Published online Jun 12, 2024.
https://doi.org/10.3346/jkms.2024.39.e203
Trends and Barriers in Diabetic Retinopathy Screening: Korea National Health and Nutritional Examination Survey 2016–2021
Abstract
Background
The prevalence of diabetes is increasing globally, highlighting the importance of preventive healthcare. This study aimed to identify the diabetic retinopathy (DR) screening rates and risk factors linked to DR screening nonadherence in the Korean population through a nationally representative sample survey.
Methods
Among the Korea National Health and Nutrition Examination Survey database from 2016 to 2021, participants aged ≥ 40 years with diabetes were included. The weighted estimate for nonadherence to DR screening within a year was calculated. Risk factor analyses were conducted using univariate and multivariate logistic regression.
Results
Among the 3,717 participants, 1,109 (29.5%) underwent DR screening within the past year, and this national estimate exhibited no statistically significant difference from 2016 to 2021 (P = 0.809). Nonadherence to annual DR screening was associated with residing in rural areas, age ≥ 80 years, low educational level, self-reported good health, absence of ocular disease, current smoking, lack of exercise and dietary diabetes treatment, and no activity limitation (all P < 0.05).
Conclusion
The recent DR screening rate in Korea was relatively low. Factors associated with apathy and complacency towards personal health were associated with the nonadherence to DR screening. Educational interventions have the potential to enhance the annual screening rate for diabetic patients.
Graphical Abstract
INTRODUCTION
Diabetic retinopathy (DR) is a prevalent microvascular complication of diabetes and a leading cause of visual impairment and blindness.1 In the US, DR affected 9.6 million people (26.4% of those with diabetes) in 2021, with 1.84 million people (5.1%) experiencing vision-threatening.2 Global population-based studies reported an annual DR incidence ranging from 2.2% to 12.7%.3 In Korea, a nationwide survey from 2017 to 2018 revealed a 24.5% prevalence of DR among those with known diabetes.4
This ocular complication of diabetes can be partially prevented through early detection and by managing modifiable risk factors, coupled with appropriate treatments like laser photocoagulation and intravitreal injections. This highlights the critical importance of DR screening, a facet currently suboptimal in Korea. In a study involving individuals aged ≥ 40 years with diabetes, only 37% underwent DR screening within the past year, as per the Korea National Health and Nutrition Examination Survey (KNHANES) 2005 and 2007–2009.5 Identifying factors associated with nonadherence to DR screening holds paramount significance in making health-related policies.
Therefore, analyzing recent trends in DR screening rates and their associated factors, especially during the coronavirus disease 2019 (COVID-19) period is necessary. To assess the latest status of DR screening in Korea, the present study utilized KNHANES VII (2016–2018) and VIII (2019–2021) database, examining sociodemographic and clinical factors related to nonadherence to DR screening.
METHODS
Study design and participants
This study utilized data from the seventh (2016–2018) and eighth KNHANES (2019–2021). KNHANES, a national cross-sectional survey and examination, has been representative of the non-institutional civilian population since 1998, assessing the health and nutritional status of adults and children in Korea. It comprises three components: the Health Interview Survey, Health Examination, and Nutrition Survey.
Among the 46,828 subjects in both periods of national health surveys, 3,739 were diagnosed with diabetes and were ≥ 40 years old. Our study focused on 3,717 participants (1,776 from KNHANES VII and 1,941 from KNHANES VIII) who responded “yes” or “no” to the question of whether they had received DR screening in the past year (Fig. 1).
Fig. 1
Flow diagrams of the study population (No. of persons).
KNHANES = Korea National Health and Nutrition Examination Survey.
Variables
Based on self-reported questionnaire-based information, the participants who answered, “yes” to the question “Within the past year, have you undergone an eye examination (fundus examination) to screen for diabetic complications?” were classified into the DR screening group, while those responding “no” were categorized into the nonadherence group.
