Barriers and enablers to physical activity in patients during hospital stay: a scoping review

A scoping review was performed to explore the nature and quantity of published literature on barriers and enablers to physical activity during a hospital stay for acute care, as perceived by hospitalized patients and their HCPs. We used the scoping review methodology suggested by Arksey and O’Malley [26] and developed further by Levac, Colquhoun, and O’Brien [27, 28]. The Joanna Briggs Institute (JBI) guidance document for the conduct of scoping reviews and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Review (PRISMA-ScR) were used to inform the methodology (Additional file 1) [29, 30]. The TDF was used to categorize the barriers and enablers extracted from the included studies [25], as described in further detail in “Collating, summarizing, and reporting the results”. No protocol was registered for this review.

A comprehensive search strategy was developed in collaboration with an experienced research librarian (FvE) of the University of Amsterdam (Additional file 2). An electronic database search of MEDLINE (through Pubmed), CINAHL Plus, Cochrane, EMBASE, PsycINFO, and the Cochrane library was performed, from the inception of the electronic databases to September 23, 2020.

All electronic database searches were combined and de-duplicated in Endnote version X9.1 (Clarivate Analytics, Philadelphia, Pennsylvania, USA) [31]. Two reviewers (SJGG and HCvDH) independently screened all titles and abstracts to determine eligibility, based on the following inclusion and exclusion criteria. Studies were considered eligible if they reported barriers or enablers to physical activity during a hospital stay as perceived by patients or HCPs. Patients had to be hospitalized in an acute care setting and HCPs had to be involved in clinical care (e.g., physicians, nurses, nursing assistants, occupational therapists, and physiotherapists). Barriers were defined as any factor reducing or negatively affecting a patient’s engagement in physical activity. Enablers were defined as any factor enhancing or positively affecting a patient’s engagement in physical activity. Barriers and enablers had to be self-reported. Studies reporting factors associated or correlated to physical activity during the hospital stay for acute care that was not self-reported were not included in this study [32]. Published full-text articles using quantitative, qualitative, or mixed-method study designs were considered, as was gray literature (i.e., academic papers, theses, and dissertations). Only studies written in English or Dutch were included. Studies reporting solely on children (<18 years), short-stay admissions (<24 h), the Intensive Care Unit, or psychiatric ward were excluded because of the differences in care and context (e.g., in terms of organization of care, length of hospital stay, patient characteristics, and care provided). Protocols and reviews were excluded as they lack empirical data. Case studies were also excluded as they often describe extreme cases that do not represent the general population of hospitalized patients. Lastly, conference abstracts were excluded.

To ensure that at least 80% agreement was reached between the reviewers in determining eligibility based on study titles and abstracts, a pilot was performed using 5% of the references. The pilot resulted in minor revisions of the inclusion and exclusion criteria, to enhance the clarity of the criterion descriptors. Full-text articles were obtained when studies fulfilled the criteria or when additional information was needed to determine eligibility. Subsequently, full-text articles were independently screened by both reviewers to determine eligibility. To ensure that at least 80% agreement was reached between reviewers in determining eligibility based on full texts, a pilot was first performed using 10% of the references.

To reduce the risk of missing relevant studies, reference lists of included studies and the reviewers’ own literature databases were screened for additional studies. Any disagreements during the study selection process were resolved by discussion, mediated by a senior researcher (AFL). The web application of Rayyan QCRI (Qatar Computing Research Institute, Hamad Bin Khalifa University) was used to facilitate the study selection process [33]. A PRISMA-ScR flowchart was created to track the screening and inclusion process of this review [30, 33].

