http://orcid.org/0000-0002-3202-6307
Figures
Abstract
The Global Burden of Disease 2010 and the WHO Global Health Estimates of years lived with disability (YLDs) uses disability-weights obtained from lay health-state descriptions, which cannot fully reflect different disease manifestations, according to severity, treatment, and environment. The aim of this study was to provide population-representative YLDs of noncommunicable diseases and injuries using a prevalence-based approach, with the disability weight measured in subjects with specific diseases or injuries. We included a total of 44969 adults, who completed the EQ-5D questionnaire as participation in the Korea National Health and Nutrition Examination Survey 2007–2014. We estimated the prevalence of each of 40 conditions identified from the noncommunicable diseases and injuries in the WHO list. Modified condition-specific disability-weight was determined from the adjusted mean difference of the EQ-5D index between the condition and reference groups. Condition-specific YLDs were calculated as the condition’s prevalence multiplied by the condition’s disability-weight. All-cause YLDs, estimated as “number of population × (1 − mean score of EQ-5D)” were 2165 thousands in 39044 thousand adults aged ≥20. The combined YLDs for all 40 conditions accounted for 67.6% of all-cause YLDs, and were 1604, 2126, 8749, and 12847 per 100000 young (age 20−59) males, young females, old (age ≥60) males, and old females, respectively. Back pain/osteoarthritis YLDs were exceptionally large (442/40, 864/146, 2037/836, and 4644/3039 per 100000 young males, young females, old males, and old females, respectively). Back pain, osteoarthritis, depression, diabetes, periodontitis, and stroke accounted for 22.3%, 9.1%, 4.6%, 3.3%, 3.2%, and 2.9% of all-cause YLDs, respectively. In conclusion, this estimation of YLDs using prevalence rates and disability-weights measured in a population-representative survey may form the basis for population-level strategies to prevent age-related worsening of disability.
Citation: Park JI, Jung HH (2017) Estimation of years lived with disability due to noncommunicable diseases and injuries using a population-representative survey. PLoS ONE 12(2): e0172001. https://doi.org/10.1371/journal.pone.0172001
Editor: Mark H. Vickers, University of Auckland, NEW ZEALAND
Received: April 27, 2016; Accepted: January 30, 2017; Published: February 14, 2017
Copyright: © 2017 Park, Jung. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Data sets used in this study are available from the Korea Centers for Disease Control and Prevention database (2007-2014). http://cdc.go.kr/CDC/contents/CdcKrContentView.jsp?cid=60940&menuIds=HOME001-MNU1130-MNU1639-MNU1748-MNU1752
Funding: The authors received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
The World Health Organization (WHO) Global Health Estimates (GHE) and World Bank-commissioned Global Burden of Disease (GBD) study measure the overall burden of disease using disability-adjusted life years (DALYs) [1]. This time-based measure combines years of life lost due to premature mortality (YLLs) and years lost due to time lived in states of less than full health (years lived with disability [YLDs]). YLDs are determined by non-fatal health outcomes of diseases and injuries; chronic noncommunicable diseases and injuries with lifelong consequences contribute markedly to non-fatal burdens of disease.
Condition-specific YLDs can be computed as the prevalence of disease or injury multiplied by the disability-weight for that condition [2,3], and their reliable quantification requires precise estimates of prevalence rates and disability-weights for those conditions. However, unfortunately, the epidemiological data currently available have limitations, including lack of information on severity distributions, inconsistent methods for measuring disability-weights, and wide variation in data sources, for most conditions [4–8]. The GBD 2010 employed lay descriptions of the consequences of various diseases and injuries for developing universal measures of disability-weights, distinct from welfare and environments [9]. The WHO GHE also used the GBD 2010-developed disability-weights to calculate global and regional YLDs, after partially revising the values. However, it is arguable whether health and welfare can be separated and whether a universal approach is possible or even desirable [10]. Additionally, a brief lay description cannot reflect various manifestations of the same disease, the effect of treatment on disability, and adaptation to environments.
The Korea National Health and Nutrition Examination Survey (KNHANES) can facilitate estimation of condition-specific YLDs for noncommunicable diseases or injuries in the general population. This large population-representative survey, conducted by the Korea Centers for Disease Control and Prevention (KCDC), used the EQ-5D questionnaire to measure health-related quality of life. The EQ-5D provides a simple descriptive profile and a single index value for health status, simplifying disability-weight calculation. The survey also includes health questionnaires and physical/laboratory examinations, allowing determination of prevalence rate and disability-weight for specific diseases and injuries. Thus, numerous condition-specific YLDs could be estimated from a single source.
The present study aimed to provide population-representative YLDs of noncommunicable diseases and injuries, based on the KNHANES data. Additionally, we compared the YLDs of this study to those of the GHE.
Materials and methods
Subjects and identification of conditions
The KNHANES is a population-based, cross-sectional study on the health and nutritional status of the non-institutionalized Korean population. The KCDC conducted the survey using a stratified, multistage, clustered probability design to select a representative, nationwide sample [11]. KNHANES comprises a health questionnaire, physical/laboratory examinations, and a nutrition survey; to date, phase I (1998), II (2001), III (2005), IV (2007−2009), V (2010−2012), and VI (2013−2015) have been executed by the Korean government. Written informed consent was obtained from each participant in the KNHANES at enrollment.
The present study protocol was approved by Kangwon national university hospital institutional review board (IRB File No.: KNUH-2015-06-001). This study was based on KNHANES 2007−2014 data, as these surveys were conducted by a single organization, KCDC, using consistent methodology since 2007, and KNHANES 2015 data were not available at the time of this analysis.
