Work with video display terminals among office employees

Work with video display terminals among office employees: I Subjective symptoms and discomfort. Scand J Work Environ Health II (1985) 457-466. Subjective symptoms and discomfort were evaluated by means of a questionnaire and compared between approximately 400 video display terminal (VDT) operators and 150 selected referents. Previous and current illnesses, educational status, and smoking and drinking habits were also studied. The results showed the VDT operators to have more eye discomfort and possibly also more musculoskeletal discomfort in the shoulders, neck, and back than the referents. The VDTgroup also reported more skin disorders. In addition, women reported consistently more disorders than men, regardless of whether or not they were employed in VDT work. Women in general displayed greater morbidity than men. Eyediscomfort, musculoskeletal discomfort, headache, and skin disorders were found to be signifi cantlycorrelated in the material. The resultsalso indicatedthat total dailyworkhours and timespent looking at the VDT screen were related to the degree of discomfort. Even when the subjects were divided into subgroups with reference to the various enterprises, the types of work and the makes of VDT, the dif ferences obtained in the degree of discomfort appeared to be due to variations in the length of work hours.

relationship of these sym ptom s and discomforts to different physical exposure factors and findings in ophthalmologic and dermatologic examinations. This first report presents questionnaire findings on subjective symptoms and discomfort and on previous and current illnesses. Other reports deal with the occupational hygiene measurements and the va rious eye and skin examinations.

Subjects
Groups investigated. The exposed group comprised all office employees doing more than 5 h of VDT work daily at an insurance company (A), an airline and a post office (B), and personnel from three daily newspapers (C). To make up the reference group, persons whose worktasks were as similar as possible to those of the exposed subjects but which did not include VDT work were selected from each of these enterprises. The sample was taken with the assistance of a group comprising representatives of employers, trade unions, and occupational safety and health services. A preliminary estimate of the size of the material, following an initial contact with the enterprises, indicated some 450 exposed employees relatively evenly distributed between the three enterprise groups (A, B, C) . It was therefore decided that the reference group was to include about ISOpersons and that, preferably, these per-  1  156  29  168  75  114  46  438  150  2  151  29  154  75  109  45  414  149  3  149  27  144   72   102  42  395  141  4  142  21  140  68  97  37  379  126 a 1 = those reported by the enterprises before the investigation, 2 = those examined in one or more of the substudies, 3 = those replying to the questionnaire, and 4 = those completing all substudies.
b A = sublects from an insurance company. B = subjects from an airline and a post office, C = subjects from three daily newspapers. sons ought to be evenly distributed between the three enterprise groups and matched with the exposed subjects in terms of age and sex. The aim of this grouping was to create about 150"quartets," each including one representative of the three exposed enterprise groups and one representative of one of the three reference groups. In practice, this grouping and matching was not feasible for all the groups, but it worked fairly well for the exposed and reference groups as a whole (tables 1 and 2). The match between the exposed and reference subjects was relatively good in group A, could be termed fair in group B, and, as regards the balance between the sexes, unsatisfactory in group C. (See figure 4 in  the Results section and table 2.) The male predominance of the exposed subjects in group C was essentially due to the majority of newspaper VDT operators being printing workers who were retrained a couple of years ago to use VDT-based computer techniques instead of lead typesetting for newspaper printing. 458 The exposed groups were also classified according to the following worktasks, as recommended in a report from the National Academy of Science in the United States (11): (i) data entry, (ii) data acquisition, (iii) interactive communication (ie, a combination of data entry and acquisition), (iv) word processing, and (v) programming . Our investigation only included the first three of these five categories. The group was also divided according to VDT make and model. Six subgroups could be defined in this way.
Nonresponse. Table 1 shows the dropout of persons who did not attend the examinations. Only 1 of the 25 persons distinguishing between those reported by the enterprises and those examined in one or more of the substudies refused to be examined. Otherwise the nonresponse between the subjects in its various phases (seetable 1) was due to incorrect particulars concerning the nature of work, changing duties within the enterprise, vacations, and sickness absence.

