- Split View
-
Views
-
Cite
Cite
Po-Wen Ku, Andrew Steptoe, Yi-Huei Chen, Li-Jung Chen, Ching-Heng Lin, Prospective association between late-life physical activity and hospital care utilisation: a 7-year nationwide follow-up study, Age and Ageing, Volume 46, Issue 3, May 2017, Pages 452–459, https://doi.org/10.1093/ageing/afw202
- Share Icon Share
Abstract
it is still equivocal whether there is a potential role of late-life physical activity in ameliorating the challenges of increasing healthcare expenditure due to the consequence of global population ageing.
this study aimed to examine the prospective association between physical activity and subsequent hospital care utilisation in older adults and to explore the optimal dose of physical activity required to reduce hospital care utilisation.
this was a prospective cohort study based on the data from the Taiwan 2005 National Health Interview Survey, which were linked to the 2005–12 claims data from the National Health Insurance system.
1,760 older adults aged 65 or more.
the frequency, duration and intensity for physical activity were assessed, and total physical activity energy expenditure was estimated. The average annualised hospital care utilisation for the period 2006 through 2012, including number of hospitalisations, number of days in hospital and the costs of hospitalisation, were calculated.
older adults engaging in at least moderate volume of physical activity (≥1,000 kcal/week) experienced fewer subsequent hospital admissions and fewer days in hospital than did sedentary individuals, after adjusting for covariates. Trends for reduced hospitalisation costs were also found. These associations persisted in sensitivity analyses, including tests of reverse causation.
this study has provided evidence that older adults who are at least moderately active may minimise utilisation of hospital care services. The findings highlight the importance of maintaining a physically active lifestyle in later life.
Introduction
There is well-documented evidence for the benefits of physical activity for a range of health conditions [1, 2], and these positive health consequences continue into older ages [3]. The rapidly increasing proportion of people aged 65 or older in many countries around the world has stimulated growing concern over the impact on healthcare systems. Increasing attention is being paid by health service providers and policymakers to the potential role of physical activity in ameliorating these challenges [4].
Although previous studies have shown that greater physical activity is associated with lower health service utilisation, few have focused on older populations [5–7]. Prospective studies examining the relationship have suffered from methodological limitations such as short follow-up periods [8, 9], failure to adjust for healthcare utilisation at baseline [10–12], self-reported healthcare use [11] and small sample sizes [9, 12]. Besides, the previous studies of this topic have been mainly conducted in North America and European countries. There are considerable socio-economic, political and cultural differences between East Asian and Western countries that may influence healthcare services and physical activity patterns. To date, there is a paucity of published data on the relationships between physical activity and healthcare utilisation among older adults in East Asia.
We took advantage of the comprehensive hospital care records in Taiwan to address some of these shortfalls by exploring the prospective associations between physical activity and subsequent health service utilisation (including hospital admissions, hospitalisation days and related costs) in a 7-year follow-up (2005–12) of a nationally representative sample of older adults. We adjusted for multiple potentially confounding factors and carried out sensitivity analyses to test for reverse causality. Public health recommendations call for 30 minutes or more of at least moderate-intensity physical activity on most days of the week (approximately ≥1,000 kcal/week) [13, 14]. However, few data are available regarding the optimal volume of physical activity required to influence healthcare use. Our second aim was therefore to identify the amount of physical activity needed to reduce hospital care utilisation.
Methods
Study design and sample
This was a prospective study involving participants aged 65 years and older in the 2005 National Health Interview Survey (NHIS) in Taiwan, which was conducted by the National Health Research Institutes and Bureau of Health Promotion, Taiwan. The NHIS comprised 24,726 participants (response rate = 80.59%) with 2,727 being aged 65 years or older. Participants were selected using multistage stratified systematic sampling design to select a nationally representative sample [15].
The NHIS data were linked to the 2005–12 claims data in the National Health Insurance Research Database. National Health Insurance in Taiwan is a public compulsory insurance system for all citizens (coverage rate = 99.6% of the total population 23.1 million in 2009) [16]. Among 2,727 older adults, 1,760 (65%) provided the consent to link their claim data.
Ethics statement
Ethics approval for this study was obtained from Taichung Veterans General Hospital Institutional Review Board, Taiwan (reference number: SE14257A-1).
Measures
Outcome variables: hospital care utilisation
Based on the claim data from the Taiwan National Health Insurance Database, the average annualised hospital care utilisation for the period 2006 through 2012, including number of hospitalisations, number of days in hospital and the costs of hospitalisation (US dollars), were calculated.
Exposure variable: physical activity
Two types of physical activity were measured in the 2005 NHIS. Leisure-time physical activity was assessed using the following questions. ‘Have you taken part in any leisure-time physical activities in the past 2 weeks?’ Respondents were asked to identify the types of leisure-time physical activity they had engaged in from 31 named activities (e.g. walking, swimming, Tai Chi, etc.) and were able to specify up to five types. Participants self-reported the frequency and duration for each activity. Metabolic equivalent (MET) intensity levels for each activity were assigned based on the compendium of physical activities [17]. Energy expenditure (in kcal) of each activity per week was calculated by activity intensity code (kcal/min) × frequency per week (times) × duration for each time (min). The energy expenditure values were then added to provide a total weekly sum for leisure-time activity. Participation in heavy physical labour during non-leisure time was assessed with the item, ‘Have you taken part in any heavy physical labour in the past 2 weeks?’ Respondents were then requested to identify the type of activity they engaged in from 10 types of activities (e.g. farm work, heavy lifting, fishing work, heavy household chores, etc.) and could specify up to five types. The total weekly amount of energy expenditure for heavy physical labour was then estimated as before. The sum of weekly total physical activity was computed to estimate total weekly energy expenditure, which was classified into four categories (no physical activity = 0, low levels = 1–999, moderate levels = 1,000–1,999 and high levels = ≥2,000 kcal/week) [14]. These measures have been employed in previous research [18, 19], with repeated measurements (n = 72) at 1, 3 and 6 months post-interview. Consistency of the main question (i.e. whether engaging in physical activity in the past 2 weeks) was 0.88, while Kappa scores for repeated measures of frequency and duration ranged between 0.41 and 0.46, providing acceptable reliability of the items utilised by the study [19, 20]. The content validity of all of the items in NHIS was reviewed by expert panels, and the whole questionnaire was then administered twice in pilot testing [21]. The physical activity questionnaires were used in the 2001, 2005, 2009 and 2013 NHISs, and functioned well. Additionally, participants with a higher level of total physical activity (especially >1000 kcal/week) in this study had lower rates of overweight/obesity impaired activities of daily living (ADL), and multiple comorbidity (all P values for chi-square tests < 0.001). This provides further evidence for the validity of this instrument.
Covariates
The following potentially relevant factors in 2005 NHIS were included as covariates based on previous studies [5, 15]: (i) socio-demographic factors: sex, age (65–74, 75–84 and ≥85 years), education level (no formal schooling and primary school or more), marital status (married/cohabitating, others), working status (retired/unable to work, homemaker and currently employed), monthly income (US dollar: <330, 330–659 and ≥660) and living status (alone versus with family/others); (ii) lifestyle behaviours: alcohol consumption (yes versus no) and smoking (current smoker, former smoker and never smoked); (iii) health status: body mass index (BMI: <18.50, 18.50–23.99, 24–26.99, ≥27 kg/m2) [22], Charlson comorbidity index (0, 1–2, ≥3), which considers the number and severity of chronic diseases [23, 24] and impaired ADL (no difficulties at all versus some or great difficulties); (iv) hospital care utilisation at baseline.
Data analysis
Descriptive statistics for each aspect of hospital care utilisation were calculated first to characterise the sample structure. Given the violation of normality, Mann–Whitney U test and Kruskal–Wallis test were used to test for differences across levels of physical activity and covariates. Variables with a P value <0.05 were included in the following regression models for adjustment.
Given that the distribution of hospital care utilisation variables was positively skewed and overdispersed, Poisson regression, negative binomial regression, zero-inflated negative binomial regression and linear regression with a log-transformed response were considered for analysing the data [25–27]. The model selections were based on the Akaike Information Criterion (AIC). The prospective associations between physical activity and hospital care services were examined using Poisson regression for number of hospitalisations, negative binomial regression for number of hospitalisation days and linear regression with a log-transformed response for hospitalisation costs respectively (Model 1).
To assess the dose–response associations between physical activity and hospital care utilisation, we tested for linear trends across activity categories. To examine the potential influence of individuals with extreme consumption of healthcare resources (outliers: >3 standard deviations from the mean) [28] at baseline and the follow-up, participants who had extreme hospitalisation costs in 2005 or during the follow-up period (2006–12) were excluded (n = 32) (Model 2). Given that the inclusion of participants with very poor health, difficulties in physical function or limited pulmonary function might affect the estimate of physical activity–hospital care utilisation relationships, several sensitivity analyses were performed to address reverse causation. First, participants who died during the first 2 years of follow-up (2005–07, n = 181) were excluded (Model 3). Then, participants who had any condition including difficulty with ADL, arthritis and chronic obstructive pulmonary diseases (COPD) (n = 475) at baseline were excluded (Model 4).
All analyses were conducted using SAS 9.4 software and a P value <0.05 was considered statistically significant.
