- Email: [email protected]
Health expectancies in the older Thai population
Archives of Gerontology and Geriatrics 53 (2011) 3–7
Contents lists available at ScienceDirect
Archives of Gerontology and Geriatrics journal homepage: www.elsevier.com/locate/archger
Health expectancies in the older Thai population Weerasak Muangpaisan a,b,*, Prasert Assantachai a, Somboon Intalapaporn a, Kathryn Richardson b, Carol Brayne b a b
Department of Preventive and Social Medicine, Faculty of Medicine, Siriraj Hospital, Mahidol University, Bangkok 10700, Thailand Institute of Public Health, University of Cambridge, Forvie Site, Robinson Way, Cambridge CB2 0SR, UK
A R T I C L E I N F O
A B S T R A C T
Article history: Received 16 October 2009 Received in revised form 30 March 2010 Accepted 1 April 2010 Available online 8 June 2010
This study aims to investigate health expectancies in five domains: cognitive health, psychological health, physical health, functional ability and self-perceived global health (SPGH) in the older Thai population. There are few studies reporting health expectancies in multidimensional health domains, most of which reported only one health dimension. The dataset used was from the Bangkok Longitudinal Study by Siriraj Hospital for the Older Men and Women (BLOSSOM), which is a community cohort study in Bangkok, Thailand. This analysis is based on the cross-sectional data in the year 2005–2006 and includes 5936 participants aged 50 years and over from community settings within six suburban areas in Bangkok. The study found that women had a longer total life expectancy (LE), but had shorter cognitive impairment-free (CIFLE), physical illness-free (PHILE) and disability-free (DIFLE) LEs, than men. However, there was no difference between the life expectancies for living with good SPGH in men and in women. Differences in health expectations might explain this finding. Health promotion and disease prevention should be initiated at a younger age and should target all health domains. ß 2010 Elsevier Ireland Ltd. All rights reserved.
Keywords: Healthy life expectancy Physical illness Cognitive impairment Psychological Disability Self-perceived global health
1. Introduction LE has now increased globally. However, living longer does not necessarily mean living with good quality of life (QoL). Healthy life expectancy (HLE) is more important. A measure that takes into account survival in and out of good health is therefore of great value in assessing and comparing health within and between geographical or social groups. LE is composed of estimated lengths of time spent in different states of health until death. These lengths of time in different states of health are health expectancies and they combine morbidity and mortality into a single composite indicator (WHO, 2008). There are two commonly used methods to calculate the health expectancy: the Sullivan method, which uses the observed agespecific prevalences of health states and the total person years lived at a particular age (Sullivan, 1971), and the multistate method, which uses a direct calculation of person years lived in the health state and requires longitudinal data to provide transition rates between health states. The latter method can take into account the reversible transition between good health and one or more disability (Jagger and Reyes-Frausto, 2002).
* Corresponding author at: Department of Preventive and Social Medicine, Faculty of Medicine, Siriraj Hospital, Mahidol University, Bangkok 10700, Thailand. Tel.: +66 2 419 8388; fax: +66 2 411 5034. E-mail address: [email protected] (W. Muangpaisan). 0167-4943/$ – see front matter ß 2010 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.archger.2010.05.012
There are few studies reporting health expectancies in multidimensional health domains. Most of the reported health expectancies show only one dimension such as disability-free life expectancies (DFLE) or cognitive-free life expectancy (CFLE). An example of a study reporting multiple health LEs is the Medical Research Council Cognitive Function and Ageing Study (MRC CFAS) which reported health expectancies in three dimensions: cognitive impairment, physical illness and functional disability (Brayne et al., 2001). SPGH is widely used to reflect the individual’s perception of health and has reported to be related to socio-economic status (O’Donnell and Propper, 1991). The patient’s subjective experience of their own health status also helps predict the risk of future mortality (Lenzen et al., 2007). SPGH can be used to calculate the health expectancy free of poor perceived health. LE is increasing over time in Asia and around the world. However, information is limited on the states of health associated with longer life. Longer LE does not necessarily imply longer periods in states of good health and there is increasing attention to quality of extended life. Research on HLE is needed for comparing the health of older populations in countries at different levels of development and for showing the dependency ratio in this population. HLE can be used to predict future needs, assess national health policy, monitor trend and inequalities. Obtaining life tables is crucial in Asian and less developed countries to enable the public policy maker to shape health plan strategies and better allocate limited resources. The objective of this paper was to
4
W. Muangpaisan et al. / Archives of Gerontology and Geriatrics 53 (2011) 3–7
investigate the health expectancies in five domains: cognitive health, psychological health, physical health, functional ability and SPGH. 2. Subjects and methods 2.1. Study design The BLOSSOM is a community cohort study in Bangkok, Thailand. The study included 5936 participants aged 50 years and over from community settings within six suburban areas in Bangkok from 2005-up to present (June 2009). The objectives of the project are to gain an understanding of the magnitude of health problems of the older population in urban areas, obtain the health determinants and predictors of the common diseases in the older population and develop strategies to promote health and prevent common diseases. Community leaders were requested to ask eligible people in their community to attend a mobile unit at a hub of the community. The hub of each community varied from a temple, school, conference hall, or leisure area. The participation rate was difficult to calculate, as the number of family members recorded in the census and by the community leaders was frequently incorrect as a result of unregistered migration. All individuals aged 50 years and over in the sampling areas were invited to participate in the study via the community leaders and community nurses of the Primary Care unit. The exclusion criteria were: (1) age under 50 years and (2) inability to answer questions independently. 2.2. Measurements This analysis considered first year data only from part of the face-to-face structured-interview which included baseline characteristics of the population, the Thai Mental State Examination (TMSE) score, the psychological domain of the Thai version of the World Health Organization Quality of Life (WHOQoL-BREF) questionnaire, the activities of daily living (ADLs) score which was translated from the ADL tool used in the Survey in Europe on Nutrition and the Elderly, a Concerted Action (SENECA), selfreported physical illnesses, special ailments (hearing problems and visual problems) and SPGH (Osler et al., 1991; Mahatnirundkul et al., 1998). SPGH was measured using the question ‘‘How satisfied are you with your health?’’, which is one of the questions in WHOQoL-BREF. The answer was rated into five levels; very dissatisfied, dissatisfied, neither satisfied nor dissatisfied, satisfied and very satisfied. 2.3. Health domains The prevalence of impairment in each health domain and poor SPGH were used to calculate the health expectancies. The cognitive health domain was based on cases of moderate to severe cognitive impairment only, as estimations of mild cognitive impairment are rather inaccurate and fluctuate (Jagger et al., 1998). Thus, a TMSE score of less than 18 was used for this purpose to make results comparable to the robust cutoff for MMSE. The psychological health domain was defined as having a poor psychological QoL score. Physical illness was based on the self-reported physical illness, and hearing and visual problems uncorrected by eyeglasses. Impairment (unable to complete, can do only with help, and can do with difficulty but without help) in self-care ability was used to define the self-care domain. Likewise, any impairment in any of the 16-item ADL scale was used for the prevalence of functional impairment to calculate the DFLE. Finally, SPGH of very dissatisfied and dissatisfied were used to define poor SPGH.
