Not all science is created equal
The first step in understanding whether a skincare product will work or not is to look at the science behind it. Studies on individual cells or animals can point us in the right direction, but the only surefire way to know how a product works is to test it on humans. In human testing, the gold standard is a Randomized Controlled Trial (RCT).
For example, let’s look at a hypothetical anti-wrinkle cream. If we wanted to test this anti-wrinkle cream using an RCT design, we would first need to get a group of people and randomly divide them into two equal-sized groups. One group would be given our anti-wrinkle (active) cream, and the other one would get a standard moisturizer (or a placebo). Ideally, this study would be blinded, meaning the participants wouldn’t know whether they are getting a standard moisturizer or the anti-wrinkle cream. The research team would not know which cream is which either – they would only know the groups in terms of treatment A and treatment B. After the study was completed and the study results were analyzed, we would know if treatment A is better than treatment B. Only then would we unmask treatments A and B, discovering which cream is which, and which treatment works better. Ideally, there would be more than one RCT with similar results. Multiple RCTs would give us a firm understanding of how this anti-wrinkle cream works on human users.
A study that was not randomized or blinded would provide less reliable evidence. If everyone in the study got the anti-wrinkle cream, this would leave the study without a placebo comparator. This type of design is known as a cohort study. Such a study would rely only on analyzing before and after, making it prone to error and bias.
Even lower on the evidence list would be a case-series study. For example, we could go to the sales counter of a skincare brand and observe the skin of all the regular users that visit that counter. We could then compare this group with a group of women that use another skincare brand. These types of study designs can be helpful, but they can also be very misleading; there could be bias in selection and other confounding variables at play. For example, if one sales counter we study is located in an affluent area and the other counter is in a lower-income area, the wealthier women may be using other high-quality products than the women from the lower-income counter. These outside factors would bias the results of the study.
The lowest level of evidence would be making conclusions about something based only on its ingredients, without human studies. This is especially important in cosmetic skincare as there is often more than one ingredient included.
