The investigators are conducting a multi-center randomized prospective trial comparing a standard physical therapy gait and balance program with a gait and balance program that includes the ActiveStep™ treadmill.

Subjects will be randomized to either have the ActiveStep™ as part of their therapy or to have their therapy without it. Data will be gathered from session notes, medical records and short interviews with the subjects at baseline and every 3 months over the telephone for the duration of the study.

Falls are the leading cause of fatal and non-fatal injuries in the elderly. Each year an estimated 1/3 of older adults fall, and the likelihood of falling increases substantially with advancing age.2 In 2005, 15,802 elderly patients (> 65 years of age) died as the result of fall-related injuries. According to the 2006 Behavioral Risk Factor Surveillance System (BRFSS) survey, 5.8 million persons aged 65 or older fell at least once during a 3 month period, which represents nearly 16 percent of the elderly population; furthermore, nearly a third of those who fell (1.8 million) sustained some form of injury that resulted in either a doctor's visit or restricted activity for at least 1 day. Falls thus represent a major cause of functional limitation and disability in the elderly. In 2000, the direct medical costs for fall-related injuries was estimated to be $19 billion. Our primary objective is to perform a multi-center, randomized clinical trial to evaluate the comparative effectiveness of a perturbation-based dynamic stability intervention (using the ActiveStep® training system) compared to a traditional physical therapy program in reducing the incidence and severity of falls in a geriatric population.

Modifiable risk factors for falls include muscle weakness, gait and balance problems, poor vision, use of psychoactive medications, and home hazards. Most effective fall-prevention interventions focus on exercise, either alone or as part of a multi-faceted program including medication management, vision correction and home modifications. One Tai Chi based program reduced the risk of falls by 55% in a randomized trial. While exercise programs appear to have efficacy in reducing fall risk, the type, frequency and duration of exercise that is adequate to significantly reduce the risk of falls remain unclear.

External perturbations such as tripping are a major contributor to falls, and trip-related falls account for as many as 20% of all hip fractures.

The ActiveStep® fall prevention training program was developed with NIH funding and consists of a series of increasingly challenging postural disturbances using a microprocessor controlled servo motor that drives a treadmill belt at defined acceleration, velocity, and distance profiles. Initial biomechanical testing of the ActiveStep® training program showed significant improvement in trunk kinematics during a single training session. The ActiveStep® training program is currently available and billable under Medicare reimbursement codes; however, there is a lack clinical data showing that it can reduce falls during patients' routine activities better than standard physical therapy interventions.

The Specific Aims are to:

1. To conduct a prospective, multi-center, pragmatic, randomized comparative effectiveness trial comparing a standard physical therapy gait and balance program with a perturbation-based dynamic stability intervention using the ActiveStep® training system. The main outcome measures were the number of subjects who reported a fall and the number who reported an injurious fall during the follow-up periods. Secondary measures included the subjects' self-reported balance confidence using the Activities-specific Balance Confidence (ABC) scale.
2. To assess the longevity of any improvement with the two types of fall prevention programs by assessing falls every three months out to at least one year. The data on time to first fall will allow insight into the potential need for and timing of "retraining" for fall prevention.
3. To evaluate for underlying patient characteristics that may predict successful prevention of falls in both treatment groups and thus help with determining appropriate selection of patients into different types of fall prevention programs.

Men and women 65 years or older who are referred for fall prevention physical therapy intervention will be eligible for the Randomized ActiveStep® Comparative Effectiveness Trial (RACE Trial). Patients will either have had a prior fall or score as a fall risk on at least one of the balance assessment measure as described below. Following education about the study and participation in the informed consent process, patients who agree to participate will be randomly assigned to either standard physical therapy gait-and-balance training (PT) or to PT+ActiveStep® in a one-to-one ratio. The study will not be blinded as there is no adequate sham-control for the ActiveStep® training; however, allocation concealment will be ensured until after patients are enrolled. The final planned target enrollment will be 550 subjects.

At the time of enrollment, all participants will undergo a fall risk assessment. This will include several standard instruments including: the Timed Up & Go Test (TUG); 42 the Dynamic Gait Index (DGI); 43 and the Berg Balance Measure. 44 In addition we will administer the Activities-specific Balance Confidence (ABC) scale, a questionnaire that assesses the level of self-confidence the patient has in their balance during 16 specific activities. 45 This scale does not assess fall risk per se but does assess fear of falling that can have a significant impact on patients' activity levels and quality of life.

The treatment protocol for the PT cohort will consist of current standard gait / balance training physical therapy as described below. The PT+ActiveStep® will receive standard PT training but with the inclusion of a number of 15 minute sessions on the ActiveStep® simulator. The content of the standard PT in the ActiveStep® group will be adjusted such that the total number of rehabilitation visits and amount of rehabilitation time will be the same for both groups, taking into account the maximum number of allowable visits for reimbursement. Thus, the ActiveStep® will substitute for some portion of the standard PT intervention.

Following the treatment phase, study subjects will be contacted by telephone every 3 months for the duration of the study for a brief interview regarding their history of falling and any significant changes in medical condition. The primary question of interest is the potential difference between treatment groups in the ability to reduce the occurrence of injurious falls subsequent to treatment. This question will be evaluated in two different ways. First, the proportion of subjects in each group experiencing a fall or an injurious fall at each timepoint will be compared; this will allow direct comparison to the three month outcomes of the pilot trial. Second, survival analysis will be used to evaluate for group differences in the likelihood of a post-treatment injurious fall over the follow-up period. In this approach the survival analysis will predict time to first injurious fall. Both of these analyses will include relevant baseline patient characteristics and pre- to post-treatment changes in functioning in an attempt to identify factors or constellations of factors that might predict subsequent fall risk. Secondary analyses will look for the occurrence of any falls using negative binomial regression to allow for the occurrence of multiple falls per subject over time.

Comorbid factors gathered as part of the follow-up process, such as medication use and co-morbidity status, will be evaluated across treatment groups using simple maximum likelihood chi-square tests. Additionally, changes in the TUG, Berg, DGI, or ABC scales before and after treatment will also be evaluated.

Inclusion Criteria:

• Age 65 and over
• Referred to physical therapy for balance training

Exclusion Criteria:

• Age < 65
• Unable to use treadmill
• Severe vertigo
• Not a candidate for gait/balance PT
• Not competent to consent to research