Primary health care-level interventions targeting health literacy and their effect on weight loss: a systematic review

Background Enhancing individual’s health literacy for weight loss is important in addressing the increasing burden of chronic disease due to overweight and obesity. We conducted a systematic review and narrative synthesis to determine the effectiveness of lifestyle interventions aimed at improving adults’ knowledge and skills for weight loss in primary health care. The literature search included English-language papers published between 1990 and 30 June 2013 reporting research conducted within Organisation for Economic Cooperation and Development member countries. Twelve electronic databases and five journals were searched and this was supplemented by hand searching. The study population included adults (≥18 years old) with a body mass index (BMI) ≥25 kg/m2 and without chronic disease at baseline. We included intervention studies with a minimum 6 month follow-up. Three reviewers independently extracted data and two reviewers independently assessed study quality by using predefined criteria. The main outcome was a change in measured weight and/or BMI over 6 or 12 months. Results Thirteen intervention studies, all targeting diet, physical activity and behaviour change to improve individuals’ knowledge and/or skills for weight loss, were included with 2,089 participants. Most (9/13) of these studies were of a ‘weak’ quality. Seven studies provided training to the intervention deliverers. The majority of the studies (11/13) showed significant reduction in weight and/or BMI in at least one follow-up visit. There were no consistent associations in outcomes related to the mode of intervention delivery, the number or type of providers involved or the intensity of the intervention. Conclusions There was evidence for the effectiveness of interventions that focussed on improving knowledge and skills (health literacy) for weight loss. However, there was insufficient evidence to determine relative effectiveness of individual interventions. The lack of studies measuring socio-economic status needs to be addressed in future research as the rates of obesity are high in disadvantaged population groups.


Background
Globally, rates of overweight (defined as a body mass index [BMI] ≥ 25 kg/m 2 ) and obesity (BMI ≥ 30 kg/m 2 ) have been escalating in association with an increasingly sedentary lifestyle and especially with an increased energy intake [1,2]. Once considered a problem only in high-income countries, obesity is on the rise in low and middle income countries [3]. The high rates have contributed to the increased prevalence of chronic diseases such as cardiovascular disease (CVD), diabetes and cancers [2].
Health literacy is the degree to which individuals have the capacity (knowledge and skills) to obtain, process and understand basic health information and services needed to make appropriate health decisions [4]. Compared with people with adequate health literacy, those with inadequate health literacy have poorer understanding of their chronic diseases [5,6], physicians' instructions [7] and healthrelated internet usage [8]. Low health literacy has been associated with more hospitalisations, greater use of emergency care, poorer adherence to medications, and is also more common among elderly persons with poorer overall health status and higher mortality rates [9]. Low health literacy has been associated with increased risk of CVD and diabetes in the Australian population [10]. Our own primary research has demonstrated an association between low health literacy and obesity among adults [11] although this has also been previously described in children [12].
Weight management comprises the primary prevention of excess weight gain, regain or loss and optimising health and reducing risk of disease (whether or not weight loss is achieved) [13]. People need strategies involving continuing lifestyle change, regular monitoring and support from primary health care (PHC) professionals who are considered to be the first line of intervention providers for weight management [14]. Multicomponent interventions targeting the three key lifestyle areas related to obesitynutrition, physical activity (PA) and psychological approaches to behavioural changeare more likely to be effective in addressing overweight and obesity than single component interventions [14]. Patients need both knowledge and skills to engage in the lifestyle change involved in these complex interventions.
Von Wagner and colleagues' review of health literacy [15] introduced a framework on the associations between health literacy and health outcomes. These are mediated by three principal domains of health actions (proposed by Paasche-Orlow and Wolf [16]), namely access and use of health care, patient-provider interactions, and self-care (management of health and illness). While acknowledging the importance of each of the three health action domains [15] in weight loss, our review focused on 'management of health and illness'. Thus the main objective of this review was to evaluate the effectiveness of lifestyle interventions which aim to achieve weight loss by enhancing individual's knowledge and/or skills for weight loss.

Study design
A systematic review with narrative synthesis.

Inclusion criteria
Studies were eligible for inclusion if they met the criteria below ( Table 1).

Types of studies
Intervention studies (experimental or quasi-experimental trials with or without a control group) with a minimum six-month follow-up published in English between 1990 and end of June 2013 within the Organisation for Economic Cooperation and Development (OECD) member countries.

