Feasibility of a mobility programme for people with dementia in the respite care setting: results of the DESKK study

DESKK was a pilot-based, quasi-experimental feasibility study that included a formative evaluation. The intervention was performed as a single-centre study at a dementia-specific RC centre.

It was planned to obtain data of 30 PwD and their caring relatives. Furthermore, RC staff were included to gain data for the formative evaluation.

The study was carried out in a dementia-specific RC centre with a ten-person capacity (common size for an RC in Germany) in the state North-Rhine-Westfalia, Germany. The dementia-specific RC was built as a pilot facility in the RC setting in 2016. The RC is an independent subunit in a long-term care centre called Haus St. Antonius Paderborn.

The infrastructure of the RC where this study took place was already partially adapted to the specific needs of PwD. For example, there were no visual barriers on the ground, and a circadian light system was installed to support the physical diurnal rhythm of PwD.

Furthermore, the care assistants in this RC performed basic mobility rehabilitation processes with some of the RC guests, but these processes were not systematically integrated in the daily care routine and were not based on assessment.

Only dyads of PwD and their caregiving relatives were planned to be included in the DESKK intervention. This was performed to analyse potential effects of the mobility programme focused on the care-related burden of the caregiving relatives.

The mobility function level of PwD is the primary outcome of the summative analysis. Mobility function was measured by the DESKK mobility assessment. This assessment battery consists of the following physical function tests: the Short Physical Performance Battery (SPPB) [15], the Box and Blocks Test (BBT) [28], the Strength-Dexterity (SD) Test [1] and the Nine-Hole Peg Test (NHPT) [28]. These tests have already been used in the German context, and the necessary information to conduct the tests is also available in German [2, 6, 31].

The DESKK mobility assessment was used to obtain empirical evidence based on the assessment scores. Based on these scores, direct training recommendations for the RC staff were developed. For these recommendations, a cut-off score was set for each area that was tested for mobility function. These cut-off scores were developed based on the literature regarding the utilized assessments. However, reference tables for these assessments were only available for seniors without dementia [1, 15, 16, 28]. For this reason, the assessment scores for seniors with “weak” physical abilities were used as reference cut-off scores in the DESKK mobility assessment. Weak was defined as the weakest third of the tested sample related to the functional abilities of the tested areas in the group of individuals aged 80 years or older.

The decision to set the cut-off-scores this way was based on the literature evidence about significantly weaker ADL abilities and a higher risk of falls in these individuals than in elderly people without dementia [32]. Scores beneath the cut-off thresholds were rated as 0, and scores above these levels were rated as 1. A score of 0 implied that training for that specific tested area was necessary, and 1 implied that physical training for that area was optional. The assessments and their cut-off scores are displayed in Table 1. After a pretest with 6 PwD before the start of the intervention, the cut-off scores for the strength-dexterity test were lowered to the weakest 25% of the population in the sample of Alcenar [1] because of a noticeable lower mean hand strength of the tested PwD.

The secondary outcomes “cognition” and “behavioural symptoms” were assessed by the Mini-Mental State Examination (MMSE) [12] and the Nurses’ Observation Scale for Geriatric Patients (NOSGER) [37] for a description of relevant study population characteristics and necessary abilities to perform the mobility programme.

The DESKK mobility exercises were developed based on existing literature [10, 21, 42] and were discussed with the RC staff and external scientists. By the start of the intervention, 16 individual exercises were included along with an exercise poster. Furthermore, an exercise manual was created, in which the exercises on the poster were described in detail, including cognitive tasks and potential exercise (easier vs. harder). These exercises are specialized to train one of the six mobility function areas (see Table 1). Each mobility function area consists of two to five different exercises, which could be chosen by a schooled RC nurse who functioned as rater to create an individualized training programme. Before the intervention started, the German Center for Neurodegenerative Diseases (Deutsches Zentrum für Neurodegenerative Erkrankungen, DZNE) scientists performed the rater schooling. Nurse assistants were not allowed to perform the mobility tests and exercise compilation. However, they were schooled to perform the compiled exercises written on a training card (for further information, see: www.​deskk.​info) with the RC guests in the facility. The empowerment of care assistants to perform the DESKK exercises during the daily care routine is a core component of the mobility programme because of limited resources of registered nurses. The exercises for the specific functional areas are all different (for individual training preferences) but were designed to train similar muscle groups. To document the daily exercise routine, a semi-standardized training protocol was developed. Furthermore, mobility courses were installed on a floor of the RC centre, including training devices for spontaneous, short-term physical training of PwD led by nurses or care assistants during regular walks inside the RC centre (e.g., from the bathroom back to the recreation area). The aim of this installation was to create a low-threshold way to encourage additional training of DESKK mobility exercises.

