An observational design with multiple measures taken at a single point in time was used as part of this feasibility study. Since the objectives of this pilot study were to test the model and assess feasibility, all outcome measures including questionnaires, physical tests, ADL and community activity outcomes were assessed once between between eight and 18 months following discharge from an acute care hospital. A delay of at least eight months between amputation and assessment was chosen to make sure that participants had time to adjust to their amputation and resume their activities, in order to get a representative picture of their effective mobility. The research protocol was approved by the Research Ethics Committee of the participating institution (Hôpital St-François D'Assise) and all participants provided their informed consent.
Ten instruments associated with the modulators of mobility were used to document the influence of 24 variables identified as factors influencing mobility [1, 8–11, 17, 23, 34, 36, 38, 39, 41–47]. Note that the term "negative" modulator refers to the idea that a personal or environmental factor, when it reaches a certain threshold, may limit or hinder mobility.
The sociodemographic, clinical, physical and psychosocial characteristics of subjects were collected using a questionnaire developed for the study. Negative modulators included age over 60 , an amputation below the knee , fatigue felt when performing ADL , daily consumption of cigarettes and alcohol [48, 49], living alone or in a care centre , a physical environment that is not very accessible [38, 39], dissatisfaction with the technical aids used , presence of another physical problem and lack of services (rehabilitation, etc.).
The Charlson Comorbidity Scale adapted to a geriatric clientele  was used to identify medical conditions via the medical record. A score of 2 or more was considered to be a negative modulator .
The Interpersonal Support Evaluation List (ISEL), developed by Cohen et al.  and adapted by McColl and Skinner , measuring three types of support--instrumental (7 items), informational (6 items) and emotional (9 items) was used. It includes a four-level Likert scale, from 0 to 3, as well as a satisfaction scale for each type of support. The total score (average of three types of support) is the variable used. A score lower than 1 indicating low support is considered to be a modulator-influencing variable.
The MOS Social Support Survey  was used as a social support measurement. The French version of this tool has good psychometric qualities. It was validated in a rehabilitative context of patients with cardiovascular disease . It includes 19 questions covering five types of support: tangible, emotional, affective, positive social interaction and informational. Scales of 0 to 5 ("never" to "always") showed an average score of 5 for each type of support. A score of < 75/95 was considered to be a negative modulator .
The Ways of Coping Questionnaire (WCQ)  was used to measure the adaptability of the individual following difficult events, such as returning home following discharge from an acute care hospital. Its internal coherence is good. The abridged version with 21 items is divided into three different aspects, i.e.: 1) distancing and avoidance, 2) looking for social support, and 3) positive re-evaluation and problem solving. A Likert-type scale allows to measure the level of use of different adaptive strategies. Adaptation was considered to be a negative modulator when Aspect 1 received an average score of > 1.5/3 or when aspects 2 and 3 received average scores of < 1.5 for aspects 2 or 3 .
The Modified Brief Pain Inventory (BPI) [57, 58] was used to evaluate pain intensity and how this pain interferes with the person's activity. It was used with various types of patients including those with neuropathic pain [59–61]. A French version of the BPI was validated [62–66]. Scores of > 5/10 (Question 5) or > 7/10 (Question 9a) on the Likert scale were considered to be negative modulators .
A Body Mass Index (BMI) of more than 30 was considered to be a negative modulator [68, 69]. Height (cm) and weight (kg) were taken from in the patient file.
The Yesavage Geriatric Depression Scale was used [70, 71]. Depression was identified starting with a score of 11/30 downward, with 92-95% sensitivity and 84-89% specificity. Above the threshold (11/30), it was considered to be a negative modulator .
The Jamar dynamometer  is valid and reliable  for measuring hand-grip strength, which is also considered to be an overall strength index for the individual. Based on the standards established according to age, strength (the average of two tests) in the 30th percentile or less was a negative modulator .
The Semmes-Weinstein Monofilament Test (10 g) was used to test protection sensitivity of the intact foot . Four sites of application were evaluated (big toe, top of the first, third and fifth metacarpals), at a rate of two actual stimulations plus one factice per site. The rest-retest reliability is good [75–77]. As soon as a monofilament was not felt, regardless of the site of application, sensitivity was considered to be a negative modulator .
