Scaling-up implementation in community hospitals: a multisite interrupted time series design of the Mobilization of Vulnerable Elders (MOVE) program in Alberta

The MOVE AB intervention was conducted on seven units at four community hospitals in Alberta, Canada. Units included: one long-term/restorative care, one general medicine, one surgical, one geriatric rehabilitation, one short stay surgical and two intensive care units. The participating units were selected by local leadership and the MOVE program was conducted between July 2015- July 2016. Patients included in the study were aged 65 years and older who were admitted to participating units. Patients receiving palliative care or on bed rest were excluded.

The implementation process was guided by the knowledge to action (KTA) cycle [19]. Each site began the MOVE AB intervention in the first week of July 2015 (see Additional file 2 for MOVE program timeline). The first 10 weeks of the study was pre-intervention phase, in which a readiness assessment was employed [20‐22]. Each participating unit completed a readiness assessment survey to understand organizational readiness for implementation. Readiness assessments were administered to staff at the unit level. As part of a sub-study, a cluster randomized trial was employed to compare the effects of using a decision support tool to select a readiness measure and selecting a readiness measure without the aid of a decision tool (findings not presented here). Two randomly selected sites selected their own readiness assessment measure from a list of validated readiness measures and the remaining two sites used the Ready, Set, Change! Decision Support tool (http://​readiness.​knowledgetransla​tion.​ca), a decision aid tool that selects the most appropriate readiness measure based on the organizational setting [20]. Implementation activities were selected using an online barriers and facilitators assessment tool, the Select, Tailor, Plan and Engage (STEP) tool informed by the theoretical domains framework [23‐25]. This online tool was completed by participating units to determine areas of improvement and to help inform the selection and development of appropriate implementation activities at their site, thus considering local context. Sites accessed all tools and resources using the MOVE online portal through the MOVE website (www.​movescanada.​ca). We monitored implementation activities using an implementation process tool.

To plan for sustainability, site implementation teams completed the NHS Sustainability Survey in both the pre-intervention and post-intervention phases [26]. The purpose of the pre-intervention NHS Sustainability Survey was to determine the overall climate of sustainability and what strategies should be implemented as part of the sustainability component of the intervention. The purpose of the post-intervention NHS Sustainability Survey was to determine whether there were changes in the climate of sustainability throughout the intervention and whether additional considerations for sustainability were necessary. The NHS Sustainability findings were disseminated to sites and utilized as an implementation support tool (Additional file 3).

Outcome ascertainment was consistent across sites. The primary outcome was the proportion of patients aged 65 and older who were mobilized at least once daily per project week, where visual audits were conducted three times per day for two days each week during the pre-intervention phase (10 weeks), intervention phase (8 weeks), and post-intervention phase (20 weeks). Eligible patients were identified and observed over time using unit -specific administrative data (e.g., daily patient census report) prior to conducting audits. Patient identification numbers and types of mobility observed were recorded using an audit tool (Additional file 4). Types of mobilization considered in the analysis included standing/walking with assistance, standing/walking supervised, and standing/walking independently. Immobile patients were those lying in bed, sitting in bed in upright position, sitting on a chair, and those who could not be observed (to allow a conservative estimate of rate of mobilization). Observational patient audits were documented by a trained research assistant at each participating site [17]. This method of mobility audit was previously compared to continuous rounding (positive likelihood ratio, 12.2; negative likelihood ratio, 0.06) [16]

Secondary outcomes were hospital length of stay (LOS) and perceptions of the intervention from site implementation teams, participating unit staff and patients/family members. LOS data were obtained from hospital administrative data along with age, gender, patient location prior to admission, and admitting diagnosis of eligible patients. Perceptions of unit staff on the involved hospital units, and of patients and their family members were collected through anonymized exit surveys (Additional file 5 and Additional file 6). In addition, the central team conducted semi-structured interviews with site implementation team members to explore their perception of the intervention and suggestions for facilitating sustainability. Implementation teams were invited to participate in interviews via email by the MOVE implementation coach. Each interview was approximately 60 minutes in length and conducted over the phone by a member of the MOVE Central Team.

A total of 42,840 mobility audits from 3,601 unique patients were completed within the four participating community hospitals in Alberta, Canada [Table 1]. These hospitals provide acute care services to communities ranging in size between 9,000 and 90,000 individuals. The seven participating units across the four hospitals varied and included surgery, geriatric rehabilitation, general medicine, intensive care, and long term/restorative care. The mean age of the included patients was 80.1 years (Standard deviation [SD] = 8.4 years) across the four sites. The most common diagnoses by patient visit were chronic obstructive pulmonary disease, heart failure, urinary system disorders, pneumonia, and ileus/intestinal obstruction.

