How Frequently Do We Touch Facial T-Zone: A Systematic Review

Background: Researchers across the world are emphasising the importance of hand-washing and limited touching of face to curb the spread of COVID-19. However, access to safe water and hygiene is inadequate in many places around the globe; hence T-zone touching restriction is considered more worthwhile compared to other prevention strategies. Aim: A systematic review was carried out to appraise the frequency of T-zone (eyes, nose, mouth, chin) touching in humans to comprehend the challenge of its restriction, and thus support public health professionals to produce evidence synthesis guidance for public. For this systemic review, data were collected by keyword searching, and several online databases were searched. The PRISMA checklist, PECO protocol and STROBE guideline were followed in this review, and pooled data were analysed in R version 4. Result: Total of 10 single arms observational studies were included. The pooled average (SD) facial self-touch per hour was 50.06 (±47) times, and a specific touch of T-zone was 68.7 (±27). T-zone self-touch within the total facial self-touch was found higher R = 0.680, with 95% CI 0.14, 0.91, P = 0.02 and X2 = 167.63, P < 0.0001. Conclusion: The review found that face-touch is a type of consistent regulatory movements. Control of T-zone touch requires extensive behaviour intervention and community awareness.


Search strategy
Several online databases were searched, including Ovid Medline (PubMed, Embase, Scopus), Science Direct, Auckland University Library (online), EBSCOhost, Google Scholar, the Web of Science and Cochrane Central Register of Controlled Trial. The keywords and search modifiers used were: "face touch", "face" OR "face" OR "self" AND "touching" OR "epidemiology" OR "frequency" OR "epidemiology" AND "COVID", "respiratory illness" "CoV" "SARS". In addition, A hand search (i.e., manual process of screening pre-defined and pre-selected peer-reviewed journals and other publications) of the references of included studies was also conducted. Several international goldstandard methodologies were implemented in this review, including PECO (participants, Exposure, comparison, and outcomes) [12], SYRINA (Systematic Review and Integrated Assessment) [13], and PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) checklist [14].

Eligibility criteria
The following inclusion criteria were considered: 1) participants: humans, both sexes, and all age groups were included; 2) exposure: touching face including T-zone; 3) outcome: frequency of touching; 4) publication date: there was no restriction on publication date. The closing updated search was carried out in May 2020; 5) language: no language restriction was applied to avoid publication bias; 6) study design: all type of study designs were accepted. All kinds of research papers were included: full publications, letters, conference papers, and theses. The exclusion criteria included: 1) study design: animal studies only; 2) participants: animals only; 3) exposure: touching of other parts of the body; 4) outcome: papers described participants' knowledge only, and did not include any findings.

Data extraction
Reviewer (JR) perused the titles and abstracts of citations identified in the search, and full manuscripts of potentially eligible articles were retrieved for review after removing the duplicates. EndNote X9 was used to remove the duplicates. Reviewer (JM) extracted the data following the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) [15] statement guideline for observational studies. Reviewers (JR, JM, BF) resolved any disputes by discussion and recurring scrutinising. The data items pulled out from the eligible paper included: title, author, year of publication, location of study, study settings, method of measurement, statistical method, sample size and characteristics, confounders/bias adjustment, and main results.

Data mining and statistical analysis
Raw data were processed into useful standard meta-data for analysis. All outcomes were transformed to "face touch per hour" in the review. Pooled data from the included studies were analysed using several statistical methods in R version 4. A Gaussian distribution was applied to understand the proportion of the scores which lie over a certain interval with high confidence based on other research conducted. Principal Component Analysis (PCA) was used to emphasise variation and boxplot was constructed to visualise the data. Hypothesis testing was done by Chisquare analyses and one-way ANOVA test. Pearson's correlation coefficient was conducted to find the degree of correlation between face touch and T-zone touch.