Results
The associations between baseline characteristics and annual mean number of hospitalisations, number of days in hospital and hospital care costs are summarised in Table 1. All independent variables, with the exception of sex, educational level, monthly income and living status, were significantly associated with hospital care utilisation variables (P < 0.05). Participants who were younger, married/cohabiting, currently working, with a higher level of physical activity, consumers of alcohol and who never smoked, and were overweight had no comorbidities and less difficulty with ADL and had lower hospital care utilisation over the follow-up period. Additionally, older women had lower numbers of hospital admission, while participants who were better educated had fewer hospital admissions and hospitalisation days but no differences in costs.
Variables in 2005 . | n . | Annual mean number of hospitalisation . | P valuea . | Annual mean number of hospitalisation days . | P valuea . | Annual mean of hospitalisation costs (USD) . | P valuea . | |||
---|---|---|---|---|---|---|---|---|---|---|
Mean . | SD . | Mean . | SD . | Mean . | SD . | |||||
Socio-demographic | ||||||||||
Sex | ||||||||||
Female | 824 | 0.4 | 0.7 | 0.026 | 6.6 | 23.6 | 0.063 | 1237.3 | 3433.6 | 0.074 |
Male | 936 | 0.5 | 0.6 | 6.4 | 20.0 | 1337.4 | 3524.6 | |||
Age (years) | ||||||||||
≥85 | 97 | 0.7 | 0.7 | <0.001 | 14.3 | 38.4 | <0.001 | 2456.9 | 6227.2 | <0.001 |
5–84 | 595 | 0.5 | 0.6 | 8.1 | 20.3 | 1562.7 | 3317.4 | |||
65–74 | 1,068 | 0.4 | 0.6 | 5.0 | 20.2 | 1033.0 | 3183.8 | |||
Education level | ||||||||||
No formal schooling | 613 | 0.5 | 0.7 | 0.007 | 8.1 | 26.1 | 0.011 | 1466.1 | 3699.0 | 0.062 |
Primary school or more | 1,144 | 0.4 | 0.6 | 5.6 | 18.9 | 1192.3 | 3359.0 | |||
Marital status | ||||||||||
Others | 602 | 0.5 | 0.7 | <0.001 | 8.6 | 27.8 | <0.001 | 1586.8 | 4149.6 | 0.001 |
Married/cohabiting | 1,158 | 0.4 | 0.6 | 5.5 | 17.7 | 1136.5 | 3068.4 | |||
Working status | ||||||||||
Retired/unable to work | 1,154 | 0.5 | 0.7 | <0.001 | 7.3 | 22.6 | <0.001 | 1425.1 | 3441.2 | <0.001 |
Housekeeper | 305 | 0.4 | 0.6 | 6.6 | 25.0 | 1222.3 | 3757.2 | |||
Currently working | 276 | 0.3 | 0.5 | 3.0 | 12.0 | 637.6 | 2023.7 | |||
Monthly income (USD) | ||||||||||
<330 | 1,189 | 0.4 | 0.7 | 0.217 | 6.5 | 20.0 | 0.255 | 1326.4 | 3444.6 | 0.375 |
330–659 | 280 | 0.5 | 0.7 | 7.8 | 29.7 | 1322.2 | 3513.9 | |||
≥660 | 281 | 0.3 | 0.5 | 5.5 | 19.6 | 1112.5 | 3656.1 | |||
Living status | ||||||||||
Alone | 176 | 0.4 | 0.6 | 0.672 | 5.4 | 13.3 | 0.695 | 1024.9 | 2278.1 | 0.592 |
With family/others | 1,584 | 0.4 | 0.7 | 6.7 | 22.5 | 1320.0 | 3589.9 | |||
Lifestyle behaviours | ||||||||||
Physical activity (kcal/week) | ||||||||||
None (0) | 605 | 0.6 | 0.8 | <0.001 | 10.7 | 30.8 | <0.001 | 1933.3 | 4632.8 | <0.001 |
Low (1–999) | 488 | 0.4 | 0.6 | 6.6 | 20.8 | 1355.4 | 3576.6 | |||
Moderate (1,000–1,999) | 262 | 0.3 | 0.5 | 3.3 | 6.9 | 830.5 | 1906.9 | |||
High (≥2,000) | 405 | 0.3 | 0.5 | 2.4 | 5.5 | 549.8 | 1268.7 | |||
Drinking | ||||||||||
Yes | 365 | 0.3 | 0.5 | 0.006 | 3.4 | 6.9 | 0.008 | 739.8 | 1426.5 | 0.009 |
No | 1,395 | 0.5 | 0.7 | 7.3 | 24.1 | 1434.6 | 3829.5 | |||
Smoking | ||||||||||
Current smoker | 311 | 0.5 | 0.8 | 0.002 | 5.8 | 10.7 | 0.001 | 1242.2 | 2440.9 | 0.002 |
Former smoker | 204 | 0.6 | 0.8 | 10.2 | 33.0 | 2005.1 | 6137.5 | |||
Never smoked | 1,203 | 0.4 | 0.6 | 6.0 | 21.5 | 1180.2 | 3382.4 | |||
Health status | ||||||||||
BMI (kg/m2) | ||||||||||
Underweight (<18.5) | 105 | 0.7 | 0.8 | 0.003 | 12.3 | 32.7 | 0.002 | 2231.6 | 5204.0 | 0.013 |
Obese (≥27) | 327 | 0.4 | 0.7 | 6.5 | 25.4 | 1167.3 | 2991.6 | |||
Overweight (24–26.99) | 468 | 0.4 | 0.6 | 4.4 | 10.0 | 1052.7 | 2828.1 | |||
Normal (18.5–23.99) | 804 | 0.4 | 0.6 | 5.8 | 16.6 | 1202.6 | 3077.8 | |||
Charlson comorbidity index | ||||||||||
≥3 | 336 | 0.7 | 0.9 | <0.001 | 10.6 | 29.7 | <0.001 | 1895.5 | 3989.3 | <0.001 |
1–2 | 708 | 0.5 | 0.6 | 8.0 | 25.3 | 1585.9 | 4291.6 | |||
0 | 716 | 0.3 | 0.5 | 3.2 | 9.6 | 714.6 | 1887.5 | |||
ADL | ||||||||||
With difficulty | 236 | 0.8 | 0.8 | <0.001 | 19.9 | 46.1 | <0.001 | 3576.5 | 7264.6 | <0.001 |
No difficulties at all | 1,524 | 0.4 | 0.6 | 4.5 | 13.6 | 936.5 | 2218.5 |
Variables in 2005 . | n . | Annual mean number of hospitalisation . | P valuea . | Annual mean number of hospitalisation days . | P valuea . | Annual mean of hospitalisation costs (USD) . | P valuea . | |||
---|---|---|---|---|---|---|---|---|---|---|
Mean . | SD . | Mean . | SD . | Mean . | SD . | |||||
Socio-demographic | ||||||||||
Sex | ||||||||||
Female | 824 | 0.4 | 0.7 | 0.026 | 6.6 | 23.6 | 0.063 | 1237.3 | 3433.6 | 0.074 |
Male | 936 | 0.5 | 0.6 | 6.4 | 20.0 | 1337.4 | 3524.6 | |||
Age (years) | ||||||||||
≥85 | 97 | 0.7 | 0.7 | <0.001 | 14.3 | 38.4 | <0.001 | 2456.9 | 6227.2 | <0.001 |
5–84 | 595 | 0.5 | 0.6 | 8.1 | 20.3 | 1562.7 | 3317.4 | |||
65–74 | 1,068 | 0.4 | 0.6 | 5.0 | 20.2 | 1033.0 | 3183.8 | |||
Education level | ||||||||||
No formal schooling | 613 | 0.5 | 0.7 | 0.007 | 8.1 | 26.1 | 0.011 | 1466.1 | 3699.0 | 0.062 |
Primary school or more | 1,144 | 0.4 | 0.6 | 5.6 | 18.9 | 1192.3 | 3359.0 | |||
Marital status | ||||||||||
Others | 602 | 0.5 | 0.7 | <0.001 | 8.6 | 27.8 | <0.001 | 1586.8 | 4149.6 | 0.001 |
Married/cohabiting | 1,158 | 0.4 | 0.6 | 5.5 | 17.7 | 1136.5 | 3068.4 | |||
Working status | ||||||||||
Retired/unable to work | 1,154 | 0.5 | 0.7 | <0.001 | 7.3 | 22.6 | <0.001 | 1425.1 | 3441.2 | <0.001 |
Housekeeper | 305 | 0.4 | 0.6 | 6.6 | 25.0 | 1222.3 | 3757.2 | |||
Currently working | 276 | 0.3 | 0.5 | 3.0 | 12.0 | 637.6 | 2023.7 | |||
Monthly income (USD) | ||||||||||
<330 | 1,189 | 0.4 | 0.7 | 0.217 | 6.5 | 20.0 | 0.255 | 1326.4 | 3444.6 | 0.375 |
330–659 | 280 | 0.5 | 0.7 | 7.8 | 29.7 | 1322.2 | 3513.9 | |||
≥660 | 281 | 0.3 | 0.5 | 5.5 | 19.6 | 1112.5 | 3656.1 | |||
