Background
Cancer-related fatigue has a major impact on psychological well-being, social relationships, work and health-related quality of life(Reference Bower1,Reference Burrell, Sasso and Greenle2) of colorectal cancer survivors. The prevalence of cancer-related fatigue among colorectal cancer survivors varies across studies from 40 to 70 % in the first 5 years after diagnosis(Reference Husson, Mols and van de Poll-Franse3,Reference Drury, Payne and Brady4) .
Observational data suggest that colorectal cancer patients with a healthy lifestyle experience less cancer-related fatigue after completion of treatment(Reference Bower1,Reference Wesselink, van Baar and van Zutphen5–Reference Eyl-Armbruster, Thong and Carr7) , where a healthy lifestyle is defined as consuming a healthy diet and/or being physically active. Results of intervention studies conducted among colorectal cancer survivors who completed treatment are not consistent but suggest potential beneficial effects of exercise programmes on cancer-related fatigue(Reference Machado, Morgado and Raposo8); only few intervention studies assessed whether a healthy diet can affect cancer-related fatigue(Reference Baguley, Skinner and Wright9). Interventions in which diet and physical activity or exercise are combined are even sparser: we identified one large and two smaller studies in specifically colorectal cancer survivors(Reference Bourke, Thompson and Gibson10–Reference Wang, Langlais and Kenfield12). The larger study was a telephone-delivered multiple health behaviour change intervention (‘CanChange’) among 410 colorectal cancer survivors diagnosed within the previous 12 months(Reference Hawkes, Chambers and Pakenham11). CanChange showed that the 6-month intervention was effective in improving physical activity and dietary habits. An effect on cancer-related fatigue was not observed, likely because participants of that study were not selected based on experiencing cancer-related fatigue and therefore did not or hardly experienced cancer-related fatigue.
The two smaller studies were both 12-week pilot/feasibility studies. The first study with eighteen participants concluded that a lifestyle intervention was feasible in colorectal cancer survivors who had surgery 6–24 months ago(Reference Bourke, Thompson and Gibson10). Results of the other study, which included fifty participants, suggest that there is a potential effect of a web-based dietary intervention on cancer-related fatigue in colorectal cancer survivors who were disease-free or had stable disease and were not undergoing chemotherapy(Reference Wang, Langlais and Kenfield12). Importantly, the studies did not select participants based on the level of cancer-related fatigue. Therefore, we argue that there is a clear need for a lifestyle intervention focused on both diet and physical activity in colorectal cancer survivors who are experiencing cancer-related fatigue.
Lifestyle interventions among colorectal cancer survivors may require a specific approach for several reasons. Colorectal cancer survivors may have disease-related barriers, such as having a stoma and/or experiencing bowel dysfunction; this may make it challenging to adhere to general advice on healthy eating and to engage in exercise(Reference Denlinger, Barsevick and Robinson13). In addition, cancer-related fatigue may limit the ability to conduct activities of daily living, such as food preparation, or participating in exercise activities(Reference Hofman, Ryan and Figueroa-Moseley14), which may ask for further adaptations to a lifestyle programme.
As lifestyle interventions often contain several interacting components and involve a range of behaviours, expertise and skills, these interventions ask for a systematic design(Reference Skivington, Matthews and Simpson15,Reference Lamort-Bouché, Sarnin and Kok16) . In the current report, we describe our approach to developing a person-centred lifestyle programme for colorectal cancer survivors based on the World Cancer Research Fund/American Institute for Cancer Research (WCRF/AICR) cancer prevention guidelines(17). Moreover, we describe the design of the SoFiT trial. The primary aim of the SoFiT trial is to test the effect of the personalised lifestyle intervention on cancer-related fatigue among colorectal cancer survivors experiencing cancer-related fatigue.
Development of the lifestyle programme
The goal of the lifestyle programme was to increase the participants’ adherence to the WCRF/AICR cancer prevention guidelines(17). These guidelines are a set of recommendations on diet, physical activity and body weight; see Table 1.
Table 1. Description of the WCRF recommendations. These recommendations are the core of the lifestyle programme that is tested in the SoFiT trial
WCRF, World Cancer Research Fund; SoFiT, Study on Fatigue: a lifestyle intervention among colorectal cancer survivors.
We aimed to design a programme that would foster lasting changes in the adherence to these guidelines and that would sustain behaviour change beyond the duration of the programme(Reference Kiernan, Brown and Schoffman18).
