Introduction

Lung cancer is associated with significant multisystem disability and impairment [1]. Exercise intolerance occurs due to several confounding factors including the underlying pathological process, effect of surgical/adjuvant treatment, existing comorbidity and a history of sedentary behaviour [1–3]. Patients experience high symptom levels, particularly dyspnoea and fatigue [4]. These exacerbate patient distress and interfere with daily activities [4] and result in a vicious spiral of decline in physical function, cardiovascular fitness and muscle strength [5]. Ramifications for the patient (activity limitations, participation restrictions and diminished health-related quality of life [HRQoL]) and health care system (hospital utilisation) ensue [1, 5, 6].

Exercise is an effective means to improve physical and psychological outcomes in lung cancer [7–10]. The body of evidence has developed rapidly, and results from primary studies and systematic reviews/meta-analyses inform the current international clinical guidelines developed by the American College of Sports Medicine and American Cancer Society [2, 11]. These guidelines send a consistent message to consumers (patients, carers, health care providers [HCPs] and policy makers): Patients with cancer should engage in 150 min of moderate-intensity physical activity (PA) and two to three resistance sessions per week and avoid sedentary time [2, 11]. However, despite the large body of evidence, translation into lung cancer clinical practice and policy is almost non-existent. Exercise is not routinely delivered as part of the lung cancer model of care worldwide, as opposed to other populations such as chronic obstructive pulmonary disease (COPD) where pulmonary rehabilitation is standard practice [8]. This is problematic because patients with lung cancer have significant unmet needs, are inactive and rarely meet PA guidelines [5, 12]. The reason for the failure of evidence translation is not clear. Therefore, the aim of this systematic review was to identify, evaluate and synthesise studies examining the barriers and enablers for patients with lung cancer to participate in PA and thereby inform clinical practice, service delivery, policy and research aiming to enhance PA levels.

Methods

Protocol

The protocol is registered with the International prospective register of systematic reviews PROSPERO (CRD4201603341). The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [13] and Enhanced Transparency of Reporting the Synthesis of Qualitative Research (ENTREQ) framework [14] were followed.

Eligibility criteria

Eligibility criteria are described in Table 1. Studies examining the barriers or enablers to patients with lung cancer participating in PA (any bodily movement produced by skeletal muscles that results in energy expenditure) [15] were included.

Table 1 Eligibility criteria for inclusion of primary studies in the systematic review
Full size table

Information sources and search

Five electronic databases (Fig. 1) were searched by one reviewer (SP) using a pre-planned, systematic, comprehensive and reproducible search strategy to identify all published studies against defined eligibility criteria. Search terms and full search strategy are available in the online supplementary Table E1. Databases were accessed via the University of Melbourne. The last search was run on 1 February 2016. Additional references were identified by cross-referencing reference lists of included articles and hand-searching personal files.

Fig. 1
figure 1

PRISMA flow diagram of study selection process. < less than, CINAHL Cumulative Index to Nursing and Allied Health Literature, EMBASE the Excerpta Medica Database, PA physical activity

Full size image

Study selection

Eligibility assessment was performed in a standardised manner (Fig. 1). Two reviewers (CG, SP) independently screened abstracts and subsequently full-text articles for inclusion against defined eligibility criteria (Table 1). Disagreement was resolved by consensus, with a third reviewer (BC) involved if necessary. At each assessment stage, the agreement between reviewers was estimated with percentage agreement and the kappa statistic using SPSS for Windows statistical software package (SPSS Inc., Version 22.0.0, Chicago, IL). All references were stored in Endnote software 2016 version X7.4.

Data collection process and data items

Data extraction for quantitative studies was performed using a bespoke data collection form by reviewers (BC, KL) and cross-checked by a second (KL, BC). Independent reviewers (CG, SP) manually extracted and cross-checked all text under ‘results/conclusions’ from qualitative studies. To avoid double counting data, multiple reports on the same participant sample were identified by juxtaposing extracted data. Collected data were stored using Microsoft® Office Excel® 2016 software.