Four instruments were used to evaluate potential mobility, i.e., capabilities to perform various activities requiring movement [27–31, 42, 79–85].
The Locomotor Capabilities Index (LCI)  is a measurement of the perception of a person's capability to perform activities with and without his/her prosthesis. The patient is asked whether he can do seven basic activities (from 'getting up from a chair' to 'going downstairs'), and seven advanced activities (from 'picking up an item from the ground when he is standing using a prosthesis' to 'walking while carrying an item'). Each task is rated on a scale (0 = no, 1 = yes with help, 2 = yes with surveillance, 3 = yes). Two sub-scores out of 21 were produced and a total score of 42. This potential mobility indicator "failed" if the score was equal to or less than 21/42, given that this threshold indicates frequent difficulties and the need for technical or human assistance .
The Timed Up and Go Test  (TUG) measures the time required to get up from a chair, walk three metres and then sit down again. The inter-rater and test-retest reliabilities are very high among ambulatory lower-limb amputees . For individuals in wheelchairs, one task of the Wheelchair Skills Test [89, 90] was used instead of the TUG, i.e., the transfer from a wheelchair to a flat surface (e.g. bed, chair), moving forward three metres in the wheelchair, turning, returning and once again going from the wheelchair to the to the initial surface. This indicator of potential mobility was considered to be "failed" if TUG > 14 seconds in both versions (prostheses and with assistive devices for walking) . For the wheelchair-adapted TUG, since no standard cut-point has been established or recommended in the literature, the threshold was set at > 34 seconds, based on the average (n = 6) obtained during the pilot project.
The Berg Balance Test (BBT)  is a measurement of seated and standing balance. It is made up of 14 subtests, rated from 0 to 4. A score lower than 45 is predictive of the risk of multiple falls.
The Amputee Mobility Predictor (AMP)  is a measurement that predicts the movement capabilities of lower-limb amputees. Its psychometric qualities were demonstrated in English and French . The AMP evaluates 21 situations requiring seated or standing balance, during movement or when reaching for objects. Scales from 0 to 2 are used, except for the last item. A failing mark in this potential mobility indicator was when the score was under 25/47. Under this threshold, the subject is unstable, loses balance and cannot reach objects .
Three instruments associated with effective mobility cover the evaluation of activities in a real environment [3, 33, 94] and the level of physical activity able to be performed.
The Assessment of Life Habits (LIFE-H) [95, 96], abridged version 3.1, was used as the main way of measuring outcomes of effective mobility. This tool was developed from the Handicap Production Process  and evaluates handicap situations that hinder social participation. It includes 200 items that could be grouped daily activities as well as social roles. The score varies from 0 (high level of handicap situations) to 9 (optimal level of social participation) for each item. The reliabilities were excellent for the overall score studied in a group of 84 seniors who had lost autonomy (CCI and confidence intervals at 95%: 0.95 (0.91 to 0.98) (test-retest) and 0.89 (0.80 to 0.93) (inter-rater) . This effective mobility indicator was considered to have "failed" for daily activities and for social roles if their average score is < 7/9, given that, at this threshold, the person can only carry out a given activity with an assistive device, or layout modifications or requires human assistance .
The Life Space Assessment (LSA)  identifies the regular mobility patterns during the month preceding the evaluation. Five levels of movement are assessed, from inside the home to outside the city. Its score varies from 0 (complete restriction of one's mobile space) to 120 (no restriction of one's mobile space). Validity, reliability and sensitivity to change in the LSA were studied in ambulatory seniors and wheelchair users [99, 100]. Metric properties in the French-Canadian version are good . This effective mobility indicator was deemed to have "failed" if the mobility space scores were below the standard values for people aged 65 and more .
The Human Activity Profile (HAP)  measures the capability to perform different levels of physical activity. It is a questionnaire with 94 questions, is self-administered and validated for different clienteles [102–104]. It begins with questions on the activities that require low energy expenditure and ends with questions associated with high energy levels. Each item must be answered using one of the following three statements: I still do this activity, I have stop doing this activity and I have never done this activity. The maximum activity score (MAS), which corresponds to the last item that the subject had done was selected. This potential mobility indicator was considered low if the MAS was lower than the standard values for the age .