Skip Navigation
Skip to contents

PHRP : Osong Public Health and Research Perspectives



Page Path
HOME > Osong Public Health Res Perspect > Volume 7(2); 2016 > Article
Original Article Accuracy of Self-reported Hypertension, Diabetes, and Hypercholesterolemia: Analysis of a Representative Sample of Korean Older Adults
Heeran Chun, Il-Ho Kim, Kyung-Duk Min
Osong Public Health and Research Perspectives 2015;7(2):108-115.
Published online: December 15, 2015
  • 23 Download
  • 33 Crossref
  • 34 Scopus

aDepartment of Health Administration, Jungwon University, Chung-buk, Korea

bSocial and Epidemiological Research, Centre for Addiction and Mental Health, Toronto, Canada

cDepartment of Psychiatry, University of Toronto, Toronto, Canada

dInstitute of Health and Environment, Graduate School of Public Health, Seoul National University, Seoul, Korea

• Received: October 14, 2015   • Revised: November 30, 2015   • Accepted: December 7, 2015

Copyright © 2016 Korea Centers for Disease Control and Prevention. Published by Elsevier Korea LLC.

This is an open access article under the CC BY-NC-ND license (

  • Objectives
    This study will assess the accuracy of self-reported hypertension, diabetes, and hypercholesterolemia among Korean older adults.
  • Methods
    Using data from the fourth Korean National Health Examination and Nutrition Survey (KNHANES IV, 2007–2009), we selected 7,270 individuals aged 50 years and older who participated in both a health examination and a health interview survey. Self-reported prevalence of hypertension (HTN), diabetes mellitus (DM), and hypercholesterolemia was compared with measured data (arterial systolic/diastolic blood pressure, fasting glucose, and total cholesterol).
  • Results
    An agreement between self-reported and measured data was only moderate for hypercholesterolemia (κ, 0.48), even though it was high for HTN (κ, 0.72) and DM (κ, 0. 82). Sensitivity was low in hypercholesterolemia (46.7%), but high in HTN and DM (73% and 79.3%, respectively). Multiple analysis shows that predictors for sensitivity differed by disease. People with less education were more likely to exhibit lower sensitivity to HTN and hypercholesterolemia, and people living in rural areas were less sensitive to DM and hypercholesterolemia.
  • Conclusion
    Caution is needed in interpreting the results of community studies using self-reported data on chronic diseases, especially hypercholesterolemia, among adults aged 50 years and older.
The accuracy of self-reported cardiovascular disease (CVD) and its determinant factors are a pivotal issue for global public health in CVD prevention and management among older populations. CVD continues to be a leading cause of morbidity and mortality in most developed and developing countries 1, 2, 3. In accompaniment with a rapidly growing elderly population concomitant with increasing life expectancy, South Korea (hereafter Korea) has seen a dramatically increasing trend in the risk of CVD. The prevalence of CVD risk factors such as of hypertension (HTN), diabetes mellitus (DM), and hypercholesterolemia among adults (30 years or older) has increased from 24.6%, 9.6%, and 10.7% in 2007, to 27.3%, 11.0%, and 14.9%, in 2013, respectively [4]. To prevent and manage the increasing health burden of CVD risk factors, the Korean government implemented the National Cardio-Cerebrovascular Disease Plan 2010–2015 [5], but patients with these chronic diseases are frequently excluded from the government's enrollment system, due to limited public knowledge and awareness of CVD risk factors [6].
Accurately assessing the prevalence and trend of these diseases is a prerequisite for societal disease management and medication compliance. While, in reality, self-reported health data through surveys are widely used due to cost-effectiveness, it is noted that complex variations between subjective and objective measures of CVD risk factors hamper us from a better understanding of the magnitude of CVD, its associated factors, and the effectiveness of government interventions for its prevention 7, 8, 9, 10. A handful of studies suggest that the prevalence of self-reported DM shows a relatively high level of agreement 11, 12, 13, 14, but high cholesterol revealed a significant discrepancy 9, 10, 11, 15, with a mixed result for high blood pressure. A Minnesota study using 2,037 participants aged 45 years or older suggests that agreement between reported and medical records was noticeable for both DM and HTN (κ 0.71–0.80) [12]. In comparison to the American research, several European findings suggest similar agreement for DM (κ 0.84–0.76), but a significantly lower agreement for HTN (κ 0.63–0.51) and hypercholesterolemia (κ 0.55–0.48) 9, 11, 13. A recent European study of 12 countries estimated that nearly 70% of European adults were unaware of having high cholesterol levels [11]. These results imply that undiagnosed cases of chronic patients could significantly deteriorate the quality of CVD primary care and intervention, when relying only on self-reported health data. However, there is a substantial knowledge gap in the accuracy of self-reported chronic diseases among Asian older adults. In addition, it is still unclear why this discrepancy in agreement exists for CVD risk factors, while emerging research shows that some socio-demographic factors such as age, sex, and education can contribute to its accuracy 9, 16, 17, 18, 19.
The central objective of this epidemiological research is to investigate the accuracy of self-reported HTN, DM, and hypercholesterolemia among Korean older adults. The specific goals of the study are to assess: (1) whether Korean older adults have a higher accuracy of reporting CVD risk factors when compared with measured data; (2) whether the extent to which the observed variation between the two measures differs in HTN, DM, and hypercholesterolemia; and (3) whether the extent may be attributable to demographic factors, socioeconomic factors, and/or health behavioral factors. The study will also examine whether there are different determinant factors between DM, HTN, and hyperlipidemia. To achieve our aims, we will use the Fourth Korean National Health and Nutritional Examination Survey (KNHANES IV), 2007–2009, the representative national data [20].
2.1 Data and study population
This study is based on data from KNHANES, conducted from 2007 to 2009. This nationally representative cross-sectional survey included health interviews, health examinations, and nutritional surveys, intended to monitor the health and nutrition status of the Korean population. The details of this survey have been published in several publications [20]. Out of 24,871 individuals for whom both reported and measured data (health interviews and health examination) were available, 8,529 (34.3%) were over the age of 50 years. We obtained 7,270 observations for the final analysis, after eliminating all missing data.
