Prevalence and 11-year incidence of common eye diseases and their relation to health-related quality of life, mental health, and visual impairment

We utilized two nationwide health examination surveys performed in Finland. They were carried out by the Finnish Institute for Health and Welfare, the first one in 2000–2001 and a follow-up in 2011 [17, 18]. In both surveys, the information on eye diseases and co-morbidities was collected in face-to-face interviews, whereas the assessment of HRQoL and mental health was based on self-administered questionnaires. The Health 2000 Survey analyzed a sample of 9922 adults aged 18 years or over living in mainland Finland. The sample was selected by a stratified two-stage cluster sampling design. The Health 2011 Survey analyzed a sample of all living members of the Health 2000 sample who had not refused to be contacted, aged 29 years or over, and a new sample of 1994 young adults aged 18 to 28 years. For this study, we only included participants aged 30 years or over in both cross-sectional and longitudinal samples. Both surveys provided a probability-clustered sampling and weighting scheme that estimates health statistics that are representative of Finnish adult population aged 30 years or over at the time of sampling [19, 20]. In addition, the scheme accounts for the oversampling of people aged 80 years or over in 2000 by doubling the sample fraction. The unweighted participation rate was 93% in the Health 2000 Survey while in the follow-up it was 73%. Separate weights were applied for the surveys to produce results representing the Finnish population at each time points [21].

Both surveys included an interview with the following questions on eye diseases: “Has a doctor diagnosed one of the following diseases: cataract, glaucoma, retinal degeneration, or other visual defect or eye trauma?” Only the individuals who had replied “yes” or “no” to at least one of these questions were chosen for the further analyses, classified as “eye disease status known”. Individuals who had only answered “no” to this set of questions were considered to not have eye diseases. There was also self-reported information on previously performed cataract operations. Only unoperated cataract patients were included in the evaluation of HRQoL, mental health, and VA, as cataract surgeries improve VA and have been demonstrated to improve QoL as well [22].

HRQoL scores were evaluated using generic preference-based 3-level version of EuroQol-5 Dimension (EQ-5D-3L, later referred as EQ-5D) and 15D questionnaires that assess physical, psychological, and social functioning and well-being [23, 24]. 15D is a self-administrated measure of HRQoL comprising one question for each of the 15 dimensions—mobility, vision, hearing, breathing, sleeping, eating, speech, excretion, usual activities, mental function, discomfort and symptoms, depression, distress, vitality, and sexual activity. Each question contains five answer options on a scale of 1 (no difficulties) to 5 (extreme difficulties). A single index score is obtained by weighting the obtained scores with population-based preference weights based on an application of the multi-attribute utility theory [25]. In this study, the 15D was weighted using Finnish preference weights with a scale of 0 (representing HRQoL equal to being dead) to 1 (representing the best possible HRQoL). Mean change/difference of ≥ 0.015 was considered to be clinically meaningful [26].

EQ-5D contains one question for each of the five dimensions: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression. Each question contains three answer options on a scale of 1 (no difficulties) to 3 (extreme difficulties), and they can be converted into EQ-5D index scores on a scale identical to the 15D index score. In this study, EQ-5D was weighted using UK time trade-off weights on a scale between − 0.59 (representing HRQoL equal to being dead) and 1 (representing the best possible HRQoL) to improve comparability with other populations [27]. Mean change/difference of ≥ 0.07 was considered to be clinically meaningful [28].

The state of mental health was assessed using two self-report questionnaires, Beck Depression Inventory (BDI) and General Health Questionnaire-12 (GHQ-12). In the first survey, BDI-21 comprising 21 questions was used to evaluate depression, whereas in the follow-up survey a shorter version, BDI-13, containing 13 questions, was used [29, 30]. A total score was calculated for both questionnaires with a scale of 0–63 for BDI-21 and 0–39 for BDI-13, where higher points indicate major depression. Total scores of ≥ 10 for BDI-21 and ≥ 5 for BDI-13 were used as cut-off points to categorize an individual as having depression [31].

GHQ-12 is a questionnaire comprising 12 questions that evaluate different dimensions of psychological distress, including depression, anxiety, social interaction, and confidence [32, 33]. The answers were dichotomized according to whether difficulties were presented or not (0 = no, 1 = yes). A total score with a scale of 0–12 was calculated using the dichotomized points, with 12 representing the highest psychological distress. A total score of > 3 was considered as indicative of psychological distress [17, 18].

