Enhancing adherence in trials promoting change in diet and physical activity in individuals with a diagnosis of colorectal adenoma; a systematic review of behavioural intervention approaches

Background Little is known about colorectal adenoma patients’ ability to adhere to behavioural interventions promoting a change in diet and physical activity. This review aimed to examine health behaviour intervention programmes promoting change in diet and/or physical activity in adenoma patients and characterise interventions to which this patient group are most likely to adhere. Methods Searches of eight databases were restricted to English language publications 2000–2014. Reference lists of relevant articles were also reviewed. All randomised controlled trials (RCTs) of diet and physical activity interventions in colorectal adenoma patients were included. Eligibility and quality were assessed and data were extracted by two reviewers. Data extraction comprised type, intensity, provider, mode and location of delivery of the intervention and data to enable calculation of four adherence outcomes. Data were subject to narrative analysis. Results Five RCTs with a total of 1932 participants met the inclusion criteria. Adherence to the goals of the intervention ranged from 18 to 86 % for diet and 13 to 47 % for physical activity. Diet interventions achieving ≥ 50 % adherence to the goals of the intervention were clinic based, grounded in cognitive theory, delivered one to one and encouraged social support. Conclusions The findings of this review indicate that behavioural interventions can encourage colorectal adenoma patients to improve their diet. This review was not however able to clearly characterise effective interventions promoting increased physical activity in this patient group. Further research is required to establish effective interventions to promote adherence to physical activity in this population.


Background
Colorectal cancer is the third most common cancer in the UK, the second most common cause of cancer death and its incidence [1] is increasing. Most colorectal cancers arise from polyps or adenomas, and high-risk adenomas (HRA) are the most likely to become cancerous [2]. One of the aims of the National Health Service Bowel Cancer Screening Programme (NHSBCSP) is to detect and remove colorectal adenomas and thus improve survival [3]. Whilst adenoma removal reduces the risk of colorectal cancer, the underlying risk factors that influence recurrence of ademona remain and the recurrence rate for adenoma has been shown to be relatively high at around 40 % after three years [4].
There is consistent evidence from observational studies that high (>500 g per week) dietary red and processed meat intake and low levels of physical activity cause colorectal cancer [5]. These risk factors are potentially modifiable and behavioural interventions which encourage change in diet and physical activity may reduce risk of recurrence of colorectal adenoma and development of colorectal cancer [6][7][8][9].
Through the introduction of the National Health Service Bowel Cancer Screening Programme the rates of detection of adenomas is likley to increase. As such identification of effective interventions to change behaviour associated with risk of colorectal adenoma in this patient group are becoming increasingly important.
Evidence suggests that interventions for populations at increased risk of disease are more likely to be successful than in healthy populations. Compared with the general population, patients with a previous diagnosis of colorectal adenoma are at increased risk of colorectal cancer. This patient population is different to the general population since they have received screening and surgical intervention to remove adenomatous polyps. As such, findings from trials of health behaviour interventions in the general population are unlikely to be generalisable to this patients group.
Inadequate adherence in clinical trials contributes to significantly increased study costs, complicates statistical analysis and threatens study validity [26][27][28]. Clinical trials of behavioural interventions frequently suffer from low levels of adherence with estimates suggesting that between 25 and 50 % of research participants are not adherent [26]. Broadly, adherence can be defined as the extent to which a trial participant acts in accordance with the instructions or recommendations of the research as specified in the study protocol.
The current literature review was undertaken to examine behavioural intervention programmes and determine adherence in RCTs promoting a reduction in consumption of red meat, elimination of processed meat and increased physical activity in individuals with a diagnosis of colorectal adenoma. The aim was to define diet and physical activity interventions to which colorectal adenoma patients are likely to adhere and to use these in the development of a large prospective RCT to assess whether the interventions are effective in changing health behaviour associated with risk of colorectal adenoma.
To achieve this aim it was necessary to i) identify RCTs of dietary and/or physical activity interventions promoting risk reduction in individuals with a diagnosis of colorectal adenoma, ii) summarise data related to protocol adherence and follow-up in these RCTs and iii) characterise the behavioural interventions or elements of these interventions which achieved and sustained maximum adherence.