Health-related quality of life was assessed using the Korean version of the EQ-5D health questionnaire. The EQ-5D comprises five dimensions: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression. Each dimension comprised three levels: no problems, moderate problems, and extreme problems. The combination of all dimensions and levels yields 243 unique health states. The EQ-5D index scores were calculated based on the Korean value set, which has been established based on a representative national sample using the time−trade-off method [12]. Scores of 1 and 0 correspond to optimal and worst health, judged to be equivalent to death, respectively.
Of the 65973 subjects participating in KNHANES 2007−2014, we excluded subjects younger than 20-years-old (n = 16444) and adults who did not complete the EQ-5D questionnaire (n = 4560). Thus, a total of 44969 subjects (18984 males and 25985 females) were included in this study (Fig 1).
Physical and laboratory examinations
Trained medical staff and medical specialists, including ophthalmologists, otolaryngologists, and dentists performed the physical examinations, following standardized procedures. Oral health examinations were conducted in mobile centers that traveled to each survey location. Dentists conducted the examinations with the participant seated in a dental chair. Before the oral examination, participants were informed about the procedures using intra-oral pictures, tooth models, and simulation patients. Pulmonary function tests were performed in participants aged ≥ 40 years, using dry rolling seal spirometers. The procedure was conducted by trained medical personnel, who underwent education sessions on pulmonary function tests and quality control prior to the study.
From July 2008 to December 2012, ophthalmological and otological examinations were conducted in the Korea National Health and Nutrition Examination Survey (KNHANES). These examinations were conducted by trained teams from mobile centers; use of such centers provided a standardized environment and equipment. Presenting visual acuity was measured using currently available refractive correction, if any, with an international standard vision chart based on the Snellen scale. Best-corrected visual acuity was measured using autorefraction and/or a pinhole. Details of ophthalmologic examinations, including autorefractometry, slit lamp biomicroscopy, fundus photography, intraocular pressure, and visual field (frequency-doubling technology) tests have been published elsewhere [13]. Each fundus image was reviewed twice: onsite by ophthalmologists or ophthalmological residents, and then by retina specialists. An audiometry test was conducted by well-trained examiners, and the air-conduction hearing threshold was measured in a soundproof booth using an automatic audiometer at 500, 1000, 2000, 3000, 4000, and 6000 Hz.
Since 2010, plain radiographs of the knee, hip, and lumbar spine have been obtained in participants aged ≥ 50 years. The radiographic images were reviewed by two radiologists. The degree of radiographic osteoarthritis was assessed according to the Kellgren−Lawrence grading system.
Blood samples were collected after at least an 8-h fast, and random spot urine samples were obtained. The samples were processed appropriately, immediately refrigerated, and transported in cold storage to the central laboratory within 24 h. Blood hemoglobin and routine chemistries, including glucose and creatinine levels, were analyzed using standard methods. From 2007 to 2012, serum ferritin levels were measured by immunoradiometric assay. Since 2008, serum creatinine levels have been standardized to isotope dilution mass spectrometry, and urine albumin levels have been measured by turbidimetric immunoassay since 2011.
Identification of conditions
We identified noncommunicable diseases and injuries from the cause list of the World Health Organization (WHO) Global Health Estimates (GHE).[2] The GHE list provides a set of mutually exclusive and collectively exhaustive categories. The causes in the list are categorized into three broad groups: (I) communicable, maternal, perinatal, and nutritional conditions; (II) noncommunicable diseases; and (III) injuries. Among a total of 79 individual diseases of group II, we selected 30 diseases that could be identified using information available from the KNHANES data. Those 30 diseases accounted for two-thirds of the total YLDs related to group II. Nine of 10 individual injuries of group III could be identified, and those accounted for nearly 100% of the total YLDs related to group III. Additionally, we selected iron-deficiency anemia, which accounted for a fourth of the total YLDs related to group I. Of the 40 identified conditions, five were divided into subcategories. Table 1 shows the study years and number of subjects analyzed for each condition.
Definition of diseases and injuries
We defined and classified iron-deficiency anemia according to the WHO criteria [14]. Diabetes mellitus was defined as fasting blood glucose levels ≥ 126 mg/dl, being on medication for raised blood glucose, or with a history of diagnosis of diabetes. Alcohol-use disorders were defined based on the Alcohol-Use Disorders Identification Test scores according to the WHO guidelines [15].
We defined glaucoma according to the criteria of the International Society of Geographical and Epidemiological Ophthalmology classification scheme: Category 1, the presence of reliable (fixation and false-positive error ≤ 1) abnormal visual field testing (≥ one location of reduced sensitivity) plus a vertical cup-to-disc ratio (VCDR) ≥ 0.7, or asymmetry of the VCDR ≥ 0.2, or the presence of optic disk hemorrhage, or a retinal nerve fiber layer defect; Category 2, a VCDR ≥ 0.9 or asymmetry of the VCDR ≥ 0.3, or the presence of a retinal nerve fiber layer defect with violation of the inferior−superior−nasal−temporal rule; or Category 3, an intraocular pressure ≥ 22 mmHg plus a visual acuity < 3/60. Cataract was defined as nuclear (Lens Opacities Classification System [LOCS] III score ≥ 4 for nuclear opalescence or nuclear color), cortical (LOCS III score ≥ 2 for cortical cataracts), posterior subcapsular (LOCS III score ≥ 2 for posterior subcapsular), or mixed (more than one type per eye) based on comparison with standard photographs. The diagnoses of diabetic retinopathy and age-related macular degeneration were made by retina specialists based on fundus photography images using protocols from the Early Treatment for Diabetic Retinopathy Study and International Age-related Maculopathy Epidemiological Study Group. We defined visual impairment as a visual acuity < 6/18 in the better eye, including uncorrected refractive errors according to the International Classification of Diseases-10.
