A Population-Based Cohort Study of the Association between Visual Loss and Risk of Suicide and Mental Illness in Taiwan

The psychosocial and health consequences of ocular conditions that cause visual impairment (VI) are extensive and include impaired daily activities, social isolation, cognitive impairment, impaired functional status and functional decline, increased reliance on others, increased risk of motor vehicle accidents, falls and fractures, poor self-rated health, and depression. We aimed to determine whether VI increases the likelihood of a poor prognosis, including mental illness, suicide, and mortality over time. In this large, location, population-based, nested, cohort study, we used data from 2000 to 2015 in the Taiwan National Health Insurance Research Database (NHIRD), which includes diagnoses of all the patients with VI. Baseline features, comorbidities, and prognostic variables were evaluated using a 1:4-matched cohort analysis. Furthermore, comparisons were performed using Cox regression and Bonferroni-correction (for multiple comparisons) to study the association between VI and poor prognosis (mental illness, suicide). The study outcome was the cumulative incidence of poor prognosis among the visually impaired and controls. A two-tailed Bonferroni-corrected p < 0.001 was considered statistically significant. Among the 1,949,101 patients enlisted in the NHIRD, 271 had been diagnosed with VI. Risk factors for poor prognosis and the crude hazard ratio was 3.004 (95% confidence interval 2.135–4.121, p < 0.001). Participants with VI had an increased risk of poor prognosis according to the sensitivity analysis, with a poor prognosis within the first year and first five years. VI was associated with suicide and mental health risks. This study revealed that patients with VI have a nearly 3-fold higher risk of psychiatric disorders, including anxiety, depression, bipolar, and sleep disorders, than the general population. Early detection through comprehensive examinations based on increased awareness in the clinical context may help maintain visual function and avoid additional complications.

disposition, hospitalization and discharge dates, length of stay and medical costs; Registry for contracted medical facilities (HOSB)the variables include hospital location and hospital level.
This study was conducted in accordance with the Declaration of Helsinki. The institutional review board of Tri-Service General Hospital at the National Defense Medical Center in Taipei, Taiwan, approved this study, and the requirement for individual consent was waived because all identifying data were encrypted (TSGHIRB No. E202216004). The NHIRD is a freely accessible database that contains de-identified patient information to protect patient anonymity.

Study Design and Participants
This study is a retrospective cohort study that uses secondary database analysis. According to the International Classification of Diseases (ICD) and Related Health Problems, the World Health Organization divides VI into five categories. The first and second categories are moderate or severe VI with excellent visual acuity equal to or better than 3/60 (0.05). The third category is blindness with visual acuity between 3/60 (0.05) to 1/60 (0.02); the fourth category is blindness with visual acuity between 1/60 (0.02) to light perception; and the fifth category is blindness with no light perception. This study included a cohort of patients from the LHID database who were newly diagnosed with VI (ICD-9-CM 369.3 and 369.4). We excluded patients diagnosed with VI before 2000, poor prognosis before VI, unknown sex, and incomplete tracking data. The inclusion and exclusion criteria are shown in Figure 1. Moreover, the date of the diagnosis of VI was used as the index date. Participants in the control group were selected from the LHID 2005 cohort. The study and control cohorts were matched 1:4 according to sex, age, and the index date.
All participants were followed up from the index date until the first diagnosis of VI, poor prognosis, death, withdrawal from the NHI program, or 31 December 2015. The covariates included sex, age group, geographical area of residence (north, center, south, and east of Taiwan), urbanization level of residence (levels [1][2][3][4], and monthly income (in New Taiwan Dollars: <18,000, 18,000-34,999, and ≥35,000). The urbanization level of residence was defined according to the population and various indicators of development. Level 1 was defined as a population of >1,250,000 with a specific designation of political, economic, cultural, and metropolitan development. Level 2 was defined as a population between 500,000 and 1,249,999, with an important role in politics, the economy, and culture. Finally, urbanization levels 3 and 4 were defined as populations between 149,999 and 499,999 and <149,999, respectively.

