Factors associated with visual function among computer-based administrative workers: a Brazilian cross-sectional study

Introduction Several studies have shown that eye and vision problems are among the most significant issues reported by individuals who use computers at work. Objectives To investigate individual and occupational environmental factors associated with visual function among workers who perform computer-based administrative tasks. Methods This is a cross-sectional study conducted in 2014-2015 with 303 workers of a public hospital in the city of São Paulo, Brazil. The participants answered a structured questionnaire, including the 25-Item National Eye Institute Visual Function Questionnaire. Statistical analyses used descriptive analysis, tests of association and multiple linear regression analysis. Results Most participants were female (61.1%); the mean age was 46.0 (standard deviation [SD]) ± 12.5, and approximately 91.7% of them reported wearing corrective lenses. Regarding visual function, the mean score at the 25-Item National Eye Institute Visual Function Questionnaire was 78.0, SD ±7.1. A regression analysis showed that visual function declined with age (ß −0.218; 95%CI −0.276--0.16l) and effort at work (ß −0.656; 95%CI −0.928--0.383). Conclusions The mean quality of visual health in the studied group was good. The younger the age and the lower the effort at work, the better the visual function. Our results point to the relevance of establishing periodical and preventive health actions, including eye health assessments.


INTRODUCTION
Information technology has expanded in the past decades, resulting in an increased use of computers at the workplace. 1A survey conducted in Europe found that around 30% of workers continually use computers throughout working hours. 2 The proper development of daily activities and work tasks depends on vision for reading and interacting with objects and people. 3he notion of visual function points to a broader conception of vision, understood as the set of mechanisms through which individuals interpret images and their visual environment. 4ye and vision problems are among the most significant health complaints of individuals who use computers at work. 5 Asthenopia, ie, visual fatigue, is one of the most common types of visual impairment.As a rule, the term asthenopia is used to designate any subjective symptom or discomfort related with the use of the eyes. 6In addition, its frequency is increasing among workers in jobs that demand high visual accuracy, such as telemarketing operators. 7,8A study conducted in Brazil found a prevalence of 54.6% of visual symptoms associated with computer use among telemarketing operators. 9omputer vision syndrome (CVS), also known as digital eye strain, is a disturbance of visual function and can be characterized as the presence of one or more symptoms derived from the use of computer screens, such as tired eyes, eyestrain, burning eyes, eye irritation, redness, blurred vision, and dry eyes. 10,11The global prevalence of CVS is estimated at over 70%. 5 Studies conducted in the United States report that 90% of the 70 million workers who use computers for more than 3 hours per day exhibited some clinical sign of CVS. 10 CVS has a multifactorial origin 12 and its known causes are categorized as intrinsic 13 or extrinsic, the latter being further divided into environmental and ocular causes. 10Among the intrinsic factors, musclerelated causes of visual fatigue prevail. 10The extrinsic ocular factors comprise a reduced blinking rate, increased exposure of eye surface, the use of contact lenses or medications, and the presence of systemic and/or external eye diseases. 10vironmental factors are related to poor workplace conditions and include lighting, 14 dust and dryness of the air, and improper shape and position of chairs. 10They might demand continuous changes in visual accommodation and convergence due to the need to focus on different distances and directions, which requires adequate coordination of eye movement for accomplishing binocular vision through image fusion. 15 possible relationship between vision disturbance and psychosocial factors at work is based on the idea that visual disorders are related to the intensity and duration of visual demands, the workers' self-perceived working conditions, and the pathophysiological characteristics of each individual's visual system. 5,16In most cases, symptoms of visual impairment develop when the visual demands posed by the work tasks exceed the visual capacity of individuals to perform them in a comfortable manner. 1 Considering the high prevalence of visual function impairment among workers who regularly use computers at work and the scarcity of studies about risk factors for this outcome, the aim of the present study was to investigate individual and occupational environmental factors associated with visual function among workers who perform computer-based administrative tasks.

STUDY POPULATION AND DESIGN
The present cross-sectional observational study was conducted between 2014 and 2015 with administrative employees at a tertiary public hospital in São Paulo, Brazil.The population was 772 workers, but only 437 met the following inclusion criteria: performing administrative tasks, using computers for at least 4 hours a day, working daytime hours, and having worked for at least 1 year on the current position.Since 125 (28.6%) workers were excluded for being on sick (n = 119) or maternity (n = 6) leaves, 312 eligible employees remained, but nine refused participation.Therefore, 303 (97.1%) employees participated in the study.
The participants worked on workstations distributed across the six floors of the administration building.According to the institutional "Environmental Risk Prevention Program" (Programa de Prevenção de Riscos Ambientais [PPRA]), the environmental lighting system included natural and artificial sources to direct the light flux.No local or supplementary light sources were present at the individual workstations.According to the PPRA, the measured illuminance in the investigated areas varied from 480 to 500 lux.

