- Email: [email protected]
Cardiorespiratory Fitness and Health Care Costs in Diabetes: The Veterans Exercise Testing Study
ARTICLE IN PRESS CLINICAL RESEARCH STUDY
Cardiorespiratory Fitness and Health Care Costs in Diabetes: The Veterans Exercise Testing Study Jonathan Myers, PhD a,b,, Christina Grne de Souza e Silva, PhD c, Rachelle Doom, MD a, Holly Fonda, MS a, Khin Chan, MD a, Shirit Kamil-Rosenberg, EdD a, Peter Kokkinos, PhD d a
Veterans Affairs Palo Alto Health Care System, Calif; b Division of Cardiovascular Medicine, Stanford University, Calif; c Heart Institute Edson Saad/Federal University of Rio de Janeiro, Brazil; d Veterans Affairs Medical Center, Washington, DC.
BACKGROUND This study aimed to determine the association between cardiorespiratory fitness and healthcare expenditures among individuals with and without diabetes. METHODS: Health care costs were quantified among 3924 consecutive men (mean age 58 ± 11 years) referred for a maximal exercise test, and compared according to presence (n = 2457) and absence (n = 1467) of diabetes and fitness. Fitness was classified into 4 categories based on age-stratified quartiles of peak metabolic equivalents: least-fit (5.1 ± 1.5 metabolic equivalents; n = 1044), moderately-fit (7.6 ± 1.5 metabolic equivalents; n = 938), fit (9.4 ± 1.5 metabolic equivalents; n = 988), and highly-fit (12.4 ± 2.2 metabolic equivalents; n = 954). Annual costs per subject were quantified over an 8-year period. RESULTS: Age, BMI, and presence of cardiovascular disease (CVD) were similar between subjects with and without diabetes. After adjusting for age and presence of CVD, annual costs per person were higher among diabetics vs. non-diabetics. Individuals with and without diabetes in the highly-fit category had annual costs (US dollars x 103) (mean ± standard deviation) that were on average $32,178 and $30,816 lower, respectively, than individuals in the least-fit category. For each 1-metabolic equivalent higher fitness, annual cost savings per person were $5,193 and $3,603 for individuals with and without diabetes, respectively. CONCLUSIONS: Higher fitness is associated with lower health care costs. Cost savings associated with higher fitness are particularly evident among individuals with diabetes. The economic burden of diabetes may be reduced through interventions that target improvements in fitness. Published by Elsevier Inc. • The American Journal of Medicine (2019) xxx:xxx-xxx KEYWORDS: Cardiorespiratory fitness; Cardiovascular disease; Diabetes; Exercise testing; Healthcare costs
INTRODUCTION Diabetes is a chronic health condition that imposes a significant societal burden through higher disability, higher mortality, higher incidence of cardiovascular disease and other comorbid conditions, and higher health care costs.1 The incidence Funding: None. Conflict of Interest: None. Authorship: All authors had access to the data and a role in writing this manuscript. Requests for reprints should be addressed to Jonathan Myers, PhD, VA Palo Alto Health Care System, Cardiology 111C, 3801 Miranda Ave, Palo Alto, CA 94304. E-mail address: [email protected] 0002-9343/Published by Elsevier Inc. https://doi.org/10.1016/j.amjmed.2019.04.006
of diabetes has grown markedly in the United States over the past 3 decades. Between 2012 and 2015, the incidence of diabetes grew by 700,000 annually in the United States, with the incidence projected to continue increasing over time as the population ages.2,3 Similar increases in the prevalence of diabetes have recently occurred throughout the Western world, and even more dramatic increases have occurred in developing countries.4,5 The growth in prevalence of diabetes has been paralleled by a progressive economic burden. In 2017, the cost of diabetes in the United States was estimated to be $327 billion, which includes $237 billion in direct medical cost and $90 billion in lost productivity.1 After adjusting for age and
ARTICLE IN PRESS
The American Journal of Medicine, Vol xxx, No xxx, ▪▪ 2019
2
gender, annual per capita health care expenditures are 2.3 times clearance to participate in exercise, or assessment of higher for people with diabetes compared with those without suspected coronary artery disease. The following subjects diabetes.1 Therefore, strategies to reduce the economic impact were excluded1: those unable to complete the test for orthoof these trends are needed. pedic, neurologic, or similar reasons2; those with an imIn recent years, a growing number of studies has reported planted pacemaker3; those who were unstable or required that cardiorespiratory fitness is a powerful risk factor for moremergent intervention; and 4) those with an exercise capacity tality and numerous health condiless than 2 metabolic equivalents. In tions, including coronary artery addition, 5 subjects in the study samCLINICAL SIGNIFICANCE disease, heart failure, kidney disple were missing data on age and ease, diabetes, cognitive dysfuncwere excluded. Presence of diabetes Among nearly 4000 subjects, a strong tion, and neurological disorders.6–9 was determined by history and physiinverse gradient was observed becal exam at the time of the exercise In many recent epidemiologic studtween levels of cardiorespiratory fittest; because the vast majority of subies, fitness was reported to be a ness from a maximal treadmill test jects were type 2 diabetics and cost stronger risk marker for mortality and health care costs, evaluated over data were not available between type and incidence of cardiovascular disan 8-year period. 1 and type 2 diabetics, they were ease than traditional risk factors This gradient was particularly evident combined. such as smoking, hypertension, among subjects with diabetes; for each and lipid disorders.6,9,10 In response to these studies, a recent American 1-metabolic equivalent higher level of Exercise testing Heart Association scientific statecardiorespiratory fitness, annual cost savA thorough clinical history, list of ment called for fitness to be considings per person (US dollars) were $5,193 medications, and cardiac risk factors ered a risk factor of equal and $3,603 for individuals with and withwere recorded prospectively at the importance to these traditional risk out diabetes, respectively. time of testing.24 Patients underwent factors.9 The growing recognition symptom-limited treadmill testing of the importance of fitness has prousing an individualized ramp treadmill protocol as previvided a foundation to encourage health care providers to proously described.25 Heart rate targets were not used as an mote physical activity, a behavior that increases fitness.9,11,12 end point. Estimated metabolic equivalents were calculated One of the key adaptations in response to an exercise profrom treadmill speed and grade.26 Quartiles of exercise capacgram is an improvement in insulin sensitivity, which reduces 13 ity were formed as described previously.27 Briefly, to minithe burden of diabetes. Both increased physical activity mize the impact of age on fitness, we first stratified the patterns and higher fitness are associated with lower mortalcohort into 4 age categories: b 50, 50-59, 60-69, and ≥ 70 ity in diabetic individuals.14–19 This reduction in mortality years. We then defined quartiles of fitness within each age catoccurs in proportion to the level of fitness and is independent egory according to metabolic equivalents achieved during the of body mass index.14–19 exercise test. Finally, we combined the respective quartiles Given the economic burden of both diabetes1 and low from all age categories to form the following fitness categofitness,20–23 it is of interest to determine the influence of fitries for the entire cohort: the least-fit category (participants ness on health care costs in patients with diabetes. However, with a peak metabolic equivalent level in the lowest 25th to our knowledge, the impact that fitness has on the relationpercentile [mean 5.1 ± 1.5; range, 2-8.1; n = 1,044]); the ship between diabetes and health care costs has not been exmoderate-fit category (those with a peak metabolic equivaplored. The knowledge that higher fitness may favorably lent level between the 26th and 50th percentile [mean 7.6 impact health care costs could provide an additional impetus ± 1.5; range, 4.7-10.3; n = 938]); the fit category (those to support regular physical activity in diabetics. In turn, this with a peak metabolic equivalent level between the 51st could have a significant impact on ameliorating the enorand 75th percentile [mean 9.4 ± 1.5; range, 6.1-12.4; n = mous burden of health care expenditures in these patients.1 988]), and the highly fit category (those with a peak metaThe current study was performed to evaluate the association bolic equivalent level in the N 75th percentile [mean 12.4 between fitness and overall health costs in patients with ± 2.2; range, 8.0-20.0; n = 954]). Age-predicted exercise cadiabetes. pacity as a continuous variable was determined using an equation derived from VA subjects.28 For perspective, an 8-minute mile would equate to approximately 12.5 metabolic equivaMETHODS lents, a 10-minute mile would equate to approximately 10 Study Sample metabolic equivalents, a 15-minute mile would equate to approximately 4.9 metabolic equivalents, and a 20-minute mile The population included 3924 consecutive patients who were would equate to approximately 3.3 metabolic equivalents. referred for an exercise treadmill test for clinical reasons at Clinical and exercise test data were entered into a unique colthe Veterans Affairs (VA) Palo Alto Health Care System, lection and reporting program that automatically generates a Calif, between January 2005 and December 2012. The mareport for distribution within the VA clinical database.24 jority of tests were performed as part of a routine evaluation,
. .
