Abstract
Background
Daily step count has emerged as a practical and objective metric of physical activity. However, evidence linking step volume to multidimensional health-related outcomes in Asian older adults remains limited. This study aimed to examine the dose–response associations between objectively measured daily step counts and quality of life, knee-specific functional outcomes, and physical performance in community-dwelling older adults, and to identify clinically meaningful step count thresholds.
Methods
This nationwide cross-sectional study included 2,603 Thai adults aged ≥ 60 years recruited using multi-stage probability sampling. Daily step counts were measured using wrist-worn accelerometers over 7 days and categorized into six groups (≤ 3,000 to > 15,000 steps/day). Outcomes included EuroQol 5-Dimension 5-Level (EQ-5D-5 L) utility, EuroQol Visual Analogue Scale (EQ-VAS), Knee Injury and Osteoarthritis Outcome Score (KOOS), Oxford Knee Score (OKS), Timed Up and Go test (TUG), Five Times Sit-to-Stand test (5×STS), gait speed, and leg strength. Multivariable linear regression models adjusted for age, sex, BMI, comorbidity, knee osteoarthritis, and geographic region were used to evaluate independent associations.
Results
The mean daily step count was 9,488 ± 5,639 steps per day. All eight outcomes demonstrated significant monotonic dose–response relationships across step count groups (p < 0.001 for trend). After full adjustment, the highest step count group (> 15,000 steps/day) had significantly better outcomes than the lowest (≤ 3,000 steps/day), including shorter TUG time (− 3.7 s), faster gait speed (+ 0.20 m/s), higher KOOS (+ 9.5 points), and higher EQ-5D-5 L utility (+ 0.087) (all p < 0.001), with several differences exceeding established minimal clinically important thresholds. Associations were most pronounced among participants aged > 75 years.
Conclusions
Higher daily step counts are independently associated with better quality of life, knee-specific function, and physical performance among community-dwelling Thai older adults. A target of 6,000–9,000 steps per day appears clinically meaningful and pragmatically attainable. Daily step count may serve as a practical and scalable metric to guide healthy aging strategies in Asian populations.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12877-026-07657-7.
Keywords: Daily step count, Older adults, Quality of life, Physical performance, Dose–response, Asia
Introduction
The global population is aging rapidly. The World Health Organization projects that the proportion of individuals aged 60 years or older will nearly double, from 12% to 22%, by 2050 [1]. Thailand exemplifies this demographic transition, with approximately 20% of its population—nearly 14 million people—now classified as older adults [2]. Promoting healthy aging has therefore become a public health priority. The WHO Healthy Aging framework conceptualizes healthy aging as the development and maintenance of functional ability that enables well-being in older age, emphasizing the importance of behavioral determinants—including physical activity—in preserving functional capacity across the life course [3]. Physical activity is a cornerstone of healthy aging, with well-established benefits for cardiovascular health, cognitive function, mental well-being, and longevity [4]. Among available forms of activity, walking is the most accessible and widely practiced by older adults. The widespread adoption of wearable devices has further enabled objective quantification of walking behavior as daily step counts.
Several large-scale epidemiological studies have demonstrated dose–response relationships between daily step counts and major health outcomes. In the Women’s Health Study (n = 16,741; mean age 72 years), mortality risk decreased progressively and plateaued at approximately 7,500 steps per day [5]. An analysis of the National Health and Nutrition Examination Survey (NHANES) cohort (n = 4,840) reported that accumulating 8,000 versus 4,000 steps per day was associated with a 51% lower risk of all-cause mortality [6], a finding subsequently supported by a meta-analysis of 15 international cohorts reporting a plateau at 6,000–8,000 steps per day among adults aged ≥ 60 years [7]. Beyond mortality, the Multicenter Osteoarthritis Study demonstrated that each additional 1,000 steps per day was associated with a 16–18% lower 2-year risk of incident functional limitation, and that fewer than 6,000 steps per day identified individuals at highest risk [8]. Collectively, these studies provide robust evidence for step count targets; however, the majority have focused primarily on mortality or isolated functional measures, leaving broader multidimensional health outcomes insufficiently explored.
Several important evidence gaps remain. First, the association between objectively measured step counts and health-related quality of life—particularly as assessed using standardized instruments such as the EuroQol 5-Dimension 5-Level (EQ-5D-5 L)—has not been adequately examined. Second, knee-specific functional scores, including the Knee Injury and Osteoarthritis Outcome Score (KOOS) and Oxford Knee Score (OKS), and comprehensive performance-based assessments—including the Timed Up and Go (TUG) test, five-times sit-to-stand (5×STS), gait speed, and leg strength—have not been evaluated concurrently in relation to step count thresholds in community-dwelling older adults. Finally, virtually all existing evidence derives from Western populations, despite potential differences in lifestyle, body composition, sociocultural context, and habitual activity patterns that may modify the step count–health relationship in Asian settings. Consistent with the International Classification of Functioning, Disability and Health (ICF) model, which frames ambulatory physical activity as a key behavioral determinant of functional capacity and participation, there is a critical need for studies that simultaneously assess multidimensional health outcomes—encompassing quality of life, disease-specific function, and objective physical performance—using objective measures in Asian older adult populations [9].
We therefore conducted a nationwide cross-sectional study in Thailand with three objectives: (1) to examine the associations between objectively measured daily step counts and quality of life, knee-specific functional scores, and performance-based physical function in community-dwelling older adults; (2) to identify step count thresholds associated with clinically meaningful differences in these outcomes; and (3) to explore potential effect modification by sex and age group. To our knowledge, this study provides the first comprehensive evidence from an Asian population linking daily ambulatory activity to a broad spectrum of health-related outcomes in older adults.
Methods
Study design and participants
This cross-sectional study was conducted as part of a nationwide health survey of older adults in Thailand between April 20, 2023, and August 19, 2025. The study protocol was approved by the Institutional Review Board of Siriraj Hospital (Protocol: CREC054/66SR-SBR3). Written informed consent was obtained from all participants. The study was conducted in accordance with the Declaration of Helsinki and reported following the STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) guidelines. Data were collected using a structured interview form developed for this study (Supplementary figure S1) and validated outcome instruments as described below.
Sampling strategy
A multi-stage probability sampling design was employed to achieve national representativeness across Thailand’s six geographic regions (North, Northeast, Central, East, West, and South). Two provinces were randomly selected within each region (12 provinces in total). Within each selected province, one district was randomly chosen, followed by random selection of three sub-districts per district (36 sub-districts in total) and one community health center per sub-district (36 study sites in total). At each site, eligible participants were randomly selected from complete community rosters. Sampling was stratified by sex to reflect the local male-to-female distribution, with a target enrollment of 80–85 participants per site. Of 2,972 eligible individuals identified from community rosters across 36 sites, 2,603 (87.6%) completed all study components and were included in the final analysis. The reasons for non-participation were failure to meet the minimum accelerometer wear-day criterion (n = 369, 12.4%). A STROBE-compliant participant flow diagram is provided in Supplementary Figure S1.
