The value of walking: a systematic review on mobility and healthcare costs

Search results

Results are presented as statistically significant differences of costs in individual studies. Due to heterogeneity in types of cost outcomes, settings, and disease areas, it was not appropriate to synthesize the results or conduct a meta-analysis of the economic findings. The initial search strategy yielded 2,771 titles after the elimination of duplicates. For the final qualitative synthesis, ten studies were included. These studies were conducted in Japan (n = 4), United States (n = 3), Brazil (n = 1), Italy (n = 1), and Germany (n = 1).

Study characteristics

Studies were conducted between 2001 and 2021. Detailed information on the individual studies is shown in Table 2. Six studies were observational cohort studies [21,22,23, 28, 29], two were cross-sectional studies [25, 27], one was a decision analysis using a Markov Model [24], and one was a microsimulation [26]. All studies (n = 10) examined the economic impact of mobility in middle-aged to aged populations. All studies focused on direct healthcare costs (including total costs but also costs by individual health sector).

Table 2 Study characteristics

The description of the source of cost data and the calculation of costs varied considerably across studies. A detailed description was absent in one study [28] but it can be assumed that estimates of hospitalization rates were based on a regional Health Service Registry. Outpatient/emergency room visits or inpatient hospital stays were gathered from institutional electronic health records [21, 25, 29], or insurance databases [22, 23]. Karl et al. [27] directly questioned patients. Hirai et al. [30] used data from the Japan Gerontological Evaluation Study (JAGES), which collected information about the costs from the municipalities that also act as insurers. Kato et al. estimated costs from public statistical data in Japan [24] and the microsimulation study by Kabiri et al. used THEMIS (The Health Economic Medical Innovation Simulation) to estimate how mobility improvements affect medical expenditures through monetized quality adjusted life years including data from MEDICARE, and MEDICAID [26].

Quality of reporting

The assessed quality of reporting is shown in Table 1. All publications outlined the background adequately to understand the research need and the research question. All studies reported cost differences due to a change in walking parameters. Costs that should be included in an analysis depend on the study perspective (refers to the point of view one takes when assessing costs), so failing to state the perspective meant that some of these studies lacked a clear rationale for the types of cost included. Five studies [24, 26, 28, 29] did not adjust costs that occurred at different points in time, and two studies [23, 26] conducted a sensitivity analysis to address a certain variety of their assumptions.

Reported results

Cost results are shown in Table 3. Perkins A, Tsuji I, Purser JL, Kato M, Turi B, Kabiri M. [21,22,23,24,25,26, 28, 29] reported that lower levels of walking ability were associated with higher health care expenditure, and one study [27] reported no statistically significant association between mobility and health care costs. Six studies reported additional associations between walking parameters and health care utilisation [21,22,23, 26, 28, 29].

Methods used to assess walking ability

The majority of walking assessments consisted of patient-reported outcomes (PROs) collected via questionnaires. These assessments included the following parameters: walking time [21, 22], and walking during leisure time as part of activities of daily life [25]. Specifically, Perkins et al. examined walking time by documenting the minutes of walking per week, using a newly developed questionnaire [21]. Tsuji et al. obtained mobility data from a survey conducted in 1994, which included a question on walking time asking how long on average patients walk a day [22]. Hirai et al. assessed walking time per day with a single question (“How long do you walk a day, on average?”). The time spent walking was categorized as > 60 min, 30–60 min, and less than 30 min per day. Turi et al. assessed walking during leisure time by using the section ‘physical activity during leisure-time’ of the ‘Baecke12’ questionnaire [25]. Walking ability was also assessed by determining walking speed using performance tests. Purser et al. used the Reubens Physical Performance Test, a supervised performance test [23]. Walking speed was examined by Bonnini et al. using a 1-km treadmill-walking test [28], and by Okayama et al. using the Endurance shuttle walk test [29]. Simulation studies did not assess walking directly. They furthermore examined quantitative risk reduction by walking derived from published studies to calculate the steps taken in their cost simulations [24, 26]. The study by Karl et al. measured the number of steps with a portable accelerometer device (Actigraph GT3X) [27].

Association with health care costsWalking time and leisure-time walking

In a cohort of community-dwelling adults older than 55 years, Perkins et al. found an association between self-reported walking time of 120 min a week or more, and a significant decrease in emergency room visits and hospital stays in the following year. Annual total ($1,856 vs $6,266 $), inpatient ( $1,184 vs $4,872), and emergency room costs ($253 vs $762) were less for those reporting 60 or more minutes of walking per week compared to those reporting less than 60 min of walking per week [21]. In a four-year-long prospective cohort study in Japanese men and women, aged 40–79 years, Tsuji et al. found that medical costs ($86 vs. $97) were 12% significantly lower per capita and month, for subjects walking for more than one hour/day than for those walking less than one hour/day [22]. Hirai et al. reported that time spent walking was negatively associated with the cumulative costs of long-term care insurance. These cumulative costs were significantly higher in those who walked for less than 30 min than in those who walked for more than 60 min. Turi et al. reported the association of self-reported walking during leisure time with total healthcare expenditure during one year prior to the date of the interview in Brazilian patients (randomly selected users of the Brazilian National Health System) aged ≥ 50 years. Individuals who ‘always walked’ were 41% less likely to be in the highest 25% quantile (an indicator of high expenditure) of incurred health-care cost when compared to individuals who ‘never walked ‘ [25] (Table 2).

Walking speed

In a frail population of hospitalised medical or surgical patients older than 65 years, Purser et al. found that when the baseline walking speed was 0.10 m/s higher, this was associated with $1,334 lower 1-year costs during the index hospitalization [23]. Bonnini et al. conducted an intervention study to evaluate the effects of an unsupervised home program in patients with cardiovascular disease, consisting of 30–60 min of brisk walking at least 3–4 days per week over 3 years, on rates of hospitalization. Between four and six years after baseline, a significant lower hospitalization rate was observed in patients that had highly improved their walking speed compared to those who had only improved their walking speed to a low extent. This resulted in an average cost reduction per patient between high and low improvers in walking between $489 and $882 [28]. Okayama et al. prospectively enrolled patients aged ≥ 70 years with advanced non-small-cell lung cancer to investigate the association of pre-treatment walking capacity with hospitalization rates and medical costs. During the first year of initial therapy, medical costs (the actual revenue the hospital was paid from the health insurance funds) did not differ between less and more mobile groups, but significantly higher additional inpatients costs ($8,076 per person) were reported for the less mobile group [29].

Number of steps

Karl et al. investigated direct medical costs of patients aged between 48 to 68 years. They used cross-sectional data of the population in the German KORA FF4 study. In a subsample of patients for whom daily step count was reported there was no statistically significant difference in costs between those who walked more than 10,000 steps per day and those who did not [27]. Kato et al. 2013 used a Markov model to simulate costs over 10 years for middle-aged Japanese patients with diabetes. They estimated that total medical costs could be 5.2% and 8.4% lower for daily step count increases of 3,000 and 5,000, respectively [24]. Kabiri et al. conducted a microsimulation study of patients aged ≥ 51 years with osteoarthritis, and reported that 554 steps more per day would be associated with a 0.9% reduction in total medical expenditure [26].

留言 (0)

沒有登入
gif