Search methods to identify relevant studies
An electronic search of eight databases (Pubmed, Cochrane, Medline, Embase, PsychINFO, HMIC, Cinahl and BNI) was conducted to capture relevant publications (searches last conducted October 2012). Detailed search strategies were developed for each database (Table 1). Searches were limited to studies involving humans, in English language and published since 2000. Significant advancement in health behaviour research and technology has been made over recent years. This time frame was chosen to enable identification of trials of health behaviour interventions which are most applicable and relevant to a contemporary cohort of patients with colorectal adenoma. All retrieved articles were reviewed to identify additional, relevant RCTs. To ensure consistency in selection, the titles and abstracts of all papers retrieved via the searches were reviewed independently by two reviewers. Papers that did not fulfil the selection criteria were excluded. Full papers were obtained for the remaining studies and two reviewers read and independently applied the selection criteria. The two reviewers met to resolve any disagreement and reach consensus.

Selection criteria
Inclusion criteria (i) RCTs with a population of adults with a previous diagnosis of colorectal adenoma without a previous diagnosis of colorectal cancer. (ii)RCTs which evaluated a behavioural intervention aiming to promote change in physical activity and/ or diet. (iii)RCTs reporting data related to adherence as either a dichotomous or continuous variable.
Other outcomes of interest were retention, attrition and reasons for drop-out. RCTs were not excluded, however, if data related to these outcomes were not reported. Meta-analysis and systematic reviews were employed as sources of additional RCTs only.

Exclusion criteria
(i) RCTs in cancer patients or cancer survivors (ii)RCTs of prevention in cancer patients (iii)RCTs in which adherence data could not be extracted.

Quality assessment
The quality of each included RCT was assessed using the Critical Appraisal Skills Programme RCT checklist [29]. The quality of each included RCT was assessed by two of the reviewers (JJ and RH) with disagreements being resolved by discussion.

Data extraction
For each of the included RCTs, the paper was read in full by two reviewers (DM and AS). Data were extracted using a proforma specifically designed to record key information related to (i) study design (ii) population characteristics (iii) characteristics of the intervention including: type of intervention; mode, location and delivery of interventions; (iv) type of intervention provider (v) duration, intensity and frequency of the intervention. Data to enable calculation of adherence, frequency and methods of assessment of adherence and reasons for drop out were also extracted.

Outcomes of interest of this review
There were four main outcomes of interest of this review. Firstly, this review focused upon whether participants received/attended the intervention or its components, as described in the study protocol. Participants needed to have attended or engaged with each of the scheduled components of the intervention to be considered fully adherent in this outcome (intervention adherence). The second outcome of interest was the extent to which participants met the dietary and/or physical activity goals of the intervention. To be classified as adherent for this outcome, participants had to adhere to ≥50 % of the diet and/or physical activity goals of the intervention. In health behaviour, it is difficult to give a precise definition or cut-off for when behaviour is deemed acceptable or not and this may vary from one context or population to another. A judgment on what such a cut-off might be was therefore required. Following much discussion and consideration, a minimum threshold of 50 % was selected because this meant at least half of the sample had achieved at least half of the intervention. This was considered in light of the fact most people in the modern Western world are sedentary and do very little physical activity-so a shift in physical activity from very little to a minimum adherence of 50 % of a physical activity intervention is not insignificant and even small changes in behaviour can be clinically worthwhile [30]. Given that participants who do well in the intervention are more likely to agree to follow-up, the third outcome was the follow up rate in the intervention group to enable comment upon the burden and acceptability of the intervention. A fourth and final outcome of interest was reported reasons for drop out.

