Determinants of compliance to the facemask directive in Greece: A population study

Purpose Primary objective of this study was to identify potential difficulties and/or discomfort when using a facemask. Moreover, to explore the impact of spectacles, contact lenses and visual acuity on the compliance to the facemask directive. Methods This is a prospective study that was conducted at the Department of Ophthalmology, University Hospital of Alexandroupolis, Greece between June 2020 and August 2020. Greek speaking citizens with permanent residency in Greece above 18 years old were included. A custom questionnaire (DeMask-20) was constructed and validated, which pertained to the perceived difficulty and discomfort when using a facemask. It contained 20 items grouped in 8 subscales (driving, near vision, distance vision, ocular discomfort, role limitation, collaboration, dependency on others, emotional stress). Perceived difficulty and discomfort when using a facemask, compliance and correlations of compliance with DeMask-20 scores, demographics, spectacle and/or contact lens use, and visual acuity were evaluated. Results The number of factors was determined through factor analysis. Cronbach’s alpha ranged from 0.716 for the “Role limitation” subscale to 0.938 for “Ocular discomfort” subscale. 1,214 participants (402 men, 812 women, mean age 36.79±12.50 years) completed the DeMask-20 instrument. Mean DeMask-20 score of all study participants was 3.79±0.71. Significant differences in DeMask-20 score were detected in gender (p = 0.009), spectacle use (p = 0.034), contact lens use (p = 0.049), and binocular distance visual acuity (bDVA) (p = 0.001). Mean compliance of all participants was 4.05±0.96. Men, people <50 years and spectacle wearers showed significantly worse compliance (p<0.05). Moreover, professional workers and professional drivers demonstrated significantly better compliance (p = 0.008 and p = 0.047). Significant correlation was detected between compliance and DeMask-20 score (p<0.001, R2 = 0.471). Significant correlations were detected with driving, near vision, distance vision, collaboration, role limitation, emotional stress (p<0.05, R2: 0.386–0.493). Conclusions Factor analysis suggested that the DeMask-20 instrument demonstrates adequate validity, while Cronbach’s alpha indicated sufficient internal consistency of all subscales. This study provided the necessary methods that could evaluate compliance trends and the efficacy of healthcare interventions against COVID-19. Our outcomes suggest that young males who use spectacles should be targeted by Greek Healthcare authorities in order to improve compliance rates.

Introduction until the end of this month [17]. Regarding facemask use directives during the period of study, it was mandatory for healthcare providers, professional drivers and passengers, for workers and general public at airports and the shop staff. As the number of cases was increasing, the authorities widened these measures, therefore from 10th August 2020 facemask usage was obligatory in places of worship, supermarket and shops for both customers and staff [18][19][20].
Within this context, the primary objective of this study was to identify potential difficulties and/or discomfort when using a facemask, evaluate the impact of spectacles, contact lenses and visual acuity, and explore their impact on compliance to the directive on facemask use.

Setting
This is a prospective study. Study protocol adhered to the tenets of the Declaration of Helsinki. The institutional review board of Democritus University of Thrace approved the study protocol. The study was conducted at the Department of Ophthalmology in the University Hospital of Alexandroupolis, Greece, between June 2020 and August 2020. Official registration number of the study is NCT04501172 (https://clinicaltrials.gov/ct2/show/NCT04501172)

Participants
Participants were contacted through social networks (Facebook) with a link to an online questionnaire. A cover letter describing the scope and eligibility criteria of the questionnaire accompanied the link. The online questionnaire was open for one month (June 2020 to July 2020). Eligibility criteria, which were described in detail in the cover letter, included age above 18 years old, adequate literacy of Greek language and permanent residence in Greece. Questionnaires completed by people younger than 18 years were excluded from data analysis. All other questionnaires were considered to meet the inclusion criteria and continued for further analysis.

