Improving walking speed reduces hospitalization costs in outpatients with cardiovascular disease. An analysis based on a multistrata non-parametric test

Background: To assess the association between walking speed (WS) and its improvement on hospitalization rates and costs in outpatients with cardiovascular disease. Methods: Six hundred forty-nine patients participating in an exercise-based secondary prevention program were studied. Patients were divided at baseline into two groups characterized by low and high WS based on the average WS maintained during a moderate 1-km treadmill-walking test. WS and other covariates were grouped into three domains (demographic factors, medical history and risk factors), and used to estimate a propensity score, in order to create homogeneous groups of patients. All-cause hospitalization was assessed 3 years after baseline as a function of WS. Hospitalization and related costs were also assessed during the fourth-to-sixth years after enrollment. To test whether the hospitalization costs were related to changes in WS after 36 months, a multistrata permutation test was performed by combining within strata partial tests. Results: The results support the hypothesis that hospitalization costs are significantly reduced in accordance with an improvement in WS. This effect is most evident among older patients, overweight or obese, smokers, and those without a history of coronary artery bypass surgery. Conclusions: The present study supports growing evidence of an inverse association between WS, risk of hospitalization and consequent health-care costs. The joint use of propensity score and multistrata permutation approaches represent a flexible and robust testing method which avoids the possible effects of several confounding factors typical of these studies.