The programme was developed based on the acknowledgement that in behaviour change interventions, relevant personal and environmental determinants of that specific behaviour need to be targeted with behaviour change techniques(Reference Kay Bartholomew, Parcel and Kok19,Reference Michie, Johnston and Rothman20) . To select effective behaviour change techniques, it is crucial to understand the factors that influence the target behaviour. This involves the identification of key personal and environmental determinants linked to the behaviour, based on theories of behaviour and behaviour change. Therefore, we conducted a systematic review to identify the most important determinants of healthy lifestyle behaviours among colorectal cancer survivors(Reference de Vries-ten Have, Winkels and Zondervan21). Complementary, we conducted interviews with eight colorectal cancer survivors, of whom six experienced cancer-related fatigue. The focus of these interviews, which lasted 30–90 min, was to assess which factors influenced compliance or non-compliance with the WCRF/AICR cancer prevention recommendations. Thematic content analysis(Reference Braun and Clarke22) was used to identify determinants of health behaviours. The determinants that emerged from the interviews and that resulted from the systematic literature review are described in the chapter Data Collection. In addition, we published a systematic review to show which behaviour change techniques are frequently used and which behaviour change techniques appear most promising to effectively change behaviour in lifestyle interventions in cancer survivors(Reference de Vries-ten Have, Winkels and Kampman23).
Within our systematic review, we observed that the behaviour change technique ‘generalisation of the target behaviour’ is a promising technique that was often used in lifestyle interventions that were effective in reducing cancer-related fatigue among cancer survivors(Reference de Vries-ten Have, Winkels and Kampman23). Generalisation of the target behaviour means that it is important to incorporate change in the own environment and in daily life. Indeed, both theory and evidence suggest that creating habits and learning behaviour in the most relevant contexts is important for long-term behaviour change(Reference Kwasnicka, Dombrowski and White24–Reference Lally and Gardner27). Moreover, behaviour change is more likely to be maintained if people are motivated by their own needs and desires(Reference Rothman, Baldwin and Hertel28), and if newly adopted behaviour reflects the values of the person and is seen as personally relevant(Reference Kwasnicka, Dombrowski and White24). For the lifestyle programme, we decided that the focus should be on supporting participants in integrating lifestyle changes into their daily lives, by considering the relevance of that behaviour for the participant, the specific barriers for that behaviour, its easiness to continue on the long term and the motivation of the participant. This led to a programme that is person-centred and autonomy supportive, which are both important predictive factors of maintenance of behaviour change(Reference Samdal, Eide and Barth29).
Our reviews(Reference de Vries-ten Have, Winkels and Zondervan21,Reference de Vries-ten Have, Winkels and Kampman23) and interviews, in combination with the ‘Theory and Technique Tool(Reference Michie, Johnston and Rothman20), set the stage for the selection of behaviour change techniques for our programme. The Theory and Technique Tool(Reference Michie, Johnston and Rothman20) serves as a guide, offering insights into the most likely links between behaviour change techniques and specific determinants. We chose behaviour change techniques that are most likely linked to the identified determinants of lifestyle behaviour, thereby enhancing the likelihood of successful and sustained behaviour change among participants. Once we established the links between the determinants and suitable behaviour change techniques, we decided on practical applications of this combination. See online Supplementary Table S2 for an overview of the behaviour change techniques per determinant and examples of how they were applied.
This development phase led to a programme that had a general core and a person-centred approach, which is described below, details can be found in online Supplementary Fig. S1.
General core of the lifestyle programme
During 6 months, participants of the lifestyle programme were coached to increase their adherence to the WCRF recommendations on diet and physical activity: weight loss was not a specific goal of the programme but may have happened as a result of adopting a healthier lifestyle.
A 6-month period with a contact moment every 2 weeks was chosen for the programme, as a previously conducted randomised controlled trial showed that a 6-month period with a contact moment every 2 weeks was sufficient to improve lifestyle behaviour among colorectal cancer survivors(Reference Hawkes, Chambers and Pakenham11).
In our programme, the participant received four face-to-face appointments with one of the two lifestyle coaches, who visited them at their home. Additionally, there were eight contact moments with the coach by telephone or video call. Both coaches graduated from a postgraduate programme for lifestyle coaching. At the beginning of the programme, the participant received a paper handbook with an introduction to the programme; information on the WCRF recommendations and on its possible health benefits and suggestions for implementation; information on the home visits and measurements; and weekly fill-in schemes for goal setting and action planning involving stating, planning, monitoring, evaluating, and revising goals, and several brochures (for a list of brochures and information, see online Supplementary Table S1).
Throughout the programme, several behaviour change techniques were applied for all participants as these are crucial for guiding behaviour change. Coaches provided information on the importance of a healthy lifestyle and its health consequences. Furthermore, behaviour change techniques around setting, reviewing and adjusting goals on behaviour and anticipated outcome (e.g. feeling more fit) were applied, since goal setting is found to be an important predictor for maintaining behaviour change in the long term(Reference Samdal, Eide and Barth29). Goal setting was always coupled with the behaviour change technique ‘Action planning’ to help individuals plan the specific actions they took to achieve their overarching goals(Reference Bailey30). Barriers for changing behaviour were also identified and formed part of action plans. Furthermore, participants monitored their own behaviour by filling in the weekly fill-in schemes in the participant handbook, and coaches gave feedback on behaviour and outcomes.
