Current physical therapy practice and implementation factors regarding the evidence-based ‘Rehabilitation of Mobility after Stroke (ReMoS)’ guideline in Germany: a cross-sectional online survey

The study was designed to evaluate and describe current physical therapy practice in German stroke rehabilitation with regard to the ReMoS guideline and represents a component of the pre-implementation planning [37]. Knowledge of local determining factors will inform the design of implementation strategies [38‐40].

This descriptive cross-sectional study was conducted as a nationwide open online survey addressed at physical therapists working in Germany.

A self-administered web-based questionnaire was developed based on the ReMoS guideline and the ‘Barriers and Facilitators Assessment Instrument (BFAI)’ [32, 41].

A preliminary questionnaire of three sections was drafted, aligned to the recommendations by Dillman et al. [42], and was critically revised by the study team as well as the ReMoS working group of the DGNR. Pre-testing took place with seven physical therapists using the cognitive technique referred to as ‘post-interview probing’ [43], whereupon additional freely formulated response options and further perceived barriers regarding the use of ReMoS recommendations were incorporated. The final online survey was prepared using the browser-based software ‘LimeSurvey’ and consisted of 12 pages (including a welcome page and a closing page, information about the study project and privacy protection, and a page for declaring informed consent) with 89 open and closed questions in a fixed sequence (see Additional file 2 for the survey in German language and Additional file 3 for the translated English version of the questionnaire). Textual information about the purpose, extent, anonymous data storage, and analysis was provided on the welcome screen. The survey questions were structured into three sections. As with earlier studies, participants were asked to rate the implementation of the ReMoS guideline recommendations regarding physical therapy in the subacute phase after stroke on an ordinal scale (always, often, sometimes and never) in the first section [44, 45]. In the second section, the validated BFAI by Peters et al. [41] was translated pragmatically to explore the determining factors of ReMoS guideline usage. The BFAI was developed to identify facilitators for and barriers against the implementation of preventive care, innovations, or guidelines. The original instrument was based on a literature review and an expert panel consensus procedure [41]. In the third and final section, characteristics of participants and their occupational circumstances (e.g. age, gender, years of job experience and occupational location) were collected. The survey extent varied as an adaptive answering procedure was applied to limit the burdens on the study participants. Participants who stated that they were not aware of the ReMoS guideline were forwarded to the final section on demographic characteristics, and in so doing were not asked about barriers and facilitators. Based on the preliminary tests, it was assumed that the questionnaire would take 20 minutes to complete. Apart from the approval of the eligibility criteria, participants always had the opportunity to refuse to answer and were able to navigate within the screens. While the survey was being completed, the level of completeness was presented visually and in numerical form. IP-addresses were not logged and no cookies were used, allowing multiple accesses to the survey via the same internet connection (e.g. from several physical therapists using the same computer). Data was stored and analysed on password-protected servers of the BTU Cottbus-Senftenberg. Intermediate storage on an access-restricted ‘LimeSurvey’ database was immediately deleted once recruitment was completed. Where participants had provided their email addresses voluntarily, these were stored separately from the survey data. The questionnaire was only available in the German language. Due to the scope of the study, an evaluation of the questionnaire validity and reliability was not conducted.

Physical therapists were invited to participate voluntarily with no provision of incentives. For accessing the survey, all participants approved the following self-reported eligibility criteria, without any further checking.

There is no register for physical therapists in Germany. However, it is estimated that about 200,000 people work in this profession [17]. As patients after stroke are treated by physical therapists with a range of treatments and in different health care settings, the precise number of eligible physical therapists is unknown. Since hypotheses were not tested, sample size calculations were not conducted. To achieve representativeness, we aimed to recruit as many eligible physical therapists as possible.

Data were collected between October and December 2019. The link to the browser-based online survey was distributed via 1) e-mail to 248 in-patient rehabilitation clinics specialised in neurorehabilitation in Germany, 2) professional societies (Deutscher Verband fuer Physiotherapie (ZVK) e.V., & Verband Physikalische Therapie – Vereinigung fuer die physiotherapeutischen Berufe (VPT) e.V.), 3), postings on the most accessed online physical therapy forum in Germany (www.​physio.​de), 4) social media accounts of the German Society for Physiotherapy Science (DGPTW e.V.) 5) cooperating partners of the BTU Cottbus-Senftenberg, and directly to professional personal contacts. Any further dissemination beyond these means was also encouraged (e.g. ‘snowball-sampling’). The survey was accessible for 6 weeks. Within that time, two acknowledgements and reminders were sent via the same channels after 2 and 4 weeks. Before access was granted to the survey, participants had to provide their written informed consent.

