Background

Many developed countries are witnessing increasing rates of diabetes and cardiovascular diseases. Evidence to support adoption of healthy lifestyles in the prevention and management of these and other long-term conditions is strong [1–3]. As an approach to managing cardiovascular disease risk, promotion of healthy lifestyles is arguably the preferred first line option over medication, which although effective in reducing cardiovascular risk [4], frequently has side-effects [5] and offers benefits only with continued adherence. A number of countries, including the UK, [6] now offer cardiovascular health checks through systematic screening programmes based in primary care. A variety of modalities exist but commonly they offer individuals a chance to modify their lifestyle behaviours and so invest in their future health.

Despite its importance, many individuals fail to engage in activities designed to help them to achieve a healthy lifestyle. For example, only around one third of patients experiencing a cardiac event take up cardiac rehabilitation [7] and up to one quarter of participants in commercial weight management programmes drop out [8]. The individual factors that can dissuade individuals from achieving lifestyle change are multiple and inter-related but broadly cluster as social, psychological and practical barriers. Compounded by the physically addictive nature of some behaviours, lifestyle change is challenging, requiring support and personal determination. The potential myriad of personal barriers to lifestyle change presented to health care providers during consultations are challenging to address [9]. Lack of skills and knowledge combined with short consultation times [10, 11] have the potential to generate generic responses that fail to meet individual needs. So, how can health care services develop a more skilled and focused approach to lifestyle behaviour change for individuals at high risk of cardiovascular events?

One option is to systematically target individually perceived key barriers to healthy lifestyles. Interventions specifically aimed at increasing initiation of lifestyle change [12–14] or uptake of associated programmes [15, 16], tend to address patient perceived barriers as they arise in consultation and the concept of barrier removal has been applied naively as a component of a more complex approach. None have specifically developed a barriers-based framework as the core of the intervention. An alternative approach would be to use knowledge of barriers and facilitators to tailor care pathways such that existing resources were more closely matched to individual need. Such an approach requires greater clarity on key influences on lifestyle behaviour change.

As the first step in developing such an approach we conducted a review of the qualitative literature reporting patient experiences of lifestyle change [17]. This revealed 348 factors that patients considered to influence lifestyle change. These were aggregated into 20 categories in which we identified five key themes: emotions; psychological beliefs; information / communication needs; support from family and friends; and transport and other costs) [17]. While this information is clinically valuable, its routine integration into relevant patient consultations requires more robust evidence. As the next step in developing a ‘lifestyle referral assessment’ we considered it necessary to identify which of the many reported influences were not just perceived by patients to be important, but have been shown to be associated with lifestyle behaviour change. We hypothesised that key influences identified in the earlier qualitative review would also likely predict uptake of lifestyle behaviour change. Previous published literature reviews on predictors of lifestyle behaviour change have focused on uptake of cardiac rehabilitation [18, 19]. Through a review of quantitative observational correlation studies, we aimed to clarify more explicitly which patient centred-factors act on uptake and completion of formal lifestyle change programmes, attempting to draw on broader literature in for example diabetes and hypertension. Our review fulfils a dual need, to provide guidance to health care staff when referring individuals to lifestyle support (increasing uptake), as well as those delivering formal programmes (maximising retention).

Methods

Searching and study selection

An electronic search strategy was developed by the authors and run in Medline, HMIC, OVID Nursing, PsycInfo and Embase from 1980 to February 2010 (available from authors on request). Search terms included cardiovascular diseases, attitude to health and prevention. The start date of 1980 was selected to reflect early influential public health policies such as the 1984 BMA drivers to commence NHS smoking cessation programmes (which started in 1998) alongside existing and expanding telephone help lines. Search results were imported into Endnote (Version X2) and independent reviewers (JM, CC, SH, KH) undertook title and abstract screening against predefined selection criteria. Reference lists of eligible studies were hand-searched for additional systematic reviews. Stages of the review process involving multiple reviewers (JM, CC, SH, KH) included group discussions to resolve any discrepancies.