ARTICLE IN PRESS Myers et al
Fitness and Health Care Costs in Diabetes
Calculation of costs Calculations of costs have been described previously in detail.20 Briefly, the VA employs a set of relational databases that provide data on costs, patterns of care, patient outcomes, and workload details of specific patient care encounters within the VA Health Care System. These include modules that record data from various departments, information from the abstract of the hospital discharge, and records of outpatient visits, including codes for the type of clinic visit, procedures, and diagnoses. The list of patients who had undergone treadmill tests during the study period was submitted to the Austin VA Automation Center, and the output generated by the center was merged with our treadmill database. Total costs for each patient were estimated during the 8year time period of data collection following the treadmill test based on a time-weighted average of the costs during the period in which a given patient was followed.
Statistical analysis Patient characteristics are presented as the percentage of the total or mean ± standard deviation. Cost data are expressed in absolute values in US dollars as totals for the 8-year observation period, and average cost per patient per year. Comparisons of costs between quartiles of fitness were performed using linear regression models. For our primary analyses, we fit 3 different regression models in order to show minimally and maximally adjusted models to examine whether the results were consistent. Model 1 adjusted only for age and age-squared, so this can be considered to be unadjusted for hypothesized confounding variables. Model 2 then adjusted for the factors we a-priori believed may be confounding the association between exercise capacity and costs.
3 Finally, model 3 additionally controlled for prevalent cardiovascular disease. Sensitivity analyses (not shown) demonstrated that results were not sensitive to excluding different proportions of high cost patients (top 10%, top 5%, top 1%); thus, the models were fit on the full population (although Figures 1 and 2 exclude the top 1% for clarity of presentation).
RESULTS Subject Characteristics Clinical and demographic characteristics of the study sample are shown in Table 1, according to age-stratified quartiles of metabolic equivalents achieved. The mean age of the sample was somewhat lower with higher fitness levels, ranging from 59 ± 11 years in the least fit group to 57 ± 12 years in the most fit group. While there was a high prevalence of cardiovascular risk factors across all levels of fitness, the fittest group had a lower proportion of subjects with a history of cardiovascular disease, hypertension, and smoking. Figure 1 shows the association between annual costs per patient per year and percentage age-predicted exercise capacity among diabetics and non-diabetics. A steep decline in costs was observed with higher exercise capacity across the distribution of fitness; notably, the decline was steeper among diabetics. Table 2 shows age- and cardiovascular disease-adjusted health care costs per patient per year stratified by quartile of fitness and presence or absence of diabetes. These relationships are shown graphically in Figure 2; overall costs were higher among diabetics regardless of fitness category, and a gradient for a reduction in costs was observed as fitness level was higher among both diabetics and non-
Figure 1 Linear regression line and 95% confidence intervals for annual costs per person vs age-predicted exercise capacity among subjects with and without diabetes.
ARTICLE IN PRESS
The American Journal of Medicine, Vol xxx, No xxx, ▪▪ 2019
4
Figure 2 Average annual health care costs per person according to presence and absence of diabetes and quartile of cardiorespiratory fitness. There was a gradient for a reduction in costs as fitness was higher in each group (P b .001).
Table 1 Characteristics of the Sample by Categories of Age-Stratified Exercise Capacity (Mean and [SD] or %) Total (n = 3,924)
Least Fit (n = 1,044)
Moderately Fit (n = 938)
Fit (n = 988)
Highly Fit (n = 954)
P Value
58 (11) 29 (5) 39 23 63 49 17
59 (11) 30 (6) 44 32 68 56 22
58 (11) 30 (5) 41 24 64 49 20
58 (11) 29 (5) 39 20 61 49 15
57 (12) 28 (4) 28 16 57 42 10
.01 b .01 b .01 b .01 b .01 b .01 b .01
22 24 26
26 32 34
23 25 27
19 21 23
19 18 19
b .01 b .01 b .01
Resting exercise test responses Heart rate (beats/min) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg)
75 (13) 131 (19) 81 (11)
76 (14) 134 (20) 80 (11)
75 (13) 132 (19) 81 (11)
74 (13) 131 (18) 81 (11)
74 (12) 129 (17) 81 (11)
b .01 b .01 b .01
Maximal exercise test responses Heart rate (beats/min) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Exercise capacity (METs) Unadjusted average annual cost per person (USD, 103)
141 (23) 177 (26) 84 (13) 9 (3) 39 (86)
128 (22) 172 (28) 84 (14) 5 (2) 58 (124)
137 (19) 178 (25) 85 (13) 8 (2) 39 (81)
146 (20) 178 (24) 84 (13) 9 (2) 33 (62)
154 (19) 181 (24) 84 (14) 12 (2) 25 (52)
b .01 b .01 .29 b .01 b .01
History and Demographics
Age Body mass index (kg/m2) Obesity (body mass index 30 kg/m2, %) CVD (%) Diabetes (present) (%) Hypertension (present) (%) Currently smoking (%) Medications Statins (%) Beta blockers (%) ACE (I) (%)
ACI-I = angiotensin converting enzyme inhibitor; CVD = cardiovascular disease; MET = metabolic equivalent; SD = standard deviation; USD = United States dollars. Age-stratified quartiles of exercise capacity.28 Analysis of variance main effect or chi-square.
ARTICLE IN PRESS Myers et al
Fitness and Health Care Costs in Diabetes
5
Table 2 Annual Costs per Person (USD) Adjusted for Age and Cardiovascular Disease According to Presence of Diabetes and Fitness Level (Mean ˘ SD) Annual Cost per Person (USD)
Total Least fit Low-fit Fit High-fit
With Diabetes (2457)
Without Diabetes (1467)
P Value
44,965 ± 96,604 59,507 ± 127,342 46,538 ± 99,844 39,575 ± 68,721 27,329 ± 53,488
37,049 ± 90,220 57,462 ± 138,237 31,336 ± 66,682 30,546 ± 62,826 26,646 ± 56,785
b .01 .04 b .01 b .01 .01
SD = standard deviation; USD= United States dollars. Mann-Whitney test.
diabetics (P b .001). Individuals with and without diabetes in the highest fitness category had annual costs that were on average $32,178 and $30,816 lower, respectively, than individuals in the lowest fitness category. Among both diabetics and non-diabetics, the largest reduction in costs was observed between the least-fit group and the next least-fit group; costs among the fittest quartile were less than half those of the least-fit quartile. For each 1-metabolic equivalent higher fitness, annual cost savings per person were $5,193 and $3,603 for individuals with and without diabetes, respectively. Table 3 shows regression-based estimates of the association between categories of fitness and total costs per patient per year among diabetics, with the least-fit group as the reference. The results were consistent across all 3 models regardless of which potential confounders were controlled for. Among diabetic patients in the highest quartile of fitness, costs were roughly $30,000 less than for those in the leastfit quartile, regardless of adjustments for potential confounders (P b .001 for each model). Compared with the least-fit group, there were also lower costs for those between moderately fit and fit quartiles.