DATA COLLECTION AND STUDY VARIABLES
Data were collected during the ophthalmological examination performed in the periodic medical consultation, when participants filled out the selfreporting questionnaire on individual characteristics (sex, age, educational level, marital status, number of household residents, family income, routine medical examinations, smoking, alcohol intake, physical activity, duration of sleep during the workweek, clinical visual disturbances, and use of glasses or contact lenses), occupational aspects/working conditions (employment status, tenure at the current position and at the institution, weekly working hours at the hospital, daily screen time, and environmental conditions such as body postures, work tools, and psychosocial stressors) and visual function.
The third part of the questionnaire focused on aspects of the psychosocial environment at work by means of the Job Stress Scale ( JSS) and the Effort-Reward Imbalance scale (ERI).The JSS is an abridged version of the Job Content Questionnaire ( JCQ) based on the demand-control model and validated for use in Brazil. 17The ratio between the demand and control scores provided a work strain score in which higher scores represented greater strain.The transcultural adaptation of ERI for Brazilian Portuguese was also employed. 18The ratio between effort and reward scores provided a score for imbalance.
The outcome variable was visual function.It was assessed with the National Eye Institute Visual Function Questionnaire (NEI VFQ-25) in its Brazilian Portuguese version. 19NEI VFQ-25 comprises 25 questions clustered into the following 12 subdomains, with scores ranging from 0 to 100%: general health, global vision, ocular pain, difficulty with near vision activities, difficulty with distance vision activities, limitations in social functioning due to vision, mental health symptoms due to vision, role limitations due to vision, dependency on others due to vision, driving difficulties, and limitations with color and peripheral vision.The global NEI VFQ-25 score ranges from 0 to 100: the higher the score, the better the visual function.

STATISTICAL ANALYSIS
Descriptive analysis was based on the calculation of means, medians, standard deviation (SD), and minimum and maximum values for quantitative variables, as well as proportions for qualitative variables.
The Kolmogorov-Smirnov test was used to investigate the adherence of NEI VFQ-25 scores to the normal distribution; the result, p = 0.289, enabled the use of parametric tests in the statistical analysis.
A univariate analysis of the factors associated with NEI VFQ-25 scores was performed using Pearson's correlation coefficient for quantitative variables; an analysis of variance (ANOVA) was performed for categorical variables with constant variance, and the Mann-Whitney (dichotomous) and Kruskal-Wallis (three or more categories) tests were done for variables without constant variance.A Tukey's post hoc test for multiple comparisons was then performed.The homogeneity of variances was assessed by Levene's test.
A multiple forward stepwise model was fit including the variables that exhibited p < 0.20 on the univariate analysis; the p-value determined the order of inclusion into the multiple model.Qualitative variables were transformed into dummy variables, considering the category with the highest mean score on NEI VFQ-25 as reference.Potential confounding and interaction effects were tested.The descriptive level of p < 0.05 was adopted.

ETHICAL ISSUES
The study was approved by the Research Ethics Committee of Escola de Saúde Pública, Universidade de São Paulo (USP) (ruling no.257,510) and the Research Ethics Committee of Hospital das Clínicas, Faculdade de Medicina, USP (ruling no.705,863) and complied with the Declaration of Helsinki.The recruited employees agreed to participate by signing an informed consent form.

DESCRIPTIVE ANALYSIS
The total population of workers comprised 437 individuals, but only 312 of them were eligible for the study.The participants were 303 (97.1%) eligible employees, and non-participants did not differ from participants as to their age range or tenure at the institution; however, there was a statistically significant difference (p = 0.013) in sex, since losses were higher among men (4.6%) compared to women (0%).
Table 1 describes the study population as to their personal characteristics.Around 61.1% of the participants were female; 34% had complete secondary education; 62% were married or had a partner; 72.9% lived in households with up to three people; and 29% reported a family income of less than 5.2 times the Brazilian minimum monthly wage.Around 66.7% of the participants reported they underwent routine medical examinations at intervals of less than 2 years; 95.7% did not smoke; 72.6% consumed alcohol one or more times per week; 58.4% had regular physical activity; 93.6% reported sleeping 6 or more hours per night during the workweek.
Table 2 shows information on continuous variables.The average age of the sample was 46.0 (SD) ± 12.5 years old, varying from 20.0 to 74.0, median 48.0 years old.The mean time working at the current position ‡ Tukey test: workers with incomplete higher education had a higher average than workers with complete higher education: p = 0.043 § Tukey test: single workers had a higher average than married workers: p = 0.001 || Times the minimum wage (MW) at the time of data collection (MW = BRL 788.00).
Regarding the participants' occupational history, Table 3 shows that 62.4% had dual employment contracts and 77.9% worked 40 hours per week.The participants were stratified according to their working conditions: 78.2% of the sample had been using computers at work for 10 years or longer; 80.9% used computers for 5 or more hours a day at work; 98.3% reported they were unable to change postures at work (Table 3).