Quality appraisal
Observational studies were assessed using a modification of the Newcastle-Ottawa Scale for use with cross-sectional and cohort studies [11]. Cross-sectional studies had a potential maximum of five stars. Figure 1 shows the details of the study selection. Through database search and other sources (such as magazines, newsletters, and tabloids), a total of 96,871 studies were retrieved. However, after removing duplicates, 8,928 reviews were included for screening. Title and abstract were screened applying the inclusion and exclusion criteria, assigning 47 for full-text review. For non-English studies, Google Translate was used for translating one French journal article. Seventeen publications were carefully reviewed, and seven were excluded due to unclear data and repetition of the same study. Finally, 10 observational studies published between 1973 to 2019 were included in the subsequent review.

Study characteristics
Included studies were single-armed observational pieces of research, and participants were university students or office employees, excluding one study [16] that took place at a petting zoo and surveyed the frequency of face touching in public visitors. Six out of 10 studies were from the USA [1,5,7,[16][17][18], one from the UK [2], one from Australia [19], one from Japan [20], and the later one [21] took place in Japan (Osaka University) and UK (Cardiff University). The participants of six studies [1,2,17,[19][20][21] were university students, and the researchers observed them during lecture time or assigned work; two studies involved health professionals [7] and medical students [19] while dealing patients; one study included public visitors [16], others involved researchers [18] and employees [5]. The number of participants in most studies was less than 50, but one [16] included 574 participants.

Summary of study results
Overall, the quality assessment of the publications based on the Newcastle-Ottawa Scale of these studies was rated as satisfactory, good or very good. Table 1 shows the result of the quality assessment of included studies. The included observational studies reporting results are summarised in Table 2. The aim of most of the studies was the observation of the frequency of face touch. One study used this outcome to investigate the cross-cultural cerebral function [21], one study to examine the accuracy of the self-report [1]. At the same time, one study compared the behaviour with primates [2], and some to evalu-ate the application of the hand hygiene concept in reality [7,19]. Participants of all studies were blinded except in four studies [1,17,18,20]; one of them [1] compared the accuracy and reactivity of self-monitoring, so the participants knew the purpose of the study. However, we have extracted the baseline data for this study [1] where two researchers observed the frequency of their face touch and they were blinded. In the other three studies [17,18,20], participants were informed that they were surveyed, but they were not particularly well-versed about the face self-touching inspection. Three studies [17,19,20] used videotape recording, while two [2,21] used PET computing facilities and magnetic tape data recorder. Another study from Japan took place in a simulated train cabin and was video monitored [20].
Face touching pattern showed differences in sex, age, hand domination and culture. Europeans were found touching face more frequently than Asians; they touched chin and mouth mostly while their counter group touched nose and eyes [21]. There were no differences in frequency in sex and hand domination; male and female frequently touched the face [16,20,21]; also, there was no difference in left-handed and right-handed participants [2]. However, sex differences were observed only in the presence of wearing cosmetics [20]. Participants who had less hand hygiene awareness were often found touching their face [7,18,20]. Health professionals (doctors, nurses, laboratory technicians) were more aware of self-touch than other staffs [7,18]. More remarkably, one study conveyed the differences in point of touch being dependant on public visibility. Participants touched or picked nose more in the amphitheatre arrangement than in the conference sitting plan [5]. The spontaneous self-face touch was compared to this behaviour in apes in one study, and the pattern of face touching was found comparable to gorillas, orangutans and chimpanzees [2]. The use of left hand to touch face was frequent [2,16], which supports the conjecture of emotional dominance in the right cerebral hemisphere and cortical functioning [3,11]. Face touch frequency also differs with the nature of the task. No task, listening to music, smartphone use, emotions, memory task were found related to spontaneous self-touch [5,17,20,21].  Selection 1) Representativeness of the exposed group.
a) Truly representative of the average person in community* b) Somewhat representative of the average person in community* c) Selected group of users d) No description of the derivation of the group 2) Selection of the non-exposed group.
a) Drawn from the same community as the exposed group* b) Drawn from a different source c) No description of the derivation of the non-exposed group a) Secured record (e.g. lab)* b) Structured interview or questionnaire* c) Written self-reports d) No description Confounder 1) Comparability of groups on the basis of the design or analysis.  T-zone = 42.2% and 57.8% = nonmucosal surface.