The DeMask-20 instrument
Literature review on a validated instrument regarding facemask-wearing trends for Greek speaking populations returned no results. Thus, an exploratory interview study was designed to create the baseline for a questionnaire development. A panel consisting of 2 ophthalmologists, 2 nurses with experience in ophthalmology outpatient care, and a psychologist were recruited for the exploratory study. A number of items covering attitudes on facemask wearing were summarized and written as interview questions. Individual interviews with 10 participants who had no previous contact with any of the members of the panel took place. The interviews were analyzed and the findings served as the basis for identifying the variables of interest that would be operationalized in specific items (questions) to be used in our instrument.
The final version of the questionnaire consisted of 2 parts. The first part pertained to the participant's demographic characteristics, with items regarding age, reported binocular distance visual acuity (bDVA), spectacle and contact lens use, potential health vulnerability [21], and compliance to the facemask wearing directive. For the enrolment in the health vulnerable group, study participant had to meet one of the following criteria: age of 65 years or older, severe heart or respiratory disease, resistant hypertension, uncontrolled diabetes mellitus, severe neurological or neuromuscular disease, kidney or liver failure, high body mass index (BMI), cancer, immunodeficiency or pregnancy [21].
The second part of the instrument consisted of twenty items that constructed 8 subscales and pertained to the potential difficulty and/or discomfort when using a facemask: DeMask-20 subscales were: a) driving (2 items), b) near vision (5 items), c) distance vision (3 items), d) ocular discomfort (3 items), e) role limitation (3 items), f) collaboration (1 item), g) dependency on others (2 items), and h) emotional stress (1 item). The 6-category ordinal polytomous items I1 to I11 were transformed to 5-category Likert-scale items for easier data interpretation and data analysis. Specifically, the categories "a) I need to remove my facemask" and "b) I have almost stopped this activity because of my vision and the use of a facemask" of the original item-version were merged into the category "1 = significant difficulty/discomfort". The numbering of the rest categories was converted accordingly (c ! 2 = great difficulty, d ! 3 = some difficulty, e ! 4 = little difficulty, f ! 5 = no difficulty). On the other hand, the original I12 to I20 were 5-category Likert scale items (1 = absolutely agree, 5 = absolutely disagree) and no conversion was necessary. A total DeMask-20 score for each participant was obtained from the average of all subscales. Items I1-I5 were optional (I1, I2 were addressed to professional drivers, and I3-I5 to professional workers), while items about demographic characteristics and I6-I20 were mandatory and had to be answered to allow the online questionnaire to be submitted.

Statistical analysis
Construct validity of the questionnaire was evaluated by exploratory factor analysis (EFA). As extraction method, Principal Component Analysis (PCA) was applied because it is one of the most simplified and commonly used methods of EFA. Initially, we used an eigenvalue (EV) > 1 (Kaiser's criterion) to determine the number of factors, in combination with a scree plot. To determine whether the data were adequate for factor analysis (FA), the Kaiser-Meyer-Olkin (KMO) measure was calculated. KMO scores between 0.8 and 1 indicate the appropriateness of the sample for FA. In addition, the Bartlett's test of sphericity was calculated. If the test was significant (p < 0.001), the data were suitable for FA [22]. Finally, since we expected that underlying factors may be related, we used oblique rotation (direct oblimin) to optimise configuration on factors (Delta = 0) [23,24]. Items were considered loaded onto a factor if values exceeded 0.40 and were considered uniquely loaded if cross-loadings on other factors were less than 0.40 [23,24]. After the number of factors had been determined, the internal consistency of DeMask-20 subscales was evaluated by Cronbach's alpha (a) estimation.
Data distribution of the questionnaire items was tested with Shapiro-Wilk test. Betweengroup comparisons of data for which the hypothesis of normality is satisfied were made using independent samples Student's t-test or one-way ANOVA. Data for which the hypothesis of normality is not satisfied were assessed with Mann-Whitney U test or Kruskal-Wallis H test. P-values lower than 0.05 were considered statistically significant. All statistical analyses were performed with SPSS Statistics for Windows software (version 20.0, IBM Corp.)