Methods
The investigation as a whole involved the following stages: questionnaire concerning subjective symptoms and discomfort, special examinations of certain eye functions during the workday, determination of the ophthalmologic routine status, occupational hygiene measurements of relevant environmental factors, and dermatologic examination of persons reporting skin disorders in the questionnaire. The questionnaire was collected and the special eye examinations were conducted the same day at the workplace, by an orthoptist. The ophthalmologic examinations, performed at the offices of specialists, and the occupational hygiene measurements took place on another day . The dermatologic examinations were performed at the workplaces.
Questionnaire. The questionnaire included questions concerning the occurrence, frequency, and intensity of ocular discomforts (smarting, itching, gritty feeling, aches, sensitivity to light, redness, teariness, and dryness) and headache. The frequency and intensity of pain or discomfort from the musculoskeletal system (hands, forearms, elbows, upper arms, shoulders, neck, and back) were obtained, as well as the occurrence of skin rashes and disorders on the face, neck, hands, and arms. Questions were also included concerning previous and present illnesses or symptoms diagnosed by physicians and any current medication, together with questions concerning smoking and drinking habits, education, present and previous employment (eg, number of years and hours per week in VDT work). Special questions concerned the use of selfcopying (carbonless) paper and the extent to which work was regarded as "interesting" and to which the respondent took a "positive" or "negative" view of work or worktasks. The replies to the question on the subject's interest in his or her work were ranked as follows: "Yes, definitely" = 1, "Yes, to some extent" = 2, "No, hardly" = 3, "No, definitely not" = 4. The answers to the question on attitudes were ranked as follows: "Definitely positive" = 1, "Moderately positive" = 2, "Moderately negative" = 3, "Definitely negative" = 4. On the day of the special eye examinations, the subjects were asked to state the occurrence of eye discomfort and headache and also to specify the hours at which they did VDT work, did other work, or took breaks.
Discomfort index. Special discomfort indices were constructed for eye discomfort and for headache and musculoskeletal discomfort. Symptom scores were obtained for the eight eye symptoms and for headache by multiplying the frequency score (occasionally = 1, weekly = 2, daily = 3) by the intensity score (negligible = 1, slight = 2, pronounced = 3). Asymptomatic subjects scored O. The eye discomfort index was the sum of the eight various eye symptom scores. The symptom score for musculoskeletal complaints was based on intensity scores alone.
Duration of workhours. The number of years in VDT work and the number of hours of VDT work done during past years were obtained from the questionnaire. Total workhours and time spent in VDT work were noted by the employees themselves on the day of the special eye examination. On another.occasion, 132 randomly selected VDT operators were followed during a normal workday by means of a special gaze direction instrument. A small infrared transmitter was fitted on one side of the frames of spectacles (persons not normally wearing spectacles were given "hollow" glasses for the day of the experiment) in such a way that the time which the wearer spent looking at the screen was recorded on a special reciever on the top of the VDT. In this way it was possible to determine total workhours and the length of time spent looking at the VDT screen.
Statistical methods. The two-sample t-test, the chisquare test, and the Wilcoxon rank-sum test were used in the statistical evaluations of differences between the various groups. Normal probability plots were used to decide whether or not to use parametric tests. All tests were two-sided.
The use of other statistical methods was also attempted, eg, linear regression analysis. However these methods did not lead to any further conclusions based on the present material. The character and size of the material precluded attempts of multidimensional stratification.
In this first and also the second paper of the investigation (5), a large number of comparisons were made, some of them without an a priori hypothesis. For this reason, reliable conclusions should be limited to statistical tests with p < 0.001.