The inclusion criteria were primary or secondary quantitative research studies examining uptake (attending at least one session) or completion (attending all sessions) of lifestyle behaviour change programmes in adults (>18 years of age) having experienced angina, myocardial infarction or transient ischemic attack, or with hypertension, diabetes type II, coronary artery disease or hypercholesterolemia. Studies were excluded if they: investigated compliance with medication for cardiovascular risk management or long term maintenance of lifestyle change (these are arguably separate bodies of research and their inclusion would result in an unwieldy report); focused on a selected population such as mental health patients (regarded as a group with specialist needs); were culturally unrepresentative of the main ethnic groups residing in Europe; included only stroke, chronic obstructive pulmonary disease, peripheral artery disease, and heart failure patients (on the basis that physical limitations would have major influences on lifestyle choices) or; consisted entirely of Diabetes type I patients (most likely to involve adolescents). Primary research studies were further required to report statistical effects (as p values or 95% confidence intervals) of investigated factors on uptake or completion of lifestyle behaviour change. Likewise, reviews were included if they were statistical meta-analyses of selected factors or narrative reviews with clear reporting of statistical effects of factors from included identifiable primary research studies. If necessary, authors were contacted to clarify reported data.

Quality assessment

A quality assessment on primary research studies was conducted using a 14-item adapted version of the STROBE guidelines for assessing observational studies [20]. One reviewer assessed all papers, with a second reviewer assessing a subset of papers. Studies were deemed poor in quality and were rejected from the review if they met less than 50% of the quality criteria. Studies meeting between 50% and 70% were regarded as adequate and those scoring more than 70% were good.

Quality assessment of identified reviews comprised four criteria that were common to two existing meta-reviews [21, 22] and a previously published checklist [23]. Reviews lacking adequate methodological description to allow us to apply these criteria were rejected.

Data extraction

Data extraction was performed independently on eligible studies by two reviewers (CC and JM). Study details were extracted and the following information related to findings was recorded: outcome type (uptake or completion), factor (e.g., age, depression, marital status), statistical method, statistical significance (p<0.05 was deemed statistically significant), direction of relationship (barrier or facilitator to lifestyle change).

Extracted factors were organised into a framework derived from our previous review [17] comprising 20 categories. We split the categories into those that were regarded as key themes (mood, information and communication, support from friends and family, psychological beliefs, and transport and other costs) and other categories. An additional category covering demographic factors was also included. Extracted organised data were checked by a second reviewer. All outcomes were recorded in Microsoft Excel 2007.

Analysis

A narrative review was conducted due to the heterogeneity between studies, in terms of the clinical populations and diversity in outcomes and outcomes assessment. Data were organised according to whether they appeared to deter, facilitate or had no relationship with either uptake or completion of lifestyle programmes. Investigated factors for which there was no evidence of relationship with uptake or completion of lifestyle programmes were recorded separately.

Finally we compared the factors reported in the framework from the qualitative literature with those examined in the quantitative studies to see whether the patient reported factors were represented in the statistical studies.

Results

The electronic search generated 16,802 hits (after de-duplication).

Systematic reviews and quality assessment

Ten reviews (seven identified from electronic database searching and three through hand searching) underwent full paper screening. Six reviews were rejected as they either provided no description of their methods or reported very limited statistical results from included primary studies [18, 24–28]. One further review was rejected because it failed to indicate which primary studies reported non-significant predictors of uptake at cardiac rehabilitation [18]. Therefore three systematic reviews [19, 29, 30] reporting predictors of uptake at cardiac rehabilitation were eligible for inclusion. Two reported meta-analyses of individual predictors of uptake (attendees versus non-attendees) of cardiac rehabilitation [29, 30]. The third [19] was a narrative review summarising findings from 15 studies [31–45] on a range of variables. The individual findings from the third review were extracted and are reported alongside the findings from the identified primary studies. A quality assessment of these individual studies was not undertaken.

