Does Sunscreen Actually Prevent Melanoma? Reading the One Randomized Trial

"Wear sunscreen to prevent skin cancer" is one of the most repeated pieces of health advice you will hear, yet only one randomized controlled trial has ever put daily sunscreen to the test against melanoma, and its answer is a cautious, heavily qualified yes. In the Nambour trial in subtropical Queensland, Australia, adults assigned to apply sunscreen every day recorded roughly half as many melanomas a decade after the study closed as adults left to their own habits. The catch is that this halving sat right on the border of statistical significance, and no trial has ever repeated it. Understanding why that single result carries so much weight, and why it still cannot settle the question, is a small lesson in how to read evidence.

Key points#

• Nambour is the only randomized trial with a melanoma endpoint; everything else is observational.
• Daily sunscreen was linked to a 50 percent lower melanoma rate, but the confidence interval ran from 0.24 to 1.02, so no benefit remains statistically possible.
• The invasive melanoma reduction did reach significance, yet it rested on just 14 total cases.
• Melanoma was a secondary outcome; the trial was built to study keratinocyte cancers, not melanoma.
• The setting (fair-skinned adults, intense year-round sun, an SPF 15 product) limits how far the numbers travel.

The melanoma result, read one number at a time#

The finding that gets quoted comes from the long follow-up, reported by Green and colleagues in the Journal of Clinical Oncology in 2011. Ten years after the trial's active phase ended, the daily-sunscreen group had 11 new melanomas and the discretionary group had 22. That translates to a hazard ratio of 0.50, meaning the observed melanoma rate was half as high with daily use.

The confidence interval is where the honesty lives. It ran from 0.24 to 1.02, and the p-value was 0.051. Because the upper bound edges just past 1.0, the data are formally compatible with no protective effect at all. The point estimate is encouraging, but with only 33 melanomas split between two groups, a handful of cases landing differently would move the whole conclusion. This is what a fragile result looks like: a promising signal that the sample was too small to nail down.

Split by depth, the picture is both sharper and shakier. Invasive melanomas numbered 3 in the sunscreen group and 11 in the discretionary group, a hazard ratio of 0.27 (95% CI 0.08 to 0.97). The upper bound stays under 1.0, so this subgroup reaches significance. But look at the width of that interval: it stretches from a 92 percent reduction at one end to a 3 percent reduction at the other, precisely because only 14 invasive cases exist to build it from. A result that wide invites interest, not confidence. The pre-invasive melanomas showed no clear effect (hazard ratio 0.73, 95% CI 0.29 to 1.81).

What the trial actually did#

To weigh those numbers, it helps to know how they were generated. As Green and colleagues described in The Lancet in 1999, the Nambour Skin Cancer Prevention Trial enrolled 1,621 adults aged 25 to 75, drawn at random from residents of Nambour, a township in subtropical southeast Queensland. It used a two-by-two factorial design: participants were assigned either to apply an SPF 15-plus sunscreen every day to the head, neck, arms, and hands or to use sunscreen at their own discretion, and separately to take 30 mg of beta-carotene or a matching placebo. The active phase ran from 1992 to 1996.

Over that stretch, daily sunscreen cut squamous-cell carcinoma tumors by about 40 percent (rate ratio 0.61, 95% CI 0.46 to 0.81). It did nothing measurable for basal-cell carcinoma (rate ratio 1.03, 95% CI 0.73 to 1.46), and beta-carotene affected neither cancer. Those keratinocyte results were the trial's original purpose and its firmest legacy: the first randomized proof that a topical sunscreen prevents a skin cancer.

Why melanoma trials are almost never run#

The obvious question is why, decades on, no one has simply run a bigger and cleaner study. The answer is a stack of practical walls. Melanoma is relatively uncommon and can take many years to surface, so a trial powered to detect a difference would need thousands of participants followed for a very long time, at enormous cost. Randomizing the behavior is also ethically awkward now: assigning a group to skip sunscreen in a high-sun region runs against decades of public health advice. Nambour cleared these hurdles almost by accident, because it was designed to study keratinocyte cancers and merely tracked melanoma along the way. That is why melanoma was a secondary outcome rather than the target, and why its numbers, though precious, were never meant to carry the full argument.

The alternative evidence base is observational: studies that follow people who choose to use sunscreen and compare them with people who do not. These are useful but vulnerable to confounding, because habitual sunscreen users and non-users tend to differ in skin type, time spent in the sun, and many other traits that also affect melanoma risk. Randomization is the tool that neutralizes those differences, which is exactly why a single small randomized signal still outranks a large pile of observational ones.

How far the finding travels#

Even if you accept the result at face value, it describes one particular world. The participants were largely fair-skinned adults living at a subtropical latitude under intense, year-round ultraviolet light, close to a worst-case setting for skin cancer and therefore the setting where any protective habit has the most room to show a benefit. The product was an SPF 15 formulation, weaker than the SPF 30 to 50 sunscreens now common. And the trial rewarded a disciplined habit of daily application to defined body areas, not the occasional beach-day use many people actually rely on. A review in the British Journal of Dermatology notes that there is no level-one evidence for protection across all melanomas, and that the trial cannot tell us how consistently sunscreen must be used to prevent the outcomes that matter most, such as metastatic disease or death.

None of this is an argument against sunscreen. Daily use has randomized evidence behind it for squamous-cell carcinoma, substantial observational support for melanoma, and a reassuring safety profile. The accurate summary is simply narrower than the slogan: one trial, not designed for the question, found a halving of melanoma that was real in direction and borderline in certainty, and no one has reproduced it. Good evidence and definitive evidence are different things, and you can hold both at once.