Living status | ||||||||||
Alone | 176 | 0.4 | 0.6 | 0.672 | 5.4 | 13.3 | 0.695 | 1024.9 | 2278.1 | 0.592 |
With family/others | 1,584 | 0.4 | 0.7 | 6.7 | 22.5 | 1320.0 | 3589.9 | |||
Lifestyle behaviours | ||||||||||
Physical activity (kcal/week) | ||||||||||
None (0) | 605 | 0.6 | 0.8 | <0.001 | 10.7 | 30.8 | <0.001 | 1933.3 | 4632.8 | <0.001 |
Low (1–999) | 488 | 0.4 | 0.6 | 6.6 | 20.8 | 1355.4 | 3576.6 | |||
Moderate (1,000–1,999) | 262 | 0.3 | 0.5 | 3.3 | 6.9 | 830.5 | 1906.9 | |||
High (≥2,000) | 405 | 0.3 | 0.5 | 2.4 | 5.5 | 549.8 | 1268.7 | |||
Drinking | ||||||||||
Yes | 365 | 0.3 | 0.5 | 0.006 | 3.4 | 6.9 | 0.008 | 739.8 | 1426.5 | 0.009 |
No | 1,395 | 0.5 | 0.7 | 7.3 | 24.1 | 1434.6 | 3829.5 | |||
Smoking | ||||||||||
Current smoker | 311 | 0.5 | 0.8 | 0.002 | 5.8 | 10.7 | 0.001 | 1242.2 | 2440.9 | 0.002 |
Former smoker | 204 | 0.6 | 0.8 | 10.2 | 33.0 | 2005.1 | 6137.5 | |||
Never smoked | 1,203 | 0.4 | 0.6 | 6.0 | 21.5 | 1180.2 | 3382.4 | |||
Health status | ||||||||||
BMI (kg/m2) | ||||||||||
Underweight (<18.5) | 105 | 0.7 | 0.8 | 0.003 | 12.3 | 32.7 | 0.002 | 2231.6 | 5204.0 | 0.013 |
Obese (≥27) | 327 | 0.4 | 0.7 | 6.5 | 25.4 | 1167.3 | 2991.6 | |||
Overweight (24–26.99) | 468 | 0.4 | 0.6 | 4.4 | 10.0 | 1052.7 | 2828.1 | |||
Normal (18.5–23.99) | 804 | 0.4 | 0.6 | 5.8 | 16.6 | 1202.6 | 3077.8 | |||
Charlson comorbidity index | ||||||||||
≥3 | 336 | 0.7 | 0.9 | <0.001 | 10.6 | 29.7 | <0.001 | 1895.5 | 3989.3 | <0.001 |
1–2 | 708 | 0.5 | 0.6 | 8.0 | 25.3 | 1585.9 | 4291.6 | |||
0 | 716 | 0.3 | 0.5 | 3.2 | 9.6 | 714.6 | 1887.5 | |||
ADL | ||||||||||
With difficulty | 236 | 0.8 | 0.8 | <0.001 | 19.9 | 46.1 | <0.001 | 3576.5 | 7264.6 | <0.001 |
No difficulties at all | 1,524 | 0.4 | 0.6 | 4.5 | 13.6 | 936.5 | 2218.5 |
USD, United States Dollar; United States Dollar to Taiwanese Dollar exchange rate = 1:30.
aTest for Mann–Whitney U test or Kruskal–Wallis test.
Variables in 2005 . | n . | Annual mean number of hospitalisation . | P valuea . | Annual mean number of hospitalisation days . | P valuea . | Annual mean of hospitalisation costs (USD) . | P valuea . | |||
---|---|---|---|---|---|---|---|---|---|---|
Mean . | SD . | Mean . | SD . | Mean . | SD . | |||||
Socio-demographic | ||||||||||
Sex | ||||||||||
Female | 824 | 0.4 | 0.7 | 0.026 | 6.6 | 23.6 | 0.063 | 1237.3 | 3433.6 | 0.074 |
Male | 936 | 0.5 | 0.6 | 6.4 | 20.0 | 1337.4 | 3524.6 | |||
Age (years) | ||||||||||
≥85 | 97 | 0.7 | 0.7 | <0.001 | 14.3 | 38.4 | <0.001 | 2456.9 | 6227.2 | <0.001 |
5–84 | 595 | 0.5 | 0.6 | 8.1 | 20.3 | 1562.7 | 3317.4 | |||
65–74 | 1,068 | 0.4 | 0.6 | 5.0 | 20.2 | 1033.0 | 3183.8 | |||
Education level | ||||||||||
No formal schooling | 613 | 0.5 | 0.7 | 0.007 | 8.1 | 26.1 | 0.011 | 1466.1 | 3699.0 | 0.062 |
Primary school or more | 1,144 | 0.4 | 0.6 | 5.6 | 18.9 | 1192.3 | 3359.0 | |||
Marital status | ||||||||||
Others | 602 | 0.5 | 0.7 | <0.001 | 8.6 | 27.8 | <0.001 | 1586.8 | 4149.6 | 0.001 |
Married/cohabiting | 1,158 | 0.4 | 0.6 | 5.5 | 17.7 | 1136.5 | 3068.4 | |||
Working status | ||||||||||
Retired/unable to work | 1,154 | 0.5 | 0.7 | <0.001 | 7.3 | 22.6 | <0.001 | 1425.1 | 3441.2 | <0.001 |
Housekeeper | 305 | 0.4 | 0.6 | 6.6 | 25.0 | 1222.3 | 3757.2 | |||
Currently working | 276 | 0.3 | 0.5 | 3.0 | 12.0 | 637.6 | 2023.7 | |||
Monthly income (USD) | ||||||||||
<330 | 1,189 | 0.4 | 0.7 | 0.217 | 6.5 | 20.0 | 0.255 | 1326.4 | 3444.6 | 0.375 |
330–659 | 280 | 0.5 | 0.7 | 7.8 | 29.7 | 1322.2 | 3513.9 | |||
≥660 | 281 | 0.3 | 0.5 | 5.5 | 19.6 | 1112.5 | 3656.1 | |||
Living status | ||||||||||
Alone | 176 | 0.4 | 0.6 | 0.672 | 5.4 | 13.3 | 0.695 | 1024.9 | 2278.1 | 0.592 |
With family/others | 1,584 | 0.4 | 0.7 | 6.7 | 22.5 | 1320.0 | 3589.9 | |||
Lifestyle behaviours | ||||||||||
Physical activity (kcal/week) | ||||||||||
None (0) | 605 | 0.6 | 0.8 | <0.001 | 10.7 | 30.8 | <0.001 | 1933.3 | 4632.8 | <0.001 |
Low (1–999) | 488 | 0.4 | 0.6 | 6.6 | 20.8 | 1355.4 | 3576.6 | |||
Moderate (1,000–1,999) | 262 | 0.3 | 0.5 | 3.3 | 6.9 | 830.5 | 1906.9 | |||
High (≥2,000) | 405 | 0.3 | 0.5 | 2.4 | 5.5 | 549.8 | 1268.7 | |||
Drinking | ||||||||||
Yes | 365 | 0.3 | 0.5 | 0.006 | 3.4 | 6.9 | 0.008 | 739.8 | 1426.5 | 0.009 |
No | 1,395 | 0.5 | 0.7 | 7.3 | 24.1 | 1434.6 | 3829.5 | |||
Smoking | ||||||||||
Current smoker | 311 | 0.5 | 0.8 | 0.002 | 5.8 | 10.7 | 0.001 | 1242.2 | 2440.9 | 0.002 |
Former smoker | 204 | 0.6 | 0.8 | 10.2 | 33.0 | 2005.1 | 6137.5 | |||
Never smoked | 1,203 | 0.4 | 0.6 | 6.0 | 21.5 | 1180.2 | 3382.4 | |||
Health status | ||||||||||
BMI (kg/m2) | ||||||||||
Underweight (<18.5) | 105 | 0.7 | 0.8 | 0.003 | 12.3 | 32.7 | 0.002 | 2231.6 | 5204.0 | 0.013 |
Obese (≥27) | 327 | 0.4 | 0.7 | 6.5 | 25.4 | 1167.3 | 2991.6 | |||
Overweight (24–26.99) | 468 | 0.4 | 0.6 | 4.4 | 10.0 | 1052.7 | 2828.1 | |||
Normal (18.5–23.99) | 804 | 0.4 | 0.6 | 5.8 | 16.6 | 1202.6 | 3077.8 | |||
Charlson comorbidity index | ||||||||||
≥3 | 336 | 0.7 | 0.9 | <0.001 | 10.6 | 29.7 | <0.001 | 1895.5 | 3989.3 | <0.001 |
1–2 | 708 | 0.5 | 0.6 | 8.0 | 25.3 | 1585.9 | 4291.6 | |||
0 | 716 | 0.3 | 0.5 | 3.2 | 9.6 | 714.6 | 1887.5 | |||
ADL | ||||||||||
With difficulty | 236 | 0.8 | 0.8 | <0.001 | 19.9 | 46.1 | <0.001 | 3576.5 | 7264.6 | <0.001 |
No difficulties at all | 1,524 | 0.4 | 0.6 | 4.5 | 13.6 | 936.5 | 2218.5 |
Variables in 2005 . | n . | Annual mean number of hospitalisation . | P valuea . | Annual mean number of hospitalisation days . | P valuea . | Annual mean of hospitalisation costs (USD) . | P valuea . | |||
---|---|---|---|---|---|---|---|---|---|---|
Mean . | SD . | Mean . | SD . | Mean . | SD . | |||||
Socio-demographic | ||||||||||
Sex | ||||||||||
Female | 824 | 0.4 | 0.7 | 0.026 | 6.6 | 23.6 | 0.063 | 1237.3 | 3433.6 | 0.074 |
Male | 936 | 0.5 | 0.6 | 6.4 | 20.0 | 1337.4 | 3524.6 | |||
Age (years) | ||||||||||
≥85 | 97 | 0.7 | 0.7 | <0.001 | 14.3 | 38.4 | <0.001 | 2456.9 | 6227.2 | <0.001 |
