Self-managed occupational therapy and physiotherapy for adults receiving inpatient rehabilitation (‘My Therapy’): protocol for a mixed-methods process evaluation

The My Therapy trial uses a stepped wedge cluster randomised design. It consists of a trial period of 54 weeks comprising nine blocks, each of six-weeks duration. The stepped wedge cluster design has a unidirectional crossover from the usual care condition (control group receiving usual care) to the experimental condition (intervention group receiving My Therapy in addition to usual care). Block one of the My Therapy trial will see all wards remaining under the control condition. From block two, and for each subsequent block, one of the eight wards will cross over to the experimental condition. The randomisation of the ward allocation has been detailed in the My Therapy Stepped Wedge Cluster Randomised Trial Protocol describing the clinical and economic outcomes [6].

The process evaluation documented in this protocol will be conducted alongside the main clinical trial using a mixed-methods design. This protocol has been developed in accordance with the SPIRIT checklist [7] and has been based on the MRC guidelines for developing and evaluating complex interventions [1]. The intention is to report on implementation, and identify contextual factors associated with variation in outcomes [8]. The contextual outcomes considered include understanding the perspectives and experiences of patients and therapists using My Therapy. The use of the logic model assisted in conceptualising components of the My Therapy program considering inputs, activities, outputs, impacts, and outcomes. This logic model informed and guided the process evaluation protocol (Fig. 1).

The main clinical trial and the process evaluation of My Therapy will take place in eight rehabilitation wards across two public and two private Victorian health networks (in Australia).

Patient participants: All participants of the stepped wedge cluster randomised trial (2,160 rehabilitation patients) will be included in the process evaluation (service profile survey and ward compliance audit - Table 1); with a subset of 120 participants undergoing more intensive evaluation (e.g., surveys and activity logs), subject to eligibility. The process evaluation will include patients aged 18 years or older, with any diagnosis, who have been admitted to one of the participating wards. The evaluation of the subset of 120 patients will exclude patients with a cognitive impairment, patients without access to Medicare (Australia’s universal health care program), as well as those who do not speak English. The cognitive exclusion is due to the burden of the data collection requirements within the process evaluation and the requirement for participants to recall their experience of rehabilitation. Cognitive impairment will be determined via health professional documentation within the medical record and/or through discussion with treating allied health, medical and nursing staff. Should a person’s cognitive skills improve, for example following resolution of delirium, patients may be invited to participate.

Staff participants: 24 registered occupational therapists and physiotherapists from participating wards, who are involved with the My Therapy trial, will be invited to participate in the process evaluation. There are no exclusion criteria for staff participation.

Patient participants: A subset of 40 patients (five from each of the eight wards) will be recruited to the process evaluation across three time points (randomisation blocks one, five and nine; total n = 120). Patients will be recruited using a convenience sampling approach, based on consecutive admissions over the study period. While it is expected this sampling method will recruit a diverse sample, participants will not be purposively recruited based on age or condition. A sample of size of 120 will provide sufficient power to detect a worthwhile effect (0.5) [9] on self-efficacy, with alpha = 0.05, power = 0.8, and assuming equal sample sizes in control and experimental groups. This sample size (n = 120) will also be sufficient to determine that 85 % of participants have the factor of interest (e.g. patient compliance with activity log) with 95 % confidence that between 78 and 92 % have the factor of interest, assuming an intra-class correlation coefficient of 0.005 and a cluster size of 15 [10].

The My Therapy program is described in detail in the My Therapy Stepped Wedge Cluster Randomised Trial Protocol describing the clinical and economic outcomes [6]. In summary, each ward will transition from the usual care condition (control group receiving usual care) to the experimental condition (intervention group receiving My Therapy in addition to usual care). My Therapy is a self-management program designed to increase dose of therapy by supporting patients to independently practice occupational therapy and physiotherapy exercises and activities outside of supervised therapy sessions. My Therapy is a partnership model based on the patient’s rehabilitation goals, abilities and willingness to participate, recommended by the treating occupational therapist or physiotherapist. Patients will be asked to complete exercises and activities selected from an online program, www.​physiotherapyexe​rcises.​com (PTX). The My Therapy exercises and activities will be sent to patients via SMS or email and viewed on a ‘smart device’ (patient’s own) or printed in hard copy and provided to the patient. In Australia, 86 % of households have internet access and 91 % of those with internet access use a mobile or smart phone [11], supporting the option of ‘smart device’ use. The recommended exercises, as well as the dose (sets and repetitions) recommended, can be customised and updated regularly by the treating therapists, as appropriate.

Within the four health networks, self-management is not currently standard practice within usual care rehabilitation; it is ad hoc, clinician-dependent and varies in terms of its definition within rehabilitation [12]. To understand the presence of self-management programs within control conditions, it is essential that therapist advice or education is differentiated. For the control condition, a self-management program, should it be provided by the treating therapist, needs to meet all of the following criteria: a written program (delivered electronically or in paper format); documented by the therapist in the medical record; include a feedback mechanism to the therapist; and be actively monitored and progressed, as clinically indicated. For wards under the experimental condition, this specifically refers to provision of a My Therapy program.

The data collection approach for the process evaluation is outlined in Table 1. There will be three layers of data collection to address the objectives of the process evaluation (ward, patient and staff) using a combination of qualitative and quantitative methods. In addition, a log of My Therapy adaptations (to the program itself or to how the program is implemented) will be maintained over the nine blocks of data collection, based on iterative feedback from staff, patients and consumer representatives.

In each of the nine blocks, a service profile survey will be completed to capture the characteristics of each rehabilitation ward (for example staffing ratios, inpatient vs. home-based ward). This will provide the opportunity to identify contextual factors associated with variation in outcomes and map barriers and enablers against the characteristics of each ward.