DISCUSSION We observed that fitness was strongly and inversely associated with total health care costs among VA subjects with and without diabetes over an 8-year period. This was
evidenced by a 57% reduction in annual costs per person between subjects in the highest-fit quartile compared with the least-fit quartile, an overall $4,704 reduction in annual costs per person for each higher metabolic equivalent achieved, and an inverse gradient between annual costs per person and age-adjusted exercise capacity (Figure 1). In addition to higher overall costs among diabetics vs non-diabetics, the reduction in costs relative to higher fitness was more pronounced among diabetics compared with non-diabetics. For each 1-metabolic equivalent higher fitness level, annual cost savings per person were $5,193 and $3,603 for individuals with and without diabetes (reflecting ≈ 11.5% and 9.7% of adjusted total costs, respectively). Moreover, in contrast to less-fit subjects, among subjects in the highestfit group, age- and cardiovascular disease-adjusted annual cost differences were minimal between diabetics and nondiabetics ($27.3 ± 53.5 vs $26.6 ± 56.8 x103, respectively). The latter finding suggests that as long as an individual is comparatively fit, the presence of diabetes has minimal impact on health care costs. Finally, unadjusted annual health care costs per person among diabetic subjects in the leastfit quartile were more than twice those for individuals in the most-fit quartile (Table 1). Even after full adjustment for relevant variables known to influence health care costs, annual costs per person in the least-fit group remained $32,178 higher than those in the fittest group, with a gradient for reduced costs as fitness was higher (Table 2, Figures 1 and 2).
Table 3 Regression-Based Association of Annual Cost per Person (USD) and Fitness Level Categories Among Individuals with Diabetes Model 1
Model 2
Fitness category
n
Estimate
95% CI
Highly-fit Fit Moderately-fit Low-fit
544 606 599 708
-33,485 -43,611 to -23,359 -25,498 -35,327 to -15,669 -18,808 -28,667 to -8949 Comparison group
Estimate
Model 3 95% CI
-29,700 -39,975 to -19,425 -23,781 -33,640 to -13,921 -17,927 -27,779 to -8075 Comparison group
Estimate
95% CI
-29,146 -39,438 to -18,854 -23,380 -33,248 to -13,513 -17,455 -27,320 to -7591 Comparison group
CI = confidence interval; USD = US dollars. Model 1 controls for age. Model 2 controls for age, hypertension, chronic heart failure, stroke family history of coronary artery disease, diabetes, smoking, and drugs. Model 3 controls for variables in models 1 and 2, and also for the presence of other cardiovascular diseases. P value b0.001.
ARTICLE IN PRESS 6 The concept that higher fitness has a marked impact on health care costs in diabetics extends the widely held view that regular exercise has numerous physiologic benefits in patients with diabetes. These benefits have been outlined by national and international guidelines on the management of diabetes (eg, American Diabetes Association29 European Association for the Study of Diabetes30), and include reductions in hemoglobin A1c, triglycerides, blood pressure, and insulin resistance. In addition to improvements in risk factors associated with cardiovascular disease and/or metabolic disease, higher fitness, exercise intervention programs, or both, have been demonstrated to improve long-term health outcomes. Prospective cohorts, including our group,14,18,19 the Health Professionals Follow-up Study,31 the Aerobics Center Longitudinal Study,15–17 and the University of Kuopio,32 have observed strong inverse associations between physical activity patterns, fitness, and mortality in patients with diabetes. Though the direct economic impact of regular exercise in diabetics is difficult to determine precisely, most of the physiologic benefits associated with regular exercise can be achieved by modest levels of activity. For example, meeting the minimal recommendations on physical activity provided by the American Diabetes Association, the European Association for the Study of Diabetes, and many other organizations (ie, 150 minutes/week of moderate activity)29,30,33,34 has been shown to improve fitness and have a favorable impact on many chronic conditions and to reduce mortality.6–12,33, 34 By extension, these adaptations would promote health benefits that would undoubtedly have a favorable impact on health care costs in patients with diabetes. Previous studies have assessed the economic implications of physical activity interventions, worksite wellness programs, and other lifestyle interventions on health care expenditures in diabetics.31,35–39 For example, the Action for Health in Diabetes (Look AHEAD) study followed 5,121 overweight or obese adults with type 2 diabetes randomly assigned to an intensive lifestyle intervention that included physical activity, or to a comparison condition of diabetes support and education.35 In addition to a marked improvement in fitness,40 they reported a mean relative per-person 10-year cost savings of $5,280 in the lifestyle intervention group. The Diabetes Prevention Program Research Group41 assessed the 10-year cost-effectiveness of lifestyle intervention vs metformin for diabetes prevention, and reported a 17% relative reduction in 10-year inpatient care costs in the lifestyle intervention group. However, little is known regarding the relation between direct measures of health care costs and objective measures of fitness and how this association is impacted by diabetes. We recently observed an inverse association between health care costs and fitness in a broad sample of VA subjects referred for exercise testing,20 and a particularly notable reduction in costs with higher fitness among overweight and obese individuals.21 The ≈ 11.5% reduction in overall costs per metabolic equivalent observed in the current study among diabetics is slightly higher than that reported from diverse samples of subjects undergoing exercise testing (5%-7%),20,22,23 but similar to that recently
The American Journal of Medicine, Vol xxx, No xxx, ▪▪ 2019 observed among obese VA subjects.21 A particularly salient observation from the current study was the fact that moving from the least-fit group to the next least-fit group yielded a cost savings of N$20,000 among both diabetics and nondiabetics (Table 2). This finding parallels the widely described concept that the largest benefits in mortality and other health outcomes occur at the low end of the fitness spectrum.6–11 Regardless of the population assessed, these results collectively suggest that considerable cost savings may be achieved by comparatively modest increases in fitness, and they provide an additional impetus for health care providers and health organizations to recommend moderate physical activity to patients with diabetes in order to improve fitness.6–12,33,34
Limitations Matching cost and treadmill data were available only for an 8-year period; a broader assessment of costs would be valuable in order to assess the many nuances regarding the relationship between fitness and health care costs, including types of costs, specific chronic conditions, and the impact of serial changes in fitness. The sample was derived from subjects referred for an exercise test, and many subjects had existing disease and comorbid conditions; therefore, the analyses may differ from those of other samples. Fitter subjects may have engaged in other healthy behaviors, such as a healthier diet, regular physician visits, or better medication adherence, but we do not have information on these behaviors. Finally, fitness is a complex attribute that is influenced by many factors in addition to physical activity patterns, and it is not possible to account for all of them.
SUMMARY Higher fitness is associated with lower health care costs, and the reduction in costs relative to fitness is more pronounced among diabetics compared with non-diabetics. In addition to its effect on insulin resistance and other health outcomes, the economic burden of diabetes may be reduced through interventions that target improvements in fitness.
References 1. American Diabetes Association. Economic costs of diabetes in the U.S. in 2017. Diabetes Care 2018;41:917-28. 2. Centers for Disease Control and Prevention. National Diabetes Statistics Report, 2017. Available at https://www.cdc.gov/diabetes/data/ statistics/statistics-report.html. Accessed December 26, 2018. 3. Boyle JP, Thompson TJ, Gregg EW, Barker LE, Williamson DF. Projection of the year 2050 burden of diabetes in the US adult population: dynamic modeling of incidence, mortality, and prediabetes prevalence. Popul Health Metr 2010;8:29. 4. Ingelfinger JR, Jarcho JA. Increase in the incidence of diabetes and its implications. N Engl J Med 2017;376:1473-4. 5. Saklayen MG. The global epidemic of the metabolic syndrome. Curr Hypertens Rep 2018;20:12. 6. Myers J, McAuley P, Lavie CJ, Despres JP, Arena R, Kokkinos P. Physical activity and cardiorespiratory fitness as major markers of
ARTICLE IN PRESS Myers et al
7.
