The Long-Term Health and Economic Value of Improved Mobility Among Older Adults in the United States

Abstract

Introduction:

Mobility impairments have substantial physical and mental health consequences, resulting in diminished quality of life. Whereas most studies on the health economic consequences of mobility limitations focus on short-term implications, this study examines the long-term value of improving mobility in older adults.

Methods:

Our six-step approach used clinical trial data to calibrate mobility improvements and estimate health economic outcomes using a microsimulation model. First, we measured improvement in steps per day calibrated with clinical trial data examining hylan G-F20 viscosupplementation treatment. Second, we created a cohort of osteoarthritis patients aged ≥51. In the third step, we estimated their baseline quality of life (QoL). Fourth, we translated steps per day improvements to changes in QoL using estimates from the literature. Fifth, we calibrated QoL in this cohort to match those in the trial. Lastly, we incorporated these data and parameters into The Health Economic Medical Innovation Simulation (THEMIS) model to estimate how mobility improvements affect functional status limitations, medical expenditures, nursing home utilization, employment, and earnings between 2012 and 2030.

Results:

In our sample of 12.6 million patients, 66.7% were female and 70% had BMI>25 kg/m². Our model predicted that a 554-steps-per-day increase in mobility would reduce functional status limitations by 5.9%, total medical expenditures by 0.9%, and nursing home utilization by 2.8%, and increase employment by 2.9%, earnings by 10.3% and monetized QoL by 3.2% over this 18-year period.

Conclusions:

Interventions that improve mobility are likely to reduce long-run medical expenditures and nursing home utilization, and increase employment.

Introduction

By 2030, mobility impairments in older persons are projected to affect 25 million Americans, or 9.3% of the population [1], and cause reductions in physical activity. Lack of physical activity often leads to greater frailty in older persons that may lead to falls [2], frequently resulting in fractures, head trauma, soft tissue injuries, or severe lacerations, and even an increased risk of dying [3–5]. Oftentimes, patients develop a fear of falling [6] and experience depression, feelings of helplessness and social isolation that ultimately reduce patients’ quality of life [7–9]. More generally, mobility is commonly considered a key component of quality of life. For example, the Euroqol-5 dimension (EQ-5D-3L—EQ-5D for brevity) questionnaire [10], a widely-used survey instrument, incorporates mobility as one of its five determinants of quality of life.

Though the burden and clinical consequences of reduced mobility on patient health outcomes and quality of life have been well established, the effect of mobility on health economic outcomes is often overlooked or largely focuses on short-term economic outcomes. Studies of disability trends among the elderly and near elderly found that mobility limitations increased between 1996 and 2010 [11]. Evidence on the short-term implications of mobility limitations has demonstrated their negative impact on a variety of economic outcomes. For instance, an analysis of data from the National Health Interview Survey demonstrated that decreased physical function reduced the likelihood of employment, decreased household income and increased missed work days for employed osteoarthritis patients [12]. Additionally, those with at least one limitation in activities of daily living (ADL), such as walking, had double the risk of admission to a nursing home [13]. Despite the well-known risks and consequences of reduced mobility, to our knowledge, no study has examined the long-term impact of maintained improvements in mobility on health economic outcomes among older population. To address this limitation, our study examined the long-term effect of improved mobility on health economic outcomes projected through 2030. Since osteoarthritis of the knee is a common contributor to decreased mobility, we focused our analysis on a nationally representative population of patients with osteoarthritis and calibrated mobility improvements for all osteoarthritis patients, using hylan G-F20 viscosupplementation. The specific type of intervention chosen, however, was not a major consideration of this study as we aimed to quantify the benefits of mobility improvements generally rather than examine the cost-effectiveness of a particular intervention. Using The Health Economic Medical Innovation Simulation (THEMIS) [14–18], we assessed the impact of improved mobility on functional limitations, medical expenses, nursing home utilization, monetized QALYs (assuming $150,000 per QALY [19]), employment, and earnings.

