Abstract
Background:
Makeup greatly impacts normal social lives but can also be a non-pharmacological form of therapy for dementia.
Objective:
To evaluate the therapeutic effect of makeup therapy.
Methods:
We carried out a prospective interventional study on female nursing home residents with dementia, focusing on the chronic therapeutic effect of makeup therapy. Thirty-four patients who received either only skin care (control group, n = 16) or skin care plus makeup therapy (makeup therapy group, n = 18) once every 2 weeks for 3 months were assessed.
Results:
Three months of makeup therapy significantly improved the Mini-Mental State Examination (MMSE) score compared with control patients (*p < 0.05). Artificial intelligence (AI) software revealed that the appearance of age decreased significantly in the makeup group compared with the control, especially among patients without depression (*p < 0.05). Furthermore, a larger AI happiness score was significantly correlated with a greater improvement of ADL in the makeup therapy group (r = 0.43, *p < 0.05).
Conclusion:
Makeup therapy had a chronic beneficial effect on the cognitive function of female dementia patients, while the chronic effect of makeup therapy on facial appearance was successfully detected by the present AI software.
INTRODUCTION
Dementia is one of the greatest social problems around the world. Although a large amount of effort has been directed toward the development of curative drugs for dementia [1–4], it has not yet been totally healed [5]. In addition, due to their side effects, pharmacological approaches cannot be applied to many patients [6]. Therefore, non-pharmacological therapy is still an important approach for treating dementia patients [7–9].
Makeup has a positive influence on social lives [10]. Although many females begin to apply makeup from adolescence [11], the frequency of makeup use decreases as they get older, especially when they develop dementia [12]. Elderly women who routinely performed skincare and added makeup showed better cognitive function than those who only performed skincare without the use of cosmetics [13], suggesting that makeup is a non-pharmacological alternative for the treatment of dementia. However, there are limited reports related to makeup therapy [14].
To evaluate the therapeutic effect of makeup therapy, we carried out a prospective interventional study on female nursing home residents with dementia. We previously reported the immediate effect of this therapy to alleviate behavioral and psychological symptoms of dementia (BPSD) [15]. In the present report, we focused on the chronic therapeutic effect of makeup therapy.
MATERIALS AND METHODS
Study participants and makeup therapy
The present report is a part of a larger project that is focusing on the chronic effect of makeup therapy whose immediate effect was reported in our previous paper [15]. As described in that paper [15], female nursing home residents who had previously been diagnosed as having dementia were enrolled in the present study from October 2020 to November 2020. They were randomly divided into a control group and a makeup therapy group. Patients received either only skin care for 5 min (control group) or skin care plus makeup therapy for 25 min (makeup therapy group). Skin care and makeup were performed by skilled therapists certified by the Japan Wellness Therapist Association, once every 2 weeks for a total duration of 3 months (Fig. 1). Makeup therapy included foundation, powder, blush for cheeks, eyebrow, eyeshadow, and lipstick makeup. To prevent COVID-19 infection, thorough infection measures were implemented, including a health check, washing hands, wearing masks, and the use of a face shield.
View of makeup therapy with thorough infection prevention measures during the COVID-19 outbreak. Permission was granted by the patient.
The present study was approved by the Okayama University Ethical Review Board (approval #OKU-2005-001) and registered in the University Hospital Medical Information Network Clinical Trials Registry (UMIN-CTR) with registration number UMIN000040180. All participants or their substitutes gave written informed consent.
Cognitive, affective, and ADL functions
Patients were evaluated by the following batteries: Mini-Mental State Examination (MMSE) [16] and Hasegawa dementia scale-revised (HDS-R) [17] for cognitive function; geriatric depression scale (GDS; cut off ≥6) [18], apathy scale (AS; cut off ≥16) [19] and Abe’s behavioral and psychological symptoms of dementia score (ABS) [20] for affective function; Alzheimer’s disease cooperative study-activities of daily living (ADL) [21] for ADL. MMSE and HDS-R were evaluated and checked twice by trained examiners, GDS and AS were declared by the patients themselves, and ABS and ADL scores were obtained by the nursing home staff. These assessments were performed before the first treatment (baseline), and 2 weeks after the last treatment (3 months after the first treatment, or the chronic phase).
AI facial analysis
The face of each patient was photographed with a web camera at the baseline and chronic phase, and analyzed with artificial intelligence (AI) software, which was a previously modified version of Microsoft Azure Face (Microsoft Corp., Washington, USA) adjusted to healthy Japanese patients (n = 143, 21–88 years of age) [22]. This AI software quickly and automatically analyzes attributes such as appearance age, gender, and emotion (neutral, happy, sad, contempt, surprise, anger, disgust, and fear). In the present study, the appearance of age and emotion were evaluated.
Statistical analysis
Comparisons of age, cognitive and affective function, ADL, and emotion between groups were carried out with Welch’s t-test or the Mann-Whitney U test. Pearson’s correlation coefficient (r) was used to assess correlations between measurements. All statistical analyses were performed with SPSS 22.0.0.0 (IBM Corp., Armonk, New York, USA). Statistical significance was assumed at p < 0.05.
RESULTS
Patients’ characteristics
A total of 36 patients were initially enrolled and divided into control (n = 17) and makeup therapy (n = 19) groups. One patient from each group dropped out due to hospitalization caused by sudden illness (control group) and treatment refusal (makeup group). Consequently, 34 patients (16 in the control group; 18 in the makeup therapy group) were enrolled in this study.