Risk of bias in individual studies

Independent reviewers (SP, CG) assessed the risk of bias of studies. Scoring agreement was estimated with percentage agreement and the kappa statistic. Case series were assessed using the Newcastle–Ottawa Quality Assessment Scale (NOS) [16]. Cross-sectional studies were assessed using the Agency for Healthcare Research and Quality Methodology Checklist for Cross-Sectional/Prevalence Studies [17]. Qualitative studies were assessed using the consolidated criteria for reporting qualitative research (COREQ) checklist [18]. Higher scores represent lower risk of bias (quantitative studies) [16, 17] or better comprehensiveness of reporting (qualitative studies) [18]. Results of studies were given the same weight regardless of their assessed risk of bias.

Synthesis of results

Quantitative data were summarised using descriptive statistics and frequencies using SPSS for Windows statistical software package (SPSS Inc., Version 22.0.0, Chicago, IL), and data on barriers and enablers to PA were synthesised in a narrative format. A meta-analysis was not possible due to variability in selection of reported outcome measures. Qualitative data on barriers and enablers were synthesised using thematic synthesis [19]. Two independent reviewers performed line-by-line coding of text from the qualitative studies (SP and CG), and similar concepts were grouped and new codes developed when necessary. Free codes were organised into descriptive major themes and sub-themes using an inductive approach [19]. Qualitative studies were cross-checked to ensure that relevant data were captured and integrated into the themes.

Results were mapped to the ‘capability, opportunity and motivation-behaviour (COM-B) system’ [20]. Capability refers to the capacity to engage in the behaviour; motivation refers to brain processes that energise and direct behaviour, and opportunity refers to factors external to the patient that act as a barrier or enabler to the behaviour [20]. The COM-B system links to the published ‘Behaviour Change Wheel’ [20] which provides a systematic framework to map behavioural analyses to specific interventions and policies aiming. Results of this review were interpreted within the COM-B system (i.e. whether barriers were related to capability, opportunity or motivational issues).

Results

Study selection

The search yielded a total of 2046 studies (Fig. 1). Assessment of title, abstract and full text resulted in 26 papers on 24 unique participant samples being selected for inclusion (Fig. 1). Agreement between reviewers of titles/abstracts was ‘substantial’ (kappa = 0.7, percentage agreement = 93%), and full-text articles were ‘almost perfect’ (kappa = 0.9, percentage agreement = 95%). Consensus was achieved on 100% of occasions when reviewers disagreed. Arbitration from a third reviewer was not required.

Study characteristics

This review included 4 prospective case series [21–24], 9 cross-sectional studies [25–35] and 11 qualitative studies (Table 2) [36–46]. Studies were conducted in eight different countries; the most common were USA (25% of studies), Canada (21%) and UK (17%) (Table 2). Most studies (79%) included only patients with lung cancer. Overall, 1074 patients with lung cancer, 169 HCPs and 23 carers were included in the review. Half of the studies included patients after treatment, and other studies included patients during (18%) or before (9%) treatment or at variable times (14%). The majority (73%) of qualitative studies used semi-structured interviews to collect data on barriers and enablers (online supplementary Table E2). All quantitative studies used questionnaires to collect data on barriers and enablers; however, there was not one consistent questionnaire used, and the majority (62%) used their own self-designed survey (online supplementary Table E2).

Table 2 Characteristics of primary studies
Full size table

Risk of bias in studies

Agreement between reviewers for scoring was almost perfect (kappa = 0.9, percentage agreement = 97%). Consensus was achieved on all occasions when reviewers disagreed. Risk of bias was predominately attributed to be due to lack of comparability for case series and inadequate reporting of the recruitment time period or handling of missing data for cross-sectional studies. Qualitative studies scored poorly for lack of reporting of the interviewers’ characteristics and the relationship between interviewers and participants. The mean (SD) risk of bias score for case series was 4.2 (1.0) out of 8 on the NOS [16], for cross-sectional studies was 8.2 (1.4) out of 11 on the Agency for Healthcare Research and Quality Methodology Checklist for Cross-Sectional/Prevalence Studies [17] and for qualitative studies was 16.4 (2.6) out of 32 on the COREQ checklist [18] showing that the overall risk of bias was moderate to high (online supplementary Table E3).