2.2 CVD risk factors
CVD risk factors included HTN, DM, and hypercholesterolemia. Reported HTN, DM, and hypercholesterolemia cases were ascertained with the following questions. Has a doctor ever told you that you have: (1) high blood pressure or HTN? (2) hyperglycemia or DM? or (3) high blood cholesterol or hypercholesterolemia? Regarding the measured CVD risk factors, HTN was defined as a systolic blood pressure ≥ 140 mmHg, or diastolic blood pressure ≥ 90 mmHg, or currently using medication due to high blood pressure (Joint National Committee-6 definition) [21]. In the case of DM, individuals with fasting blood glucose levels of 126 mg/mL or those under treatment are categorized as DM [22]. Hypercholesterolemia is considered as a fasting blood cholesterol level ≥ 240 mg/mL or using medication for the condition [23].
2.3 Predictors
Sociodemographic and behavioral characteristics were taken into account as predictors of the discrepancy between reported and measured data. Sociodemographic variables included sex, marital status, region (urban/rural), education (elementary school/high school/college or higher), employment status (yes/no), and household income. The equalized household income levels were categorized into tertiles (high/middle/low), after household income was divided by the square root of the household size. This was to adjust for differences in disposable income by the numbers of people in the household [24].
We considered three behavioral factors: smoking (never/past/current); binge drinking (yes/no); and physical activity (yes/no). Binge drinking was defined as consuming alcohol more frequently than twice a week and more than seven glasses (50 mL; 5 glasses for women) on each occasion. Physical activity includes three types of exercise: (1) intensive activity for at least 20 minutes on 3 or more days a week; (2) moderate activity for at least 20 minutes on 5 days or more a week; and (3) walking for at least 30 minutes on 5 days or more a week. Self-rated health was divided into three groups: very good and good, average, and poor and very poor.
2.4 Statistical analysis
Descriptive statistics for proportions and mean were presented for all variables. Kappa statistics were employed to measure the accordance of reported and measured values. Sensitivity and specificity were calculated to examine the accuracy of reported HTN, DM, and hypercholesterolemia, compared with the gold standard measures of these CVD risk factors. We used multiple logistic regression models to assess factors associated with an individual's awareness of CVD risk factors. Awareness rates and sensitivity indicated the proportion of participants who reported having HTN, DM, and hypercholesterolemia among the cases of clinically measured CVD risk factors. SAS 9.3 was used for these statistical analyses.
Table 1 shows the descriptive characteristics, the prevalence of CVD risk factors (HTN, DM, hypercholesterolemia), as well as the results of κ, sensitivity, and specificity tests between self-reported and measured diseases among participants aged 50 years or older. The first two columns of Table 1 show general characteristics of the study population. The sample consisted of similar proportions of three age groups: 50–59 years (36.8%), 60–69 years (34.7%), and 70 years and older (28.6%). The proportion of women was higher than men (57.4% vs 42.6%). Nearly three in four respondents were currently married, while one in four was divorced, separated, or widowed. The majority of participants (63.1%) lived in urban areas, compared to 37% rural dwellers. While the highest percentage of respondents (56.7%) had an education level of elementary school or less, only 9.1% of them had earned a college degree or higher, and 17.1% had received a high school diploma. About half of the respondents were in paid employment. In terms of health behaviors, 16.6% were current smokers, 22.6% were past smokers. About 7.8% were binge drinkers, and 41.3% did not exercise. More than one third of participants reported their health was very good or good, and another third reported having poor or very poor health.
The results in the right column show the prevalence and values of κ sensitivity and specificity between measured and self-reported HTN, DM, and hypercholesterolemia. The prevalence of three measured CVD risk factors was higher than that of reported risk factors, especially in reported hypercholesterolemia. In general, both the reported and measured prevalence in HTN and DM seems to be much higher among those who were older, previously married, and had lower socioeconomic status (lower education, lower income, not employed). For hypercholesterolemia, the reported and measured prevalence shows a similar trend to those of HTN and DM, but the measured prevalence differs across sex (women: 21%, men: 10.9%), marital status (previously married: 20.0%, married: 15.6), district (urban: 18.4%, rural: 13.7%), and employment status (not employed: 19.4%, employed: 14%). It is noteworthy that these differences were much higher than those of HTN and DM.
Regarding the results of κ and sensitivity tests, whereas HTN and DM showed substantial agreement (κ: 0.72 and 0.82, sensitivity: 73.0 and 79.3, respectively), hypercholesterolemia showed moderate agreement (κ: 0.48, sensitivity: 46.7). When considering demographic and socioeconomic factors overall, higher kappa and sensitivity between measured and reported HTN was found in the following three demographics: the middle age group (60–69 years), women, and the unemployed group. However, the differences were not found in DM, except not employed. By contrast, the agreement and sensitivity of hypercholesterolemia was higher among those with a higher level of education (κ: 0.54 vs. 0.46, sensitivity: 57.4 vs. 41.6) and a higher income (κ: 0.49 vs. 0.44, sensitivity: 52.1 vs. 39.8), compared with reference groups. Interestingly, the κ and specificity in hypercholesterolemia were higher among those who were non-employed than employed, in all three CVD risk factors.