From 5936 participants 839 were excluded from the analysis; 20 for being under 50 years old and 819 for having incomplete data used in the analysis. The source of most of the missing data was from the health questionnaires (n = 513, 8.7%) and the baseline variables (n = 306, 5.2%). Age, sex, marital status and income distributions were similar in participants with missing and complete data. However, among the available data for educational level, a greater percentage of subjects with missing data had no formal education compared to subjects with complete data (15.8% vs. 8.5%). 2.4. Statistical analysis Frequencies and proportions of the main baseline characteristics were calculated and stratified by sex. Age was categorized into four groups: 50–59, 60–69, 70–79 and 80 years old and over. The independent-samples t-test was used to compare the mean age across men and women. The x2-test was used to compare the categorical variables of age, sex, educational level, marital status, living status, and income across men and women. Health life expectancies were calculated using the method developed by Sullivan (1971). The abridged life table for the Thai population, by sex, developed by the WHO 2006 was used due to the small number of participants in each year of age and to be comparable with similar studies (World Health Organization, 2008). For this study, only LE for those aged 50 years and over were relevant. Due to the lack of sex- and age-specific total population and total deaths in the WHO life table for the population aged over 80 years old, data from the Department of Local Administration, Ministry of Interior, Thailand (2006) were used for the calculation in that age group. An age interval of 5 years was set for each age group, except for the final age interval, which was assumed to have length 10 years.
3. Results The baseline characteristics of the subjects with complete data are summarized in Table 1. There were 1357 men (26.6%) and 3740 women (73.4%). Men were significantly older (Table 1). Women had a lower educational level with 74.7% studying for 4 years or less compared to 54.9% in men (p < 0.001). In terms of marital status, men were more likely to be married (81.8% vs. 46.9%), while
Table 1 Baseline characteristics of the participants with complete data. N (%) unless otherwise indicated, mean S.D., or n (%). Baseline characteristics
Men
Women
Total
p
Number Age, years
1357 64.1 8.2
3740 62.1 8.5
5097 62.6 8.5
<0.001
<0.001
Education Did not study 1–4 years Primary school Secondary school Bachelor degree/higher
39 705 105 391 117
(2.9) (52.0) (7.7) (28.8) (8.6)
395 2398 212 580 155
(10.6) (64.1) (5.7) (15.5) (4.1)
434 3103 317 971 272
(8.5) (60.9) (6.2) (19.1) (5.3)
Marital status Single Married Widowed Divorced or separated Living alone
59 1110 122 66 46
(4.3) (81.8) (9.0) (4.9) (3.4)
408 1753 1160 419 187
(10.9) (46.9) (31.0) (11.2) (5.0)
467 2863 1282 485 233
(9.2) (56.2) (25.2) (9.5) (4.6)
Income Inadequate Adequate Saving
0.02 <0.001
226 (16.7) 923 (68.0) 208 (15.3)
821 (22.0) 2431 (65.0) 488 (13.0)
1047 (20.5) 3354 (65.8) 696 (13.7)
<0.001
W. Muangpaisan et al. / Archives of Gerontology and Geriatrics 53 (2011) 3–7
5
Table 2 Total LE together with the various sub-categories of it in absolute values and % of LE. Age groups
Males LEa
50–54
55–59
60–64
65–69
70–74
75–79
80+
26.9
23.2
19.7
16.3
13.2
10.4
9.6
b
CIFLE %CIFLE
25.0 92.9
21.3 91.8
17.7 89.8
14.1 86.5
10.8 81.8
7.8 75.0
7.5 78.1
PIFLEc %PIFLE
25.5 94.8
21.8 94.0
18.2 92.4
14.8 90.8
11.6 87.9
8.7 83.7
9.1 94.8
PHILEd %PHILE
4.4 16.4
3.2 13.8
2.3 11.7
1.7 10.4
1.2 9.1
0.7 6.7
0.6 6.3
SCLEe %SCLE
21.8 81.0
18.1 78.0
14.7 74.6
11.3 69.3
8.4 63.6
5.5 52.9
4.6 47.9
DFLEf %DFLE
16.5 61.3
13.2 56.9
10.0 50.8
7.2 44.2
4.8 36.4
2.6 25.0
2.1 21.9
HDLEg %HDLE
3.8 14.1
2.7 11.6
1.9 9.6
1.3 8.0
0.8 6.1
0.4 3.8
0.2 2.1
SPGHLEh %SPGHLE
22.5 83.6
19.2 82.8
16.0 81.2
13.0 79.8
10.1 76.5
7.6 73.1
8.1 84.4
Females LEa
29.4
25.3
21.4
17.7
14.2
11.0
9.6
b
CIFLE %CIFLE
21.9 74.5
17.7 70.0
13.6 63.6
14.1 79.7
10.5 73.9
7.4 67.3
7.0 72.9
PIFLEc %PIFLE
27.1 92.2
23.0 90.9
19.1 89.3
15.4 87.0
11.8 83.1
8.4 76.4
8.9 92.7
PHILEd %PHILE
3.4 11.6
2.5 9.9
1.8 8.4
1.2 6.8
0.8 5.6
0.5 4.5
0.3 3.1
SCLEe %SCLE
20.4 69.4
16.5 65.2
12.8 59.8
9.3 52.5
6.4 45.1
3.9 35.5
3.2 33.3
DFLEf %DFLE
11.6 39.5
8.6 34.0
6.0 28.0
3.9 22.0
2.2 15.5
1.1 10.0
0.8 8.3
HDLEg %HDLE
2.1 7.1
1.4 5.5
0.9 4.2
0.5 2.8
0.3 2.1
0.2 1.8
0.1 1.0
23.7 80.6
20.3 80.2
17.0 79.4
13.7 77.4
10.6 74.6
7.7 70.0
8.1 84.4
SPGHLEh %SPGHLE
Notes: %Values are expressed as % of LE. a LE in years. b TMSE score <18. c Poor psychological QoL. d Presence of physical illness. e Self-care life expectancy in years. f Functional impairment-free life expectancy. g Any of ill-health in cognitive, psychological, physical health and functional impairment. h Healthy life expectancies by SPGH of ‘‘being in good health’’ (very satisfied, satisfied and neither dissatisfied nor satisfied).