Synthesis of results

Six major themes and 18 sub-themes for barriers and enablers to PA were identified across the 24 studies (Table 3, Fig. 2). Patient quotes from primary qualitative studies are provided to reflect themes.

Table 3 Summary of results from studies on barriers (−) and enablers (+) to PA
Full size table
Fig. 2
figure 2

Results: factors (barriers and enablers) influencing physical activity in lung cancer mapped using the COM-B model [20]. COM-B ‘capability, opportunity and motivation-behaviour’ system, HCP health care provider, PA physical activity

Full size image

Theme 1: Patient motivation and beliefs

The anticipated impact of PA was reported by patients [25, 34, 36, 37, 39, 40, 42, 44–46], and this was predominately seen as an enabler and motivator for PA. Anticipated benefits included improvements in physical health, facilitating return to work and being able to fight the disease. Other patient beliefs were that PA would help gain control over the disease, be a distraction, relieve stress, improve HRQoL and provide social benefits especially from exercising in a group.

“You hope that by exercising, you can improve physically and your quality of life, and also maybe prevent other illnesses” [36]

The expected impact of PA on symptoms varied with anticipated improvements [25, 42, 45] and harm [37, 44, 46] reported by patients. For the latter, patients expressed fear and avoidance of PA in the belief that it could worsen their symptoms, which may then not resolve.

Patients experienced positive outcomes from PA, and this was seen as an enabler to continue PA [23–25, 36, 40, 44, 45]. They reported improvements in physical health, psychological health and symptoms with PA, as well as functional deterioration following cessation of PA.

“The Wii is – is a fantastic machine. I’m still using it, and I have almost 213 hours on it now. I look forward to using it in the morning because it gets endorphins in the brain going, and it starts me for the rest of the day” [40]

Theme 2: Patient pre-diagnosis PA habits and perceived relevance

Previous PA history strongly influenced patients’ current participation in PA [25, 36, 37, 41, 42, 44, 45]. Patients who had exercised previously were more motivated to exercise after diagnosis. Patients described themselves, as being ‘exercisers’ or ‘non-exercisers’, and the non-exercisers preferred to do usual activities for PA, such as walking, as opposed to structured exercise. The experience of physical discomfort following PA (as a result of being unaccustomed to PA) was reported as a barrier and hindered adherence. Some patients recognised cancer as an opportunity for behaviour change and a trigger to start exercising; others expressed lack of interest and were not willing to try something new. Perceived relevance was identified [37, 43] as an important factor for enabling PA [37, 43], and unless benefits were thought to be achieved in the short term, patients felt that they were unlikely to participate in PA especially given their poor prognosis.

“I’ve never exercised before and I have never wanted to” [45]

Theme 3: Physical influences

Symptoms were reported as barriers to PA [23, 25, 35–37, 39, 40, 44, 45] and resulted in less PA, poor adherence and withdrawal from programs. Fatigue was the most commonly cited symptom, and symptoms during chemotherapy were a common barrier (Table 3). Medical complications including post-operative complications delaying mobilisation and hospital readmissions were barriers to PA [23, 34, 36, 37, 39–42, 45]. In six studies [25, 34, 37, 42–44], comorbidities were noted as a barrier to PA. Older age, lower education and smoking history were seen in patients who were less active or less motivated to be active/change behaviour [31, 34].