Methods

Our analytic approach relied on a six-step process to model the effect of improved mobility, as measured by increased steps per day, on health economic outcomes for older persons (Figure 1). This process involved aligning data from multiple sources because no single source of data included steps per day, health-related quality of life, and measures of the downstream effects of increased mobility (e.g. functional status, medical expenditures, and labor force participation). First, we searched for a study that showed mobility improvements with an osteoarthritis mobility-improving treatment. We identified mobility improvements based on a clinical trial, MARCHE, that estimated the impact of treating osteoarthritis patients with a hylan G-F20 viscosupplementation injection on mobility, measured as steps per day, after 90 days [20, 21]. Second, we selected a nationally representative cohort of osteoarthritis patients aged 51 and older, based on the 2012 Medical Expenditure Panel Survey (MEPS) data. Third, we estimated a model to predict patient utility based on EQ-5D data of the osteoarthritis cohort as a function of individual characteristics, including functional status as assessed by ADL and instrumental activities of daily living (IADL) limitations. Fourth, to translate improvements in steps per day to changes in health economic outcomes, we identified models in the literature that estimated patient quality of life using the EQ-5D questionnaire among patients in different steps-per-day groups. Using these models, we extrapolated changes in steps per day from the clinical trial to changes in EQ-5D for the nationally representative osteoarthritis cohort. Fifth, we identified the levels of functional limitation (ADL and IADL limitations) in the osteoarthritis cohort that would represent the mobility levels observed in the pre- and post-treatment populations from the trial. Lastly, simulated two populations (with and without treatment) using THEMIS, allowing us to estimate the effect of improved functional status on health economic outcomes.

Figure 1. Steps to estimate the impact of improved mobility on health-related economic outcomes.

* Steps 4–6 were performed separately for each simulation scenario –standard of care (baseline) and mobility improvement– by steps-per-day group and sex. The proportions of males and females in each steps-per-day group in each scenario were obtained from the MARCHE trial.

Abbreviations: OA = osteoarthritis; EQ-5D = the Euroqol-5 dimension (EQ-5D) questionnaire designed to measure quality of life; ADL = activities of daily living; IADL = instrumental activities of daily living; THEMIS = The Health Economics Medical Innovation Simulation; MEPS = Medical Expenditures Panel Survey.

We estimated the effect of changes in improved mobility by comparing the results from the “status quo” population (pre-treatment mobility levels) to the “mobility improvement” population (mobility levels consistent with hylan G-F20 treatment), assuming that all patients with osteoarthritis pursued mobility-improving treatment and maintained their improved mobility relative to their mobility trend over time through recurrent treatments. Each methodology step is described in more detail below.

Step 1: Measure mobility improvements

We obtained the effect of osteoarthritis treatment on patient mobility using data on steps per day from the MARCHE trial (Supplement 1). The trial was a pre-post cohort study that measured the effect of using hylan G-F20 on patient’s mobility as measured by steps per day. We selected this trial because the number of steps per day captured patient’s mobility, irrespective of speed and endurance. Steps per day also incorporates patient’s lifestyle choices in addition to their physical ability to move—which is commonly assessed through other tests such as the 6-minute walk distance [22]. Prior to treatment, patients in the study walked between 643 and 9,242 steps per day with a median of 4,125 (interquartile range: 3,000–5,449). After treatment, patients walked between 812 and 15,939 steps per day with a median of 4,919 steps per day (interquartile range: 3,089–7,640). On average, patients treated with hylan G-F20 walked 554 more steps per day (p=0.001).

Step 2: Create the cohort of osteoarthritis patients

We next created the cohort of osteoarthritis patients using data from the 2012 MEPS. We selected MEPS because it is drawn from a nationally representative sample of households with information on patient demographics, health conditions, health care utilization and expenditures. Osteoarthritis was identified using ICD-9 diagnosis codes (Supplement 2). The characteristics of this cohort are presented in Table 1.

Table 1.

Characteristics of osteoarthritis patients aged 51 and older in the 2012 MEPS

	Male	Female	Total
N	4,181,674	8,443,867	12,625,540
Age
51–60 years	33.3%	29.5%	30.8%
61–70 years	32.1%	33.1%	32.7%
71–80 years	20.6%	21.6%	21.3%
≥81 years	14.0%	15.8%	15.2%
BMI
<25 kg/m2	22.1%	33.8%	29.9%
≥25 kg/m2	77.9%	66.2%	70.1%

Notes: MEPS = Medical Expenditures Panel Survey; BMI = body mass index.