Eye discomfort
One or more of the eight eye discomfort symptoms were reported by 64 070 of the exposed employees as against 46 % of the referents (p < 0.001). The exposed group had a higher level of discomfort with reference to all symptoms, and the differences were significant for five of them (one with p < 0.05 and four with p < 0.01) (figure 1).
When the material was separated by enterprise (A, B, and C) differences in discomfort were found for enterprises A and B but not for C (figure 2). Enterprise C also differed from the others as to sex distribution (figure 3). The relative number of men in the exposed group was 55 % in enterprise C compared to 16 and 12 % in enterprise A and B, respectively.
The frequency of discomfort was also found to vary between the men and women, for both the exposed and % 100 figure 4). This was true of all the symptoms, the difference being statistically significant (p < 0.01) for four of them, and for eye discomfort in general (one or more of the eight symptoms) (p < 0.001). The sex-related differences in discomfort could have been the reason no differences in discomfort could be established between the exposed and reference subjects in enterprise C. Subjective discomfort and attitudes Various discomfort scores, separated with reference to VOT use and nonuse and to sex, are presented in figure 5. As has already been mentioned, the exposed subjects had significantly higher eye discomfort scores than the referents, and the women experienced more discomfort than the men (figure 5A). The musculoskeletal discomfort scores were also found to be higher for the women, but they did not differ between the exposed subjects and the referents. However, there was a tendency towards higher discomfort scores for proximal locations (shoulder, neck and back) among the exposed group when it was compared to the reference group ( figure 6). The women reported more headache and a higher frequency of skin disorders than the men (figures 5C and 0). Skin disorders were more widely reported by the exposed employees than the referents , especially where women were concerned.
Significant correlation coefficients (0.38-0.45, with p < 0.001) were found between the scores of eye discomfort, headache, and musculoskeletal discomfort. The frequency of skin disorders was also associated (p < 0.001) with the other discomfort scores.
The answers obtained for the " interest" and " attitude " questions, ie, questions as to whether the employees found their work "interesting" and to what extent they took a " positive" or " negative" attitude towards it, did not reveal any differences between the exposed and reference employees or between the sexes. There was, however, a strong association between the scores of the attitude and interest questions (indicating that the two questions probably measured the same expression of "attitude/interest").
Age did not correlate with the subjective discomforts, but interest and a positive attitude towards work increased with advancing age, to reach a maximum at age 40-50 years.

General health status
There was no significant difference between the VOT group and the referents with regard to the occurrence of current and previous diseases or to prescribed medication. There was a general increase in morbidity among the women, a finding reported also in other investigations . The use of oral contraceptives was not related to high scores.  Figure 4. Frequency of eye discomfort and the various eye symptoms in the exposed (diagonally str iped columns) and reference (unstriped columns) groups distributed according to sex (0 =men, '" =women). The sig· niflcance ind ications refer to compar isons between the women in the exposed and reference groups . (.. p < 0.01, .. , p < 0.001)

Smoking and drinking habits
As usual, a relation was found between alcohol consumption and smoking, and also alcohol consumption was higher among the men. There was no difference in smoking or drinking habits between the VDT group and the reference group, and there was no relation between smoking or drinking and the subjective complaints studied.