Primary studies and quality assessment

In total 166 papers underwent full paper screening of which 33 [46–78] met the inclusion criteria. Reasons for rejection included: paper predating or reported in the review by Cooper et al. [19] (n=72), paper not assessing cardiovascular prevention initiatives (n=19), outcomes were medication adherence or maintenance with no mention of uptake or completion (n=20); absence of statistical analysis between groups (n=18 papers), and other (n=4). The main characteristics of the primary studies are summarised in Table 1.

Table 1 Main characteristics of included 32 primary studies
Full size table

One paper which met the inclusion criteria was found to be poor in quality (scoring 5/14 (36%) and was rejected [58]. Five papers were considered to be of adequate quality [61, 62, 72–74] and 27 were good (Additional file 1).

Summary of findings

In total 374 factors were extracted from the 32 included primary studies. Twenty-three studies reported 253 factors relating to uptake, with the remaining 121 factors from 11 studies relating to completion. Five factors were co-reported in two systematic reviews and were therefore excluded. Three studies [51, 55, 76] reported on both uptake and completion of programmes. All studies were concerned with cardiac rehabilitation despite the broad scope of the terms defined in the electronic search strategy. Most studies reporting uptake failed to define what was meant by attendance or participation but all studies included a comparator group of non-attendees. In two studies, completion of programmes was defined as completing at least half [76] or two-thirds [55] of all the sessions with the comparator groups for these variables being those who dropped out earlier (as opposed to those who did not start the programme). All other studies reported completion as attendance at all sessions.

Key themes

Key themes that contained factors most consistently associated with uptake at lifestyle programmes were ‘friends and family support’, ‘transport and other costs’, and ‘psychological beliefs’ (Table 2). Problems with transport, perception of greater consequences to illness and attribution of more symptoms to illness were most consistently predictive of uptake. Absence of a partner, lack of employment, transport / distance problems, low self-efficacy and perceptions of less control of the illness were commonly predictive of non-uptake. Earlier studies suggest that lower educational attainment may predict non-uptake, however this has not been supported by the majority of more recent studies. Anxiety and depression were the most frequently reported emotional factors. None of the studies suggested that anxiety deters uptake of lifestyle change with one (out of eight) even suggesting it to be facilitative. Just under half the studies examining the relationship between depression and uptake found that affected individuals may be less likely to start a lifestyle programme. One study [74] indicated a facilitative role of depression in uptake, however it was not found to be independently predictive. Programme drop-out appears to be consistently associated with the presence of depression (Table 3).

Table 2 Studies reporting factors (organised into key themes) relating to uptake of lifestyle programmes
Full size table
Table 3 Studies reporting factors (organised into key themes) relating to completion of lifestyle programmes
Full size table

Within the five themes there were a number of factors that appeared to have no evidence of any relationship with uptake or completion of programmes. These included for example, knowledge of smoking cessation, living alone and occupation (Table 4).

Table 4 Factors from key themes showing no evidence of relationship with uptake / completion of CR programmes [references]
Full size table

Concordance between the lay concepts that defined the qualitative framework [17] and the factors that were reported in the quantitative observational studies appeared good for only one of the five key themes (‘transport and other costs’). For example, emotional factors in the qualitative framework included fear, lack of motivation, confidence and embarrassment while the quantitative studies reported only on anxiety and depression. Factors relating to ‘friends and family support’ were limited to partner status in the quantitative data as compared with the qualitative framework which reported more nuanced factors relating to the quality of support.

Other categories

Numerous factors relating to ‘physical wellbeing’ were reported in the quantitative studies. Despite this, the relationships between uptake of cardiac rehabilitation and either previous cardiac events, cardiac procedures or clinical risk factors remain unclear (Table 5). In contrast, the presence of co-morbid long-term conditions and poor physical functioning appears to deter uptake. While most of the evidence on body weight / mass index indicates that this has no relationship with uptake, there may an increased likelihood of overweight individuals dropping out of programmes (Table 6). Many of the factors relating to ‘physical wellbeing’ showed no evidence of relationship with either uptake or completion of lifestyle change programmes (Table 7). There was a paucity of studies reporting the relationship between uptake or completion of lifestyle change support and factors relating to ‘referrals’, ‘culture’, ‘social support’, ‘the role of the health care professional’, ‘attitudes to rehabilitation’, ‘attitudes to exercise’ and ‘balancing and integrating health care needs with everyday life’. Current smoking status (the only factor relating to ‘personal choices and cultural preferences’) produced conflicting results with regards to uptake and completion. As with ‘physical wellbeing’ there has been extensive research into the relationship between uptake, and both gender and age. Although there may be a trend towards non-uptake among females and older people, the majority of studies suggested no relationship.