5–84 | 595 | 0.5 | 0.6 | 8.1 | 20.3 | 1562.7 | 3317.4 | |||
65–74 | 1,068 | 0.4 | 0.6 | 5.0 | 20.2 | 1033.0 | 3183.8 | |||
Education level | ||||||||||
No formal schooling | 613 | 0.5 | 0.7 | 0.007 | 8.1 | 26.1 | 0.011 | 1466.1 | 3699.0 | 0.062 |
Primary school or more | 1,144 | 0.4 | 0.6 | 5.6 | 18.9 | 1192.3 | 3359.0 | |||
Marital status | ||||||||||
Others | 602 | 0.5 | 0.7 | <0.001 | 8.6 | 27.8 | <0.001 | 1586.8 | 4149.6 | 0.001 |
Married/cohabiting | 1,158 | 0.4 | 0.6 | 5.5 | 17.7 | 1136.5 | 3068.4 | |||
Working status | ||||||||||
Retired/unable to work | 1,154 | 0.5 | 0.7 | <0.001 | 7.3 | 22.6 | <0.001 | 1425.1 | 3441.2 | <0.001 |
Housekeeper | 305 | 0.4 | 0.6 | 6.6 | 25.0 | 1222.3 | 3757.2 | |||
Currently working | 276 | 0.3 | 0.5 | 3.0 | 12.0 | 637.6 | 2023.7 | |||
Monthly income (USD) | ||||||||||
<330 | 1,189 | 0.4 | 0.7 | 0.217 | 6.5 | 20.0 | 0.255 | 1326.4 | 3444.6 | 0.375 |
330–659 | 280 | 0.5 | 0.7 | 7.8 | 29.7 | 1322.2 | 3513.9 | |||
≥660 | 281 | 0.3 | 0.5 | 5.5 | 19.6 | 1112.5 | 3656.1 | |||
Living status | ||||||||||
Alone | 176 | 0.4 | 0.6 | 0.672 | 5.4 | 13.3 | 0.695 | 1024.9 | 2278.1 | 0.592 |
With family/others | 1,584 | 0.4 | 0.7 | 6.7 | 22.5 | 1320.0 | 3589.9 | |||
Lifestyle behaviours | ||||||||||
Physical activity (kcal/week) | ||||||||||
None (0) | 605 | 0.6 | 0.8 | <0.001 | 10.7 | 30.8 | <0.001 | 1933.3 | 4632.8 | <0.001 |
Low (1–999) | 488 | 0.4 | 0.6 | 6.6 | 20.8 | 1355.4 | 3576.6 | |||
Moderate (1,000–1,999) | 262 | 0.3 | 0.5 | 3.3 | 6.9 | 830.5 | 1906.9 | |||
High (≥2,000) | 405 | 0.3 | 0.5 | 2.4 | 5.5 | 549.8 | 1268.7 | |||
Drinking | ||||||||||
Yes | 365 | 0.3 | 0.5 | 0.006 | 3.4 | 6.9 | 0.008 | 739.8 | 1426.5 | 0.009 |
No | 1,395 | 0.5 | 0.7 | 7.3 | 24.1 | 1434.6 | 3829.5 | |||
Smoking | ||||||||||
Current smoker | 311 | 0.5 | 0.8 | 0.002 | 5.8 | 10.7 | 0.001 | 1242.2 | 2440.9 | 0.002 |
Former smoker | 204 | 0.6 | 0.8 | 10.2 | 33.0 | 2005.1 | 6137.5 | |||
Never smoked | 1,203 | 0.4 | 0.6 | 6.0 | 21.5 | 1180.2 | 3382.4 | |||
Health status | ||||||||||
BMI (kg/m2) | ||||||||||
Underweight (<18.5) | 105 | 0.7 | 0.8 | 0.003 | 12.3 | 32.7 | 0.002 | 2231.6 | 5204.0 | 0.013 |
Obese (≥27) | 327 | 0.4 | 0.7 | 6.5 | 25.4 | 1167.3 | 2991.6 | |||
Overweight (24–26.99) | 468 | 0.4 | 0.6 | 4.4 | 10.0 | 1052.7 | 2828.1 | |||
Normal (18.5–23.99) | 804 | 0.4 | 0.6 | 5.8 | 16.6 | 1202.6 | 3077.8 | |||
Charlson comorbidity index | ||||||||||
≥3 | 336 | 0.7 | 0.9 | <0.001 | 10.6 | 29.7 | <0.001 | 1895.5 | 3989.3 | <0.001 |
1–2 | 708 | 0.5 | 0.6 | 8.0 | 25.3 | 1585.9 | 4291.6 | |||
0 | 716 | 0.3 | 0.5 | 3.2 | 9.6 | 714.6 | 1887.5 | |||
ADL | ||||||||||
With difficulty | 236 | 0.8 | 0.8 | <0.001 | 19.9 | 46.1 | <0.001 | 3576.5 | 7264.6 | <0.001 |
No difficulties at all | 1,524 | 0.4 | 0.6 | 4.5 | 13.6 | 936.5 | 2218.5 |
USD, United States Dollar; United States Dollar to Taiwanese Dollar exchange rate = 1:30.
aTest for Mann–Whitney U test or Kruskal–Wallis test.
The overall prevalence of engaging in physical activity at moderate or higher levels (i.e. ≥1,000 kcal/week) was 37.9% (see Table 2). Participants with moderate or high levels of activity were more likely to be male, younger, better educated, married/cohabitating, currently working, with higher income, currently drinking and smoking, overweight, and had less comorbidity and no difficulties in ADL.
Variables in 2005 . | Total, N . | Overall physical activity . | P for chi-square test . | |||||||
---|---|---|---|---|---|---|---|---|---|---|
None . | Low . | Moderate . | High . | |||||||
N . | % . | N . | % . | N . | % . | N . | % . | |||
Socio-demographic | ||||||||||
Sex | ||||||||||
Female | 824 | 326 | (39.6) | 245 | (29.7) | 111 | (13.5) | 142 | (17.2) | <0.001 |
Male (≥1,000) | 936 | 279 | (29.8) | 243 | (26.0) | 151 | (16.1) | 263 | (28.1) | |
Age (years) | ||||||||||
≥85 | 97 | 58 | (59.8) | 19 | (19.6) | 8 | (8.2) | 12 | (12.4) | <0.001 |
75–84 | 595 | 247 | (41.5) | 171 | (28.7) | 79 | (13.3) | 98 | (16.5) | |
65–74 | 1,068 | 300 | (28.1) | 298 | (27.9) | 175 | (16.4) | 295 | (27.6) | |
Education level | ||||||||||
No formal schooling | 613 | 275 | (44.9) | 162 | (26.4) | 72 | (11.7) | 104 | (17.0) | <0.001 |
Primary school or more | 1,144 | 328 | (28.7) | 325 | (28.4) | 190 | (16.7) | 301 | (26.3) | |
Marital status | ||||||||||
Others | 602 | 252 | (41.9) | 177 | (29.4) | 83 | (13.8) | 90 | (15.0) | <0.001 |
Married/cohabiting | 1,158 | 353 | (30.5) | 311 | (26.9) | 179 | (15.5) | 315 | (27.2) | |
Working status (H1, H1A) | ||||||||||
Retired/unable to work | 1,154 | 402 | (34.8) | 329 | (28.5) | 191 | (16.6) | 232 | (20.1) | <0.001 |
Housekeeper | 305 | 140 | (45.9) | 94 | (30.8) | 36 | (11.8) | 35 | (11.5) | |
Currently working | 276 | 57 | (20.7) | 57 | (20.7) | 33 | (12.0) | 129 | (46.7) | |
Monthly income (NT dollar) | ||||||||||
0–9,999 | 1,189 | 460 | (38.7) | 324 | (27.2) | 155 | (13.0) | 250 | (21.0) | <0.001 |
10,000–19,999 | 280 | 81 | (28.9) | 86 | (30.7) | 46 | (16.4) | 67 | (23.9) | |
≥20,000 | 281 | 60 | (21.4) | 77 | (27.4) | 58 | (20.6) | 86 | (30.6) | |
Living status | ||||||||||
Alone | 176 | 55 | (31.3) | 45 | (25.6) | 30 | (17.0) | 46 | (26.1) | 0.507 |
With family/others | 1,584 | 550 | (34.7) | 443 | (28.0) | 232 | (14.6) | 359 | (22.7) | |
Lifestyle behaviours | ||||||||||
Drinking | ||||||||||
Yes | 365 | 93 | (25.5) | 83 | (22.7) | 64 | (17.5) | 125 | (34.2) | <0.001 |
No | 1,395 | 512 | (37.6) | 405 | (29.0) | 198 | (14.2) | 280 | (20.1) | |
Smoking | ||||||||||
Current smoker | 311 | 102 | (32.8) | 76 | (24.4) | 48 | (15.4) | 85 | (27.3) | 0.027 |
Former smoker | 204 | 69 | (33.8) | 50 | (24.5) | 24 | (11.8) | 61 | (29.9) | |
Never smoked | 1,203 | 418 | (34.7) | 350 | (29.1) | 186 | (15.5) | 249 | (20.7) | |
Health status | ||||||||||
BMI (kg/m2) | ||||||||||
Underweight (<18.5) | 105 | 59 | (56.1) | 20 | (19.0) | 9 | (8.6) | 17 | (16.2) | <0.001 |