8.
9.
10.
11. 12. 13.
14.
15.
16.
17.
18.
19.
20.
21.
22. 23.
24. 25.
Fitness and Health Care Costs in Diabetes
cardiovascular risk: their independent and interwoven importance to health status. Prog Cardiovasc Dis 2015;57:306-14. Kokkinos PF, Faselis C, Myers JN, et al. Cardiorespiratory fitness and incidence of major adverse cardiovascular events in US veterans: a cohort study. Mayo Clin Proc 2017;92:39-48. Harber MP, Kaminsky LA, Arena R, et al. Impact of cardiorespiratory fitness on all-cause and disease-specific mortality: advances since 2009. Prog Cardiovasc Dis 2017;60:11-20. Ross R, Blair S, Arena R, et al. Importance of assessing cardiorespiratory fitness in clinical practice: a case for fitness as a clinical vital sign: a scientific statement from the American Heart Association. Circulation 2016;134:e653-99. DeFina LF, Haskell WL, Willis BL, et al. Physical activity versus cardiorespiratory fitness: two (partly) distinct components of cardiovascular health? Prog Cardiovasc Dis 2015;57:324-9. Myers J. The new AHA/ACC guidelines on cardiovascular risk: when will fitness get the recognition it deserves? Mayo Clin Proc 2014;89:722-6. Lavie CJ, Archer E, Nauman J. Arrival and survival of the fittest. Am Heart J 2018;196:153-5. Sampath Kumar A, Maiya AG, Shastry BA, et al. Exercise and insulin resistance in type 2 diabetes mellitus: a systematic review and metaanalysis. Ann Phys Rehabil Med 2019;62:98-103. McAuley PA, Myers JN, Abella JP, Tan SY, Froelicher VF. Exercise capacity and body mass as predictors of mortality among male veterans with type 2 diabetes. Diabetes Care 2007;30:1539-43. Church TS, LaMonte MJ, Barlow CE, Blair SN. Cardiorespiratory fitness and body mass index as predictors of cardiovascular disease mortality among men with diabetes. Arch Intern Med 2005;165:2114-20. Church TS, Cheng YJ, Earnest CP, et al. Exercise capacity and body composition as predictors of mortality among men with diabetes. Diabetes Care 2004;27:83-8. Wei M, Gibbons LW, Kampert JB, Nichaman MZ, Blair SN. Low cardiorespiratory fitness and physical inactivity as predictors of mortality in men with type 2 diabetes. Ann Intern Med 2000;132:605-11. Kokkinos P, Myers J, Nylen E, et al. Exercise capacity and all-cause mortality in African American and Caucasian men with type 2 diabetes. Diabetes Care 2009;32:623-8. Kokkinos P, Myers J, Faselis C, Doumas M, Kheirbek R, Nylen E. BMI-mortality paradox and fitness in African American and Caucasian men with type 2 diabetes. Diabetes Care 2012;35:1021-7. Myers J, Doom R, King R, et al. Association between cardiorespiratory fitness and health care costs: the Veterans Exercise Testing Study. Mayo Clin Proc 2018;93:48-55. ) Grüne de Souza de Silva C, Kokkinos P, Doom R, et al. Association between cardiorespiratory fitness, obesity, and health care costs: the Veterans Exercise Testing Study. Int J Obesity, in press. Weiss P, Froelicher V, Myers J, Heidenreich P. Health care costs and exercise capacity. Chest 2004;126:608-13. Bachman JM, DeFina LF, Franzini L, et al. Cardiorespiratory fitness in middle age and health care costs in later life. J Amer Coll Cardiol 2015;66:1876-85. Shue P, Froelicher VF. EXTRA: an expert system for exercise test reporting. J Non-Invasive Testing 1998;II-4:21-7. Myers J, Buchanan N, Walsh D, et al. Comparison of the ramp versus standard exercise protocols. J Am Coll Cardiol 1991;17:1334-42.
7 26. Glass S, Dwyer GB. ACSM’s Metabolic Equations Handbook. Philadelphia, PA: Lippincott, Williams & Wilkins. 2007. 27. Kokkinos P, Faselis C, Myers J, Sui X, Zhang J, Blair S. Age-specific exercise capacity threshold for mortality risk assessment in male veterans. Circulation 2014;130:653-8. 28. Morris CK, Myers J, Froelicher VF, Kawaguchi T, Ueshima K, Hideg A. Nomogram based on metabolic equivalents and age for assessing aerobic capacity in men. J Am Coll Cardiol 1993;22:175-82. 29. Sheri R, Colberg SR, Sigal RJ, et al. Physical activity/exercise and diabetes: a position statement of the American Diabetes Association. Diabetes Care 2016;39:2065-79. 30. Inzucchi SE, Bergenstal RM, Buse JB, et al. Management of hyperglycemia in type 2 diabetes: 2015: a patient-centered approach: update to a position statement of the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care 2015;38:140-9. 31. Tanasescu M, Leitzmann MF, Rimm EB, Hu FB. Physical activity in relation to cardiovascular disease and total mortality among men with type 2 diabetes. Circulation 2003;107:2435-9. 32. Hu G, Pekka Jousilahti Barengo NC, Qiao Q, Lakka TA, Tuomilehto J. Physical activity, cardiovascular risk factors, and mortality among Finnish adults with diabetes. Diabetes Care 2005;28:799-805. 33. Fletcher GF, Blair SN, Blumenthal J, et al. Statement on exercise: benefits and recommendations for physical activity programs for all Americans: a statement for health professionals by the Committee on Exercise and Cardiac Rehabilitation of the Council on Clinical Cardiology, American Heart Association. Circulation 1992;86:340-4. 34. U.S. Department of Health and Human Services, Physical Activity Guidelines for Americans, US Department of Health and Human Services. 2018https://health.gov/paguidelines/second-edition/. Accessed January 5, 2019. 35. Espeland MA, Glick HA, Bertoni A, et al. Impact of an intensive lifestyle intervention on use and cost of medical services among overweight and obese adults with type 2 diabetes: the Action for Health in Diabetes. Diabetes Care 2014;37:2548-56. 36. Nguyen HQ, Ackermann RT, Berke EM, et al. Impact of a managedMedicare physical activity benefit on health care utilization and costs in older adults with diabetes. Diabetes Care 2007;30:43-8. 37. Mori DL, Silberbogen AK, Collins AE, Ulloa EW, Brown KL, Niles BL. Promoting physical activity in individuals with diabetes: telehealth approaches. Diabetes Spectrum 2011;24:127-35. 38. White ND, Lenz TL, Skrabal MZ, Skradski JJ, Lipari L. Long-term outcomes of a cardiovascular and diabetes risk-reduction program initiated by a self-insured employer. Am Health Drug Benefits 2018;11: 177-83. 39. Bunting BA, Lee G, Knowles G, Lee C, Allen P. The Hickory Project: Controlling healthcare costs and improving outcomes for diabetes using the Asheville Project model. Am Health Drug Benefits 2011;4:343-50. 40. Wing RR, Bolin P, Brancati FL, et al. Cardiovascular effects of intensive lifestyle intervention in type 2 diabetes. N Engl J Med 2013;369: 145-54. 41. Diabetes Prevention Program Research Group. The 10-year costeffectiveness of lifestyle intervention or metformin for diabetes prevention: an intent-to-treat analysis of the DPP/DPPOS. Diabetes Care 2012;35:723-30.