Step 3: Estimate EQ-5D based on the MEPS data

We next created a general regression model to estimate EQ-5D measures for the cohort of osteoarthritis patients in MEPS. The purpose of this model was to include variables that impacted quality of life, such as patient demographics, ADL limitations and IADL limitations. Unlike steps per day, these variables are included in large datasets such as the MEPS and Health and Retirement Study (HRS), which also include other parameters to estimate health economic outcomes. We developed the regression model using 2001–2003 MEPS data to predict EQ-5D for osteoarthritis patients (MEPS-EQ-5D) (Supplement 3). Data from this time period were selected because it is the time period during which MEPS included the EQ-5D questionnaire in surveys. The number of ADL and IADL limitations were included along with other covariates in this regression model. Decreasing the number of ADL and IADL limitations in this model resulted in improved EQ-5D values.

Step 4: Translate changes in steps per day into quality-of-life improvements

We translated improvements in mobility—measured as steps per day in the MARCHE trial—to the EQ-5D, a quality of life metric reported in MEPS. Prior studies have found that mobility and quality of life, measured by the EQ-5D, are significantly correlated among older persons. Further, there is a significant association between mobility and each component of the EQ-5D [23]. Since data are not available to directly convert steps per day to the EQ-5D, we first converted steps per day to the SF-36 (Supplement 4), and then converted the SF-36 to EQ-5D. First, we used data from a study by Yasunaga et al. [24], which estimated variations in SF-36 score across different mobility ranges measured by steps per day, to convert steps per day to the SF-36.

We obtained the number of females and males in the MARCHE trial, at initial assessment and after mobility-improving treatment, that fell into four groups according to the number of steps per day (Figure 2), and assigned similar proportion of patients in each steps-per-day group to the osteoarthritis cohort in MEPS, assuming similar mobility improvements in the MEPS cohort for each sex. We then assigned the values for eight SF-36 dimensions using the values associated with each dimension described in Yasunaga et al. (Supplementary Table 2) to the MEPS osteoarthritis patients in each steps-per-day group (Supplementary Table 3).

Figure 2. Percentage of osteoarthritis patients in each steps-per-day group under each simulation scenario.

Note: The ranges of each steps-per-day group were defined by Yasunaga et al. (24). According to these ranges, we identified the number of individuals who fell into each group according to the MARCHE trial results. Baseline was the steps per day walked by those receiving the standard of care treatment, and mobility improvement was the steps per day walked by patients after treatment with hylan G-F 20. Panel A and B represent data on female and male osteoarthritis patients, respectively.

Second, we estimated the impact of changes in SF-36 on the overall patient quality of life measured by EQ-5D-based utility values [10]. Using a regression model from Ara et al., which was based on 6,350 patients [25], we converted SF-36 values to EQ-5D-based utility values. By applying Ara’s EQ-5D regression equation to patients in each steps-per-day group, we estimated the quality of life for the cohort of osteoarthritis patients (Ara-EQ-5D). Supplement 5 provides additional details for this stage.

Step 5: Determine the underlying functional status profile of the populations

Next, we translated the mobility improvements shown by increased number of steps per day in the MARCHE trial to the MEPS osteoarthritis cohort. We determined the appropriate level of functional limitations (measured by prevalence of ADL and IADL limitations) that would yield levels of utility comparable to the Ara-EQ-5D values for each steps-per-day group. The number of ADL and IADL limitations for each individual were conditional on having any limitations, and were the main parameters in the estimation of EQ-5D in the cohort of osteoarthritis population (Supplement 3). For example, if the MEPS-EQ-5D value for a steps-per-day group among females was higher than the Ara-EQ-5D value for the same group, the prevalence of ADL or IADL limitations was increased to eliminate (or at least reduce) the discrepancy.

Step 6: Translate quality-of-life improvements into health economic outcomes

Finally, we used THEMIS, a microsimulation that tracks individuals 51 years and older, to translate changes in quality of life to health economic outcomes. THEMIS incorporates large datasets such as the HRS and MEPS, with capabilities to make long-run projections of health economic outcomes (Supplement 6) [16, 17, 26–29]. As a result, we were able to incorporate the MEPS osteoarthritis cohort with calibrated prevalence of ADL and IADL limitations for each steps-per-day group in each scenario along with the regression model to predict MEPS-EQ-5D THEMIS. As previously mentioned, the MEPS-EQ-5D regression model was designed to calculate the impact of the reduction in the prevalence of osteoarthritis patients with ADL and IADL limitations (as a measure for improved mobility) to changes in quality of life. Through the calibration process (step 5), we rescaled the values of MEPS-EQ-5D to match Ara-EQ-5D by calibrating the prevalence of ADL and IADL limitations for each steps-per-day group under each simulation scenario.

