Which providers can bridge the health literacy gap in lifestyle risk factor modification education: a systematic review and narrative synthesis

The Australian National Primary Health Care Strategy [1] and Council of Australian Governments (COAG) Australian Better Health Initiative (ABHI) include key priority areas that focus on improved chronic disease prevention and screening of those with at least one risk factor for chronic disease [1]. Integral to achieving this is to increase health literacy particularly in relation to modifying the behavioural risk factors of smoking, nutrition, alcohol, physical activity and weight (SNAPW). The SNAPW risk factors are major contributing factors to the development of chronic disease worldwide [2‐4] and are the focus of a number of international policy initiatives such as the US Healthy People 2010 initiative.

Health literacy is described as the cognitive and social skills which determine the motivation and ability of individuals to gain access to, understand and use information in ways which promote and maintain good health [5]. Low levels of literacy in the Australian population are a significant problem with recent figures indicating that less than half (48%) of the adult population reached the minimum level of literacy and numeracy required to function on a daily basis in today’s society [6‐8]. This is consistent with research from UK and USA where 46% and 47% of the population respectively achieved the minimum level of literacy necessary [9, 10]. The picture is even worse in people from low socioeconomic backgrounds and this further compounds their disadvantage [6, 11]. Health literacy, as defined by Nutbeam [12] is likely to be present at much lower levels than literacy and numeracy. High levels of health literacy are associated with specific health promoting behaviours such as eating five portions of fruit and vegetables per day or being a non-smoker independently of age, education, gender, ethnicity or income [13, 14].

Simply providing people with information alone about modifying SNAPW risk factors is not usually enough to bring about lifestyle change [15]. Rather, a partnership approach between patients and providers, based on shared decision making and good communication, may be necessary for developing a sense of confidence and ability to change [12, 16]. Without adequate health literacy people may not have the capacity to self-manage their health and prevent the development of chronic disease through lifestyle risk factor modification.

In response to the National Primary Health Care Strategy [1] and National Preventative Health Strategy [17] there is a drive to improve the health literacy of Australians. Primary care is ideally placed to support lifestyle risk factor management and health literacy as 86% of the Australian population visit their GP at least once per year [18]. However addressing health literacy and SNAPW risk factor management in general practice is difficult; the average consultation time with a GP is 7–8 min shorter than the time necessary to provide smoking cessation counselling [19]. The tyranny of the urgent means that people may only present to the GP when sick leaving little or no time for prevention [20‐22].

The developing role of practice nurses and allied health professionals in the prevention of chronic disease provides an opportunity to tackle SNAPW risk factor management and poor health literacy in those at risk of developing chronic disease. We know from a previous systematic review on skill mix that substituting GPs with health professionals such as nurses or pharmacists can be effective in disease management and health promotion in older people [23, 24]. However it is not clear what impact the type of provider, such as dietician, diabetes educator or GP may have on the development of health literacy and associated SNAPW risk factor modification.

The aim of this systematic review and narrative synthesis is to determine how effective primary healthcare providers are at improving the health literacy of patients to make SNAPW lifestyle changes. A second aim is to discuss the drivers and barriers for health professionals trying to improving health literacy and risk factor modification in primary care.

A systematic review was undertaken. Studies were identified by searching Medline, Embase, Cochrane Library, CINAHL, Joanna Briggs Institute, Psychinfo, Web of Science, Scopus, APAIS, Australian Medical Index, Community of Science and Google Scholar from 1 January 1985 to 30 June 2009. Health literacy and related concepts were found to be poorly indexed in many of the databases so a list of key words and text words were developed and retested for use in the different databases, terms used in the Medline search are listed in Table 1. Hand searches were also conducted of four key journals: Patient Education and Counselling, Health Education and Behaviour, American Journal of Preventive Medicine and Preventive Medicine. Systematic reviews identified in the process were read and all papers that met the inclusion criteria for this review were added to the list of papers. The bibliographies of experimental papers included were screened to identify additional studies.

There were several key definitions used to scope and focus the review.

Studies were included in the review if they were published in English, between 1985 and June 2009, included males and females aged 18 years and over with at least one SNAPW risk factor for the development of a chronic disease. The interventions had to be implemented within primary health care as defined and the studies had to report an outcome measure associated with health literacy (knowledge, skills, attitudes, self efficacy, stages of change, motivation and patient activation) and a measure of SNAPW behaviour change. We could not identify established tools for measuring interactive and critical health literacy so we looked to the self management literature for instruments that measure the concepts of self-efficacy, patient motivation, confidence and broader social support such as the Diabetes Self Efficacy Scale, the Social Support Survey and measures of Prochaska and DiClemente’s Stages of Change Model [26].