The average age of patients was 89.8±4.1 years in the control group and 89.0±3.9 years in the makeup therapy group. In both groups, baseline cognitive scores of MMSE and HDS-R, as well as the ADL score, declined (13.3±6.4, 9.7±6.3, and 9.3±5.7, respectively in the control group; 11.4±6.7, 9.0±6.5 11.7±6.0, respectively in the makeup therapy group), while baseline GDS and AS were mildly elevated (8.1±3.9 and 23.7±9.2, respectively in the control group; 7.3±3.0 and 24.5±7.5, respectively in the makeup therapy group). There were no significant differences in the following parameters between the two groups: age, baseline cognitive function, affective function, and ADL scores. There were no adverse events, including skin ailments or COVID-19 infection.
Cognitive, affective, and ADL scores
After 3 months of treatment, makeup therapy significantly improved MMSE score compared to the control, decreasing by 0.1±4.0 in the control group but increasing by 4.9±5.2 in the makeup therapy group (*p < 0.05; Fig. 2). HDS-R improved in the makeup group (4.0±6.2) and was slightly higher than the control (0.4±5.3, p = 0.09). In both groups, even though GDS and AS improved while ABS and ADL worsened after 3 months, none of these parameters were significantly different between the two groups.
Chronic effects of makeup therapy, showing a greater improvement of MMSE score in the makeup therapy group compared with the control (*p < 0.05).
AI facial analysis
At the baseline, age gap (appearance age –real age) was not significantly different between the control group (1.4±5.3) and makeup therapy group (3.1±4.0, Table 1). After 3 months of treatment, age gap barely changed from the baseline (0.0±1.9) in the control group (Table 1). On the other hand, age gap decreased slightly in the makeup therapy group (–1.2±2.1, p = 0.12). In particular, among patients without depression (GDS scores of 0–4), appearance of age decreased significantly in the makeup group (0.88±0.76) compared with the control group (0.88±76, *p±70.05; Fig. 3). Emotion (neutral, happy, sad, contempt, surprise, anger, disgust, and fear) was not different between both groups at the baseline or after 3 months of treatment. However, greater happiness was significantly correlated with higher ADL in the makeup therapy group (r = 0.43, *p±70.05; Fig. 3).
AI face analysis
Age gap = appearance age - real age (years). AI, articicial intelligence.
AI facial analysis, showing a significant decrease of the appearance of age after makeup therapy compared with the control in patients without depression (*p < 0.05).
DISCUSSION
The present study demonstrates the chronic effect of makeup therapy, which improved the MMSE score of female dementia patients (Fig. 2). AI facial analysis revealed that makeup therapy made them appear younger than control patients, especially those without depression (Table 1, Fig. 3). Furthermore, increased AI happiness score after makeup therapy was significantly correlated with an improvement of ADL (Fig. 4).
The chronic effect of makeup therapy benefitted the cognitive function of dementia patients rather than their affective function or ADL (Fig. 2). This was different from its immediate effect, which mainly improved BPSD [15]. Magnetoencephalography demonstrated that a combination of non-pharmacological therapy reduced alpha activity in the right temporal lobe and fusiform gyrus as well as higher low gamma activity in the right angular gyrus [23], and topographic near-infrared spectroscopy revealed that makeup therapy activated the cerebral frontal lobe [24]. In the present study, since we examined the chronic effect of makeup therapy 2 weeks after the last treatment to eliminate its immediate effect, each immediate beneficial effect of makeup therapy on BPSD diminished within 2 weeks.
A significant correlation between change in AI happiness and ADL score (*p < 0.05).
The appearance of age was judged to be younger by AI software immediately after makeup therapy [15]. The present AI facial analysis was performed 2 weeks after the last treatment, when makeup had already been removed. Therefore, the present appearance of age and emotion judged by AI software might reflect the physical or psychological condition of the participants [22, 25]. Facial appearance is correlated with cardiovascular disease risk and survival [26, 27], suggesting that facial appearance can be a screening biomarker of health condition, and may be used to evaluate the therapeutic effect of makeup therapy. Umeda-Kameyama et al. demonstrated that cognitive function is strongly correlated with the appearance of age [28]. The difference between appearance of age of the makeup therapy and control groups was significant only among patients without depression, possibly because appearance of age was younger, even in the control treatment (skin care) among patients with depression (Fig. 3). Dementia patients with depression may benefit more from this intervention [29]. Furthermore, AI facial emotion analysis (Fig. 4) may be superior to self-reported scales such as GDS or AS for evaluating the affective function of dementia patients because of its independence on verbal ability or cognitive function [30].
The limitation of the present study is small sample size (n = 34) and limited study sites (n = 2). A multicenter, randomized clinical trial with a larger number of participants would provide greater and more robust evidence of the benefit of makeup therapy.
In conclusion, makeup therapy had a chronic beneficial effect on the cognitive function of female dementia patients. The present AI software revealed that the appearance of age was chronically reduced by makeup therapy, especially in patients without depression, and AI happiness score after makeup therapy was correlated with an improvement of ADL.
Footnotes
ACKNOWLEDGMENTS
We are grateful to Asai T, Ebata N, Hirata M, Koganei K, Nishida M, Sato K, Yamamoto T, Mori I, Maeda T, Shimada Y, and the nursing home staff of Kandenjoylife for their support. This work was partly supported by a Grant-in-Aid from the Japan Society for Dementia Prevention, a Grant-in-Aid for Scientific Research (C) 20K09370, 20K12044, Challenging Research 21K19572, Young Research 20K19666, 21K15190, and by Grants-in-Aid from the Research Committees (Toba K, and Tsuji S) from the Japan Agency for Medical Research and Development.
