I had a conversation last spring with a dermatology resident who told me, with conviction, that she had a “Fitzpatrick III skin type” and that this informed her choice of moisturiser. I asked her whether she meant her tendency to tan, her tendency to burn, her melanin content, her tendency to develop post-inflammatory hyperpigmentation, or her oiliness. She paused for a long time and then said, “All of them?”
I was not trying to embarrass her. I was trying to point at something I have seen in dermatology offices, on Reddit, in skincare YouTube, and in product marketing for at least a decade. The Fitzpatrick scale, which was designed by Thomas Fitzpatrick in 1975 to predict response to UV-A phototherapy dosing in patients with psoriasis and vitiligo, has been pressed into service as a general skin-type taxonomy that it was never built for. The slippage is consequential. It affects product recommendations, post-inflammatory hyperpigmentation risk assessment, laser settings in cosmetic dermatology, and how clinicians and consumers talk about skin colour.
I want to walk through what Fitzpatrick actually wrote in 1975 and 1988, what the scale was for, why it has been widely misused, and what better alternatives exist for the questions people are usually trying to answer.
What the original paper actually said
Thomas Fitzpatrick first introduced the scale informally in 1975 and published the more cited version in 1988 in Archives of Dermatology. The 1988 paper is the canonical reference. It is a short paper and it is specific about what the scale was for.
Fitzpatrick designed the scale to answer a single question: how much UV-A do you give a patient in a phototherapy session before they burn? Phototherapy at the time was being used for psoriasis and vitiligo, the dose was based on minimal erythema dose (MED), and there was clinical demand for a way to predict MED before exposing the patient. Fitzpatrick noticed that asking patients two questions (do you burn easily, and do you tan?) predicted MED reasonably well. He categorised the answers into six types:
- Type I: always burns, never tans
- Type II: usually burns, tans minimally
- Type III: sometimes burns, tans moderately
- Type IV: rarely burns, tans well
- Type V: very rarely burns, tans darkly
- Type VI: never burns, deeply pigmented
The original paper is explicit. The scale is a response-to-UV scale, based on self-reported sun-burning and tanning behaviour. It is not a skin-colour scale. It is not a skin-type scale in the sense of “dry/oily/combination”. It is not a melanin-content scale. It does not predict post-inflammatory hyperpigmentation risk, sebum production, barrier function, or anything else other than the response to UV exposure.
Fitzpatrick himself, in the 1988 paper, was clear about the limits. The scale was developed in a predominantly white American patient population. Types V and VI were added later, partly because of acknowledged inadequacy in describing patients with darker skin. The scale assumes a relationship between burning behaviour and skin colour that is true on average and frequently wrong in individual cases. East Asian and South Asian patients in particular are poorly described by the scale, because the relationship between visible skin tone and burning behaviour does not follow the European pattern the scale was built around.
How the misuse compounds
The scale entered general dermatology not because anyone made a deliberate decision to extend it, but because there was no other widely adopted system and clinicians needed shorthand for talking about skin colour and pigmentation behaviour. The first round of misuse was using Fitzpatrick to describe skin colour directly, which it does only loosely. The second round was using it to predict post-inflammatory hyperpigmentation risk. The third was using it to set parameters for cosmetic laser treatments. The fourth was using it in cosmetic product marketing to describe target audiences.
Each of these uses is a step further from what the scale was designed to do.
The cosmetic-laser application is the most clinically consequential. Laser settings (specifically for hair removal, pigmentation treatment, and resurfacing) depend on a real prediction of melanin content in the epidermis, because the laser energy interacts with melanin and the risk of post-treatment hyperpigmentation and burns depends on it. Using Fitzpatrick as a proxy for melanin content works reasonably well at the extremes (Type I and Type VI) and progressively worse in the middle, where the relationship between self-reported burning behaviour and actual epidermal melanin content is loose. Patients who self-identify as Type III or Type IV can have widely varying actual melanin content, and laser settings calibrated to a Fitzpatrick estimate alone are a recurring source of adverse events in cosmetic dermatology.
Roberts, in a 2009 critique published as the Roberts Skin Type Classification System, made this point at length. Susan C. Taylor and others have made similar arguments for skin of colour patient populations. The Lancer Ethnicity Scale exists. The Goldman classification exists. The Glogau scale (which addresses photoaging rather than skin colour) exists for the actual photoaging question. None of these has dislodged Fitzpatrick from clinical and consumer use, partly because Fitzpatrick is simpler and partly because change is slow.
What to use instead, depending on the question
This is the part I think most articles on this topic skip. The right answer is not “use a different single scale”. The right answer is to figure out which question you are trying to answer and use the system that was built for that question.
If you want to predict your response to sun exposure and your sunburn risk: Fitzpatrick is actually fine for this. It was built for this. The variable to capture is “how easily do I burn”, and Fitzpatrick captures it.