Sample size
The multi-stage cluster sampling design targeted approximately 80–85 participants per site across 36 sites, yielding an anticipated sample size of approximately 2,900 individuals. An a priori sample size calculation based on multivariable linear regression with 10 covariates, a minimum detectable effect size of Cohen’s f² = 0.02, a two-sided α of 0.05, and 80% power (G*Power 3.1) estimated a minimum required sample of approximately 995 participants. The target of 2,900 was conservatively set to account for the multi-stage cluster design effect (DEFF ≈ 1.5–2.0) and an anticipated 10–15% exclusion rate. A post hoc analysis confirmed that the final analytic sample of 2,603 participants provided greater than 95% power, retained here as supplementary confirmation only.
Eligibility criteria
Eligible participants were Thai nationals aged ≥ 60 years who had resided in the selected community for at least six months and were able to provide written informed consent and communicate in Thai. Participants were excluded if they had severe cardiopulmonary disease precluding safe participation in physical performance tests, neurological or otological disorders affecting balance, neuromuscular disorders preventing completion of assessments, or bedridden status. Use of personal walking aids was permitted; participants were instructed to use their habitual device during all performance-based assessments. Walking aid type and frequency of use were not systematically recorded.
Exposure: daily step count
Daily ambulatory activity was objectively measured using Fitbit Versa 4 smartwatches (Fitbit International Ltd, San Francisco, CA, USA), which have demonstrated acceptable validity for step counting in community-dwelling older adults when compared with direct visual count (intraclass correlation coefficient [ICC] = 0.88) [10]. Participants were instructed to wear the device on their non-dominant wrist for seven consecutive days, removing it only for bathing or swimming, and to maintain their usual daily activities during the monitoring period. A valid wear day was defined as ≥ 10 h of wear time. Participants with ≥ 4 valid days were included in the analyses, consistent with established conventions [11]. Average daily step count was calculated as total accumulated steps divided by the number of valid wear days. For primary analyses, step counts were categorized into six groups to facilitate clinical interpretability and comparability with prior work: [6] Group 1, ≤ 3,000; Group 2, 3,001–6,000; Group 3, 6,001–9,000; Group 4, 9,001–12,000; Group 5, 12,001–15,000; and Group 6, > 15,000 steps per day.
Outcomes
Quality of life was assessed using the EQ-5D-5 L [12], a standardized instrument evaluating five dimensions (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression). Responses were converted to a utility index score using the validated Thai value set [13], with scores ranging from − 0.421 (worst health state) to 1.0 (perfect health). Participants also rated their overall health using the EQ visual analogue scale (EQ-VAS; range 0–100).
Knee-specific functional outcomes were assessed using the validated Thai versions of the KOOS [14] and OKS [15]. The KOOS comprises 42 items across five subscales (pain, symptoms, activities of daily living, sport/recreation, and knee-related quality of life), each scored from 0 to 100 (higher scores indicate better outcomes). The OKS is a 12-item questionnaire scored from 0 to 48 (higher scores indicate better outcomes).
Performance-based tests were administered by trained research assistants following standardized protocols. Prior to data collection, all assistants completed a two-day training program; inter-rater reliability was confirmed as acceptable for all measures (ICC ≥ 0.85). Questionnaire instruments were administered following standardized scripts. Participants were permitted to wear their customary footwear and to use walking aids as required.
The TUG test was used to assess mobility and fall risk [16]. From a seated position on a standard chair (height approximately 46 cm), the participant rose, walked 3 m at their usual pace, turned, returned, and sat down; the mean time of two trials was recorded. A time of ≥ 13.5 s indicates increased fall risk. The 5×STS was used to measure lower extremity strength [17]. The participant stood up and sat down five times as quickly as possible with arms crossed over the chest; completion time was recorded. Gait speed was measured over a 6-meter walkway at a self-selected pace with a 2-meter acceleration phase; the mean speed of two trials was calculated in meters per second. A gait speed of < 0.8 m/s indicates increased risk of adverse outcomes [18]. Leg strength was assessed using a back and leg dynamometer with the participant standing on the device platform, back straight against a wall, knees flexed approximately 110–120°, exerting maximum isometric leg extension force; the highest value from two trials was recorded in kilograms.
Covariates
Covariates were collected through a structured interview and data collection form developed specifically for this study (Supplementary File 1), including age, sex, BMI, education level, marital status, occupation, physician-diagnosed comorbidities (hypertension, diabetes mellitus, dyslipidemia, cardiovascular disease, chronic respiratory disease, and musculoskeletal conditions), CCI, current medication use, smoking status, alcohol consumption, presence of knee osteoarthritis, and knee pain severity assessed using an 11-point numeric rating scale (0–10).
Statistical analysis
Continuous variables were reported as mean ± standard deviation (SD) or median and interquartile range (IQR), as appropriate. Categorical variables were reported as frequency and percentage. Differences in daily step count across demographic and clinical subgroups were assessed using the Mann–Whitney U test (for two groups) or Kruskal–Wallis test (for three or more groups). Differences in health-related outcomes across step count categories were evaluated using the Kruskal–Wallis test, with ordered trends assessed using the Jonckheere–Terpstra test.
Multivariable linear regression models were fitted to examine the independent associations between step count category (categorical, Group 1 as reference) and each of the eight health-related outcomes. Models were adjusted for age, sex, BMI, CCI, knee osteoarthritis status, and geographic region. Regression coefficients (β) with 95% confidence intervals (CI), adjusted R², and partial eta-squared (η²p) were reported to reflect both the direction and magnitude of associations.
Multicollinearity among covariates was assessed using the Variance Inflation Factor (VIF); all VIF values were below 2.4, indicating no meaningful collinearity. Residual assumptions were evaluated for each model: Breusch–Pagan tests indicated heteroscedasticity in seven of eight models (p < 0.05), consistent with the distributional properties of health-outcome data in large cross-sectional studies; this affects standard error precision but does not bias regression coefficients. Residual non-normality on formal testing was expected given the large sample size (n > 2,500) and does not invalidate inference under the Central Limit Theorem.