During the first appointment, the focus of the coach was on building trust with the participant and on obtaining a clear impression of the participant. Moreover, the goal of that first appointment was to discuss mutual expectations and to get commitment of the participant for the programme. During this first meeting, the coach also discussed a baseline report that contained the results on the diet quality of the participant (see later for details about dietary assessment), physical activity level of the participant (see later for details on assessment), determinants of behaviour and other information (body weight and BMI, stoma, allergies/intolerances, cancer-related side effects and co-morbidities). Based on this, participant and coach together set two or three goals towards reaching the WCRF recommendations (see Table 1), and two or three goals on an anticipated outcome of behaviour (e.g. experiencing less fatigue or feeling more active). The coach and participant made an action plan for the next 2 weeks.
The next appointments focused on evaluating and monitoring goals, on providing feedback, and on changing and/or adding goals when possible. Barriers were identified, and action or coping plans were made together with the participant. Additional goals towards reaching the WCRF recommendations could be added throughout the programme. After 3 months, the coach presented a mid-term report to the participant, compiled from the data that the participant provided at the 3-month assessment. Similar to the baseline report, this report contained information on diet quality, physical activity, determinants of behaviour, and body weight and BMI. The coach used this information to discuss progress and to plan the content of the individual programme for the next 3 months. During the following appointments, participants worked on the goals again. Now, the focus was also on relapse prevention. At the final appointment, the progress and programme were evaluated and a long-term planning was made to prevent relapse and maintain a healthy lifestyle after the programme.
Person-centred approach of the programme
The programme was person-centred as it was (1) tailored towards the current lifestyle and personal characteristics of the participant, (2) targeting personal behavioural determinants and (3) taking into account the preferences, opportunities and disease-related barriers of the participant.
Personalisation to current lifestyle and personal characteristics
To tailor the programme to the current lifestyle and personal characteristics of the participant, the following information of the participant was used at baseline and at 3 months: diet quality, physical activity level, weight/BMI, any co-morbidities, allergies and intolerances, cancer-related side effects, and having a stoma.
Personalisation to the determinants of behaviour
To tailor the programme to individual behavioural determinants of the participant, we created a questionnaire to assess these behavioural determinants (see more details on this questionnaire later). In this questionnaire, we concentrated on a specific set of determinants that we deemed crucial to achieve sustained long-term changes in dietary intake and physical activity. The participant completed this questionnaire at baseline and after 3 months. The coaches received an overview of determinants of behaviour of the participant and used practical applications of matched behaviour change techniques to target those determinants; see online Supplementary Table S2.
Preferences, opportunities and disease-related barriers
The preferences and opportunities of the participant were taken into account in shaping the coaching. The coach and participant together decided on which WCRF recommendations were targeted and how these were targeted considering the challenges participants had (e.g. lack of financial resources) by setting SMART goals (Specific, Measurable, Achievable, Relevant and Time-Bound) and by creating action plans. Moreover, it is often crucial to address other issues that affected the participant, to enable the participant to change their lifestyle behaviour(Reference Polhuis, Bouwman and Vaandrager31). This could be attention for mental health (e.g. anxiety for recurrence), sleep problems, weight management, acknowledgement for the cancer-related fatigue and acceptance of self, the disease and their place in society (e.g. not being able to work anymore). This mostly involved giving the participant room to express such problems.
Methods of the randomised controlled trial
To represent the patient voice throughout all aspects of the study, a patient representative of a national patient organisation ‘Stichting Darmkanker’ was involved in the writing of the project proposal to acquire funding, in the writing of the protocol for medical-ethical approval, in testing questionnaires, with recruitment efforts, and other study procedures. The study was approved by the Medical Research Ethics Committee (Committee on Research Involving Human Subjects region Arnhem-Nijmegen, NL75999.091.21, nr 2021-8182). All participants provide written informed consent.
Design of the trial
The SoFiT study was a parallel randomised controlled trial with two groups: an intervention and a control group. For a timeline of the research activities of the study, see online Supplementary Fig. S2. The intervention group received the 6-month programme described above. Participants in the control group did not receive that programme but were contacted every 1·5 months to promote retention. At 1·5 months, participants in the control group received a newsletter in which general tips were given about books to read (tips are about novels, not related to lifestyle) or podcasts to listen to (podcasts are entertaining stories, not related to lifestyle). At 3 months, participants in the control group received a phone call to collect study-related data (see Table 2). At 4·5 months, participants in the control group received a newsletter with general information on the importance of a control group. After completion of data collection at 6 months, the control group received the information booklets that the intervention group also received, and two lifestyle coaching sessions with a coach.