The developed DESKK programme was tested in the RC centre from October 2017 until September 2018. Four group interviews were conducted during the intervention phase at intervals of 6 to 10 weeks to obtain data for the formative evaluation. The time intervals were extended during the intervention period. In the group interviews, the intervention processes, regarding barriers and facilitators, were discussed with the involved RC staff. The interviews were held with the help of a self-developed semi-structured interview schedule (attached as supplementary file) that is based on parts of the Consolidated Framework for Implementation Research (CFIR) model [8]. The framework can be used modularly, i.e., in individual parts, since not all CFIR constructs fit every study design [8]. To select the appropriate CFIR constructs, they were first examined for their relevance and suitability for the study design, the intervention and the setting (primary “inner setting”), based on given CFIR literature [24].

The construct areas were chosen to reflect the structural and personal requirements (with their linked processes) to implement the programme as intended. The chosen construct areas are displayed in Table 2.

The group discussions were documented with audio recordings, which were transcribed afterwards. In addition, seven single interviews were conducted using the same method with selected RC staff members, who also took part in the group discussions.

This was done to obtain a deeper insight into individual ratings regarding the project components than would be possible in group interviews because of potential hierarchical influences between the interview participants. Furthermore, a short questionnaire that included quantitative structured, ordinal scaled items was completed by the RC staff to provide the opportunity to reflect changes in essential aspects of the acceptance of the programme over time.

Based on the interview schedule, which was structured on the chosen CFIR constructs, a code tree was created with a deductive procedure using a structured content analysis according to Mayring [27]. Thereby, the interview transcripts were coded by two independent scientists. The material was divided into content sections and codes. Through a circular process, the statements of the interviewees were reduced to the essential content. The coding and analysis processes were performed with the help of MaxQDA 13 qualitative analysis software [41]. At regular intervals, the codes and the coded material were synthesized during a discussion process between the participating project scientists.

The DESKK mobility assessment was used to test the mobility function levels of PwD at the beginning of the training period in the RC centre (t0), after two weeks (t1) and, if possible, after three (t2) and four weeks (t3) of their stay. The training schedule was planned with a minimum of 20 min of active training per day. The mobility function levels of the included PwD and the changes over time between t0 and t1 were analysed using SPSS 21 (IBM Corp., Armonk, NY, USA). In addition to descriptive statistics, a Wilcoxon-Mann-Whitney significance test was performed to identify changes in the mobility function level of PwD between t0 and t1.

The data displayed below were obtained from the single and group interviews with the RC staff. An overview of the interviewee characteristics is provided in Tables 3 and 4.

The following themes were structured based on the CFIR content areas and their definitions:

The average length of stay in the RC centre of the included PwD was 17.7 [SD: 6.7] days. The sociodemographic characteristics of the included PwD are displayed in Table 5. In terms of the sociodemographic data, the study participants had a mean age of 81 years and exhibited moderate cognitive impairments. Approximately two-thirds of the participants were female. With an average score of 15.6 measured by the NOSGER, mild to moderate behavioural symptoms were found in the sample [38]. In particular, apathy and defence reactions were most prominent behavioural symptoms in the sample (n = 12).

The qualitative data obtained in this study provided a detailed insight into the processes, barriers and facilitators of the implementation of the DESKK mobility programme into the daily care routine. The analysed dementia-specific structures of the RC centre are not common for RC centres in Germany. However, it is likely that this intervention could be acceptable and beneficial in any RC centre that has the training and capacity to provide person-centred dementia care. Detailed information about the given Ressources in this specific RC centre are displayed on the concept website www.​deskk.​info.

The time required to perform the DESKK mobility assessment (approximately 15 min after three months of intervention) makes it an efficient instrument compared with other mobility assessments [19].

The connection between assessment scores and suggested mobility exercises is an essential component of the DESKK mobility programme; simple cut-off scores guide nurses to make specific exercise recommendations. There is a great need for such practical approaches in this field, and the related solutions have been inadequate until now [4].