Educational status
The subjects were divided into the following six educational groups: elementaryschool, junior high school, senior high school, compulsory comprehensive school, vocational school, and university or college. These groups did not differ in terms of discomfort frequencies or scores. The results revealed one difference connected with educational status, ie, the higher the level of education, the higher the rate of alcohol consumption, but this difference was not related to age, sex, or exposure.
Duration of workhours for the exposed groups Figure 7 shows the duration of workhours for all the VDT operators and for the female VDT operators in relation to sex and occurrence of subjective symptoms and discomforts . (See the Methods section.) As regards duration (figure 7A), the subject s reporting eye discomfort were found to have worked somewhat longer than those without eye discomfort (not significant), but otherwise there were no observable differences worth mentioning. More substantial differences were noted with regard to the number of hours per week devoted to VDT work (figure 7B). Women worked slightly fewer hours per week than men (not significant), and, where eye discomfort and musculoskeletalcomplaints were concerned, those with complaints worked longer hours at the VDT (p < , and headache (C) and the frequency of skin disorders (D) for the exposed (diagonally striped columns) and reference (unstrlped columns) groups [whole group (All) and men (0) and women ("'), respectively) and forthe men and women in the two groups combined (horizontally striped columns).  Figure 6. Location of musculoskeletal complaints. Score differences between the entire exposed (VDT) and reference groups and between the men and the women of the two groups (* p < 0.05, * * * p < 0.001) The measurements undertaken with the special gaze direction instrument (figure 7E) revealed several significant differences of interest. Subjects with eye discomfort and musculoskeletal discomfort spent more time looking at the screen and had longer workhours on that particular day. Those reporting headache also had longer workhours than the others (figure 7F).
Figures 7E and F are based on registered work duration, in contrast to the (subjective)reported work duration in figures 7A-D. Thus, eg, the data on workhours given in figure 7F are considered more reliable than the data on workhours in figure 7D. The results in figures 7E and F must still, however, be interpreted with some caution, because they concern a subgroup with possibly a shorter worktime (figure 7F: 6 h, registered) than for the whole exposed group (figure 7D: slightly above 7 h, reported).
An overall assessment of the results suggests a doseeffect relationship, especially as regards the measurements obtained with the gaze direction instrument.

462
Group classification of the material The various enterprises. The enterprise classification is a modified version of the group classification described in the Subjects section and used in describing the findings on eye discomfort. Thus group B has been divided into B 1 (= the post office) and B 2 (= the airline). Enterprise A is the insurance company, and C represents the three newspapers. As can be seen from figure 8 on subjective symptoms and discomfort, there were certain differences between these enterprises. Almost invariably, enterprise A had the highest and enterprise C the lowest discomfort scores, in comparisons both between enterprises generally and between the exposed women employed by them. The differences between enterprises A and C were statistically significant. Figure 9 shows the results of the gaze direction measurements undertaken at the various enterprises. It was found that both the length of time for which the gaze was directed at the screen and the total number of workhours were significantly shorter in the newspaper enterprises (C) than in the other enterprises. What is more, enterprise A obtained relatively high scores for total number of workhours.
In summary, the various enterprises obtained different discomfort scores, but these differences may be related to the various durations of workhours.
Different types of work. The types of VDT work included in our investigation were confined to data entry (type 1), data acquisition (type 2), and the combination of data entry and acquisition, ie, interactive communication (type 3). The newspaper enterprises (C) had data entry only and the airline (B 2 ) data acquisition only. Therefore, it was relevant to compare data acquisition (type 2) with interactive communication (type 3) since both types of work occurred to the same extent in the insurance company (A) and at the post office (B 1 ) . On account of the sex differences in subjective symptoms and discomfort, we have chosen to confine our comparisons to the largest group (women) having these two different types of job. The results can be seen from figure 10. The discomfort scores were consistently higher for job type 3 (musculoskeletal complaints p < 0.05, otherwise not significant). Job types 2 and 3 also differed with regard to registered number of workhours. Employees doing a combination of data entry and acquisition work had a much longer workday than those with acquisition work only (p < 0.01). As might be expected, however, acquisition involved more time looking at the screen (not significant). Thus one cannot discount the possibility of the differences in subjective disorders between types of work obtained in our study being referable to differences in the duration of workhours.
Different makes and models of video display terminals. Altogether the operators taking part in our investigation used 11 different makes or models of VDT. . Duration of work with a video disp lay term inal (VDl) and different measures of VDT work and total worktime for t he exposed SUbjects report ing (+) and not report ing (-) discomforts. The two lett-hand columns for each discomfort are for both the men (,I) and the women ( 9) , and the two rlqht-nand columns are for women only. E and F refer to a subgroup only, 132 randomly selected VDT operators . In A through D reported work durat ion is used, while E and F use registered work dura- [The possibility that the degree of discomfort is related to display and lighting parameters is examined in another report from this study (5).)

