Understanding participation of liver transplant recipients in an exercise intervention RCT: a cross-sectional study of barriers and motivators

Abstract

Background

Understanding barriers to participate in exercise intervention studies is critical for guiding approaches to increase sample representativeness, to mitigate selection bias, and to achieve the required sample size for study purposes. This study compared physical activity levels and exercise-related perceptions, including barriers and facilitators, between liver transplant recipients who were willing versus not-willing to engage in an exercise intervention randomised controlled trial (RCT).

Methods

A comparative cross-sectional study was conducted using baseline data of the PHOENIX-Liver trial, a RCT evaluating a 6-month home-based exercise intervention followed by a physical activity maintenance program. Two informed consents were obtained from eligible liver transplant recipients (3–5 months post-transplant; >18 years) either willing (n = 64) or not willing (n = 46) to participate. Clinical and sociodemographic characteristics were collected from electronic medical records and questionnaires. Physical activity was assessed objectively (accelerometry) and subjectively (Physical Activity Vital Sign Questionnaire). Exercise-related barriers and facilitators were measured using the Motivators and Barriers Questionnaire.

Results

42% of 110 eligible individuals declined participation, most often due to time constraints (40%), transportation challenges (25%), and lack of motivation (15%). Non-participants were significantly more likely to have lower education (p = 0.009), lower income (p = 0.004), and live farther from the study site (p = 0.004). Self-reported physical activity levels did not differ between groups, but accelerometry showed that non-participants spend less time in moderate-to-vigorous physical activity (p = 0.05) and took fewer daily steps (p = 0.01). Multivariate logistic regression showed trends for greater distance from centre (p = 0.056), lower education (p = 0.082), and lack of motivation/time (p = 0.071) with unwillingness to participate. High income was independently associated with meeting WHO physical activity guidelines (p = 0.049).

Conclusion

A substantial proportion of eligible individuals declined participation. Some sociodemographic and behavioural barriers point to potential selection bias. Addressing these through tailored study designs could enhance engagement, improve representativeness and support the recruitment of representative patient populations in future exercise interventions.

Clinical trial registration number

NCT06302205 (trial registration date: 06/12/2023).

Supplementary Information

The online version contains supplementary material available at 10.1186/s13102-025-01369-y.

Background

With 41,111 procedures performed in 2023, liver transplantation is the second most frequently performed solid organ transplant worldwide, following kidney transplantation [1]. It is a life-saving intervention that restores liver function and significantly improves survival and quality of life in persons with end-stage liver disease. However, despite successful transplantation, long-term survival rates and overall quality of life remain suboptimal compared to the general population [2], largely due to complications such as cardiovascular [3] and metabolic diseases [4]. Low levels of physical fitness and physical activity are important modifiable risk factors contributing to these poor outcomes [5, 6].

Regular physical activity plays a crucial role in long-term recovery following major surgery, including liver transplantation [7, 8]. The World Health Organization (WHO) recommends that individuals with chronic conditions engage in at least 150 min of moderate-intensity or 75 min of vigorous-intensity aerobic activity per week, supplemented by muscle strengthening activities and postural balance exercises at least twice per week [9, 10]. Importantly, even lower levels of physical activities can still yield significant health benefits [11]. In liver transplant recipients, regular physical activity has been shown to improve cardiorespiratory fitness, walking capacity, and overall quality of life, and reduces cardiovascular and all-cause mortality [12]. Despite these well-established benefits, the majority of transplant recipients do not meet recommended physical activity levels [13]. People with end-stage liver disease are among the most sedentary individuals with chronic illness [14]. Although some improvement in physical activity levels is observed after transplantation, activity levels often remain below both WHO physical activity guidelines and those of the general population [6, 13]. The reasons behind these persistently low levels of physical activity remain poorly understood [15].

Moreover, strong evidence from large-scale powered randomized controlled trials (RCTs) evaluating the effectiveness of exercise interventions on physical fitness and long-term physical activity behaviour in liver transplant recipients remain scarce [6, 12]. Recruitment of patients for exercise trials is often challenging, with markedly low participation rates reported in transplant and other clinical populations, typically ranging between 38 and 55% [16, 17]. In the general population, factors such as age, lifestyle, low motivation, and socioeconomic status contribute to reduced levels of physical activity [18], and these same factors may further restrict trial participation, introducing selection bias. In transplant populations [15], as in other chronic illness groups [19, 20], additional disease-related barriers exist, with compromised health status representing a particularly important limiting factor. Therefore, a better understanding of the barriers and facilitators to participation in exercise trials is essential, yet have not been systematically examined in liver transplant recipients. Such insights are crucial for designing exercise studies that can be generalized to the overall transplant population, and for the subsequent successful implementation of exercise in clinical practice.

This sub-study of the PHOENIX-Liver trial aimed to enhance understanding of the trial’s representativeness and to identify barriers and facilitators affecting recruitment of adult liver transplant recipients into a RCT involving a 6-month home-based exercise intervention followed by a physical activity maintenance program. We also explored characteristics associated with meeting WHO physical activity guidelines to improve understanding of broader physical activity engagement in this population. These findings are intended to inform more inclusive and pragmatic recruitment strategies for future exercise and physical activity trials in liver transplant populations, thereby enhancing generalizability and implementation potential in real-world clinical settings.

