cleanser and a 20 percent salicylic acid professional peel are not the same product, even though the ingredient is the same.

The same logic applies to preservatives. Phenoxyethanol is capped at 1 percent in the EU, must be avoided in leave-on products for children under three, and is monitored by ANSM post-market. That entire framework exists to keep the actual exposure inside a safe window. A hazard-only score does not capture any of this.

Cases Where EWG Contradicts Dermatology Consensus

Several ingredient categories illustrate the gap between the app and mainstream dermatology practice.

Parabens. Methylparaben and ethylparaben are rated 4 to 7 on Skin Deep depending on version, while the SCCS reconfirmed their safety at current EU concentration limits and the FDA has published multiple statements finding no causal link with breast cancer. Our full review of the paraben evidence base covers the SCCS 2020 and 2023 opinions in detail.

Sunscreens. Some organic UV filters (octinoxate, octocrylene, homosalate) carry high EWG scores, mostly on endocrine signals from studies at doses irrelevant to topical use. Meanwhile the EADV and every national dermatology society continue to recommend daily broad-spectrum photoprotection because UV-induced photoaging and non-melanoma skin cancer risk are far better characterised than any theoretical filter concern. See our EU versus US sunscreen filter comparison for the regulatory picture.

Retinol. Rated 9 on Skin Deep for decades on the basis of pregnancy exposure warnings, yet universally recognised by dermatology as one of the most evidence-supported topical actives for photoaging and acne. Pregnancy caution is real and legitimate, but a single blanket score does not communicate that nuance.

Fragrance. Rated 8 as a category. Legitimate concern exists for contact dermatitis in sensitised individuals, but the score does not distinguish IFRA-restricted formulations from unrestricted ones, and it does not separate a hypoallergenic parfum from an allergenic essential oil blend. Our post on fragrance-free versus unscented labelling covers the practical distinction.

What Regulators Actually Do

Cosmetic safety in the EU is governed by Regulation 1223/2009, enforced nationally by bodies such as ANSM in France. Before a cosmetic is placed on the market, the manufacturer must produce a Cosmetic Product Safety Report that includes exposure calculations, margin of safety values and toxicological review. The Responsible Person is legally accountable. Post-market surveillance is handled through cosmetovigilance networks reporting to ANSM.

In the United States, the FDA regulates cosmetics under the FD&C Act, updated by the Modernization of Cosmetics Regulation Act (MoCRA) that came fully into force in 2024. Facility registration, product listing and adverse event reporting are now federal requirements. The FDA works with the Cosmetic Ingredient Review (CIR) expert panel for ingredient assessments.

France also mobilises the Haute Autorité de Santé and Inserm for broader health guidance, and PubMed indexes the peer-reviewed literature that feeds into all of these frameworks. None of these bodies uses a one to ten hazard score. They all use risk-based frameworks with concentration limits, use conditions and post-market data.

Fear Marketing and the Cost to Consumers

Hazard-only scoring is easy to weaponise as marketing. Brands and retailers building around a "clean" positioning frequently cite EWG ratings or a proprietary equivalent to justify banning ingredients that are perfectly safe at cosmetic use conditions. The result is not neutral. It steers consumers away from evidence-based actives (retinoids, some preservatives, well-tolerated organic UV filters) and toward alternatives that are sometimes less effective, less well studied or more allergenic.

For a broader look at how this framing distorts purchasing decisions, our analysis of clean beauty marketing and its impact on natural skincare claims unpacks the incentive structures at work.

The cost is not only economic. Under-protected skin, incomplete acne routines and confused choices in pregnancy skincare have real dermatological consequences, and they are frequently traceable to fear-driven ingredient bans.

How Dermatologists Actually Evaluate an Ingredient

Board-certified dermatologists in 2026 use a very different mental checklist:

1. Is there a plausible mechanism of action for the claim, in peer-reviewed literature indexed by PubMed?
2. Are there controlled clinical studies (ideally randomised) at cosmetic-relevant concentrations?
3. What is the regulatory status in the EU, US and other major markets, and are there SCCS, CIR or FDA opinions on file?
4. What is the safety margin at the concentration used in a typical product?
5. Are there real-world post-market signals (cosmetovigilance, published case reports)?
6. Does the product match the patient's skin type, barrier status and other actives?

None of this maps to a single one to ten score. That is why the reflex "check it on EWG" workflow tends to disappear once someone has been through dermatology training. It solves the wrong problem.

For readers building a routine from scratch, our evidence-based acne routine and the evidence-based rosacea routine illustrate the actual decision process without any hazard score in sight.

When the App Can Still Be Useful

The Skin Deep database is not useless. Read as a starting list of ingredients that may deserve a closer look, particularly for people with a documented contact allergy history, it can flag common sensitisers and prompt useful questions. As a data aggregator, it beats scrolling through fifteen tabs of raw INCI documentation.

The mistake is to treat the numeric score as a verdict. A high EWG rating on an ingredient your dermatologist has recommended is not a reason to stop using it. A low EWG rating on a product is not a guarantee it suits your barrier. The database is best used as one input among several, alongside regulatory status, clinical evidence and personal history.

FAQ

Is EWG Skin Deep a reliable source for cosmetic safety?

It is a hazard aggregator built by an advocacy group, not a regulatory or peer-reviewed source. Its ratings frequently disagree with SCCS, CIR, ANSM and FDA assessments, mostly because it does not model exposure. Use it as one input, not as a verdict.

Why do dermatologists ignore EWG ratings?

Because dermatology practice is based on risk assessment, not hazard flagging. A hazard-only one to ten score cannot capture concentration, residence time, exposure route or patient context. Clinical evidence and regulatory opinions carry far more weight in a dermatology consultation.

Are ingredients rated 7 to 10 on Skin Deep dangerous?

Not automatically. Retinol, some preservatives and some UV filters carry high EWG scores but remain approved and recommended at cosmetic concentrations by every major regulator and dermatology society. Context and dose decide risk.

Does the FDA or ANSM use EWG scores?

No. The FDA regulates cosmetics under the FD&C Act as updated by MoCRA. ANSM enforces EU Regulation 1223/2009 in France. Both use full risk assessments with concentration limits, use conditions and post-market surveillance. Neither cites EWG Skin Deep in its official work.

What should I use instead of EWG to check an ingredient?

Regulatory databases such as the EU CosIng portal, SCCS opinions on the European Commission website, CIR reports, FDA guidance, ANSM cosmetovigilance data and PubMed for the primary literature. For everyday decisions, a consultation with a dermatologist beats any database.

Is EWG useful for allergy screening?

Yes, with limits. It can flag common contact allergens, but a proper patch test with a dermatologist remains the gold standard for identifying your personal sensitisers. The app is a screening prompt, not a diagnosis.

Does a low EWG score mean a product will work for my skin?

No. Efficacy is a separate question from safety. A product can be scored 1 on Skin Deep and still be the wrong choice for your skin type, barrier state or active goals. Match the formula to your skin, not to a hazard label.

Sources and Further Reading

• European Commission, Scientific Committee on Consumer Safety opinions.
• American Academy of Dermatology, patient guidance and position statements on ingredient safety, aad.org.
• European Academy of Dermatology and Venereology, congress abstracts and position papers, eadv.org.
• ANSM, cosmetovigilance dossiers, ansm.sante.fr.
• Inserm and Haute Autorite de Sante, public health syntheses on cosmetic exposure.
• US FDA, cosmetics regulation under FD&C Act and MoCRA.
• PubMed, primary toxicology and dermatology literature.