Data from the online survey were exported to Microsoft Excel and checked for plausibility. Data analysis was conducted with R and R Studio version 3.4.0. All the questionnaires which were answered at least in part (confirmation of the eligibility criteria and response to at least one item) were analysed. We did not correct for any missing data. Approval for further contact and feedback on the questionnaire were not analysed. Participation and completion rates were calculated on the basis of the number of survey accesses, agreements to participate, and the number of participants submitting the final survey page.

Demographic characteristics were analysed based on the item response and data level. Categorical variables are reported as absolute and relative frequencies, while for continuous variables the means and standard deviations are presented.

Guideline adherence can be defined as “the degree of conformity between the knowledge, cognition and/or action of an agent” with guideline recommendations [46]. As was the case in earlier studies, the evaluation of guideline adherence was based on the actions of physical therapists with regard to individual guideline recommendations [27, 47]. Frequencies and percentages were calculated per guideline recommendation, within each domain, and for overall guideline adherence. To evaluate the levels of guideline adherence per ReMoS guideline recommendation, results were dichotomized as “always” and “often” being guideline adherent and “sometimes” and “never” being not guideline adherent. Guideline adherent physical therapy practice was defined as 80% of the responders reporting frequent implementation of guideline recommendations. This level was characterised as being ‘excellent’ by Donohue et al., 2014 [47].

The first author of this paper coded the meaning of the freely written text entries regarding 1) the physiotherapeutic interventions designed to achieve post-stroke mobility-related rehabilitation goals, and 2) the access options to the guideline. The codings were then analysed for their frequencies within the guideline domains and reported descriptively.

Responses to the BFAI questionnaire were analysed and presented for each item as absolute and relative frequencies based on the item response rate. To explore barriers and facilitators to ReMoS guideline use, responses of “fully agree” and “agree” as well as “disagree” und “fully disagree” were merged. Responses to the option “ do not agree nor disagree” were not included within this dichotomization, consistent with the recommendations of the questionnaire developers [41].

Sample characteristics were described descriptively according to the data scale level, where for each item an appropriate point measure as well as a scattering measure are given.

Participants were on average 42 years of age (standard deviation 11.4) years old, mostly female (n = 68, 69.4%), employed in a rehabilitation centre (50, 51.5%) and had no academic degree (n = 57, 60.1%)). Table 1 provides detailed information on participants´ characteristics.

The ReMoS guideline was unknown to 52.9% (n = 55/104) of the responders. Twenty-five persons indicated how they became aware of the guideline and reported academic education (n = 7), online research (n = 6), their workspace (n = 5) and others (n = 14) as their main familiarising sources.

Overall guideline adherence was 34.8%. The highest levels of adherence were reported within the domain ‘Improving the ability to walk in (partially) ambulant patients’, and the lowest levels were reported within the domain ‘Achieving the ability to walk for non-ambulant patients’.

Guideline adherent physical therapy as defined in this study was identified for only intensive walking training for ‘Improving of balance and reducing the risk of falls’ (84.3%, n = 86/102)). The lowest levels of adherence were reported for walking training with functional electric stimulation within the domains ‘Improving of walking distance’ (1.9%, n = 2/108), ‘Improving the ability to walk in (partially) ambulant patients’ (5.5%, n = 8/145) and a combination of end-effector-based devices with functional electric stimulation for ‘Improving of walking speed’ (2.4%, n = 3/124).

Intensive walking training was reported to be used frequently, but not in combination with technical support or functional electrical stimulation as recommended in the ReMoS guideline. Most of the study participants reported using walking aids for ‘Improving the ability to walk in (partially) ambulant patients’ as well as acoustic feedback while walking. Further results are presented in Table 2.

Sixty-seven participants provided more detailed insights into their daily practice by responding to open-ended questions on therapeutic goal-directed interventions. Interventions based on traditional rehabilitation concepts (e.g. Bobath, Vojta, Proprioceptive Neuromuscular Facilitation) were reported to be used often for ‘Achieving the ability to walk in non-ambulant patients’. Exercises on unstable or varying surfaces were reported to be used for ‘Improving of balance and reducing the risk of falls’ (see Additional file 3).

