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

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.

Table 1. Search terms

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.

Results of the search

Figure 1 shows the outcome of the search process and application of the selection criteria.
The electronic searches identified 2221 potentially relevant articles. Following removal
of 805 duplicates, 1416 papers remained. A 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]–35].

Fig. 1. Results of the search strategy

Table 2. Reason for exclusion of papers

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).

Table 3. Characteristics of included trials

Table 4. Characteristics of the intervention

Table 5. Adherence outcomes

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 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 PREVENT 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].