All covariates were complete for all 2,603 participants (0% missing). Missing data were limited to outcome variables only (range: 0.1%–9.6%) and were handled using per-outcome complete-case analysis, yielding analytic samples of 2,566–2,601 for seven outcomes and 2,354 for leg strength. For leg strength (9.6% missing), a sensitivity analysis using multiple imputation by chained equations (MICE; 20 imputed datasets) confirmed that complete-case estimates were not meaningfully biased (Supplementary Table S1). All analyses were performed using Python (SciPy 1.17.0; Statsmodels 0.14.6). Statistical significance was defined as p < 0.05 (two-tailed).
Results
Participant characteristics and daily step count
A total of 2,972 individuals were identified from community rosters and invited to participate. Of these, 2,603 (87.6%) completed all components and were included in the final analysis (Supplementary Figure S1). The mean age was 70.2 ± 6.7 years (median 69.0, IQR 66.0–74.0 years), and 64.2% were female. By BMI category, 10.0% were underweight, 34.2% were of normal weight, 37.9% were overweight, and 17.9% were obese. By comorbidity burden, 36.0% had low (CCI 1–2), 56.4% had moderate (CCI 3–4), and 7.6% had high (CCI ≥ 5) burden. The prevalence of knee osteoarthritis was 47.4%. Occupationally, 40.1% were not working, 39.0% were farmers, and 20.9% reported other occupations (Table 1).
Table 1.
Daily step count by participant characteristics (N = 2,603)
| Characteristic | n | % | Mean ± SD (steps/day) |
Median (IQR) (steps/day) |
P-value |
|---|---|---|---|---|---|
| Overall | 2,603 | (100%) | 9,488 ± 5,639 | 8,423 (5,472–12,293) | |
| Sex | |||||
| Male | 933 | (35.8%) | 10,522 ± 6,063 | 9,509 (6,076–14,226) | |
| Female | 1,670 | (64.2%) | 8,910 ± 5,302 | 7,818 (5,281–11,457) | < 0.001 |
| Age group (mean 70.2 ± 6.7 years; median 69.0, IQR 66.0–74.0) | |||||
| 60–65 years | 609 | (23.4%) | 11,372 ± 5,950 | 10,387 (6,970–14,536) | |
| 66–70 years | 896 | (34.4%) | 10,243 ± 5,527 | 9,112 (6,344–12,751) | |
| 71–75 years | 615 | (23.6%) | 8,752 ± 5,217 | 7,700 (5,263–10,938) | |
| >75 years | 483 | (18.6%) | 6,648 ± 4,611 | 5,638 (3,352–8,972) | < 0.001 |
| BMI category (mean 23.8 ± 4.2 kg/m²; median 23.6, IQR 20.8–26.3) | |||||
| Underweight (< 18.5 kg/m²) | 261 | (10.0%) | 9,737 ± 6,685 | 8,004 (4,690–13,427) | |
| Normal weight (18.5–22.9 kg/m²) | 890 | (34.2%) | 10,010 ± 6,031 | 8,853 (5,576–13,258) | |
| Overweight (23.0–27.4 kg/m²) | 986 | (37.9%) | 9,455 ± 5,133 | 8,644 (5,846–12,036) | |
| Obese (≥ 27.5 kg/m²) | 466 | (17.9%) | 8,420 ± 5,086 | 7,301 (4,970–10,652) | < 0.001 |
| Occupation | |||||
| Not working | 1,044 | (40.1%) | 7,695 ± 4,826 | 6,880 (4,211–9,971) | |
| Farmer | 1,015 | (39.0%) | 11,069 ± 5,993 | 10,178 (6,645–14,500) | |
| Other | 544 | (20.9%) | 9,980 ± 5,428 | 8,904 (6,315–12,474) | < 0.001 |
| Geographic region | |||||
| Northern | 442 | (17.0%) | 10,038 ± 5,873 | 8,836 (6,107–12,782) | |
| Northeastern | 451 | (17.3%) | 10,019 ± 5,786 | 9,012 (5,768–13,728) | |
| Central | 376 | (14.4%) | 8,794 ± 5,519 | 7,924 (4,952–11,597) | |
| Eastern | 451 | (17.3%) | 8,596 ± 5,026 | 7,591 (5,072–11,236) | |
| Western | 459 | (17.6%) | 9,604 ± 5,786 | 8,313 (5,760–11,919) | |
| Southern | 424 | (16.3%) | 9,788 ± 5,646 | 8,822 (5,610–12,752) | < 0.001 |
| Charlson Comorbidity Index (mean 3.0 ± 1.1; median 3.0, IQR 2.0–4.0) | |||||
| Low burden (CCI 1–2) | 937 | (36.0%) | 11,142 ± 5,742 | 9,955 (6,954–14,213) | |
| Moderate burden (CCI 3–4) | 1,468 | (56.4%) | 8,785 ± 5,461 | 7,600 (4,972–11,364) | |
| High burden (CCI ≥ 5) | 198 | (7.6%) | 6,869 ± 4,198 | 6,109 (3,544–9,881) | < 0.001 |
| Knee osteoarthritis | |||||
| No knee OA | 1,370 | (52.6%) | 9,783 ± 5,707 | 8,773 (5,774–12,621) | |
| Knee OA present | 1,233 | (47.4%) | 9,160 ± 5,546 | 7,972 (5,263–11,996) | 0.003 |
Data are presented as mean ± standard deviation (SD) and median (interquartile range [IQR]) for daily step count within each subgroup
BMI categories follow Asian WHO criteria: underweight < 18.5; normal weight 18.5–22.9; overweight 23.0–27.4; obese ≥ 27.5 kg/m²
P-values based on Mann–Whitney U test (two groups) or Kruskal–Wallis test (three or more groups)
Step count based on ≥ 4 valid Fitbit wear days (≥ 10 h wear time per day)
Abbreviations: BMI body mass index, CCI Charlson Comorbidity Index, IQR interquartile range, OA osteoarthritis, SD standard deviation
The mean daily step count was 9,488 ± 5,639 steps per day (median 8,423, IQR 5,472–12,293). The distribution across step count groups was: Group 1, n = 222 (8.5%); Group 2, n = 549 (21.1%); Group 3, n = 646 (24.8%); Group 4, n = 500 (19.2%); Group 5, n = 285 (10.9%); and Group 6, n = 401 (15.4%). Male participants accumulated more steps per day than female participants (median 9,509, IQR 6,076–14,226 vs. 7,818, IQR 5,281–11,457 steps per day; p < 0.001). Step count declined progressively with age, from 10,387 (IQR 6,970–14,536) steps per day among those aged 60–65 years to 5,638 (IQR 3,352–8,972) steps per day among those aged > 75 years (p < 0.001). Farmers accumulated the highest median step count (10,178, IQR 6,645–14,500 steps per day), approximately 1.5-fold higher than participants who were not working (6,880, IQR 4,211–9,971 steps per day; p < 0.001). Participants with high comorbidity burden (CCI ≥ 5) accumulated significantly fewer steps than those with low burden (CCI 1–2) (median 6,109, IQR 3,544–9,881 vs. 9,955, IQR 6,954–14,213 steps per day; p < 0.001) (Table 1).