Table 2. Timing of the measurements and assessments in the SoFiT trial, a randomised controlled study among colorectal cancer survivors
FACIT, Functional Assessment of Chronic Illness Therapy; SoFiT, Study on Fatigue: a lifestyle intervention among colorectal cancer survivors; SQUASH, Short QUestionnaire to ASsess Health-enhancing physical activity; PSQI, Pittsburgh Sleep Quality Index; CSD, Consensus Sleep Diary; FACT-C questionnaire, Functional Assessment of Cancer Therapy – Colorectal questionnaire; PHQ, Patient Healthcare Questionnaire; GAD, Generalized Anxiety Disorder.
* Participants completed surveys online or via paper, except for the SQUASH questionnaire and the survey on medicine and vitamin use, which were assessed by the research team during the home visits and via telephone. All measurements were done by the research team during the home visits.
Population and recruitment
Inclusion criteria were as follows: adult colorectal cancer survivors with stage I–III disease who completed colorectal cancer treatment at least 6 months and no more than 5 years ago and who were experiencing cancer-related fatigue as assessed during screening. We assessed cancer-related fatigue with the Functional Assessment of Chronic Illness Therapy (FACIT) – Fatigue Scale(Reference Van Belle, Paridaens and Evers32) and defined cancer-related fatigue as a score lower than 34(Reference Van Belle, Paridaens and Evers32). Additional inclusion criteria were as follows: persons had to live on a reasonable distance from the research centre (within approximately 1·5 h driving by car from Wageningen University & Research), persons had to be willing to be randomised into either the intervention or control group, and persons had to be able to speak, write and read Dutch.
Exclusion criteria are as follows: participation in another study that could interfere with the current study, excessive alcohol consumption (more than four glasses/d on average), chronic recreational drug use, and unwilling or unable to comply with the intervention (e.g. through dementia or mental illness).
Recruitment took place via three routes: through the Prospective National Colorectal Cancer (PLCRC) cohort(Reference Derksen, Vink and Elferink33,Reference Burbach, Kurk and Coebergh van den Braak34) , through regional hospitals, and through newsletters, websites, and social media channels of patient organisations.
Recruitment through PLCRC was as follows. PLCRC is an open cohort of colorectal cancer patients diagnosed across >60 hospitals in the Netherlands(Reference Derksen, Vink and Elferink33,Reference Burbach, Kurk and Coebergh van den Braak34) and consists of >10 000 participants. Upon recruitment into that PLCRC cohort, all participants are asked whether they would like to be informed about future studies of other research institutes.
The PLCRC research team selected which cohort members signed informed consent to be invited for future research studies. Out of those cohort members, the PLCRC research team assessed who completed treatment 6 months to 5 years ago, and who lives on a reasonable distance from the SoFiT research centre. Those cohort members received an information package about the SoFiT study from the PLCRC research team. This contained an invitation letter, flyer and postcard to send back to the study team to express their interest.
Recruitment through regional hospitals was as follows. A member of the hospital research team regularly checked whether there were colorectal cancer survivors who completed cancer treatment at least 6 months ago. Those colorectal cancer survivors received the information package about the study.
Persons interested in hearing more about the study were requested to contact the SoFiT research team at Wageningen University & Research. The research team provided written and oral information about the study, screened the person for eligibility and answered any questions that the person had. Eligibility screening involved completing an online (or paper) questionnaire through Castor Electronic Data Capture (Castor EDC). This questionnaire consisted of the thirteen-item FACIT-Fatigue Scale(Reference Van Belle, Paridaens and Evers32) to assess whether the person was experiencing cancer-related fatigue and of questions to assess the other inclusion and exclusion criteria. Eligible persons who signed written informed consent were included in the study.
Sample size calculation
The intended sample size is based on an estimated differential change in cancer-related fatigue from baseline to 6 months between intervention and control group of 3 points on the FACIT-Fatigue Scale(Reference Minton and Stone35,Reference Cella, Eton and Lai36) . A change of 3 points is considered a clinically important difference(Reference Minton and Stone35,Reference Cella, Eton and Lai36) . In the calculations, we used an sd of 6·7 which was inferred from a 12-week physical activity intervention aimed to decrease cancer-related fatigue in colorectal cancer survivors(Reference Kim, Lee and Lee37). Assuming an α of 0·05 and a power of 80 %, the required sample size was estimated as 78 per group. Comparable lifestyle interventions show dropout rates of about 15 %. We increased the estimated sample size by 15 % to account for possible dropout, resulting in a total sample size of 184 participants, 92 per group. If persons were diagnosed with cancer, cancer recurrence or metastasis during the 6 months of the study, participants were taken out of the study and not replaced.
Data collection
Measurements and assessments were done at baseline, at 3 months, 6 months and 12 months; see Table 2 for an overview of which measurements and assessments were done. At baseline and 6 months, a member of the research team collected data from participants during a home visit; shortly before or after those visits, participants completed questionnaires via paper or online. At 3 and 12 months, data participants completed questionnaires via paper or online and via telephone. All data were entered and/or collected through Castor EDC.
Data are only accessible to members of the research team. Depending on the role of the team member, access can be restricted to specific subgroups or items. Data on adverse events were reported to the Medical Ethical Committee once a year, and data on serious adverse events were reported to the Medical Ethical Committee as soon as the research team was notified about those.