Intervention studies were included in the review if they were randomised, quasi randomised controlled trials, controlled before and after studies or interrupted time series. In addition non-experimental studies were included in an extraction of barriers and facilitators of health literacy and SNAPW risk factor management, see Table 2 for organisational framework for the review.

The papers were screened by two researchers (AW and JT). A 10% sample of excluded studies was reviewed by a third reviewer (MH). Verification and data extraction were performed by two researchers (AW and JT), a quality assessment was performed using a published checklist [27] ( Additional file 1 and Additional file 2) by one reviewer (SD) and a 20% overlapping sample by a second researcher (AW). Data were extracted (AW and JT) into an MS Access™ database and included variables such as type of health professional, intervention description, duration and frequency of intervention and outcomes of interest. Interventions were coded into categories (group education, motivational interviewing and counselling, written material, mixed intervention, telephone or computer) and the intensity scored using a combination of frequency and duration of intervention. High intensity interventions were those with at least eight hours or contacts, medium intensity interventions had more than three hours or contacts but less than eight and low intensity interventions were those with up to three hours or contacts.

A vote counting approach to the synthesis was used. Each of the outcome measures of interest such as change in a SNAPW behaviour or health literacy measures were coded as significantly improved or not significantly improved based on the results reported in the paper for each outcome of interest. The outcomes were coded as a statistically significant improvement if the paper reported a positive change with a p ≤ 0.05. The tables report the total number of studies reporting that outcome measure as the denominator and the numerator is the total number of studies with a significant improvement in that outcome measure. This approach to the analysis has been used in other systematic reviews of complex interventions [23, 28, 29]

Drivers and barriers for providers involved in developing SNAPW health literacy were extracted from the 42 descriptive papers identified during the search by one researcher (SD) and the findings coded using the definitions from the Institute for Healthcare Improvement [25] and synthesised by two researchers (MH and SD). This review was conducted as part of a larger policy relevant review [30] and funded by a Stream 13 grant from the Australian Primary Health Care Research Institute.

The database searches yielded 4691 papers that were assessed for inclusion in the review and after the screening and verification stages data were extracted from 52 papers that described intervention studies to address health literacy and lifestyle risk factor modification provided by different health professionals, see Figure 1 for PRISMA [31] flow chart. The characteristics of the included studies are in Table 3. In addition to the 52 intervention studies qualitative data on drivers and barriers were extracted from the 42 papers identified describing descriptive studies of health literacy and SNAPW risk factor modification, including facilitators and barriers.

Most of the studies (71%, 37/52) demonstrated an improvement in health literacy, see Table 4. Overall, health literacy and SNAPW risk factor were both improved for 61% (14/23) of interventions to address nutrition, 54% (15/28) for physical activity, 43% (3/7) for weight and 40% (6/15) for smoking. When interventions were grouped according to the health professional providing the intervention, 33% (3/9) of the studies reporting interventions provided by doctors resulted in an improvement in health literacy compared to interventions provided by other health professionals such as dieticians, educators or nurses (92% 11/12) and multidisciplinary teams (91% 10/11). When the interventions were categorised into low, medium and high intensity it became clear that different types of health professionals tended to provide interventions of varying intensities according to our definition. For example, 71% (5/7) of the interventions provided by doctors were categorised as low intensity. These interventions tended to be motivational interviewing and counselling around smoking cessation and physical activity prescription and were often only one session with goal setting and were described as brief interventions [19, 32‐35]. In contrast 80% (8/10) of the interventions provided by nurses, dieticians or educators and 90% (9/10) provided by multidisciplinary teams were categorised as medium or high intensity. These interventions were often motivational counselling or group education programs that took place over a number of weeks and targeted smoking, nutrition or physical activity [36‐52]. These interventions improved health literacy (10/11) although the effect on SNAPW risk factors was a little less with 8/11 reporting an improvement. Of the studies involving a lay worker, alone or as part of a multi-disciplinary team, 71% (5/7) targeted people from ethnic minority backgrounds. Overall, the included studies were of medium quality (36/52), 11 were high quality and five studies were of low quality. See Table 5 for a summary of the included studies.

A number of barriers and drivers were identified that related to the providers ability to provide SNAPW health literacy interventions. The barriers and drivers can be grouped under three main headings: provider context, provider costs and interaction between providers and patients. There were 32 papers describing provider factors, 27 describing provider and service context and 20 that described barriers and drivers at the provider patient interface. The barriers and drivers are listed in Table 6. Provider barriers included lack of knowledge or skills in preventive medicine and provider attitudes to providing this type of care. Linked to this were barriers and drivers around provider context such as support for professional development and funding mechanisms for health education. Many of the drivers and barriers around the patient provider interface relate to their relationship, trust and continuity of care.