Obese (≥27) | 327 | 108 | (33.0) | 92 | (28.1) | 62 | (19.0) | 65 | (19.9) | |
Overweight (24–26.99) | 468 | 130 | (27.8) | 136 | (29.1) | 89 | (19.0) | 113 | (24.1) | |
Normal (18.5–23.99) | 804 | 272 | (33.8) | 227 | (28.2) | 98 | (12.2) | 207 | (25.7) | |
Charlson comorbidity index | ||||||||||
≥3 | 336 | 154 | (45.8) | 91 | (27.1) | 43 | (12.8) | 48 | (14.3) | <0.001 |
1–2 | 708 | 263 | (37.1) | 193 | (27.3) | 100 | (14.1) | 152 | (24.5) | |
0 | 716 | 188 | (26.3) | 204 | (28.5) | 119 | (16.6) | 205 | (28.6) | |
ADL | ||||||||||
Some or great difficulties | 236 | 183 | (77.5) | 41 | (17.4) | 9 | (3.8) | 3 | (1.3) | <0.001 |
No difficulties at all | 1,524 | 422 | (27.7) | 447 | (29.3) | 253 | (16.6) | 402 | (26.4) |
Variables in 2005 . | Total, N . | Overall physical activity . | P for chi-square test . | |||||||
---|---|---|---|---|---|---|---|---|---|---|
None . | Low . | Moderate . | High . | |||||||
N . | % . | N . | % . | N . | % . | N . | % . | |||
Socio-demographic | ||||||||||
Sex | ||||||||||
Female | 824 | 326 | (39.6) | 245 | (29.7) | 111 | (13.5) | 142 | (17.2) | <0.001 |
Male (≥1,000) | 936 | 279 | (29.8) | 243 | (26.0) | 151 | (16.1) | 263 | (28.1) | |
Age (years) | ||||||||||
≥85 | 97 | 58 | (59.8) | 19 | (19.6) | 8 | (8.2) | 12 | (12.4) | <0.001 |
75–84 | 595 | 247 | (41.5) | 171 | (28.7) | 79 | (13.3) | 98 | (16.5) | |
65–74 | 1,068 | 300 | (28.1) | 298 | (27.9) | 175 | (16.4) | 295 | (27.6) | |
Education level | ||||||||||
No formal schooling | 613 | 275 | (44.9) | 162 | (26.4) | 72 | (11.7) | 104 | (17.0) | <0.001 |
Primary school or more | 1,144 | 328 | (28.7) | 325 | (28.4) | 190 | (16.7) | 301 | (26.3) | |
Marital status | ||||||||||
Others | 602 | 252 | (41.9) | 177 | (29.4) | 83 | (13.8) | 90 | (15.0) | <0.001 |
Married/cohabiting | 1,158 | 353 | (30.5) | 311 | (26.9) | 179 | (15.5) | 315 | (27.2) | |
Working status (H1, H1A) | ||||||||||
Retired/unable to work | 1,154 | 402 | (34.8) | 329 | (28.5) | 191 | (16.6) | 232 | (20.1) | <0.001 |
Housekeeper | 305 | 140 | (45.9) | 94 | (30.8) | 36 | (11.8) | 35 | (11.5) | |
Currently working | 276 | 57 | (20.7) | 57 | (20.7) | 33 | (12.0) | 129 | (46.7) | |
Monthly income (NT dollar) | ||||||||||
0–9,999 | 1,189 | 460 | (38.7) | 324 | (27.2) | 155 | (13.0) | 250 | (21.0) | <0.001 |
10,000–19,999 | 280 | 81 | (28.9) | 86 | (30.7) | 46 | (16.4) | 67 | (23.9) | |
≥20,000 | 281 | 60 | (21.4) | 77 | (27.4) | 58 | (20.6) | 86 | (30.6) | |
Living status | ||||||||||
Alone | 176 | 55 | (31.3) | 45 | (25.6) | 30 | (17.0) | 46 | (26.1) | 0.507 |
With family/others | 1,584 | 550 | (34.7) | 443 | (28.0) | 232 | (14.6) | 359 | (22.7) | |
Lifestyle behaviours | ||||||||||
Drinking | ||||||||||
Yes | 365 | 93 | (25.5) | 83 | (22.7) | 64 | (17.5) | 125 | (34.2) | <0.001 |
No | 1,395 | 512 | (37.6) | 405 | (29.0) | 198 | (14.2) | 280 | (20.1) | |
Smoking | ||||||||||
Current smoker | 311 | 102 | (32.8) | 76 | (24.4) | 48 | (15.4) | 85 | (27.3) | 0.027 |
Former smoker | 204 | 69 | (33.8) | 50 | (24.5) | 24 | (11.8) | 61 | (29.9) | |
Never smoked | 1,203 | 418 | (34.7) | 350 | (29.1) | 186 | (15.5) | 249 | (20.7) | |
Health status | ||||||||||
BMI (kg/m2) | ||||||||||
Underweight (<18.5) | 105 | 59 | (56.1) | 20 | (19.0) | 9 | (8.6) | 17 | (16.2) | <0.001 |
Obese (≥27) | 327 | 108 | (33.0) | 92 | (28.1) | 62 | (19.0) | 65 | (19.9) | |
Overweight (24–26.99) | 468 | 130 | (27.8) | 136 | (29.1) | 89 | (19.0) | 113 | (24.1) | |
Normal (18.5–23.99) | 804 | 272 | (33.8) | 227 | (28.2) | 98 | (12.2) | 207 | (25.7) | |
Charlson comorbidity index | ||||||||||
≥3 | 336 | 154 | (45.8) | 91 | (27.1) | 43 | (12.8) | 48 | (14.3) | <0.001 |
1–2 | 708 | 263 | (37.1) | 193 | (27.3) | 100 | (14.1) | 152 | (24.5) | |
0 | 716 | 188 | (26.3) | 204 | (28.5) | 119 | (16.6) | 205 | (28.6) | |
ADL | ||||||||||
Some or great difficulties | 236 | 183 | (77.5) | 41 | (17.4) | 9 | (3.8) | 3 | (1.3) | <0.001 |
No difficulties at all | 1,524 | 422 | (27.7) | 447 | (29.3) | 253 | (16.6) | 402 | (26.4) |
Variables in 2005 . | Total, N . | Overall physical activity . | P for chi-square test . | |||||||
---|---|---|---|---|---|---|---|---|---|---|
None . | Low . | Moderate . | High . | |||||||
N . | % . | N . | % . | N . | % . | N . | % . | |||
Socio-demographic | ||||||||||
Sex | ||||||||||
Female | 824 | 326 | (39.6) | 245 | (29.7) | 111 | (13.5) | 142 | (17.2) | <0.001 |
Male (≥1,000) | 936 | 279 | (29.8) | 243 | (26.0) | 151 | (16.1) | 263 | (28.1) | |
Age (years) | ||||||||||
≥85 | 97 | 58 | (59.8) | 19 | (19.6) | 8 | (8.2) | 12 | (12.4) | <0.001 |
75–84 | 595 | 247 | (41.5) | 171 | (28.7) | 79 | (13.3) | 98 | (16.5) | |
65–74 | 1,068 | 300 | (28.1) | 298 | (27.9) | 175 | (16.4) | 295 | (27.6) | |
Education level | ||||||||||
No formal schooling | 613 | 275 | (44.9) | 162 | (26.4) | 72 | (11.7) | 104 | (17.0) | <0.001 |
Primary school or more | 1,144 | 328 | (28.7) | 325 | (28.4) | 190 | (16.7) | 301 | (26.3) | |
Marital status | ||||||||||
Others | 602 | 252 | (41.9) | 177 | (29.4) | 83 | (13.8) | 90 | (15.0) | <0.001 |
Married/cohabiting | 1,158 | 353 | (30.5) | 311 | (26.9) | 179 | (15.5) | 315 | (27.2) | |
Working status (H1, H1A) | ||||||||||
Retired/unable to work | 1,154 | 402 | (34.8) | 329 | (28.5) | 191 | (16.6) | 232 | (20.1) | <0.001 |
Housekeeper | 305 | 140 | (45.9) | 94 | (30.8) | 36 | (11.8) | 35 | (11.5) | |
Currently working | 276 | 57 | (20.7) | 57 | (20.7) | 33 | (12.0) | 129 | (46.7) | |
Monthly income (NT dollar) | ||||||||||
0–9,999 | 1,189 | 460 | (38.7) | 324 | (27.2) | 155 | (13.0) | 250 | (21.0) | <0.001 |
10,000–19,999 | 280 | 81 | (28.9) | 86 | (30.7) | 46 | (16.4) | 67 | (23.9) | |
≥20,000 | 281 | 60 | (21.4) | 77 | (27.4) | 58 | (20.6) | 86 | (30.6) | |
Living status | ||||||||||
Alone | 176 | 55 | (31.3) | 45 | (25.6) | 30 | (17.0) | 46 | (26.1) | 0.507 |