Methods

Study design

This comparative cross-sectional sub-study is nested within the ongoing PHOENIX-Liver study [21], a randomized controlled trial designed to evaluate the effects of a 6-month home-based exercise and subsequent physical activity intervention in de novo adult liver transplant recipients [21]. The study received ethics approval (B3222022000943) from the local ethical committee of UZ/KU Leuven on December 13, 2022 and adhered to the International Conference for Harmonization of Good Clinical Practice guidelines. The protocol was registered with ClinicalTrials.gov (Registry number: NCT06302205) and the detailed trial protocol has been published elsewhere [21]. While this sub-study is cross-sectional and was conducted and reported in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines [22], the parent trial adheres fully to CONSORT guidelines for randomized controlled trials [23]. Completed STROBE and CONSORT checklists are included as supplementary files with the manuscript.

Participants enrolled in PHOENIX-Liver undergo a comprehensive test battery spread across five visits, each lasting approximately three hours, at University Hospitals Leuven. Assessments include physical fitness and cardiovascular evaluations. The home-based exercise program consists of aerobic training (3 sessions per week, gradually increasing from 25 to 40 min) and strength training (2 sessions per week, 20–30 min per session) and aims to improve the physical fitness of the patients. After completion of the 6-month home-based exercise intervention, participants receive a personalised physical activity maintenance program, developed together with the participant, to support a long-term physically active life. Motivational interviewing techniques are employed to foster intrinsic motivation [15, 21]. In addition, the physical activity program incorporates an array of behaviour change techniques, including education, goal setting and planning, feedback and monitoring, social support, and restructuring of the physical and social environment [21].

Dual informed consent procedure

A key methodological innovation of this sub-study is the use of a dual informed consent procedure, allowing for systematic comparison between participants who agreed to the RCT and those who declined. Eligible liver transplant recipients were first invited to provide full consent (ICF-A) for participation in the PHOENIX-Liver intervention study. If a participant declined ICF-A, they were subsequently invited to provide partial consent (ICF-B) for participation in the comparative cross-sectional ‘Non-Intervention Participant Profiling’ sub-study only. This approach allowed for the profiling of both willing and unwilling participants, offering a unique opportunity to assess potential selection bias in trial recruitment and to gain insight into real-world barriers to participation in post-transplant exercise interventions. All participants provided written informed consent prior to inclusion.

Study population and recruitment

Between March 2023 and March 2025, all adult liver transplant recipients from the University Hospitals of Leuven (Belgium) (n = 162) who were 3–5 months post-transplantation and met the eligibility criteria (n = 110) (see Table 1) were invited to participate in the PHOENIX-Liver study.

Table 1.

Inclusion and exclusion criteria for PHOENIX-Liver RCT

Inclusion criteria
□	Liver, combined liver-kidney, combined liver-intestine, and combined liver-pancreas transplant recipients from the University Hospitals of Leuven
□	Men and women ≥ 18 years of age
□	Transplant vintage of three to five months
□	Individuals receiving a re-transplantation are also considered eligible
□	Written informed consent

Exclusion criteria
□	Positive cardiopulmonary exercise test (electrocardiogram or clinical)
□	Unstable angina
□	Life-threatening arrhythmias
□	Uncontrolled hypertension/diabetes
□	Haemoglobin A1c ≥ 9%
□	Severe pulmonary disease (forced expiratory volume in 1 s < 50%)
□	Musculoskeletal disorders or any other medical reasons by the physician considered a contraindication for moderate or high-intensity physical exercise
□	Combined thoracic-abdominal organ transplantation
□	Ongoing treatment for malignancies
□	Unable to understand Dutch
□	Unable to provide informed consent independently or participate in a training intervention independently due to cognitive dysfunction
□	No access to smartphone and/or computer with internet access
□	Progressive neurodegenerative disorders (e.g., Parkinson’s disease) that act as confounding factors on the primary study outcome (i.e., peak oxygen uptake)

The recruitment strategy involved multiple partners and contact moments to maximize awareness and engagement. Eligible individuals were informed about the study by a member of the research team on at least three separate occasions, beginning at the time of waitlist registration and continuing until 3 to 5 months following transplantation. In addition, various healthcare professionals involved in the patient’s care trajectory, such as the transplant surgeon, rehabilitation physician, and physiotherapist, also introduced and supported the study invitation. To further promote visibility, study flyers were displayed in the hospital department where liver transplant recipients are admitted post-operatively, and a dedicated study website was made available. This integrated approach aimed to optimize informed decision-making and improve participation rates among eligible individuals.

Data collection

Data were collected at baseline, between three to five months post-transplantation, from both participants and non-participants. Self-administered questionnaires were completed either digitally or on paper, in the hospital or at home, according to participant’s preference. Additional study data were collected from electronic health records and from actigraphy measurements. All data were managed in a secure web-based platform (REDCap), ensuring secure and efficient data capture.

Outcome measures

Sociodemographic and clinical characteristics

Sociodemographic data were collected via self-reported questionnaires (Supplement I) and included information on age, sex, religion, social support, housing situation, caregiving responsibilities, children, education level, employment status, distance from the participant’s home to the study location, and monthly income. Smoking behaviour, together with clinical data including weight and the primary cause of liver failure, was obtained from electronic medical records.