Randomisation and blinding
After collecting baseline data during the home visit, the participant was randomised to intervention or control group (1:1 allocation ratio) through the randomisation module of Castor EDC using a stratified block-randomisation procedure. Stratification was done for the level of cancer-related fatigue (≤20, or >20 on the FACIT-Fatigue Scale) and for whether or not chemotherapy was received as part of treatment (chemotherapy yes/no). Permuted blocks of 4, 6 of 8 were used.
The nature of the intervention did not allow us to blind participants to treatment allocation. However, the baseline measurements were carried out blinded and participants were immediately afterwards randomised so that they were aware of their group allocation at the end of the first home visit. The primary outcome of the trial could not be collected blinded, as this is self-reported cancer-related fatigue: participants completed the FACIT-Fatigue questionnaire online or via paper at home, before or shortly after the measurement visit. After completion of data collection, the database will be locked and pseudo-anonymised by an independent researcher to ensure that data analysis is conducted blinded. Full details of the statistical analysis plan of any primary or secondary outcome or mediation analyses will be finalised prior to database lock and unblinding, as previously recommended(Reference Petersen, Kris-Etherton and McCabe38).
Primary and secondary outcomes
Primary outcome: cancer-related fatigue
Cancer-related fatigue was assessed with the FACIT-Fatigue Scale. This scale comprises of thirteen items that assess cancer-related fatigue and its impact. Each item is scored on a five-point Likert scale ranging from: ‘Not at all’ to ‘Very much’(Reference Yellen, Cella and Webster39). Lower scores mean higher cancer-related fatigue levels. We chose the FACIT-Fatigue Scale based on two reviews(Reference Minton and Stone35,Reference Agasi-Idenburg, Velthuis and Wittink40) that concluded that this questionnaire is valid and user-friendly, relatively brief, and can easily be combined with other quality-of-life instruments.
Secondary outcome: changes in lifestyle behaviour
To assess how lifestyle behaviour changed during the 6 months of the study, we assessed dietary intake, physical activity level and anthropometrics at baseline and 6 months.
Diet. To assess changes in dietary intake, participants completed a semi-quantitative FFQ(Reference Feunekes, Van Staveren and Graveland41,Reference Verkleij-Hagoort, de Vries and Stegers42) . Participants reported the intake of foods and drinks consumed during the previous month.
In addition, participants completed a brief dietary screener ‘Eetscore’, which was developed to assess diet quality by comparing intake with the Dutch food-based dietary guidelines 2015(Reference de Rijk, Slotegraaf and Brouwer-Brolsma43). This Eetscore estimates adherence to each of the fifteen components of the Dutch dietary guidelines and includes an additional sixteenth component on unhealthy choices of foods and drinks. For each of these sixteen components, ten points can be awarded, resulting in a total score of 0 (no adherence) to 160 (complete adherence to all guidelines). The coach used the Eetscore to tailor the intervention to the current dietary intake of the participant. As timing of dietary intake may be relevant in the context of cancer-related fatigue(Reference Chong, Eussen and van Roekel44), data on timing of dietary intake was assessed with a twenty-six-item questionnaire on meal regularity, meal frequency and meal clock time(Reference Berendsen, Boss and Smits45).
Physical activity. Physical activity level was assessed in two ways: subjectively via the Short QUestionnaire to ASsess Health-enhancing physical activity (SQUASH)(Reference Wendel-Vos, Schuit and Saris46) questionnaire and objectively via an accelerometer (ActivPalTM Micro3). The questions in SQUASH focus on commuting, work/school activities, household activities, leisure time and sports in a representative week. Scores are assigned to the reported activities based on intensities in MET and translated to minutes of physical activity. The coach used the information from the SQUASH questionnaire to tailor the intervention. Participants wore the accelerometer for nine consecutive days on the midline of the thigh. This is in line with previous research(Reference Aguilar-Farias, Martino-Fuentealba and Salom-Diaz47) which showed that 5 days with at least 1 weekend day is needed for reliable estimation of activities and considered that the days of placement and removal are not used for analyses.
Anthropometrics. A researcher measured the body weight of the participant with a calibrated scale, waist circumference with a tape measure and height with a stadiometer. At 3 months and 12 months, participants were asked to weigh themselves using their own weighing scales and self-report their weight.
Secondary outcome: echo intensity of skeletal muscle and skeletal muscle cross-sectional circumference and thickness
We previously showed that the prevalence of cancer-related fatigue was higher among colorectal cancer patients who had more fat infiltration in skeletal muscle(Reference van Baar, Bours and Beijer48),(Reference Aubrey, Esfandiari and Baracos49). Lifestyle may affect fat infiltration in skeletal muscle(Reference Aubrey, Esfandiari and Baracos49). In the current study, we aim to assess whether a potential change in cancer-related fatigue is mediated by changes in fat infiltration in skeletal muscle.