In both surveys, the distance VA was measured by a study nurse binocularly at 4 m with current visual correction. Illumination was set to ≥ 350 lx on the modified logMAR letter chart published by Precision Vision [19, 20, 34]. All VA values are presented as Snellen decimal equivalents. Low VA values outside the modified logMAR letter chart that could not be determined were reported as 0.01. The classified VA values were following: VA ≥ 1.0 (good vision), VA 0.63–0.8 (adequate vision), VA 0.32–0.5 (weak vision), VA 0.125–0.25 (impaired vision), and VA < 0.1 (severe vision loss or blindness) [8]. Habitual binocular distance VA ≤ 0.25 was considered as impaired vision. We found the binocular evaluation of VA important because the relation of vision and HRQoL was studied.

Common diseases assessed in the interview (data available from 7371 to 7385 and 5714 to 5720 participants in 2000 and 2011) were accounted for their potential impact on the HRQoL and mental health. The diseases were classified into major co-morbidity groups according to Taipale and colleagues [8]: heart diseases (myocardial infarction, angina pectoris, heart failure, arrhythmias, and “other heart disorders”); respiratory diseases (asthma, chronic obstructive pulmonary disease, chronic bronchitis, and “other pulmonary disease”); vascular diseases (stroke and varicose veins in lower limbs); musculoskeletal conditions (rheumatoid arthritis, osteoarthrosis, fractures, and osteoporosis); psychiatric conditions (psychotic disorders, depression, anxiety, psychoactive substance abuse, and “other psychiatric disease”). Furthermore, hypertension, diabetes, Parkinson’s disease, and unspecified cancer were included as separate groups.

Co-morbidity status was determined according to Taipale and colleagues [8] so that individuals were considered to have co-morbidity if they reported any of the conditions included in the co-morbidity group. When analyzing new incident diagnoses during the follow-up period, each condition was scrutinized in 2000 baseline and in 2011 follow-up. If the subject reported at least one new condition included in the co-morbidity group in 2011, they were classified as having incident co-morbidity, regardless of the presence of other conditions of that specific co-morbidity group at baseline.

The data were analyzed using R software version 3.5.1 [35], and it included both cross-sectional survey samples for all cross-sectional and longitudinal analyses. The sampling design, the oversampling of individuals aged 80 years or over, and the loss to follow-up were accounted for by using Survey package 3.37 for R [36] and weighting scheme calculated by the Finnish Institute for Health and Welfare.

For the prevalence and incidence analyses, population totals and ratios were estimated using functions svytotal and svyratio included in the Survey package. Individuals with missing data in analyzed variables were excluded. As the distribution of the continuous variable data was skewed, Mann–Whitney U test was used for the between-group comparisons. The impact of age, gender, eye diseases, impaired distance VA, and co-morbidities on HRQoL and mental health were estimated through linear regression, and standardized regression coefficients were calculated using lm.beta package 1.5-1 for R [37]. Multicollinearity in regression analyses was measured through variance inflation factors using car package 2.1-5 for R [38, 39]. Odds ratios (ORs) with 95% confidence intervals were calculated using logistic regression analysis, adjusted for age, gender, and co-morbidities. For all analyses, a two-tailed p value of < 0.05 was considered to be statistically significant.

Figure 1 presents the number of the individuals with self-reported eye disease in the two surveys that were included in the analyses. In total, 8028 individuals aged 30 years or over had been invited in the 2000 survey and 8006 in the 2011 survey. Eye disease status data was obtained from 7379 and 5710 individuals, of whom 4620 took part in both time points and were included in the 11-year follow up. Table 1 shows the number, mean age, and the gender distribution of the survey samples, individuals with/without eye diseases, and individuals with impaired/good distance VA in both surveys and in the 2011 follow-up. It also includes the number of individuals with eye disease status known who had available data on HRQoL, mental health, and distance VA. The data in all analyses were compared between individuals with eye diseases and those with no eye diseases, and individuals with impaired distance VA and those with good distance VA.

The estimated prevalence and incidence of glaucoma, cataract, RD, and visual impairment in the Finnish adult population is shown in Table 2. The prevalence of cataract and RD increased between the time points, whereas glaucoma and visual impairment decreased. The prevalence and incidence of all eye diseases and visual impairment increased with age, and they appeared to be more common in women, particularly in age group 75 + years (Fig. 2).

EQ-5D and 15D index scores were lower (p < 0.0001) in individuals with eye disease or visual impairment compared to those with no eye disease or with good distance VA in both time points, indicating lower HRQoL in eye disease patients and visually impaired (Fig. 3). This difference was also clinically meaningful in both time points. However, the mean values of these scores were higher in 2011 than in 2000 in all eye disease groups (p < 0.01) and individuals with visual impairment (p < 0.05). Individuals with no eye diseases and those with good distance VA had better HRQoL in 2011 only according to 15D score (p = 0.0001 and p = 0.036, respectively). Moreover, the improvement of mean HRQoL seen in all eye disease groups and those with visual impairment was clinically meaningful between the time points, except for glaucoma with EQ-5D.