Cardiorespiratory Fitness and Health Care Costs in Diabetes: The Veterans Exercise Testing Study Jonathan Myers, PhD a,b,, Christina Grne de Souza e Silva, PhD c, Rachelle Doom, MD a, Holly Fonda, MS a, Khin Chan, MD a, Shirit Kamil-Rosenberg, EdD a, Peter Kokkinos, PhD d a
Veterans Affairs Palo Alto Health Care System, Calif; b Division of Cardiovascular Medicine, Stanford University, Calif; c Heart Institute Edson Saad/Federal University of Rio de Janeiro, Brazil; d Veterans Affairs Medical Center, Washington, DC.
BACKGROUND This study aimed to determine the association between cardiorespiratory fitness and healthcare expenditures among individuals with and without diabetes. METHODS: Health care costs were quantified among 3924 consecutive men (mean age 58 ± 11 years) referred for a maximal exercise test, and compared according to presence (n = 2457) and absence (n = 1467) of diabetes and fitness. Fitness was classified into 4 categories based on age-stratified quartiles of peak metabolic equivalents: least-fit (5.1 ± 1.5 metabolic equivalents; n = 1044), moderately-fit (7.6 ± 1.5 metabolic equivalents; n = 938), fit (9.4 ± 1.5 metabolic equivalents; n = 988), and highly-fit (12.4 ± 2.2 metabolic equivalents; n = 954). Annual costs per subject were quantified over an 8-year period. RESULTS: Age, BMI, and presence of cardiovascular disease (CVD) were similar between subjects with and without diabetes. After adjusting for age and presence of CVD, annual costs per person were higher among diabetics vs. non-diabetics. Individuals with and without diabetes in the highly-fit category had annual costs (US dollars x 103) (mean ± standard deviation) that were on average $32,178 and $30,816 lower, respectively, than individuals in the least-fit category. For each 1-metabolic equivalent higher fitness, annual cost savings per person were $5,193 and $3,603 for individuals with and without diabetes, respectively. CONCLUSIONS: Higher fitness is associated with lower health care costs. Cost savings associated with higher fitness are particularly evident among individuals with diabetes. The economic burden of diabetes may be reduced through interventions that target improvements in fitness. Published by Elsevier Inc. • The American Journal of Medicine (2019) xxx:xxx-xxx KEYWORDS: Cardiorespiratory fitness; Cardiovascular disease; Diabetes; Exercise testing; Healthcare costs
INTRODUCTION Diabetes is a chronic health condition that imposes a significant societal burden through higher disability, higher mortality, higher incidence of cardiovascular disease and other comorbid conditions, and higher health care costs.1 The incidence Funding: None. Conflict of Interest: None. Authorship: All authors had access to the data and a role in writing this manuscript. Requests for reprints should be addressed to Jonathan Myers, PhD, VA Palo Alto Health Care System, Cardiology 111C, 3801 Miranda Ave, Palo Alto, CA 94304. E-mail address: [email protected] 0002-9343/Published by Elsevier Inc. https://doi.org/10.1016/j.amjmed.2019.04.006
of diabetes has grown markedly in the United States over the past 3 decades. Between 2012 and 2015, the incidence of diabetes grew by 700,000 annually in the United States, with the incidence projected to continue increasing over time as the population ages.2,3 Similar increases in the prevalence of diabetes have recently occurred throughout the Western world, and even more dramatic increases have occurred in developing countries.4,5 The growth in prevalence of diabetes has been paralleled by a progressive economic burden. In 2017, the cost of diabetes in the United States was estimated to be $327 billion, which includes $237 billion in direct medical cost and $90 billion in lost productivity.1 After adjusting for age and
ARTICLE IN PRESS
The American Journal of Medicine, Vol xxx, No xxx, ▪▪ 2019
2
gender, annual per capita health care expenditures are 2.3 times clearance to participate in exercise, or assessment of higher for people with diabetes compared with those without suspected coronary artery disease. The following subjects diabetes.1 Therefore, strategies to reduce the economic impact were excluded1: those unable to complete the test for orthoof these trends are needed. pedic, neurologic, or similar reasons2; those with an imIn recent years, a growing number of studies has reported planted pacemaker3; those who were unstable or required that cardiorespiratory fitness is a powerful risk factor for moremergent intervention; and 4) those with an exercise capacity tality and numerous health condiless than 2 metabolic equivalents. In tions, including coronary artery addition, 5 subjects in the study samCLINICAL SIGNIFICANCE disease, heart failure, kidney disple were missing data on age and ease, diabetes, cognitive dysfuncwere excluded. Presence of diabetes Among nearly 4000 subjects, a strong tion, and neurological disorders.6–9 was determined by history and physiinverse gradient was observed becal exam at the time of the exercise In many recent epidemiologic studtween levels of cardiorespiratory fittest; because the vast majority of subies, fitness was reported to be a ness from a maximal treadmill test jects were type 2 diabetics and cost stronger risk marker for mortality and health care costs, evaluated over data were not available between type and incidence of cardiovascular disan 8-year period. 1 and type 2 diabetics, they were ease than traditional risk factors This gradient was particularly evident combined. such as smoking, hypertension, among subjects with diabetes; for each and lipid disorders.6,9,10 In response to these studies, a recent American 1-metabolic equivalent higher level of Exercise testing Heart Association scientific statecardiorespiratory fitness, annual cost savA thorough clinical history, list of ment called for fitness to be considings per person (US dollars) were $5,193 medications, and cardiac risk factors ered a risk factor of equal and $3,603 for individuals with and withwere recorded prospectively at the importance to these traditional risk out diabetes, respectively. time of testing.24 Patients underwent factors.9 The growing recognition symptom-limited treadmill testing of the importance of fitness has prousing an individualized ramp treadmill protocol as previvided a foundation to encourage health care providers to proously described.25 Heart rate targets were not used as an mote physical activity, a behavior that increases fitness.9,11,12 end point. Estimated metabolic equivalents were calculated One of the key adaptations in response to an exercise profrom treadmill speed and grade.26 Quartiles of exercise capacgram is an improvement in insulin sensitivity, which reduces 13 ity were formed as described previously.27 Briefly, to minithe burden of diabetes. Both increased physical activity mize the impact of age on fitness, we first stratified the patterns and higher fitness are associated with lower mortalcohort into 4 age categories: b 50, 50-59, 60-69, and ≥ 70 ity in diabetic individuals.14–19 This reduction in mortality years. We then defined quartiles of fitness within each age catoccurs in proportion to the level of fitness and is independent egory according to metabolic equivalents achieved during the of body mass index.14–19 exercise test. Finally, we combined the respective quartiles Given the economic burden of both diabetes1 and low from all age categories to form the following fitness categofitness,20–23 it is of interest to determine the influence of fitries for the entire cohort: the least-fit category (participants ness on health care costs in patients with diabetes. However, with a peak metabolic equivalent level in the lowest 25th to our knowledge, the impact that fitness has on the relationpercentile [mean 5.1 ± 1.5; range, 2-8.1; n = 1,044]); the ship between diabetes and health care costs has not been exmoderate-fit category (those with a peak metabolic equivaplored. The knowledge that higher fitness may favorably lent level between the 26th and 50th percentile [mean 7.6 impact health care costs could provide an additional impetus ± 1.5; range, 4.7-10.3; n = 938]); the fit category (those to support regular physical activity in diabetics. In turn, this with a peak metabolic equivalent level between the 51st could have a significant impact on ameliorating the enorand 75th percentile [mean 9.4 ± 1.5; range, 6.1-12.4; n = mous burden of health care expenditures in these patients.1 988]), and the highly fit category (those with a peak metaThe current study was performed to evaluate the association bolic equivalent level in the N 75th percentile [mean 12.4 between fitness and overall health costs in patients with ± 2.2; range, 8.0-20.0; n = 954]). Age-predicted exercise cadiabetes. pacity as a continuous variable was determined using an equation derived from VA subjects.28 For perspective, an 8-minute mile would equate to approximately 12.5 metabolic equivaMETHODS lents, a 10-minute mile would equate to approximately 10 Study Sample metabolic equivalents, a 15-minute mile would equate to approximately 4.9 metabolic equivalents, and a 20-minute mile The population included 3924 consecutive patients who were would equate to approximately 3.3 metabolic equivalents. referred for an exercise treadmill test for clinical reasons at Clinical and exercise test data were entered into a unique colthe Veterans Affairs (VA) Palo Alto Health Care System, lection and reporting program that automatically generates a Calif, between January 2005 and December 2012. The mareport for distribution within the VA clinical database.24 jority of tests were performed as part of a routine evaluation,
. .