For health behavioral factors and self-rated health, the prevalence of measured HTN and DM was highest among past smokers (51.9% and 18.6%, respectively), binge drinkers (52.5% and 16.2%, respectively), and the no physical activity group (50.3% and 16.3%, respectively). By contrast, compared to the counterparts, the kappa and sensitivity in HTN and DM were higher among non-binge drinkers and never smoked. For hypercholesterolemia, measured prevalence was higher among never smoked and binge drinkers. The prevalence, κ, sensitivity in HTN, DM, and hypercholesterolemia were the highest among those who had poor and very poor self-rated health (κ: 0.77, 0.85, and 0.52; sensitivity: 80.6, 87.5, and 52.6, respectively). It is noteworthy that specificity for all three diseases shows a high accuracy of > 95%, regardless of the individual's socioeconomic or health behavioral factors.
Table 2 shows the results of multiple logistic regressions, examining the sociodemographic and behavioral factors associated with the sensitivity of CVD risk factors, after eliminating participants who did not possess each chronic disease according to the health examination. The total samples used in this analysis were 3,547 for HTN, 1,118 for DM, and 1,214 for hypercholesterolemia. In general, compared to participants aged 50–59 years, those aged 60–69 years had a high level of sensitivity for all three CVD risk factors (HTN, DM, hypercholesterolemia), while only adults aged 70 years or more were more aware of their HTN symptoms. Women have greater sensitivity to HTN than men [odds ratio (OR): 1.37, 95% confidence interval (CI): 1.07–1.74], but men were more sensitive to hypercholesterolemia (OR: 0.60, 95% CI: 0.39–0.93). Marital status did not influence sensitivity to all three risk factors. Compared with adults living in urban areas, rural dwellers were less likely to be aware of CVD risk factors, especially DM and hypercholesterolemia. According to one's education level, a socioeconomic gradient was found in awareness of HTN and hypercholesterolemia, but not in DM. Compared to those with college education or higher, ORs for sensitivity progressively decreased with a lower educational level. Participants who were not employed seemed to be more aware of their HTN and hypercholesterolemia than those who were employed.
Regarding health behaviors, only smoking was related to sensitivity of hypercholesterolemia. Sensitivity in past and current smokers was lower than for those who have never smoked (OR and 95% CI for past smoker: 0.60: 0.38–0.93; current smoker: 0.41: 0.26–0.66). Regardless of CVD risk factors, poor and very poor self-rated health was significantly associated with a higher rate of sensitivity than good and very good self-rated health (OR and 95% CI for HTN: 2.07, 1.71–2.50; DM: 3.23, 2.21–4.72; and hypercholesterolemia: 2.09, 1.55–2.80).
This study was conducted to determine the accuracy of self-reported HTN, DM, and hypercholesterolemia when compared with global gold standard of measured data: representative national data from KNHANES IV. Our study's findings show that sensitivity of self-reported data for HTN and DM among Korean older adults aged 50 years and older was relatively high (73.0% for HTN and 79.3% for DM), but that of hypercholesterolemia was low (46.7%). Specificity, by contrast, provided accurate results for all three diseases (> 95%). Reliability, based on κ agreement, showed similar results: κ = 0.72, 0.82, and 0.48 for HTN, DM, and hypercholesterolemia, respectively. Reliability and validity tests show that self-reported data are relatively good for HTN and DM, but reported hypercholesterolemia suggests a significant reporting error. In terms of accuracy of reported DM, our results highly agreed with previous studies from other regions, such as Europe and the USA, which show high level of accuracy. Moreover, our findings were similar to those from European research of the agreement between self-reported and measured data for HTN and hypercholesterolemia. A Dutch study [13] showed the highest concordance for DM (κ = 0.84), a moderate level for HTN (κ = 0.63), and the lowest level for hypercholesterolemia (κ = 0.48). A study of the Spanish adult population [9] also suggested good agreement for self-reported DM (κ = 0.78), but moderate agreement for HTN (κ = 0.51), as well as poor agreement for hyperlipidemia (κ = 0.27). While respondents' awareness of the nature of their diseases can affect their accurate self-reporting, it is assumed that intervention programs for CVD risk factors—public education campaigns, knowledge translation, and the infrastructure of health examination programs across countries—may influence these differences in awareness of CVD risk factors.
Given the significance of preventing or controlling CVD risk factors, important modifiable risk factors in specificity for these diseases have not been adequately assessed. Our multiple logistic regression analysis shows that demographic and socioeconomic factors, behavioral factors, as well as self-rated health status are more likely to be significantly related to specificity in at least one of the three chronic diseases, but the direction of association differs according to the disease. Regarding chronological age, the elderly group aged 60–69 years turned out to have a higher awareness of all three diseases in comparison to adults aged 50–59 years and those aged 70 years or older, which proved inconsistent with previous studies 8, 18, 25. Further analysis shows that those aged 60–69 years appear to have more medical check-ups; thus it is plausible that this age group seems to be more aware of the possibility of developing a chronic disease. When it comes to sex differences in awareness of diseases, our study shows that women are more vigilant for HTN, but negligent for hypercholesterolemia, which is in considerably disagreement with previous findings from other countries 18, 21, 25. One of our study's notable findings was that, although females reported a greater prevalence of hypercholesterolemia, they were less aware of their symptoms than men. For all three diseases, rural people were less sensitive than urbanites, but about DM and hypercholesterolemia most especially. A possible explanation is that, when compared with urban elderly, rural dwellers may suffer from a scarcity of information, knowledge, or accessibility to health care services. Our further analysis (data not shown) reveals that the rural elderly may have difficulties interpreting the results of medical check-ups, even though the Korean government provides, free of charge, medical check-ups every 2 years for all Koreans over the age of 40 years, regardless of where they live.