Each unit completed the STEP Tool, which provides sites with a list of strategies that implementation teams could choose from based on each site’s identified barriers and facilitators. Each site was unique in terms of their overall patient, provider, setting, and implementation barriers. Sites were able to prioritize key barriers (e.g., patient/family beliefs about mobilization, presence of other unit priorities) and select appropriate strategies. There were similarities and differences in terms of barriers and chosen strategies between and across sites. All sites chose to use champions and documentation (i.e., use of whiteboards, log sheet), in addition to the mandatory components of educational meetings and audit-feedback. Sites were able to tailor materials to the local context, for example they could use local logos and embed MOVE key messages into existing program materials. For example, a surgical unit incorporated MOVE strategies into their recovery patient materials. Other commonly utilized strategies included reminders and educational materials. All sites also had patient education activities (e.g. bedside education) and educational materials (e.g. pamphlets, posters). For a complete list of selected implementation strategies, please see Additional file 7.

There are an estimated 6% (95% CI: [1, 11], p-value=0.0173) more out of bed audits at the end of the intervention period compared to end of pre-intervention period (Fig. 1). A decrease in rate of mobilization was observed post intervention implementation, where a slope difference of -0.68% (95% CI [-1.17, -0.18], p = 0.0073) was estimated between pre- and post-intervention implementation. Consequently, an estimated -5% (95% CI: [-10, -0.1], p-value=0.0443) fewer out of bed daily audits were observed at the end of the post-implementation period compared to during intervention.

During the intervention period, there was a non-statistically significant slight decrease in median LOS observed, where patients on average stayed an estimated 3.59 fewer days (95% CI:[− 15.06, 7.88], p-value = 0.5401) in hospital compared to post-intervention implementation (Fig. 2). An increase in median LOS was observed immediately post-intervention followed by a slight increasing trend. Patients stayed 2.17 (95%CI:[− 20.49, 16.15], p = 0.8162) less days in hospital post-intervention implementation compared to the pre-intervention implementation.

Exit surveys were completed with 29 staff members (8 nurses; 5 physicians; 2 administrators; 2 physiotherapists; 11 other allied health professionals) and 45 patients (39 patients; 3 family members/caregivers; 3 unidentified). For staff and patient participants stratified by site, please see Additional file 8.

Most staff perceived the educational activities as effective in increasing staff knowledge on early mobilization. Specifically, survey respondents perceived staff education (72%; n = 21) and educational materials (76%; n = 22) as being effective in increasing their knowledge. A total of 48% (n = 14) of staff respondents believed that patient and family members’ knowledge about early mobilization increased. Similarly, 48% (n = 14) of staff respondents perceived patient and family members support for early mobilization increased as a result of MOVE.

A total of 51% (n = 23) of patients believed education activities and materials were effective in increasing their knowledge about the importance of moving during their hospital stay. Although fewer patients (40%; n = 18) perceived the educational activities to be effective in helping them move during their hospital stay, 59% (n = 24) identified that staff always or frequently encouraged them to perform more mobilization activities.

Nine site implementation team members representing all four sites participated in interviews to provide feedback on their implementation experience with MOVE. See Additional file 9 for participant breakdown by profession and site. Keeping staff abreast of MOVE activities, outlining the intended purpose of each step, and providing updates were all ways identified to help support staff participation. Most sites held implementation planning meetings. Involving front-line staff, especially champions, in the planning and implementation activities was seen as a key strategy and helpful at all sites.

Interview participants identified that education sessions with staff were a key strategy to do early on to disseminate information and engage frontline staff. Regular, dedicated time with unit staff to discuss MOVE implementation was perceived to be important for making staff aware of the intervention, and was an opportunity to identify complementary initiatives to support mobilization with both staff and patients. Interview participants expressed the overall importance of embedding the practice change into work processes and organizational culture, so that MOVE no longer felt like a program or new initiative.

Consistent communication between the central coaches and the sites was also seen as an effective way of providing ongoing support. Steering committee meetings gave sites the opportunity to share and learn from each other, including discussing common challenges and potential solutions. Interview participants also felt that the feedback they received on intervention progress throughout implementation was helpful.

All sites (n = 4) completed the NHS Sustainability survey during pre- and post-intervention to identify strengths and weaknesses in implementation planning and predict the likelihood of sustainability (Additional file 3). Based on the NHS Sustainability Guide, an overall score of 55 or higher suggests that a site is likely on track for sustainability, a score between 45 and 55 suggests making small changes to get on track for sustainability, while a score of 45 or lower recommends a site create a sustainability plan to increase the likelihood initiative will sustain. Based on pre-intervention survey results, two sites, Site B and Site C had a sustainability score above 55 (62.9 and 85.1), Site A had a sustainability climate score of 45.2 (between 45 and 55), while Site D had a score of 23 (45 or lower). Common strengths from sites were strong senior leadership engagement and adaptability of sites in facing ongoing changes in staff, leadership, organization structures etc. Challenges during pre-intervention were lack of infrastructure to enable sustainability (e.g., policies and procedures to reflect new processes) and lack of staff involvement and training to sustain the process. During post –intervention, sites with lower sustainability scores demonstrated an increase in sustainability climate where Site D had a sustainability climate score of 38.6 and Site A had a sustainability score of 62.9. Site B demonstrated a small decrease in sustainability with a score of 60.6 whereas Site C had a substantiality score of 40.9. During post-intervention, a common strength amongst sites was clinical leadership engagement and a common challenge identified was lack of senior leadership engagement.