FACTORS ASSOCIATED WITH VISUAL FUNCTION -UNIVARIATE ANALYSIS
The mean NEI VFQ-25 score of personal characteristics was higher among single employees, those who underwent routine medical examinations, and those who did not wear corrective lenses (Table 1).
Table 2 shows that older employees showed lower NEI VFQ-25 scores.Tenure at both the current position and the institution exhibited statistically significant associations with visual function.Regarding psychosocial factors at work, visual function was inversely correlated with ERI dimensions; the NEI VFQ-25 score was lower when greater effort, greater overcommitment, and greater effort-reward imbalance were observed.
Table 3 describes the results of the univariate analysis of occupational history and working conditions.Employment status exhibited a significant association with visual function; the mean NEI VFQ-25 score was lower among workers with dual employment contracts than among the other employees.The mean NEI VFQ-25 score was lower among employees who reported using computers for 10 years or longer.Acoustic comfort where the computer was located at work was associated with visual function; the mean scores were lower when the acoustic conditions were optimal.

MULTIPLE LINEAR REGRESSION ANALYSIS (JOINT ANALYSIS OF FACTORS ASSOCIATED WITH VISUAL FUNCTION)
As shown in Table 5, the multiple linear regression analysis indicated that the factors independently associated with visual function were age and effort at work.The NEI VFQ-25 score exhibited a decrease of 0.218 per additional year of age and of 0.656 per additional point in the effort at work score.The adjusted coefficient of determination of the model (r²a) was 0.20.Analysis of residuals showed that errors adhered to the normal curve, therefore the model did not show bias.