Effect of the prevalence of ADL and IADL limitations on health economic outcomes

In THEMIS, changes in the prevalence of ADL and IADL affected individual’s health related economic outcomes such as the risk of living in a nursing home, employment status, earnings, and medical expenditures between 2012 and 2030. THEMIS utilizes a number of projection techniques and multiple datasets such as the HRS, MEPS, and Medicare Current Beneficiary Survey, to estimate the trends of health economic outcomes in the United States. The health economic outcomes of each individual were simulated according to estimated regression models, with various coefficients including but not limited to the individual’s demographics, comorbidities, and functional status, (i.e., the number of ADL and IADL limitations, various lifetime trajectory of health, functional status, and medical expenditures for individuals in THEMIS). Therefore, the changes in the prevalence of ADL and IADL limitations translated into different health and economic status of individuals.

Since the status quo and mobility improvement simulation scenarios were different in their adjustments for the prevalence of having ADL or IADL limitations, we estimated the impact of improved mobility by comparing the outcomes of two simulation scenarios. These outcomes included the total number of individuals with functional status limitations, monetized QALYs (assuming $150,000 per QALY [19]), number of people living in nursing homes, medical expenditures, and changes in employment status and earnings due to improved mobility through 2030. The total number of individuals with functional limitations were measured as the total number of individuals with any ADL or IADL limitations. Nursing home living status was defined as new entry into a short or long-term care facility from the community. The estimation of medical expenditures included total expenditures and costs to Medicare (Parts A, B, and D) and Medicaid. The impacts on fiscal outcomes of improved earnings and productivity were also included in our projections.

Sensitivity analyses

We conducted two sensitivity analyses to analyze the impact of changing our model input parameters on the estimated changes of health economic outcomes. First, we tested the hypothesis of parameter equivalence using two-sample t-tests to demonstrate statistical similarities between the outcomes of two simulation scenarios, by bootstrapping outcomes’ means and standard errors with replacement for 1,000 replicates. Second, we analyzed the impact of improved mobility on health economic outcomes assuming that individuals maintained the treatment effect through recurring treatments over one-, five-, and ten- year periods as opposed to the original eighteen years.

Results

Of the 12,625,540 patients in the 2012 MEPS data, 66.7% were female, 84.8% were younger than 81 years, and 70.1% had a BMI over 25 kg/m². Mobility improvements reduced the number of osteoarthritis patients with functional status limitations. According to our 18-year simulation, osteoarthritis patients experienced approximately 125 million patient-years of ADL limitations under the status quo scenario (Supplementary Figure 2). After we applied our mobility improvement simulation in which patients walked an additional 554 steps per day on average, this figure decreased by 5.9% (−7.4 million) to 118 million patient-years of ADL limitations. On a per person basis, an average of 89.7% of patients experienced any ADL limitation each year under the status quo scenario, while in the mobility improvement scenario that number was reduced to 83.8%. Over time, a larger share of patients experienced ADL limitations in each scenario, though in each year the percentage of individuals with ADL limitations was lower under the mobility improvement scenario compared to the status quo scenario (Supplementary Figure 3). These improvements in ADL limitations translated to a cumulative 3.2% improvement in quality of life ($11.3 to $11.7 trillion) from 2012 to 2030 when we assumed a value of $150,000 per QALY (Supplementary Figure 4).

Under the status quo scenario, the cumulative healthcare expenditures were $4,837 billion from 2012 to 2030 with the majority of expenditures incurred by Medicare (Supplementary Figure 5). Medicare and Medicaid healthcare expenditures were estimated at $2,580 billion and $825 billion respectively. Under the mobility improvement scenario, however, overall healthcare expenditures were reduced to $4,793 billion, generating savings of $43.9 billion (−0.91%) over 18 years. Though the distribution of expenditures among payers matched that of the status quo scenario, expenditures were estimated at $2,551 billion for Medicare ($28.7 billion reduction) and $806 billion for Medicaid ($19.0 billion reduction). The projected annual savings in healthcare expenditures are illustrated in Figure 3.

Figure 3. Estimated medical savings from mobility improvements among osteoarthritis patients (2012–2030).

Increasing the number of steps per day for osteoarthritis patients by 554 on average, decreased overall nursing home utilization. Under the status quo scenario, the number of patients who required a nursing home was estimated at 342,543 patients in 2012, with an upward trend in nursing home use each year as patients aged (Supplementary Figure 6). Nursing home use also increased under the mobility improvement scenario, however, the rate of increase was slower, especially in later years. Thus, patients with improved mobility required fewer nursing home stays, especially later in life. Overall, this led to a 2.8% decrease in nursing home utilization for patients with mobility improvements relative to those in the status quo scenario.