Statistical Analyses
To investigate the association between VI and risk of suicide and mental illness, we conducted the following statistical analyses to compare the clinical characteristics of the participants in the case and control groups. The clinical characteristics of the participants are expressed numerically. We compared the distribution of categorical characteristics and baseline comorbidities between the case and control groups using Fisher's exact test and the chi-squared test. Continuous variables are presented as means and standard deviations and compared using the t-test.
As the primary goal of this study was to determine whether the clinical characteristics of patients are associated with poor prognosis, the Cox regression analyses were used to determine the risk of poor prognosis; the results are presented as hazard ratios (HRs) with the associated 95% CIs.
Associations between time-to-event outcomes and clinical characteristics were examined using the Kaplan-Meier method and multivariate Cox regression analysis with stepwise selection; the results are presented as adjusted HRs with the corresponding 95% CIs. The poor prognosis incidence (per 10 5 person-years) was calculated based on sex, age, and comorbidities for each cohort. Adjustments were made for age, sex, and concomitant comorbidities for inclusion in the multivariate model.
Bonferroni-correction for multiple comparisons was applied. A two-tailed Bonferronicorrected p < 0.001 was considered statistically significant. All statistical analyses were performed using IBM SPSS Statistics for Windows version 22.0 (released 2013, IBM Corp., Armonk, NY, USA).

Results
Among the 1,949,101 patients in the LHID 2005 from the NHIRD, 539 had been diagnosed with VI. In total, 271 patients were assigned to the study cohort, and 1084 age-, sex-, and comorbidity-matched patients were assigned to the comparison cohort ( Figure 1). The baseline data of the patients and control groups are shown in Table 1. The average age of the VI cohort was 39.01 ± 16.83 years, and the proportion of male patients was 52.03%. Among the study population, the majority of patients were aged 19 years (36.53%); 27.31% of patients were 65 years and older, 20.66% were 20-44 years, and 15.50% were 45−64 years old. Our findings show no significant differences in sex, age, thyrotoxicosis, septicemia, tumor, or season between the groups with and without VI after matching. Furthermore, we analyzed individual outcomes, such as dementia, eating disorders, SRD, psychotic disorders, autism, and other mental disorders ( Table 2). In general, VI was associated with an increased risk of individual outcomes, including dementia, eating disorders, SRD, and other mental disorders, even after excluding individuals with poor prognoses within the first year and first five years. Moreover, VI was associated with an increased risk of anxiety, depression, bipolar disorder, and sleep disorders, as well as individual types of all-cause mortality. Figure 2 shows the Kaplan-Meier survival curve of patients with poor prognosis stratified by visual loss using the log-rank test; patients with visual loss had a significantly higher cumulative risk of developing a poor prognosis 16 years after the index date (logrank test, p < 0.001).     Table 3 shows the Cox regression analysis of the factors associated with the risk of poor prognosis. The crude HR was 3.004 (95% CI: 2.135-4.121, p < 0.001). After adjusting for sex, age group, geographical area of residence, urbanization level of residence area, and monthly income, the adjusted HR was 2.956 (95% CI: 1.984-3.960, p < 0.001). Age, urbanization level, and level of care correlated with poor prognosis. Moreover, poor prognosis tended to occur in patients older than 45 years, regardless of urbanization level and level of care (p < 0.05). Notably, this study included more males than females. Moreover, the level of urbanization increased with the frequency of chronic diseases (e.g., HTN, renal disease), which more likely resulted in the crude and adjusted HR of poor prognosis (p < 0.05).
The patients were stratified by the variables presented in Table 3, and adjusted hazard ratios of different subgroups were calculated ( Table 4). The visual loss group encountered 49 medical events due to first diagnosed poor prognosis in the 2552.73 person-years (PY) observed, representing a rate of 936.93 per 105 PYs; the group without visual loss encountered 97 medical events in the 10,352.96 person-years (PY) observed, representing a rate of 1614 per 105 PYs. After Bonferroni-correction for multiple comparisons, when compared to those without visual loss, patients with visual loss poor prognosis showed patients with visual loss poor prognosis ratio of 2.956 (95% CI 1.984-3.960, p < 0.001).   This study was designed to analyze the short-, medium-, and long-term effects of VI on patients. Factors in the poor prognosis subgroups (overall, 1-year, and 5-year subgroups) posed a significant risk compared to the group without VI (Table 5). Visual loss was associated with an increased risk of poor prognosis (aHR, 2.956; 95% CI, 1.984-3.960; p < 0.001); post-traumatic stress disorder (PTSD)/acute stress disorder (ASD) in VI was 5.835 fold higher (95% CI, 3.866−7.892; p < 0.001) than those without VI. In the present study, visual loss was associated with an increased risk of any poor prognosis from the subgroups, as well as with an increased risk of specific types of personality disorders, including mental disorders, anxiety, depression, bipolar, sleep disorders, PTSD/ASD, dementia, eating disorders, SRD, psychotic disorders, autism, other mental disorders, and suicide (Table 5). After Bonferroni-correction for multiple comparisons, notable increases in the first year in SRD with VI was 5.745-fold (95% CI, 3.882-7.759; p < 0.001), PTSD/ASD was 5.526-fold (95% CI, 3.884-7.801; p < 0.001), suicide mortality was 2.402-fold (95% CI, 1.362-3.780; p < 0.001), and non-suicide mortality was 3.765-fold (95% CI, 2.087-7.184; p < 0.001). After the fifth year, the increase in depression was 5.219-fold (95% CI, 3.467-6.920; p < 0.001).
These associations remained significant after the visual loss diagnoses in the first year, persisting even after the first five years after experiences with visual loss, according to the sensitivity analysis (Table 5).  Table 3. CI = confidence interval, DM: diabetes mellitus; HTN: hypertension; CLD: chronic liver disease.

