How to interpret skin care research data

• Before reading research abstracts or articles on a particular topic, read an overview of the issue written in plain English. Many skin rejuvenation methods are discussed in Anti-Aging Treatments section of this site. For example, before reading research abstracts on laser resurfacing, read our article about it.
• There are two major types of research articles: a study report and a scientific review. A study report is essentially an account of a research study performed by the authors. A review, on the other hand, is a summary and analysis of key studies in a particular (usually narrow) field of research. In a review, the authors mainly discuss other people's studies. If written by knowledgeable and impartial researchers, reviews are useful because they allow the reader to quickly assimilate the results of multiple studies. Still, reviews are a third party interpretation of the research data and cannot fully replace reading the actual studies. In most cases you can quickly figure out if an article is a review. Many reviews contain the word "review" in the title or publication attributes. Also, if the abstract does not contain specific, quantitative experimental findings, chances are the publication is a review.
• Make sure that you understand the meaning of the terms used. If you do not understand some terms, do not try to guess their meaning from the context - it is easier to get it wrong than right. In fact, some very common words take on a different meaning when used in scientific articles. For example, in a scientific paper, "significant improvement" does not necessarily mean a substantial improvement. In most cases, it only means that the observed improvement was statistically significant, i.e. not likely to have happened by chance. In statistical terms, even a 3% improvement can sometimes be considered significant as long as it has been proved to fall outside the range of random fluctuations. In real terms, however, such an improvement is not worth the time or money.
• Always try to figure out the quantitative magnitude of the observed effects (such as reduction in wrinkle deapth or increased skin elasticity). Determine how these effects were measured. Physical measurements, especially using several independent methods, are more credible than visual observations or self-assessments of the study participants.
• Abstracts are adequate for scanning research data. However, if you come across a particularly interesting or impressive new treatment (and especially if you are seriously considering this treatment for yourself), get the full text of the research articles. Unfortunately, due to copyrighting restrictions we can only provide abstracts on this site. You can obtain the full text of research articles in medical libraries or online at such sites as medscape.com (depending on the journal, they may charge a fee per article).
• Try to find several studies on your subject of interest. Only multiple studies by unrelated, independent teams can fully substantiate the treatment's effectiveness. It is best to rely on the studies performed by the teams that are not affiliated with or funded by the pharmaceutical or skin care companies. Unfortunately, such studies are harder to come by in our market driven world.

How to determine the value and credibility of a research study

Not all scientific studies are created equal: some deserve more credence than others. To distinguish among "the good, the bad and the ugly," we recommend the following.

• Look where a study has been published. The publication source should by no means be your only criterion. However, just as newspapers, TV shows or websites, not all scientific journals are created equal. Respectable peer reviewed journals (e.g. Nature, JAMA, Lancet, etc.) tend to be more rigorous in screening out flawed studies because their reputation is their main asset. Hence the studies they publish tend to be of better quality.
• Long-term studies are more valuable that the shorter ones. Some treatments work well in the short run but quickly lose their effectiveness or even lead to long-term damage. While a short-term study is better than nothing at all, the evidence of long-term benefits and safety (12 month or longer) is far more valuable.
• Studies with larger number of participants tend to be more reliable since a larger sample provides more reliable statistics.
• Studies in cell tissue culture or animal models are far less reliable that those in human subjects. In most cases, an animal or tissue culture study is an indicator of whether a human study is worthwhile, but by no means the evidence that the treatment will work in humans too.
• Studies with better controls hold more value. The golden standard of medical research is a so-called randomized double-blind placebo-controlled study. In such a study the subjects are randomly assigned to two similar groups. One group gets an active treatment and the other, a.k.a. the control group, gets a placebo (hence the term placebo-controlled). Furthermore, neither the subjects nor the administering physicians know which batch has the active substance and which is a placebo (hence double-blind). In the studies of topical skin treatments, for example, the placebo may be the base cream (i.e. inactive ingredients mixed together). Notably, many skin care studies use each subject as her own control by applying different agents (e.g. wrinkle cream vs inactive base cream) to different sides of the face. However, this tends to be less reliable because the subjects often have to apply the treatment themselves and may confuse or forget the application rules. Unfortunately, randomized double-blind placebo-controlled studies are relatively rare in skin care research because they are costly and harder to perform properly.
• Studies with a thorough statistical analysis of results tend to be more reliable. Imagine that we have just finished a study and observed some effect, such as improved skin elasticity. How do we know that this is a real effect rather than a transient natural fluctuation in people's skin properties or a measurement error? Statistics to the rescue! It turns out that statistical methods allow us to estimate the probability (p) that the observed effect resulted from random fluctuations or errors rather than biological activity of the treatment. The smaller the p, the more reliable the result. You should give more credence to the studies where such probability is less than 5% (p < 0.05) or, even better, less than 1% (p < 0.01). Whenever the researches confirm their results with a rigorous statistical analysis they almost always include that value of p in the abstract of the article. Hence if the abstract does not include the value of p, chances are that such analysis has not been done (or perhaps it was, but didn't come out favorably).

This article is by no means an exhaustive tutorial on how to interpret scientific research and distinguish between reliable evidence and junk science. In fact, a comprehensive discussion of this issue would take a sizeable book. While the first-hand analysis of research can give you an edge in your quest for the truth, it is no simple matter and should be done carefully and thoroughly. When deriving conclusions, one should take into account the limitations of both the research itself and one's expertise in interpreting it.