While numerous studies have highlighted a socioeconomic gradient in health, our results also observed that highly educated groups were more sensitive to HTN and hypercholesterolemia. The socioeconomic gap in our study can be attributable to broader inequalities in socioeconomic forces; in particular, evidence shows that being highly educated is closely related to an increase in health-related information and knowledge, which lead to a higher awareness of chronic diseases 2, 16, 19, 25. Income levels are also known to be a relevant factor associated with higher awareness and better management [26], but our study found this link only for hypercholesterolemia. In contrast to previous findings [2], economically inactive groups in our study were more likely to be aware of HTN and hypercholesterolemia. It can be proposed that, more than likely they are spending more time on health management in comparison to the employed who have difficulty in balancing work with family life, for Koreans have the longest workweek of all OECD countries.
In contrast to other studies 17, 27, health-related behavioral factors such as drinking and physical activities were not substantially associated with sensitivity for CVD risk factors. Only smoking influenced an awareness of CVD risk factors on health-related behaviors. In particular, hypercholesterolemia's sensitivity between reported and measured data was significantly lower among past and current smokers than among nonsmokers. In addition to these factors, our findings show that individuals with average and poor self-rated health had higher rate of agreement and of sensitivity for all three CVD risk factors, when compared with those with good self-rated health (e.g., OR in poor self-rated health: 2.07 for HTN, 3.23 for DM, and 2.09 for hypercholesterolemia). This result implies that self-rated health may be a useful indicator of awareness of CVD-related factors.
The limitations of this study are as follows. First, we cannot exclude the possibility of an over- or under-diagnosis of directly measured data. For example, in the case of white coat syndrome for HTN or DM, which is an anxiety-induced false positive reaction 25, 28, the sensitivity level of self-reported data may be decreased. However, since our classification criteria for directly measured data did not cover all diagnosis criteria, except for fasting blood glucose levels, we also used the oral glucose tolerance test for DM diagnosis. Second, physician-related factors such as age or hospital setting were not controlled. The heterogeneity of disease criteria is limited to the results garnered from one country to another. Nevertheless, this study has strengths; it is based on a nationwide, largescale population, sampling scheme with a high response rate (74%), and the results of this study can be generalized to the Korean population as a whole. Moreover, this study's results provide diverse factors including sociodemographic and behavioral factors that influence awareness of CVD risk factors, which can use intervention programs of these diseases.
In conclusion, this study of Korean older adults reveals that reliability and validity of self-reported data for HTN and DM are generally acceptable, but for hypercholesterolemia is relatively poor. Caution is needed when interpreting the results of community studies using self-reported data on CVD risk factors, especially hypercholesterolemia, among adults aged 50 years and older. This result can be employed to find controllable risk factors for cardiovascular management that is specific to subgroups. In general, the elderly aged 60–69 years, who are highly educated, live in rural areas, are not engaged in the labor market, and have “poor” self-rated health are more likely to be aware of their three CVD risk factors. Compared to men, women are more likely to be vigilant of HTN, but less likely to be aware of hypercholesterolemia. Past and current smokers are particularly less sensitive for hypercholesterolemia. Our study's results imply the importance of health intervention programs and highlight the necessity for improving public knowledge and information. This requires further studies to elaborate more on the predictors of sensitivity and their strength of association. The results can also be used to find controllable risk factors specific to subgroups for cardiovascular management.
The authors declare that there is no conflict of interest regarding the publication of this paper.
This research was supported by the National Research Foundation of Korea (NRF-2013R1A1A1076139; NRF-2014S1A3A2035458).
  • 1. Kim E.J., Yoon S.J., Jo M.W.. Measuring the burden of chronic diseases in Korea in 2007. Public Health 127(9): 2013 Sep;806-813. PMID: 23434039.Article
  • 2. Lopez A.D., Mathers C.D., Ezzati M.. Global and regional burden of disease and risk factors, 2001: systematic analysis of population health data. Lancet 367(9524): 2006 May 27;1747-1757. PMID: 16731270.Article
  • 3. Yach D., Hawkes C., Gould C.L.. The global burden of chronic diseases: overcoming impediments to prevention and control. JAMA 291(21): 2004 Jun 2;2616-2622. PMID: 15173153.Article
  • 4. Korea Center for Diseases Control & Prevention . Summary of 2013 Korea National Health and Nutrition Survey (KNHANES) results. 2014, Department of Health and Nutrition Survey. Seoul (Korea).
  • 5. Lee K.S.. Second comprehensive plan on cardio-cerebrovascular diseases (2011–2015). 2011, Ministry of Health and Welfare. Kwachen.
  • 6. Mukattash T.L., Shara M., Jarab A.S.. Public knowledge and awareness of cardiovascular disease and its risk factors: a cross-sectional study of 1000 Jordanians. Int J Pharm Pract 20(6): 2012 Dec;367-376. PMID: 23134095.Article
  • 7. Ueshima H., Sekikawa A., Miura K.. Cardiovascular disease and risk factors in Asia: a selected review. Circulation 118(25): 2008 Dec 16;2702-2709. PMID: 19106393.Article
  • 8. Bowlin S.J., Morrill B.D., Nafziger A.N.. Validity of cardiovascular disease risk factors assessed by telephone survey: the Behavioral Risk Factor Survey. J Clin Epidemiol 46(6): 1993 Jun;561-571. PMID: 8501483.Article
  • 9. Huerta J.M., Tormo M.J., Egea-Caparrós J.M.. Accuracy of self-reported diabetes, hypertension and hyperlipidemia in the adult Spanish population. DINO study findings. Rev Esp Cardiol 62(2): 2009 Feb;143-152. PMID: 19232187.Article
  • 10. Taylor A., Dal Grande E., Gill T.. Comparing self-reported and measured high blood pressure and high cholesterol status using data from a large representative cohort study. Aust N Z J Public Health 34(4): 2010 Aug;394-400. PMID: 20649780.Article
  • 11. Tolonen H., Koponen P., Mindell J.S.. Under-estimation of obesity, hypertension and high cholesterol by self-reported data: comparison of self-reported information and objective measures from health examination surveys. Eur J Public Health 24(6): 2014 Dec;941-948. PMID: 24906846.Article
