Effect of Makeup Use on Depressive Symptoms: An Open, Randomized and Controlled Trial

The sample size was defined by the maximum error criterion (3 points on the average of the ZSDS). A total of 240 participants were identified and recruited from a previous national survey conducted by our group (observational, cross-sectional, population-based study) based on a representative sample of the target population (n = 2400). The aim of this earlier survey was to estimate the prevalence of depressive symptoms, and the results demonstrated the existence of an association between a lower prevalence of symptoms suggestive of mild depression among participants who frequently used makeup [13].

The ZSDS (also referred to here as the Zung Scale) was used to assess the presence and intensity of depressive symptoms, constituting the primary endpoint of the study. The Zung Scale was developed by the psychiatrist William Zung, from Duke University, in 1965 [14, 15] and continues to be substantially used [16]. It is based on self-reporting the frequency (Likert scale: almost never/sometimes/a lot of the time/most of the time) of 20 different feelings and emotions, of which ten are negative (points given range from 1 to 4 according to the answer obtained) and ten are positive (points given range from 4 to 1 according to the answer obtained). Ultimately, the raw score can vary between 20 and 80 points and is then converted to a percentage index [17]. In a subgroup of participants, a mirror test was also carried out to assess self-image, as well as salivary cortisol measurement to assess stress levels. The rationale underlying the mirror test is to confront the subject with their own reflection, which is relevant for studies involving cosmetics as it is a powerful psychological trigger and has a significant emotional charge, activating aspects related to self-acceptance and thereby characterizing itself as a mild stressor [18]. Salivary cortisol functions as a biomarker in stress research, as a measure of the adaptation of the hypothalamic–pituitary–adrenal axis to stress, contributing to the way in which environmental events predispose an individual to depression [19].

The inclusion criteria were: self-reported infrequent use of makeup (never or almost never/only on special occasions/only on weekends) and a Zung Scale score between 30 and 60, indicating minimum depressive symptoms (in order to measure the effect of introducing frequent use of makeup).

The exclusion criteria were a self-report of having received from a doctor, at some point of life, a diagnosis of mood disorders, current or previous use (90 days before the interview) of any medication with action on the central nervous system (for this approach, non-technical language was used, asking them about the use of medicines considered to be “soothing”, to “facilitate sleep” or any other medicine for “some emotional problem”).

In addition, those presenting with an initial gross score of < 30 or > 60 on the Zung Scale (first application of Zung Scale [Zung 1] of between 0.38 and 0.75), were excluded, as illustrated in Fig. 1.

The recruited participants were randomized (simple alternation of groups according to order of recruitment, carried out centrally in the Data Processing centre) between “test” and “control” groups. Among the 120 participants recruited for inclusion the test (intervention) group, 95 were ultimately included in the study. Those who lived in São Paulo City were invited to constitute a subgroup for "in-person" participation (n = 22); the remaining 73 participants were allocated to the subgroup for remote/online participation. Using the same approach, from the 120 participants randomized for the control group, 50 were ultimately included in the study. Those from São Paulo City were invited to participate in the "in-person" subgroup (n = 16), and 34 were part of the subgroup for “remote/online participation" (Fig. 2).

After signing the ICF, participants in the test group (both in-person and online) received an incentive to adopt the routine of frequently using makeup (intervention) after attending a makeup workshop and being supplied with a sufficient variety and volume of makeup products to comply with the study guidelines for 60 days.

For the subgroup “in-person”, the makeup workshop was carried out in a face-to-face class on self-makeup with a professional teacher/makeup artist. The participants were encouraged to perform a first makeup application, assisted by students of the Aesthetics Course and they received a product kit, as shown in Fig. 3. They also received a link [20] for a 20-min video lesson on self-makeup taught by the same teacher. This lesson contained explanations on the use of the products and the techniques for applying the products, with three makeup options: basic application for day-to-day use; basic application with eye enhancement; application for nights. This video lesson also encouraged them to view videos sent to them 15 and 45 days after the workshop, each lasting 10 s, to reinforce the use of makeup and to rewatch the video class in case of doubts.

For the subgroup with online participation, after signing the ICF, the same product kit and a video lesson were sent. After 15 and 45 days, the videos of encouragement were also sent.

After signing the ICF, participants in the control group received only guidance on completing follow-up questionnaires, without any encouragement to use products or change habits.

A review of the existing depression screen guidelines and instruments [21] concluded that the 2-item Patient Health Questionnaire (PHQ-2) provided a suitable primary screening tool for depression that was reliable, brief and easy to administer. If positive, this should be followed by a suitable instrument, such as the Patient Health Questionnaire (PHQ-2 and PHQ-9), the General Health Questionnaire (GHQ-28) or the “Zung Self-Rating Depression Scale (ZSDS). The authors of the review concluded that the first three instruments were more sensitive and that the fourth (ZSDS) was the most specific [21]. The ZSDS was chosen for the present study based on its extensive use and validation in female health [22], including its Brazilian Portuguese translation [23]. The primary outcome was therefore the change in the level of depressive symptoms measured by the ZSDS at baseline and after 30 and 60 days. All participants answered the online questionnaire with the ZSDS, at the time points indicated in Table 1.