Construct validity and reliability
Factor analysis revealed six factors with EV > 1: EV factor 1 = 7.933, EV factor 2 = 2.568, EV factor 3 = 1.644, EV factor 4 = 1.511, EV factor 5 = 1.163, and EV factor 6 = 1.036, which explained 79.27% of the variance of the items (Fig 1). The Kaiser-Meyer-Olkin (KMO) measure, representing the sampling adequacy for the analysis, was 0.835. The Bartlett's test of sphericity was significant (p < 0.0001), rejecting the null hypothesis that our items are uncorrelated, and indicating that FA would be useful as a data reduction technique. The pattern matrix (Table 1) demonstrates the items that are loaded to each factor after rotation. All loading values of the items were above 0.7. Item I17 was not included in the subscale "Ocular discomfort", which included items I18, I19, I20, due to its low loading value (0.496) and was evaluated as a distinct subscale which contained a single item. The same was applied for item I5 that demonstrated low loading values to several factors. Reliability analysis was done by Cronbach's alpha estimation as an index of internal consistency for each subscale (Table 2) [25]. Cronbach's alpha ranged from 0.716 for the "Role limitation" subscale to 0.938 for "Ocular discomfort" subscale. Thus, it becomes obvious that the majority of the subscales presented high internal consistency.

Discussion
The COVID-19 pandemic has introduced the necessity of facemask use as a type of personal protection equipment (PPE) for the reduction of the SARS-CoV-2 transmission. A great variety of facemask types is available for the general public [9]. Among them, the traditional medical facemasks, known as "surgical masks", some more specialized masks such as FFP2, FFP3, N95, KN95, but also homemade (cloth) non-certified facemasks have become part of daily life [9].
Despite the significant role that facemasks play during this pandemic due to the beneficial impact on the prevention of the virus SARS-CoV-2 transmission, they had traditionally been associated with discomfort and increased difficulty in certain activities of daily living [15,[26][27][28][29]. The perceived difficulty when wearing a facemask could easily contribute to reduced compliance to the facemask-wearing directive and potentially increase the rate of virus transmission. Moreover, former researchers reported that people who are using spectacles and/or contact lenses perceive significantly more difficulty when compared to the rest of the population, primarily due to fogging of glasses and intense tear evaporation, especially when the facemask is not properly fitted [15].
Within this context, we attempted to measure the perceived difficulty and/or discomfort of Greek people when wearing a facemask, and explore potential correlations with compliance to the facemask-wearing directive by the Ministry of Health and Welfare. Special attention was given to identify whether spectacle, contact lens use and suboptimal visual acuity contribute to lower levels of compliance.
Since no relevant validated instrument existed for Greek-speaking patients, we constructed the DeMask-20 questionnaire, which quantified the perceived difficulty when wearing a facemask in 20 items, grouped in 8 subscales. Factor analysis suggested that the DeMask-20 instrument demonstrates adequate validity, while Cronbach's alpha indicated sufficient internal consistency for all subscales.
Our participants presented an average DeMask-20 score of 3.79 indicating that Greek people do actually perceive a variable amount of difficulty and discomfort when wearing a facemask. Women reported significantly worse scores than men, identifying difficulty in collaborating with peers, and due to ocular discomfort and emotional stress. Spectacles and contact lenses also contributed to worse DeMask-20 scores, primarily due to difficulty in distance and near vision activities (for spectacle users) and due to distance vision activities, collaboration and driving (for contact lens users). Moreover, lower levels of visual acuity were associated with worse DeMask-20 scores. Despite the different methods, our outcomes are in accordance to former publications, which also revealed a negative impact on the quality of life when using a facemask. Morishima et al. [15] performed a repeated cross-sectional survey in Japan in 2009, 2012 and 2015 for the use of facemask for the protection from H1N1 and common cold viruses. According to their outcomes, the most common problem was humidity in the facemask, fogging up of glasses, difficulty in breathing for both genders, and makeup coming off for women. Similarly, Lim et al. [26] and Ong et al. [14] analyzed the impact of PPE such as N95 facemask on the development of headaches of healthcare workers while attending to patients during the 2003 SARS epidemic and COVID-2019 pandemic, respectively. From these surveys, it