Ultrasound of skeletal muscle. Information on echo intensity of skeletal muscle, skeletal muscle thickness and circumference, and thickness of subcutaneous fat(Reference Young, Jenkins and Zhao50–Reference Perkisas, Bastijns and Baudry52) was obtained through ultrasound assessment in brightness mode (B-mode). We used ultrasound as this is a portable, non-invasive technique that does not involve radiation. Echo intensity of skeletal muscle is considered indicative of fat infiltration in skeletal muscle(Reference Young, Jenkins and Zhao50,Reference Pillen and van Alfen53) . Ultrasound assessments of the rectus femoris, lateral gastrocnemius and biceps brachialis were done during the home visits using a portable ultrasound, the Terason® uSmart 3300 with Terason® linear transducer 15WL4 (Terason). The settings were sat at a gain level of 58 dB, dynamic range at 72 and transducer frequency to 8 MHz giving a frequency of 23 Hz with OmniBeam switched on.
Secondary outcome: symptom clusters
Cancer-related fatigue is rarely an isolated symptom and occurs often in symptom clusters with symptoms such as disturbed sleep quality(Reference Legg, Meertens and van Roekel54), emotional distress and other concerns(Reference Berger, Mitchell and Jacobsen55). We will assess the effect of the lifestyle intervention on this cluster of symptoms, including sleep outcomes, health-related quality of life, colorectal cancer-related concerns, and depression and anxiety.
Sleep quality and duration. Sleep quality was assessed with the self-reported Pittsburgh Sleep Quality Index (PSQI)(Reference Buysse, Reynolds and Monk56). This validated questionnaire contains nineteen questions on seven domains: subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleep medication and daytime dysfunction. The scores range from 0 to 21, with higher scores indicating worse sleep quality. Participants also completed the fifteen-item Consensus Sleep Diary (CSD)(Reference Carney, Buysse and Ancoli-Israel57) to obtain information about sleep characteristics. Participants completed this diary during the period that they wore the accelerometer.
Health-related quality of life and colorectal cancer-related concerns. The FACIT-Fatigue Scale was used in combination with the Functional Assessment of Cancer Therapy – Colorectal (FACT-C) questionnaire(Reference Ward, Hahn and Mo58). The FACT-C contains thirty-six items and was used to assess physical well-being, social/family well-being, emotional well-being and functional well-being and to assess concerns about colorectal cancer-specific issues; the recall period is 7 days. Higher scores mean better quality of life.
Depression and anxiety. Participants completed the Patient Healthcare Questionnaire (PHQ)-8(Reference Kroenke, Spitzer and Williams59,Reference Kroenke and Spitzer60) to assess signs of possible depressive disorders and the Generalized Anxiety Disorder (GAD)-7 questionnaire to identify probable cases of generalised anxiety disorder(Reference Spitzer, Kroenke and Williams61). Both questionnaires have a recall period of 2 weeks, and higher scores mean higher indication of possible depressive disorder and generalised anxiety disorder.
Additional outcomes and measurements
Behavioural determinants
Behavioural determinants were assessed for two reasons: to tailor the programme to the determinants of the participant and to estimate the influence of those determinants on behaviour change. Participants completed a questionnaire with questions on selected determinants: knowledge, motivation, attitude, task and barrier self-efficacy, skills, perceived outcomes/benefits, physical environment, social influence, habits for dietary behaviour and physical activity and identity/values/norms (see online Supplementary Material for the questionnaire). The questions are based on an adapted version of the Determinants of Physical Activity Questionnaire(Reference Taylor, Lawton and Conner62) and the ‘Self-Report Behavioural Automaticity Index’ for the determinant ‘habit’(Reference Gardner, Abraham and Lally63). We based our questions for our specific determinants on these existing questionnaires to reflect the aim of the intervention to increase adherence to the WCRF guidelines. Examples of included questions are: ‘I find it important to live healthily’ (determinant: attitude), and ‘The people in my surroundings who are important to me support me or encourage me to live healthily’ (determinant: social influence). Each determinant is scored on a scale from 1 to 7. Higher scores indicate a more positive evaluation of that underlying determinant of behaviour (e.g. better knowledge about guidelines, sufficient skills to prepare healthy foods, etc.), and a score below 5·5 on any of the determinants was considered a point of attention in the lifestyle programme.