Our model also projected that improved mobility would increase employment over an 18-year time horizon. Specifically, patients aged ≥51 years contributed 44.0 and 43.3 million years of employment to the workforce, which translated to $751 and $829 billion in earnings, under the status quo and mobility improvement scenarios, respectively (Supplementary Figures 7–8). Thus, we projected that improved mobility would increase the number of years that individuals were employed by 1.2 million (2.9%) and the overall earnings by $78 billion (10.3%) over this 18-year time span. Employment and earnings of patients with mobility improvements was higher than the status quo scenario cohort in all years (Supplementary Figures 9–10). Overall, employment and earnings fell over time in both groups as patients either withdrew from the labor force, lost employment or died.

Table 2 summarizes the results for ADL limitations, monetized QALYs, medical expenditure, and nursing home utilization, employment and earnings from 2012 to 2030. The total value of improved mobility—incorporating gains in QALYs and earnings, and savings in total medical expenditures—was $481.6 billion across the 18 years of the simulation. The largest benefit (in percentage terms) was increased earnings (10.3%). ADL limitations and nursing home use declined by 5.9% and 2.8%. Total medical expenditures and government medical expenditures by Medicare and Medicaid decreased by 0.9%, 1.1% and 2.3% respectively.

Table 2.

Estimated cumulative changes in model outcomes due to mobility improvements among osteoarthritis patients (2012–2030)

Outcome	Cumulative Change	Cumulative Percent Change
Number of Patient-Years with ADL Limitations	−7.4 (million)	−5.94%
Monetized QALYs*	360.1 ($billion)	3.18%
Total Medical Expenditures	−43.9 ($billion)	−0.91%
Medicare Expenditures	−28.7 ($billion)	−1.11%
Medicaid Expenditures	−19.0 ($billion)	−2.30%
Number of Patient-Years in Nursing Home	−0.5 (million)	−2.82%
Number of Patient-Years of Employment	1.2 (million)	2.87%
Total Earnings	77.7 ($billion)	10.34%
Total Value to Society**	481.6 ($billion)

^*Assumed $150,000 per QALY.

^**Total value to society was calculated as the sum of gains in earnings and monetized QALYs and savings in total medical expenditures.

Abbreviations: ADL = activities of daily living; QALY = quality-adjusted life year

Sensitivity analyses

We rejected the null hypothesis of statistical similarity based on our bootstrapping results around the parameters used for the estimation of model results for two scenarios, indicating that differences between scenarios were statistically significant at 95% confidence level (p <0.001 for all outcomes) (Supplementary Table 5). In our second sensitivity analysis, improving mobility over shorter time periods mirrored the effect of reductions over an 18-year time horizon, and the relative magnitude of change in each outcome (except employment) was larger (Supplementary Table 6). From 2012 to 2022, the largest percentage change occurred in total earnings (8.4%). Number of patient-years with ADL limitations declined by 6.4%, nursing home use also declined by 3.9% and employment increased by 1.5%. Medicare and Medicaid expenditures decreased by 1.7% and 3.5%, respectively. Detailed results on each health economic outcome during this time frame are provided in Supplementary Figures 11–16.

Discussion

We found that improved mobility statistically significantly reduces functional limitations, medical expenses, nursing home utilization and increases quality of life and earnings among osteoarthritis patients over the next 18 years after treatment. We estimated that improved mobility would add $481.6 billion of value to society over this time period. To our knowledge, no prior studies have examined the long-term impact of improved mobility on health economic outcomes in older patients with osteoarthritis. Our study quantifies this long-term impact among a nationally representative cohort of osteoarthritis patients by analyzing the social value of improving mobility as a whole, instead of only considering disease-specific healthcare costs.

Other studies have determined the short-term economic impact of improvements in mobility. Using data from the National Health Interview Survey, Dall et al. arrived at similar conclusions that improved physical function is associated with higher likelihood of employment, increased household income and decreased missed work days for employed osteoarthritis patients [12]. Studies examining factors that predict nursing home admissions found that risk of admission more than doubled for patients with 1–2 ADL limitations (OR: 2.45; 2.02–2.97) and more than tripled for those with ≥3 ADL limitations (OR: 3.25; 2.59–4.09) versus those with no ADL limitations. This finding aligns with our result that improving mobility reduces nursing home stays.