The results from this review highlight the complex relationships between providers and interventions to develop health literacy of patients to make SNAPW lifestyle changes. The relative effectiveness of non medical members of the primary health care medical team compared with doctors in improving health literacy was confounded by the intensity of the intervention (in terms of hours or number of contacts). Thus effectiveness in terms of improvement of health literacy may be related to capacity of the provider (time as well as skills and attitudes) to undertake more than brief interventions. For some SNAPW lifestyle changes, such as smoking cessation interventions, low intensity interventions resulted in behaviour change but not necessarily improvements in health literacy.

Shared decision making and good communication are important to developing a sense of trust and partnerships to develop health literacy [12, 16] and the more intensive interventions may provide a platform for this to occur. The results from the driver and barrier extraction highlight the importance of continuity of care, the provider patient relationship and opportunity for follow up. This would support the suggestion that developing health literacy around SNAPW takes time and therefore a medium to high intensity intervention is required. Many of the barriers to shared decision making in practice, such as time, are more acute for doctors than for other health professionals [84].

Because of the nature of general practice, interventions involving doctors tend to be brief interventions and focus on issues such as smoking and physical activity prescription [19, 32‐35]. In Australia, initiatives such as Lifescripts (evidence based interventions to support lifestyle risk factor assessment and management), and 45+ health checks (health assessments targeting people aged 45+ at risk of developing a chronic disease) aim to support these brief interventions in primary care. Referral to programs for dietary education would provide patients at risk with a more intensive intervention but in Australia GPs only refer around 10% of their at risk patients to such programs [18, 85] and GPs do not have capacity to provide more intensive interventions themselves [86]. A recent randomised controlled trial of lifestyle risk factor management in Australian general practice found that brief advice for physical activity resulted in an increase in patient self-reported activity but only those patients referred to the group programs demonstrated an improvement in diet and weight (Harris et al, MJA in press).

Creating a time where issues such as health literacy or lifestyle risk factor management can be addressed without the pressure to treat an acute problem is important and may offer an explanation as to why these more intensive approaches might be effective. Health screening programs delivered in primary health care could provide an opportunity for other members of the health care team such as practice nurses to be involved in the assessment, brief intervention, referral and group programs located at the practice. Allied health professionals such as dieticians, educators or physiotherapists could also be involved in providing education and health coaching. In addition to this there needs to be a shift in patient attitudes to using primary health care services for prevention of chronic illness. Research has shown that low health literacy is associated with poorer uptake of screening for colorectal cancer [87], breast cancer and prevention measures such as flu vaccination [88].

At a policy level there needs to be greater understanding of the skills and intensity of interventions required to improve health literacy and for SNAPW risk factor modification. For example, brief interventions can be very effective for smoking cessation [82] and this can be provided by a GP or practice nurse [51]. For more complex interventions such as dietary advice and weight loss then well trained health professionals who are able to deliver interventions of the appropriate intensity are required. Many of these interventions were group based programs which also provided peer support to the participants [36, 37, 39‐42, 44‐49, 52]. Educating health professionals about the impact of health literacy on a range of behaviours is important if they are to be better able to support their patients to manage their health [20, 22, 89]. Many of the current tools to measure health literacy may be impractical for use as a screening tool in general practice but are useful as broad guidelines to help health professionals understand the impact of low health literacy on their patient’s health status [8]. Internationally, a number of governments have policy in place to address health inequities that result from poor health literacy [9, 90, 91].

The main limitation of this review was that whilst there were 52 studies included, once the principal health professionals providing the intervention were identified the numbers of papers in each group were small and most of the included studies were of moderate quality, only 11/52 were assessed as being of high quality. The heterogeneity of the interventions identified also meant that a meta-analysis was not appropriate. In addition, the details of providers and description of their characteristics and role in the intervention was not systematically reported by the studies included in the review. Another limitation was the way in which health literacy is and is not measured in studies of lifestyle risk factor modification. In order to capture the complex definition of health literacy proposed by Nutbeam [5, 12] then the measures need to go beyond simple measures of functional health literacy. The measures used in many of these studies included self-efficacy and patient activation in order to include those that addressed critical health literacy.

The results of this review highlight the importance of the provider being able to provide moderate to high intensive interventions to address health literacy to make SNAPW lifestyle risk factor changes. As the context of the primary health care setting makes it difficult for GPs to provide the intensity of intervention required to influence health literacy and behaviour change it is important the referral mechanisms to intensive programs or other health professionals are available.