“I have had my knees replaced. I have got arthritis, also. It is hard to blame this [on] lung cancer or chemo or being 66 years old. I mean it is some of all three, I’m sure” [44]

Theme 4: Psychological influences

Psychological factors were identified by patients as barriers to PA [25, 29, 31, 37, 40, 42, 46]. These included fear of exercise, lack of confidence, depression, anxiety, feeling overwhelmed, difficulty adapting to diagnosis and introversion. These factors were especially noted to be barriers at time of diagnosis and during treatment.

“I don’t trust my body. Maybe at the last minute I will be paralysed by anxiety or I will be bothered by persistent constipation or a bronchial infection” [37]

Theme 5: Social influences

Carers and relatives were noted as a source of motivation and encouragement for PA [23, 25, 34, 37, 43]. Patients relied on caregivers to recall information. However, in two studies [37, 43], carers and relatives were seen as a barrier. Some patients reported that carers/relatives believed that they were not able to be active and that PA would be harmful. Patients identified that encouragement of PA by HCPs was a key enabler and lack of it was a large barrier [24, 25, 37, 44].

“When you have a prognosis of six months, it is hard to motivate yourself to get in shape, knowing that at some point the disease will come…When you are sick, you lose a lot of psychic autonomy and your ability for decision making…However, if there is a professional PA teacher to push you, there is no need for questioning or thinking about this issue.” [37]

Theme 6: Environmental and structural influences

In 15 studies [21, 23–28, 30, 36–38, 42–45], patients reported preferences for PA, especially relevant for delivery of exercise programs, which included the venue of exercise, supervision, structure, schedule and exercise type. These factors were noted to either promote or hinder adherence to PA and are described in detail in Table 3. There were conflicting results regarding whether exercising at home was preferred. Preference for home-based exercise ranged from 20 to 100% of surveyed patients across the five studies [24, 25, 27, 28, 30]. Two studies [42, 43] highlighted that home-based exercise was an enabler in that it was more convenient, less burdensome (no travel/parking required) and more personal. However, a third study [45] identified home-based exercise as a barrier in that it required the patient to be self-motivated to perform the exercises. Social engagement and supervision were highly sought after by patients and a benefit of hospital-based exercise. Patients felt that supervision provided a sense of security, leadership and means to be monitored/progressed. Transportation was a barrier [23, 25, 39, 42] particularly when patients were medically unable to drive.

“They wouldn’t let me drive. I didn’t have any way to get to town” [42]

Exercising in a group compared to alone was explored in seven studies [25, 27, 28, 30, 36, 43, 45] with conflicting results. However, generally, patients viewed group classes as positive in that it promoted social engagement and allowed them to be supervised (preference ranged from 19 to 27% of surveyed patients).

“Training in a group is much better because you know that someone is keeping an eye out whether you come or not. It’s a little more difficult when you have to do it on your own.” “We inspire each other to come and it’s nice to be expected and a part of the group” [45]

The type and format of PA preference were investigated in 10 studies [24, 25, 27, 28, 30, 37, 42–45]. Consistently walking was the preferred type of PA (42–89% of surveyed patients), and patients also wanted a variety of exercises to prevent boredom (56–61%). Moderate-intensity exercise was the most preferred intensity (44–62%). Overall, patients reported that they wanted an individualised program, with achievable/meaningful goals, and centred on symptom management and functional independence. The ability to adapt the program based on current symptoms and comorbidities was also important. Convenience of exercise (being able to fit into daily routine) was an enabler, and lack of it was a barrier. Time required to exercise was a barrier to PA [36, 42, 45]. Weather was also a barrier [25, 37, 44], and patients had low resilience to adapt in bad weather.