To explore the factors associated with DR screening, we collected sociodemographic, socioeconomic, and health status variables including: year, residential area (urban/rural, classified by administrative district), sex (female/male), age group (40–59/60–79/≥80 years), household income (lowest/2nd–4th/highest quintile), education level (elementary school or lower/middle school/high school/university or higher), marital status (living with spouse or not), household members (single/multiple), self-reported health status (healthy/fair/unhealthy), ocular disease (cataract, age-related macular degeneration [AMD], and glaucoma), smoking status (never/former/current smoker), binge alcohol user, non-pharmacologic treatment for diabetes, obesity (body mass index ≥ 25 kg/m2 or not), activity limitation, economic status (unemployed/employed), perceived stress (feeling a lot/little), physical activity, comorbidities, and cancer.
To define ocular disease, comorbidities, and cancer variables, we relied on participants’ responses to the question: “Have you ever been diagnosed with a disease by a doctor before?” Ocular disease was established based on whether the individual had a previous diagnosis from a doctor for cataract, AMD, or glaucoma. Participants were categorized into two groups: those with at least one of these three conditions and those with none. Comorbidities included hypertension, dyslipidemia, stroke, ischemic heart disease, arthritis, osteoporosis, pulmonary tuberculosis, asthma, thyroid disease, depression, kidney disease, and liver cirrhosis. The number of comorbidities was considered a continuous variable. Cancer variables were coded as “yes” if the individual had been diagnosed with any of the following: gastric cancer, liver cancer, colon cancer, breast cancer, cervical cancer, lung cancer, thyroid cancer, or other types of cancer. If none were present, it was categorized as “no.”
All respondents were asked whether they had smoked a total of 100 cigarettes in their lifetime. Those who answered “yes” and currently smoked were categorized as “current smokers,” while those who had quit smoking were classified as “former smokers.” Respondents who had smoked fewer than 100 cigarettes in their lifetime were labelled as “never smokers.” Binge alcohol users were defined as individuals who consumed more than seven drinks on a single occasion for men and five drinks on a single occasion for women two or more times within a week.6
We evaluated diabetes management as either pharmacologic treatment alone (utilizing insulin or oral glucose-lowering medications) or a supplementary non-pharmacologic approach (involving exercise or dietary methods).
In the KNHANES, an activity limitation was defined as a current restriction in a respondent’s daily and social activities due to health issues, encompassing both physical and mental aspects.6
Physically active status was defined as engaging in at least 2 hours 30 minutes of moderate-intensity physical activity, or 1 hour 15 minutes of high-intensity physical activity, or a combination of moderate- and high-intensity physical activity (1 minute of high-intensity physical activity = 2 minutes of moderate-intensity physical activity) per week.6
Statistical analysis
The data were analyzed with SPSS version 27.0 software (IBM Corp., Armonk, NY, USA), integrating variables of strata, cluster, and weight. Clinical characteristics were compared between individuals who underwent DR screening in the past year and those who did not. Multivariate logistic regression analysis was conducted to identify predictive factors associated with nonadherence to DR screening, adjusting for covariates determined in the univariate analysis. The multivariate analysis utilized data from 3,149 subjects who completed the related survey. All statistical tests were two-sided, with differences considered statistically significant at a significance level of 0.05.
Ethics statement
This study was reviewed and approved by the Institutional Review Board of the Seoul National University Bundang Hospital and the requirement for informed consent was waived (IRB X2401-876-901).
RESULTS
A total of 3,717 participants were included in the study. There were 1,893 women, with a mean age of 64.2 ± 0.2 years. Table 1 illustrates the characteristics of the study population. The mean number of comorbidities was 1.8 ± 1.3. Overall, 1,109 (29.5%) participants underwent a DR screening within the past year. The percentage of participants receiving DR screening showed a decline after in 2019, yet there was no statistically significant difference between the years (P = 0.809, χ2 test) or between KNHANES VII and VIII (30.4% vs. 28.8%, P = 0.374, χ2 test) (Fig. 2).