Methods of synthesis
Since the focus of this review was identification and characterisation of behavioural interventions that maximise adherence in RCTs promoting behavioural change in adenoma patients, it was not appropriate to conduct a statistical analysis. Data were therefore subject to a narrative synthesis.  further 1206 of these articles were excluded following review of the title or abstract and 196 articles were excluded after a full review of the article. The reasons for exclusion are provided in Table 2. The 14 remaining articles reported on nine RCTs which included individuals with a diagnosis of colorectal adenoma. Two of these RCTs were excluded from further review because they reported on RCTs of a dietary supplement and two RCTs were excluded because calculation of adherence was not possible. Five RCTs of a diet and/or physical activity intervention in colorectal adenoma patients were included in the current review [31][32][33][34][35].

Description of included trials
The characteristics of the five RCTs included are summarised in Tables 3, 4

and 5. The Minnesota Cancer
Prevention Research Unit (Minnesota CPRU) [31] trial and the Polyp Prevention Trial (PP trial) [32,36] evaluated the impact of a behavioural intervention upon diet alone and the Bowel Health for Better Health (BHBH) [34], PREVENT [33] and the BeWEL [35] trials examined the impact of a behavioural intervention upon diet and physical activity (Tables 3, 4 and 5). In total, 1932 adenoma patients were randomised to receive these behavioural interventions. The majority of trial participants were aged 40 years or more, Caucasian and had received at least 15 years of education. All five publications reported that the behavioural interventions were successful in achieving change in diet and/or physical activity in adenoma patients ( Table 3).

Characteristic of the behavioural intervention
In all five RCTs, participants were asked to meet or exceed current diet and/or physical activity recommendations for risk reduction at the general population level ( Table 4). The intervention in each of the five RCTs comprised a combination of behavioural, educational and affective approaches to promote behavioural change. Behavioural components of the intervention were based upon cognitive behavioural psychology and employed techniques such as negotiation and goal setting and encouraged planning, self monitoring and skill building. In addition, the Minnesota CPRU, PREVENT and BeWEL trials provided positive reinforcement and feedback. The Minnesota CPRU trial also used fridge magnets and birthday cards as      Subjective and objective, self reported daily diary and food frequency questionnaire measurement body weight, waist circumference, blood pressure, and of biological markers (e.g., total, low and high density SenseWear armband worn for 7 days to measure daily expenditure and minutes of moderate intensity exercise. Trained lifestyle counsellor recorded attendance a Intervention is effective for promoting behavioural change in adenoma patients based upon ≤50 adherence to the behavioural goals of the intervention memory aids to maintain motivation and adherence. Tool kits of items such as pedometers and shopping bags and water bottles with trial logos were provided to participants of the BeWEL trial. Other equipment such as weighing scale, kitchen gadgets, physical activity equipment (e.g., exercise DVDs, hand weights and hoola hoops) were available, on loan also. The educational materials delivered as part of the diet intervention generally provided information on nutrition and advice on ways to modify lifestyle to concur with target recommendations of the intervention. To highlight the importance of risk factor reduction, the PREVENT intervention provided information on personalised risk profiles in addition to distribution of general literature related to cancer prevention. Affective components of the intervention focused upon development of coping skills, confidence and self efficacy and provision of emotional support. In the Minnesota CPRU, BHBH and BeWEL trials support from a friend or partner was encouraged. Diet interventions were delivered by dedicated dieticians and/ or nutritionists. Trained lifestyle counsellors delivered the diet and physical activity intervention in the BeWEL trial. No exercise experts were involved with development and/ or delivery of the physical activity interventions. The interventions were delivered at individual counselling session in the Minnesota CPRU, PP, BHBH and BeWEL trials. The PREVENT trial employed a combination of individual and group sessions.