Physical activity

Physical activity levels were evaluated subjectively using the Physical Activity Vital Sign questionnaire (Supplement II). It assesses the frequency (days per week) and duration (minutes per week) of moderate-to-vigorous physical activity (MVPA), as well as the number of days per week participants engaged in strength training [24]. In accordance with the WHO guidelines [9], participants achieving ≥ 150 min of MVPA and engaging in strength training on at least two days per week were classified as meeting recommended physical activity levels [9].

In a subset of participants (n = 37) physical activity levels were also evaluated objectively using the ActiGraph wGT3X-BT triaxial accelerometer (ActiGraph LLC, Pensacola, FL, USA), a widely validated tool in physical activity research [25]. At the start of the trial, ethical approval for accelerometer-based assessments had not yet been obtained for participants who provided partial informed consent (ICF-B), restricting the collection of these data. Formal approval to include accelerometer monitoring in this subgroup was granted 15 months after trial initiation (6 June 2024). Consequently, accelerometer data were only available for participants enrolled after this amendment. Participants were instructed to wear the device at the waist for one week, providing data on average daily step count and time spent in MVPA [25]. Raw accelerometer data were processed using ActiLife software (version 6). Non-wear time was determined according to the algorithm by Choi et al. [26], with non-wear defined as ≥ 90 consecutive minutes of zero activity counts, allowing for brief interruptions of 1–2 min with minimal movement (counts between 0 and 100). Data were considered valid if the participant wore the device for at least 600 min (10 h) per day on ≥ 4 days, including at least 1 weekend day and 3 weekdays. Activity intensity was categorized using the vector magnitude cut-points proposed by Sasaki et al. [27]. MVPA was defined as ≥ 2690 counts per minute, with bouts identified using standard bout detection settings (minimum 10-minute duration, allowing for 1–2 min below threshold within the bout). Non-wear time was excluded from all analyses.

To gain insights into the barriers and motivators related to physical activity behaviour, the Motivators and Barriers Questionnaire (Supplement III) was administered. This self-report questionnaire includes 31 barrier items and 23 motivator items, each rated on a 4-point Likert scale ranging from 1 (“not at all”) to 4 (“very much”). While originally developed for individuals undergoing haemodialysis, the questionnaire has recently been used in solid organ transplant recipients, including 146 liver transplant patients [28, 29]. Based on the component analysis by van Adrichem et al. [29], the items are categorized into four barrier domains and four motivator domains. Higher scores on barrier items reflect greater perceived barriers, whereas higher scores on motivator items indicate stronger motivation to engage in physical activity.

Reasons for non-participation were assessed using a single-item questionnaire (Supplement IV) asking: “What is the main reason you are not participating in this study?” Participants could choose from predefined response options, including lack of motivation, lack of transportation, lack of time, fear that the program would be too intense, fear of injury, fatigue, dislike of exercise, a preference not to disclose a reason, or other (with open field).

Statistical analyses

Population characteristics were summarized using means (standard deviations) or medians [interquartile ranges] for continuous variables and frequencies (percentages) for categorical variables. Between-group comparisons were performed using independent Student’s t-test or Mann-Whitney U-tests for continuous variables and Chi-square tests or Fisher’s exact tests for categorical variables. To account for multiple testing across these comparisons, false discovery rate correction was applied using the Benjamini-Hochberg procedure [30]. Both unadjusted p-values and false discovery rate-adjusted q-values are reported. The q-value provides an estimate of the probability that a significant result is a false positive, taking into account that multiple statistical tests are being performed [30]. A p- or q-value < 0.05 was considered statistically significant.

To identify factors associated with willingness to participate in the training intervention RCT, we first performed univariate logistic regression analyses. False discovery rate correction was also applied to these analyses to adjust for multiple comparisons. Variables with a q-value < 0.05 in the univariate analyses were subsequently included into a multivariable logistic regression model. Odds ratios (ORs) with 95% confidence intervals (CIs) are reported; these are referred to as unadjusted ORs for univariate analyses and adjusted ORs (AORs) for multivariable analyses.

In a separate analysis, predictors of meeting the WHO physical activity guidelines [9], defined as engaging in at least 150 min of MVPA per week and performing muscle-strengthening exercises on at least 2 days per week, were explored. Variables with q < 0.20 in the univariate analysis and/or clinical relevance were included in the multivariable model.

Sensitivity analyses were conducted to assess the robustness of findings. Outliers, defined as values exceeding ± 3 standard deviations from the mean, were removed from continuous outcomes, and primary analyses were repeated to check whether the results were influenced by extreme values.

All statistical analyses were performed using R version 4.4.1 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Participant recruitment and eligibility

From the start of recruitment in March 2023 until March 2025, 162 liver transplant recipients at University Hospitals Leuven were screened for eligibility. Of these, 31% (51/162) ineligible for study participation (Fig. 1) due to palliative care policy or patient death (16%, 8/51), cardiovascular contraindications (16%, 8/51) such as unstable angina or life-threating arrhythmias (8%, 4/51), language barriers (12%, 6/51), or age under 18 years (10%, 5/51).

Fig. 1.

Flow chart PHOENIX-Liver study. ICF, informed consent form; LT, liver transplantation; LTRs, liver transplant recipients

Of the 110 eligible liver transplant recipients, 58% (64/110) provided informed consent to participate in the intervention study (ICF-A). A total of 42% (46/110) declined participation in the intervention RCT; of these, 87% (40/46) consented to participate in the “Non-Intervention Participant Profiling” component by signing informed consent B (ICF-B). The remaining 13% (6/46) declined participation in both components of the study.