women were more likely to be single (10.9% vs. 4.3%), widowed (31.0% vs. 9.0%), divorced or separated (11.2% vs. 4.9%). Table 2 and Fig. 1 show the total life expectancies with each type of health domain and with any health domain. Overall, women had a longer LE than men in every age group except for those aged 80+ years. This is due to the common data being used for this age group as the data was not available by sex. At the age of 50–54 years, men had a further LE of 26.9 years and women had 29.4 years. At this age, men could expect to spend 25.0, 25.5, 4.4, 21.8, 16.5, and 3.8 years free of cognitive impairment, psychological impairment, physical illness, self-care functional impairment, total functional impairment, and any impairment of health areas, respectively. Although women had a longer total LE, they had fewer years than men free from cognitive impairment, physical illness, functional impairment and any impairment in health areas. The psychological impairment-free LE in women was longer than
Fig. 1. ELE in (a) men and (b) women.
men in the age group 50–74 years and shorter than men in the age groups 75–79 years and 80+ years. The most striking feature is that the physical illness-free LE for both sexes was very short compared to other domains. Even in the younger age group (50–54 years), men had 4.4 years and women had 3.4 years of LE free from physical illness. When each domain was considered together, the ill-health free LE was only 3.5 years in men and 1.8 years in women between the ages of 50–54 years. This was because many of the younger participants had self-reported physical illness, hearing, and visual problems although they were free from impairment in the other domains. Although women had shorter health expectancies in all health states, they did not have a marked difference in LE with good SPGH. The percentage of years free from poor SPGH declined slightly with increasing age from 83.6% at the age of 50–54 years to 76.5% at age 75–79 years in men and from 80.6 to 70.0% in women, respectively. At the age of 80 years and over, the percentage increased to 84.4% for both men and women, which is likely due to the different data source used to construct the life table for this age group. 4. Discussion The life table constructed in this study emphasises that, apart from the psychological health expectancies in the young old (50– 69 years), women had disease-free life expectancies shorter than men in every health domain. However, both men and women had similar proportions of LE appreciating ‘‘being in good health’’. Women might accept some limitations of health status better than men. The study population demographic demonstrates the traditional Thai belief in the past that education is less important in women than men. Previous studies show that low socio-economic status including educational attainment and social class is a
6
W. Muangpaisan et al. / Archives of Gerontology and Geriatrics 53 (2011) 3–7
significant risk factor for negative health outcomes such as cognitive impairment, functional impairment, some specific diseases, and mortality (Letenneur et al., 1999; Melzer et al., 2000; Van Lenthe et al., 2004). The marked difference in reports from women in this study could be linked to education. Attention to ensuring greater appropriate education for both sexes may lead to improvements in health in the longer term. The healthy LEs measured in this study highlight early exposure to health problems particularly physical health in Thailand. As health promotion and prevention strategies are developed, the maintenance of good health states and care-plans which begin at an earlier age, before people become elderly, are needed. The mental health expectancy may be calculated, based on two different classes of indicators: (1) disease-specific life expectancies such as dementia-free LE and depression-free LE and (2) the functional impact of mental ill-health, such as LE free of social isolation or LE in good perceived mental health (Jagger et al., 1998). The first type of calculation needs clinical diagnosis using the International Classification of Disease (ICD) or the Diagnostic and Statistical Manual of Mental Disorders (DSM) to allow temporal stability and cross-population comparability. In this paper, the ‘‘no cognitive impairment’’ state was used as there was no clinical diagnosis of dementia in the study. However, it still provides the information as a symptom-free LE. Likewise, the choice of using ‘‘poor psychological QoL’’ as an indicator for psychological health in the life table may be less comparable to other countries. However, the ultimate goal of the health expectancies are to give outcome measures for evaluating population health which provide status relative to LE and take into account QoL. A further strength of this life table is that it provides multiple dimensions of health in the older populations. Most reported health expectancies show only one dimension such as DFLE or cognitive-free LE. The life table in this paper provides mental health expectancies (cognitive and psychological health), physical health expectancies, DFLE and SPGH expectancy. It is the first life table that combines every dimension of health in the calculation. As it is independent of the size of the populations and of their age structure, health expectancies can be used for the direct comparison of different groups that make up populations and also can contribute to public health policies. The limitations of the life table in this paper using Sullivan’s method should be taken into account. First, the calculation of health expectancies using prevalence data does not incorporate transition rates in and out of the health states. Cases in one health state can recover, and this is relatively common (Bowling and Grundy, 1997; Mendes de Leon et al., 1997). Reversible causes of cognitive impairment account for 0–23% for partial and 0–10% for full reversal (Weytingh et al., 1995). Also, recovery from disability is well-recognized especially in acute medical conditions such as stroke. With cross-sectional surveys transition rates cannot be estimated. Second, the prevalence of health states came from the community survey only and did not include those in institutions. Life tables for a population should include representative sampling from the total population to avoid bias in estimating health problems. Also, there could be a volunteer effect as people who participated in the BLOSSOM might be healthier than those who did not. Third, the definitions of health states in this study are different from other studies. Different health state definitions are common across studies and make direct comparison of health expectancies across studies difficult. However, the life table provides useful public health information. Finally, the prevalence of the health states in this study came from the cross-sectional survey of one study, not a national database. The generalizability of the studied population to the national level must be considered. There have been only three studies publishing life tables of health expectancies in Thailand (Jitapunkul and Chayovan, 2000;
Jitapunkul et al., 2001, 2003). The first study estimated HLE comparing the status of ‘‘being in good health’’ in older Thai people between 1986 and 1995. This study was based on two cross-sectional surveys, used self-perceived health as an assessment tool and calculated HLE by the Sullivan’s method. However, the questions used in both surveys were different and the scaling method of the scores was also different (Jitapunkul and Chayovan, 2000). This could lead to varied prevalence estimates of ‘‘being in good health’’. The second study estimated active life expectancies in 723 participants aged 60 years and over in central region of Thailand (Jitapunkul et al., 2001). This study used the modified Barthel ADL index and Chula ADL index to survey the prevalence of disability. The limitations were mainly due to a small number of participants in each age strata which was likely to make the estimates imprecise. Another study by Jitapunkul et al. (2003) estimated the DFLE by Sullivan’s method in Thai elderly people aged 60 years and over. There was a lack of information about mental health problems, physical illnesses and SPGH in this study. Comparison of health expectancies across studies is difficult as a result of different methods in developing the life table, different definitions of health states and also secular trends and geographic variations. Direct comparisons of the DFLE between the studies by Jitapunkul et al. (2003) and the BLOSSOM are difficult as the definitions of disability used and the measurement scales used to collect the prevalence data differ. However, both studies used the same face-to-face interview method and similar study populations of the older population in the community. The population in the study by Jitapunkul et al. (2003) also lived in more rural areas, whereas the population in the BLOSSOM study lived in more urban areas. Nevertheless, the main finding in the BLOSSOM study is consistent with the other studies in that men had better DFLE, whereas the finding from the studies by Jitapunkul et al. (2003) showed the opposite. Differences between the findings in the studies by Jitapunkul et al. (2003) and the BLOSSOM study might be due to the different scales used, different data collection processes, or changes in the country’s demographic profile over this 10-year period. Health expectancies can also be calculated using multistate modelling if longitudinal data is available. The health expectancies in the Melton Mowbray Ageing Project (MMAP) were calculated from longitudinal data using this method (Sauvaget et al., 2001). In multistate modelling, both the probability of becoming impaired in health states (decrement) and the probability of recovering from the impairment (increment) are used in the calculations. Thus, estimations are more accurate as transitions between health states (e.g., recovery) are common. The definition used for active life expectancy in the MMAP was independency in all of the seven ADLs qualified an individual to be active. The scale used to define disability in the BLOSSOM study is similar to the MMAP in that the individual had to be independent in all items to qualify as not impaired. However, the MRC CFAS health expectancy estimations by Brayne et al. (2001) used less strict criteria. Cross-sectional data were used to calculate a life table, hence transition states were not possible, as in this study. The MRC CFAS had strength in including institutional residents which makes it more generalizable to the whole population. The BLOSSOM study did not collect data from institutional carehomes; however, the proportion of the older population who reside in institutional care-homes is much smaller than compared to Western countries. The MMAP reported two domains of health expectancies (cognitive impairment-free life expectancy: CIFLE, and active LE), the MRC CFAS reported three domains (CIFLE, active LE, and physical illness-free), whereas this study reports four domains of health (adding the psychological domain) and one SPGH.