With family/others | 1,584 | 550 | (34.7) | 443 | (28.0) | 232 | (14.6) | 359 | (22.7) | |
Lifestyle behaviours | ||||||||||
Drinking | ||||||||||
Yes | 365 | 93 | (25.5) | 83 | (22.7) | 64 | (17.5) | 125 | (34.2) | <0.001 |
No | 1,395 | 512 | (37.6) | 405 | (29.0) | 198 | (14.2) | 280 | (20.1) | |
Smoking | ||||||||||
Current smoker | 311 | 102 | (32.8) | 76 | (24.4) | 48 | (15.4) | 85 | (27.3) | 0.027 |
Former smoker | 204 | 69 | (33.8) | 50 | (24.5) | 24 | (11.8) | 61 | (29.9) | |
Never smoked | 1,203 | 418 | (34.7) | 350 | (29.1) | 186 | (15.5) | 249 | (20.7) | |
Health status | ||||||||||
BMI (kg/m2) | ||||||||||
Underweight (<18.5) | 105 | 59 | (56.1) | 20 | (19.0) | 9 | (8.6) | 17 | (16.2) | <0.001 |
Obese (≥27) | 327 | 108 | (33.0) | 92 | (28.1) | 62 | (19.0) | 65 | (19.9) | |
Overweight (24–26.99) | 468 | 130 | (27.8) | 136 | (29.1) | 89 | (19.0) | 113 | (24.1) | |
Normal (18.5–23.99) | 804 | 272 | (33.8) | 227 | (28.2) | 98 | (12.2) | 207 | (25.7) | |
Charlson comorbidity index | ||||||||||
≥3 | 336 | 154 | (45.8) | 91 | (27.1) | 43 | (12.8) | 48 | (14.3) | <0.001 |
1–2 | 708 | 263 | (37.1) | 193 | (27.3) | 100 | (14.1) | 152 | (24.5) | |
0 | 716 | 188 | (26.3) | 204 | (28.5) | 119 | (16.6) | 205 | (28.6) | |
ADL | ||||||||||
Some or great difficulties | 236 | 183 | (77.5) | 41 | (17.4) | 9 | (3.8) | 3 | (1.3) | <0.001 |
No difficulties at all | 1,524 | 422 | (27.7) | 447 | (29.3) | 253 | (16.6) | 402 | (26.4) |
Variables in 2005 . | Total, N . | Overall physical activity . | P for chi-square test . | |||||||
---|---|---|---|---|---|---|---|---|---|---|
None . | Low . | Moderate . | High . | |||||||
N . | % . | N . | % . | N . | % . | N . | % . | |||
Socio-demographic | ||||||||||
Sex | ||||||||||
Female | 824 | 326 | (39.6) | 245 | (29.7) | 111 | (13.5) | 142 | (17.2) | <0.001 |
Male (≥1,000) | 936 | 279 | (29.8) | 243 | (26.0) | 151 | (16.1) | 263 | (28.1) | |
Age (years) | ||||||||||
≥85 | 97 | 58 | (59.8) | 19 | (19.6) | 8 | (8.2) | 12 | (12.4) | <0.001 |
75–84 | 595 | 247 | (41.5) | 171 | (28.7) | 79 | (13.3) | 98 | (16.5) | |
65–74 | 1,068 | 300 | (28.1) | 298 | (27.9) | 175 | (16.4) | 295 | (27.6) | |
Education level | ||||||||||
No formal schooling | 613 | 275 | (44.9) | 162 | (26.4) | 72 | (11.7) | 104 | (17.0) | <0.001 |
Primary school or more | 1,144 | 328 | (28.7) | 325 | (28.4) | 190 | (16.7) | 301 | (26.3) | |
Marital status | ||||||||||
Others | 602 | 252 | (41.9) | 177 | (29.4) | 83 | (13.8) | 90 | (15.0) | <0.001 |
Married/cohabiting | 1,158 | 353 | (30.5) | 311 | (26.9) | 179 | (15.5) | 315 | (27.2) | |
Working status (H1, H1A) | ||||||||||
Retired/unable to work | 1,154 | 402 | (34.8) | 329 | (28.5) | 191 | (16.6) | 232 | (20.1) | <0.001 |
Housekeeper | 305 | 140 | (45.9) | 94 | (30.8) | 36 | (11.8) | 35 | (11.5) | |
Currently working | 276 | 57 | (20.7) | 57 | (20.7) | 33 | (12.0) | 129 | (46.7) | |
Monthly income (NT dollar) | ||||||||||
0–9,999 | 1,189 | 460 | (38.7) | 324 | (27.2) | 155 | (13.0) | 250 | (21.0) | <0.001 |
10,000–19,999 | 280 | 81 | (28.9) | 86 | (30.7) | 46 | (16.4) | 67 | (23.9) | |
≥20,000 | 281 | 60 | (21.4) | 77 | (27.4) | 58 | (20.6) | 86 | (30.6) | |
Living status | ||||||||||
Alone | 176 | 55 | (31.3) | 45 | (25.6) | 30 | (17.0) | 46 | (26.1) | 0.507 |
With family/others | 1,584 | 550 | (34.7) | 443 | (28.0) | 232 | (14.6) | 359 | (22.7) | |
Lifestyle behaviours | ||||||||||
Drinking | ||||||||||
Yes | 365 | 93 | (25.5) | 83 | (22.7) | 64 | (17.5) | 125 | (34.2) | <0.001 |
No | 1,395 | 512 | (37.6) | 405 | (29.0) | 198 | (14.2) | 280 | (20.1) | |
Smoking | ||||||||||
Current smoker | 311 | 102 | (32.8) | 76 | (24.4) | 48 | (15.4) | 85 | (27.3) | 0.027 |
Former smoker | 204 | 69 | (33.8) | 50 | (24.5) | 24 | (11.8) | 61 | (29.9) | |
Never smoked | 1,203 | 418 | (34.7) | 350 | (29.1) | 186 | (15.5) | 249 | (20.7) | |
Health status | ||||||||||
BMI (kg/m2) | ||||||||||
Underweight (<18.5) | 105 | 59 | (56.1) | 20 | (19.0) | 9 | (8.6) | 17 | (16.2) | <0.001 |
Obese (≥27) | 327 | 108 | (33.0) | 92 | (28.1) | 62 | (19.0) | 65 | (19.9) | |
Overweight (24–26.99) | 468 | 130 | (27.8) | 136 | (29.1) | 89 | (19.0) | 113 | (24.1) | |
Normal (18.5–23.99) | 804 | 272 | (33.8) | 227 | (28.2) | 98 | (12.2) | 207 | (25.7) | |
Charlson comorbidity index | ||||||||||
≥3 | 336 | 154 | (45.8) | 91 | (27.1) | 43 | (12.8) | 48 | (14.3) | <0.001 |
1–2 | 708 | 263 | (37.1) | 193 | (27.3) | 100 | (14.1) | 152 | (24.5) | |
0 | 716 | 188 | (26.3) | 204 | (28.5) | 119 | (16.6) | 205 | (28.6) | |
ADL | ||||||||||
Some or great difficulties | 236 | 183 | (77.5) | 41 | (17.4) | 9 | (3.8) | 3 | (1.3) | <0.001 |
No difficulties at all | 1,524 | 422 | (27.7) | 447 | (29.3) | 253 | (16.6) | 402 | (26.4) |
The results of the fully adjusted multivariable regression models (Model 1 in Table 3) show that those who engaged in moderate or higher levels of physical activity had a significantly lower number of hospital admissions and fewer days in hospital, than those who were not active. Differences were substantial, with an approximate 30% reduction in number of hospital admissions and number of days in hospital among more active participants. It seemed that older adults with high levels of physical activity (i.e. ≥2000 kcal/week) did not show a further large reduction in hospital care utilisation in comparison with those with moderate levels of physical activity (i.e. 1,000–1,999 kcal/week). Although the tests for linear trend supported an inverse dose–response relationship between physical activity energy expenditure and hospital care costs, it appeared that only individuals with high levels of physical activity had a significant but small reduction of expenditure (approximately 10%). A similar pattern was also observed in the sensitivity analyses (Models 2–4), although effects for moderate physical activity were not significant in Model 4, probably because of the smaller sample size.