Table 1
Characteristics of study population (N = 3,717)
Fig. 2
The rate of diabetic retinopathy screening from 2016 to 2021. The error bars represent the 95% confidence interval.
DR = diabetic retinopathy, KNHANES = Korea National Health and Nutrition Examination Survey.
The Supplementary Table 1 presents the odds ratios for factors associated with nonadherence to DR screening identified in the univariate analysis. Among them, living in rural area (adjusted odds ratio [aOR], 1.32; 95% confidence interval [CI], 1.04–1.68) than those living in urban area, aged ≥ 80 years (aOR, 2.96; 95% CI, 1.93–4.55) than those who were aged 40–59 years, elementary school or lower education level (aOR, 1.64; 95% CI, 1.18–2.28) than university or higher, self-reported “healthy” (aOR, 1.69; 95% CI, 1.21–2.36) or “fair” status (aOR, 1.26; 95% CI, 1.01–1.56) than “unhealthy” status, having no ocular disease (aOR, 1.59; 95% CI, 1.29–1.96) than having any of the three conditions (cataract, AMD, and glaucoma), current smoker (aOR, 1.80; 95% CI, 1.26–2.56) than never smoker, without exercise and dietary treatment for diabetes (aOR, 1.60; 95% CI, 1.21–2.13), and without activity limitation (aOR, 1.50; 95% CI, 1.15–1.95) were constantly significant in the multivariate analysis (Table 2). The subanalysis that divided KNHANES VII and VIII revealed similar results (Supplementary Tables 2 and 3).
Table 2
Factors associated with nonadherence to diabetic retinopathy screening in multivariate logistic regression analysis (N = 3,149)
DISCUSSION
In this study, we examined trends in DR screening rates and identified sociodemographic, socioeconomic, and health-related factors influencing DR screening in South Korea using a nationally representative database.
Among Korean adults aged ≥ 30 years, the estimated prevalence of diabetes was approximately 6.05 million people in 2020.7 Considering the presence of approximately 4,000 ophthalmologists in Korea at that time, a rough calculation suggests an estimate of around 1,500 adult patients of clinical burden per ophthalmologist per year. Additionally, DR screening in Korea is now being performed not only by ophthalmologists but also by various physicians who diagnosed diabetes using fundus photography. Therefore, the screening capacity of the current Korean eye care system seems sufficient for every diabetic patient. The primary challenge lies in the low DR screening rate.
A previous study utilizing KNHANES III and IV (2005, 2007–2009) database revealed that 37% (998/2660) of diabetic participants aged ≥ 40 years had undergone DR screening within a year.5 The authors identified factors such as age > 65 years, living in urban areas, high educational levels, and self-reported “unhealthy” status as favorable for receiving annual DR screening. Compared to this a decade-old result, our study revealed a substantial decrease in the DR screening rate (29.5%) using the same database setting. Another paper using 2006–2015 National Health Insurance Service–National Sample Cohort data and using procedure codes for fundus examination, fundus photography, and/or fundus fluorescein angiography to define DR screening found a steady increase in fundus screening for DR, but still less than 30%.8 In the US, 63.4% of adults aged ≥ 20 years diagnosed with diabetes were adherent to diabetic eye examinations within the past year.9 While our study could not provide specific reasons for the recent decrease in DR screening rates compared to previous studies, efforts to increase the DR screening rate need to be intensified because of its negative impact on medical costs, quality of life, and mortality.10, 11
The impact of the COVID-19 pandemic on healthcare utilization in South Korea was comparatively less significant than the global average, despite the implantation of closure and distancing policies.12 Park et al.11 reported that there was no correlation with the number of DR screenings in Korea before and after COVID-19 when assessed within a brief period (2019 and 2020). In Hong Kong, effective infection control measures enabled the continuous provision of DR screening throughout the lockdown period due to COVID-19 outbreak.13 Similarly, in our study, which covered an extended period pre- and post-COVID-19, no statistically significant difference was observed in the yearly rates of DR screening. Overall, it appears that in some countries, the COVD-19 pandemic did not significantly impact the DR screening rates, possibly due to the perception that ophthalmic visits pose a relatively lower risk of COVID-19 exposure.11
We identified several risk factors associated with nonadherence to DR screening. These include living in rural areas, aged ≥80 years, low education levels, self-reporting as healthy or fair status, having no ocular disease, being a current smoker, lacking supplemental exercise and dietary treatment for diabetes, and having no activity limitation. Some of these findings align with previous reports.5, 9, 11 Interestingly, it was revealed that individuals who perceived themselves as healthy and reported no particular issues with their eyes or body tended not to receive DR screening. Considering the risk factors highlighted in our study, low concern and complacency towards personal health seem to hinder DR screening. This underscores the importance of educational interventions to enhance awareness and encourage regular DR screening. These educational measures require efforts not only from ophthalmologists but also from internists who diagnose diabetes. Many diabetic patients may not be aware of the risks associated with DR without a physician’s warning. In the US, the lack of health insurance is a significant barrier to DR screening (76% of insured vs. 36% of uninsured Americans).9 However, since all Korean residents are legally registered in the National health Insurance Service, insurance issues do not impact DR screening in Korea.