W. Muangpaisan et al. / Archives of Gerontology and Geriatrics 53 (2011) 3–7
5. Conclusions and recommendations Women had poorer health states and a shorter LE in good health despite longer total LEs. Despite this, women had similar expectancies in good SPGH to men. Life expectancies free from physical illness were short even in the population aged 50–54 years old. This provides further evidence of the need for health promotion and disease prevention to be initiated at younger ages targeting all health domains. Conflict of interest statement None. Acknowledgement This study was supported by the Thai Annual Government Statement of Expenditure. References Bowling, A., Grundy, E., 1997. Activities of daily living: changes in functional ability in three samples of elderly and very elderly people. Age Ageing 26, 107–114. Brayne, C., Matthews, F.E., Mcgee, M.A., Jagger, C., 2001. Health and ill-health in the older population in England and Wales. The Medical Research Council Cognitive Function and Ageing Study (MRC CFAS). Age Ageing 30, 53–62. Department of Local Administration, Ministry of Interior, 2006. Number of births, deaths, registered-in and registered-out from registration record by region and sex, 2005–2006. Jagger, C., Reyes-Frausto, S., 2002. Monitoring Health by Healthy Active Life Expectancy. . Jagger, C., Ritchie, K., Bronnum-Hansen, H., Deeg, D., Gispert, R., Grimley Evans, J., Hibbett, M., Lawlor, B., Perenboom, R., Polge, C., Van Oyen, H., 1998. Mental health expectancy—the European perspective: a synopsis of results presented at the Conference of the European Network for the Calculation of Health Expectancies (Euro-REVES). Medical Research Council Cognitive Function and Ageing Study Group. Acta Psychiatr. Scand. 98, 85–91. Jitapunkul, S., Chayovan, N., 2000. Healthy life expectancy of Thai elderly: did it improve during the soap-bubble economic period? J. Med. Assoc. Thai. 83, 861–864.
7
Jitapunkul, S., Krungkraipetch, N., Kamolratanakul, P., Dhanamun, B., 2001. Dependence and active life expectancy of the elderly population living in the central region of Thailand. J. Med. Assoc. Thai. 84, 349–356. Jitapunkul, S., Kunanusont, C., Phoolcharoen, W., Suriyawongpaisal, P., Ebrahim, S., 2003. Disability-free life expectancy of elderly people in a population undergoing demographic and epidemiologic transition. Age Ageing 32, 401–405. Lenzen, M.J., Scholte op Reimer, W.J., Pedersen, S.S., Boersma, E., Maier, W., Widimsky, P., Simoons, M.L., Mercado, N.F., Wijns, W., 2007. The additional value of patient-reported health status in predicting 1-year mortality after invasive coronary procedures: a report from the Euro Heart Survey on Coronary Revascularisation. Heart 93, 339–344. Letenneur, L., Gilleron, V., Commenges, D., Helmer, C., Orgogozo, J.M., Dartigues, J.F., 1999. Are sex and educational level independent predictors of dementia and Alzheimer’s disease? Incidence data from the PAQUID project. J. Neurol. Neurosurg. Psychiatry 66, 177–183. Mahatnirundkul, S., Tantipiwatanasakul, W., Poompaisalchai, W., Wongsuwan, G., Prommanajirungkul, R., 1998. Comparison of the WHOQOL-100 and the WHOQOL-BREF (26 items). J. Ment. Health Thai. 5, 4–15. Melzer, D., McWilliams, B., Brayne, C., Johnson, T., Bond, J., 2000. Socioeconomic status and the expectation of disability in old age: estimates for England. J. Epidemiol. Commun. Health 54, 286–292. Mendes de Leon, C.F., Beckett, L.A., Fillenbaum, G.G., Brock, D.B., Branch, L.G., Evans, D.A., Berkman, L.F., 1997. Black-white differences in risk of becoming disabled and recovering from disability in old age: a longitudinal analysis of two EPESE populations. Am. J. Epidemiol. 145, 488–497. O’ Donnell, O., Propper, C., 1991. Equity and the distribution of UK National Health Service resources. J. Health Econ. 10, 1–19. Osler, M., De Groot, L.C., Enzi, G., 1991. Life-style: physical activities and activities of daily living. Euronut SENECA investigators. Eur. J. Clin. Nutr. 45 (Suppl. 3), 139– 151. Sauvaget, C., Jagger, C., Arthur, A.J., 2001. Active and cognitive impairment-free life expectancies: results from the Melton Mowbray 75+ health checks. Age Ageing 30, 509–515. Sullivan, D.F., 1971. A single index of mortality and morbidity. HSMHA Health Rep. 86, 347–354. Van Lenthe, F.J., Schrijvers, C.T., Droomers, M., Joung, I.M., Louwman, M.J., Mackenbach, J.P., 2004. Investigating explanations of socio-economic inequalities in health: the Dutch GLOBE study. Eur. J. Public Health 14, 63–70. Weytingh, M.D., Bossuyt, P.M., Van Crevel, H., 1995. Reversible dementia: more than 10% or less than 1%? A quantitative review. J. Neurol. 242, 466–471. WHO (World Health Organization), European Life Expectancy Monitoring Unit, 2008. Health Expectancy Calculation by the Sullivan Method: A Practical Guide. Health Statistics And Health Information System: Life Table. WHO, Geneva.