  • 12. Okura Y., Urban L.H., Mahoney D.W.. Agreement between self-report questionnaires and medical record data was substantial for diabetes, hypertension, myocardial infarction and stroke but not for heart failure. J Clin Epidemiol 57(10): 2004 Oct;1096-1103. PMID: 15528061.Article
  • 13. El Fakiri F., Bruijnzeels M.A., Hoes A.W.. No evidence for marked ethnic differences in accuracy of self-reported diabetes, hypertension, and hypercholesterolemia. J Clin Epidemiol 60(12): 2007 Dec;1271-1279. PMID: 17998082.Article
  • 14. Goldman N., Lin I.F., Weinstein M.. Evaluating the quality of self-reports of hypertension and diabetes. J Clin Epidemiol 56(2): 2003 Feb;148-154. PMID: 12654409.Article
  • 15. Martin L.M., Leff M., Calonge N.. Validation of self-reported chronic conditions and health services in a managed care population. Am J Prev Med 18(3): 2000 Apr;215-218. PMID: 10722987.Article
  • 16. Huang P.Y., Buring J.E., Ridker P.M.. Awareness, accuracy, and predictive validity of self-reported cholesterol in women. J Gen Intern Med 22(5): 2007 May;606-613. PMID: 17370032.Article
  • 17. Jeong J.Y., Choi Y.J., Jang S.N.. Awareness, treatment, and control rates of hypertension and related factors of awareness among middle aged adult and elderly in Chuncheon: Hallym Aging Study (HAS). J Prev Med Public Health 40(4): 2007 Jul;305. -312. [In Korean]. PMID: 17693734.Article
  • 18. McDonald M., Hertz R.P., Unger A.N.. Prevalence, awareness, and management of hypertension, dyslipidemia, and diabetes among United States adults aged 65 and older. J Gerontol A Biol Sci Med Sci 64(2): 2009 Feb;256-263. PMID: 19181717.
  • 19. Song Y., Ma W., Yi X.. Chronic diseases knowledge and related factors among the elderly in Jinan, China. PLoS ONE 8(6): 2013 Jun 24;e68599PMID: 23826394.Article
  • 20. Kweon S., Kim Y., Jang M.J.. Data resource profile: the Korea National Health and Nutrition Examination Survey (KNHANES). Int J Epidemiol 43(1): 2014 Feb;69-77. PMID: 24585853.Article
  • 21. Lee H.Y., Park J.. The Korean Society of Hypertension Guidelines for the Management of Hypertension in 2013: its essentials and key points. Pulse 3:2013 April;21-28. PMID: 26587454.Article
  • 22. Korean Diabetes Association . Diabetes management guidelines. Seoul. 2013, Summary 10p.
  • 23. Shin H.. Edited by Sung J.: Dyslipidemia management guideline. 2nd ed.2009, Korean Society of Lipidology and Atherosclerosis. Seoul: Chungun, 6p.
  • 24. Organisation for Economic Co-operation and Development (OECD) . OECD Economic Outlook. 1997, OECD. Paris (France), p 50.
  • 25. Fagard R.H., Cornelissen V.A.. Incidence of cardiovascular events in white-coat, masked and sustained hypertension versus true normotension: a meta-analysis. J Hypertens 25(11): 2007 Nov;2193-2198. PMID: 17921809.Article
  • 26. Mosca I., Bhuachalla B.N., Kenny R.A.. Explaining significant differences in subjective and objective measures of cardiovascular health: evidence for the socioeconomic gradient in a population-based study. BMC Cardiovasc Disord 13:2013 Aug 30;64PMID: 24119371.Article
  • 27. Ahn S., Smith M.L., Cho J.. Hypertension awareness and associated factors among older Chinese adults. Front Public Health 1:2013 Dec;67PMID: 24350235.Article
  • 28. Campbell L.V., Ashwell S.M., Borkman M.. White coat hyperglycaemia: disparity between diabetes clinic and home blood glucose concentrations. BMJ 305(6863): 1992 Nov 14;1194-1196. PMID: 1467722.Article
Table 1
Discrepancy between reported and measured hypertension, diabetes, and hypercholesterolemia and associated factors in Korean elderly, aged 50 years or more, Fourth Korean National Health Examination and Nutrition Survey 2007–2009.
Hypertension (%)
Diabetes (%)
Hypercholesterolemia (%)
n (%) R M κ Sensitivity Specificity R M κ Sensitivity Specificity R M κ Sensitivity Specificity
Total 7,270 (100) 36.4 48.8 0.72 73.0 98.5 13.6 15.4 0.82 79.3 98.4 11.7 16.7 0.48 46.7 95.4
 Age (y) 50–59 2,674 (36.8) 23.9 36.5 0.67 63.5 98.8 8.7 11.2 0.77 70.7 99.1 13.1 16.8 0.45 46.9 93.7
60–69 2,519 (34.7) 40.5 51.5 0.76 77.5 98.9 17.5 18.3 0.84 85.0 97.6 14.2 18.6 0.53 53.1 94.7
≤ 70 2,077 (28.6) 47.5 61.4 0.69 75.8 97.5 15.2 17.2 0.81 79.3 98.2 6.6 14.2 0.44 36.3 98.3
 Sex Male 3,096 (42.6) 34.4 48.9 0.67 68.3 98.1 14.9 17.0 0.81 78.9 98.2 9.2 10.9 0.47 48.5 95.5
Female 4,174 (57.4) 37.9 48.7 0.76 76.5 98.8 12.7 14.2 0.82 79.8 98.5 13.4 21.0 0.48 46.0 95.2
 Marital status Married 5,426 (74.6) 33.7 45.8 0.71 71.6 98.4 13.4 15.3 0.81 78.7 98.5 11.8 15.6 0.48 48.7 95.0
others 1,844 (25.4) 44.4 57.5 0.73 76.4 99.1 14.2 15.5 0.82 81.1 98.1 11.1 20.0 0.46 42.1 96.6
 Region Urban 4,584 (63.1) 36.4 48.7 0.73 73.6 98.9 14.6 16.3 0.82 80.8 98.3 13.6 18.4 0.50 50.3 94.7
Rural areas 2,686 (37.0) 36.3 49.0 0.70 72.0 98.0 11.8 13.8 0.80 76.5 98.5 8.4 13.7 0.42 38.5 96.4
 Education ≥ College 675 (9.3) 31.0 40.7 0.75 73.8 98.5 10.8 12.2 0.79 76.8 98.3 15.0 17.0 0.54 57.4 93.8
High school 1,279 (17.6) 33.9 46.1 0.71 71.5 98.4 12.1 14.7 0.82 77.1 99.1 15.5 16.3 0.49 55.8 92.3
Middle school 1,194 (16.4) 30.7 42.1 0.73 71.6 99.0 13.2 14.6 0.84 81.6 98.5 12.1 16.2 0.49 49.2 95.0
≤ Elementary school 4,122 (56.7) 39.7 52.9 0.71 73.7 98.4 14.6 16.4 0.81 79.7 98.1 9.8 16.9 0.46 41.6 96.7
 Employment Yes 3,643 (50.1) 30.3 43.8 0.68 67.5 98.6 10.8 12.6 0.79 75.8 98.6 9.9 14.0 0.44 43.1 95.5
No 3,627 (49.9) 42.5 53.8 0.75 77.6 98.4 16.4 18.2 0.83 81.8 98.2 13.4 19.4 0.50 49.3 95.2
 Income High 2,370 (32.6) 31.2 43.8 0.72 70.4 99.3 10.6 12.7 0.79 75.1 98.7 14.1 17.1 0.49 52.1 93.8
Middle 2,581 (35.5) 36.0 47.7 0.73 73.8 98.4 14.2 16.3 0.83 80.3 98.7 12.1 17.8 0.49 47.2 95.5
Low 2,319 (31.9) 42.1 55.1 0.70 74.4 97.6 16.0 17.1 0.81 81.6 97.6 8.7 15.1 0.44 39.8 96.9
 Smoking Never smoker 4,441 (61.1) 37.4 48.9 0.75 75.3 98.9 12.8 14.5 0.82 80.1 98.5 13.2 19.2 0.49 48.3 95.2
Past smoker 1,622 (22.3) 38.4 51.9 0.68 71.5 97.3 16.5 18.6 0.83 81.1 98.3 10.4 12.8 0.46 46.9 95.0
Current smoker 1,207 (16.6) 30.1 44.3 0.67 66.2 98.7 12.5 14.5 0.76 73.7 97.9 7.9 12.9 0.41 37.8 96.6
 Binge drinking No 6,702 (92.2) 36.5 48.5 0.73 74.0 98.7 13.6 15.3 0.82 79.9 98.4 11.7 17.0 0.48 46.7 95.5
Yes 568 (7.8) 34.5 52.5 0.58 62.4 96.3 13.2 16.2 0.77 72.8 98.3 11.1 13.2 0.44 46.7 94.3
 Physical activity Yes 4,267 (58.7) 35.3 47.8 0.72 72.4 98.6 13.3 14.8 0.81 79.5 98.2 11.4 16.1 0.48 47.0 95.5
No 3,003 (41.3) 37.9 50.3 0.72 73.9 98.5 14.1 16.3 0.82 79.1 98.6 12.1 17.5 0.47 46.4 95.2
 Self-rated health Very good/good 2,584 (35.5) 30.5 45.1 0.67 66.3 98.9 8.5 11.3 0.77 69.8 99.3 9.8 15.8 0.43 40.3 96.0
Average 2,148 (29.6) 33.3 46.3 0.70 70.3 98.6 11.4 13.6 0.78 74.0 98.4 11.4 16.3 0.48 46.4 95.4
Poor/very poor 2,538 (34.9) 45.0 54.7 0.77 80.6 98.0 20.7 21.1 0.85 87.5 97.2 13.8 18.0 0.52 52.6 94.8