DISCUSSION
Our results show a satisfactory quality of visual health among computer-based workers of this public hospital.The factors that remained independently associated with visual function were age and effort at work.
Most of the participants were women, were in the fifth decade of life, and wore corrective lenses.The mean visual function score was similar to those reported by studies with healthy people in Armenia 20 and people with visual impairment in the United States, 21 but lower than those in American studies with healthy people 21,22 and a German population-based study. 23The visual function domains with higher scores were color vision, dependency, and social functioning, which was similar to what was observed in a German study. 24General health was among the three domains with the lowest scores, just as with participants from Armenia 20 and Germany. 24ome individual, clinical, and work-related variables exhibited statistically significant associations with visual function in our univariate analysis.For example, the sex of participants is not a consensus in the scientific literature -it was associated with visual outcomes in some studies, 7,8 but did not pose influence in others. 21Marital status was not associated with visual function in American studies. 21Wearing corrective lenses was associated with visual function in Africa 11 and Asia. 8Screen time (in Spain 7 and Ethiopia 11 ) and time in certain occupations (in Sri Lanka 8 ) were associated with visual outcomes such as CVS or asthenopia.However, these variables were eliminated following multiple statistical modeling as a function of the greater effect of age and effort at work in our Brazilian group.The final model showed significance for age and effort, where older age and greater effort corresponded to poorer visual function.
This study showed that, for each additional year of age, there was a statistically significant decrease in mean global NEI VFQ-25 scores.These results are similar to those found in the American 21 and Armenian 20 publications.Among our participants, six out of ten had presbyopia.This demonstrated that the workers comprised an aging group in their 4th and 5th decades of life, which was compatible with the expected outcome. 25Presbyopia corresponds to the difficulty of clearly distinguishing nearby objects due to the inability to focus the eye to meet the visual demand close by.In our sample group, ocular pain was the most affected domain in visual function, which may be related to exposure to continuous computerbased work associated with presbyopia and other clinical visual disturbances.
Regarding psychosocial factors at work, only effort at work (from the effort-reward imbalance model) was associated with visual function.Effort-reward imbalance and overcommitment were not associated in the multiple model.Similarly to another study that analyzed visual function, ie CVS, the dimensions of the demand-control model were not associated with the ophthalmic outcome (asthenopia). 5One study that assessed other psychological factors and visual fatigue among bank employees who used computers at work found that social support, group conflict, poor selfesteem, job dissatisfaction, and skills under-utilization behaved as predictors of visual complaints. 26xperimental studies have shown effects of mental overload on visual function in men. 27In Norway, young women showed a deficit in visual function when exposed to stress, with effects such as a transient increase in trapezius muscle activity and a more forward leaning posture to try to increase productivity (in relation to reading speed) in people with normal vision. 28This impact may be greater in older people with a history of visual impairment, even if corrected.Screen brightness also had a negative influence on visual health, but at the public hospital where this study was conducted, the workstation illuminance assessment (based on institutional documents [PPRA]) showed results of 480 to 500 lux, which are in line with the Brazilian standards.
The following factors were determinants of greater effort at work in the studied population: interruptions at work, working after hours, increase of work demands in the past years, insufficient time for the actual workload, and responsibility at work (data not shown).We believe that the employees' responses might have been due to changes made at the institution during the period of data collection.Indeed, by that time (2014-2015) the hospital underwent the implementation and certification of several quality management systems, which might have affected several work processes and changed the hospital organizational charts and procedures.These changes might have further increased a pre-existent excessive workload.In addition, some older employees went into retirement but no new workers were hired in their place.This possibly increased the pressure on the remaining employees to maintain and increase productivity, while meeting deadlines.As a result, the participants' responses reflected this additional degree of effort at work.
A relevant characteristic within this scenario of greater effort at work was an ever-increasing use of computers.Most work requiring computers is associated with considerable mental and especially cognitive demands.Changes in work processes and an increasing use of computers result in greater need for visual efficiency and activation of the nervous system components that coordinate eye movement and accommodation. 16,27,28he relationships of personal characteristics of workers, psychosocial risk factors, and environmental factors at the workplace should be accurately evaluated for the purpose of health promotion and prevention of ocular symptoms. 1The results of this study thus showed that the effects of computer use on workers' visual function need to be accurately and regularly assessed and followed up.
Important preventive recommendations in CVS management are: a) regular work evaluations and corrections of environmental conditions 29 such as screen brightness and contrast adjustment; b) actions to control and reduce negative psychosocial conditions related to work effort 29 ; c) promoting eye health education among computer users on preventive strategies that encompass environmental factors and health promotion, including self-assessment 11,29 and visual rest, such as taking micro breaks, blinking frequency, etc.; d) and the inclusion of the eye exam in periodic examinations of workers. 10,29In this last recommendation, the occupational physician must be qualified to perform the measurement of visual acuity.The worker should be referred for a complete ophthalmological examination with an ophthalmologist when: a) presenting visual acuity equal to or less than 20/30 (Snellen's table) in at least one eye, with or without visual symptoms, or a difference in visual acuity between both eyes of two or more lines; and/or b) presenting strabismus.
In the 21st century, telework is a rapidly growing strategy for increasing productivity.It has become mandatory for the job market, as seen during the COVID-19 pandemic.Computer-based tasks performed at home can be a part of a company policy to improve work-life balance.However, working at home is not without health risks, as it may be performed under sub-optimal work conditions.Recommendations for workers, employers, and public authorities must address advantages and disadvantages of telework. 30Suggestions, as well as the support of companies, are welcome to the implementation of home office improvements.
The limitations of the present study include its cross-sectional design, which does not allow the establishment of causal inferences, and the lack of performance of diagnostic tests for dry eye or systemic disorders.Using self-reporting questionnaires could lead to bias, but most of them presented good internal consistency.In regard to its strengths, the present study contributes to the understanding of a subject that is still scarcely investigated; the high rate of participation allows inferring that the results have adequate internal validity, and the robust strategy for statistical analysis minimized bias.

CONCLUSIONS
The results of the present study showed that despite the aging of the study population, the quality of their visual health was good.They worked at a public hospital, following standards for the promotion and protection of visual health through annual eye examination during a periodic occupational medical consultation.The factors that remained independently associated with visual function were age and effort at work, being that the younger the age and the lower the effort at work, the better the visual function.The results point to the relevance of establishing periodical and preventive health actions, including eye health assessments.Even though its implementation is not a rule, the psychosocial environment at work should be evaluated, especially looking at effort at work, as a strategy for promotion of visual function and prevention of visual disturbances.

Table 2 .
Analysis of correlations between quantitative variables and visual function in computer-based administrative workers of a public hospital, São Paulo, 2015 (n = 303) * Pearson's correlation coefficient.SD = standard deviation.

Table 1 .
Continuedwas 15.8 ± 10.0 years, varying from 0.6 to 44.5, median 15.0 years, and the mean time working at the institution was 18.7 ± 10.2 years, varying from 0.6 to 44.5, median 20.1 years.

Table 4 .
Scores of visual function scale domains in computer-based administrative workers of a public hospital, São Paulo, 2015 SD = standard deviation.

Table 5 .
Univariate and multiple linear regression analyses of visual function in computer-based administrative workers of a public hospital and independent variables, São Paulo, 2015 (n = 303)