Counts describe this library, not the strength of evidence.
How to read this evidence
The step-count literature moves from specific older populations to broader international cohorts and multiple outcomes. Associations support attention to sustained activity; they do not establish a universal daily threshold or an individual survival forecast.
A meta-analysis of 25 randomised trials (7,861 older adults) found physical activity improved global Active Ageing (SMD 1.38, 95% CI 0.54-2.22), social participation, mental health and some physical tests, but not gait speed, handgrip strength or balance. Results are domain-specific and the best type, frequency and dose remain unknown.
Among 15,197 people with knee or hip osteoarthritis in a digital education and exercise program, the share meeting recommended activity levels rose from 51.4% at enrolment to 58.5% at one year, and all four activity trajectory groups reported comparable improvements in patient-reported outcomes. This registry had no comparison group, so the improvements cannot be credited to the program.
Among 1118 community-dwelling adults aged 50 and older in Singapore, life space (how far people move in daily life) was a stronger predictor of self-rated health and physical performance than questionnaire-reported physical activity. Both exposures and outcomes were measured at one time, so the study shows association only, and the abstract gives no effect sizes.
In a two-wave survey of 597 Chinese secondary school students, feeling supported in physical education class was linked to more sustained extracurricular activity through motivation, perceived benefits of sport, and loss aversion (both mediating effects p < 0.001). The abstract gives no effect sizes, and survey data cannot prove that class support causes more activity.
Across 22 reports (7 randomized), acute exercise showed no clear effect on seroprotection after influenza vaccination, for example H1N1 RR 0.88 (95% CI 0.65-1.21; k = 3). Evidence certainty was low or very low and trials were few, so the data neither prove equivalence nor rule out a benefit.
Among 408 healthcare students at four West Bank universities, 68.1% reached a high physical activity level on the WHO GPAQ, with medical students less active than nursing and physical therapy students (p < 0.001). The data are self-reported from one cross-sectional survey in 2025, so they show patterns in this group, not what drives activity.
Among 6,530 university students given the same target of 45 extracurricular sessions, those using intelligent-assistant support did 0.58 fewer total sessions than those self-scheduling, with a slightly higher share of running (+2.13 percentage points). The groups were not randomised, so differences may reflect who chose or was assigned each approach.
In 58 women with rheumatoid arthritis (RA) from western Mexico, lower physical activity was associated with higher CRP (beta = -0.312, p = 0.023), and more sedentary time and weaker handgrip were associated with higher disease activity and disability. This is a small cross-sectional study, so it cannot show whether activity changes the disease or the disease limits activity.
In 45 children and adolescents with autism, those who had done Halliwick-based aquatic activity for at least 3 months scored higher on every KIDSCREEN-52 domain, but only six domains reached statistical significance. This is a cross-sectional comparison of two small groups, so it cannot show that swimming improved quality of life.
From interviews with 14 Saudi adults with type 2 diabetes, hypertension or cardiovascular disease who kept using the Sehhaty app's activity function, four themes emerged, including an expectation that programs fit prayer times, family routines and summer heat. All 14 were sustained users, so the findings say nothing about how acceptable such programs are to the wider population.
Among 677 Chinese college students, higher physical activity was linked to lower internet addiction scores (continuous B = -2.031 per 1,000 MET-min/week). The proposed route through body awareness and emotion regulation depended on how the analysis was specified, and a single survey cannot show cause.
In a survey of 3,009 middle-school students in 6 Chinese provinces, higher health literacy was associated with healthier eating, but this link weakened as physical activity participation rose, and school connectedness showed an opposing indirect pathway. This is a cross-sectional questionnaire study, so it cannot show cause, and the abstract gives no effect sizes.
Among 1,390 undergraduates in Jiangsu Province, perceived academic demands were positively associated with psychological distress (β = 0.382), mostly through chronic time pressure (indirect B = 0.182), while the added path through leisure-time physical activity was tiny (B = 0.010). This cross-sectional survey cannot show cause or order of events.
Of 150 patients surveyed a mean 6.8 years after partial (unicompartmental) knee replacement, 33.3% reported current sport or physical activity, and 74.0% of those did more than 3 h/week. Earlier activity levels were recalled from memory, so the before-and-after comparisons are weak.
In 45 sedentary college students who each did four endurance protocols, threshold training gave the highest mean heart rate, interval training (HIIT) the highest peak heart rate (THR vs HIIT p = 0.032 and p = 0.003) and the highest perceived exertion, while low-intensity high-volume work felt most pleasant. These are acute responses within single sessions, not training results.
Among 2,137 female college students, more physical activity was associated with fewer autistic traits (β = -0.197), and social anxiety accounted for 31% of the total association. This was a one-time self-report survey, so it cannot show that exercise lowers autistic traits, and the traits were measured by a brief screening questionnaire, not a diagnosis.
In a self-report survey of adolescents, higher physical activity was linked to lower suicidal ideation and higher regulatory emotional self-efficacy, which accounted for part of the link. Because the abstract gives no sample size or effect sizes and the design is cross-sectional, it cannot show that exercise reduces suicidal thoughts.
In 813 female university students followed across three waves over 6 months, higher restrained eating predicted more later physical activity (β = 0.229 and 0.164, both p < 0.001) and higher activity predicted more later restrained eating (β = 0.098, p < 0.01, and 0.431, p < 0.001). Both were self-reported, and the pattern fits weight and shape control, not necessarily health.