In each of the nine blocks, a ward compliance audit will also be completed to measure patient participation in a self-management program on a set day. The audit will capture information regarding the number of patients on the ward and the number of patients with a self-management program through staff feedback (Table 1).

Patient evaluation (Table 1) will be completed in blocks one, five and nine. This will include: an audit of PTX to measure patient completion (i.e., adoption) of My Therapy exercises and activities (experimental condition wards only); a patient survey, including an exercise self-efficacy scale, to understand the acceptability and impact of self-management / My Therapy, as well as the patient barriers and enablers to participation; and an activity log to measure patient fidelity with completing My Therapy tasks and other self-management exercises. A small subset of patients will also complete activity monitoring to further understand patient fidelity with the My Therapy program.

The Spinal Cord Injury Exercise Self Efficacy Scale (ESES) will be used to measure participants’ exercise behaviours and exercise self-efficacy, enabling researchers to understand the person’s confidence and ability to self-manage their My Therapy program (or self-management activities) [13]. The ESES will be completed on the final day of the activity log recording. The ESES has 10 items, each rated on a 4-point Likert scale [13]. Scoring can range from 10 to 40, with a higher score indicating higher exercise self-efficacy. Acceptable reliability and validity has been noted with the SCI population [13], and the tool has been applied to patients with neurological conditions [14], congenital heart disease [15], and chronic obstructive pulmonary disease [16]. In cases where one or more items of the ESES have missing data, but over half of the items have been completed, the average score across all completed items will be used for the missing item scores, to calculate an overall ESES score.

Twelve patients, from two of the four health services, will be asked to participate in physical activity monitoring (during blocks one, five and nine; total n = 36). The participants in block one will be under control conditions, the participants in block five will have patients under both experimental and control conditions, and participants in block nine will all be under experimental conditions. This will allow us an opportunity to compare the alternate conditions at different timepoints. Participants will be asked to wear an accelerometer-based activity monitor (activPAL, PAL Technologies Limited) on the anterior, middle thigh for 24-hour monitoring over seven consecutive days. Monitors will be placed in a ziplock bag, placed on a small piece of gauze to protect the skin and covered in waterproof dressing. For patients under the experimental conditions (My Therapy), the monitor will measure the daily number of steps and standing time, as an indicator of patient fidelity with the recommended My Therapy program. The activPAL provides a valid and reliable measure of physical activity in older adults [17]. It has detected hypothesised increases in walking during inpatient rehabilitation, providing evidence of construct validity [18].

A survey and a focus group will be completed with participating occupational therapists and physiotherapists. Staff will firstly complete a survey followed by participation in a focus group. The staff online survey will be based on the validated Determinants of Implementation Behaviour Questionnaire (DIBQ) [19]. The survey will be sent to staff participants by researchers via email, prior to each of the focus groups. Survey responses will assist in directing the topics discussed within the focus groups. Staff focus groups will be divided into staff in the control condition and staff in the experimental conditions to ensure there is no cross contamination of views and opinions. The focus groups will be conducted and recorded over a video conference platform. A semi-structured interview guide will include questions on barriers, enablers and acceptability of My Therapy and what could make My Therapy delivery better for future implementation. For therapists whose wards remain under usual care conditions, the questions will be reframed requesting they consider any recommended ‘self-management’ programs as defined by the set criteria.

We will also complete an audit of PTX to measure staff fidelity with prescribing and modifying My Therapy tasks and exercise in blocks one, five and nine. In addition, we will complete an audit of timetabling each block for usual care supervised therapy to capture any changes in usual care occupational therapy or physiotherapy service provision by staff (Table 1).

Descriptive statistics will be used for all quantitative data, in line with recommendations by Moore, Audrey [1]. Independent group t-tests or chi-squared tests will also be considered for quantitative data as appropriate. If data is not normally distributed, non-parametric equivalent tests will be used. Study data will be collected and managed using REDCap (Research Electronic Data Capture) electronic data capture tools hosted at Monash University and managed by Helix [20, 21], and exported to SPSS for statistical analysis. All analyses will assume a significance of p < 0.05. For participants who use activity monitors, should they be discharged during the seven day period while it is intended they are wearing the activity monitor, a daily average will be calculated based on the available data with a minimum of three 24-hour days of activity data required for inclusion [22].

Qualitative data from patient surveys will be entered into REDCap and uploaded to NVivo (QSR International) for analysis. Qualitative data from staff focus groups will be transcribed and, following member-checking to confirm the transcript reflects participants’ intended meaning, will be uploaded to NVivo (QSR International), along with qualitative/free-text data from staff surveys. Themes will be identified through thematic analysis, adopting an interpretive description methodological approach [23]. The interpretative description approach allows for health researchers to explore meanings and explanations that may result in clinical implications through inductive analytic processes [23]. All qualitative data will be coded independently by two members of the research team. A third researcher will be engaged should consensus not be reached on the coding framework. Participants will have the opportunity to view the transcript of their focus group or interview to check that their views have been adequately represented. The process of member checking to achieve credibility and confirmability has been considered the ‘gold standard’ for establishing trustworthiness within qualitative research [24, 25]. The concept of trustworthiness in qualitative research has been framed to also consider transferability and dependability [26]. To address the requirements of transferability, a detailed description of the sample population will be provided in addition to providing verbatim quotations. Confirmability and dependability will also be achieved by the triangulation of both quantitative and qualitative data, and by reporting a detailed description of the research process [25]. This will enable credibility, breadth and depth of the analysis of the implementation of the My Therapy program.