Results of pooled data
Average touch of self-face per hour was 50.07 times (SD = 47), and T-zone (eyes, nose, mouth and chin) was 68.7 (27). Detailed results were presented in Table 3, and the distribution of the standard deviation of T-zone touch stressed in Figure 2. There was a lack of symmetry in the data distribution. The skewness factor was -0.05, which indicates that a distribution of the database did not follow any normal distribution. Kurtosis value for the T-zone was -0.89, indicating a lack of outliers in the database as the extreme values are less than those of normal distribution. The correlation of T-zone touched in the total face touch was compared with various statistical methods and shown in the Table 4. Correlation coefficient R = 0.680 with 95% CI 0.14, 0.91, P = 0.02.
The one-way analysis of variance (ANOVA) for the face touched in T-zone is presented in Table 6. The comparisons shown borderline significance value (P = 0.0643).
The Chi-square test for the given probabilities for mouth, nose, eyes separately, and T-zone touch are elaborated in Table 5. Assuming an alpha of 0.05, we reject the null hypothesis.
The mean value for the components of T-zone is summarised in Figure 3. It was found, based on the sample study, that chin (29%) and nose (27%) were frequently touched compared to other parts of the face. The overall mean and SD distributions are shown in the Figure 4.

Discussion
This review has identified a small body of evidence, primarily from observational studies, which helped us to produce a conjecture that the prevention of self-inoculation of COVID-19 requires extensive behaviour control approach to avoid pandemic related economic and health consequences. Overall, this review has recommended that face touch is elicited in every human being without any stimulation and is a type of consistent regulatory movements like posture changing. This is not to suggest that self-touching should be recognised; self-touch exhibits not only a person's anxiety or uneasiness but mostly some comprehensive state of emotional and working memory homeostasis [11]. Hand hygiene (i.e. hand wash), social distancing and avoiding T-zone touch are suggested by the clinical researchers as fundamental instrumental acts to defend from COVID-19. However, in cases of running water scarcity, hand-washing strategy to reduce transmis- Table 3: Mean (SD) of frequency of face and T-zone touch.      sion of infection will be ineffective [22,23]. Similarly, social distancing makes people slow down and reconsider several issues that they take for granted, but could increase anxiety, stress and other consequences [24,25]. Relatively, awareness of T-zone touch could significantly reduce the infection rate.

Variables Mean SD Minimum Median Maximum
To increase compliance with public health speak may need to introduce where the message could include both positive descriptive norms (information on desirable typical behaviour) and prescriptive injunctive norms (social approval for such action). This can be achieved by targeting behaviour alteration domains: environment, habits and motivations. Scientists recommend several ways to implement these practices: counter habit (i.e., training to redirect the impulses), behavioural and physical barriers (e.g., hand clasped sitting posture, wearing makeup/masks/face shields/hand gloves), and mindfulness (e.g., wearing hand gloves/perfumes in hands to remind as the hand is nearer to the face) [9].
In conclusion, the results of this review suggest that without reinforcing specific "behaviour control approach" T-zone touch restriction, which is the only beneficial approach compared to other mitigations, would be a highly challenging intrigue in public health.

Notes
1 "Face-touching was defined as touching any part of one's face, head, hair or neck (or extension of this area, such as glasses or earrings) above an imaginary plane resting on the shoulders, with one's hand (or extensions of the hand, such as a pen or drink container)" Nelson RO, Boykin RA, Hayes SCJBr, therapy. Longterm effects of self-monitoring on reactivity and on accuracy. 1982; 20(4): 357-63. 2 When a person's contaminated hand/s transmit germs from one part of the body to another, it is termed as self-inoculation.