The behavioural determinants that we examined were chosen as follows. In our systematic review(Reference de Vries-ten Have, Winkels and Zondervan21), we identified knowledge, motivation, social support, norms and influence, beliefs (included under attitude), skills, environmental factors (i.e. weather and lack of access to facilities and equipment), dealing with symptoms related to the disease and/or treatment (included under barrier self-efficacy), and perceived outcomes/benefits as important population-specific determinants of behaviour. We also included attitude (i.e. including general beliefs about healthy living), self-efficacy (i.e. task self-efficacy), habits and identity/values/norms as those factors were deemed important according to the colorectal cancer survivors whom we interviewed. Specifically, changing habits is important because substituting unhealthy habitual behaviour with healthy habits significantly predicts long-term behaviour change(Reference Wood and Neal64,Reference Verplanken and Orbell65) . In addition, the determinant ‘identity, values & norms’ is important, since behaviour that reflects one’s identity and values and adheres to social norms is more meaningful, creates a sense of belonging and becomes integrated into a person’s sense of self(Reference Kwasnicka, Dombrowski and White24).
Biospecimens
Participants were asked to collect faecal samples into collection tubes (Sarstedt faeces tubes with sterile screw caps). The participant collected the samples within 3 days prior to the home visit and stored the sample in an air-tight plastic bag in a freezer. A member of the study team transported the samples on dry ice to the research facility where it is stored at −80°C until further analysis. We plan to use 16S rRNA sequencing data for the taxonomic characterisation of gut microbiota, possibly with shotgun metagenome sequencing for further characterisation.
Blood pressure
Blood pressure was measured during the home visits by a member of the research team using a sphygmomanometer (OMRON M2, Omron) on the non-dominant arm. The participant was asked to sit at rest for 5 min before blood pressure was measured.
Haemoglobin
Haemoglobin was assessed as we previously identified this as a potential contributing factor for cancer-related fatigue in this population(Reference Kiebach, de Vries-ten Have and van Duijnhoven66). A finger-prick blood sample was taken during the home visits by a member of the research team and analysed immediately to assess haemoglobin as an indicator of anaemia using a HemoCue Hb 201+ (HemoCue AB).
Physical performance
The following tests were conducted during the home visits. We chose tests that were feasible in an area with limited space and that required minimal materials.
1: Three-minute step test: indicates cardiopulmonary fitness by heart rate measurement for 1 min after the completion of the 3 min of stepping(Reference Bennett, Parfitt and Davison67).
2: Five times sit-to-stand test: assesses the time necessary to achieve the standing position after the fifth repetition(Reference Bohannon68), which is considered a measure of strength of the lower extremities.
3: Tandem test: measures balance while standing in three different positions for at least 10 s(Reference Koyama, Tanabe and Itoh69).
4: Chair sit-and-reach test: measures the flexibility of lower extremity and lower back(Reference Jones, Rikli and Max70).
Participants sat on a chair with one leg bent at a 90° angle and the other leg extended. They reached towards their toes with the hand on the same side as the extended leg. The distance between the hand and toes was measured. This was repeated on the other side. The side with the greatest stretching was measured twice, and the mean value will be used.
5: Hand grip strength using Jamar dynamometer(Reference Mathiowetz, Weber and Volland71): participants sat in a chair with the dominant arm unsupported next to their body positioned at a 90° angle. Measurements were taken twice, and the mean value will be used.
Other study parameters return
Sociodemographic information (age, sex, marital status, education, employment and smoking) is collected via standardised questionnaires online, and other personal parameters (medicine use and use of dietary supplements) are collected via standardised questionnaires during the baseline visit; during the following assessments/visits, the research team asks whether there are changes in those personal factors. Clinical data are collected through questionnaires and linkage with the Netherlands Cancer Registry.
Process evaluation
During and after the trial, we will investigate the experiences of twenty participants and the lifestyle coaches with the programme by means of interviews. With these interviews, we gather information on the barriers and facilitators experienced during the intervention, about the reach, dose delivered and received, about the fidelity and about the acceptability of the intervention(Reference Moore, Audrey and Barker72).
Costs and cost-effectiveness
We collected the following data on costs through an online self-reported questionnaire: medical costs and costs for lifestyle behaviours, productivity at work(Reference Bouwmans, Krol and Severens73), unpaid/volunteer work, and data on how much participants would want to pay for participation in the lifestyle programme. These data will allow us to evaluate cost-effectiveness of the 6-month lifestyle programme using a similar approach as previously used by researchers of our group(Reference Dorhout, Haveman-Nies and van Dongen74,Reference van Dongen, Haveman-Nies and Wezenbeek75) .
Data analysis of primary outcome
The primary study outcome is cancer-related fatigue at 6 months. To test the effect on the primary outcome, we will perform an ANCOVA, assuming that the assumptions of normality of residuals and homogeneity of variance are met. In the ANCOVA analysis that will be conducted on an intention-to-treat basis, we will compare cancer-related fatigue at 6 months between intervention and control group and adjust for cancer-related fatigue at baseline, and results will be reported as effect sizes and 95 % CI. We will include the stratification factors used during randomisation in the ANCOVA model(Reference Kahan and Morris76), assuming similar variance in cancer-related fatigue across stratification factors. We will apply multiple imputations to impute missing data of cancer-related fatigue at baseline or 6 months.