Disabling hearing impairment was defined as an audiometric International Society of Otolaryngology value (average of values at 500, 1000, 2000, 4000 Hz) ≥ 41 decibels in the better ear, according to the WHO classification [16]. Chronic obstructive pulmonary disease was defined and classified based on pulmonary function test results according to the Global Initiative for Chronic Obstructive Lung Disease [17]. Chronic kidney disease was classified into risk categories according to the Kidney Disease Improving Global Outcomes guidelines [18]. Osteoarthritis was defined as the presence of knee or hip pain with Kellgren−Lawrence grading scales score ≥ 2 on the corresponding radiographic images. We assessed periodontitis using the WHO Community Periodontal Index [19]. Periodontal disease was defined as a Community Periodontal Index score ≥ 3 with symptoms of difficulty in chewing and recent toothache.
We defined cancers, depression, ischemic heart disease (myocardial infarction or angina), stroke, current asthma, current peptic ulcer, cirrhosis, atopic dermatitis, and rheumatoid arthritis as a physician-based diagnosis of each disease. Unintentional or intentional injuries were defined based on self-reported questionnaires.
We summarized the definition of each condition in Table 2.
Statistical analysis and computation of disability-weights and YLDs
Statistical analyses were performed with SPSS (version 22.0). Since the KCDC conducted the KNHANES using a complex survey design, we used SPSS Complex Samples modules to produce reliable point estimates and robust standard errors.
We computed YLDs as follows. First, we estimated the prevalence of each condition in every 2 × 2 age−sex group, arranged according to an age cutoff of 60 years and sex, as well as in the total population. Composite sample weights were introduced separately in each of the analyses to provide representative estimates of the Korean population. We calculated a total of nine composite sample weights by multiplying the survey sample weights by the year weights, according to KCDC’s guidebook. The survey sample weight for each examination was computed using the sampling rate, response rate, and age−sex proportion of the Korean population. The year weight for each examination was determined by the number of households that participated in that year’s examination.
Second, a general linear model was used to test the effect of each condition on the EQ-5D index, introducing the composite sample weight. Each adjusted mean difference from the reference group was computed using the estimated marginal means of the EQ-5D index; these were estimated as the mean value, averaged over all cells generated by the age (and sex for the analysis of the total population) category, after subdividing age into seven 10-year-width categories, from 20−29 years through to 70−79 years, as well as ≥ 80 years. For comparison, all other subjects without each condition served as the reference group. The differences in these age (and sex)-adjusted mean scores between the condition and reference groups were obtained for each of the 2 × 2 age−sex groups, as well as for the total population. As in the GBD 2010 and GHE [2,3], it was assumed that the conditions co-occurred independently of each other within the age (and sex) category. However, for osteoarthritis and back pain (as well as caries and periodontitis), which were most prevalent and which were substantially correlated with each other, both conditions were introduced into the model together, to adjust for the effect of dependence. If the two-tailed P value exceeded 0.10 (one-tailed P value > 0.05), the value of the adjusted mean difference was excluded from further analyses.
Third, the “adjusted mean difference of the EQ-5D index” was used to establish a condition-specific disability-weight. To estimate the comorbidity-adjusted effect of each condition on disability (“condition specific DW”), the disability-weight that included the condition of interest (“combined DW”) was compared with the disability-weight that excluded the condition of interest (“comorbid DW”). Our approach was comparable to that of the GBD 2010 and GHE. Assuming that comorbid conditions change the quantitative score for the health-related quality of life multiplicatively rather than additively,
Assuming that comorbidities were independently distributed in the condition and reference groups within the age (and sex) category, the value of “1 – comorbid DW” was replaced with the estimated “marginal mean of EQ-5D index in the reference group”, and the value of “1 – combined DW” was replaced with the “marginal mean in the group with the condition of interest”.
We computed YLDs as the prevalence of each condition multiplied by the condition’s disability-weight; this prevalence-based method had also been used in both the GBD 2010 and GHE:
The value of “combined DW – comorbid DW” was replaced with the “adjusted mean difference of the EQ-5D index” between the condition and reference groups in our study. The same value had been calculated from “condition specific DW × (1 – comorbid DW)” in the GBD 2010 and GHE [2,3].
All-cause YLDs were estimated as the number of the population multiplied by the mean score of the combined disability-weight (= “1 – mean score of EQ5D index”).
We used the Korean population count released by Statistics Korea for 2012 in computing YLDs. Our YLDs were then compared with those of the WHO GHE 2014, the WHO’s most recent update of the GHE for 2012.
Results
Subject characteristics and EQ-5D index
The unweighted/weighted mean age of the study population was 50.4/45.6 years, and 57.8%/50.7% were female. The mean score of unweighted EQ-5D index of the study sample was 0.930, and the mean score of weighted EQ-5D index for the Korean population was 0.945. The mean score of the EQ-5D index decreased with age, and the age-related decrease in EQ-5D was more marked in females (Fig 2).
Prevalence and modified disability-weight
Table 3 shows the prevalence rates and the modified condition-specific disability-weights for noncommunicable diseases and injuries. Diabetes, alcohol-use disorders, hearing impairment, chronic obstructive pulmonary disease (COPD), chronic kidney disease (CKD), osteoarthritis, back pain, and periodontitis were common conditions, with a prevalence ≥ 5%. Iron-deficiency anemia was most prevalent in young females. Depression was more prevalent in females than in males. Alcohol-use disorders, road injury, other mechanical injuries, other unintentional injuries, and violence were most prevalent in young males. The prevalence rates of diabetes, visual/hearing impairment, ischemic heart disease, stroke, COPD, CKD, osteoarthritis, back pain, periodontitis, and edentulism were markedly higher in old than in young people. Osteoarthritis and back pain were very common in old females. Dental caries was common in all of the age−sex groups.