The most frequently reported reasons for declining participation in the exercise intervention were lack of time (40%), transportation difficulties for the study assessments (25%), and a lack of motivation to exercise (15%). Additional reasons included concerns that the training program would be too intense (13%), already being physical active (13%), fatigue (8%), pain (3%), do not like to sport (3%), and fear of injury (3%). Four participants (10%) chose not to disclose their reasons for non-participation (Fig. 1).

Sociodemographic and clinical differences between participants and non-participants

As shown in Table 2, no significant differences were observed between the group of participants and non-participants regarding sex, age, body mass index, or aetiology of liver failure. However, the proportion of individuals with lower educational levels (5% (3/61) vs. 25% (10/40), q = 0.034), and those reporting a low monthly income (< 1000 EUR/month) (0% (0/57) vs. 13% (5/39), q = 0.020) was significantly higher in the group of non-participants. Similarly, the average distance from the participant’s home to the study site (50 [IQR: 31–88] vs. 72 [IQR: 54–125], q = 0.020) was significantly larger in the group of non-participants.

Table 2.

Socio-demographic and clinical characteristics of willing and unwilling participants

	Total sample	Participants	Non-participants	p-value	q-value
	(n = 104)	(n = 64)	(n = 40)
Age (years)	62 [53–69]	61 [51–68]	64 [55–69]	0.245	0.383
Sex				0.401	0.525
Male	68 (65%)	44 (69%)	24 (60%)
Female	36 (35%)	20 (31%)	16 (40%)
Weight (kg)	72 [63–81]	73 [65–82]	70 [59–81]	0.229	0.374
Body mass index (kg/m²)	24.0 [20.9–27.5]	23.8 [21.3–27.5]	24.0 [20.5–27.6]	0.63	0.693
Home-to-site distance (km)	59 [35–118]	50 [31–88]	72 [54–125]	0.004*	0.020*
Aetiology				0.07	0.14
Lifestyle-related	61 (59%)	33 (52%)	28 (70%)
Non-lifestyle-related	43 (41%)	31 (48%)	12 (30%)
Smoking				0.84	0.875
Non-smoker	59 (57%)	37 (58%)	22 (55%)
Ex-smoker	45 (43%)	27 (42%)	18 (45%)
Religion				0.163	0.301
Atheism	36 (36%)	24 (40%)	12 (30%)
Christianity	56 (56%)	35 (57%)	21 (53%)
Islam	5 (5%)	2 (3%)	3 (8%)
Buddhism	1 (1%)	0 (0%)	1 (3%)
Other	2 (2%)	0 (0%)	2 (5%)
Social support1				0.297	0.439
Yes	98 (96%)	61 (98%)	37 (93%)
No	4 (4%)	1 (2%)	3 (8%)
Household				0.587	0.676
Co-habituating	87 (84%)	52 (81%)	35 (88%)
Single	17 (16%)	12 (19%)	5 (13%)
Caregiving responsibilities				0.827	0.875
Yes, spend a lot/moderate amount of energy caring for others	31 (31%)	18 (30%)	13 (33%)
No, almost not	70 (69%)	43 (70%)	27 (68%)
Children				0.826	0.875
No Children	31 (30%)	14 (22%)	7 (18%)
1–2 children	64 (62%)	38 (59%)	26 (65%)
≥ 3 children	19 (18%)	12 (19%)	7 (18%)
Education level				0.009*	0.034*
No/primary education	13 (13%)	3 (5%)	10 (25%)
Secondary/post-secondary education	49 (49%)	30 (49%)	19 (48%)
Higher (non)-university education	39 (39%)	28 (46%)	11 (28%)
Employment status				0.826	0.875
Not working	73 (70%)	44 (69%)	29 (73%)
Working	31 (30%)	20 (31%)	11 (28%)
Monthly income2				0.004*	0.020*
< 1000 EUR/month	5 (5%)	0	5 (13%)
1000–2500 EUR/month	69 (72%)	40 (70%)	29 (74%)
> 2500 EUR/month	22 (23%)	17 (30%)	5 (13%)

Values reported in median [IQR] or number (percentage). Percentages were calculated based on the available data for each category. Missing data were as follows: religion (willing group: 3 participants (5%); unwilling group: 1 participant (3%)), social support (willing group: 2 participants (3%)), caregiving responsibilities (willing group: 3 participants (5%)), education level (willing group: 3 participants (5%)), and monthly income (willing group: 7 participants (11%); unwilling group: 1 participant (3%))

¹Social support status was assessed with the question: Suppose you need care for a long time, do you have someone who would assist you regularly?

²Monthly income categories refer to net monthly income

Comparison of physical activity levels between participants and non-participants

Based on the self-reported data on physical activity, 34% (34/104) of the liver transplant recipients was meeting the WHO guidelines of at least 150 min of MVPA per week and at least 2 times per week strength exercises. Overall, recipients reported a median of 120 min of MVPA per week [IQR: 60–205], with 41% (41/104) achieving at least 150 min of MVPA. The median frequency of strength training was 3 days per week [IQR: 2–3], and 75% (76/104) reported engaging in strength exercises at least 2 times per week. As shown in Table 3, no statistically significant differences were observed in subjectively reported physical activity levels on the Physical Activity Vital Sign questionnaire between participants and non-participants.

Table 3.