M = measured; R = reported.

Table 2
Results from multivariate logistic analysis of sensitivity.
n = 3,547 (%) OR 95% CI n = 1,118 (%) OR 95% CI n = 1,214 (%) OR 95% CI
Age (y) 50–59 975 27.5 Ref 300 26.8 Ref 450 37.1 Ref
60–69 1,297 36.6 1.92 1.57–2.35 460 31.1 2.30 1.56–3.39 469 38.6 1.45 1.09–1.94
≤ 70 1,275 36.0 1.65 1.31–2.08 358 32.0 1.47 0.95–2.28 295 24.3 0.82 0.57–1.19
Sex Male 1,515 42.7 Ref 525 47.0 Ref 336 27.7 Ref
Female 2,032 57.3 1.37 1.07–1.74 593 53.0 0.74 0.46–1.19 878 72.3 0.60 0.39–0.93
Marital status Married 2,486 70.1 Ref 832 74.4 Ref 846 69.7 Ref
others 1,061 29.9 0.99 0.82–1.21 286 25.6 1.11 0.75–1.64 368 30.3 0.95 0.71–1.27
Region Urban 2,231 62.9 Ref 748 66.9 Ref 845 69.6 Ref
Rural areas 1,316 37.1 0.93 0.78–1.10 370 33.1 0.70 0.50–0.99 369 30.4 0.71 0.54–0.93
Education ≥ College 275 7.8 Ref 82 7.3 Ref 115 9.5 Ref
High school 590 16.6 0.78 0.56–1.09 188 16.8 0.86 0.45–1.64 208 17.1 0.93 0.58–1.50
Middle school 503 14.2 0.69 0.49–0.98 174 15.6 1.01 0.51–1.99 193 15.9 0.71 0.43–1.17
≤ Elementary school 2,179 61.4 0.57 0.41–0.79 674 60.3 0.76 0.40–1.43 698 57.5 0.55 0.35–0.88
Employment Yes 1,595 45.0 Ref 459 41.1 Ref 510 42.0 Ref
No 1,952 55.0 1.29 1.08–1.53 659 58.9 1.04 0.73–1.47 704 58.0 1.28 0.98–1.67
Income High 1,038 29.3 Ref 301 26.9 Ref 405 33.4 Ref
Middle 1,231 34.7 1.10 0.90–1.34 421 37.7 1.30 0.88–1.91 460 37.9 0.89 0.66–1.19
Low 1,278 36.0 0.97 0.78–1.20 396 35.4 1.26 0.82–1.93 349 28.8 0.71 0.51–1.00
Smoking Never smoker 2,171 61.2 Ref 642 57.4 Ref 851 70.1 Ref
Past smoker 841 23.7 1.03 0.81–1.31 301 26.9 0.89 0.56–1.42 207 17.1 0.60 0.38–0.93
Current smoker 535 15.1 0.87 0.67–1.12 175 15.7 0.63 0.38–1.03 156 12.9 0.41 0.26–0.66
Binge drinking No 3,249 91.6 Ref 1,026 91.8 Ref 1,139 93.8 Ref
Yes 298 8.4 0.85 0.65–1.11 92 8.2 0.82 0.48–1.42 75 6.2 1.09 0.64–1.84
Physical activity Yes 2,038 57.5 Ref 630 56.4 Ref 688 56.7 Ref
No 1,509 42.5 0.97 0.83–1.14 488 43.7 0.92 0.68–1.26 526 43.3 0.99 0.78–1.26
Self-rated health Very good/good 1,164 32.8 Ref 291 26.0 Ref 409 33.7 Ref
Average 994 28.0 1.25 1.04–1.51 292 26.1 1.24 0.85–1.81 349 28.8 1.34 0.99–1.82
Poor/very poor 1,389 39.2 2.07 1.71–2.50 535 47.9 3.23 2.21–4.72 456 37.6 2.09 1.55–2.80