The modified disability-weights for visual impairments, stroke, osteoarthritis, back pain, and self-harm were distinctly larger than those for other conditions. When the disability-weight for each condition was compared between the age−sex groups, stroke, asthma, atopic dermatitis, rheumatoid arthritis, osteoarthritis, back pain, caries, periodontitis, falls, and other mechanical injuries had larger disability-weights in old than in young people. The disability-weights for iron-deficiency anemia and lung cancers were significant only in males. Diabetes had a relatively small disability-weight in young males, as compared to other age−sex groups.
All-cause and condition-specific YLD
All-cause YLDs in 39044 thousand adults aged ≥20 years were 2165 thousand years. The combined YLDs (the sum of each condition-specific YLD) from noncommunicable diseases and injuries were similar to all-cause YLDs in old males, whereas those in young females accounted for about 50% of all-cause YLDs (Table 4).
YLDs due to back pain and osteoarthritis were about 487745 and 199650, respectively, and were largest among all the condition-specific YLDs, particularly in old females. Additionally, depression, diabetes, stroke, and periodontitis had YLDs > 50000 years.
The aggregate of condition-specific YLDs for each age−sex group was similar to the condition-specific YLDs calculated using the prevalence and disability-weight determined in the total population.
Comparison with WHO estimates
Table 5 shows the estimates converted to YLDs per 100000 people. The magnitude of the combined YLDs in our study was less than the WHO estimates. The ratio difference between our combined YLDs and the WHO YLDs was largest in young males, whereas it was minimal in old females. Many condition-specific YLDs in our study were less than those of the WHO estimates. The YLDs for iron-deficiency anemia, malignancies, diabetes, depression, alcohol-use disorders, visual/hearing impairment, ischemic heart disease, COPD, and injuries were substantially smaller than those of the GHE. However, the YLDs for stroke, peptic ulcer, osteoarthritis, back pain, caries, and periodontitis were larger than those of the WHO estimates. Particularly, the YLDs for back pain and osteoarthritis in old females were markedly larger than those of the WHO estimates.
Rank and percentage of YLD
The overall rank and the percentage of YLDs in this study differed from the WHO estimates (Table 6). Alcohol-use disorders, COPD, and injuries ranked lower in our study, and depression, diabetes, alcohol-use disorders, COPD, and injuries accounted for a reduced percentage of all-cause YLDs, than in the WHO estimates. In contrast, osteoarthritis, stroke, and peptic ulcer ranked higher and accounted for a greater percentage of all-cause YLDs than in the WHO estimates.
The five leading causes of YLDs were back pain, depression, alcohol-use disorders, diabetes, and stroke, in young males, whereas these were back pain, depression, osteoarthritis, alcohol-use disorders, and caries in young females. Back pain, osteoarthritis, stroke, hearing impairment, and CKD were the top-five YLD causes in old males, while back pain, osteoarthritis, periodontitis, stroke and diabetes ranked highest in old females.
Discussion
We here readily estimated population-level YLDs for noncommunicable diseases and injuries using a prevalence-based approach, in which we measured the disability-weight in subjects with specific disease or injury. To the best of our knowledge, this is the first report to date estimating YLDs for numerous conditions using prevalence rates and disability-weights both measured in a representative sample. This new approach revealed that the increase of combined YLDs from noncommunicable diseases with ageing, which is more distinct in females, was mostly due to the exceptionally large YLDs ascribed to back pain and osteoarthritis, particularly in old females.
Back pain and osteoarthritis were very common, particularly in old females, and the disability-weights for those diseases were exceptionally large. The disability-weights were also large for visual impairments (except in uncorrected refractive errors) and stroke, but the prevalence rates of these conditions were much lower than those of back pain or osteoarthritis. Diabetes, alcohol-use disorders, hearing impairment, COPD, CKD, and periodontitis were also very common, but the disability-weights were not as large as those for back pain or osteoarthritis.
Our study findings differed somewhat from those of the GHE. The condition-specific YLDs for many diseases and injuries (except for back pain, osteoarthritis, periodontitis, stroke, CKD, caries, or peptic ulcer) were lower than those of the GHE, and the combined YLDs for all conditions in most age−sex groups (except in old females) were also lower than those of the WHO estimates. In contrast to the WHO estimates, the combined YLDs in our study differed markedly between males and females. The differences in male and female health-related quality of life have been observed not only in our study of the Korean population, but also in studies of other ethnic populations [20–22].
The GBD 2010 used lay descriptions of the symptoms and dysfunctions resulting from diseases or injuries to estimate disability-weights for those conditions, and obtained highly consistent values across surveys performed in diverse communities [9]. Although these brief descriptions were straightforward, they may fail to reflect different manifestations of any given disease in terms of severity, treatment, or environment. A number of studies have evaluated DALYs or YLDs for various diseases [23–27], and most have used health state descriptions to estimate disability-weights, following the GBD and WHO method. It is doubtful that the abstract values of disability-weights obtained using descriptions that assume typical manifestations of diseases could reflect a real-life health state. Previous YLDs differed markedly across previous studies, particularly in the case of mild disease [28]. The manifestations of disease may vary from asymptomatic to apparently symptomatic depending on individual conditions and environments. In general, mild disease states, with no or vague symptoms, are common, whereas severe states, with apparent symptoms, are relatively uncommon. If a disability-weight for a severe, uncommon disease state is used to estimate the YLDs of a mild common disease state, the YLDs may be erroneously overestimated. The lack of population information on the severity distribution of most conditions may frequently lead to mismatch errors between prevalence rates and disability-weights in the calculation of YLDs. These types of mismatch errors can be avoided by measuring both values from a single study sample.