Table 5 Studies reporting factors from other categories relating to uptake of cardiac rehabilitation
Full size table
Table 6 Factors reported in other patient centred categories relating to completion of programmes
Full size table
Table 7 Factors from other categories showing no evidence of relationship with uptake / completion of CR programmes [references]
Full size table

As with the key themes, concordance between the factors that informed the qualitative framework [17] and those reported in the quantitative data was in general poor. Patient reported factors in for example ‘personal choices and cultural preferences’ related to unwillingness to change habits and feelings of resentment from enforced changes in lifestyle. These were not addressed by quantitative studies, which only referred to smoking status.

Discussion

To our knowledge this is the first systematic review of the cardiovascular literature that attempts to map quantitatively defined factors to those that patients perceive to be important in deciding upon lifestyle behaviour change. The review is enhanced by this approach as it demonstrates the divergent foci of the two methodologies and reinforces the need to examine both bodies of evidence.

The value of our review is that it can inform a consideration of the skills required to fulfil increasingly active public health agendas across developed countries. For example, the UK Department of Health’s ‘Every Contact Counts’ (http://www.dh.gov.uk/health/2012/01/forum-response/initiative accessed 13/1/2012) is one of a number of policy led changes to the delivery of health promotion and prevention [79, 80]. A key requirement is for front line health care providers to seize every opportunity to promote and support individuals to adopt healthy lifestyles. Understanding key factors that influence lifestyles may improve the quality of these interactions and so increase initiation of lifestyle change.

Comparison with existing literature

An earlier review of factors associated with uptake of cardiac rehabilitation [19] found age, deprivation, beliefs and physician recommendations to be the main predictors. However in our review the majority of studies found no relationship between age and uptake of cardiac rehabilitation and we would suggest other factors such as referral patterns (potentially favouring younger patients) and lack of work commitments (in older patients) might contribute to this unclear picture. As in the Cooper review [19] we also found strong evidence that beliefs have an important role in influencing patient decisions about lifestyle change, as was support from family and friends. Previous reviews have been limited to uptake (often termed ‘attendance’) at cardiac rehabilitation. We have extended this to include completion thus enabling us to consider the trajectory of influences on lifestyle change beyond the initial stages. Although we could not glean much from the programme completion data, it does appear that in certain circumstances (depression, personal beliefs and possibly partner status) factors that deter uptake may also increase the likelihood of programme drop-out.

Ideally intervention studies aiming to support lifestyle behaviour change should attempt to address a range of potential barriers. A brief examination of previous intervention studies that have specifically sought to address barriers to lifestyle change reveals limited awareness of the main predictors of uptake of lifestyle change. Assessment of social (family) support and education were reported in two studies [14, 81]. Mood problems, transport, costs, and beliefs lacked mention despite them having a key role in influencing uptake and completion of lifestyle change.

Summary of main findings

Our initial hypothesis that key patient perceived barriers (relating to emotions, psychological beliefs, information / communication needs, support from family and friends, and transport and other costs) would be predictive of uptake of lifestyle behaviour change appears to hold true. In particular, problems with transport and attribution of greater consequences and more symptoms to illness were most consistently predictive of uptake. Other factors provide less clear evidence as to their role in either uptake or completion despite, as is the case with factors related to physical wellbeing, extensive examination. In comparison with uptake, completion of lifestyle behaviour change has received relatively little attention. From the data that is available, there is a suggestion that some factors mainly predictive of uptake are not necessarily the same as those predicting completion.