Contents lists available at ScienceDirect
Archives of Gerontology and Geriatrics journal homepage: www.elsevier.com/locate/archger
Health expectancies in the older Thai population Weerasak Muangpaisan a,b,*, Prasert Assantachai a, Somboon Intalapaporn a, Kathryn Richardson b, Carol Brayne b a b
Department of Preventive and Social Medicine, Faculty of Medicine, Siriraj Hospital, Mahidol University, Bangkok 10700, Thailand Institute of Public Health, University of Cambridge, Forvie Site, Robinson Way, Cambridge CB2 0SR, UK
A R T I C L E I N F O
A B S T R A C T
Article history: Received 16 October 2009 Received in revised form 30 March 2010 Accepted 1 April 2010 Available online 8 June 2010
This study aims to investigate health expectancies in five domains: cognitive health, psychological health, physical health, functional ability and self-perceived global health (SPGH) in the older Thai population. There are few studies reporting health expectancies in multidimensional health domains, most of which reported only one health dimension. The dataset used was from the Bangkok Longitudinal Study by Siriraj Hospital for the Older Men and Women (BLOSSOM), which is a community cohort study in Bangkok, Thailand. This analysis is based on the cross-sectional data in the year 2005–2006 and includes 5936 participants aged 50 years and over from community settings within six suburban areas in Bangkok. The study found that women had a longer total life expectancy (LE), but had shorter cognitive impairment-free (CIFLE), physical illness-free (PHILE) and disability-free (DIFLE) LEs, than men. However, there was no difference between the life expectancies for living with good SPGH in men and in women. Differences in health expectations might explain this finding. Health promotion and disease prevention should be initiated at a younger age and should target all health domains. ß 2010 Elsevier Ireland Ltd. All rights reserved.
Keywords: Healthy life expectancy Physical illness Cognitive impairment Psychological Disability Self-perceived global health
1. Introduction LE has now increased globally. However, living longer does not necessarily mean living with good quality of life (QoL). Healthy life expectancy (HLE) is more important. A measure that takes into account survival in and out of good health is therefore of great value in assessing and comparing health within and between geographical or social groups. LE is composed of estimated lengths of time spent in different states of health until death. These lengths of time in different states of health are health expectancies and they combine morbidity and mortality into a single composite indicator (WHO, 2008). There are two commonly used methods to calculate the health expectancy: the Sullivan method, which uses the observed agespecific prevalences of health states and the total person years lived at a particular age (Sullivan, 1971), and the multistate method, which uses a direct calculation of person years lived in the health state and requires longitudinal data to provide transition rates between health states. The latter method can take into account the reversible transition between good health and one or more disability (Jagger and Reyes-Frausto, 2002).
* Corresponding author at: Department of Preventive and Social Medicine, Faculty of Medicine, Siriraj Hospital, Mahidol University, Bangkok 10700, Thailand. Tel.: +66 2 419 8388; fax: +66 2 411 5034. E-mail address: [email protected] (W. Muangpaisan). 0167-4943/$ – see front matter ß 2010 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.archger.2010.05.012
There are few studies reporting health expectancies in multidimensional health domains. Most of the reported health expectancies show only one dimension such as disability-free life expectancies (DFLE) or cognitive-free life expectancy (CFLE). An example of a study reporting multiple health LEs is the Medical Research Council Cognitive Function and Ageing Study (MRC CFAS) which reported health expectancies in three dimensions: cognitive impairment, physical illness and functional disability (Brayne et al., 2001). SPGH is widely used to reflect the individual’s perception of health and has reported to be related to socio-economic status (O’Donnell and Propper, 1991). The patient’s subjective experience of their own health status also helps predict the risk of future mortality (Lenzen et al., 2007). SPGH can be used to calculate the health expectancy free of poor perceived health. LE is increasing over time in Asia and around the world. However, information is limited on the states of health associated with longer life. Longer LE does not necessarily imply longer periods in states of good health and there is increasing attention to quality of extended life. Research on HLE is needed for comparing the health of older populations in countries at different levels of development and for showing the dependency ratio in this population. HLE can be used to predict future needs, assess national health policy, monitor trend and inequalities. Obtaining life tables is crucial in Asian and less developed countries to enable the public policy maker to shape health plan strategies and better allocate limited resources. The objective of this paper was to
4
W. Muangpaisan et al. / Archives of Gerontology and Geriatrics 53 (2011) 3–7
investigate the health expectancies in five domains: cognitive health, psychological health, physical health, functional ability and SPGH. 2. Subjects and methods 2.1. Study design The BLOSSOM is a community cohort study in Bangkok, Thailand. The study included 5936 participants aged 50 years and over from community settings within six suburban areas in Bangkok from 2005-up to present (June 2009). The objectives of the project are to gain an understanding of the magnitude of health problems of the older population in urban areas, obtain the health determinants and predictors of the common diseases in the older population and develop strategies to promote health and prevent common diseases. Community leaders were requested to ask eligible people in their community to attend a mobile unit at a hub of the community. The hub of each community varied from a temple, school, conference hall, or leisure area. The participation rate was difficult to calculate, as the number of family members recorded in the census and by the community leaders was frequently incorrect as a result of unregistered migration. All individuals aged 50 years and over in the sampling areas were invited to participate in the study via the community leaders and community nurses of the Primary Care unit. The exclusion criteria were: (1) age under 50 years and (2) inability to answer questions independently. 2.2. Measurements This analysis considered first year data only from part of the face-to-face structured-interview which included baseline characteristics of the population, the Thai Mental State Examination (TMSE) score, the psychological domain of the Thai version of the World Health Organization Quality of Life (WHOQoL-BREF) questionnaire, the activities of daily living (ADLs) score which was translated from the ADL tool used in the Survey in Europe on Nutrition and the Elderly, a Concerted Action (SENECA), selfreported physical illnesses, special ailments (hearing problems and visual problems) and SPGH (Osler et al., 1991; Mahatnirundkul et al., 1998). SPGH was measured using the question ‘‘How satisfied are you with your health?’’, which is one of the questions in WHOQoL-BREF. The answer was rated into five levels; very dissatisfied, dissatisfied, neither satisfied nor dissatisfied, satisfied and very satisfied. 2.3. Health domains The prevalence of impairment in each health domain and poor SPGH were used to calculate the health expectancies. The cognitive health domain was based on cases of moderate to severe cognitive impairment only, as estimations of mild cognitive impairment are rather inaccurate and fluctuate (Jagger et al., 1998). Thus, a TMSE score of less than 18 was used for this purpose to make results comparable to the robust cutoff for MMSE. The psychological health domain was defined as having a poor psychological QoL score. Physical illness was based on the self-reported physical illness, and hearing and visual problems uncorrected by eyeglasses. Impairment (unable to complete, can do only with help, and can do with difficulty but without help) in self-care ability was used to define the self-care domain. Likewise, any impairment in any of the 16-item ADL scale was used for the prevalence of functional impairment to calculate the DFLE. Finally, SPGH of very dissatisfied and dissatisfied were used to define poor SPGH.