ARTICLE IN PRESS Myers et al
Fitness and Health Care Costs in Diabetes
Calculation of costs Calculations of costs have been described previously in detail.20 Briefly, the VA employs a set of relational databases that provide data on costs, patterns of care, patient outcomes, and workload details of specific patient care encounters within the VA Health Care System. These include modules that record data from various departments, information from the abstract of the hospital discharge, and records of outpatient visits, including codes for the type of clinic visit, procedures, and diagnoses. The list of patients who had undergone treadmill tests during the study period was submitted to the Austin VA Automation Center, and the output generated by the center was merged with our treadmill database. Total costs for each patient were estimated during the 8year time period of data collection following the treadmill test based on a time-weighted average of the costs during the period in which a given patient was followed.
Statistical analysis Patient characteristics are presented as the percentage of the total or mean ± standard deviation. Cost data are expressed in absolute values in US dollars as totals for the 8-year observation period, and average cost per patient per year. Comparisons of costs between quartiles of fitness were performed using linear regression models. For our primary analyses, we fit 3 different regression models in order to show minimally and maximally adjusted models to examine whether the results were consistent. Model 1 adjusted only for age and age-squared, so this can be considered to be unadjusted for hypothesized confounding variables. Model 2 then adjusted for the factors we a-priori believed may be confounding the association between exercise capacity and costs.
3 Finally, model 3 additionally controlled for prevalent cardiovascular disease. Sensitivity analyses (not shown) demonstrated that results were not sensitive to excluding different proportions of high cost patients (top 10%, top 5%, top 1%); thus, the models were fit on the full population (although Figures 1 and 2 exclude the top 1% for clarity of presentation).
RESULTS Subject Characteristics Clinical and demographic characteristics of the study sample are shown in Table 1, according to age-stratified quartiles of metabolic equivalents achieved. The mean age of the sample was somewhat lower with higher fitness levels, ranging from 59 ± 11 years in the least fit group to 57 ± 12 years in the most fit group. While there was a high prevalence of cardiovascular risk factors across all levels of fitness, the fittest group had a lower proportion of subjects with a history of cardiovascular disease, hypertension, and smoking. Figure 1 shows the association between annual costs per patient per year and percentage age-predicted exercise capacity among diabetics and non-diabetics. A steep decline in costs was observed with higher exercise capacity across the distribution of fitness; notably, the decline was steeper among diabetics. Table 2 shows age- and cardiovascular disease-adjusted health care costs per patient per year stratified by quartile of fitness and presence or absence of diabetes. These relationships are shown graphically in Figure 2; overall costs were higher among diabetics regardless of fitness category, and a gradient for a reduction in costs was observed as fitness level was higher among both diabetics and non-
Figure 1 Linear regression line and 95% confidence intervals for annual costs per person vs age-predicted exercise capacity among subjects with and without diabetes.
ARTICLE IN PRESS
The American Journal of Medicine, Vol xxx, No xxx, ▪▪ 2019
4
Figure 2 Average annual health care costs per person according to presence and absence of diabetes and quartile of cardiorespiratory fitness. There was a gradient for a reduction in costs as fitness was higher in each group (P b .001).
Table 1 Characteristics of the Sample by Categories of Age-Stratified Exercise Capacity (Mean and [SD] or %) Total (n = 3,924)
Least Fit (n = 1,044)
Moderately Fit (n = 938)
Fit (n = 988)
Highly Fit (n = 954)
P Value
58 (11) 29 (5) 39 23 63 49 17
59 (11) 30 (6) 44 32 68 56 22
58 (11) 30 (5) 41 24 64 49 20
58 (11) 29 (5) 39 20 61 49 15
57 (12) 28 (4) 28 16 57 42 10
.01 b .01 b .01 b .01 b .01 b .01 b .01
22 24 26
26 32 34
23 25 27
19 21 23
19 18 19
b .01 b .01 b .01
Resting exercise test responses Heart rate (beats/min) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg)
75 (13) 131 (19) 81 (11)
76 (14) 134 (20) 80 (11)
75 (13) 132 (19) 81 (11)
74 (13) 131 (18) 81 (11)
74 (12) 129 (17) 81 (11)
b .01 b .01 b .01
Maximal exercise test responses Heart rate (beats/min) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Exercise capacity (METs) Unadjusted average annual cost per person (USD, 103)
141 (23) 177 (26) 84 (13) 9 (3) 39 (86)
128 (22) 172 (28) 84 (14) 5 (2) 58 (124)
137 (19) 178 (25) 85 (13) 8 (2) 39 (81)
146 (20) 178 (24) 84 (13) 9 (2) 33 (62)
154 (19) 181 (24) 84 (14) 12 (2) 25 (52)
b .01 b .01 .29 b .01 b .01
History and Demographics
Age Body mass index (kg/m2) Obesity (body mass index 30 kg/m2, %) CVD (%) Diabetes (present) (%) Hypertension (present) (%) Currently smoking (%) Medications Statins (%) Beta blockers (%) ACE (I) (%)
ACI-I = angiotensin converting enzyme inhibitor; CVD = cardiovascular disease; MET = metabolic equivalent; SD = standard deviation; USD = United States dollars. Age-stratified quartiles of exercise capacity.28 Analysis of variance main effect or chi-square.
ARTICLE IN PRESS Myers et al
Fitness and Health Care Costs in Diabetes
5
Table 2 Annual Costs per Person (USD) Adjusted for Age and Cardiovascular Disease According to Presence of Diabetes and Fitness Level (Mean ˘ SD) Annual Cost per Person (USD)
Total Least fit Low-fit Fit High-fit
With Diabetes (2457)
Without Diabetes (1467)
P Value
44,965 ± 96,604 59,507 ± 127,342 46,538 ± 99,844 39,575 ± 68,721 27,329 ± 53,488
37,049 ± 90,220 57,462 ± 138,237 31,336 ± 66,682 30,546 ± 62,826 26,646 ± 56,785
b .01 .04 b .01 b .01 .01
SD = standard deviation; USD= United States dollars. Mann-Whitney test.
diabetics (P b .001). Individuals with and without diabetes in the highest fitness category had annual costs that were on average $32,178 and $30,816 lower, respectively, than individuals in the lowest fitness category. Among both diabetics and non-diabetics, the largest reduction in costs was observed between the least-fit group and the next least-fit group; costs among the fittest quartile were less than half those of the least-fit quartile. For each 1-metabolic equivalent higher fitness, annual cost savings per person were $5,193 and $3,603 for individuals with and without diabetes, respectively. Table 3 shows regression-based estimates of the association between categories of fitness and total costs per patient per year among diabetics, with the least-fit group as the reference. The results were consistent across all 3 models regardless of which potential confounders were controlled for. Among diabetic patients in the highest quartile of fitness, costs were roughly $30,000 less than for those in the leastfit quartile, regardless of adjustments for potential confounders (P b .001 for each model). Compared with the least-fit group, there were also lower costs for those between moderately fit and fit quartiles.