“I was real afraid of the effect that winter was going to have on the lungs. The cold air did affect the breathing a lot when I went out. So, yes, I was quite afraid, especially of exercise. Because you just didn’t know.” [44]

There were diverse results with regards to patient preferences for when to commence a PA program and when to receive PA advice [25–28, 30, 42, 43]. Overall, the preferences were highest before treatment for both participation in a program (26–68%) and for receiving advice (14–68%). However, in two studies [42, 43], patients highlighted that they would not like receiving advice or to perform PA before or during treatment because they felt overwhelmed and incapable of taking on new information. Patients who received surgery preferred to start PA during adjuvant treatment, whereas patients who did not have surgery preferred to start after completion of treatment. Patients strongly relied on verbal information [27, 28, 30] predominately from their physician on PA [26, 28, 30, 39, 44]. Whilst patients were receptive to receiving PA information (23–85%) [27, 28, 30], few had received any advice. Patients were confused about how hard to exercise. Contradictory advice from different HCPs (across or within professions) was a frustration [36, 43]. Data from HCP (as opposed to patients) [32, 33] demonstrated a lack of understanding of the potential benefits of PA from the multidisciplinary team, lack of knowledge of services and lack of services as barriers to PA. Telerehabilitation was identified by HCPs as a potential strategy for provision of information [38].

Discussion

This review has identified the barriers and enablers to patients with lung cancer participating in PA, the evidence of which predominately comes from studies including patients themselves rather than from HCPs or carers. Lack of PA is a significant issue in lung cancer [5, 12], and limited evidence translation into clinical practice necessitates the development of new strategies and models of care aiming to enhance PA. Importantly, this review has identified a number of potentially modifiable patient-reported barriers and a variety of enablers, which should be targeted and utilised to inform future clinical practice. Implementing evidence into practice is a highly challenging task, and one strategy that has been developed to facilitate this is the COM-B model [20] and associated Behavioural Change Wheel [20]. Our results can be interpreted within the COM-B model. For example, capability-related barriers to PA included symptoms and comorbidities, opportunity-related barriers included lack of access to services, and motivation-related barriers included fear of PA and perceived relevance. In contrast, enablers for PA mapped to these domains included personal preferences for PA, encouragement from carers and HCPs (opportunity) and anticipated benefits from PA (motivation). To change practice or behaviour, one or multiple aspects of the COM-B model can be targeted with interventions (in research or clinical practice). Based on this framework and our results, we recommend that potential interventions to target in lung cancer are education (of patients, carers and HCPs), persuasion (of HCPs and patients), training (HCPs), environmental restructuring, modelling and enablement [20]. These strategies could be used to inform the development of new interventions attempting to reduce barriers and take advantage of enablers, and thus increase PA in lung cancer. Such an intervention could target a number of these factors at once. For example, a multidimensional interventional program which could be tested in the future may be one that educates HCPs working in the area of lung cancer about PA through an online training program, that, in turn, facilitates HCPs to educate their patients/carers about PA, combined with a system of integrated online prompts during patient consults for the HCP to ask their patients about their current PA levels.

Results show clear messages from patients about their desire for individualised care. This is evident across a number of aspects spanning from the delivery of exercise programs, type of PA and timing relative to treatment. Whilst there were a small number of trends found in patient preferences, for example, patients predominately preferred walking as their main type of PA, overall there was inconsistency within and across primary studies. We hypothesise that this is due to the diversity of patient personal preferences, which is also evident within the general population. Therefore, in attempt to enable patients with lung cancer to become and stay active, flexibility needs to be added into the delivery of health care services. An appealing model is one where patients could select from a variety of options, including when, where and how they will exercise. Individualised care is also highly important given the elderly age and high incidence of multimorbidities especially COPD in this population. In COPD, pulmonary rehabilitation was traditionally implemented in a standardised format, with little variability across institutions or countries. There is now a recent shift towards more individualised care in COPD, especially given the highly complexity of patients and impact of multimorbidities. Telerehabilitation is being actively investigated in COPD pulmonary rehabilitation [47] and may be a potential strategy to facilitate patients with lung cancer to exercise as well, particularly for those who have barriers to access services. In the COPD population, there are a number of different pulmonary rehabilitation models in use or being researched at present in attempt to improve PA levels. These include the use of smart phone applications to stimulate PA [48], communication with therapist via the internet for motivation [48] and supervised home exercise training with real-time remote videoconferencing [49]. These models could be applied to the lung cancer population as well. Flexibility is easier to implement in clinical practice, but challenging to use in a RCT design; however, studies investigating the efficacy of this model in lung cancer would be highly beneficial. Potentially, a stepped care model may be appropriate in lung cancer, where patients could be offered an escalating number of treatment options depending on their own need at that point in time.