A few studies have assessed DALYs or YLDs using a disability-weight that was directly measured in patients with specific diseases or injuries [5,29–33]. However, most of them evaluated YLDs for only one or two conditions, and even did so without reasonable reference groups. The aggregate of separate estimations of YLDs for various diseases, without relevant mutual exclusions between disease categories, may result in overestimation of total YLDs, due to the duplications of the YLDs. The GHE list, from which we identified specific diseases and injuries for this study, may provide mutually exclusive and aggregative categories. In our study, the disability-weight was measured in subjects with a specific disease or injury; both the prevalence and disability-weight were measured in a representative sample, and the YLDs from dozens of conditions in the GHE list were estimated from the same source. Thus, our estimates of the condition-specific and combined YLDs were more likely to reflect the real health state of the population and to overcome erroneous estimation due to mismatch errors or duplicated counts.
There was a marked, age-related increase in YLDs ascribed to osteoarthritis and back pain, with a notable difference between males and females. The YLDs from osteoarthritis and back pain were exceptionally large, particularly for older females. The combined YLDs from both these conditions accounted for 31.5% of all-cause YLDs in adults aged ≥ 20 years. In the Global Burden of Disease (GBD) 2010 and WHO GHE, back pain and osteoarthritis also ranked highest. However, the YLDs from those were not as marked as in our estimates, and did not differ between the sexes. Our finding of the sex differences in YLDs for back pain and osteoarthritis could possibly explain the worse health-related quality of life in females, which has also been demonstrated in previous studies performed in other ethnic groups [20–22]. For osteoarthritis, the differences in YLDs for the two sexes resulted mainly from the difference in prevalence rates. The prevalence of osteoarthritis in females was markedly higher than that in males, while the disability-weight was similar between the sexes. As we confirmed osteoarthritis from radiographs as well as from symptoms, our prevalence estimate is reliable.
On the other hand, back pain, the single highest-ranked condition, was common in young people as well as in old people. However, the disability-weight from back pain was relatively small in young people as compared to old people. Back pain has diverse causes, including osteoarthritis, herniated disks, instability, spinal stenosis, and the sequelae of spine surgery, and is most frequently diagnosed as “nonspecific back pain” [34]. In our study, in half of old people with back pain, this disorder was accompanied by radiographic osteoarthritis, but this accompanying rate was sharply decreased in young people. Further research is urgently needed to define the broad category of “back pain” better.
Additionally, the YLD estimates of our study incorporate the effect of current treatments as well as the severity of the disorder itself. This point should be considered when interpreting our results. Back pain and osteoarthritis should receive greater emphasis in terms of disability, particularly in older women. However, preventative strategies or brief supportive care, rather than traditional or specialized treatments, may be more effective in reducing osteoarthritis and back pain [35,36].
Diabetes is another important cause of disability. The YLDs from diabetes were 72000 years and accounted for 3.3% of all-cause YLDs. Diabetes is a common disease that has various complications. The GBD 2010 and WHO GHE used discrete disability-weights according to the complication of diabetes (uncomplicated, diabetic foot, and diabetic neuropathy). However, it may be difficult to establish the distribution of complications at the age, sex, and regional level. In the KNHANES sample, the microvascular complication rates of diabetes differed according to age and sex. Diabetic retinopathy/nephropathy (urine albumin-to-creatinine ratio ≥ 30 mg/g) was observed in 13.7%/19.1%, 13.9%/19.7%, 20.2/29.7%, and 19.5/24.7% of the young-male, young-female, old-male, and old-female diabetics, respectively. In our study, the disability-weight ascribed to diabetes was 0.015, 0.021, 0.026, and 0.021, in the young males, young females, old males, and old females, respectively. We believe that our data represent more reliable disability-weights and YLDs for diabetes at the age and sex level.
Depression and alcohol-use disorders are well known to be major contributors to disability. The YLDs from depression/alcohol-use disorders accounted for 4.6%/2.0% of all-cause YLDs, respectively, and the values were notably different from the WHO’s global (10.8%/4.3%) and regional (7.5%/7.6%) estimates. This large difference was caused by the marked differences in the disability-weights between the GHE and our study. The GHE disability-weights for major depression and alcohol-use disorders ranged from 0.159 to 0.655, but the overall disability-weights for depression and alcohol-use disorders were 0.072 and 0.017, respectively, in our study. The disability due to these mental and behavioral disorders could easily be affected by the social or cultural environment. Moreover, the severity of the disorders could be differently regarded by the patients themselves and by those around them. We obtained disability-weights from a self-reported questionnaire (EQ-5D), using the Korean value set that was established based on a representative national sample. The disability-weights and YLDs of our research therefore incorporate cultural effects and self-assessments.
The overall YLDs from visual impairment (including uncorrected refractive errors) accounted for 2.3% of all-cause YLDs. In adults aged ≥ 20 years, the overall prevalence of visual impairments (including blindness) with best-corrected visual acuity < 8/16 in the better eye was 1.58%, and the prevalence of uncorrected refractive errors was 3.85%. When each cause was calculated as a percentage of total causes of visual impairment (excluding uncorrected refractive errors), the causes were cataract (62.0%), glaucoma (10.1%), age-related macular degeneration (8.3%), diabetic retinopathy (3.6%), and undetermined causes (16.0%). Visual impairment from undetermined causes did not decrease the EQ-5D index scores in our study (data not shown), although that was the largest global cause of YLDs ascribed to visual impairment according to the GHE. Cataracts, uncorrected refractive errors, and macular degeneration were the top three contributors to YLDs due to visual impairment. A total of 29487 adults underwent ophthalmologic examinations in the KNHANES from July 2008 to December 2012, and trained medical staff and ophthalmologists conducted the examinations using standardized equipment and protocols. Our results may be helpful for the estimation of the global or regional burden of visual impairment.