Setting
The intervention needed to be delivered in PHC, or by PHC professionals outside PHC, or to individuals who were referred to the study by PHC professionals. The definition of PHC used was: Socially appropriate, universally accessible, scientifically sound first level care provided by health services and systems with a suitably trained workforce comprised of multi-disciplinary teams supported by integrated referral systems in a way that: gives priority to those most in need and addresses health inequalities; maximises community and individual self-reliance, participation and control; and involves collaboration and partnership with other sectors to promote public health. [17] Types of participants Men and women (≥18 years) with a BMI ≥ 25 kg/m 2 at baseline and without chronic disease who were in 'treatment' for weight reduction.

Types of interventions
Interventions aiming to achieve weight reduction through changing diet and/or PA with or without psychological approaches to behaviour change by improving the participants' knowledge and/or skills for weight loss.

Outcomes
Measured change in body weight or BMI (in kg/m 2 ) between baseline and follow-up, at least once, at six months or beyond post intervention. For controlled studies these outcomes are compared between intervention and comparison groups. The outcomes were classified as statistically significant if the weight or BMI reduction reported was p < 0.05.

Excluded studies
Studies were excluded if they included pregnant women or individuals diagnosed with CVD, diabetes, cancer or other chronic conditions or where pharmaceutical or surgical interventions were employed for weight loss.

Search strategy
The electronic search covered the period from 1990 to June 2013. . The selected studies were also used for identifying earlier and more recent publications. Several relevant websites of key government, international bodies and non-government organisations were searched for grey literature. Experts in this area of research were also contacted for any relevant literature.

Identification of relevant studies
CJ, JL and NF independently carried out initial screening of the retrieved titles and abstracts (where available) against agreed a priori criteria summarised in Table 1 (Step 1). A 10% random sample of excluded studies was reviewed by CS and JL (Step 2). Full-text copies of potentially eligible papers were obtained and independently assessed (Step 3) by seven reviewers (CJ, CS, JL, JT, NF, MH and SD). Data were independently extracted by three reviewers (CJ, MH and NF) into a summary table (Step 4). Any disagreements in specific study inclusion and/or data extraction were resolved through consensus by discussion.

Quality assessment of studies
Quality assessment for each included study was carried out by CJ and SD using a standard checklist [18] and checked by NF. An overall methodological rating of strong, moderate or weak was achieved in six sections: 1) selection bias, 2) study design, 3) confounders, 4) blinding, 5) data collection methods, and 6) withdrawals and dropouts [18].

Data extraction
Data were systematically extracted on the following domains: Study characteristics: year of publication, design, recruitment method, location (country and setting), number of participants, study duration and length of follow-up (points of follow-up measurement), proportion of subjects lost to follow-up and appropriate control or 'usual care' group (where applicable).
Intervention intensity: Interventions were categorised into low, medium, high or very high intensity depending on the duration of contact between the provider and the participant or the number of points of contact between the two during the intervention period: low intensity: ≤4 hours of contact or 6 points of contact between the provider(s) and the participants; medium intensity: >4 hours and <8 hours of contact or 10 points of contact between the provider(s) and the participants; high intensity: ≥8 hours and <12 hours of contact or 12 points of contact between the provider(s) and the participants; and very high intensity: ≥12 hours of contact or 14 points of contact between the provider(s) and the participants.
Intervention characteristics: professional background of individuals delivering the intervention, mode of administration, component, dose of delivery (frequency and duration), and focus (knowledge, skills, behavioural change).

Analysis
Change in measured weight and/or BMI was compared over 6 and/or 12 months. A meta-analysis could not be performed due to small number of studies identified and heterogeneity amongst the studies. A narrative synthesis approach was used.

Trial identification
The process of identifying and selecting papers for inclusion in this review is illustrated in Figure 1. Titles and abstracts (where available) of 2,286 papers were screened and 255 papers identified for full-text assessment. Of these, 18 papers were eligible for data extraction. From these, a further 179 papers were identified for a full text review. After excluding studies which did not meet the inclusion criteria, 13 papers were included in this review [19][20][21][22][23][24][25][26][27][28][29][30][31] (Figure 1).
In almost all studies, patients were recruited from a PHC service, which was also the intervention setting. The exception was one study for which participants were recruited from community and the intervention was delivered in PHC [22].
The total number of participants across all studies was 2,089 (mean n = 161). Retention rates varied from 45 to 100% with eight studies retaining >80% of participants at the final follow-up.
From the methodological quality assessment, most of the studies were scored to be 'weak' in quality with only four studies scoring as 'strong' and none as 'moderate' ( Table 2).