Recently, artificial intelligence (AI)-based analysis technology to detect diabetic complications has significantly improved, supporting physicians in making clinical decisions. In the near future, widespread adoption of AI-based medical image analysis technology is expected to increase the screening and diagnosis rates of DR.14
Our study had some limitations. First, this study relied on cross-sectional data, limiting the ability to establish causation despite confirming associations, which is an inherent constraint within this dataset. Additionally, since KNHANES relies on self-reported questionnaires, there is a possibility of receiving inaccurate information from respondents. Second, despite the various risk factors found in this study, the precise causes behind the low DR screening rate remain undetermined. Further studies should incorporate inquiries regarding the reasons for not undergoing DR screening to provide comprehensive insights into this aspect.
In this study, using a nationwide survey database, we observed recent low adherence to DR screening and identified individuals at risk of nonadherence to screening. Currently, the Korean Ophthalmological Society is conducting campaigns to promote fundus examinations as routine health checkups; however, there remains a considerable distance to cover. Considering the substantial burden of DR, impacting not only society but also individual patients, urgent medical, political, and social support is imperative to augment annual DR screening.
SUPPLEMENTARY MATERIALS
Factors associated with nonadherence to diabetic retinopathy screening in univariate logistic regression analysisSupplementary Table 1
Factors associated with nonadherence to diabetic retinopathy screening in logistic regression analysis (Korea National Health and Nutrition Examination Survey VII)Supplementary Table 2
Factors associated with nonadherence to diabetic retinopathy screening in logistic regression analysis (Korea National Health and Nutrition Examination Survey VIII)Supplementary Table 3
Funding:This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. RS-2023-00210974) and Seoul National University Bundang Hospital (Grant No. SNUBH-18-2023-0014).
Author Contributions:
Conceptualization: Kim MS.
Data curation: Kim MS.
Formal analysis: Kim MS.
Funding acquisition: Kim MS.
Investigation: Kim MS.
Methodology: Kim MS.
Writing - original draft: Kim MS.
Writing - review & editing: Kim MS, Park SJ, Joo K, Woo SJ.
References
-
Korea Disease Control and Prevention Agency. Guideline for Data Users of Korea National Health and Nutrition Examination Survey. Cheongju, Korea: Korea Centers for Disease Control and Prevention; 2023.
-
-
Yoo KJ, Lee Y, Lee S, Friebel R, Shin SA, Lee T, et al. The road to recovery: impact of COVID-19 on healthcare utilization in South Korea in 2016–2022 using an interrupted time-series analysis. Lancet Reg Health West Pac 2023;41:100904
-
Figures
1 / 2
Tables
1 / 2
ORCID IDs
Funding Information
-
National Research Foundation of Korea
RS-2023-00210974
-
Seoul National University Bundang Hospital
SNUBH-18-2023-0014