Health-related outcomes by step count group
All eight health-related outcomes differed significantly across step count groups (Kruskal–Wallis and Jonckheere–Terpstra p < 0.001 for all; Table 2). Among the eight outcomes assessed, physical performance measures showed the largest absolute differences across step count categories. Specifically, mean TUG time improved from 15.6 ± 5.5 s (Group 1, ≤ 3,000 steps/day) to 10.4 ± 2.3 s (Group 6, > 15,000 steps/day), gait speed improved from 0.62 ± 0.22 to 0.89 ± 0.25 m/s, and leg strength improved from 41.0 ± 19.6 to 58.5 ± 27.8 kg across the same step count range. Knee-specific function scores also improved progressively, with KOOS total score increasing from 71.7 ± 19.9 in Group 1 to 86.0 ± 16.8 in Group 6, and OKS from 35.4 ± 11.6 to 42.5 ± 9.4. Quality of life improved in parallel, with EQ-5D-5 L utility scores increasing from 0.825 ± 0.188 to 0.932 ± 0.113, and EQ-VAS from 73.4 ± 16.4 to 79.8 ± 15.2.
Table 2.
Health-related outcomes by daily step count group (N = 2,603)
| Daily Step Count Group (steps/day) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Domain | Outcome (direction) | n | Overall (N = 2,603) |
Group 1 ≤ 3,000 (n = 222) |
Group 2 3,001–6,000 (n = 549) |
Group 3 6,001–9,000 (n = 646) |
Group 4 9,001–12,000 (n = 500) |
Group 5 12,001–15,000 (n = 285) |
Group 6 > 15,000 (n = 401) |
P
(KW) |
P
(trend) |
|
| Quality of life | ||||||||||||
| ↑ EQ-5D-5 L utility index † | 2,595 | 0.899 ± 0.145 | 0.825 ± 0.188 | 0.874 ± 0.171 | 0.901 ± 0.136 | 0.919 ± 0.109 | 0.919 ± 0.136 | 0.932 ± 0.113 | < 0.001 | < 0.001 | ||
| ↑ EQ-VAS (0–100) † | 2,581 | 77.8 ± 15.4 | 73.4 ± 16.4 | 75.9 ± 16.1 | 78.1 ± 15.0 | 78.8 ± 14.5 | 79.5 ± 15.2 | 79.8 ± 15.2 | < 0.001 | < 0.001 | ||
| Knee-specific function | ||||||||||||
| ↑ KOOS total score (0–100) † | 2,601 | 81.0 ± 18.3 | 71.7 ± 19.9 | 77.5 ± 18.8 | 80.2 ± 18.1 | 83.8 ± 16.5 | 84.3 ± 17.2 | 86.0 ± 16.8 | < 0.001 | < 0.001 | ||
| ↑ Oxford Knee Score (0–48) † | 2,601 | 39.9 ± 10.4 | 35.4 ± 11.6 | 38.5 ± 10.5 | 39.5 ± 10.7 | 41.2 ± 9.9 | 41.2 ± 9.7 | 42.5 ± 9.4 | < 0.001 | < 0.001 | ||
| Physical performance | ||||||||||||
| ↓ TUG test (seconds) † | 2,586 | 11.72 ± 3.42 | 15.64 ± 5.52 | 12.38 ± 3.28 | 11.75 ± 3.15 | 10.85 ± 2.41 | 10.75 ± 2.36 | 10.41 ± 2.29 | < 0.001 | < 0.001 | ||
| ↓ 5×STS (seconds) † | 2,569 | 15.92 ± 5.05 | 19.44 ± 7.36 | 16.82 ± 4.87 | 16.11 ± 5.00 | 14.97 ± 4.39 | 14.85 ± 4.38 | 14.48 ± 3.86 | < 0.001 | < 0.001 | ||
| ↑ Gait speed (m/s) † | 2,563 | 0.809 ± 0.249 | 0.622 ± 0.215 | 0.761 ± 0.235 | 0.810 ± 0.246 | 0.852 ± 0.240 | 0.844 ± 0.238 | 0.892 ± 0.247 | < 0.001 | < 0.001 | ||
| ↑ Leg strength (kg) † | 2,354 | 51.7 ± 25.0 | 41.0 ± 19.6 | 45.9 ± 21.7 | 50.1 ± 24.2 | 54.7 ± 24.2 | 57.6 ± 27.1 | 58.5 ± 27.8 | < 0.001 | < 0.001 | ||
Data are presented as mean ± SD. All p-values are two-sided
P (KW) = overall difference across step count groups assessed by Kruskal–Wallis test
P (trend) = monotonic ordered trend across step count groups assessed by Jonckheere–Terpstra test
Non-parametric tests were used throughout, given non-normal distributions confirmed by Shapiro–Wilk test
Step count groups based on Paluch et al. (2022): Group 1, ≤ 3,000; Group 2, 3,001–6,000; Group 3, 6,001–9,000; Group 4, 9,001–12,000; Group 5, 12,001–15,000; Group 6, > 15,000 steps/day
† Missing data: EQ-5D-5 L utility, n = 8; EQ-VAS, n = 22; TUG, n = 17; 5×STS, n = 34; gait speed, n = 40; leg strength, n = 249
↑ Higher score = better outcome. ↓ Lower score = better outcome
Abbreviations: EQ-VAS EuroQol visual analogue scale, KW Kruskal–Wallis, KOOS Knee Injury and Osteoarthritis Outcome Score, OKS Oxford Knee Score, TUG Timed Up and Go test, 5×STS Five Times Sit-to-Stand test
Multivariable linear regression
After adjustment for covariates, step count group remained independently associated with all eight health-related outcomes (Table 3). The largest and most clinically meaningful differences were observed between Group 6 (> 15,000 steps/day) and Group 1 (≤ 3,000 steps/day). For physical performance, the adjusted TUG time difference was − 3.7 s (95% CI − 4.2 to − 3.2; p < 0.001), exceeding the established MCID of 2.0–3.4 s [19]. Gait speed improved by + 0.20 m/s (95% CI 0.163–0.235; p < 0.001), KOOS total score by + 9.5 points (95% CI 6.7–12.3; p < 0.001), and EQ-5D-5 L utility by + 0.087 units (95% CI 0.063–0.110; p < 0.001), all exceeding established clinically meaningful thresholds. The only non-significant association was EQ-VAS in Group 2 versus Group 1 (adjusted β = +1.9, 95% CI − 0.4 to 4.3; p = 0.109). Adjusted R² values ranged from 0.058 (EQ-VAS) to 0.417 (leg strength), with corresponding η²p values indicating small-to-moderate effect sizes, and greater explanatory power for objective performance outcomes than for subjective measures.