Both reviewers (SJGG and HCvDH) independently extracted data using a custom-built data extraction form. Characteristics of included studies (author(s), year of publication, type of study, study aim, method, population, setting, and study sample) were extracted according to the JBI Guidance document for the conduct of scoping reviews [29]. Barriers and enablers identified in the results sections of the included studies were extracted using an iterative data extraction process. Barriers and enablers reported by patients and HCPs were extracted separately. Different extraction methods were used for qualitative and quantitative studies [34]. From qualitative studies, all barriers and enablers reported by patients or HCPs were extracted. For quantitative studies, the approach described by Weatherson [35] was used, meaning that barriers and enablers were extracted if ≥50% of participants agreed that the factor influenced patients’ physical activity behavior. For example, in a survey with dichotomous answering options (agree/disagree), the factor “discussing physical activity during physician rounds increases patients’ physical activity levels” was not extracted as an enabler if 42% of the HCPs agreed. Some questionnaire measures contained an intermediate category, such as 5-point Likert-scale questions with answering options: 1 = strongly agree, 2 = somewhat agree, 3 = neither agree nor disagree, 4 = somewhat disagree, and 5 = strongly disagree. Barriers or enablers were only extracted if at least 50% of participants somewhat agreed or strongly agreed that they perceived it as a barrier or enabler [35]. For example, if 60% of the HCPs agreed (18% somewhat agreed and 42% strongly agreed) that “discussing physical activity during physician rounds increases patients’ physical activity levels” was an enabler, this factor was extracted as enabler [35]. If a quantitative study included open-ended questions, the responses were extracted as in qualitative studies.

To ensure the reliability of the data extraction process, the reviewers first extracted data from five randomly selected articles [14‐16, 19, 36] and discussed their findings to resolve disagreements and improve the preliminary data extraction table. This process was then repeated with five other articles [17, 21, 37‐39], after which both researchers agreed on the data extraction and no further changes to the data extraction table were required. Finally, each reviewer independently extracted half of the remaining articles and then critically reviewed the extraction of the other half performed by the other reviewer. Disagreements were resolved by discussion and rereading source material, and two senior researchers were consulted in case of discrepancies (AFL and MvdS).

Both reviewers (SJGG and HCvDH) independently coded the extracted barriers and enablers and categorized them into the 14 TDF domains [25, 40, 41]. The theoretical definitions and component constructs of the domains as presented in Additional file 3 were used to guide the coding process. Barriers and enablers were coded separately for patients and HCPs and were coded to more than one domain if the content suited multiple domains. To increase inter-coder reliability, the two reviewers (SJGG and HCvDH) met to discuss coding discrepancies and to iteratively modify the coding structure after every ten articles. Discrepancies were solved by discussion and rereading the articles. If necessary, a senior researcher (MvdS) was consulted to discuss and resolve discrepancies. This process was repeated until a final TDF categorization had been obtained. Two senior researchers (AFL and MvdS) supervised the categorization process. The entire authorship team reviewed the final categorization. MAXQDA Analytics Plus 2020 (VERBI Software, 2018, Berlin, Germany) was used to facilitate data coding and the categorization process. The numbers of different barriers and enablers assigned per TDF domain as well as the number of articles reporting on barriers and enablers per TDF domain were presented separately for patients and HCPs. Finally, a descriptive summary of the reported barriers and enablers was composed for patients and HCPs.

Additional file 4 presents an overview of the included studies. Of the 56 studies, 32 used a qualitative study design [14‐20, 38, 42, 44‐46, 49‐54, 57, 61, 67, 68, 70, 72‐76, 78‐80, 83], seven a quantitative study design [21, 37, 39, 63, 66, 69, 82], and 17 a mixed-methods study design [11, 36, 43, 47, 48, 55, 56, 58‐60, 62, 64, 65, 71, 77, 81, 84]. Nineteen studies reported barriers and enablers as perceived by patients [14, 20, 36, 37, 39, 48, 51, 61, 63‐65, 67, 70, 73, 78, 79, 81, 83], 23 reported those perceived by HCPs [16‐18, 21, 38, 42, 43, 47, 50, 52‐55, 57, 59, 60, 68, 69, 74‐76, 80, 82], and 14 reported those perceived by patients and HCPs [11, 15, 19, 44‐46, 49, 56, 58, 62, 66, 71, 72, 77, 84]. Sample sizes varied between n = 6 and n = 345 patients and between n = 5 and n = 261 HCPs. Two studies did not specify the sample size [11, 77], and one study only specified the number of included sites [47]. Further descriptions of the populations and settings included are provided in Additional file 4. The included studies were published between 2003 and 2020, and only seven studies were published before 2010 [15, 52, 56, 75, 77, 78, 82].

The results of the data extraction process are presented in Additional file 5. After coding and discussing all extracted fragments containing barriers and enablers, SJGG and HCvDH reached a consensus on 1316 barriers and enablers. Two hundred sixty-four (20.2%) patient-reported barriers and 415 (31.7%) HCP-reported barriers were coded. Two hundred twenty-eight (17.3%) patient-reported enablers and 409 (31.2%) HCP-reported enablers were coded.