If you want to predict your risk of post-inflammatory hyperpigmentation from inflammation (acne, eczema, ingrown hairs, picking): Fitzpatrick is a crude proxy. The Roberts Hyperpigmentation Scale (R0-R4) is more direct. Patients in the R3-R4 range (people whose skin is genetically predisposed to deposit melanin in response to inflammation) have a much higher risk of PIH and need to be much more cautious with skincare actives that cause inflammation. Roberts also separates this from the photoaging scale and the hyperpigmentation scale into different axes.
If you want to characterise oiliness and sensitivity: the Baumann skin type system (16 types across four axes: dry/oily, sensitive/resistant, pigmented/non-pigmented, wrinkled/tight) is more useful than any single scale. Leslie Baumann published this in 2006. The questionnaire is publicly available. I have my caveats about it (the four axes are not independent, the validation is brand-affiliated) but it is a better starting point than “Fitzpatrick III, combination skin”.
If you want to predict melanin content for laser settings: spectrophotometric measurement (Cortex DermaSpectrometer, Mexameter, or similar in-clinic devices) is what good cosmetic dermatologists use. The numerical melanin index (M-index) is what you actually want. Fitzpatrick is a proxy that is acceptable when you cannot measure directly.
If you want to talk about skin colour: use skin colour. Munsell color chips, the New Zealand skin type system, the Crayola crayon analogy if you must. Skin colour is observable. You do not need a scale.
The contrarian view I hold
I think we should mostly stop using Fitzpatrick in non-phototherapy contexts and accept that there is no single scale that does the job. The skincare and cosmetic dermatology fields are going to keep using Fitzpatrick because it is a shared vocabulary, and I understand the cost of breaking the shared vocabulary. But I would like to see more product recommendations, more laser settings, more PIH risk assessments, and more dermatology training pivot toward task-specific scales.
What I find most frustrating is the slippage in consumer skincare. A product labelled “for Fitzpatrick IV-VI” is, at best, telling you the brand has thought about people with darker skin tones. At worst it is using Fitzpatrick to make a claim about post-inflammatory hyperpigmentation behaviour or melanin response that the scale was never built to support.
The dermatology resident I spoke to last spring is a smart clinician and was using the same shorthand her training had given her. The shorthand is widely used. It is also wrong.
What I would tell my past self
If I could go back to the version of me who started writing about skincare and assumed Fitzpatrick was the canonical skin-type scale, I would say three things.
Fitzpatrick is a scale for predicting sunburn. That is what the original paper says. That is what it is good for. That is its scope.
Almost every other question you might want a “skin type” scale to answer has a better scale somewhere. The Roberts scale for hyperpigmentation. The Baumann system for oiliness and sensitivity. Direct melanin measurement for laser. Direct observation for skin colour.
When someone tells you their Fitzpatrick type, they are usually answering a question about skin colour or PIH risk by referring to a sun-burn scale. The conversation is more useful if you ask them what they actually want to know.
I do not expect Fitzpatrick to disappear. I do think we can use it more carefully.
FAQ
Is Fitzpatrick racist?
That is the wrong frame, I think. The scale was developed in a predominantly white patient population and is poorly calibrated for darker skin, especially South Asian and East Asian patients. Whether that is racism or methodological narrowness depends on what you mean by the word. The clinical consequence is the same. The scale undersells the diversity it was extended to cover and should be used carefully in skin-of-colour patients.
My dermatologist says I am Fitzpatrick II. Should I push back?
Not necessarily. If the context is sun exposure or photoprotection counselling, Fitzpatrick II is a reasonable shorthand. If the context is laser settings or hyperpigmentation risk, ask for a more specific assessment.
What is my Fitzpatrick if I am South Asian?
The honest answer is that the scale fits poorly. Many South Asian patients self-report as IV or V but have burning behaviour, tanning behaviour, and PIH risk that the scale does not predict well. The Roberts scale and direct measurement are better starting points.
Does any single scale work for all skin types?
No, and I think that is the point. The reason we have multiple scales is that “skin type” is shorthand for a bundle of different variables (UV response, melanin content, sebum production, sensitivity, hyperpigmentation behaviour, photoaging) that do not vary together cleanly. A single scale that captured all of these would be either wrong or extremely long.
Should I trust products marketed by Fitzpatrick type?
I would treat the marketing carefully. A product marketed for “Fitzpatrick IV-VI” is, at best, communicating that the brand has tested on people with deeper skin tones and considered the formulation for darker skin. At worst it is using Fitzpatrick as a stand-in for hyperpigmentation behaviour the formula has not actually been validated against.
References
Fitzpatrick TB. (1988). The validity and practicality of sun-reactive skin types I through VI. Arch Dermatol.
Roberts WE. (2009). The Roberts Skin Type Classification System. J Drugs Dermatol / Roberts skin classification literature, 2009.
Taylor SC, et al. Skin of color classification literature, 2000s.
Baumann L. (2006). The Skin Type Solution / Baumann skin type system literature.