Intervention adherence
Intervention adherence was reported in the Minnesota CPRU, PREVENT and BeWEL trials only. Full intervention adherence was not, however, achieved in either of these trials. In the Minnesota CPRU trial, 93 % intervention adherence was reported based upon attendance at all four intervention visits. The PREVENT trial reported that 60 % of participants received four of the five counselling telephone calls. The BeWEL trial reported that 97 % attended all the face to face sessions (3 sessions) and 59 % completed all of the 9 planned telephone calls ( Table 5).

Adherence to the behavioural goals of the intervention
Across the five RCTs, adherence to the dietary goals of the intervention ranged from 18 to 86 % and adherence to the physical activity goals of the intervention ranged from 13 to 47 % in the RCTs encouraging increased physical activity (Table 5).
In terms of effectiveness, the Minnesota CPRU, BHBH and BeWEL interventions were successful in achieving ≥ 50 % adherence to the behavioural goals of the intervention. In the Minnesota CPRU, diet only interventions achieved 86 % adherence to the fruit and vegetable goals of the intervention. The BHBH intervention, which promoted change in both diet and physical activity, was more effective with respect to diet, achieving 84 % adherence to the fruit and vegetable goals, 53 % adherence to the fibre goals and only 47 % adherence to the physical activity goals of the intervention. The BeWEL diet intervention achieved 73 % adherence to the fruit and vegetable goals. The PRE-VENT intervention, which promoted change in both diet and physical activity, was ineffective and failed to achieve adherence of ≥50 % with respect to any of the behavioural goals of the intervention. The effectiveness of the PP intervention could not be defined because adherence was assessed at multiple points and divided into three subgroups based upon total number of goals met during the trial period (Table 5).

Follow-up rate
Follow-up rate was generally high, ranging from 78 to 89 % in the RCTs of promoting change in diet and 78 %-91 % in RCTs encouraging change in both diet and physical activity. The reasons for withdraw or loss to follow-up were reported in the Minnesota CPRU, BeWEL and PP trials only. The Minnesota trial reported that 2 % of participants were inappropriately randomized and a further 10 % withdrew or were lost to follow-up. In the PP trial, 4 % were lost to follow-up. In the BeWEL trial, 9 % withdrew (Table 5).

Reasons for drop out
Only the BeWEL and PP trials reported reasons for drop out. 7 % of the PP trial participants discontinuing due to illness, no longer wishing to participate or moving to a health centre not participating in the trial (Table 5).

Methodological quality of the included trials
A meta analysis of trial data was not possible due to the heterogeneity in trial design and outcomes reported. Data related to trial quality was therefore subject to narrative synthesis. Trial quality was assessed using the Critical Appraisal Skills Programme RCT checklist and all trials were considered to be of high quality (scores ranging from 7.5 to 9 out of 10). The lack of reporting of research personnel blinding and reasons for participant withdraw from the study were the most commonly recorded methodological weaknesses. Two of the RCTs also failed to provide details of the required sample size and/or to comment upon whether the study was adequately powered to detect a significant difference between the two study arms [31,32].