. | Annual mean number of hospitalisation . | Annual mean number of hospitalisation days . | Annual mean of hospitalisation costs . | |||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
n . | RRs . | 95% CI . | P value . | n . | RRs . | 95% CI . | P value . | n . | RRs . | 95% CI . | P value . | |
Model 1 | 1,637 | 1,637 | 1,640 | |||||||||
PA | 0.002a | 0.024a | 0.033a | |||||||||
None (ref.) | 1.00 | ─ | ─ | 1.00 | ─ | ─ | 1.00 | ─ | ─ | |||
Low | 0.87 | (0.71–1.05) | 0.145 | 1.09 | (0.88–1.36) | 0.422 | 1.00 | (0.91–1.08) | 0.908 | |||
Moderate | 0.71 | (0.55–0.92) | 0.010 | 0.67 | (0.51–0.87) | 0.003 | 0.98 | (0.89–1.09) | 0.734 | |||
High | 0.72 | (0.57–0.92) | 0.008 | 0.65 | (0.50–0.83) | <0.001 | 0.89 | (0.80–0.99) | 0.033 | |||
Model 2 | 1,605 | 1,605 | 1,608 | |||||||||
PA | 0.001a | 0.002a | 0.047a | |||||||||
None (ref.) | 1.00 | ─ | ─ | 1.00 | ─ | ─ | 1.00 | ─ | ─ | |||
Low | 0.89 | (0.73–1.09) | 0.249 | 1.04 | (0.84–1.28) | 0.722 | 1.01 | (0.92–1.10) | 0.884 | |||
Moderate | 0.73 | (0.56–0.95) | 0.018 | 0.74 | (0.57–0.96) | 0.025 | 0.99 | (0.89–1.10) | 0.816 | |||
High | 0.69 | (0.54–0.89) | 0.004 | 0.70 | (0.55–0.89) | 0.003 | 0.90 | (0.81–1.00) | 0.053 | |||
Model 3 | 1,464 | 1,464 | 1,466 | |||||||||
PA | 0.001a | 0.001a | 0.024a | |||||||||
None (ref.) | 1.00 | ─ | ─ | 1.00 | ─ | ─ | 1.00 | ─ | ─ | |||
Low | 0.90 | (0.73–1.12) | 0.343 | 1.10 | (0.89–1.37) | 0.364 | 1.02 | (0.93–1.11) | 0.714 | |||
Moderate | 0.73 | (0.55–0.98) | 0.033 | 0.74 | (0.57–0.97) | 0.026 | 0.98 | (0.88–1.09) | 0.734 | |||
High | 0.70 | (0.54–0.91) | 0.009 | 0.70 | (0.55–0.89) | 0.004 | 0.89 | (0.80–0.99) | 0.031 | |||
Model 4 | 1,203 | 1,203 | 1,203 | |||||||||
PA | 0.003a | 0.006a | 0.051a | |||||||||
None (ref.) | 1.00 | ─ | ─ | 1.00 | ─ | ─ | 1.00 | ─ | ─ | |||
Low | 0.84 | (0.66–1.08) | 0.168 | 1.04 | (0.80–1.35) | 0.773 | 0.97 | (0.88–1.08) | 0.609 | |||
Moderate | 0.76 | (0.56–1.03) | 0.075 | 0.78 | (0.58–1.05) | 0.102 | 1.00 | (0.88–1.12) | 0.942 | |||
High | 0.68 | (0.51–0.90) | 0.007 | 0.72 | (0.55–0.96) | 0.024 | 0.87 | (0.78–0.98) | 0.026 |
. | Annual mean number of hospitalisation . | Annual mean number of hospitalisation days . | Annual mean of hospitalisation costs . | |||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
n . | RRs . | 95% CI . | P value . | n . | RRs . | 95% CI . | P value . | n . | RRs . | 95% CI . | P value . | |
Model 1 | 1,637 | 1,637 | 1,640 | |||||||||
PA | 0.002a | 0.024a | 0.033a | |||||||||
None (ref.) | 1.00 | ─ | ─ | 1.00 | ─ | ─ | 1.00 | ─ | ─ | |||
Low | 0.87 | (0.71–1.05) | 0.145 | 1.09 | (0.88–1.36) | 0.422 | 1.00 | (0.91–1.08) | 0.908 | |||
Moderate | 0.71 | (0.55–0.92) | 0.010 | 0.67 | (0.51–0.87) | 0.003 | 0.98 | (0.89–1.09) | 0.734 | |||
High | 0.72 | (0.57–0.92) | 0.008 | 0.65 | (0.50–0.83) | <0.001 | 0.89 | (0.80–0.99) | 0.033 | |||
Model 2 | 1,605 | 1,605 | 1,608 | |||||||||
PA | 0.001a | 0.002a | 0.047a | |||||||||
None (ref.) | 1.00 | ─ | ─ | 1.00 | ─ | ─ | 1.00 | ─ | ─ | |||
Low | 0.89 | (0.73–1.09) | 0.249 | 1.04 | (0.84–1.28) | 0.722 | 1.01 | (0.92–1.10) | 0.884 | |||
Moderate | 0.73 | (0.56–0.95) | 0.018 | 0.74 | (0.57–0.96) | 0.025 | 0.99 | (0.89–1.10) | 0.816 | |||
High | 0.69 | (0.54–0.89) | 0.004 | 0.70 | (0.55–0.89) | 0.003 | 0.90 | (0.81–1.00) | 0.053 | |||
Model 3 | 1,464 | 1,464 | 1,466 | |||||||||
PA | 0.001a | 0.001a | 0.024a | |||||||||
None (ref.) | 1.00 | ─ | ─ | 1.00 | ─ | ─ | 1.00 | ─ | ─ | |||
Low | 0.90 | (0.73–1.12) | 0.343 | 1.10 | (0.89–1.37) | 0.364 | 1.02 | (0.93–1.11) | 0.714 | |||
Moderate | 0.73 | (0.55–0.98) | 0.033 | 0.74 | (0.57–0.97) | 0.026 | 0.98 | (0.88–1.09) | 0.734 | |||
High | 0.70 | (0.54–0.91) | 0.009 | 0.70 | (0.55–0.89) | 0.004 | 0.89 | (0.80–0.99) | 0.031 | |||
Model 4 | 1,203 | 1,203 | 1,203 | |||||||||
PA | 0.003a | 0.006a | 0.051a | |||||||||
None (ref.) | 1.00 | ─ | ─ | 1.00 | ─ | ─ | 1.00 | ─ | ─ | |||
Low | 0.84 | (0.66–1.08) | 0.168 | 1.04 | (0.80–1.35) | 0.773 | 0.97 | (0.88–1.08) | 0.609 | |||
Moderate | 0.76 | (0.56–1.03) | 0.075 | 0.78 | (0.58–1.05) | 0.102 | 1.00 | (0.88–1.12) | 0.942 | |||
High | 0.68 | (0.51–0.90) | 0.007 | 0.72 | (0.55–0.96) | 0.024 | 0.87 | (0.78–0.98) | 0.026 |
Model 2 = Model 1 excluding those with an extremely high hospitalisation cost in 2005 and the follow-up period, 2006–12;Model 3 = based on Model 2 excluding those died in 2005–2007;Model 4 = based on Model 2 excluding those with ADL, arthritis and COPD in 2005.PA, physical activity.Covariates in all models: baseline hospital care utilisation, sex (only adjusting for annual mean number of hospitalisation), age, education level (only adjusting for annual mean number of hospitalisation and annual mean number of hospitalisation days), marital status, working status, alcohol consumption, smoking, BMI, Charlson comorbidity index, and ADL.
aP value for trend using linear regression with multivariable adjustment.