Patient- and HCP-reported barriers were assigned to 13 of the 14 TDF domains. An overview of the TDF coding of all barriers is provided in Additional file 6 and summarized in Fig. 2. Patient-reported barriers were assigned most frequently to the TDF domains ECR (n = 148, 56.1%), Social Influences (n = 32, 12.1%), and Beliefs about Consequences (n = 25, 9.5%). Of the other 11 domains, the largest numbers of barriers were assigned to the domains Emotion (n = 16, 6.1%) and SPRI (n = 10, 3.8%). HCP-reported barriers were assigned most frequently to the TDF domains ECR (n = 210, 50.6%), MADP (n = 45, 10.8%), and SPRI (n = 31, 7.5%). Of the other 11 domains, the largest numbers of barriers were assigned to the domains Beliefs about Consequences (n = 27, 6.5%) and Emotion (n = 22, 5.3%). No patient- and HCP-reported barriers were assigned to the domain Optimism.

The TDF domains to which barriers were most frequently assigned are highlighted below. The domain ECR had the majority of both patient- and HCP-reported barriers assigned to it and covered four main topics: (1) patient-related factors (e.g., medical factors, age, language barriers), (2) care processes and organizational characteristics (e.g., prescribed immobility, communication, hospital culture, bed-centered care), (3) physical environment of the hospital (e.g., room, unit, hospital), and (4) resources (e.g., limited time, staffing, equipment) (Additional file 6). Patient-reported barriers assigned to the domain Social Influences included interpersonal processes between patients, visitors, and HCPs that negatively influence physical activity, such as lack of encouragement and assistance and providing more care than necessary. Patient-reported barriers assigned to the domain Beliefs about Consequences included the belief that physical activity results in negative consequences (e.g., injuries, falling, or missing meals and care), the belief that rest is needed for recovery, and the belief that patients may be inconveniencing busy HCPs. Most of the HCP-reported barriers assigned to the domain MADP related to prioritization. A high workload and safety considerations resulted in physical activity receiving a lower priority than medical treatment or rest. HCP-reported barriers assigned to the domain SPRI included the passive and dependent attitude patients adopt during hospitalization (e.g., the idea that patients should remain in bed, personality, and character traits). In addition, HCPs mentioned the role they fulfill regarding physical activity (e.g., lack of role clarity in improving physical activity, attributing responsibility to others, and nurses lacking autonomy in deciding how and when to mobilize patients).

Patient- and HCP-reported enablers were assigned to 11 and 13 of the 14 TDF domains, respectively. An overview of the TDF-coding of all enablers is provided in Additional file 7 and summarized in Fig. 3. Patient-reported enablers were most frequently assigned to the TDF domains ECR (n = 67, 30.2%), Social Influences (n = 54, 24.3%), and Goals (n = 32, 14.4%). Of the remaining 11 domains, the largest numbers of enablers were assigned to the domains Knowledge (n = 24, 10.5%) and Beliefs about Consequences (n = 17, 7.7%). No patient-reported enablers were assigned to the domains Reinforcement, MADP, and Emotion. HCP-reported enablers were most frequently assigned to the TDF domains ECR (n = 143, 35.0%), Social Influences (n = 76, 18.6%), and Behavioral Regulation (n = 54, 13.2%). Of the remaining 11 domains, the largest numbers of enablers were assigned to the domains SPRI (n = 45, 11%) and Knowledge (n = 19, 4.7%). No HCP-reported enablers were assigned to the domain Optimism.

In line with the categorization of the barriers, most patient- and HCP-reported enablers were assigned to the domain ECR and covered the same four main topics: (1) patient-related factors, (2) care processes and organizational characteristics, (3) physical environment of the hospital, and (4) resources (Additional file 7). Patient- and HCP-reported enablers assigned to the domain Social Influences included interpersonal processes between patients and visitors or HCPs that positively influence physical activity, such as being encouraged and assisted. Patients also described that other patients motivated them to perform more physical activity, while HCPs described how leadership and multidisciplinary collaboration enabled them to improve patients’ physical activity. Patient-reported enablers assigned to the domain Goals included the importance of having a goal (e.g., experiencing the positive effects of physical activity or preventing the negative effects of physical inactivity). This domain also included the importance of having autonomy and being involved in physical activity-related decision-making. HCP-reported enablers assigned to the domain Behavioural Regulation included strategies aimed at regulating behavior, such as providing education, appointing mobility champions, making performance and expectations visible, creating a habit, and using mobility documentation tools, reminders, daily schedules, exercise programs, and mobility audits.