DISCUSSION We observed that fitness was strongly and inversely associated with total health care costs among VA subjects with and without diabetes over an 8-year period. This was
evidenced by a 57% reduction in annual costs per person between subjects in the highest-fit quartile compared with the least-fit quartile, an overall $4,704 reduction in annual costs per person for each higher metabolic equivalent achieved, and an inverse gradient between annual costs per person and age-adjusted exercise capacity (Figure 1). In addition to higher overall costs among diabetics vs non-diabetics, the reduction in costs relative to higher fitness was more pronounced among diabetics compared with non-diabetics. For each 1-metabolic equivalent higher fitness level, annual cost savings per person were $5,193 and $3,603 for individuals with and without diabetes (reflecting ≈ 11.5% and 9.7% of adjusted total costs, respectively). Moreover, in contrast to less-fit subjects, among subjects in the highestfit group, age- and cardiovascular disease-adjusted annual cost differences were minimal between diabetics and nondiabetics ($27.3 ± 53.5 vs $26.6 ± 56.8 x103, respectively). The latter finding suggests that as long as an individual is comparatively fit, the presence of diabetes has minimal impact on health care costs. Finally, unadjusted annual health care costs per person among diabetic subjects in the leastfit quartile were more than twice those for individuals in the most-fit quartile (Table 1). Even after full adjustment for relevant variables known to influence health care costs, annual costs per person in the least-fit group remained $32,178 higher than those in the fittest group, with a gradient for reduced costs as fitness was higher (Table 2, Figures 1 and 2).
Table 3 Regression-Based Association of Annual Cost per Person (USD) and Fitness Level Categories Among Individuals with Diabetes Model 1
Model 2
Fitness category
n
Estimate
95% CI
Highly-fit Fit Moderately-fit Low-fit
544 606 599 708
-33,485 -43,611 to -23,359 -25,498 -35,327 to -15,669 -18,808 -28,667 to -8949 Comparison group
Estimate
Model 3 95% CI
-29,700 -39,975 to -19,425 -23,781 -33,640 to -13,921 -17,927 -27,779 to -8075 Comparison group
Estimate
95% CI
-29,146 -39,438 to -18,854 -23,380 -33,248 to -13,513 -17,455 -27,320 to -7591 Comparison group
CI = confidence interval; USD = US dollars. Model 1 controls for age. Model 2 controls for age, hypertension, chronic heart failure, stroke family history of coronary artery disease, diabetes, smoking, and drugs. Model 3 controls for variables in models 1 and 2, and also for the presence of other cardiovascular diseases. P value b0.001.
ARTICLE IN PRESS 6 The concept that higher fitness has a marked impact on health care costs in diabetics extends the widely held view that regular exercise has numerous physiologic benefits in patients with diabetes. These benefits have been outlined by national and international guidelines on the management of diabetes (eg, American Diabetes Association29 European Association for the Study of Diabetes30), and include reductions in hemoglobin A1c, triglycerides, blood pressure, and insulin resistance. In addition to improvements in risk factors associated with cardiovascular disease and/or metabolic disease, higher fitness, exercise intervention programs, or both, have been demonstrated to improve long-term health outcomes. Prospective cohorts, including our group,14,18,19 the Health Professionals Follow-up Study,31 the Aerobics Center Longitudinal Study,15–17 and the University of Kuopio,32 have observed strong inverse associations between physical activity patterns, fitness, and mortality in patients with diabetes. Though the direct economic impact of regular exercise in diabetics is difficult to determine precisely, most of the physiologic benefits associated with regular exercise can be achieved by modest levels of activity. For example, meeting the minimal recommendations on physical activity provided by the American Diabetes Association, the European Association for the Study of Diabetes, and many other organizations (ie, 150 minutes/week of moderate activity)29,30,33,34 has been shown to improve fitness and have a favorable impact on many chronic conditions and to reduce mortality.6–12,33, 34 By extension, these adaptations would promote health benefits that would undoubtedly have a favorable impact on health care costs in patients with diabetes. Previous studies have assessed the economic implications of physical activity interventions, worksite wellness programs, and other lifestyle interventions on health care expenditures in diabetics.31,35–39 For example, the Action for Health in Diabetes (Look AHEAD) study followed 5,121 overweight or obese adults with type 2 diabetes randomly assigned to an intensive lifestyle intervention that included physical activity, or to a comparison condition of diabetes support and education.35 In addition to a marked improvement in fitness,40 they reported a mean relative per-person 10-year cost savings of $5,280 in the lifestyle intervention group. The Diabetes Prevention Program Research Group41 assessed the 10-year cost-effectiveness of lifestyle intervention vs metformin for diabetes prevention, and reported a 17% relative reduction in 10-year inpatient care costs in the lifestyle intervention group. However, little is known regarding the relation between direct measures of health care costs and objective measures of fitness and how this association is impacted by diabetes. We recently observed an inverse association between health care costs and fitness in a broad sample of VA subjects referred for exercise testing,20 and a particularly notable reduction in costs with higher fitness among overweight and obese individuals.21 The ≈ 11.5% reduction in overall costs per metabolic equivalent observed in the current study among diabetics is slightly higher than that reported from diverse samples of subjects undergoing exercise testing (5%-7%),20,22,23 but similar to that recently
The American Journal of Medicine, Vol xxx, No xxx, ▪▪ 2019 observed among obese VA subjects.21 A particularly salient observation from the current study was the fact that moving from the least-fit group to the next least-fit group yielded a cost savings of N$20,000 among both diabetics and nondiabetics (Table 2). This finding parallels the widely described concept that the largest benefits in mortality and other health outcomes occur at the low end of the fitness spectrum.6–11 Regardless of the population assessed, these results collectively suggest that considerable cost savings may be achieved by comparatively modest increases in fitness, and they provide an additional impetus for health care providers and health organizations to recommend moderate physical activity to patients with diabetes in order to improve fitness.6–12,33,34
Limitations Matching cost and treadmill data were available only for an 8-year period; a broader assessment of costs would be valuable in order to assess the many nuances regarding the relationship between fitness and health care costs, including types of costs, specific chronic conditions, and the impact of serial changes in fitness. The sample was derived from subjects referred for an exercise test, and many subjects had existing disease and comorbid conditions; therefore, the analyses may differ from those of other samples. Fitter subjects may have engaged in other healthy behaviors, such as a healthier diet, regular physician visits, or better medication adherence, but we do not have information on these behaviors. Finally, fitness is a complex attribute that is influenced by many factors in addition to physical activity patterns, and it is not possible to account for all of them.
SUMMARY Higher fitness is associated with lower health care costs, and the reduction in costs relative to fitness is more pronounced among diabetics compared with non-diabetics. In addition to its effect on insulin resistance and other health outcomes, the economic burden of diabetes may be reduced through interventions that target improvements in fitness.
References 1. American Diabetes Association. Economic costs of diabetes in the U.S. in 2017. Diabetes Care 2018;41:917-28. 2. Centers for Disease Control and Prevention. National Diabetes Statistics Report, 2017. Available at https://www.cdc.gov/diabetes/data/ statistics/statistics-report.html. Accessed December 26, 2018. 3. Boyle JP, Thompson TJ, Gregg EW, Barker LE, Williamson DF. Projection of the year 2050 burden of diabetes in the US adult population: dynamic modeling of incidence, mortality, and prediabetes prevalence. Popul Health Metr 2010;8:29. 4. Ingelfinger JR, Jarcho JA. Increase in the incidence of diabetes and its implications. N Engl J Med 2017;376:1473-4. 5. Saklayen MG. The global epidemic of the metabolic syndrome. Curr Hypertens Rep 2018;20:12. 6. Myers J, McAuley P, Lavie CJ, Despres JP, Arena R, Kokkinos P. Physical activity and cardiorespiratory fitness as major markers of
ARTICLE IN PRESS Myers et al
7.