The combined YLDs from all injuries accounted for 2.6% of all-cause YLDs. Falls and road injury accounted for 95.7% of the total YLDs from all injuries. Falls and road injury were common in old women and in young men, respectively. Although the sequelae of injuries may have a wide spectrum of severity, it may be very hard to identify the severity distribution of sequelae at the population level. Previous studies performed in European countries have suggested that injuries are main contributors to YLDs, as well as years of life lost (YLLs) [33,37]. Those studies analyzed the data based on the disability-weights obtained from patients in hospital or emergency settings. The disability-weights obtained from hospitalized patients or emergency department attendances would reflect disability for severe injuries, but would not represent disability for injuries of various states. As mentioned earlier, if the disability-weight for severe injuries was applied to a mild state, the YLDs may be overestimated. In contrast, our research was based on the KNHANES, which involved non-institutionalized civilians only, and investigated recent (within 1 year) injuries. People with severe conditions or lifelong sequelae were more likely to be excluded, and our results may underestimate the YLDs from injuries.
Stroke/ischemic heart disease accounted for 166/98 YLDs per 100000 adults and 2.9%/1.7% of all-cause YLDs in our study, whereas they accounted for 130/186 YLDs per 100000 adults and 1.2%/1.7% of all-cause YLDs in the WHO’s regional estimates. Percutaneous coronary intervention is well known to provide a benefit in terms of quality of life in patients with ischemic heart disease [38]. The existence of effective treatment may result in the contrasting YLDs between diseases. Disability due to hearing impairment could also be affected by the availability of medical resources. A substantial improvement in the mental health quality of life after cochlear implant or hearing aid use has been reported in patients with hearing impairments [39]. It is quite possible that the YLDs from hearing impairment are smaller (our estimates) than expected (WHO’s global estimates) with the aid of these modalities. The YLDs from chronic obstructive pulmonary disease was less than the WHO’s global or regional estimates. In old females the degree of decreased forced expiratory volume in 1 second did not correlate with the severity of disability. It remains possible that the assessment of chronic obstructive pulmonary disease based only on the results of a pulmonary function test cannot readily estimate the severity in old females. Chronic kidney disease also contributes to YLDs. Interestingly, chronic kidney disease with moderately increased risk significantly contributed to YLDs only in old men, whereas a more advanced state (with high risk) did not in old men. The potential overestimation of the glomerular filtration rate in the case of muscle wasting, a common problem in the elderly suffering from kidney disease, may influence the association between disability and estimated kidney function in old people. Peptic ulcer accounted for 1.5% of all-cause YLDs, and the value was larger than the WHO’s global (0.1%) and regional (0.2%) estimates. We believe that the relatively larger YLDs due to peptic ulcer reflect the regional differences in the characteristics of peptic ulcers [40]. Oral health disorders (dental caries, periodontitis, and edentulism) were important contributors to YLDs. Oral health disorder is a preventable disease and is related to general hygiene. It is therefore necessary to emphasize the importance of oral hygiene. The magnitude of YLDs due to cancers was relatively small. Malignancy is a well-known major contributor to YLLs, but the YLDs due to malignancy is not thought be large. Nevertheless, colorectal and lung cancers in old males and breast cancer in females were significant contributors to YLDs. Interestingly, iron-deficiency anemia, contributed to YLDs only in males. The causes of iron deficiency, physical activity, or comorbidities may be involved in the association between iron-deficiency anemia and YLDs. Overall, these findings suggest that multifactorial processes are involved in the determination of health-related quality of life in the general population.
Taken together, the YLDs estimated in this study differed somewhat from those of the GBD 2010 and WHO GHE. The differences might arise from three different sources. First, we estimated the prevalence rates of 40 conditions in a single representative sample, whereas those of the GBD and GHE were obtained from various sources. Additionally, we confirmed many diseases by objective physical/laboratory findings along with patient’s symptoms, to obtain more precise estimates of disease prevalence. Our estimates might provide a more consistent and reliable source for comparing disease burdens among numerous conditions. Second, we generated disability-weights from the EQ-5D index directly measured in a large sample from the KNHANES, whereas the GBD 2010 measured disability-weights using lay health-state descriptions, which could not reflect various manifestations of the same disease particularly in terms of severity. Additionally, we computed disability-weights separately in each age and sex category (young males, young females, old males, and old females). Disease burdens estimated in our study might incorporate age-specific effects as well as the severity of the disorder. Finally, we could overcome erroneous estimation of YLDs from mismatch errors between prevalence rates and disability-weights through measuring them in a single study sample. The GBD rankings are based on epidemiological data that may not be sufficiently robust for the calculation of the YLDs; the lack of reliable information on severity distributions may lead to mismatch errors in the calculation of YLDs. We believe that our YLD estimates are more likely to reflect real health states of the population.
On the other hand, the use of computed disability-weights from people having health conditions is not necessary an advantage compared to the weights that use lay health-state descriptions as a basis. Since persons with health conditions tend to underestimate their level of disability, the use of weights based on lay descriptions is the most conservative approach. In addition, disability-weights based on lay descriptions have been preferred for the estimation of disease burdens because they take into account the opinion of the general population.