This review found that patients preferred to receive information about PA initially from their treating doctor and prior to treatment commencing, reflecting the perceived high importance patients place on their primary doctor for their overall cancer management and is not a new finding for the cancer literature. Capitalising on the opportunity of lung cancer doctors to discuss and recommend PA during patient consultations represents a powerful strategy to directly influence behaviour. Patients deemed to be inactive could subsequently be referred to a physiotherapist, kinesiologist or accredited exercise specialist for a detailed assessment and treatment plan. In order to optimise success of this approach, efforts to enhance the education and training of the multidisciplinary team regarding PA, to address the barrier of perceived lack of knowledge on the part of HCPs, would be required. Symptom modification from PA was also evident as both a barrier and enabler. We found that many patients expected positive outcomes from PA including reduction in symptoms, yet others feared and avoided PA due to the belief that PA may worsen symptoms. In addition, carers often blocked PA with the thought that it may be harmful. Evidence from systematic reviews shows that exercise is associated with strong and consistent improvements in fitness, muscle strength, HRQoL and symptoms in cancer [2, 11], and therefore, this is an important message to be relayed to patients and carers by all HCPs as patients move along the care pathway. Reinforcement of this message by the entire multidisciplinary team is important.

Our results are not dissimilar to findings from other cancer populations [50, 51]. In breast cancer, factors influencing PA are psychological (motivation, fear, dislike), physical (ageing, cancer treatment, comorbidities, fatigue, weight gain) and contextual and environment barriers [50]. Similarly, in prostate cancer, factors include clinician and spouse influences, comorbidities, fatigue and lack of advice [51]. This suggests that a larger approach may be needed to effectively implement exercise into the model of cancer care, more broadly that for just the lung cancer population.

This review is strengthened by inclusion of qualitative and quantitative data [14]. In particular, the ‘patient voice’ has featured prominently in this review. However, there is a clear gap in the data from the limited number of studies including HCPs. These studies found important barriers, and further research is required to explore these issues. We followed a robust protocol, review guidelines [13, 14] and incorporated duplicate screening, data extraction and risk of bias assessment to enhance review rigour. Only one study was excluded through non-publication in English. All studies were included regardless of risk of bias, and therefore, results from primary studies should be taken with caution. The study is also limited by the lack of randomised trials and overall low level of evidence of included primary studies; however, these are clear gaps in the literature which should be addressed with future research.

This systematic review has identified the volume of literature reporting on barriers and enablers to physical activity in lung cancer. Barriers and enablers are multidimensional and span diverse factors. Patient barriers include physical capability (symptoms, comorbidities), psychological influences, previous sedentary lifestyle and perceived relevance. Enablers include anticipated benefits of PA, opportunity for behaviour change and social influences from HCPs and carers. Patients strongly desire individualised PA programs. Considering these factors in a structured behavioural change framework has elucidated potential mechanisms for enhancing interventions and clinical services to increase PA in patients with lung cancer.

COM-B system, capability, opportunity, motivation-behaviour; COPD, chronic obstructive pulmonary disease; COREQ, consolidated criteria for reporting qualitative research; ENTREQ, Enhanced Transparency of Reporting the Synthesis of Qualitative Research; HCPs, health care providers; HRQoL, health-related quality of life; NOS, Newcastle–Ottawa Quality Assessment Scale; PA, physical activity; PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses; PROSPERO, international prospective register of systematic reviews