Self-copying (carbon less) paper
Since the use of self-copying paper has been reported as being related to skin and eye discomfort (6), questions were included concerning the frequenc y with which self-copyingpaper was handled and the approximate number of sheets with which the employees came into contact every day. There was no diffe rence between the sexes, but the exposed subjects had more contact with paper of this kind than the referents. The subjective disorders reported were not, however, related to the use of self-copying paper.
For reasons already stated, only the results obtained for the women at the insurance company and the post office were used when the different screens were compar ed. Thus the comparison included six different types of screen, marked K, L, M, N, 0, and P in figures 11 and 12. This comparison revealed certain differences. Screen K, for example, was definitely more "discomfort prone" than screen P (p < 0.001), but there were other differences as well (eg, screens K and L were associated with more discomfort than screens o and P). In the determination of the duration of workhours (figure 12) both shorter total workhours and shorter duration of gaze at the screen were recorded for screen P in comparison with the material as a whole (both p < 0.001). Thus, again, the differences between certain screens as regards the degree

Discussion
The results of the present study showed the VDT operators to suffer from more eye discomfort and possibly also more musculoskeletal discomfort in their shoulders, neck, and back than the referents. The VDT group also reported more skin disorders. In addition, women reported consistently more complaints than men, regardless of whether they were doing VDTwork or not , and women displayed greater morbidity than men. These findings agree closelywith results from a number of other studies , eg, concerning the frequency of eye discomfort (2,10,12,15). Starr (16), however, did not find the frequency of eye complaints to be significantly higher among exposed subjects than among  referents . An elevated frequency of musculoskeletal complaints for the shoulders , neck, and back has been reported in several studies (2,4,16), whereas no difference was reported by Gould & Grischkowsky (3). As regards skin disorders, a significantly higher frequency has been reported by Murray et al (8), though only at one of three workplaces investigated , while Frank (2) did not report any differences between exposed and reference subjects . A higher frequency of discomfort among women than men as regards musculoskeletal and eye complaints has been reported by Wallin et al (17) and is implied by the results of another investigation (8), Both the total number of workhours per workday and the time spent at the vnT screen seemed to be related to subjective disorders. Similar results have been reported by Wallin et al (17) concerning eye dis-comfort, musculoskeletal discomfort, and headache; by Mellner & Moberg (7) concerning eye and musculoskeletal complaints; and by Rey & Meyer (12) concerning eye discomfort. Other studies (13,14,15) have failed however to reveal such relations. When the subjects in our study were divided into groups according to enterprise, type of VDT work, and make of VDT, it seemed that the differences thus obtained were also referable to variations in the duration of workhours. There were significant correlations between the scores of the different types of discomfort studied. Thus, for example, persons with eye discomfort tended also to report headache and/or musculoskeletal complaints . Similar results have been obtained, for example, by Smith et al (l4).and, as regards the correlation between eye discomfort and musculoskeletal complaints, by Wallin et al (17). There are several possible explanations for these relations, eg, the aforementioned correlations with the duration of workhours. It is also possible that suboptimal reading conditions on the screen could cause eye discomfort and musculoskeletal discomforts due to attempts to compensate for poor image quality. There are indications of a relationship of this kind in an investigation by Smith et al (14) . Special mechanisms of a psychological nature may also contribute to differences in the discomfort reported by exposed subjects (ie, VDT operators) and reference subjects.
Little is known today about the causes of the elevated discomfort-proneness and the morbidity of women. Are women constitutionally more sensitive than men, and is this why they react more vigorously to environmental hazards? Do women have more strenuous work, even though there is supposed to be equality between the sexes? Is the total living situation of women today more conducive to discomfort? Many women wor k part-time (see figure 7, duration of workhours) so as to adjust to the needs of the home and family , and, at the same time, have to measure up to high standards at work.
In addition to these psychosocial uncertainties, we do not know enough about occupational hygiene aspects and medical factors with a bearing on the occurrence of discomfort in connection with VDT work and differences in this respect between men and women. Our continuing studies will be concerned with these problems.