From 5936 participants 839 were excluded from the analysis; 20 for being under 50 years old and 819 for having incomplete data used in the analysis. The source of most of the missing data was from the health questionnaires (n = 513, 8.7%) and the baseline variables (n = 306, 5.2%). Age, sex, marital status and income distributions were similar in participants with missing and complete data. However, among the available data for educational level, a greater percentage of subjects with missing data had no formal education compared to subjects with complete data (15.8% vs. 8.5%). 2.4. Statistical analysis Frequencies and proportions of the main baseline characteristics were calculated and stratified by sex. Age was categorized into four groups: 50–59, 60–69, 70–79 and 80 years old and over. The independent-samples t-test was used to compare the mean age across men and women. The x2-test was used to compare the categorical variables of age, sex, educational level, marital status, living status, and income across men and women. Health life expectancies were calculated using the method developed by Sullivan (1971). The abridged life table for the Thai population, by sex, developed by the WHO 2006 was used due to the small number of participants in each year of age and to be comparable with similar studies (World Health Organization, 2008). For this study, only LE for those aged 50 years and over were relevant. Due to the lack of sex- and age-specific total population and total deaths in the WHO life table for the population aged over 80 years old, data from the Department of Local Administration, Ministry of Interior, Thailand (2006) were used for the calculation in that age group. An age interval of 5 years was set for each age group, except for the final age interval, which was assumed to have length 10 years.
3. Results The baseline characteristics of the subjects with complete data are summarized in Table 1. There were 1357 men (26.6%) and 3740 women (73.4%). Men were significantly older (Table 1). Women had a lower educational level with 74.7% studying for 4 years or less compared to 54.9% in men (p < 0.001). In terms of marital status, men were more likely to be married (81.8% vs. 46.9%), while
Table 1 Baseline characteristics of the participants with complete data. N (%) unless otherwise indicated, mean S.D., or n (%). Baseline characteristics
Men
Women
Total
p
Number Age, years
1357 64.1 8.2
3740 62.1 8.5
5097 62.6 8.5
<0.001
<0.001
Education Did not study 1–4 years Primary school Secondary school Bachelor degree/higher
39 705 105 391 117
(2.9) (52.0) (7.7) (28.8) (8.6)
395 2398 212 580 155
(10.6) (64.1) (5.7) (15.5) (4.1)
434 3103 317 971 272
(8.5) (60.9) (6.2) (19.1) (5.3)
Marital status Single Married Widowed Divorced or separated Living alone
59 1110 122 66 46
(4.3) (81.8) (9.0) (4.9) (3.4)
408 1753 1160 419 187
(10.9) (46.9) (31.0) (11.2) (5.0)
467 2863 1282 485 233
(9.2) (56.2) (25.2) (9.5) (4.6)
Income Inadequate Adequate Saving
0.02 <0.001
226 (16.7) 923 (68.0) 208 (15.3)
821 (22.0) 2431 (65.0) 488 (13.0)
1047 (20.5) 3354 (65.8) 696 (13.7)
<0.001
W. Muangpaisan et al. / Archives of Gerontology and Geriatrics 53 (2011) 3–7
5
Table 2 Total LE together with the various sub-categories of it in absolute values and % of LE. Age groups
Males LEa
50–54
55–59
60–64
65–69
70–74
75–79
80+
26.9
23.2
19.7
16.3
13.2
10.4
9.6
b
CIFLE %CIFLE
25.0 92.9
21.3 91.8
17.7 89.8
14.1 86.5
10.8 81.8
7.8 75.0
7.5 78.1
PIFLEc %PIFLE
25.5 94.8
21.8 94.0
18.2 92.4
14.8 90.8
11.6 87.9
8.7 83.7
9.1 94.8
PHILEd %PHILE
4.4 16.4
3.2 13.8
2.3 11.7
1.7 10.4
1.2 9.1
0.7 6.7
0.6 6.3
SCLEe %SCLE
21.8 81.0
18.1 78.0
14.7 74.6
11.3 69.3
8.4 63.6
5.5 52.9
4.6 47.9
DFLEf %DFLE
16.5 61.3
13.2 56.9
10.0 50.8
7.2 44.2
4.8 36.4
2.6 25.0
2.1 21.9
HDLEg %HDLE
3.8 14.1
2.7 11.6
1.9 9.6
1.3 8.0
0.8 6.1
0.4 3.8
0.2 2.1
SPGHLEh %SPGHLE
22.5 83.6
19.2 82.8
16.0 81.2
13.0 79.8
10.1 76.5
7.6 73.1
8.1 84.4
Females LEa
29.4
25.3
21.4
17.7
14.2
11.0
9.6
b
CIFLE %CIFLE
21.9 74.5
17.7 70.0
13.6 63.6
14.1 79.7
10.5 73.9
7.4 67.3
7.0 72.9
PIFLEc %PIFLE
27.1 92.2
23.0 90.9
19.1 89.3
15.4 87.0
11.8 83.1
8.4 76.4
8.9 92.7
PHILEd %PHILE
3.4 11.6
2.5 9.9
1.8 8.4
1.2 6.8
0.8 5.6
0.5 4.5
0.3 3.1
SCLEe %SCLE
20.4 69.4
16.5 65.2
12.8 59.8
9.3 52.5
6.4 45.1
3.9 35.5
3.2 33.3
DFLEf %DFLE
11.6 39.5
8.6 34.0
6.0 28.0
3.9 22.0
2.2 15.5
1.1 10.0
0.8 8.3
HDLEg %HDLE
2.1 7.1
1.4 5.5
0.9 4.2
0.5 2.8
0.3 2.1
0.2 1.8
0.1 1.0
23.7 80.6
20.3 80.2
17.0 79.4
13.7 77.4
10.6 74.6
7.7 70.0
8.1 84.4
SPGHLEh %SPGHLE
Notes: %Values are expressed as % of LE. a LE in years. b TMSE score <18. c Poor psychological QoL. d Presence of physical illness. e Self-care life expectancy in years. f Functional impairment-free life expectancy. g Any of ill-health in cognitive, psychological, physical health and functional impairment. h Healthy life expectancies by SPGH of ‘‘being in good health’’ (very satisfied, satisfied and neither dissatisfied nor satisfied).
