Table 2 Study characteristics
From: The value of walking: a systematic review on mobility and healthcare costs
Publication | Title | Country | Database and sample size | Study design | Study population and setting | Method & Comparison | Outcome measures | Cost data |
|---|---|---|---|---|---|---|---|---|
Walking Time | ||||||||
Perkins et al. 2001 [21] | Assessing the Association of Walking with Health Services Use and Costs among Socioeconomically Disadvantaged Older Adults | USA | n = 1,088 Regenstrief Physical activity and Health Survey (RPAHS) | Cohort study | Community dwelling patients > 55 years | Multivariate models assessing the association between walking and health services use and costs (adjusting for sociodemographic characteristics, chronic disease, health status, and previous utilisation) | Health services use, and total costs | Health services use and cost data obtained from the RPAHS for the 12 months following interviews (Primary care, emergency, hospital costs as charges calculated on an annual basis) |
Tsuji et al. 2003 [22] | Impact of walking upon medical care expenditure in Japan | Japan | n = 27,431 NHI claims history files | Prospective cohort study | Japanese men and women, aged 40–79 years. National Health Insurance (NHI) beneficiaries in rural Japan | Logistic regression model and multivariate models to adjust for the effect of potential socio economic confounders. Persons walked for < 30 min, 30 min – 1 h, > 1 h (per day) | Medical care Expenditure | Insurance claims history; uniform national fee schedule determines prize for each service prospectively collected for 4 years (charges for outpatient and inpatient care) |
Hirai et al. 2021 [30] | Physical Activity and Cumulative Long-Term Care Cost among Older Japanese Adults: A Prospective Study in JAGES | Japan | n = 34,797 Japan Gerontological Evaluation Study (JAGES) | Prospective cohort study | Community-dwelling people aged 65 years or older, with no physical or cognitive disabilities | Generalized linear model with Tweedie distribution and log-link function, adjusted for socio economic confounders. Comparison of individuals by 3 categories (< 30 min, 30–59 min, 60 min) of time spent walking | Cost of long-term care insurance services | The outcome variable was the cumulative cost of LTCI services during the follow-up period. Documented in the Japan Gerontological Evaluation Study (JAGES) |
Walking in Leisure Time | ||||||||
Turi et al. 2015 [25] | Walking and health care expenditures among adult users of the Brazilian public healthcare system: retrospective cross-sectional study | Brazil | n = 963 Basic Health Units | Retrospective cross-sectional study | Patients aged ≥ 50 years | Logistic regression model and multivariate models to adjust for the effect of potential socio economic confounders. Walking (never, seldom, sometimes, often, always) during leisure-time and healthcare expenditure in primary care | Total medical expenditures | Expenditure on consultations, laboratory tests and medical consultations transformed into currency by standard table (provided by Brazilian government) |
Walking Speed | ||||||||
Purser et al. 2005 [23] | Walking speed predicts health status and hospital costs for frail elderly male veterans | USA | n = 1,388 Department of Veterans Affairs (VA) multicenter clinical trial | Cohort study | Medical or surgical patients > 65 years. Geriatric Evaluation and Management (GEM) program | Multivariate models to adjust for the effect of potential socio economic confounders, assessing baseline in gait speed and absolute change over 1 year | Inpatient health services use, and total costs | Data on length of stay, number, and charges of inpatient consultations, rehabilitation and social work visits as available from Veterans Affairs databases |
Bonnini et al. 2020 [28] | Improving walking speed reduces hospitalization costs in outpatients with cardiovascular disease. An analysis based on a multistrata non-parametric test | Italy | n = 649 Exercise-based secondary prevention program | Prospective cohort study | Patients participating in an exercise-based secondary prevention program (average age 63) | Multi-strata permutation test after propensity score matching. Patients divided at baseline into two groups characterized by low and high WS (based on the average WS maintained during a moderate 1-km treadmill-walking test) | All-cause hospitalization and related costs | Hospitalization related costs (Process of transformation of hospitalization rates to costs is not described.) |
Okayama et al. 2021 [29] | Clinical impact of walking capacity on the risk of disability and hospitalizations among elderly patients with advanced lung cancer | Japan | n = 60 Shizuoka Cancer Center | Prospective cohort study | Patients aged ≥ 70 years with advanced non-small-cell lung cancer (NSCLC) | Recurrent event analysis comparing (not adjusted for socioeconomic confounders) comparing Mobile group vs less mobile group | Length of hospital stay, inpatient medical costs | Inpatient medical costs. Electronic health records of hospitals, actual revenue that the hospital was paid from the health insurance funds for a given inpatient stay |
Number of Steps | ||||||||
Kato et al. 2013 [24] | Effects of walking on medical cost: A quantitative evaluation by simulation focusing on diabetes | Japan | n = 1,000 hypothetical subjects | 10 year- Markov model | Hypothetical subjects representing middle aged Japanese people | Markov Model (rates of events were determined on papers and statistical data published in Japan) Patients after 10 years with 0 steps, + 3,000 steps, and + 5,000 steps | Total number of events during 10 years, Medical costs during 10 years | Medical costs for diabetes (inpatient and outpatient costs) associated with each health status events estimated from public statistical data in Japan |
Karl et al. 2018 [27] | Direct healthcare costs associated with Device-based assessment, and self-reported physical activity: results from a cross-sectional population-based study | Germany | n = 477 KORA FF4 study | Retrospective cross-sectional study | Patients aged between 48 and 68 years | Two-part gamma regression model adjusted for potential socioeconomic confounders comparing inactive participants to active subjects (very low moderate-vigorous physical activity (MVPA) vs. very high MVPA) | Total healthcare costs | Direct medical costs (based on physician visits, ambulatory hospital visits, and drugs) calculated using national unit costs, as recommended by the Working Group Methods in Health Economic Evaluation (AG MEG) |
Kabiri et al. 2018 [26] | The Long-Term Health and Economic Value of Improved Mobility among Older Adults in the United States | USA | n = 12.6 million Medical Expenditure Panel Survey (MEPS 2012) | Micro Simulation | Patients ≥ 51 years with osteoarthritis | Six-step process to model the effect of improved mobility though improvements in quality of life measures on health economic outcomes. Comparing “status quo” population (pre-treatment mobility levels) with the “mobility improvement” population | Medical expenditures, Nursing home utilisation | THEMIS microsimulation tracked individuals to translate changes in quality of life to health economic outcomes derived from MEPS data base (included costs were not described) |