Intervention characteristics
The lifestyle interventions varied in the number of contacts with participants, mode of delivery, intervention providers, behaviour-change techniques (Table 3) and the duration of final follow-up. The modal period of intervention delivery was 6 months (n = 4), ranging from  A number of modes of intervention delivery were used: All the studies involved face-to-face contact, along with telephone and mail in some cases, except one that utilised the Internet and delivered the intervention through a website [28]. Of the 12 studies using face-to-face education, six used group sessions only [19,21,22,24,26,27], three used a combination of one-to-one and group sessions [20,23,29] and three individual contacts only [25,30,31].
The studies utilised a range of providers. There were six studies with one or one main deliverer of the intervention [19,24,25,28,30,31]. Of these, three studies used providers other than PHC professionals to promote weight reduction. One study evaluated an intervention delivered in the PHC setting by non-National Health Service staff (health promotion counsellors) [25], another used an Internet-based weight-control package in a community setting [28] and the third study, [30] evaluated the effectiveness of medical assistants as weight loss counsellors. Five studies had 2-3 intervention providers delivering the intervention [21][22][23]27,29] and in two studies, even though there were multiple intervention providers, only one led the group sessions [20,26].
Seven of the 13 studies provided training to health professionals and educators to deliver the intervention, however, only five provided details of the training [19,24,25,30,31], the duration of which ranged from three hours to nine sessions of six hours each. Two papers provided no details of the training [23,27].
Of the seven studies that included patients at risk of developing type 2 diabetes, three provided patients with information on diabetes prevention [21,26,31]. The one study where the participants had metabolic syndrome used a general recommendation-based program of lifestyle intervention carried out by trained professionals [23].

Types of interventions
All the reviewed studies included interventions that, in combination, focussed on diet, PA and psychological approaches to health behaviour change. All studies explicitly stated that they targeted participants' dietary knowledge except one where the emphasis was on meal replacement [29]. In six studies participants were provided educational resources/tools [19,23,24,26,28,30] and in one study participants attended dietitian-supervised cooking classes [22]. Participants were encouraged to keep food records [29], given diaries [19,26] and PA logbooks [26], provided with analysed nutritional data and brief comments on food diaries [21] or given the opportunity to review their completed food diaries and PA records with education deliverers [30].
In eight studies PA education [20,[23][24][25]27,28,30,31] was given. One study also provided individualised advice on exercise [23]. In five studies participants could attend PA session(s) [19,21,22,24,29].   A lifestyle change personal goal setting approach was used by a number of studies to motivate participants to progress from intention to actual behaviour change [19,[27][28][29]31]. Motivational interviewing [25,31], group motivational discussions [21], telephone motivation sessions along with motivation letters [24], and websitegenerated motivational statements [28] were used to modify participants' behaviour and achieve weight loss. Participants were encouraged to self-monitor behaviour [19] and also develop cognitive and behavioural skills for managing diet and PA [25] or to self-manage weight [28]. In one study, [27] regular self-assessment was used to empower participants to take responsibility for own decisions and make informed choices. Other behaviour techniques included planning [25,29] and problem solving and environmental control [29].
The social support enhanced by the group setting in 9/13 studies mentioned earlier was further emphasised in two studies by encouraging participants to seek support from their own social networks [24,27].
Of the 11 studies with positive results, three were rated as 'strong' (all RCTs) [21,23,25] and the other eight were rated as 'weak' quality. In one of these studies (Greaves et al.) there was an increased proportion of patients achieving weight targets in the intervention but not control group [25]. However there was no significant change in mean weight in the intervention group and this group had lower initial weight compared to the control group (Table 4). Of the two studies that did not show significant reduction in either weight or BMI, one was rated as 'weak' [28] and the other as 'strong' quality [31].
There were no consistent differences in the effectiveness of interventions by their mode of delivery (except for the one Internet only intervention which was not effective), provider, behavioural intervention or intensity (Tables 4-6).