Table 3 displays the reported barriers and facilitators of ReMoS guideline usage. Only participants who reported as being aware of the guideline were directed to this section of the survey. In terms of professional related factors, the majority of the respondents disagreed with the statements on lacking knowledge (n = 32, 84.2%), the ReMoS-guideline is not fitting into daily routines (n = 30, 81.1%), having problems with changing routines (n = 30, 71.5%) or a general reluctance to work according to protocols (n = 30, 71.4%), which may facilitate the ReMoS guideline use. Further, participants acknowledged the guideline being a good starting point for their self-study (n = 23, 57.5%), leaving adequate space to draw their own conclusions (n = 19, 46.4%) and to weigh the wishes of the patients (n = 18, 43.9%). Twenty respondents (60.6%) agreed on a need for financial compensation as barrier when having to work according to the ReMoS guideline.

Of the additional barriers, the availability of end-effector or exoskeleton-based devices and treadmills were relevant factors to 56.8–84.1% of the responders. Only 31.8% (n = 14) and 15.6% (n = 7) of the responders, respectively, felt that they required knowledge of specific recommendations or that they were unaware of the link between the guideline and the phases of neurorehabilitation in Germany. Involvement of other professions was reported to be relevant by 46.7% (n = 21) of the responders.

Our sample is comparable to the general population of physical therapists in Germany in terms of age, sex distribution and working setting, but substantially over represents those physical therapists with an academic qualification (39.4% in our sample compared to approximately 3% nationally) [17, 48].

Although the ReMoS guideline was published 4 years before the survey, many physical therapists were still unaware of it. Since no systematic or theory-based approach was used to disseminate or implement the ReMoS guideline, physical therapists may not have been reached via typical passive dissemination routes [49]. Also, structural limitations may also have been an underlying factor, as Bahns et al. [44] also reported with the lack of awareness of a guideline for non-specific low back pain among German physical therapists. Most physical therapists in Germany graduate from vocational schools, where the training content is based on the 1994 Training and Examination Ordinance for Physiotherapists (PhysTh-AprV). This may result in a lack of conveyance of knowledge about guidelines as well as a limited acquisition of competence for evidence-based practice. Based on current German law, patients with statutory health insurance have a right to care that is consistent with the current state of scientific knowledge (Social Code Book Five - Statutory Health Insurance § 135a). An amendment of the professional laws, which is expected in 2024, might make the full or partial academic training of physiotherapists in Germany permanent. Until then, however, owners of physiotherapy practices, managers of therapeutic facilities, and the physiotherapists themselves should prioritise the financing of evidence-based, practice-oriented, advanced training.

As guidelines represent only one way by which evidence-based knowledge is transferred into practice, adherence to the ReMoS guideline was calculated regardless of whether physical therapists were aware of the guideline or not. In the light of the self-selected participation and the self-reported information, frequencies of adherence for some recommendations can be considered as good to excellent [25, 47]. Because our sample conveniently over-represented professionals with an academic degree, our results are also likely to overestimate the true level of guideline adherence amongst all physical therapists [44].

Similar to our findings and consistent with Donnellan et al. [50], adherence to guidelines seems to vary across recommendations amongst European nurses, speech therapists, and physical therapists involved in acute stroke care and stroke rehabilitation [28, 47, 51‐53]. The varying adherence to the recommendations in our survey might in part be explained by the underrepresentation of other health professions involved in post-stroke mobility rehabilitation. For example, sports therapists can also implement recommendations such as gait training with technical support, and functional electrical stimulation can be performed by masseurs and balneotherapists in Germany. The ReMoS guideline itself does not distinguish responsibilities.

Although Germany differs significantly from other countries regarding the education level for physiotherapists and autonomy in professional practice, this alone cannot explain the restricted adherence to the guidelines. Studies in Europe and worldwide point to the limitations of guideline-based physiotherapy in post-stroke rehabilitation [29, 54, 55].

However, the heterogeneous use of the concept ‘guideline adherence’ and the insufficiently standardised evaluation and operationalisation of the concept should be looked at critically. The resulting restricted comparability of study results has been criticised by experts within the context of physiotherapy research [56, 57]. Varying methods and benchmarks for evaluating guideline adherence can also be seen specifically in the context of stroke rehabilitation. This applies to physiotherapy in stroke research, where as an alternative to our and similar studies using self-reported data, retrospective [27, 28, 58] and prospective audits [25, 29] are used to describe practices related to guideline usage.

Barriers to guideline adherence and evidence-based physiotherapy have been evaluated elsewhere. These include factors related to health care professional characteristics, the context, and the guideline itself [50, 59‐61].