All eye disease groups and visually impaired had worse (p < 0.0001) BDI-21 total scores compared to individuals with no eye diseases or those with good distance VA (Fig. 3). Similar difference was found for BDI-13 total scores in all eye disease groups (p < 0.0001) and visually impaired (p = 0.002). Because the scales of the BDI questionnaires between the time points were not comparable, the mean change between time points was not evaluated. The effect of various eye diseases on GHQ-12 varied: it was most severe in RD (p < 0.0001 in both time points) and least severe in glaucoma (p = 0.037, only in 2011). Only individuals with RD (p = 0.004) and those with no eye disease or with good distance VA (p < 0.0001) showed better GHQ-12 scores in 2011 than in 2000. All eye disease groups had worse (p < 0.0001) distance VA compared to those with no eye diseases in both time points. All groups showed better mean distance VA in 2011 than in 2000. Overall, RD was associated with lowest scores in all these parameters compared to other eye diseases in 2000, but in 2011 this difference was significant only in EQ-5D (p = 0.010) and distance VA (p < 0.0001). However, visual impairment showed the highest impact on these parameters compared to all eye diseases in both time points, excluding BDI-21 and BDI-13, in which no significant difference was found between eye diseases and visual impairment.

The individual HRQoL dimensions were evaluated using ORs by comparing those with difficulties to those without difficulties. In addition, mental health was assessed by using the cut-off points for GHQ-12 and BDI total scores. For HRQoL, the most affected dimensions in individuals with eye disease and those with visual impairment were usual activities and mobility in EQ-5D, and vision, usual activities, and vitality in 15D (Table 3). There was variation in how the dimensions were affected between the eye diseases: among those with RD the majority of the individual dimensions were affected in both EQ-5D and 15D, whereas among those with unoperated cataract none of the EQ-5D dimensions differed from those with no eye diseases, even though the index score was significantly lower. Pain/discomfort and anxiety/depression in EQ-5D was only affected in those with RD in 2000, and anxiety/depression in EQ-5D was affected in visually impaired in both time points. The vision dimension in 15D was most affected in all eye diseases and visually impaired. Individuals with visual impairment showed high odds for having difficulties, including eight-fold increase in usual activities (EQ-5D) and sixfold increase in self-care (EQ-5D) in 2000 compared to individuals with good distance VA. Self-care (EQ-5D) was affected in individuals with visual impairment but not in those with eye disease. For mental health, depression in 2000 (BDI-21 ≥ 10) was more prevalent among individuals with RD and unoperated cataract compared to those with no eye diseases. In 2011, BDI-13 was not significantly affected in any of the groups. Psychological distress (GHQ-12 > 3) was found more prevalent among individuals with RD in 2000 and visual impairment in both time points.

The effect of the awareness of the eye diseases on the HRQoL and mental health was evaluated using linear regression analyses, including age, gender, and co-morbidities (Tables 4 and 5). After these corrections, the impact of impaired distance VA on HRQoL was more significant than any of the eye diseases. Only RD showed significant impact on 15D, GHQ-12, and BDI-21 of all eye diseases in 2000, whereas in 2011 only unoperated cataract showed significant impact on 15D of all eye diseases. In addition to visual impairment, psychiatric disorder and Parkinson’s disease had high impact on HRQoL. However, the overall effect and/or association of all these diseases and visual impairment on HRQoL and mental health were lower in 2011 than in 2000. No significant change was observed in the outcome when only statistically significant (p < 0.05) factors were included as explanatory variables in stepwise-insertion analysis. Multicollinearity ranged from 1.007 to 1.520 denoting no or very little multicollinearity.

Furthermore, the longitudinal setting was utilized when observing the change in the HRQoL, GHQ-12 scores, and distance VA in individuals who had same eye status in both time points (Fig. 4). Individuals with visual impairment in both time points were not included as their number was low (n = 8). All groups, including those with no eye diseases and with good distance VA, showed a small decline in the HRQoL values, with clinically meaningful decline in 15D values in all eye disease groups. All eye disease groups showed worsening in the GHQ-12 total score and all groups showed decrease in the distance VA. The impact of aging was visualized (Fig. 5), which shows that the decline in HRQoL and distance VA, and worsening in GHQ-12 is associated with aging.

Figures 6 and 7 show the overall shape of association between HRQoL, mental health, and distance VA in both time points. The decrease in the HRQoL and the worsening in mental health are associated with decreasing distance VA in all groups, including those with no known eye diseases.