Discussion
Our paper has reviewed publications reporting lifestyle interventions at the primary health care-level that aimed to increase adults' knowledge and skills for weight loss.
We identified only 13 studies fulfilling the inclusion criteria, the majority of which demonstrated a positive impact on weight loss. Only one included study reported outcome data for both 6 and 12 month periods [30]. Consistent with earlier systematic review evidence [32,33], this study showed weight regain after an initial weight loss which was, though, in keeping with the study hypothesis.
The studies included in our review were heterogeneous, the only commonality being the intervention focus, which in every case was to change diet and PA behaviours in conjunction with behaviour change. Included amongst the modes of delivery were individual and/or group sessions or the Internet with the number of sessions ranging from five to 104 conducted over 3 to 12 months. Educators were drawn from different professional backgrounds and included both health and nonhealth professionals. Only seven studies mentioned training the educators in the intervention delivery, suggesting a need for evaluated programs to more explicitly describe the training provided to the intervention deliverers.
This review reinforces earlier systematic reviews' findings which support the efficacy of combining both dietary and PA interventions together with behaviour modification [34,35]. However, these earlier reviews did not necessarily include only those lifestyle interventions which specifically aimed to impact individual's knowledge and skills required for weight loss. Despite the diversity in the types of interventions included in this review, the results provide evidence to support the role of lifestyle interventions aiming to change individual's knowledge and/or skills in weight loss. Eleven out of 13 studies reported positive intervention effects. In the small number of studies identified, no one type or component of lifestyle intervention emerged as the most effective model and we were thus unable to determine what constituted the successful aspect of the intervention. By the same token, our review could not ascertain why interventions in two studies [28,31] failed to accomplish their objectives.
Reviews on the relationship between health literacy level and health outcomes have shown a consistent association between low health literacy and poorer health-related knowledge and comprehension [9,36]. None of the studies included in this review reported on participants' health literacy. There is thus a need to address this gap in research and to develop weight loss interventions which specifically target people's health literacy. Similarly, none of the included studies had specifically targeted or measured outcomes in disadvantaged socio-economic population. This is despite obesity being particularly prevalent among those in the most disadvantaged socio-economic groups [14] and disadvantaged populations struggling most with limited health literacy [37]. Future research is needed to evaluate interventions with this oft-neglected population group. Other approaches which develop and draw upon the   health literacy located in wider community networks may need to be considered in future research with socially disadvantaged groups [38].

Strengths and limitations
The included studies were based in a large number of countries with different health systems, obesity issues and population characteristics including Australia, the UK, the USA and European countries. The review is limited, however, by the small number of studies that met our selection criteria and were of high quality. With this limitation, it was not only difficult to confidently identify the effectiveness of the weight loss interventions, it was also not possible to identify individual intervention components associated with success. Our search was not limited to studies that had tested participants' health literacy for weight loss at baseline so we cannot ascertain if participants had low health literacy at the start. In addition, the included studies did not explicitly measure improvements in health literacy for weight loss. Though the vast majority of studies did lead to weight loss, we are unable to report specifically on health literacy improvements.
Most of the studies did not state the participants' ethnicity, education and socio-economic status. This limits our capacity to understand the generalisability of our findings to people with different socio-economic status and ethnic backgrounds. Obesity often coexists with other chronic conditions; however, our review excluded patients with existing chronic diseases thus limiting the scope of the review findings to people who are in otherwise good health.
The heterogeneity of the studies precluded us from performing a meta-analysis. Furthermore, the review was not limited to RCTs. On the one hand, this allowed us to examine the efficacy of interventions in less controlled situations, while on the other hand studies displaying biases were also included. Most of the studies did not report on the reasons for participant dropout and more than half of the studies did not conduct the analysis on an intention-to-treat basis.
We included studies where the minimum follow-up period was six months and used outcome data only for 6 and 12 months. However, weight loss achieved in the first 6 months after intervention is often regained in the subsequent months as demonstrated by one of the reviewed studies [30]. Thus for studies where the final follow-up was at six months, we cannot exclude the possibility of weight regain over longer follow-up periods.

Conclusions
Improving health literacy for weight loss in a PHC setting is a complex task and difficult to achieve. Health literacy in weight loss requires not only an understanding of what is required to lose weight but also an insight into the factors that prevent individuals from weight loss and promote weight regain. As our adopted framework demonstrates, these are important prerequisites for building motivation for change and the ability to achieve health goals. Some promising results were found for complex multicomponent lifestyle interventions conducted at the PHC-level that focussed on weight loss. However, there was insufficient evidence to discriminate between the effectiveness of intervention components. Interventions of at least medium intensity delivered by a range of health professionals, which addressed both diet and PA and using behavioural strategies, were effective. More research is needed to explore the pathway between the intervention components, health literacy, behaviour, weight loss and, in the longer term, maintenance of these losses. The lack of studies measuring socioeconomic status needs to be addressed in future.