Table 3.
Multivariable linear regression: associations between step count group and health-related outcomes
| Domain | Outcome | n | Step Group (ref = Group 1) | Unadjusted β | Unadjusted 95% CI | Unadjusted P | Adjusted β | Adjusted 95% CI | Adjusted P | η²p | R²adj |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Quality of Life | |||||||||||
| ↑ EQ-5D-5 L utility | 2,595 | Group 2 | +0.049 | (+0.027, +0.071) | <0.001 | +0.039 | (+0.018, +0.060) | <0.001 | 0.135 | 0.135 | |
| Group 3 | +0.076 | (+0.055, +0.098) | <0.001 | +0.066 | (+0.045, +0.087) | <0.001 | |||||
| Group 4 | +0.094 | (+0.071, +0.116) | <0.001 | +0.076 | (+0.054, +0.098) | <0.001 | |||||
| Group 5 | +0.094 | (+0.069, +0.119) | <0.001 | +0.077 | (+0.052, +0.101) | <0.001 | |||||
| Group 6 | +0.107 | (+0.084, +0.130) | <0.001 | +0.087 | (+0.063, +0.110) | <0.001 | |||||
| ↑ EQ-5D VAS | 2,581 | Group 2 | +2.5 | (+0.1, +4.9) | 0.038 | +1.9 | (-0.4, +4.3) | 0.109 | 0.058 | 0.058 | |
| Group 3 | +4.7 | (+2.4, +7.1) | <0.001 | +4.0 | (+1.6, +6.3) | 0.001 | |||||
| Group 4 | +5.4 | (+3.0, +7.8) | <0.001 | +4.2 | (+1.7, +6.7) | <0.001 | |||||
| Group 5 | +6.1 | (+3.4, +8.8) | <0.001 | +4.8 | (+2.0, +7.6) | <0.001 | |||||
| Group 6 | +6.4 | (+3.9, +8.9) | <0.001 | +5.0 | (+2.4, +7.6) | <0.001 | |||||
| Knee Function | |||||||||||
| ↑ KOOS total | 2,601 | Group 2 | +5.9 | (+3.1, +8.6) | <0.001 | +4.8 | (+2.3, +7.3) | <0.001 | 0.252 | 0.252 | |
| Group 3 | +8.5 | (+5.8, +11.3) | <0.001 | +6.6 | (+4.1, +9.1) | <0.001 | |||||
| Group 4 | +12.1 | (+9.3, +14.9) | <0.001 | +8.7 | (+6.1, +11.3) | <0.001 | |||||
| Group 5 | +12.6 | (+9.5, +15.7) | <0.001 | +8.5 | (+5.6, +11.4) | <0.001 | |||||
| Group 6 | +14.3 | (+11.4, +17.3) | <0.001 | +9.5 | (+6.7, +12.3) | <0.001 | |||||
| ↑ Oxford Knee Score | 2,601 | Group 2 | +3.2 | (+1.6, +4.8) | <0.001 | +2.6 | (+1.1, +4.1) | <0.001 | 0.181 | 0.181 | |
| Group 3 | +4.2 | (+2.6, +5.8) | <0.001 | +3.3 | (+1.8, +4.8) | <0.001 | |||||
| Group 4 | +5.8 | (+4.2, +7.5) | <0.001 | +4.3 | (+2.8, +5.9) | <0.001 | |||||
| Group 5 | +5.8 | (+4.0, +7.6) | <0.001 | +4.0 | (+2.2, +5.7) | <0.001 | |||||
| Group 6 | +7.1 | (+5.4, +8.8) | <0.001 | +4.9 | (+3.2, +6.6) | <0.001 | |||||
| Physical Performance | |||||||||||
| ↓ TUG (seconds) | 2,586 | Group 2 | -3.26 | (-3.75, -2.76) | <0.001 | -2.71 | (-3.18, -2.25) | <0.001 | 0.286 | 0.286 | |
| Group 3 | -3.89 | (-4.38, -3.41) | <0.001 | -3.04 | (-3.50, -2.58) | <0.001 | |||||
| Group 4 | -4.80 | (-5.30, -4.29) | <0.001 | -3.65 | (-4.14, -3.17) | <0.001 | |||||
| Group 5 | -4.89 | (-5.44, -4.33) | <0.001 | -3.58 | (-4.12, -3.05) | <0.001 | |||||
| Group 6 | -5.23 | (-5.75, -4.71) | <0.001 | -3.68 | (-4.19, -3.17) | <0.001 | |||||
| ↓ 5×STS (seconds) | 2,569 | Group 2 | -2.62 | (-3.40, -1.84) | <0.001 | -2.09 | (-2.83, -1.35) | <0.001 | 0.195 | 0.195 | |
| Group 3 | -3.33 | (-4.09, -2.57) | <0.001 | -2.56 | (-3.30, -1.83) | <0.001 | |||||
| Group 4 | -4.47 | (-5.26, -3.68) | <0.001 | -3.33 | (-4.09, -2.56) | <0.001 | |||||
| Group 5 | -4.60 | (-5.47, -3.72) | <0.001 | -3.28 | (-4.13, -2.44) | <0.001 | |||||
| Group 6 | -4.96 | (-5.78, -4.14) | <0.001 | -3.29 | (-4.10, -2.47) | <0.001 | |||||
| ↑ Gait speed (m/s) | 2,563 | Group 2 | +0.139 | (+0.101, +0.178) | <0.001 | +0.126 | (+0.092, +0.159) | <0.001 | 0.335 | 0.335 | |
| Group 3 | +0.188 | (+0.150, +0.225) | <0.001 | +0.146 | (+0.113, +0.179) | <0.001 | |||||
| Group 4 | +0.230 | (+0.191, +0.269) | <0.001 | +0.175 | (+0.141, +0.210) | <0.001 | |||||
| Group 5 | +0.222 | (+0.179, +0.265) | <0.001 | +0.157 | (+0.119, +0.195) | <0.001 | |||||
| Group 6 | +0.270 | (+0.230, +0.310) | <0.001 | +0.199 | (+0.163, +0.235) | <0.001 | |||||
| ↑ Leg strength (kg) | 2,354 | Group 2 | +4.8 | (+0.5, +9.2) | 0.030 | +5.2 | (+1.7, +8.6) | 0.003 | 0.417 | 0.417 | |
| Group 3 | +9.1 | (+4.8, +13.3) | <0.001 | +7.5 | (+4.1, +10.9) | <0.001 | |||||
| Group 4 | +13.6 | (+9.3, +18.0) | <0.001 | +9.6 | (+6.1, +13.1) | <0.001 | |||||
| Group 5 | +16.5 | (+11.7, +21.3) | <0.001 | +10.2 | (+6.3, +14.1) | <0.001 | |||||
| Group 6 | +17.4 | (+12.9, +21.9) | <0.001 | +9.8 | (+6.1, +13.5) | <0.001 | |||||
Data are presented as regression coefficient (β) with 95% confidence interval (CI). Group 1 (≤ 3,000 steps/day) is the reference category. All models adjusted for age, sex, body mass index, Charlson Comorbidity Index, knee osteoarthritis status, and geographic region. Adjusted β and 95% CI shown in bold
Step count groups: Group 2, 3,001–6,000; Group 3, 6,001–9,000; Group 4, 9,001–12,000; Group 5, 12,001–15,000; Group 6, > 15,000 steps/day
η²p, partial eta-squared (approximated from R²adj)
R²adj, adjusted R-squared. ↑ Higher = better outcome; ↓ Lower = better outcome
Abbreviations: CI confidence interval, EQ-VAS EuroQol visual analogue scale, KOOS Knee Injury and Osteoarthritis Outcome Score, OKS Oxford Knee Score, TUG Timed Up and Go test, 5×STS Five Times Sit-to-Stand test
Dose–response relationships and subgroup analyses