Physical activity levels of participants and non-participants according to the physical activity vital sign questionnaire

	Total sample	Participants	Non-Participants (n = 40)	P-value
	(n = 104)	(n = 64)
Median total minutes of MVPA per week – [IQR]	120 [60–205]	120 [60–240]	105 [60–175]	0.175
< 150 min of MVPA per week - n (%)	60 (59%)	34 (56%)	26 (65%)
≥ 150 min of MVPA per week - n (%)	41 (41%)	27 (44%)	14 (35%)
Median days of strength training per week – [IQR]	3 [2–3]	2 [1–3]	3 [2–4]	0.1
< 2 days per week strength training - n (%)	25 (25%)	17 (28%)	8 (20%)
≥ 2 days per week strength training - n (%)	76 (75%)	44 (72%)	32 (80%)
≥ 150 min of MVPA per week + ≥ 2 days per week strength training - n (%)	34 (34%)	23 (38%)	11 (28%)	0.39

Values reported in medians [IQR] or number (percentage). Percentages were calculated based on the available data. Physical Activity Vital Sign questionnaire data were missing for 3 participants

Abbreviations: IQR, interquartile range; MVPA, Moderate to vigorous physical activity

In a subsample with ActiGraph data (participants [ICF-A], n = 30; non-participants [ICF-B], n = 7), the median daily step count was 3,788 steps/day [IQR: 2,602–6,988], and the median weekly MVPA time was 93 min [IQR: 36–284]. When examined by group, participants had a median daily step count of 4,472 steps/day [IQR: 2,967–8,902] and a median weekly MVPA of 136 min [IQR: 57–342], whereas non-participants had a median daily step count of 2,990 steps/day [IQR: 1,370–3,297] and a median weekly MVPA of 25 min [IQR: 15–125] (Fig. 2).

Fig. 2.

Self-reported and objectively assessed physical activity in de novo liver transplant recipients at 3–5 months post-transplant. (A) Average weekly minutes of moderate-to-vigorous physical activity (MVPA) as assessed subjectively via the Physical Activity Vital Sign (PAVS) questionnaire and objectively via ActiGraph (AG) in n = 37 liver transplant recipients; (B) Average daily step counts as measured by ActiGraph in participants (n = 30) and non-participants (n = 7)

A discrepancy was observed between self-reported and objectively measured weekly MVPA (Fig. 2). Overall, 65% (24/37) of the subsample overestimated their activity on the self-report questionnaire. Among non-participants, 86% (6/7) overestimated their activity, compared with 60% (18/30) of participants.

Comparison of motivators and barriers to physical activity between participants and non-participants

Motivators and barriers to physical activity (MBQ) domain scores between participants and non-participants are illustrated in Fig. 3A-B, with the distribution of scores per motivator and barrier domain provided in Supplement V and VI, respectively. The main motivators for physical activity among liver transplant recipients were related to health and physical outcomes as well as external influences. Participants in the intervention study scored significantly higher in the motivator domain of health and physical outcomes compared to the non-participants (2.7 [IQR: 2.4–2.9] vs. 2.4 [IQR: 1.9–2.8], p = 0.008). No significant differences were observed between the two groups in the other motivator domains.

Fig. 3.

Motivator and Barrier questionnaire domain scores in participants and non-participants. Data represent median values with 95% confidence intervals for motivator domains (A) and barrier domains (B) for physical activity in participants (n = 64) and non-participants (n = 40). *p < 0.05

Overall, median scores across the barrier domains were low, suggesting that most barriers were perceived as ‘not at all’ or ‘slightly’ limiting among the total group of liver transplant recipients. However, individuals of the non-participant group reported significantly higher scores in the domains of low expectations and self-confidence (0.429 [IQR: 0-0.571] vs. 0.143 [IQR: 0-0.429], p = 0.030) and lack of motivation or time (0.500 [IQR: 0–1.00] vs. 0 [IQR: 0-0.250], p ≤ 0.001) when compared to the group of participants.

Factors associated with willingness to participate in the exercise study and adherence to WHO physical activity guidelines

Table 4 presents the results of the univariate logistic regression analyses conducted to identify factors associated with (1) willingness to participate in the training intervention and (2) meeting the WHO physical activity guidelines, as measured with the Physical Activity Vital Sign questionnaire.

Table 4.

Univariate logistic regression analysis of factors associated with willingness to participate in the training study, and meeting WHO physical activity guidelines