In line with previous studies, we used the BFAI to quantitatively evaluate the influencing factors of guideline use in stroke rehabilitation [51, 62]. In our sample, many factors were found to be similar to those identified by Otterman et al. [51] and van Peppen et al. [62],e.g., physical therapists’ sense of having enough knowledge about the guideline, that the guideline leaves enough room to respect patients’ preferences and allows physical therapists to draw their own conclusions, as well as the agreement regarding financial compensation. In contrast to the previously mentioned studies, participants in our study did not report inadequate time or problems with changing routines.

As reported by Ajimsha et al., the lack of availability of therapeutic equipment can also lead to a limited application of ReMoS recommendations [27]. The authors there report that the lack of therapeutic devices explains adherence rates of 0% regarding electromyographic biofeedback, robotic, and virtual reality training. Less equipment-intensive interventions, such as circuit class training for walking (95%), was on the other hand used more often [27]. This emphasises the value of the current literature where barriers and facilitators have been assessed not in relation to an entire guideline, but rather in relation to individual recommendations [63‐65]. As the authors of the ReMoS guideline recommend several resourceful interventions targeting different rehabilitation goals, a more refined exploration at the level of individual recommendations might also prove useful.

Knowledge about factors influencing guideline usage in stroke rehabilitation usually emerges from multidisciplinary qualitative or mixed-methods approaches, where aspects such as ‘practical familiarity with the recommended treatments’, environmental factors (e.g. time and resources), guideline characteristics (e.g. ‘limited applicability of recommendations’), or organisational or personal factors with respect to patients, relatives and therapists [59, 66‐73] are reported on. Building upon our results, additional approaches such as focus groups [72, 73], semi-structured interviews [67, 69], or both [68], should serve to broaden or further specify aspects of the ReMoS guideline.

Using online-survey methods for this cross-sectional design is afflicted with several limitations.

Sampling was specifically directed towards physiotherapists who work in outpatient and inpatient rehabilitation settings with patients in a subacute phase following a stroke. Therefore, the results of the survey do not represent the entire population of German physiotherapists. Although we attempted to distribute the survey to a large number of physical therapists using an open sampling approach, the number of study participants (n = 170) was small compared to the total number of approximately 200,000 physical therapists working in Germany [17, 74]. However, surveys conducted in Germany also recruited comparatively small samples related to the total number of physical therapists in Germany [44, 75]. As there is no register for physiotherapists in Germany, it is unclear how large the proportion of those working with neurological patients is. In addition, only the perspective of physiotherapists was considered and other professional groups that might be involved in implementing the ReMoS guideline were omitted.

Furthermore, it is unclear how valid the results regarding barriers and facilitators are, since the translation, cultural adaption and validation of the BFAI are all still pending. Since only those physical therapists who were unaware of the ReMoS guideline were directed directly to the final section on participant characteristics, their perception of the barriers against and facilitators for using the ReMoS guideline remains unknown.

By using an open-access online survey and distributing it through multiple channels, as well as sending multiple reminders, we attempted to avoid any sampling bias. At the same time, multiple responding could not be avoided because of the design of the open survey as being without restricted access. Apart from employing a ‘snowball-sampling’ paradigm, additional ways of collecting survey data, such as via postal letters or via phone calls, might have further increased the participation rate.

The survey aimed to capture several aspects of physiotherapy practice as related to the ReMoS guideline. Despite the adaptive answering procedure, the number of questions and the predominantly closed question format may have negatively influenced the flow of participants and the overall response rate.

Because of the self-selected participation and the self-reported data, our results will not provide the whole picture of stroke rehabilitation in Germany. Acquisition bias and social desirability bias due to self-reported answering could well be ameliorated by chart audits or clinical observations [76].

Future studies might employ qualitative approaches to acquire a comprehensive understanding of determining factors, and might also enable the development of multifaceted interventions to implement the ReMoS guideline in a targeted manner. A consistent definition of guideline adherence by health care professionals will lead to an enhanced comparability of the results.

As the survey was not designed to investigate adherence to ReMoS recommendations during the chronic phase after stroke, and as it did not consider the multidisciplinarity of stroke rehabilitation, further investigations across the stroke rehabilitation continuum in Germany are needed. Given the paucity of implementation studies regarding stroke rehabilitation and the fact that the most effective implementation interventions have yet to be discovered, implementation researchers should investigate the feasibility, effectiveness and sustainability of implementation interventions under the specific conditions of the German health care system [77, 78].

In the German context, a secure legal legitimisation of academic physiotherapy education as well as a targeted dissemination of guidelines to users should increase the awareness of these guidelines [79]. Politicians and managers of rehabilitation facilities need to be aware of existing problems and invest in the rehabilitation infrastructure as well as health care professionals to ensure a guideline adherent and thus evidence-based rehabilitation.