women were more likely to be single (10.9% vs. 4.3%), widowed (31.0% vs. 9.0%), divorced or separated (11.2% vs. 4.9%). Table 2 and Fig. 1 show the total life expectancies with each type of health domain and with any health domain. Overall, women had a longer LE than men in every age group except for those aged 80+ years. This is due to the common data being used for this age group as the data was not available by sex. At the age of 50–54 years, men had a further LE of 26.9 years and women had 29.4 years. At this age, men could expect to spend 25.0, 25.5, 4.4, 21.8, 16.5, and 3.8 years free of cognitive impairment, psychological impairment, physical illness, self-care functional impairment, total functional impairment, and any impairment of health areas, respectively. Although women had a longer total LE, they had fewer years than men free from cognitive impairment, physical illness, functional impairment and any impairment in health areas. The psychological impairment-free LE in women was longer than
Fig. 1. ELE in (a) men and (b) women.
men in the age group 50–74 years and shorter than men in the age groups 75–79 years and 80+ years. The most striking feature is that the physical illness-free LE for both sexes was very short compared to other domains. Even in the younger age group (50–54 years), men had 4.4 years and women had 3.4 years of LE free from physical illness. When each domain was considered together, the ill-health free LE was only 3.5 years in men and 1.8 years in women between the ages of 50–54 years. This was because many of the younger participants had self-reported physical illness, hearing, and visual problems although they were free from impairment in the other domains. Although women had shorter health expectancies in all health states, they did not have a marked difference in LE with good SPGH. The percentage of years free from poor SPGH declined slightly with increasing age from 83.6% at the age of 50–54 years to 76.5% at age 75–79 years in men and from 80.6 to 70.0% in women, respectively. At the age of 80 years and over, the percentage increased to 84.4% for both men and women, which is likely due to the different data source used to construct the life table for this age group. 4. Discussion The life table constructed in this study emphasises that, apart from the psychological health expectancies in the young old (50– 69 years), women had disease-free life expectancies shorter than men in every health domain. However, both men and women had similar proportions of LE appreciating ‘‘being in good health’’. Women might accept some limitations of health status better than men. The study population demographic demonstrates the traditional Thai belief in the past that education is less important in women than men. Previous studies show that low socio-economic status including educational attainment and social class is a
6
W. Muangpaisan et al. / Archives of Gerontology and Geriatrics 53 (2011) 3–7
significant risk factor for negative health outcomes such as cognitive impairment, functional impairment, some specific diseases, and mortality (Letenneur et al., 1999; Melzer et al., 2000; Van Lenthe et al., 2004). The marked difference in reports from women in this study could be linked to education. Attention to ensuring greater appropriate education for both sexes may lead to improvements in health in the longer term. The healthy LEs measured in this study highlight early exposure to health problems particularly physical health in Thailand. As health promotion and prevention strategies are developed, the maintenance of good health states and care-plans which begin at an earlier age, before people become elderly, are needed. The mental health expectancy may be calculated, based on two different classes of indicators: (1) disease-specific life expectancies such as dementia-free LE and depression-free LE and (2) the functional impact of mental ill-health, such as LE free of social isolation or LE in good perceived mental health (Jagger et al., 1998). The first type of calculation needs clinical diagnosis using the International Classification of Disease (ICD) or the Diagnostic and Statistical Manual of Mental Disorders (DSM) to allow temporal stability and cross-population comparability. In this paper, the ‘‘no cognitive impairment’’ state was used as there was no clinical diagnosis of dementia in the study. However, it still provides the information as a symptom-free LE. Likewise, the choice of using ‘‘poor psychological QoL’’ as an indicator for psychological health in the life table may be less comparable to other countries. However, the ultimate goal of the health expectancies are to give outcome measures for evaluating population health which provide status relative to LE and take into account QoL. A further strength of this life table is that it provides multiple dimensions of health in the older populations. Most reported health expectancies show only one dimension such as DFLE or cognitive-free LE. The life table in this paper provides mental health expectancies (cognitive and psychological health), physical health expectancies, DFLE and SPGH expectancy. It is the first life table that combines every dimension of health in the calculation. As it is independent of the size of the populations and of their age structure, health expectancies can be used for the direct comparison of different groups that make up populations and also can contribute to public health policies. The limitations of the life table in this paper using Sullivan’s method should be taken into account. First, the calculation of health expectancies using prevalence data does not incorporate transition rates in and out of the health states. Cases in one health state can recover, and this is relatively common (Bowling and Grundy, 1997; Mendes de Leon et al., 1997). Reversible causes of cognitive impairment account for 0–23% for partial and 0–10% for full reversal (Weytingh et al., 1995). Also, recovery from disability is well-recognized especially in acute medical conditions such as stroke. With cross-sectional surveys transition rates cannot be estimated. Second, the prevalence of health states came from the community survey only and did not include those in institutions. Life tables for a population should include representative sampling from the total population to avoid bias in estimating health problems. Also, there could be a volunteer effect as people who participated in the BLOSSOM might be healthier than those who did not. Third, the definitions of health states in this study are different from other studies. Different health state definitions are common across studies and make direct comparison of health expectancies across studies difficult. However, the life table provides useful public health information. Finally, the prevalence of the health states in this study came from the cross-sectional survey of one study, not a national database. The generalizability of the studied population to the national level must be considered. There have been only three studies publishing life tables of health expectancies in Thailand (Jitapunkul and Chayovan, 2000;
Jitapunkul et al., 2001, 2003). The first study estimated HLE comparing the status of ‘‘being in good health’’ in older Thai people between 1986 and 1995. This study was based on two cross-sectional surveys, used self-perceived health as an assessment tool and calculated HLE by the Sullivan’s method. However, the questions used in both surveys were different and the scaling method of the scores was also different (Jitapunkul and Chayovan, 2000). This could lead to varied prevalence estimates of ‘‘being in good health’’. The second study estimated active life expectancies in 723 participants aged 60 years and over in central region of Thailand (Jitapunkul et al., 2001). This study used the modified Barthel ADL index and Chula ADL index to survey the prevalence of disability. The limitations were mainly due to a small number of participants in each age strata which was likely to make the estimates imprecise. Another study by Jitapunkul et al. (2003) estimated the DFLE by Sullivan’s method in Thai elderly people aged 60 years and over. There was a lack of information about mental health problems, physical illnesses and SPGH in this study. Comparison of health expectancies across studies is difficult as a result of different methods in developing the life table, different definitions of health states and also secular trends and geographic variations. Direct comparisons of the DFLE between the studies by Jitapunkul et al. (2003) and the BLOSSOM are difficult as the definitions of disability used and the measurement scales used to collect the prevalence data differ. However, both studies used the same face-to-face interview method and similar study populations of the older population in the community. The population in the study by Jitapunkul et al. (2003) also lived in more rural areas, whereas the population in the BLOSSOM study lived in more urban areas. Nevertheless, the main finding in the BLOSSOM study is consistent with the other studies in that men had better DFLE, whereas the finding from the studies by Jitapunkul et al. (2003) showed the opposite. Differences between the findings in the studies by Jitapunkul et al. (2003) and the BLOSSOM study might be due to the different scales used, different data collection processes, or changes in the country’s demographic profile over this 10-year period. Health expectancies can also be calculated using multistate modelling if longitudinal data is available. The health expectancies in the Melton Mowbray Ageing Project (MMAP) were calculated from longitudinal data using this method (Sauvaget et al., 2001). In multistate modelling, both the probability of becoming impaired in health states (decrement) and the probability of recovering from the impairment (increment) are used in the calculations. Thus, estimations are more accurate as transitions between health states (e.g., recovery) are common. The definition used for active life expectancy in the MMAP was independency in all of the seven ADLs qualified an individual to be active. The scale used to define disability in the BLOSSOM study is similar to the MMAP in that the individual had to be independent in all items to qualify as not impaired. However, the MRC CFAS health expectancy estimations by Brayne et al. (2001) used less strict criteria. Cross-sectional data were used to calculate a life table, hence transition states were not possible, as in this study. The MRC CFAS had strength in including institutional residents which makes it more generalizable to the whole population. The BLOSSOM study did not collect data from institutional carehomes; however, the proportion of the older population who reside in institutional care-homes is much smaller than compared to Western countries. The MMAP reported two domains of health expectancies (cognitive impairment-free life expectancy: CIFLE, and active LE), the MRC CFAS reported three domains (CIFLE, active LE, and physical illness-free), whereas this study reports four domains of health (adding the psychological domain) and one SPGH.