Across all six step count groups, all eight health-related outcomes demonstrated statistically significant monotonic dose–response relationships (Jonckheere–Terpstra p < 0.001 for all). For objective physical performance measures (TUG, 5×STS, gait speed, and leg strength), the greatest improvements were observed between Groups 1 and 3, after which the magnitude of improvement plateaued or increased more gradually, suggesting a potential threshold effect. In contrast, knee-specific functional outcomes and quality-of-life measures exhibited more gradual, near-linear improvements across the entire step count spectrum without a clear plateau (Fig. 1).
Fig. 1.
Health-Related Outcomes by Daily Step Count Group. Adjusted mean values (95% confidence intervals) for eight health-related outcomes across six daily step count groups in 2,603 community-dwelling Thai older adults. Panels show: a EQ-5D-5 L utility index; b EQ-VAS score (0–100); c KOOS total score (0–100); d Oxford Knee Score (0–48); e Timed Up and Go test (seconds); f Five Times Sit-to-Stand test (seconds); g Gait speed (m/s); h Leg strength (kg). All models were adjusted for age, sex, body mass index, Charlson Comorbidity Index, knee osteoarthritis status, and geographic region. Error bars represent 95% confidence intervals. P for trend assessed by Jonckheere–Terpstra test (all p < 0.001). ↑ Higher score = better outcome; ↓ Lower score = better outcome
When stratified by sex, dose–response patterns were largely parallel between males and females, indicating no significant interaction for most outcomes. However, statistically significant sex-by-step count interactions were observed for leg strength (p < 0.001), with a steeper dose–response gradient among males, and for TUG performance (p = 0.007), with relatively greater benefit observed in females (Fig. 2).
Fig. 2.
Health-Related Outcomes by Daily Step Count Group, Stratified by Sex. Adjusted mean values (95% confidence intervals) stratified by sex. Blue line: male participants (n = 933); red line: female participants (n = 1,670). Panels show: a EQ-5D-5 L utility index; b EQ-VAS score (0–100); c KOOS total score (0–100); d Oxford Knee Score (0–48); e Timed Up and Go test (seconds); f Five Times Sit-to-Stand test (seconds); g Gait speed (m/s); h Leg strength (kg). Outcomes and covariate adjustments are as described in Fig. 1. P for interaction was assessed by adding a sex × step count group interaction term to each regression model. Statistically significant sex-by-step count interactions were observed for leg strength (p < 0.001) and TUG performance (p = 0.007); no significant interaction was observed for the remaining six outcomes (all p > 0.05). ↑ Higher score = better outcome; ↓ Lower score = better outcome
In age-stratified analyses, significant age-by-step count interactions were observed for several physical performance outcomes. Gait speed demonstrated the strongest interaction (p < 0.001), with a pronounced dose–response relationship among participants aged > 75 years, whereas minimal change was observed among those aged 60–65 years. A similar age-dependent pattern was evident for leg strength (p = 0.001). In contrast, TUG and 5×STS improved consistently with higher step counts across all age groups, with no significant age interaction detected (p > 0.05 for both) (Fig. 3).
Fig. 3.
Health-Related Outcomes by Daily Step Count Group, Stratified by Age Group. Adjusted mean values (95% confidence intervals) stratified by four age groups: 60–65 years (n = 609), 66–70 years (n = 896), 71–75 years (n = 615), and > 75 years (n = 483). Outcomes and covariate adjustments are as described in Fig. 1. P for interaction was assessed by adding an age group × step count group interaction term to each regression model. Significant age-by-step count interactions were observed for gait speed (p < 0.001) and leg strength (p = 0.001), with steeper dose–response gradients in older participants. No significant age interactions were detected for TUG or 5×STS (both p > 0.05). ↑ Higher score = better outcome; ↓ Lower score = better outcome
Discussion
This large-scale cross-sectional study of 2,603 community-dwelling Thai older adults demonstrates robust, dose-dependent associations between daily step count and multiple health-related outcomes spanning quality of life, knee-specific function, and physical performance. All eight outcomes improved progressively across step count groups, and these associations remained statistically significant after adjustment for demographic and clinical confounders. Collectively, our findings support a daily step target of 6,000–9,000 steps as an evidence-informed and contextually appropriate recommendation for older Thai adults, consistent with recent international consensus [6].
Age-stratified analyses revealed significant associations between step count and health-related outcomes across all age groups, with notably steeper gradients among older participants. This pattern suggests that daily step count serves not only as a behavioral metric of physical activity, but also as a sensitive marker of functional reserve in advanced age. The stronger associations observed among the oldest participants highlight the potential utility of step count monitoring for identifying vulnerability and tracking functional decline. Longitudinal studies are warranted to determine whether these associations reflect causal benefits of increased ambulatory activity.