Several sensitivity analyses are planned for the primary outcome. As sensitivity analysis, we will conduct a ‘complete case’ analysis in which we will only include participants of whom we have complete data on cancer-related fatigue at baseline and 6 months. Moreover, we will conduct a per-protocol analysis in which we will only include those participants who were adherent to the protocol. Full adherence is defined as having at least eleven sessions with the coach.
Data analysis of secondary outcomes and mediation analyses
Secondary outcomes
We will explore the effect of the intervention on skeletal muscle fat infiltration (echo intensity data), sleep quality and behaviour, quality of life, depression and anxiety using a series of ANCOVA models using the same approach as described for the primary outcome. In addition, we will explore the prolonged effects of the intervention at 12 months.
Mediation analyses
We will examine whether changes in dietary behaviour, physical activity and/or skeletal muscle fat infiltration act as mediators in the relationship between the lifestyle intervention and cancer-related fatigue calculating percentile bootstrap CI for indirect effects(Reference Andrew77). We also plan to evaluate whether changes in behavioural outcomes are mediated by changes in behavioural determinants.
Discussion
Cancer-related fatigue can affect many aspects of health-related quality of life. The SoFiT trial aims to assess whether a personalised lifestyle intervention focused on a healthier dietary intake, and more physical activity can help to reduce cancer-related fatigue among colorectal cancer survivors. The SoFiT lifestyle programme has several unique features. The first is that it is rooted in behaviour change principles. The systematic design of the programme is based on behavioural determinants of a healthy lifestyle in colorectal cancer survivors and on evidence- and theory-based behaviour change techniques. Furthermore, the programme has a person-centred approach as it is tailored to the lifestyle of the individual, to behavioural determinants, and to the preferences, opportunities, and disease-related barriers of the individual. Tailored approaches are needed to take into account the complexity of lifestyle behaviours: the precision health approach(Reference Hekler, Tiro and Hunter78).
The second unique feature is that only persons who experience cancer-related fatigue were eligible for the study. This is an important strength as very few previous trials were conducted among populations experiencing cancer-related fatigue. As a consequence, participants in previous trials were mostly high functioning and were not or hardly experiencing fatigue, which limited the possibility of whether lifestyle can affect cancer-related fatigue.
The third unique feature is that we assessed a wide range of outcomes. Cancer-related fatigue often occurs in symptom clusters with symptoms such as disturbed sleep quality(Reference Legg, Meertens and van Roekel54), emotional distress and other concerns(Reference Berger, Mitchell and Jacobsen55,Reference Han, Yang and Syrjala79) , and we therefore took a wide variety of outcomes into account in this trial.
The fourth unique aspect is that we included outcomes that focus on physiological changes while we also assess changes in behaviour and behavioural determinants to gain more insight into the mechanisms of action. As a physiological outcome, we assess skeletal muscle echo intensity as an indicator of skeletal muscle fat infiltration. Previous studies on body composition among colorectal cancer patients/survivors used clinical computed tomography images to gather information on fat infiltration in skeletal muscle(Reference van Baar, Bours and Beijer48,Reference van Baar, Beijer and Bours80,Reference Brown, Caan and Meyerhardt81) . In the context of intervention studies, the use of computed tomography images to track changes in fat infiltration is problematic: computed tomography includes exposure to radiation, and a computed tomography machine is not portable. We therefore chose to use ultrasound to obtain information on skeletal muscle echo intensity as an indicator of fat infiltration in skeletal muscle.
Recruitment for the SoFiT trial started in November 2021 and was completed by June 2023. We expect to complete data collection of the 12-month time point by June 2024.
Acknowledgements
The authors would like to thank the following persons: Petra Kellerhuis of Stichting Darmkanker; Annelies Visser, Anne Marie Bloo, Jolanda Aerts, Iris Krabbenborg, Lara Schepers, Sharon Bloemhof, and all Master students Nutrition and Health who contributed to the data collection for the project. The authors thank the project team of PLCRC (Prospectief Landelijk CRC Cohort) and collaborators at Hospital Gelderse Vallei Ede, Rijnstate Hospital Arnhem, FlevoHospital Almere, Hospital Deventer and Slingeland Hospital Doetinchem for their contribution to the recruitment.
This work was supported by grant [IIG_2019_1981] from the World Cancer Research Fund (WCRF) and Wereld Kanker Onderzoek Fonds (WKOF), as part of the World Cancer Research Fund International Grant Programme and by internal funding of the Division of Human Nutrition and Health from Wageningen University & Research. Funders are not involved in the study design, in the collection, analysis, and interpretation of data, or in the publications that will result from this study.
Formulation of research questions and design of the study: ALL.; carrying out the study: J. dV.-tH., K. M., A. V., L. W. and R. W.; writing of the manuscript: J. dV.-tH. and R. W; all authors provided input on and approved the final version of the manuscript.
The authors declare no conflicts of interest.
Supplementary material
For supplementary material/s referred to in this article, please visit https://doi.org/10.1017/S0007114524001107
Open access