There are several points to consider when interpreting our results. First, we did not investigate all the causes of YLD in the GHE list. In addition to the investigated diseases, anxiety disorders, migraine, schizophrenia, drug-use disorders, and gynecologic disorders are important contributors to the global disease burden. We could not include such conditions, due to the lack of relevant information in the KNHANES data. However, the combined YLDs from the conditions included in this study accounted for 61.7% of all-cause YLDs in the WHO’s global estimates. Additionally, we investigated visual/hearing impairments and oral health disorders, which were not included in the WHO’s regional estimates. Second, certain diseases were defined by a physician-based diagnosis of the disease, while many other diseases were confirmed by physical/laboratory examinations, in conjunction with the patient’s history. The prevalence of a diagnosed disease could different from the true prevalence. Some diagnoses may have been incorrect or missed. Third, as the survey enrolled non-institutionalized individuals who volunteered to participate, persons with severe conditions were more likely to be excluded. Thus, the estimates would have been underestimated in the KNHANES. Finally, the present study included subjects who resided in Korea, and the disability-weights were generated based on the EQ5D index scores calculated using the Korean value set, which were closer to values of the Japanese study than those of studies in western countries [12]. Since there is likely to be regional or ethnic differences in disease burdens, it is difficult to draw general conclusions applicable to the global population. Nevertheless, the values of prevalence, disability-weight, and YLDs determined in our study may be helpful in estimating non-fatal burdens of diseases in East Asia, where populations with similar ethnic and cultural backgrounds reside.
Conclusions
This relatively simple, prevalence-based approach, using a population-representative survey, could readily estimate YLDs reflecting the real health state of the general population. The results of this study may form the basis for population-level strategies to prevent age-related worsening of disability, which is more severe in females.
Acknowledgments
We express our gratitude to all the survey respondents and to members of the KNHANES.
Author Contributions
- Conceptualization: HHJ JIP.
- Formal analysis: HHJ JIP.
- Investigation: HHJ JIP.
- Methodology: HHJ JIP.
- Supervision: HHJ.
- Visualization: HHJ JIP.
- Writing – original draft: HHJ JIP.
- Writing – review & editing: HHJ JIP.
References
- 1.
World Health Organization. WHO | Global Burden of Disease (GBD) [Internet]. World Health Organization; 30 Jun 2014 [cited 26 Sep 2015]. Available: http://www.who.int/healthinfo/global_burden_disease/gbd/en/
- 2.
WHO methods and data sources for global burden of disease estimates 2000-2011 [Internet]. Department of Health Statistics and Information Systems, World Health Organization; 2013 Nov. Report No.: Global Health Estimates Technical Paper WHO/HIS/HSI/GHE/2013.4. Available: http://www.who.int/healthinfo/statistics/GlobalDALYmethods_2000_2011.pdf
- 3. Vos T, Flaxman AD, Naghavi M, Lozano R, Michaud C, Ezzati M, et al. Years lived with disability (YLDs) for 1160 sequelae of 289 diseases and injuries 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380: 2163–2196. pmid:23245607
- 4. Murray CJ, Lopez AD. Regional patterns of disability-free life expectancy and disability-adjusted life expectancy: Global Burden of Disease Study. Lancet. 1997;349: 1347–1352. pmid:9149696
- 5. Begg SJ, Vos T, Barker B, Stanley L, Lopez AD. Burden of disease and injury in Australia in the new millennium: measuring health loss from diseases, injuries and risk factors. Med J Aust. 2008;188: 36–40. pmid:18205562
- 6. Michaud CM, McKenna MT, Begg S, Tomijima N, Majmudar M, Bulzacchelli MT, et al. The burden of disease and injury in the United States 1996. Popul Health Metr. 2006;4: 11. pmid:17049081
- 7. Dodhia H, Phillips K. Measuring burden of disease in two inner London boroughs using Disability Adjusted Life Years. J Public Health. 2008;30: 313–321.
- 8. Kominski GF, Simon PA, Ho A, Luck J, Lim Y-W, Fielding JE. Assessing the burden of disease and injury in Los Angeles County using disability-adjusted life years. Public Health Rep. 2002;117: 185–191. pmid:12357003
- 9. Salomon JA, Vos T, Hogan DR, Gagnon M, Naghavi M, Mokdad A, et al. Common values in assessing health outcomes from disease and injury: disability weights measurement study for the Global Burden of Disease Study 2010. Lancet. 2012;380: 2129–2143. pmid:23245605
- 10. Voigt K, King NB. Disability weights in the global burden of disease 2010 study: two steps forward, one step back? Bull World Health Organ. 2014;92: 226–228. pmid:24700983
- 11. Kweon S, Kim Y, Jang M-J, Kim Y, Kim K, Choi S, et al. Data resource profile: the Korea National Health and Nutrition Examination Survey (KNHANES). Int J Epidemiol. 2014;43: 69–77. pmid:24585853
- 12. Lee Y-K, Nam H-S, Chuang L-H, Kim K-Y, Yang H-K, Kwon I-S, et al. South Korean time trade-off values for EQ-5D health states: modeling with observed values for 101 health states. Value Health. 2009;12: 1187–1193. pmid:19659703
- 13. Yoon KC, Choi W, Lee HS, Kim S-D, Kim S-H, Kim CY, et al. An Overview of Ophthalmologic Survey Methodology in the 2008-2015 Korean National Health and Nutrition Examination Surveys. Korean J Ophthalmol. 2015;29: 359–367. pmid:26635451
- 14.
Iron Deficiency Anaemia Assessment, Prevention, and Control A guide for programme managers [Internet]. World Health Organization; 2001. Available: http://apps.who.int/iris/bitstream/10665/66914/1/WHO_NHD_01.3.pdf
- 15.
Babor TF, Higgins-Biddle JC, Saunders JB, Monteiro MG. The Alcohol Use Disorders Identification Test Guidelines for Use in Primary Care [Internet]. Department of Mental Health and Substance Dependence, World Health Organization; 2001. Available: http://apps.who.int/iris/bitstream/10665/67205/1/WHO_MSD_MSB_01.6a.pdf
- 16.
WHO | Deafness and hearing loss [Internet]. World Health Organization; 2 Mar 2015 [cited 5 Jan 2016]. Available: http://www.who.int/mediacentre/factsheets/fs300/en/
- 17.
Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease (updated 2016) [Internet]. Global Initiative for Chronic Obstructive Lung Disease; 2016. Available: http://www.goldcopd.org/uploads/users/files/GOLD_Report 202016.pdf
- 18. Levin A, Stevens PE. Summary of KDIGO 2012 CKD Guideline: behind the scenes, need for guidance, and a framework for moving forward. Kidney Int. 2014;85: 49–61. pmid:24284513
- 19. Petersen PE, Ogawa H. Strengthening the prevention of periodontal disease: the WHO approach. J Periodontol. 2005;76: 2187–2193. pmid:16332229
- 20.
Szende A, Williams A. Measuring self-reported population health: an international perspective based on EQ-5D [Internet]. EuroQol Group; 2004. Available: http://www.euroqol.org/fileadmin/user_upload/Documenten/PDF/Books/Measuring_Self-Reported_Population_Health_-_An_International_Perspective_based_on_EQ-5D.pdf
- 21. Sullivan PW, Lawrence WF, Ghushchyan V. A national catalog of preference-based scores for chronic conditions in the United States. Med Care. 2005;43: 736–749. pmid:15970790
- 22.
Tomlin, Stephania, Radomiljac, Ali and Kay, Alison. Health and Wellbeing of Adults in Western Australia 2014, Overview and Trends [Internet]. Department of Health, Western Australia; 2015. Available: http://ww2.health.wa.gov.au/~/media/Files/Corporate/Reports and publications/Population surveys/2041-HWSS-Adults-WA-Overview-and-Trends.ashx
- 23. Kruijshaar ME, Hoeymans N, Spijker J, Stouthard MEA, Essink-Bot M-L. Has the burden of depression been overestimated? Bull World Health Organ. 2005;83: 443–448. pmid:15976895
- 24. Yoon S-J, Bae S-C, Lee S-I, Chang H, Jo HS, Sung J-H, et al. Measuring the burden of disease in Korea. J Korean Med Sci. 2007;22: 518–523. pmid:17596664
- 25. Baltussen RMPM, Sanon M, Sommerfeld J, Würthwein R. Obtaining disability weights in rural Burkina Faso using a culturally adapted visual analogue scale. Health Econ. 2002;11: 155–163. pmid:11921313
- 26. Schwarzinger M, Stouthard MEA, Burström K, Nord E. Cross-national agreement on disability weights: the European Disability Weights Project. Popul Health Metr. 2003;1: 9. pmid:14633276
- 27. Havelaar AH, de Wit MA, van Koningsveld R, van Kempen E. Health burden in the Netherlands due to infection with thermophilic Campylobacter spp. Epidemiol Infect. 2000;125: 505–522. pmid:11218201
- 28. Haagsma JA, Polinder S, Cassini A, Colzani E, Havelaar AH. Review of disability weight studies: comparison of methodological choices and values. Popul Health Metr. 2014;12: 20. pmid:26019690
- 29. Darbà J, Kaskens L, Pérez-Álvarez N, Palacios S, Neyro JL, Rejas J. Disability-adjusted-life-years losses in postmenopausal women with osteoporosis: a burden of illness study. BMC Public Health. 2015;15: 324. pmid:25880810
- 30. Holtslag HR, van Beeck EF, Lichtveld RA, Leenen LP, Lindeman E, van der Werken C. Individual and population burdens of major trauma in the Netherlands. Bull World Health Organ. 2008;86: 111–117. pmid:18297165
- 31. Brennan DS, Spencer AJ, Roberts-Thomson KF. Quality of life and disability weights associated with periodontal disease. J Dent Res. 2007;86: 713–717. pmid:17652197
- 32. Jia T-W, Zhou X-N, Wang X-H, Utzinger J, Steinmann P, Wu X-H. Assessment of the age-specific disability weight of chronic schistosomiasis japonica. Bull World Health Organ. 2007;85: 458–465. pmid:17639243
- 33. Lyons RA, Kendrick D, Towner EM, Christie N, Macey S, Coupland C, et al. Measuring the population burden of injuries—implications for global and national estimates: a multi-centre prospective UK longitudinal study. PLoS Med. 2011;8: e1001140. pmid:22162954
- 34. Clark LL, Hu Z. Low back pain, active component, U.S. Armed Forces, 2010-2014. MSMR. 2015;22: 8–11.
- 35. Fransen M, Simic M, Harmer AR. Determinants of MSK health and disability: lifestyle determinants of symptomatic osteoarthritis. Best Pract Res Clin Rheumatol. 2014;28: 435–460. pmid:25481425
- 36. Arora P, Vasa P, Brenner D, Iglar K, McFarlane P, Morrison H, et al. Prevalence estimates of chronic kidney disease in Canada: results of a nationally representative survey. CMAJ. 2013;185: E417–23. pmid:23649413
- 37. Polinder S, Meerding WJ, Mulder S, Petridou E, van Beeck E, EUROCOST Reference Group. Assessing the burden of injury in six European countries. Bull World Health Organ. 2007;85: 27–34. pmid:17242755
- 38. Weintraub WS, Spertus JA, Kolm P, Maron DJ, Zhang Z, Jurkovitz C, et al. Effect of PCI on quality of life in patients with stable coronary disease. N Engl J Med. 2008;359: 677–687. pmid:18703470
- 39. Contrera KJ, Betz J, Li L, Blake CR, Sung YK, Choi JS, et al. Quality of life after intervention with a cochlear implant or hearing aid. Laryngoscope. 2016;
- 40. Lam SK. Differences in peptic ulcer between East and West. Baillieres Best Pract Res Clin Gastroenterol. 2000;14: 41–52. pmid:10749088