This is the first scoping review on patient- and HCP-reported barriers and enablers relating to physical activity during the hospital stay for acute care using a TDF analysis. A strength of this study is that it was designed and conducted according to the systematic scoping review methodology and that it followed the PRISMA-ScR statement recommendations [26‐30]. Secondly, almost all aspects of data collection, data extraction, and data analysis were carried out independently by two researchers, with a third party available in case of disagreements. Thirdly, given the extensive, thorough search strategy in multiple databases, along with the inclusion of quantitative, qualitative as well as mixed-methods study designs, we were able to present a complete overview of all barriers and enablers reported in the current literature. Fourthly, an additional strength of this study is the use of the TDF as a theoretical framework to categorize barriers and enablers. The use of the TDF ensured that the reviewers assessed barriers and enablers from a broad perspective, thereby also exploring underexposed domains.

We also recognize some limitations. While the use of the TDF facilitates reviewers in exploring barriers and enablers from a broad perspective, it does not provide an explanation as to how barriers and enablers are connected and influence one another. Another limitation of this study is that barriers and enablers are presented based on the number of articles in which they have been reported. As the frequency of reporting is primarily a function of the methods used to present the data, this alone should not be used as a proxy of importance. In other words, a barrier that has only been reported once may be just as relevant as one that has been reported many times. Furthermore, a secondary analysis of differences in perceived barriers and enablers among patient subgroups or among professions could not be performed due to the lack of detailed reporting in the included studies. Lastly, as this was a scoping review, no quality appraisal of included articles was performed [30].

Our findings provide a comprehensive overview of barriers and enablers to physical activity during a hospital stay for acute care. The large number of barriers and enablers we found, distributed across many TDF domains, highlight the complexity of physical activity behavior during the hospital stay and the need for tailored interventions. A context-based assessment should be performed to determine which barriers and enablers can be targeted in a specific clinical setting. Our comprehensive overview will enable clinicians and researchers to perform this context-based assessment from a broad perspective and support them in establishing a behavioral diagnosis of what needs to change in a specific context in order to improve physical activity behavior during the hospital stay.

Subsequently, clinicians and researchers will be able to link relevant barriers and enablers to specific intervention strategies and behavior change techniques (BCTs) [25, 41, 94]. An example of a framework that could be used to assist clinicians and researchers in selecting appropriate BCTs is the Behaviour Change Wheel (BCW) [22]. Our TDF-based overview provides the initial step in developing and implementing theory-informed behavior change interventions aimed at improving physical activity during the hospital stay [41].

Given the large number of factors influencing the physical activity behavior of hospitalized patients, we recommend that clinicians and researchers develop and implement interventions targeted at multiple barriers and enablers. Previous research suggests that developing and implementing such tailored multimodal interventions may be more effective than unimodal interventions [95]. Moreover, given a large number of barriers and enablers assigned to the Environmental Context and Resources and Social Influences context in our review, we suggest that clinicians and researchers should always consider incorporating intervention strategies targeting these TDF domains in their multimodal interventions.

Future research should focus on exploring relationships between barriers and enablers both within and between TDF domains. Revealing these relationships may facilitate the assessment of barriers and enablers in specific clinical settings and may increase the effectivity of future tailored multimodal interventions. Future research is also needed to explore the differences in perspectives perceived by different patient subgroups (e.g., age, sex, pathologies). Similarly, more research is needed to investigate differences in perceived barriers and enablers among professions and how these differences relate to their role in improving physical activity during the hospital stay. Additionally, further research is needed to develop and validate a TDF-based questionnaire that could facilitate the context-based assessment of barriers and enablers across all TDF domains. Further research is needed to retrospectively identify which barriers and enablers to physical activity during the a hospital stay have been targeted in previously described intervention studies [94], so clinicians may be better able to implement these interventions in other contexts. Finally, there is a need for research assessing the effectiveness of tailored multimodal interventions that target context-based barriers and enablers to physical activity in hospitalized patients.