Variable	Participation in the study		Meeting the WHO PA guidelines
	OR [95%CI]	q-value	OR [95%CI]	q-value
Age (years)	0.987 [0.955–1.020]	0.598	0.995 [0.961–1.029]	0.894
Female	1.833 [1.115–3.015]	0.543	1.043 [0.439–2.479]	0.923
Body mass index (kg/m2)	1.031 [0.943–1.127]	0.635	0.984 [0.893–1.085]	0.894
Home-to-site distance (km)	0.988 [0.978–0.997]	0.050*	1.002 [0.992–1.011]	0.894
Life-style related liver disease	0.771 [0.344–1.731]	0.635	0.634 [0.276–1.458]	0.599
History of smoking	0.892 [0.402–1.978]	0.778	1.481 [0.645-3.400]	0.602
Employment	1.222 [0.506–2.951]	0.723	1.896 [0.776–4.629]	0.453
Education	6.444 [1.648–25.195]	0.042*	7.333 [0.9111–59.021]	0.453
High income	2.805 [0.935–8.414]	0.17	4.029 [1.485–10.936]	0.102
Motivator: health & physical outcomes (score: 0–4 scale)	3.761 [1.463–9.666]	0.042*	2.039 [0.788–5.274]	0.453
Motivator: external influences (score: 0–4 scale)	1.558 [0.847–2.866]	0.308	1.051 [0.558–1.983]	0.923
Motivator: groups activities and financial resources (score: 0–4 scale)	1.313 [0.801–2.151]	0.475	1.234 [0.740–2.059]	0.649
Motivator: Psychological outcomes (score: 0–4 scale)	0.907 [0.567–1.45]	0.723	0.934 [0.567–1.538]	0.894
Barrier: fear of negative effects (score: 0–4 scale)	0.509 [0.234–1.108]	0.2	0.466 [0.177–1.224]	0.453
Barrier: Physical limitations (score: 0–4 scale)	0.438 [0.194–0.989]	0.141	0.457 [0.178–1.172]	0.453
Barrier: low expectations and self-confidence (score: 0–4 scale)	0.207 [0.052–0.833]	0.097	0.487 [0.119–1.990]	0.599
Barrier: lack of motivation or time (score: 0–4 scale)	0.185 [0.065–0.527]	0.036*	0.575 [0.233–1.424]	0.563
Adherence WHO PA guidelines	1.596 [0.671–3.793]	0.475	-	-

Abbreviations: OR, odd ratio; CI, confidence interval; Kg, kilogram; Km, kilometre; m, meter; PA, physical activity; WHO, world health organization. *p < 0.05

In the univariate analyses (Table 4), several variables were significantly associated with non-participation. These included a greater distance to the study site (q = 0.050), lower education level (q = 0.042), lower scores on the motivator domain ‘health & physical outcomes’ (q = 0.042), and higher scores on the barrier domain ‘lack of motivation or time’ (q = 0.036).

In the multivariable analysis (Supplement VII), none of the variables remained statistically significant after adjustment for confounders. However, a strong trend was observed for distance to the study site: with each additional kilometre, the odds of participation decreased by approximately 1.2% (AOR = 0.988, 95% CI [0.976–1.000], p = 0.056). Additionally, trends were observed for higher education (AOR = 7.295, 95% CI [0.777–68.500], p = 0.082), and the barrier domain ‘lack of motivation or time’ (AOR = 0.304, 95% CI [0.0.084–1.108], p = 0.071).

In the univariate analysis for meeting the WHO physical activity guidelines (Table 4), no significant independent factors were identified. However, in the multivariable logistic regression analysis (Table 5), high monthly income was significant associated with meeting the WHO physical activity guidelines (AOR = 3.082; 95% CI [1.002–9.473]; p = 0.049).

Table 5.

Multiple logistic regression analysis to identify factors associated with meeting WHO physical activity guidelines

Variable	Meeting WHO PA guidelines
	AOR [95%CI]	p-value
Employment	1.755 [0.620–4.969]	0.289
Education	2.611 [0.277–24.619]	0.402
High income	3.082 [1.002–9.473]	0.049*
Motivator: Health & physical outcomes (score: 0–4 scale)	1.256 [0.383–4.112]	0.707
Barrier: Lack of motivation or time (score: 0–4 scale)	0.812 [0.268–2.456]	0.712

Abbreviations: AOR, adjusted odd ratio; CI, confidence interval PA, physical activity; WHO, world health organization

Discussion

This study is the first to evaluate factors associated with willingness to participate in an exercise intervention trial specifically among liver transplant recipients. While physical activity is increasingly recognized as a cornerstone of post-transplant recovery [6], engagement in exercise-based RCTs remains limited in transplant and other clinical populations [16, 17]. Understanding participation dynamics is therefore essential, not only to improve inclusivity and representativeness of RCTs, but also to support the effective integration of physical activity into post-transplant care. Our findings provide valuable insight into the sociodemographic, logistical, and behavioural barriers that influence participation, and underscore the structural challenges faced by underrepresented liver transplant recipients.

By examining both participants and non-participants in the PHOENIX-Liver trial, this study helps to identify key obstacles to engagement and reveal which groups are at greater risk of being excluded from such interventions. Our findings suggest that practical and motivational barriers, such as time constraints, transportation difficulties, and lower intrinsic motivation, play a major role in limiting participation. This aligns with previous research showing that logistical challenges and psychosocial factors are common reasons for non-participation in exercise interventions among clinical populations [17, 31, 32]. Compared to participants, we found that non-participants were living further away from the study site and were more likely to have a lower socioeconomic status based on education level and income. In addition, non-participants engaged less in objectively assessed physically activity. These findings emphasize the need to address specific barriers experienced by these individuals to ensure equitable access to research participation.