. | Annual mean number of hospitalisation . | Annual mean number of hospitalisation days . | Annual mean of hospitalisation costs . | |||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
n . | RRs . | 95% CI . | P value . | n . | RRs . | 95% CI . | P value . | n . | RRs . | 95% CI . | P value . | |
Model 1 | 1,637 | 1,637 | 1,640 | |||||||||
PA | 0.002a | 0.024a | 0.033a | |||||||||
None (ref.) | 1.00 | ─ | ─ | 1.00 | ─ | ─ | 1.00 | ─ | ─ | |||
Low | 0.87 | (0.71–1.05) | 0.145 | 1.09 | (0.88–1.36) | 0.422 | 1.00 | (0.91–1.08) | 0.908 | |||
Moderate | 0.71 | (0.55–0.92) | 0.010 | 0.67 | (0.51–0.87) | 0.003 | 0.98 | (0.89–1.09) | 0.734 | |||
High | 0.72 | (0.57–0.92) | 0.008 | 0.65 | (0.50–0.83) | <0.001 | 0.89 | (0.80–0.99) | 0.033 | |||
Model 2 | 1,605 | 1,605 | 1,608 | |||||||||
PA | 0.001a | 0.002a | 0.047a | |||||||||
None (ref.) | 1.00 | ─ | ─ | 1.00 | ─ | ─ | 1.00 | ─ | ─ | |||
Low | 0.89 | (0.73–1.09) | 0.249 | 1.04 | (0.84–1.28) | 0.722 | 1.01 | (0.92–1.10) | 0.884 | |||
Moderate | 0.73 | (0.56–0.95) | 0.018 | 0.74 | (0.57–0.96) | 0.025 | 0.99 | (0.89–1.10) | 0.816 | |||
High | 0.69 | (0.54–0.89) | 0.004 | 0.70 | (0.55–0.89) | 0.003 | 0.90 | (0.81–1.00) | 0.053 | |||
Model 3 | 1,464 | 1,464 | 1,466 | |||||||||
PA | 0.001a | 0.001a | 0.024a | |||||||||
None (ref.) | 1.00 | ─ | ─ | 1.00 | ─ | ─ | 1.00 | ─ | ─ | |||
Low | 0.90 | (0.73–1.12) | 0.343 | 1.10 | (0.89–1.37) | 0.364 | 1.02 | (0.93–1.11) | 0.714 | |||
Moderate | 0.73 | (0.55–0.98) | 0.033 | 0.74 | (0.57–0.97) | 0.026 | 0.98 | (0.88–1.09) | 0.734 | |||
High | 0.70 | (0.54–0.91) | 0.009 | 0.70 | (0.55–0.89) | 0.004 | 0.89 | (0.80–0.99) | 0.031 | |||
Model 4 | 1,203 | 1,203 | 1,203 | |||||||||
PA | 0.003a | 0.006a | 0.051a | |||||||||
None (ref.) | 1.00 | ─ | ─ | 1.00 | ─ | ─ | 1.00 | ─ | ─ | |||
Low | 0.84 | (0.66–1.08) | 0.168 | 1.04 | (0.80–1.35) | 0.773 | 0.97 | (0.88–1.08) | 0.609 | |||
Moderate | 0.76 | (0.56–1.03) | 0.075 | 0.78 | (0.58–1.05) | 0.102 | 1.00 | (0.88–1.12) | 0.942 | |||
High | 0.68 | (0.51–0.90) | 0.007 | 0.72 | (0.55–0.96) | 0.024 | 0.87 | (0.78–0.98) | 0.026 |
. | Annual mean number of hospitalisation . | Annual mean number of hospitalisation days . | Annual mean of hospitalisation costs . | |||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
n . | RRs . | 95% CI . | P value . | n . | RRs . | 95% CI . | P value . | n . | RRs . | 95% CI . | P value . | |
Model 1 | 1,637 | 1,637 | 1,640 | |||||||||
PA | 0.002a | 0.024a | 0.033a | |||||||||
None (ref.) | 1.00 | ─ | ─ | 1.00 | ─ | ─ | 1.00 | ─ | ─ | |||
Low | 0.87 | (0.71–1.05) | 0.145 | 1.09 | (0.88–1.36) | 0.422 | 1.00 | (0.91–1.08) | 0.908 | |||
Moderate | 0.71 | (0.55–0.92) | 0.010 | 0.67 | (0.51–0.87) | 0.003 | 0.98 | (0.89–1.09) | 0.734 | |||
High | 0.72 | (0.57–0.92) | 0.008 | 0.65 | (0.50–0.83) | <0.001 | 0.89 | (0.80–0.99) | 0.033 | |||
Model 2 | 1,605 | 1,605 | 1,608 | |||||||||
PA | 0.001a | 0.002a | 0.047a | |||||||||
None (ref.) | 1.00 | ─ | ─ | 1.00 | ─ | ─ | 1.00 | ─ | ─ | |||
Low | 0.89 | (0.73–1.09) | 0.249 | 1.04 | (0.84–1.28) | 0.722 | 1.01 | (0.92–1.10) | 0.884 | |||
Moderate | 0.73 | (0.56–0.95) | 0.018 | 0.74 | (0.57–0.96) | 0.025 | 0.99 | (0.89–1.10) | 0.816 | |||
High | 0.69 | (0.54–0.89) | 0.004 | 0.70 | (0.55–0.89) | 0.003 | 0.90 | (0.81–1.00) | 0.053 | |||
Model 3 | 1,464 | 1,464 | 1,466 | |||||||||
PA | 0.001a | 0.001a | 0.024a | |||||||||
None (ref.) | 1.00 | ─ | ─ | 1.00 | ─ | ─ | 1.00 | ─ | ─ | |||
Low | 0.90 | (0.73–1.12) | 0.343 | 1.10 | (0.89–1.37) | 0.364 | 1.02 | (0.93–1.11) | 0.714 | |||
Moderate | 0.73 | (0.55–0.98) | 0.033 | 0.74 | (0.57–0.97) | 0.026 | 0.98 | (0.88–1.09) | 0.734 | |||
High | 0.70 | (0.54–0.91) | 0.009 | 0.70 | (0.55–0.89) | 0.004 | 0.89 | (0.80–0.99) | 0.031 | |||
Model 4 | 1,203 | 1,203 | 1,203 | |||||||||
PA | 0.003a | 0.006a | 0.051a | |||||||||
None (ref.) | 1.00 | ─ | ─ | 1.00 | ─ | ─ | 1.00 | ─ | ─ | |||
Low | 0.84 | (0.66–1.08) | 0.168 | 1.04 | (0.80–1.35) | 0.773 | 0.97 | (0.88–1.08) | 0.609 | |||
Moderate | 0.76 | (0.56–1.03) | 0.075 | 0.78 | (0.58–1.05) | 0.102 | 1.00 | (0.88–1.12) | 0.942 | |||
High | 0.68 | (0.51–0.90) | 0.007 | 0.72 | (0.55–0.96) | 0.024 | 0.87 | (0.78–0.98) | 0.026 |
Model 2 = Model 1 excluding those with an extremely high hospitalisation cost in 2005 and the follow-up period, 2006–12;Model 3 = based on Model 2 excluding those died in 2005–2007;Model 4 = based on Model 2 excluding those with ADL, arthritis and COPD in 2005.PA, physical activity.Covariates in all models: baseline hospital care utilisation, sex (only adjusting for annual mean number of hospitalisation), age, education level (only adjusting for annual mean number of hospitalisation and annual mean number of hospitalisation days), marital status, working status, alcohol consumption, smoking, BMI, Charlson comorbidity index, and ADL.
aP value for trend using linear regression with multivariable adjustment.
Discussion
This population-based 7-year follow-up study revealed that older adults engaging in regular physical activity, especially at moderate or higher levels (≥1,000 kcal/week), have less utilisation of hospital care services, including hospital admissions and hospitalisation days over subsequent years. Trends for reduced hospitalisation costs were also found, but this association was only identified in those reporting high levels of physical activity (≥2,000 kcal/week). These results were observed in multivariable regression models after adjusting for the baseline outcome measurements, socio-demographic variables, lifestyle behaviours and comorbid conditions. Sensitivity analyses for assessing reverse causation provided further support to the robustness of these findings.
This study demonstrates that late-life physical activity is associated with lower utilisation of subsequent hospital care services, supporting the previous findings of prospective studies based on community-dwelling older adults [11, 12]. The protective effect of physical activity on subsequent hospital care utilisation in this study, although substantial, is smaller than that described in previous studies [10–12]. This may be due to the more comprehensive covariates (such as baseline hospital care utilisation) included in the present study. In contrast, the results of previous studies were estimated without controlling for the baseline status of hospital care utilisation, which may lead to the overestimation of the strength of associations. Differences in the prospective associations of physical activity with hospital care utilisation may also be due to methodological variations in quantifying and reporting physical activity and the heterogeneity of samples.
This study indicated that older adults who are at least moderately active (the equivalent of walking for roughly ≥30 minutes/day) [29] had lower utilisation of subsequent hospital care services such as hospital admission and hospitalisation days. Notably, more intense physical activity (i.e. expending ≥2,000 kcal/week) did not yield a further large reduction in hospital care utilisation in comparison with moderate levels of physical activity (i.e. 1,000–1,999 kcal/week). This echoes the recommendations for the public that call for ≥30 minutes of at least moderate-intensity physical activity on most days of the week [3]. These findings were further confirmed using sensitivity analyses in Models 2–4 and were not due to the inclusion of very high hospital care users (Model 2).
The trend across physical activity categories and hospital care costs was significant in this study, but the significant and small effect on cost reduction (approximately 10%) was only found in the group reporting high physical activity. The association remained significant even when participants who had extreme hospitalisation costs were excluded. It is possible that the determinants of hospital care spending are different from those of hospital admission and hospitalisation days. Costs not only relate to the severity of the condition but also to patients’ treatment preferences and physicians’ beliefs and financial incentives [30]. These underlying factors may attenuate the relationship of physical activity with hospitalisation costs.
Although this prospective study was conducted adjusting for a comprehensive range of potential confounders based on a nationally representative sample, it inevitably has some limitations. First, the assessment of physical activity, including leisure-time, household and work-related physical activity, were self-reported, so may have been susceptible to recall bias, especially in ageing populations. Physical activity was assessed using a single measurement, but levels may change over time. Finally, although the hospital care data were based on the compulsory National Health Insurance program for all citizens, 35% of participants in 2005 Taiwan NHIS did not provide the consent to link their claim data. There were no significant differences between individuals who did and did not consent in terms of age at baseline. However, those who did not consent to link the claim data were more likely to be female, have lower education attainment, be physically inactive and experience more chronic conditions and difficulties with ADL (all P values for chi-square tests < 0.05, data not shown). Because physical inactivity and hospital care utilisation are likely to be more common in this sector of the population, any selection bias may have led to underestimation of the association between physical activity and hospital care utilisation. This is an observational study, which cannot establish causality. A well-designed large-scale randomised controlled trial with a representative sample would be needed to verify these findings.
This study has provided evidence that older adults who are at least moderately active may reduce utilisation of hospital care services. The findings highlight the importance of maintaining a physically active lifestyle in later life.
Acknowledgements
The authors would like to thank The Health and Welfare Data Science Centre, Ministry of Health and Welfare (HWDC, MOHW), for support and the Healthcare Service Research Centre (HSRC) of Taichung Veterans General Hospital for statistical support.
Conflicts of interest
None were reported.
Funding
This study was supported by grants from the Taiwan Ministry of Science and Technology (NSC 101-2314-B-075A-008; MOST 103-2314-B-075A-006).
Comments