CI = confidence interval; OR = odds ratio; Ref = reference category.

Multiple logistic analysis include all variables—demographic (age, sex, marital status, region), socioeconomic (education, employment status, income), health behaviors (smoking binge drinking, physical activity), and self-rated health.

Figure & Data



    Citations to this article as recorded by  
    • Vascular risk factors and stroke risk across the life span: A population-representative study of half a million people
      Raed A Joundi, Scott B Patten, Jeanne VA Williams, Eric E Smith
      International Journal of Stroke.2022; : 174749302110706.     CrossRef
    • The difference between hypertension determined by self-report versus examination in the adult population of the USA: Continuous NHANES 1999–2016
      Leanna Delhey, Catherine Shoults, Kemmian Johnson, Mohammed Orloff, Mohammed F Faramawi, Robert Delongchamp
      Journal of Public Health.2021; 43(2): 316.     CrossRef
    • Bi-directional association between allergic rhinitis and diabetes mellitus from the national representative data of South Korea
      Tae Kyung Lee, Ye Jin Jeon, Sun Jae Jung
      Scientific Reports.2021;[Epub]     CrossRef
    • Cut-off points of anthropometric markers associated with hypertension and diabetes in Peru: Demographic and Health Survey 2018
      Akram Hernández-Vásquez, Diego Azañedo, Rodrigo Vargas-Fernández, Juan Pablo Aparco, Raul Martín Chaparro, Marilina Santero
      Public Health Nutrition.2021; 24(4): 611.     CrossRef
    • Examining elevated blood pressure and the effects of diabetes self-management education on blood pressure among a sample of Marshallese with type 2 diabetes in Arkansas
      Pearl A. McElfish, Christopher R. Long, Zoran Bursac, Aaron J. Scott, Harish E. Chatrathi, Ka‘imi A. Sinclair, Nirav Nagarsheth, Mikaila Calcagni, Jay Patolia, Marie-Rachelle Narcisse, Solveig A. Cunningham
      PLOS ONE.2021; 16(4): e0250489.     CrossRef
    • Self-Reported Modifiable Risk Factors of Cardiovascular Disease among Older Adults in Malaysia: A Cross-Sectional Study of Prevalence and Clustering
      Ying Ying Chan, Norhafizah Sahril, Muhammad Solihin Rezali, Lim Kuang Kuay, Azli Baharudin, Mohamad Aznuddin Abd Razak, Mohd Shaiful Azlan Kassim, Muhammad Fadhli Mohd Yusoff, Mohd Azahadi Omar, Noor Ani Ahmad
      International Journal of Environmental Research an.2021; 18(15): 7941.     CrossRef
    • Social engagement and allostatic load mediate between adverse childhood experiences and multimorbidity in mid to late adulthood: the Canadian Longitudinal Study on Aging
      Leslie Atkinson, Divya Joshi, Parminder Raina, Lauren E. Griffith, Harriet MacMillan, Andrea Gonzalez
      Psychological Medicine.2021; : 1.     CrossRef
    • Usefulness of Relative Handgrip Strength as a Simple Indicator of Cardiovascular Risk in Middle-Aged Koreans
      Won Bin Kim, Jun-Bean Park, Yong-Jin Kim
      The American Journal of the Medical Sciences.2021; 362(5): 486.     CrossRef
    • Agreement Between Self-Reported Information and Administrative Data on Comorbidities, Imaging and Treatment in Denmark – A Validation Study of 38,745 Patients with Knee or Hip Osteoarthritis
      Halit Selçuk, Ewa M Roos, Dorte T Grønne, Martin T Ernst, Søren T Skou
      Clinical Epidemiology.2021; Volume 13: 779.     CrossRef
    • Validity of self-reported diabetes varies with sociodemographic charecteristics: Example from Iran
      Mehdi Moradinazar, Yahya Pasdar, Farid Najafi, Ebrahim Shakiba, Behrooz Hamzeh, Mehnoosh Samadi, Maryam Mirzaei, Annette J. Dobson
      Clinical Epidemiology and Global Health.2020; 8(1): 70.     CrossRef
    • Dietary patterns and cardiovascular disease in Greek adults: The Hellenic National Nutrition and Health Survey (HNNHS)
      Dimitra Karageorgou, Emmanouella Magriplis, Ioanna Bakogianni, Anastasia V. Mitsopoulou, Ioannis Dimakopoulos, Renata Micha, George Michas, Triantafyllia Ntouroupi, Sophia M. Tsaniklidou, Konstantina Argyri, Michail Chourdakis, Demosthenes B. Panagiotakos
      Nutrition, Metabolism and Cardiovascular Diseases.2020; 30(2): 201.     CrossRef
    • Accuracy of self‐reported hypertension: Effect of age, gender, and history of alcohol dependence
      Jeannette L. Wellman, Brian Holmes, Shirley Y. Hill
      The Journal of Clinical Hypertension.2020; 22(5): 842.     CrossRef
    • Spatial Epidemiology of Diabetes and Tuberculosis in India
      Daniel J Corsi
      JAMA Network Open.2020; 3(5): e203892.     CrossRef
    • Credibility of self-reported health parameters in elderly population
      Roi Amster, Iris Reychav, Roger McHaney, Lin Zhu, Joseph Azuri
      Primary Health Care Research & Development.2020;[Epub]     CrossRef
    • What are the illnesses associated with frailty in community-dwelling older adults: the Korean Frailty and Aging Cohort Study
      Sunyoung Kim, Hee-Won Jung, Chang Won Won
      The Korean Journal of Internal Medicine.2020; 35(4): 1004.     CrossRef
    • The Demographic Representativeness and Health Outcomes of Digital Health Station Users: Longitudinal Study