Individuals with lower education levels and income were underrepresented in our study sample. These findings suggest that lower socioeconomic status—reflected in both educational attainment and income—was associated with reduced study participation. This pattern is consistent with prior research showing that socioeconomically disadvantaged populations are often underrepresented in clinical studies [33], which poses a threat to the generalizability and equity of health research outcomes. To better reach individuals with lower education levels, recruitment strategies should be tailored to their specific context [17, 33, 34]. This includes using of clear and accessible (lay) language, simplifying study materials, providing visual aids, and increasing contact moments for personal explanation [17, 34]. In the PHOENIX-Liver study, for example, liver transplant recipients were approached on multiple occasions throughout their care trajectory and informed by various healthcare providers (e.g., transplant surgeon, rehabilitation physician, physiotherapist, study staff) [21]. Additionally, a shortened and adapted information leaflet was provided alongside the official informed consent, and visibility was enhanced through flyers on the ward and a study website [21]. Beyond informational barriers, financial constraints may also hinder participation [17, 34]. In our study, a greater proportion of individuals in the non-participant group had a monthly income below €1000, suggesting that economic factors such as transportation costs or time investment may have discouraged engagement. Practical strategies—such as reimbursing travel expenses, offering home-based assessments or interventions, and minimizing other study-related costs—may help to reduce these barriers [35, 36]. Importantly, we also observed that individuals with a higher monthly income showed a significant association with meeting the WHO physical activity guidelines toward greater adherence. This is in line with previous findings showing that lower socioeconomic status has not only been associated with reduced study participation, but also with lower physical activity levels, as shown in general populations [37]. This highlights a critical paradox: the individuals who may benefit most from physical activity interventions are often the least likely to engage in them. Addressing these structural inequities requires not only tailored recruitment and study design strategies but also broader policy-level actions. For example, improving reimbursement policies for physiotherapy or subsidizing access to community-based exercise facilities could help reduce financial barriers to sustained physical activity among transplant recipients with limited resources [38, 39].

Our findings revealed overall low activity levels in this post-transplant population. The median daily step count was 3,788 steps/day, and only 34% of individuals met the WHO physical activity guidelines—defined as at least 150 min of MVPA per week and strength training on at least two days per week. These results are consistent with previous studies in liver transplant recipients [5, 40]. For instance, Tanaka et al. reported an average of 3,887steps/day in liver transplant recipients 6 months post-transplantation [40]. Interestingly, individuals in the non-participant group appeared to be less active than the participants, a trend also observed in previous studies across various populations [41]. Although only a minority of participants met the WHO guidelines, it is important to note that even activity levels below these thresholds can confer substantial health benefits [11]. Currently, no post-transplant-specific physical activity guidelines exist [15], highlighting the need for tailored recommendations that account for the unique needs and limitations of transplant recipients. Furthermore, growing evidence suggests that using a single cutoff point to define “meeting” physical activity guidelines may not be the most effective approach for promoting activity at the population or individual level. Emphasizing that any amount of physical activity is beneficial can empower inactive individuals to gain health benefits, even when the recommended target range is perceived as out of reach [42]. Developing post-transplant-specific guidelines that consider these nuances could help clinicians prescribe appropriate activity levels, address barriers to physical activity, and promote sustained engagement in this population.

Moreover, a direct comparison between objective and subjective physical activity measurement showed that 85% of non-participants overestimated their physical activity levels on the Physical Activity Vital Sign questionnaire compared to objectively measured data. These findings suggest that, in this population, self-reported physical activity through the Physical Activity Vital Sign questionnaire may not be sufficiently reliable to identify individuals who would most benefit from a physical activity intervention. However, this interpretation should be made with caution, given the small subsample size with available accelerometer data. These findings highlight the need for further research with larger cohorts to validate whether subjective tools, such as the Physical Activity Vital Sign questionnaire, can serve as reliable and practical tools for screening physical activity levels in this population.

In our total cohort, health and physical benefits emerged as the most frequently reported motivator for engaging in physical activity. This is consistent with previous research in both solid organ transplant recipients [43] and other chronic disease populations [44], where the health and physical benefits of exercise are commonly identified as key facilitators of physical activity. Notably, this motivator was even more pronounced among participants who were willing to participate in the intervention. These observations highlight the importance of strengthening health literacy, for example through targeted education [45, 46], to enhance patients’ understanding of the benefits of physical activity and support more informed decision-making. Higher health literacy has been associated with more advanced phases of motivational readiness for physical activity [47] as well as higher overall physical activity levels [48–50]. Such insights offer important implications not only for recruitment and retention strategies in future exercise interventions but also for motivating individuals to engage in physical activity more broadly. Emphasizing the health and physical benefits of exercise during clinical consultations by various health professionals and throughout study recruitment may enhance motivation among liver transplant recipients. Importantly, this should be combined with other behaviour change techniques, including motivational interviewing, goal setting, self-monitoring of physical activity, and follow-up prompts, all of which can be tailored to meet patients’ changing needs to improve intrinsic motivation and promote long-term engagement in physical activity [15].

Overall, reported barriers to engage in physical activity were relatively low in our cohort. The generally low barrier scores across both participants and non-participants may partly reflect social desirability bias, defined as an individual’s tendency to respond in a manner viewed favourably by society [51]. The use of self-report questionnaires with fixed-scale scoring options may increase the likelihood of socially desirable responding [52, 53]. Social desirability bias has previously been demonstrated in the context of physical activity, often leading patients to over-report their activity levels [54]. Alternatively, it is possible that liver transplant recipients perceive each individual barrier as only slightly limiting, but the cumulative effect of multiple minor barriers may still contribute to low overall physical activity levels. In our cohort, the most frequently cited barrier was physical limitations. Although comparative data in liver transplant recipients remain limited, this observation aligns with findings from studies in other chronic disease populations, where physical limitations are often reported as significant barriers to physical activity [55–57]. In our study, individuals who scored higher on specific barrier domain ‘lack of motivation or time’ were significantly less likely to participate, as evidenced by univariate analyses.