was concluded that de novo PPE-associated headaches or exacerbation of pre-existing headache disorders are developed in the majority of healthcare workers, which leads to frequent abuse of analgesics. Another phenomenon described in the literature during COVID-19 pandemic is retroauricular maskinduced dermatitis, which is caused by ear loop facemasks [13]. Ocular problems due to facemask use were reported by Moshirfar et al. [27] who indicated that, during the COVID-19 pandemic, an increase in ocular irritation and dry eye disease symptoms was observed among people using a facemask regularly, patients and healthcare workers. Among the possible explanations of this phenomenon is the tear film evaporation accelerated by the increased airflow toward the eyes, which may result in irritation or inflammation of the ocular surface when it lasts for hours or days. An additional interesting explanation about the corneal irritation among staff members using taped facemasks for the prevention of air convection toward the eyes is the fact that the adhesion of the tape to the skin of the upper cheek may prevent the lower eyelid from normal excursion resulting in mechanical ectropion with secondary lagophthalmos. In fact, the same authors hypothesized that dry eye caused by evaporating of the tear film, an essential barrier against pathogenic invasion, but simultaneously the increase of eye rubbing because of the ocular discomfort could result in a higher vulnerability to pathogens through the eyes.
Regarding compliance to the facemask-wearing directive, 76% of study participants declared full or almost full compliance, 18.1% sometimes, and 5.1% no compliance. Women complied more than men despite worse DeMask-20 scores, age was positively correlated with compliance, and professionals who were obliged to wear a facemask in their working  environment (including professional drivers) presented better compliance, as well. As expected, spectacles were associated with significantly worse compliance.
Compliance was significantly correlated with the total DeMask-20 score and almost all subscale scores. This outcome provides essential new information for some of the fundamental reasons that explain why a person complies with the facemask directive, or not. Within this context, it provides the necessary data to the Healthcare authorities to implement strategies or interventions to improve compliance rates. Therefore, the primary target group for the Greek Ministry of Health and Welfare should be males, below 50 years old, who wear spectacles. Secondary, they should focus on people who: a) have average or poor visual acuity and experience significant difficulty both in near and distance vision activities, and, b) who have a pre-existing ocular surface disease that is most likely to be exacerbated by the facemask. Poor compliance is most likely also on people who present difficulty when collaborating with others, and those who experience significant emotional stress. Last but not least, poor compliance is expected in people who believe that facemask use reduces their opportunities for personal growth.
Former investigators reported similar results regarding compliance to the facemask-wearing directive. Sim et al. indicated personal discomfort and sense of embarrassment as the primary reasons for reduced compliance [30]. Regarding healthcare providers primary reasons for reduced compliance were discomfort, breathing problems and shortness of breath [31][32][33]. Moreover, young age and male gender were associated with reduced compliance [34]. In fact, according to Capraro &Barcelo, men presented significantly lower rates of compliance than women, especially when facemasks were not obligatory by law. Interestingly, it has also been found that focusing on community protection was associated with higher compliance to the facemask directive than focusing on protecting the individuals themselves [35].
Prior to the interpretation of our results, certain limitations of our study have to be noted. Although we have a robust number of participants, our sample was not stratified. Moreover, since participants were contacted via Facebook, only patients who use Social Media were represented. However, taking into account the significant penetrance of the internet and the social media to the Greek society, we are confident that our outcomes could be generalized for Greece.

Conclusions
To our knowledge, this is the first study that assesses compliance to the facemask-wearing directive in Greece. Moreover, to our knowledge, this is the first study to construct a validated instrument that evaluates the perceived difficulty when wearing a facemask. Within this context, we provided the necessary methods that could evaluate the compliance trends and the efficacy of healthcare interventions in the COVID-19 era. Nevertheless, our outcomes suggest