The shape of the dose–response relationship varied by outcome type. Objective physical performance measures (TUG, gait speed, and leg strength) demonstrated steep improvements between ≤ 3,000 and 6,001–9,000 steps per day, followed by attenuated gains at higher step volumes, suggesting a potential threshold effect. In contrast, quality-of-life and knee-specific outcomes improved more gradually across the full step count spectrum. Importantly, differences between the lowest and highest step count groups exceeded established clinically meaningful thresholds: the adjusted TUG difference of − 3.7 s surpassed the MCID of 2.0–3.4 s [19], the gait speed improvement of + 0.27 m/s exceeded the commonly cited MCID of approximately 0.10 m/s [20] by 2.5-fold, and the EQ-5D-5 L utility gain of 0.087 was nearly twice the established minimal important difference [21]. These clinically meaningful thresholds were approached or achieved by participants accumulating 6,001–9,000 steps per day compared with those accumulating ≤ 3,000 steps per day, reinforcing this range as a meaningful and attainable target. These findings are consistent with recent large-scale evidence: a UK Biobank study (n = 72,174) reported optimal mortality reduction at 9,000–10,500 steps per day with a minimal benefit threshold of 4,000–4,500 steps per day [22], while a meta-analysis of 35 cohorts identified 7,000 steps per day as associated with clinically meaningful reductions in all-cause mortality (47%), cardiovascular disease (25%), and dementia (38%) [23]. The convergence of these findings across Western and Asian populations strongly supports 6,000–9,000 steps per day as an evidence-informed and achievable target for older adults.
Male participants accumulated more daily steps and demonstrated superior physical performance overall, consistent with established normative data for Thai older adults [24]. However, the health benefits of increased step count were largely consistent across sexes, with significant sex-by-step count interactions limited to leg strength and TUG performance. Although male participants typically achieve higher step volumes, our data suggest that a target of 6,000–9,000 steps per day is appropriate for both sexes, with individuals encouraged to exceed this range when feasible.
The mean daily step count in our cohort (9,488 ± 5,639 steps per day) was substantially higher than values reported in Western populations of comparable age, including U.S. adults (approximately 5,100 steps per day) [25], likely reflecting the predominantly rural composition of our sample, where agricultural occupations necessitate higher habitual walking. Farmers accumulated approximately 50% more steps per day than non-working participants, consistent with previous normative data from our group [26]. Importantly, most existing step-based evidence derives from Western cohorts, limiting generalizability to Asian populations that differ in body composition, cultural activity patterns, and disease epidemiology. Our study addresses this gap by providing the first comprehensive dose–response data from a large Asian sample, supporting the development of regionally relevant physical activity guidance. Urban Thai populations, exposed to more sedentary environments, may exhibit substantially lower step counts and therefore warrant targeted public health interventions.
The observed associations between higher daily step counts and better health-related outcomes are consistent with several plausible biological mechanisms. Regular ambulatory activity is hypothesized to confer benefits through musculoskeletal pathways — enhancing lower limb muscle strength and joint lubrication [27] — as well as cardiovascular and metabolic pathways, including improved cardiorespiratory fitness and insulin sensitivity [28]. Neurological mechanisms, notably upregulation of brain-derived neurotrophic factor (BDNF), may further support balance control and gait automaticity [29], while anti-inflammatory effects of habitual walking may attenuate the low-grade systemic inflammation that drives sarcopenia and functional decline in aging populations [30]. Together, these pathways provide a plausible mechanistic basis for the dose–response associations observed across all eight outcomes.
Daily step count represents a simple, objective, and scalable metric for risk stratification in older adults. In primary care settings and community health programs — including Thailand’s Community Health Volunteer (CHV) network — routine step count monitoring may facilitate early identification of at-risk individuals and support targeted behavioral interventions, aligning with Thailand’s national strategy for active aging.
Several limitations should be considered. First, the cross-sectional design precludes causal inference; reverse causation is particularly plausible, as individuals with better physical function may be more capable of accumulating higher step counts. Second, step count was assessed over a single 7-day monitoring period, which may not fully capture long-term habitual activity patterns, and wrist-worn accelerometers may overestimate step counts compared with hip-worn devices, particularly during slow walking in older adults. Both issues represent potential sources of non-differential misclassification that would most likely attenuate rather than inflate the observed associations. Third, step volume was examined without consideration of intensity (e.g., cadence or moderate-to-vigorous activity), which recent evidence suggests may independently contribute to health benefits [31]. Fourth, residual confounding from unmeasured factors — including diet, psychological status, household income, neighbourhood walkability, and access to green space — cannot be excluded despite adjustment for multiple covariates. Fifth, the predominantly rural sample may limit generalizability to urban populations. Sixth, enrollment required travel to community health centers and completion of physical performance tests, which may have deterred less mobile individuals from participating, potentially overrepresenting healthier and more active older adults (healthy volunteer bias); however, the response rate of 87.6% reduces this concern. Seventh, missing outcome data were limited to 0.1%–1.3% for seven outcomes; leg strength had 9.6% missing data attributable to participants unable to complete dynamometry testing, and a sensitivity analysis using multiple imputation confirmed the robustness of complete-case findings (Supplementary Table S1). Finally, walking aid type and frequency of use were not recorded, precluding formal assessment of its potential influence on TUG and gait speed measurements.
Conclusions
This nationwide cross-sectional study demonstrates that daily step count is independently and robustly associated with better quality of life, knee-specific function, and objective physical performance among community-dwelling Thai older adults. A daily target of 6,000–9,000 steps appears clinically meaningful and pragmatically attainable for the majority of older adults in this population, with the strongest associations evident among those of more advanced age. To our knowledge, these findings provide the first comprehensive dose–response data from an Asian population to inform regionally appropriate step count recommendations.
Daily step count represents a simple, objective, and scalable metric for risk stratification and health monitoring in older adults, with potential for integration into primary care and community health settings. These associations, while robust after adjustment for multiple confounders, are observational in nature and do not establish causality. Longitudinal studies and randomized intervention trials are warranted to determine whether increasing step counts leads to sustained improvements in health-related quality of life and physical function in older Asian populations.
Supplementary Information
Acknowledgements
The authors would like to thank Ms. Nichakorn Khomawut for her contribution to this study, particularly for her assistance in data collection and statistical analysis.
Authors’ contributions
R.N., K.C., and C.P. conceived and designed the study. R.N., C.A., and K.A. were involved in data collection and recruitment of participants. R.N. and C.P. performed the statistical analysis. R.N. and K.C. wrote the main manuscript text. C.A., K.A., and C.P. prepared the figures and tables. All authors reviewed and approved the final manuscript.
Funding
A grant from the National Research Council of Thailand (NRCT), Bangkok, Thailand. The funder had no influence on the study design or protocol, the interpretation of data, the final conclusions drawn, or the decision to publish., R016641051.