Notably, the reasons for non-participation in this study often overlapped with the reported barriers to physical activity. Interestingly, despite being less likely to be employed, unwilling participants most frequently cited time constraints as their reason for non-participation. A possible explanation is that they may be less willing to invest time to physical activity rather than genuinely having too little time. These constraints reflect the need for interventions that are flexible and minimally disruptive to individuals’ schedules. Additionally, transportation challenges were the second most commonly reported reason for non-participation in the PHOENIX-Liver trial. These challenges were exclusively related to the requirement for participants to travel to the study site for study assessments, as the exercise intervention itself was home-based. This is consistent with our logistic regression analysis, which identified greater distance from the study site as a significant predictor of reduced willingness to participate. Addressing these challenges by decentralizing assessments or incorporating remote evaluation methods could help reduce logistical obstacles and facilitate participation, particularly for individuals living farther from the study site or those without reliable access to transportation.

Efforts to mitigate these barriers can enhance engagement in exercise and physical activity interventions for liver transplant recipients, extending beyond research and even beyond the liver transplant setting. Namely, these findings may also be relevant for other transplant populations, as physical inactivity is both common [15] and a significant risk factor among all transplant recipients [58]. To maximize uptake and sustained adherence to an active lifestyle, exercise and physical activity interventions should be person-centred and tailored to overcome individuals perceived barriers, ensuring they are both feasible and adaptable to each individual’s unique need. A comprehensive, multi-level approach that integrates individuals’ needs with healthcare support and policy improvements may ultimately lead to sustained engagement in physical activity, improving long-term health outcomes and overall well-being in this population [15].

This study has several limitations that should be acknowledged. First, although both subjective and objective measures of physical activity were collected, objective data using ActiGraph devices were only available in a relatively small subsample. Despite this small sample size, significant differences were found between participants and non-participants. However, limited statistical power may restrict the strength of conclusions that can be drawn from these findings. Consequently, the majority of physical activity data relied on self-reported questionnaire, which may be prone to social desirability bias and recall bias, potentially leading to overestimation of activity levels. Second, as this study focused on willingness to participate in an intervention rather than actual behaviour over time, we were unable to assess whether non-participants eventually engaged in physical activity independently after refusing study participation. Third, the decision to dichotomize physical activity data in the logistic regression analyses (e.g., ≥ 150 min/week of MVPA) may have reduced variability between participants and non-participants. Using continuous physical activity data may be more clinically meaningful, as even activity levels below the WHO guidelines can confer health benefits [11]. Fourth this was a single-centre study, with recruitment limited to a single tertiary transplant hospital in Belgium, which may restrict the generalizability of our findings to other settings or populations. Finally, the intervention itself is highly time-consuming and intensive (5 days/week), which may have influenced willingness to participate and might limit the applicability of our findings to less intensive or more flexible programs.

Despite these limitations, the study also has notable strengths. A major strength is its innovative methodology, which, for the first time, maps selection bias in this context by including both participants and non-participants, as well as incorporating assessments of motivators and barriers. This approach offers a comprehensive perspective on factors influencing study participation and potentially also the uptake of a physical activity lifestyle in this population, serving as a valuable example for future clinical research aiming to improve inclusivity and representativeness. Additionally, we were able to verify several self-reported variables with data from the electronic health records, which strengthens the accuracy of medical and sociodemographic information.

Conclusion

This cross-sectional analysis within the PHOENIX-Liver trial highlights key factors influencing participation in an exercise and physical activity intervention among liver transplant recipients. Sociodemographic disparities, particularly lower socioeconomic status, were strongly associated with unwillingness to participate, underscoring the need for targeted strategies to improve accessibility. Non-participants also exhibited lower physical activity levels and perceived more barriers, such as time constraints, low motivation, and transportation challenges, which frequently overlapped with reasons for nonparticipation. Addressing these barriers requires a person-centred approach that considers individual, healthcare system, and policy-level factors. By implementing tailored interventions, flexible study designs, and structural support, engagement in physical activity can be enhanced, ultimately improving health outcomes and quality of life for liver transplant recipients.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AOR

Adjusted odd ratio

CI

Confidence interval

ICF

Informed consent form

IQR

Interquartile range

LT

Liver transplantation

MBQ

Motivators and barriers to physical activity

MVPA

Moderate-to-vigorous physical activity

OR

Odd ratio

RCT

Randomised controlled trial

WHO

World Health Organization

Author contributions

HV, LV, and SL were responsible for patient recruitment and data collection. SL performed the data analysis and interpretation. SL and VC drafted the initial manuscript. All authors contributed to critical revision of the manuscript and approved the final version. VC is the guarantor of the study.

Funding

This research project is fully funded by support from KU Leuven (C2-M, grant number: C26M/21/001) and FWO (S006722N). The funding source of the present project had no role in the design of this study and will not have any role during its execution, analyses, interpretation of the data, or decision to submit results.

Data availability

The datasets generated and/or analysed during the current study will be made available in the KU Leuven institutional repository Lirias. The data are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Ethics approval (B3222022000943) for the study has been granted by the local ethical committee of UZ/KU Leuven on 13/12/2022. An amendment for accelerometer-based assessments (ICF-B subgroup) was approved on 6 June 2024. The study complies with the International Conference for Harmonization of Good Clinical Practice guidelines and was conducted in accordance with the Declaration of Helsinki. All participants provided written informed consent prior to inclusion.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.