Although representation of patient reported factors in the quantitative studies was in general not good, in some instances this may not be problematic. For example, the relationship between depression and, to a lesser extent anxiety, and uptake / completion was a strong feature of the cohort studies but not the qualitative literature. Rates of co-morbid depression in adults with diabetes and heart disease are thought to be double those seen in non-affected populations [82, 83]. The increased risk of depression and thereafter poorer outcomes reinforces the importance of recognition of mood problems in these populations. The apparent lower profile of these types of mood problems in the qualitative literature may be due to self selection (of those with low mood out of interviewing) or stigma of personal disclosure during interviews.

Conversely, for other categories such as ‘formal support’, ‘friends and family support’, and ‘social support’, cohort studies had failed to address issues that appear to be of central importance to patients. Thus network size and marital status are reported but perceived support not.

Strengths and limitations

Despite our aim to examine lifestyle change in programmes other than cardiac rehabilitation, this was not achieved. We used the same search strategy to identify qualitative studies examining patient perceived influences in uptake of lifestyle change and found studies relating to diabetes and hypertension. This perhaps highlights the paucity of research into long-term conditions and changing lifestyle behaviours through formal programmes. In spite of this limitation we consider that these findings will be generalisable to the primary care population with risk factors for cardiovascular disease for two reasons. Firstly, the categories were derived from studies representing a broad range of cardiovascular related conditions. Secondly studies included in the current review reported patients attending outpatient cardiac rehabilitation who would therefore experience similar practical and social difficulties as those attempting to attend lifestyle change programmes in the community. Although the studies were of adequate quality, many failed to define what they meant by attendance and participation. We therefore assumed that the presence of a non-attending comparator group meant that these studies related to uptake. However, we do not know if uptake related to attendance at one or multiple sessions and we would question the meaningfulness of including data where attendance at for example one session was commonplace.

We did not formally investigate the possibility of publication bias because the majority of studies reported multiple factors. However, we did observe a broad range of significant and non-significant findings indicating little evidence for publication bias.

The studies reported in the current review that were derived from the systematic review by Cooper et al. [19] were not assessed for quality by either ourselves or in the previous review. We intended for our work to be an update and expansion of the original review and therefore did not re-examine earlier studies. This may have affected the interpretation of the results in some categories. For example, the ‘information and communication’ the majority of studies indicating that lower educational status correlated with poor uptake were reported in the Cooper review [19] but later studies have not confirmed this relationship.

Conclusions

In conducting this review, our aim was to gather robust information that would support the development of a lifestyle assessment to be used at the point of for example referral to lifestyle support. We consider that the evidence gathered supports the findings from our previous qualitative review that at least five areas (emotional status, access to transport and cost of programmes, knowledge, understanding and beliefs about the condition and healthy lifestyles, and the quality of personal support) are influential in lifestyle change. Narrowing down the field of factors that influence lifestyle behaviour change should support health care providers to garner the knowledge and skills required to achieve more productive consultations. Further work is required to formalise these areas into an approach that can characterise patient responses in these areas and guide decision making about the most suitable type (formal group programme versus individual support) and level (supported change versus self-managed change) of support. Further research is needed to examine the nature of potential relationships between the factors identified in this and our previous review, and completion of lifestyle change programmes other than cardiovascular rehabilitation. In the meantime we recommend that health care providers use the information provided in this review (see summary in Table 8) to guide decisions about referral to lifestyle support.

Table 8 Summary of clinical messages from key themes
Full size table

Authors' information

Jenni Murray, BSc, MSc, Phd, Senior Research Fellow.

Cheryl Craigs, PgDip, BSc, MSc, Doctoral Student.

Kate Hill, BSc, MSc, Phd, Senior Research Fellow.

Stephanie Honey, BSc, MSc, Phd, Research Fellow.

Allan House BSc. MBBS, MRCP, MRCPsych DM, Professor of Liaison Psychiatry.

Declarations

Ethical approvals: not required.