W. Muangpaisan et al. / Archives of Gerontology and Geriatrics 53 (2011) 3–7
5. Conclusions and recommendations Women had poorer health states and a shorter LE in good health despite longer total LEs. Despite this, women had similar expectancies in good SPGH to men. Life expectancies free from physical illness were short even in the population aged 50–54 years old. This provides further evidence of the need for health promotion and disease prevention to be initiated at younger ages targeting all health domains. Conflict of interest statement None. Acknowledgement This study was supported by the Thai Annual Government Statement of Expenditure. References Bowling, A., Grundy, E., 1997. Activities of daily living: changes in functional ability in three samples of elderly and very elderly people. Age Ageing 26, 107–114. Brayne, C., Matthews, F.E., Mcgee, M.A., Jagger, C., 2001. Health and ill-health in the older population in England and Wales. The Medical Research Council Cognitive Function and Ageing Study (MRC CFAS). Age Ageing 30, 53–62. Department of Local Administration, Ministry of Interior, 2006. Number of births, deaths, registered-in and registered-out from registration record by region and sex, 2005–2006. Jagger, C., Reyes-Frausto, S., 2002. Monitoring Health by Healthy Active Life Expectancy. . Jagger, C., Ritchie, K., Bronnum-Hansen, H., Deeg, D., Gispert, R., Grimley Evans, J., Hibbett, M., Lawlor, B., Perenboom, R., Polge, C., Van Oyen, H., 1998. Mental health expectancy—the European perspective: a synopsis of results presented at the Conference of the European Network for the Calculation of Health Expectancies (Euro-REVES). Medical Research Council Cognitive Function and Ageing Study Group. Acta Psychiatr. Scand. 98, 85–91. Jitapunkul, S., Chayovan, N., 2000. Healthy life expectancy of Thai elderly: did it improve during the soap-bubble economic period? J. Med. Assoc. Thai. 83, 861–864.
7
Jitapunkul, S., Krungkraipetch, N., Kamolratanakul, P., Dhanamun, B., 2001. Dependence and active life expectancy of the elderly population living in the central region of Thailand. J. Med. Assoc. Thai. 84, 349–356. Jitapunkul, S., Kunanusont, C., Phoolcharoen, W., Suriyawongpaisal, P., Ebrahim, S., 2003. Disability-free life expectancy of elderly people in a population undergoing demographic and epidemiologic transition. Age Ageing 32, 401–405. Lenzen, M.J., Scholte op Reimer, W.J., Pedersen, S.S., Boersma, E., Maier, W., Widimsky, P., Simoons, M.L., Mercado, N.F., Wijns, W., 2007. The additional value of patient-reported health status in predicting 1-year mortality after invasive coronary procedures: a report from the Euro Heart Survey on Coronary Revascularisation. Heart 93, 339–344. Letenneur, L., Gilleron, V., Commenges, D., Helmer, C., Orgogozo, J.M., Dartigues, J.F., 1999. Are sex and educational level independent predictors of dementia and Alzheimer’s disease? Incidence data from the PAQUID project. J. Neurol. Neurosurg. Psychiatry 66, 177–183. Mahatnirundkul, S., Tantipiwatanasakul, W., Poompaisalchai, W., Wongsuwan, G., Prommanajirungkul, R., 1998. Comparison of the WHOQOL-100 and the WHOQOL-BREF (26 items). J. Ment. Health Thai. 5, 4–15. Melzer, D., McWilliams, B., Brayne, C., Johnson, T., Bond, J., 2000. Socioeconomic status and the expectation of disability in old age: estimates for England. J. Epidemiol. Commun. Health 54, 286–292. Mendes de Leon, C.F., Beckett, L.A., Fillenbaum, G.G., Brock, D.B., Branch, L.G., Evans, D.A., Berkman, L.F., 1997. Black-white differences in risk of becoming disabled and recovering from disability in old age: a longitudinal analysis of two EPESE populations. Am. J. Epidemiol. 145, 488–497. O’ Donnell, O., Propper, C., 1991. Equity and the distribution of UK National Health Service resources. J. Health Econ. 10, 1–19. Osler, M., De Groot, L.C., Enzi, G., 1991. Life-style: physical activities and activities of daily living. Euronut SENECA investigators. Eur. J. Clin. Nutr. 45 (Suppl. 3), 139– 151. Sauvaget, C., Jagger, C., Arthur, A.J., 2001. Active and cognitive impairment-free life expectancies: results from the Melton Mowbray 75+ health checks. Age Ageing 30, 509–515. Sullivan, D.F., 1971. A single index of mortality and morbidity. HSMHA Health Rep. 86, 347–354. Van Lenthe, F.J., Schrijvers, C.T., Droomers, M., Joung, I.M., Louwman, M.J., Mackenbach, J.P., 2004. Investigating explanations of socio-economic inequalities in health: the Dutch GLOBE study. Eur. J. Public Health 14, 63–70. Weytingh, M.D., Bossuyt, P.M., Van Crevel, H., 1995. Reversible dementia: more than 10% or less than 1%? A quantitative review. J. Neurol. 242, 466–471. WHO (World Health Organization), European Life Expectancy Monitoring Unit, 2008. Health Expectancy Calculation by the Sullivan Method: A Practical Guide. Health Statistics And Health Information System: Life Table. WHO, Geneva.