Data were analyzed by using JMP statistical discovery software version 9.0 (SAS Institute). Independent-sample t tests were used to determine whether the parametric differences between those with and without an intention to seek a medical opinion for suspected dry eye disease, as well as between those with >5 mm (normal) and ≤5 mm (abnormal) tear production, were significant. P<.05 was considered significant.

Figure 2 depicts the flow diagram of the study. Of the 1689 visitors to the survey website, 980 (58.0%) volunteers, including 440 men (44.9%), 539 women (55.1%), and one person of unidentified gender, completed the Web-based self-screening questionnaire. Their demographic data are summarized in Table 1. Ocular fatigue (61.4%), dry eye sensation (40.7%), and sensitivity to bright light (31.3%) were the major symptoms.

Of the 979 gender-identified respondents, 355 participants had subjective dry eye symptomatology. The symptoms were significantly more common among women than among men (226 female participants, 63.7%, P=.0001, Fisher’s exact test). Ninety-two participants (25.9%) did not intend to seek medical attention for the symptoms (Table 2). Men were significantly less inclined to visit a clinic for definitive diagnosis (P=.006, Fisher’s exact test). Further, significantly fewer participants in their 20s intended to see a doctor (P<.001, chi-square test).

Of the 355 participants with subjective dry eye symptomatology, 99 (27.9%) participants, including 24 men (24%) and 75 women (76%), performed the Schirmer test and answered the paper-based lifestyle questionnaire. Most of these participants were healthy: only 19.2% had a systemic disease. Contact lenses were used by 40.4%, and 10.1% had a smoking history (Table 3).

Thirty-two participants had tear production of ≤5 mm (Table 4). No significant differences were noted in the duration of VDT use (P=.82, Wilcoxon rank-sum test), use of contact lenses (P=.67, Fisher’s exact test), duration of contact lens use (P=.98, t test), and smoking history (P=.49, Fisher’s exact test) between the >5 mm and ≤5 mm tear production groups. A significant difference was found only in exercise habits: a lower level of physical activity was significantly associated with a lower value in the Schirmer test (P=.02, Wilcoxon rank-sum test).

In this study, a Web-based self-screening questionnaire was used to draw attention to dry eye disease among general Internet users and identify those at higher risk. Of the 1689 visitors to the survey website, 58.0% (980/1689) completed the first phase of the study, 36.2% (355/980) had five or more dry eye symptoms, 27.9% (99/355) completed the second phase of the study, and 32.3% (32/99) had tear production of ≤5 mm. This self-screening tool therefore identified a few individuals with probable dry eye disease. We speculate that a high number of general Internet users have latent dry eye symptomatology.

In the first phase of this study, 63.7% (226/355) of the female participants reported five or more dry eye symptoms, supporting previous conclusions that dry eye disease is more common in women [1-3,5-7,36].

Overall, the participants in phase 2 were generally healthy and health conscious, with a low percentage of smokers and good reported dietary habits (data not shown). A lower level of physical activity was significantly associated with a lower value in the Schirmer test. Notably, 25.9% of the 355 participants who had five or more dry eye symptoms, especially men and people in their 20s, did not intend to see a doctor. The self-screening questionnaire may have slightly improved their health consciousness but was not sufficient to persuade them to seek medical care. Therefore, only some VDT users may be motivated enough to visit a doctor despite having symptoms. Ophthalmologists should promote awareness of dry eye disease more proactively, and the public should be educated about the condition for early detection and intervention.

First, because the participants were recruited from general Internet users in a limited target area, a considerable part of the high-risk population was excluded. The recruitment method would have also introduced selection bias. Second, the participants with fewer than five dry eye symptoms were excluded from the physical phase of the study. Therefore, parametric differences between these participants and the participants with subjective dry eye symptomatology were not analyzed. Third, the second-phase results were based on the Schirmer test, which measures tear production; therefore, the evaporative aspects of dry eye disease were overlooked. Finally, the Web-based self-screening questionnaire has not yet been validated; the validation process is now underway under the supervision of the Japan Dry Eye Research Society.

Future research should include both biologic measures (break-up time and fluorescein staining) and the Schirmer test to diagnose dry eye disease. In addition, a normal control group and general Internet users in a wider geographic area should be included. Moreover, the self-screening system needs to be improved. As the next research step, the Web-based self-screening questionnaire should be optimized and validated to ensure its effectiveness in increasing awareness of dry eye disease especially among VDT users.

The proposed Web-based self-screening questionnaire seems to be a promising tool to raise awareness of dry eye disease among general Internet users and identify those with a higher risk of developing the condition. The factors that encourage people with probable dry eye symptomatology to seek medical help are yet to be identified. Further research is required to provide sufficient information on the disease and diagnostic tests, determine symptom severity via a validated self-screening tool, and introduce improved diagnostic measures.