Summary of main findings
This review identified two behavioural interventions that were effective in achieving ≥50 % adherence to a diet intervention and encouraging change in fruit and vegetable intake in colorectal adenoma patients. The effective diet interventions were grounded in social cognitive theory with the initial intervention counselling session being delivered individually during a clinic based consultation. These interventions also encouraged participants to identify social support and provided personalised, printed educational materials and recipes to aid behavioural change.
However, the physical activity interventions reviewed did not achieve similar levels of adherence to the goals of the intervention and as such, were ineffective for promoting increased physical activity in colorectal adenoma patients. Inaccuracies in self reporting of adherence due to recall bias and discrepancies between researcher and participant definition of moderate intensity activity may explain low adherence rates to these physical activity interventions. Of additional note is the lack of involvement of an exercise specialist in development and/or delivery of these physical activity interventions.
Intervention adherence could not be compared across all five RCTs due to either lack of reporting or heterogeneity in reporting. Furthermore, full intervention adherence was not achieved in any of the RCTs reviewed.
Of particular note, the follow-up rate in all five trials was high which may indicate that some aspects of the interventions used in these trials are acceptable to this patient group. However since the trials all employed complex interventions and do not report adherence to individual components of the intervention it is difficult to identify which of the components were more acceptable than others.
Overall, data relating to intervention adherence and reasons for drop out provided little insight with respect to the characteristics of the interventions to which this patient group are most likely to adhere. The five RCTs reviewed were relatively homogeneous with respect to the demographics of the populations studied and the nature, content and target recommendations of the behavioural interventions. However, these RCTs were heterogeneous regarding the timing of the intervention in relation to the diagnosis of colorectal adenoma, the duration of the RCT and the intensity of delivery of the intervention. Overall, the methodological quality of the included RCTs was good.
The physical activity and diet interventions were very similar with respect to the behavioural, educational and affective strategies employed to promote behavioural change. Furthermore, these strategies have been shown to be effective in increasing adherence to physical activity in cancer patients and survivors [37,38]. The reasons why the physical activity interventions were less effective in this population are therefore unclear. It is possible that colorectal adenoma patients perceive change in diet to be more easily achievable than change in physical activity, and as such, require greater motivation, self efficacy and confidence to adhere to target recommendations for physical activity. A more structured, maximum contact, patient focused, physical activity programme than that provided by the PREVENT and BHBH interventions may therefore be necessary to promote adherence in this patient group.
There is evidence to suggest that colorectal adenoma patients are largely unaware of the implications of their diagnosis and may not view themselves as being at increased risk of colorectal cancer [39][40][41][42]. Since people are more amenable to behaviour change following a health event or scare [43], it is possible that improved communication of personal risk with respect to recurrence of colorectal adenoma and progression to colorectal cancer at the time of colorectal adenoma diagnosis will enhance adherence to behavioural interventions in this population. Personalized risk information was provided in one of the behavioural interventions; unfortunately this intervention was ineffective with regard to promoting adherence or change in this population. However this intervention did target five other risk factors for colorectal adenoma recurrence in addition to encouraging increased physical activity. This provides another possible reason why it was ineffective for promoting change in physical activity in colorectal adenoma patients. Interventions promoting change in multiple risk factors are inherently more complex to deliver and assess and results are more difficult to interpret. From the patient perspective, change in multiple behaviours is also much more challenging and additional barriers and facilitators to change need to be considered when designing behavioural interventions to promote change in multiple risk factors.

Strengths and limitations
This is the first systematic review to examine behavioural intervention programmes and determine adherence to interventions which promote change in diet and/or physical activity in adenoma patients. As such, this review has provided a useful insight into the ability of adenoma patients to adhere to diet and physical activity interventions. Further research is however required to identify physical activity interventions to which colorectal adenoma patients are likely to adhere. Only three of the five RCTs eligible for inclusion in this review examined the impact of a behavioural intervention upon physical activity making it difficult to draw any meaningful conclusions. Moreover, only one RCT promoting a reduction in red meat consumption was identified meaning the aims of this review were not fully met. However, this review did capture all RCTs of diet and/or physical activity in colorectal adenoma patients published in the last 14 years. There is a risk of publication bias because unpublished RCTs were not included. Similarly, limiting literature searches to English language publications may impact upon language bias. However, the effect of this is likely to be small as it is unlikely many if any studies were missed that would have been included in this review.

Conclusion
This review identified two interventions which were effective in colorectal adenoma patients for promoting change in diet and successfully achieved at least 50 % adherence to the goals of the diet intervention. However, this review failed to identify effective interventions for promoting adherence to physical activity in this patient group. Derivation of a physical activity intervention to which colorectal adenoma patients are likely to adhere was, therefore, not possible. Future research should focus upon interventions promoting change in physical activity alone and which involve an exercise specialist in the design and delivery of the intervention. Provision of personalized risk information should also be considered to promote adherence to physical activity interventions in this patient group.