8.
9.
10.
11. 12. 13.
14.
15.
16.
17.
18.
19.
20.
21.
22. 23.
24. 25.
Fitness and Health Care Costs in Diabetes
cardiovascular risk: their independent and interwoven importance to health status. Prog Cardiovasc Dis 2015;57:306-14. Kokkinos PF, Faselis C, Myers JN, et al. Cardiorespiratory fitness and incidence of major adverse cardiovascular events in US veterans: a cohort study. Mayo Clin Proc 2017;92:39-48. Harber MP, Kaminsky LA, Arena R, et al. Impact of cardiorespiratory fitness on all-cause and disease-specific mortality: advances since 2009. Prog Cardiovasc Dis 2017;60:11-20. Ross R, Blair S, Arena R, et al. Importance of assessing cardiorespiratory fitness in clinical practice: a case for fitness as a clinical vital sign: a scientific statement from the American Heart Association. Circulation 2016;134:e653-99. DeFina LF, Haskell WL, Willis BL, et al. Physical activity versus cardiorespiratory fitness: two (partly) distinct components of cardiovascular health? Prog Cardiovasc Dis 2015;57:324-9. Myers J. The new AHA/ACC guidelines on cardiovascular risk: when will fitness get the recognition it deserves? Mayo Clin Proc 2014;89:722-6. Lavie CJ, Archer E, Nauman J. Arrival and survival of the fittest. Am Heart J 2018;196:153-5. Sampath Kumar A, Maiya AG, Shastry BA, et al. Exercise and insulin resistance in type 2 diabetes mellitus: a systematic review and metaanalysis. Ann Phys Rehabil Med 2019;62:98-103. McAuley PA, Myers JN, Abella JP, Tan SY, Froelicher VF. Exercise capacity and body mass as predictors of mortality among male veterans with type 2 diabetes. Diabetes Care 2007;30:1539-43. Church TS, LaMonte MJ, Barlow CE, Blair SN. Cardiorespiratory fitness and body mass index as predictors of cardiovascular disease mortality among men with diabetes. Arch Intern Med 2005;165:2114-20. Church TS, Cheng YJ, Earnest CP, et al. Exercise capacity and body composition as predictors of mortality among men with diabetes. Diabetes Care 2004;27:83-8. Wei M, Gibbons LW, Kampert JB, Nichaman MZ, Blair SN. Low cardiorespiratory fitness and physical inactivity as predictors of mortality in men with type 2 diabetes. Ann Intern Med 2000;132:605-11. Kokkinos P, Myers J, Nylen E, et al. Exercise capacity and all-cause mortality in African American and Caucasian men with type 2 diabetes. Diabetes Care 2009;32:623-8. Kokkinos P, Myers J, Faselis C, Doumas M, Kheirbek R, Nylen E. BMI-mortality paradox and fitness in African American and Caucasian men with type 2 diabetes. Diabetes Care 2012;35:1021-7. Myers J, Doom R, King R, et al. Association between cardiorespiratory fitness and health care costs: the Veterans Exercise Testing Study. Mayo Clin Proc 2018;93:48-55. ) Grüne de Souza de Silva C, Kokkinos P, Doom R, et al. Association between cardiorespiratory fitness, obesity, and health care costs: the Veterans Exercise Testing Study. Int J Obesity, in press. Weiss P, Froelicher V, Myers J, Heidenreich P. Health care costs and exercise capacity. Chest 2004;126:608-13. Bachman JM, DeFina LF, Franzini L, et al. Cardiorespiratory fitness in middle age and health care costs in later life. J Amer Coll Cardiol 2015;66:1876-85. Shue P, Froelicher VF. EXTRA: an expert system for exercise test reporting. J Non-Invasive Testing 1998;II-4:21-7. Myers J, Buchanan N, Walsh D, et al. Comparison of the ramp versus standard exercise protocols. J Am Coll Cardiol 1991;17:1334-42.
7 26. Glass S, Dwyer GB. ACSM’s Metabolic Equations Handbook. Philadelphia, PA: Lippincott, Williams & Wilkins. 2007. 27. Kokkinos P, Faselis C, Myers J, Sui X, Zhang J, Blair S. Age-specific exercise capacity threshold for mortality risk assessment in male veterans. Circulation 2014;130:653-8. 28. Morris CK, Myers J, Froelicher VF, Kawaguchi T, Ueshima K, Hideg A. Nomogram based on metabolic equivalents and age for assessing aerobic capacity in men. J Am Coll Cardiol 1993;22:175-82. 29. Sheri R, Colberg SR, Sigal RJ, et al. Physical activity/exercise and diabetes: a position statement of the American Diabetes Association. Diabetes Care 2016;39:2065-79. 30. Inzucchi SE, Bergenstal RM, Buse JB, et al. Management of hyperglycemia in type 2 diabetes: 2015: a patient-centered approach: update to a position statement of the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care 2015;38:140-9. 31. Tanasescu M, Leitzmann MF, Rimm EB, Hu FB. Physical activity in relation to cardiovascular disease and total mortality among men with type 2 diabetes. Circulation 2003;107:2435-9. 32. Hu G, Pekka Jousilahti Barengo NC, Qiao Q, Lakka TA, Tuomilehto J. Physical activity, cardiovascular risk factors, and mortality among Finnish adults with diabetes. Diabetes Care 2005;28:799-805. 33. Fletcher GF, Blair SN, Blumenthal J, et al. Statement on exercise: benefits and recommendations for physical activity programs for all Americans: a statement for health professionals by the Committee on Exercise and Cardiac Rehabilitation of the Council on Clinical Cardiology, American Heart Association. Circulation 1992;86:340-4. 34. U.S. Department of Health and Human Services, Physical Activity Guidelines for Americans, US Department of Health and Human Services. 2018https://health.gov/paguidelines/second-edition/. Accessed January 5, 2019. 35. Espeland MA, Glick HA, Bertoni A, et al. Impact of an intensive lifestyle intervention on use and cost of medical services among overweight and obese adults with type 2 diabetes: the Action for Health in Diabetes. Diabetes Care 2014;37:2548-56. 36. Nguyen HQ, Ackermann RT, Berke EM, et al. Impact of a managedMedicare physical activity benefit on health care utilization and costs in older adults with diabetes. Diabetes Care 2007;30:43-8. 37. Mori DL, Silberbogen AK, Collins AE, Ulloa EW, Brown KL, Niles BL. Promoting physical activity in individuals with diabetes: telehealth approaches. Diabetes Spectrum 2011;24:127-35. 38. White ND, Lenz TL, Skrabal MZ, Skradski JJ, Lipari L. Long-term outcomes of a cardiovascular and diabetes risk-reduction program initiated by a self-insured employer. Am Health Drug Benefits 2018;11: 177-83. 39. Bunting BA, Lee G, Knowles G, Lee C, Allen P. The Hickory Project: Controlling healthcare costs and improving outcomes for diabetes using the Asheville Project model. Am Health Drug Benefits 2011;4:343-50. 40. Wing RR, Bolin P, Brancati FL, et al. Cardiovascular effects of intensive lifestyle intervention in type 2 diabetes. N Engl J Med 2013;369: 145-54. 41. Diabetes Prevention Program Research Group. The 10-year costeffectiveness of lifestyle intervention or metformin for diabetes prevention: an intent-to-treat analysis of the DPP/DPPOS. Diabetes Care 2012;35:723-30.