Various interventions can be implemented to attain the benefits of mobility improvements. Knee replacement [30], viscosupplementation treatments [31], and chronic disease prevention programs such as EnhanceFitness [32] and Fit and Strong![33], have successfully demonstrated improvements on a number of physical function measures. Other studies have shown that interventions such as those previously described are cost-saving [34, 35].

When making coverage decisions for interventions that improve mobility, it is unclear whether payers and policymakers are fully internalizing the long-run benefits of interventions that improve mobility. With healthcare expenditures projected to rise 6.0% on average through 2025 [36], investing in mobility improving interventions will be essential to partially offset some of the likely rise in healthcare expenditures in the coming years. Furthermore, with the baby boomer generation aging into prime ages for nursing home use and increased healthcare utilization, demand for healthcare services and nursing homes will likely increase in the coming years. Improving mobility may reduce the burden on providers and payers, especially Medicare. On the other hand, some payers that invest in interventions to improve mobility may not see the benefits of these interventions. Commercial insurers often see an annual turnover of 20% [37]. Among Medicaid enrollees, 43% of adults and 26% of children disenroll within 12 months. These high rates of churn may make investments in mobility improvements—while valuable from a patient and societal perspective—a poor investment for many payers.

There are several limitations to our study. First, mobility improvements were calibrated based on a single treatment, hylan G-F20, while other treatments might have a higher or lower impact on mobility. Second, although the MARCHE trial examined patients with osteoarthritis of the knee, we used the MEPS data, which only identified patients with osteoarthritis in general, to create our cohort of patients. Thus, our model assumed that all osteoarthritis patients experience a mobility improvement, even though the intervention used to calibrate the mobility improvement was specific to osteoarthritis of the knee. As 56% of symptomatic osteoarthritis incidence occurs in the knee [38], a back-of-the-envelope calculation of the benefits of mobility improved for patients with knee osteoarthritis would apply that percentage to figures from the boarder osteoarthritis cohort estimated in this study. Third, our model assumed that mobility improvement would be maintained by individuals over time through recurrent treatments. In reality, the treatment effect of many mobility interventions may not be maintained over long periods of time. We included a sensitivity analysis around the duration of maintained improved mobility due to treatment, in order to address this limitation and analyze its impacts on model outcomes. Fourth, our model used two studies to translate steps per day to EQ-5D as a measure of quality of life. These studies included different sample populations without osteoarthritis, which could impact the accuracy of EQ-5D estimation for osteoarthritis patients in THEMIS. Fifth, the magnitude of change in steps per day was based on a clinical trial that measured mobility improvements; however, the effectiveness of mobility interventions may be different in the real world. Further, although THEMIS controls for other factors (e.g., age, race, gender, comorbid conditions) when predicting the relationship between mobility and health economic outcomes, these relationships represent current correlations in the data; the true causal effect of improved mobility on quality of life in the real-world may differ from our estimates. Sixth, our model examined only changes in medical expenditures due to mobility improvements rather than all health care costs associated with the osteoarthritis disease and its treatment. Seventh, when considering our analysis in a real-world setting, it is unclear whether our estimates should be viewed as conservative. A new cohort of patients with osteoarthritis will be treated with a mobility improving treatment in each year. Since our model only followed patients with osteoarthritis in 2012, we did not observe the impact of this additive effect. Our simulation showed diminishing returns to investments in mobility over time for most health economic outcomes. Yet in a real-world setting, the benefits of treating patients today accrue above and beyond the improvements reaped by those treated in the prior years. Finally, our study measures only the benefits of improved patient mobility but does not consider the cost of interventions that brought about this change; future research should measure the cost effectiveness of specific mobility interventions, including hylan G-F20 viscosupplementation.

Conclusion

Increased mobility as measured by steps per day is associated with improved quality of life and fewer ADL limitations. Further, improved mobility decreased health care expenditures and nursing home utilization, and increased quality of life, employment and earnings. These findings suggest that the societal benefits of improved mobility may be underappreciated by providers, payers, and policymakers.

Highlights:

• Patients with mobility limitations are at high risk of morbidity, mortality and decreased quality of life.

• This paper expands on prior research demonstrating the short-term health economic consequences of mobility impairments by estimating the long-term value in older adults.

• Interventions that improve patient mobility are likely to reduce functional status limitations, decrease medical expenditures and nursing home utilization, and increase employment.