This study incorporates a prospective economic analysis alongside a randomized controlled trial (RCT). The economic analysis takes the perspective of the health-care system and compares two different physiotherapy programs (movement strategy training (MST) and progressive resistance strength training (PST)) with usual PD care (control group receiving a PD education program without falls education). The three interventions are compared to assess relative cost effectiveness where benefits are measured in terms of two intermediate clinical outcomes, 'falls prevented' and 'injurious falls prevented' and an economic outcome endpoint HRQoL.
Movement strategy training (MST)
The MST intervention group receives an 8 week outpatient program consisting of a once-weekly physical therapy session (usually provided in groups of 4), which comprises strategies to prevent falls, enhance balance and improve mobility as defined by Morris . Education about risk factors for falls and strategies for reducing falls in the home is included. Participants also complete a home program of exercises once a week. During the 8 weeks each individual also receives one home visit by an occupational therapist (OT), physical therapist or nurse where compliance with the program is monitored.
Progressive resistance strength training (PRT)
The PRT intervention group receive an 8 week outpatient program consisting of a once weekly 60 minute individualized strengthening program. Resistance is progressively increased and all exercises are functional in design. Participants are also instructed on how to complete the exercises once a week at home. Appropriate equipment including a weighted vest is provided for this purpose. As with the MST intervention, people in the PRT intervention group also receive a home visit.
The control intervention is an 8 week, once weekly, 60 minute outpatient social activity and PD education program, with an additional activity performed once a week at home. During the program, each person also receives a home visit (as described above).
The study does not restrict participants from accessing their usual care or any other activities. It is predicted that a small number of participants might access other physical therapy services, or home modification during the intervention and 12-month follow-up period. The baseline, 3 month and 12 month follow-up questionnaires were designed to capture this information.
Participants are recruited into the RCT from specialist outpatient movement disorder clinics and private neurologists and physical therapists in the Melbourne metropolitan area as well as from advertising in local newspapers. All participants recruited for the RCT are included in the cost effectiveness analysis. To be included, participants are required to have idiopathic PD, be willing and able to provide informed consent, able to walk and safely participate in an exercise program and have the ability to travel to the outpatient clinic for the 8 weeks of therapy and for 4 testing sessions. People are excluded if they scored less than 24 on the Mental State Examination (MMSE) , if they are Hoehn and Yahr Stage V  (bed or wheelchair-bound), or if they are on major tranquilizers.
Due to the paucity of trials in people with PD using prospective falls as an outcome measure, the sample size calculation for the study was inherently speculative. Falls rates were initially determined based on data from a previous study of people with PD and older people without PD . The minimum sample size per group required to obtain a statistically significant difference in falls frequency between the control group (with an estimated 60% falls rate) and the PST or MRT groups (with a 40% rate) was calculated as 110 participants (α = 0.05, β = 80%, 10% drop-out rate). However, a recent blinded interim assessment of the falls data indicated a far higher falls rate than initially expected (over 800 falls reported by 93 participants enrolled). An interim analysis by an external data monitoring committee is scheduled in the near future and will consider intervention efficacy, safety and sample size.
All participants are tested four times by a trained blinded independent assessor: prior to and at the completion of each 8 week intervention, and 3 months and 12 months after completion of the intervention.
Ethics approvals were obtained from the Southern Health Human Research Ethics Committee (HREC Number 060035) and The University of Melbourne Health Sciences Human Ethics Sub-Committee (Number 0828579) of Australia.
Measures of outcomes
An intervention aimed at improving strength or mobility and reducing falls in people with PD is likely to have an immediate effect on both health outcomes and HRQoL. We have therefore incorporated a range of outcome measures aimed at detecting these changes in our core set of fall outcome measures. The primary clinical outcome measure is falls frequency (number of falls per person). A fall has been broadly defined as an unexpected event in which the participant comes to rest on the ground, floor or lower level [31, 32]. This definition is consistent with the literature on falls prevention. Changes in HRQoL will be measured using both a PD-specific measure (PDQ-39)  and a generic utility instrument (Euroqol-5D)  administered at baseline, immediately post-intervention, 3 months, and 12 months.
Measures of cost
From the health care-system perspective costs incorporate the direct costs of each intervention, including the costs of running the interventions themselves, the costs to the participants of accessing the intervention and any relevant downstream health costs or savings. These costs (and savings) will be quantified, valued, and aggregated to determine total and marginal cost estimates per unit of outcome.
Direct costs of delivering the interventions will include the resource costs of running the physiotherapy training, the Falls Prevention Education Program, and conducting the home visit. Cost categories will include staff costs (therapists and allied health assistants), equipment (such as weight training equipment used in the PRT intervention), space, capital, overheads, and consumables (education materials). Program development costs will be attributed to each intervention, capturing the costs of training staff to deliver the interventions and developing the intervention protocols.
The costs/savings of health services utilization resulting from events relating to falls and related injuries will be valued to capture the changes in service use associated with reduced falls that result from the interventions. This requires identifying:
• healthcare costs resulting from events relating to falls and related injuries;
• the number and severity of injuries resulting from falls (by injury type);
• the number of hospitalizations and length of stay attributable to falls related injuries; and
• the number of people requiring institutional care following falls injuries.
The direct costs or out-of-pocket payments associated with the interventions that are borne by patients and/or their families will be included. From the health care system perspective the costs of access (transportation) to and from the intervention is likely to be the primary direct personal cost associated with the intervention.
Indirect costs associated with the interventions will be included in the analysis. These are the opportunity cost for participants and their carers of the intervention (patient time), and productivity gains/losses in work and leisure that result from the intervention.
Exclusion of trial costs
All setup costs and other costs related to the clinical trial will be excluded. These include costs associated with trial administration, data collection, and actual measurement of clinical outcomes as per the trial protocol.