Cortisol is a hormone released in stress situations [24]. Its role in the genesis of depression is considered to be important, including a temporal dimension related to the phases of life and specific variations during environmental events. Cortisol level can also predispose to or trigger the beginning of a depression episode [19].

Saliva was collected in a saliva collection device (Salivette®) for cortisol measurement following the ingestion of water and slight rinsing with mouthwash to eliminate any food residues. Participants were instructed not to touch the collection cotton of the device, but to chew it for 2 min and then deposit it in the proper sterile container and add some additional saliva. The samples were collected by qualified personnel, properly packed and transported, then centrifuged at 3000 rpm for 10 min at room temperature; the supernatant was transferred into the sample cup and measured in a CObas 8000 modular analyzer (Roche Diagnostics, Basel, Switzerland) for cortisol level, including immunoassay with electrochemiluminescence reading (Roche Diagnostics).

The mirror test was originally developed by American psychologist Gordon Gallup Jr. for testing animal cognition [25]. Although the purpose used in the present study has not been validated, we considered its use as an opportunity to assess its applicability in such studies as it has been frequently used by companies in the field of beauty and aesthetics. The rationale underlying mirror testing relates to the person’s confrontation with their own reflection. This confrontation is relevant for studies involving cosmetics because it is a powerful psychological trigger characterized as a mild stressor, with significant emotional load and activation aspects related to self-acceptance [18].

The mirror test was applied in a neutral room with air conditioning and the participants answered the following open questions: (1) What do you see in front of the mirror? Describe; and (2) How would someone else describe you on a first date? The answers were recorded and transcribed. The adjectives contained in the answers were organised and associated with scores, as shown in Table 2, for statistical analysis. Higher scores were attributed to the most positive adjectives.

The secondary outcomes were only measured in the in-person participants: salivary cortisol concentration, at baseline and after 30 and 60 days, and self-perception of the image from the mirror test, at baseline and after 60 days.

The results were analysed using SPSS v.21 software (SPSS IBM Corp., Armonk, NY, USA). Student's t-test was used to assess two means of independent samples; the paired t-test was used to assess the means of dependent samples; and the repeated measures analysis of variance (ANOVA) was used for intragroup comparisons between more than two means of dependent samples. The significance level was set at 5%.

Among the 95 participants in the test group, despite stimuli and guidance, only 82 participants reported an increase in the frequency of makeup use, and only 66 of these effectively reached the level of “frequent use”. In comparison, from the 50 participants in the control group, none of whom received stimulus or guidance, 20 participants reported a spontaneous increase in the frequency of makeup use, with 11 of them achieving the level of “frequent use”. The diagram in Fig. 4 shows that the procedures “change or not” the frequency of makeup use determined new subgroups. The analysis was kept for the originally defined subgroups only in those cases where the change in frequency of makeup use over time would not be a determining factor of bias.

Regarding depressive symptoms, Fig. 5 shows that, considering the total number of participants and comparing the averages of the three ZSDS measurements, significant reductions (P = 0.04) were identified in the intragroup comparison between the second and third measures compared to the first, in the subgroup that did increase their frequency of makeup use. No difference was observed among the participants who did not increase the frequency of makeup use.

Considering the total number of participants and analyzing the subgroups “adopted or not the frequent use of makeup”, Fig. 6 shows a significant reduction between the average of the second (P = 0.04) and third (P = 0.003) measurements compared to the first, among those who started frequently using makeup. Additionally, for this subgroup a significant reduction (P = 0.04) was identified in the average between the third versus the second measurement. No difference was observed among the participants who did not reach frequent use of makeup (3 or more times per week).

Comparison of average salivary cortisol levels in the test group revealed a significant reduction in cortisol level after the makeup workshop + first makeup compared to before these interventions (paired t-test, P = 0.000) (Fig. 7). Comparison of average salivary cortisol levels between the test and control groups revealed a significant reduction in cortisol level in the test group after the makeup workshop + first makeup compared to the control group (Student’s t-test, P = 0.019). This result indicates that the interventions had an impact on the participants.

Analysis of the variation of these average levels at the follow-up time points of 30 and 60 days revealed no significant differences (Fig. 8, suggesting that the effect generated by the makeup workshop and first makeup application was not maintained over time, being restricted to the period of the activities. The role of cortisol, especially in its proximal dimension [19], was notable between the dosages before and after the makeup workshop and first makeup application. The non-routine situation in which the participants took part may be related to this result, as the effect did not last into the follow-up period.

In the mirror test, there was a significant increase in the score obtained in the test group before and after the makeup workshop and first makeup application (paired t-test, P = 0.000) and compared to the control Group (t-est, P = 0.002) (Fig. 9). The two comparisons indicate that there was a positive impact of the intervention, reinforced by the fact that the initial average in the test group was lower than that in the control group.

Figure 10 shows no significant variation in the average score between the first and second measurements in the subgroups that do not increase the frequency of makeup use. On the other hand, in the group that did increase the frequency of makeup use, there was a significant increase (P = 0.000) between the first and second measurements. This was also observed among the second measurements of the subgroup that reached the frequent use status of makeup use (P = 0.000), compared to those that did not.