Data availability
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
The study protocol was approved by the Institutional Review Board of Siriraj Hospital (Protocol: CREC054/66SR-SBR3). Written informed consent was obtained from all participants.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.World Health Organization. Ageing and health. Geneva: WHO; 2022. [Google Scholar]
- 2.National Statistical Office of Thailand. The 2021 Survey of the Older Persons in Thailand. Bangkok: National Statistical Office; 2022. [Google Scholar]
- 3.World Health Organization. World report on ageing and health. Geneva: WHO. 2015. ISBN: 9789241565042.
- 4.World Health Organization. WHO Guidelines on Physical Activity and Sedentary Behaviour. Geneva: WHO; 2020. [PubMed] [Google Scholar]
- 5.Lee IM, Shiroma EJ, Kamada M, Bassett DR, Matthews CE, Buring JE. Association of step volume and intensity with all-cause mortality in older women. JAMA Intern Med. 2019;179(8):1105–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Saint-Maurice PF, Troiano RP, Bassett DR Jr, et al. Association of daily step count and step intensity with mortality among US adults. JAMA. 2020;323(12):1151–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Paluch AE, Bajpai S, Bassett DR, et al. Daily steps and all-cause mortality: a meta-analysis of 15 international cohorts. Lancet Public Health. 2022;7(3):e219–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.White DK, Tudor-Locke C, Zhang Y, et al. Daily walking and the risk of incident functional limitation in knee OA: an observational study. Arthritis Care Res (Hoboken). 2014;66(9):1328–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.World Health Organization. International Classification of Functioning, Disability and Health: ICF. Geneva: WHO. 2001. ISBN: 9241545429.
- 10.Fokkema T, Kooiman TJM, van der Krijnen WP, de Groot M. Reliability and validity of ten consumer activity trackers depend on walking speed. Med Sci Sports Exerc. 2017;49(4):793–800. [DOI] [PubMed] [Google Scholar]
- 11.Tudor-Locke C, Burkett L, Reis JP, et al. How many days of pedometer monitoring predict weekly physical activity in adults? Prev Med. 2005;40(3):293–8. [DOI] [PubMed] [Google Scholar]
- 12.EuroQol Group. EuroQol—a new facility for the measurement of health-related quality of life. Health Policy. 1990;16(3):199–208. [DOI] [PubMed] [Google Scholar]
- 13.Pattanaphesaj J, Thavorncharoensap M, Ramos-Goñi JM, et al. The EQ-5D-5L valuation study in Thailand. Expert Rev Pharmacoecon Outcomes Res. 2018;18(5):551–8. [DOI] [PubMed] [Google Scholar]
- 14.Reosanguanwong K, Suphakitchanusan W, Lertwanich P. Cross-cultural adaptation and validation of the Thai version of the Knee Injury and Osteoarthritis Outcome Score for Patellofemoral Pain and Osteoarthritis (KOOS-PF). Orthop J Sports Med. 2024;12(10):23259671241279465. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Charoencholvanich K, Pongcharoen B. Oxford knee score and SF-36: translation and reliability for use with total knee arthroplasty patients in Thailand. J Med Assoc Thai. 2005;88(9):1194–202. [PubMed] [Google Scholar]
- 16.Podsiadlo D, Richardson S. The timed Up & Go: a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc. 1991;39(2):142–8. [DOI] [PubMed] [Google Scholar]
- 17.Bohannon RW. Reference values for the five-repetition sit-to-stand test: a descriptive meta-analysis of data from elders. Percept Mot Skills. 2006;103(1):215–22. [DOI] [PubMed] [Google Scholar]
- 18.Studenski S, Perera S, Patel K, et al. Gait speed and survival in older adults. JAMA. 2011;305(1):50–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Aghayev E, Elfering A, Schwarzenbach O, et al. Assessment of the minimum clinically important difference in the Timed Up and Go Test after surgery for lumbar degenerative disc disease. Spine. 2017;42(13):1001–6. [Google Scholar]
- 20.Bohannon RW, Glenney SS. Minimal clinically important difference for change in comfortable gait speed of adults with pathology: a systematic review. J Eval Clin Pract. 2014;20(4):295–300. [DOI] [PubMed] [Google Scholar]
- 21.Walters SJ, Brazier JE. Comparison of the minimally important difference for two health state utility measures: EQ-5D and SF-6D. Qual Life Res. 2005;14(6):1523–32. [DOI] [PubMed] [Google Scholar]
- 22.Ahmadi MN, Rezende LFM, Ferrari G, Del Pozo Cruz B, Lee IM, Stamatakis E. Do the associations of daily steps with mortality and incident cardiovascular disease differ by sedentary time levels? A device-based cohort study. Br J Sports Med. 2024;58(5):261–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Ding D, Nguyen B, Nau T, et al. Daily steps and health outcomes in adults: a systematic review and dose-response meta-analysis. Lancet Public Health. 2025;10(8):e668–81. [DOI] [PubMed] [Google Scholar]
- 24.Chandhanayingyong C, Adulkasem N, Asavamongkolkul A, et al. Establishing normative values for performance-based tests in older Thai adults: a nationwide cross-sectional study. Arch Phys Med Rehabil. 2024;105(6):1133–41. [DOI] [PubMed] [Google Scholar]
- 25.Tudor-Locke C, Johnson WD, Katzmarzyk PT. Accelerometer-determined steps per day in US adults. Med Sci Sports Exerc. 2009;41(7):1384–91. [DOI] [PubMed] [Google Scholar]
- 26.Howard RA, Freedson PS. Potential issues with measurement of physical activity using accelerometers. Med Sci Sports Exerc. 2009;41(2):272–4. [Google Scholar]
- 27.Vina J, Sanchis-Gomar F, Martinez-Bello V, Gomez-Cabrera MC. Exercise acts as a drug; the pharmacological benefits of exercise. Br J Pharmacol. 2012;167(1):1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Gill JM, Malkova D. Physical activity, fitness and cardiovascular disease risk in adults: interactions with insulin resistance and obesity. Clin Sci (Lond). 2006;110(4):409–25. PMID: 16526946. [DOI] [PubMed] [Google Scholar]
- 29.Gholami F, Mesrabadi J, Iranpour M, Donyaei A. Exercise training alters resting brain-derived neurotrophic factor concentration in older adults: A systematic review with meta-analysis of randomized-controlled trials. Exp Gerontol. 2025;199:112658. PMID: 39674562. [DOI] [PubMed]
- 30.Beckwée D, Delaere A, Aelbrecht S, et al. Exercise Interventions for the Prevention and Treatment of Sarcopenia. A Systematic Umbrella Review. J Nutr Health Aging. 2019;23(6):494–502. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Mandini S, Conconi F, Mori E, et al. Walking and hypertension: greater reductions in subjects with higher baseline systolic blood pressure following six months of guided walking. PeerJ. 2018;6:e5471. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.