Discussion
People with VI are at a higher risk of poor mental health outcomes, as well as physical comorbidities [19][20][21]. According to a study published in 2000, two-thirds (66.6%) of patients admitted for inpatient treatment for depression showed a diminished perception of ambient light [22]. According to a cross-sectional study comprising 213 participants (with depression and without depression), patients with depression were 4.5 times more likely to report lower perception of ambient light from age-related eye illness than those without depression. However, the mechanism underlying the association between depression and impaired perception remains unknown [23].
Previous studies have shown an association between VI and self-rated health [16,24,25]. Reduced vision had an independent influence on global health ranking by those under the age of 80 years [24]. Participants in the study who reported VI were twice as likely as those who did not report VI to indicate poor self-rated health (OR = 2.13, 95% CI: 1.94-2.33). Selfreported health is thought to reflect physical health conditions. When reporting self-rated health, a respondent may also consider healthy or unhealthy habits and activities [23].
Previous studies have reported that increased exposure to potentially stressful situations is associated with an increased risk of mental health problems in persons with VI. Studies have linked visual impairments to a high prevalence of post-traumatic stress disorder [25], higher risk of depression and anxiety [26,27], and burdensome life experiences such as loneliness [28]. While it has been discovered that mental health issues are more common in young persons with VI than in older persons with VI [29], studies have also indicated that older adults with VI have a higher prevalence of a range of mental health disorders when compared with similarly aged people in the general population [30]. Therefore, people with VI are more likely to have mental health problems, regardless of age; this is in line with the results of our study.
The present study found that, in addition to depression, VI may directly increase the risk of suicide (HR: 1.50, 95% CI: 0.90-2.49). When non-ocular characteristics, including medical comorbidities and self-rated health, were considered, individuals with VI had a 64% (HR: 1.64, 95% CI: 0.99-2.72) greater risk of death by suicide [31]. While blindness due to age-related illnesses (e.g., age-related macular degeneration), diabetic retinopathy, and glaucoma is not reversible, many impairments may be cured or prevented entirely, which may prevent development of several mental health disorders [32]. Nevertheless, despite their frequent contact with patients and understanding of the devastating effects of VI, ophthalmologists seldom diagnose or treat depression [33]. This decision making involves ophthalmologists' ability to diagnose depression and suicidal behavior, as well as knowing when to send patients for psychiatric examination and care. Further research is needed to determine the frequency of cases with psychological difficulties that ophthalmologists encounter and appropriately manage. In addition, efforts to teach ophthalmologists in residency to effectively manage suicidal behaviors caused by VI are warranted.
PTSD may follow an exceptionally threatening or horrifying event, where the person experiencing it feels a severe threat of injury or death. Common symptoms of PTSD are re-experiencing the event in the form of flashbacks or nightmares, avoidance of stimuli associated with the event, alterations in cognition and mood, and increased arousal and reactivity [34]. As identified in a recent review, the one study that assessed PTSD prevalence specifically in people with VI was concerned with adolescents in a war conflict area [35]. This study found a lower prevalence of PTSD among those with impaired vision or hearing compared with those without impairments (4.2% versus 11.4%), which was explained by a lower exposure to traumatic events among those with VI [36]. However, in previous reviews, the prevalence estimates of PTSD in populations prone to VIs (older people, primary care patients) have ranged from 1.7% to 32.5% [37,38], which is both lower and higher, respectively, than those found in general population samples [39,40]. Thus, more studies are warranted to conclude whether individuals with VI are at a higher risk of PTSD.
While the current state of knowledge suggests that people with a visual impairment are more vulnerable to potentially traumatic events, few studies have examined the prevalence of PTSD. PTSD can occur after an exceptionally threatening or horrifying event in which the person experiencing it felt a severe threat of injury or death. Common symptoms of PTSD include re-experiencing the event in the form of flashbacks or nightmares, avoidance of stimuli associated with the event, and altered perceptions of the event [34].