      Leah Flitcroft, Won Sun Chen, Denny Meyer
      Journal of Medical Internet Research.2020; 22(6): e14977.     CrossRef
    • Self‐reported diabetes, hypertension and hypercholesterolemia among older persons in Malaysia
      Nur Liana Ab Majid, Wan Shakira Rodzlan Hasani, Halizah Mat Rifin, Tania Gayle Robert Lourdes, Miaw Yn Jane Ling, Thamil Arasu Saminanthan, Hasimah Ismail, Ahzairin Ahmad, Muhammad Fadhli Mohd Yusoff
      Geriatrics & Gerontology International.2020; 20(S2): 79.     CrossRef
    • Association between number of medications used and nutritional markers among elderly persons with chronic diseases: National Health Survey (2013)
      Isabel Cristina Bento, Mary Anne Nascimento Souza, Sérgio Viana Peixoto
      Revista Brasileira de Geriatria e Gerontologia.2019;[Epub]     CrossRef
    • Prevalence, determinants and association of unawareness of diabetes, hypertension and hypercholesterolemia with poor disease control in a multi-ethnic Asian population without cardiovascular disease
      Ryan E. K. Man, Alvin Hong Wei Gan, Eva K. Fenwick, Alfred Tau Liang Gan, Preeti Gupta, Charumathi Sabanayagam, Nicholas Tan, Kah Hie Wong, Tien Yin Wong, Ching-Yu Cheng, Ecosse L. Lamoureux
      Population Health Metrics.2019;[Epub]     CrossRef
    • Differential self-report error by socioeconomic status in hypertension and hypercholesterolemia: INSEF 2015 study
      Irina Kislaya, Hanna Tolonen, Ana Paula Rodrigues, Marta Barreto, Ana Paula Gil, Vânia Gaio, Sónia Namorado, Ana João Santos, Carlos Matias Dias, Baltazar Nunes
      European Journal of Public Health.2019; 29(2): 273.     CrossRef
    • Validity of Self-reported Hypertension and Factors Related to Discordance Between Self-reported and Objectively Measured Hypertension: Evidence From a Cohort Study in Iran
      Farid Najafi, Yahya Pasdar, Ebrahim Shakiba, Behrooz Hamzeh, Mitra Darbandi, Mehdi Moradinazar, Jafar Navabi, Bita Anvari, Mohammad Reza Saidi, Shahrzad Bazargan-Hejazi
      Journal of Preventive Medicine and Public Health.2019; 52(2): 131.     CrossRef
    • Public health monitoring of hypertension, diabetes and elevated cholesterol: comparison of different data sources
      Laura Paalanen, Päivikki Koponen, Tiina Laatikainen, Hanna Tolonen
      European Journal of Public Health.2018; 28(4): 754.     CrossRef
    • Low potassium and high sodium intakes: a double health threat to Cape Verdeans
      Daniela Alves, Zélia Santos, Miguel Amado, Isabel Craveiro, António Pedro Delgado, Artur Correia, Luzia Gonçalves
      BMC Public Health.2018;[Epub]     CrossRef
    • Diet-Related Risk Factors for Incident Hypertension During an 11-Year Follow-Up: The Korean Genome Epidemiology Study
      Hye Lee, Hyesook Park
      Nutrients.2018; 10(8): 1077.     CrossRef
    • Evaluation of the association between the number of natural teeth and anemia among Korean adults using nationally representative data
      Kyungdo Han, Jun-Beom Park
      Journal of Periodontology.2018; 89(10): 1184.     CrossRef
    • Diabetes Fact Sheet in Korea, 2016: An Appraisal of Current Status
      Jong Chul Won, Jae Hyuk Lee, Jae Hyeon Kim, Eun Seok Kang, Kyu Chang Won, Dae Jung Kim, Moon-Kyu Lee
      Diabetes & Metabolism Journal.2018; 42(5): 415.     CrossRef
    • Accuracy of self-reported hypertension
      Vivian S.S. Gonçalves, Keitty R.C. Andrade, Kenia M.B. Carvalho, Marcus T. Silva, Mauricio G. Pereira, Tais F. Galvao
      Journal of Hypertension.2018; 36(5): 970.     CrossRef
    • Trends in Determinants of Hypercholesterolemia among Chinese Adults between 2002 and 2012: Results from the National Nutrition Survey
      Peng-kun Song, Hong Li, Qing-qing Man, Shan-shan Jia, Li-xiang Li, Jian Zhang
      Nutrients.2017; 9(3): 279.     CrossRef
    • High Level Physical Activity and Prevalence of Cardiovascular Disease Using the Korea National Health and Nutrition Examination Survey Data, 2007-2013
      Kyounghoon Park, Byung-Joo Park
      Journal of Preventive Medicine and Public Health.2017; 50(5): 320.     CrossRef
    • Association between underweight and tooth loss among Korean adults
      In-Seok Song, Kyungdo Han, Jae-Jun Ryu, Jun-Beom Park
      Scientific Reports.2017;[Epub]     CrossRef
    • Potential use of telephone-based survey for non-communicable disease surveillance in Sri Lanka
      H. M. M. Herath, N. P. Weerasinghe, T. P. Weerarathna, A. Hemantha, A. Amarathunga
      BMC Public Health.2017;[Epub]     CrossRef
    • Level of Agreement and Factors Associated With Discrepancies Between Nationwide Medical History Questionnaires and Hospital Claims Data
      Yeon-Yong Kim, Jong Heon Park, Hee-Jin Kang, Eun Joo Lee, Seongjun Ha, Soon-Ae Shin
      Journal of Preventive Medicine and Public Health.2017; 50(5): 294.     CrossRef
    • Evaluation of Self-assessment in Cardiovascular Diseases Among Korean Older Population
      Hae-Wol Cho, Chaeshin Chu
      Osong Public Health and Research Perspectives.2016; 7(2): 75.     CrossRef

    PHRP : Osong Public Health and Research Perspectives