Another essential aspect of the DESKK mobility assessment was the use of objective functional physiotherapeutic measurements instead of questionnaires, as are commonly used in nursing practices throughout Germany and abroad [11, 22]. A main advantage of objective functional measurements compared with questionnaire-based subjective measures lies in the higher validity and reliability of the objective measures. Subjectively rated questionnaires depend on the individual feelings of the rater related to the specific test discipline (e.g., climbing stairs). The related high risk of scoring bias is widely known [14], but nevertheless, objective functional measurements are seldom used in nursing practice. A reason could be that functional measurements are not often incorporated into nursing education and that the functional measurements were developed to be conducted by therapists rather than by nurses [33].

Most exercises were well accepted by PwD, but some exercises were not. A workbench exercise was not well accepted, and the same result was seen in an exercise with clothespins (see results). The work duties of this sample seemed to not be associated with positive memories as predicted. The literature promotes connecting the activities of PwD with their former habits/activities [3]. In the field of physical activity, it could make sense to differ between former hobbies and work-related activities. No literature on this topic was found, but it would be an interesting research question.

The result that ball games worked especially well for PwD is not surprising because many people had ball game experiences in their earlier lives, and these experiences were mostly related to leisure activities. This might explain why ball-related games are one of the most commonly used physical exercises in the field of dementia care [3].

The use of the Nintendo Wii® seemed to cause problems for some RC staff members. However, the use of this gaming console in dementia care has already been evaluated in different studies in the caregiving setting, and few problems for nurses occurred as a result of its use [20, 23].

A very important aspect of why some exercises and the daily documentation of the DESKK mobility programme in the training protocol were not always performed as intended was related to an unexpectedly high staff fluctuation during the intervention period. The rate of programme use was low for most of the part-time RC staff, which is a well-known inhibitor of the successful implementation of complex interventions in daily routines [29].

The exercises themselves were rated by the staff as well suited for PwD in most cases. However, it is not clear whether the preferences had a real impact on the motivation of PwD to perform specific mobility exercises. In the literature, the effectiveness of preference-based exercise compilations remains unclear, and further research is needed in this area [39].

The training protocol was primarily used for empirical data collection but was also intended to act as an activity protocol for the implementation of the programme. As already pointed out, the documentation of the protocol by staff was mostly fragmented and was occasionally imprecise. Some training sessions were not documented in the paper and pencil protocol, or the exact training time was sometimes missing. As reasons for this behaviour the nurses mentioned that the protocol was a paper and pencil version, but they were accustomed to documenting all daily care routines in the electronic system and thereby forgot to fill out the training protocol sometimes. Based on this feedback, the training protocol was embedded in the electronic documentation system after the end of the study.

The sociodemographic findings are quite similar to findings of many other dementia studies in which PwD participated in mobility programmes [13]. The age range between 75 and 85 years also reflects the largest group of people affected by dementia in Germany [7].

The relatively short time since the DESKK participants were diagnosed with dementia might be explained by the fact that many official dementia diagnoses in Germany occur in already advanced stages of the disease or by the fact that PwD are never officially medically diagnosed and, instead, the diagnosis is just determined by the general practitioner or the caregiving relative [25].

The summative evaluation of the programme took into account the general limitation of the small sample and the very heterogeneous mobility function capabilities of the included PwD. As already noted, the sample was smaller than originally planned (goal of n = 30 PwD), but in general, a small and heterogeneous sample (see Tables 5 and 6) is expected and common in feasibility studies [30]. In particular, the high level of heterogeneity in the participants’ physical function areas was intended this way. The inclusion criteria were very broad in terms of the physical abilities of PwD to evaluate the adaptability of the mobility programme to different physical ability levels. Interestingly, the different physical capability levels did not notably affect the programme execution. The basic inclusion criteria ensured that PwD could participate in the DESKK mobility programme. The exercises themselves all included variations to decrease or increase the training intensities to adapt to the individual mobility level.

Because of the chosen study design (feasibility study), the quantitative data can only provide first hints about the impact the DESKK programme may have on function and may have a high risk of bias. Tables 6 and 7 are primarily displayed to show the functional characteristics of included study participants and changes during the intervention time. There is no intention to prove the effectiveness of the physical exercises but rather to evaluate their feasibility in the RC setting for PwD. Indeed, most of the literature-based chosen exercises were previously used and evaluated positively in terms of their general effectiveness in improving the mobility function abilities of PwD [9, 13, 35].

Nevertheless, the initial summative findings for the DESKK mobility programme are positive, especially related to upper limb functionality and considering the short individual intervention time for most of the included PwD.