According to a recent analysis, the one study that explicitly investigated PTSD prevalence in persons with a visual impairment was concerned with teenagers in a military conflict region [35]. This study revealed a lower prevalence of PTSD among people with impaired vision or hearing compared to those without impairments (4.2% versus 11.4%) [36], which was explained by the impaired group's lesser exposure to stressful situations. PTSD prevalence estimates in populations prone to visual impairments (older persons, primary care patients) have ranged from 1.7% to 32.5% [37,38], which is both greater and lower than that seen in general population samples [39,40]. More study is needed to determine whether those with visual impairments are more likely to develop PTSD. More study is needed to determine which life experiences may be to blame for the probable difference in PTSD prevalence between the VI population and the general population.
Synthesis the above studies, PTSD is prone to forming in people with VI. Nonetheless, the degree and impact of PTSD in this population remains uncertain, and further investigation is necessary to advance knowledge regarding these aspects. This would entail conducting larger comparative studies using dependable methods and valid assessment tools. The current diagnostic instruments for traumatic events and PTSD must be validated and, if necessary, modified for individuals with VI. First, the influence of VI in the manifestation of PTSD symptoms, particularly intrusions, avoidance, and hyperarousal, must be considered. Second, PTSD and traumatic brain injury often co-occur, particularly in ex-service personnel, and traumatic brain injury may result in VI. Consequently, it may become challenging to attribute a symptom to a specific diagnosis. Third, PTSD diagnosis may be associated with an increase in vision problems due to heightened awareness and reporting of vision problems, neurophysiological manifestations, and medication side effects. Mental health professionals, vision rehabilitation specialists, and eye care providers should be acquainted with these factors to improve identification and treatment of PTSD in this population.
The major strength of our study is its population-based design. However, this study has some limitations. First, the NHIRD did not annually assess important risk factors associated with disabling eye conditions such as smoking, a risk factor for cataracts, and age-related maculopathy. Second, the NHIRD does not provide detailed information on variables, including socioeconomic factors, occupation, unhealthy behaviors, amount of alcohol consumption, and the genetic background of the subjects, that may affect the association between VI and poor mental health outcomes. Lastly, the study participants were selected based on their medical records in the NHIRD, so our ICD−9-CM code does not identify congenital VI and does not include an assessment of some risk factors associated with suicide, such as depression. Other than that, the fact that not all suicides are reported as suicides, may underestimate the data between VI and suicides. The bias caused by unknown confounders could not be avoided in this retrospective cohort study despite meticulous adjustments. Nevertheless, multivariate logistic regression models and Bonferroni-correction for multiple comparisons were used to adjust our results. The recommendation for future studies with higher methodological rigor is appropriate to address the limitations of the current study.

Conclusions
This study revealed that patients with VI have a nearly three-fold higher risk of mental disorders, including anxiety, depressive, bipolar, and sleep disorders, than the general population. After the sensitivity test, excluding the first year, patients with VI were found to have a nearly five-fold higher risk for PTSD/ASD. When the first five years were excluded, the data showed patients with VI to have a nearly five-fold higher risk for depression. Traumatic experiences appear to have a significant impact on the mental health in people with VI, and these results highlight their need for mental health care. The high prevalence of PTSD lends credence to the suggestion for a better-adapted health-care system for persons with visual impairment. People who are visually impaired may have a greater threshold for asking for assistance. Equally significant, health workers are unaware of the mental challenges connected with visual impairments [41].
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/healthcare11101462/s1, Table S1. Abbreviation, ICD-9-CM, and definition. Funding: This study was supported by a grant from the Tri-Service Hospital Research Foundation (TSGH-B-112020). The funders had no role in the study design, data collection and analysis, decision to publish, or manuscript preparation.

Institutional Review Board Statement:
The study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board of Tri-Service General Hospital (TSGHIRB No. E202216004). The need for individual consent was waived because all identifying data were encrypted.