Interventions to enhance adherence to dietary advice for preventing and managing chronic diseases in adults

Sophie Desroches; Annie Lapointe; Stéphane Ratté; Karine Gravel; France Légaré; Stéphane Turcotte

doi:10.1002/14651858.CD008722.pub2

Intervenciones para mejorar el cumplimiento del asesoramiento dietético para la prevención y el control de las enfermedades crónicas en adultos

Declaraciones de intereses de los autores

Versión publicada: 28 febrero 2013 Historial de versiones

https://doi.org/10.1002/14651858.CD008722.pub2

Contraer todo Desplegar todo

Resumen

disponible en

Antecedentes

Se ha reconocido que el cumplimiento deficiente puede ser un riesgo grave para la salud y el bienestar de los pacientes, y el mayor cumplimiento del asesoramiento dietético es un componente crítico para la prevención y el control de las enfermedades crónicas.

Objetivos

Evaluar los efectos de las intervenciones para mejorar el cumplimiento del asesoramiento dietético para la prevención y el control de las enfermedades crónicas en adultos.

Métodos de búsqueda

Se hicieron búsquedas en las siguientes bases de datos electrónicas hasta el 29 de septiembre de 2010: The Cochrane Library (número 9, 2010), PubMed, EMBASE (Embase.com), CINAHL (Ebsco) y PsycINFO (PsycNET) sin ninguna restricción de idioma. También se revisaron: a) las conferencias relevantes, los simposios y las actas de coloquios y los resúmenes de años recientes; b) los registros en la Web de los ensayos clínicos; y c) las bibliografías de estudios incluidos.

Criterios de selección

Se incluyeron ensayos controlados aleatorios que evaluaron las intervenciones para mejorar el cumplimiento del asesoramiento dietético para la prevención y el control de las enfermedades crónicas en adultos. Los estudios fueron elegibles cuando el resultado primario era el cumplimiento del asesoramiento dietético por parte del paciente. Se define "paciente" como un adulto que participa en un estudio de la prevención de enfermedades crónicas o del tratamiento de enfermedades crónicas que incluye el asesoramiento dietético.

Obtención y análisis de los datos

Dos revisores evaluaron de forma independiente la elegibilidad de los estudios. También evaluaron el riesgo de sesgo y extrajeron los datos mediante una versión modificada de la plantilla de extracción de datos del Grupo de Revisión Cochrane de Consumidores y Comunicación (Cochrane Consumers and Communication Review Group). Las discrepancias de criterio se resolvieron mediante discusión y consenso, o con un tercer revisor. Debido a que los estudios difirieron ampliamente en lo que se refiere a las intervenciones, las medidas de cumplimiento de la dieta, el asesoramiento dietético, la naturaleza de las enfermedades crónicas y la duración de las intervenciones y el seguimiento, se realizó un análisis cualitativo. Los estudios incluidos se clasificaron de acuerdo a la función de la intervención y los resultados se presentan en una tabla narrativa mediante el recuento de votos para cada categoría de intervención.

Resultados principales

Se incluyeron 38 estudios con 9445 participantes. Entre los estudios que midieron los resultados del cumplimiento de la dieta entre un grupo de intervención y un grupo de control/atención habitual, 32 de cada 123 resultados del cumplimiento de la dieta favorecieron al grupo de intervención, cuatro favorecieron al grupo de control mientras que 62 no tuvieron ninguna diferencia significativa entre los grupos (la evaluación fue imposible para 25 resultados del cumplimiento de la dieta debido a que no se proporcionaron los datos ni los análisis estadísticos necesarios para la comparación entre los grupos). Las intervenciones que mostraron una mejoría en al menos un resultado del cumplimiento de la dieta son las que incluyeron: seguimiento telefónico, video, contrato, información, herramientas nutricionales y las intervenciones más complejas incluidas las intervenciones múltiples. Sin embargo, estas intervenciones tampoco mostraron diferencias en algunos resultados del cumplimiento de la dieta en comparación con un grupo de control/atención habitual lo cual proporciona resultados no concluyentes acerca de la intervención más efectiva para mejorar el asesoramiento dietético. La mayoría de los estudios que informaron un resultado del cumplimiento de la dieta que favorece al grupo de intervención en comparación con el grupo de control/atención habitual en el corto plazo tampoco informaron efectos significativos en puntos temporales posteriores. Los estudios que investigaron intervenciones como sesiones grupales, sesiones individuales, recordatorios, restricción y técnicas de cambio del comportamiento no informaron ningún resultado del cumplimiento de la dieta que mostrara una diferencia estadísticamente significativa a favor del grupo de intervención. Finalmente, los estudios generalmente fueron de corta duración y de baja calidad, y las medidas del cumplimiento variaron ampliamente.

Conclusiones de los autores

Se necesitan estudios adicionales a largo plazo y de buena calidad que utilicen medidas más estandarizadas y validadas del cumplimiento para identificar las intervenciones que deben usarse en la práctica para mejorar el cumplimiento del asesoramiento dietético en el contexto de diversas enfermedades crónicas.

Resumen en términos sencillos

disponible en

Intervenciones para mejorar el cumplimiento del asesoramiento dietético para la prevención y el control de las enfermedades crónicas en adultos

Las enfermedades crónicas son la principal causa de mortalidad a nivel mundial. Aunque la adopción de una dieta saludable se reconoce como un componente importante para la prevención y el tratamiento, muchos individuos en riesgo o que presentan enfermedades crónicas no cumplen el asesoramiento dietético recomendado. Los métodos utilizados para facilitar los cambios en los hábitos alimentarios mediante el asesoramiento dietético (definido en esta revisión como “intervenciones”) podría mejorar el cumplimiento del asesoramiento dietético por parte de los pacientes. Por lo tanto, se examinaron los ensayos de las intervenciones que procuran mejorar el cumplimiento del asesoramiento dietético para la prevención y el control de las enfermedades crónicas en adultos.

Se identificaron 38 estudios con 9445 participantes que examinaban varios tipos de intervenciones para mejorar el cumplimiento del asesoramiento dietético con objeto de prevenir y controlar muchas enfermedades crónicas. Las enfermedades crónicas principales incluidas fueron las enfermedades cardiovasculares, la diabetes, la hipertensión y las nefropatías. Las intervenciones que mostraron una mejoría en al menos un resultado del cumplimiento de la dieta son las que incluyeron: seguimiento telefónico, video, contrato, información, herramientas nutricionales y las intervenciones más complejas incluidas las intervenciones múltiples. Sin embargo, estas intervenciones tampoco mostraron diferencias en algunos resultados del cumplimiento de la dieta en comparación con un grupo de control/atención habitual lo cual da lugar a resultados no concluyentes acerca de la intervención más efectiva para mejorar el asesoramiento dietético. Es interesante destacar que todos los estudios que incluyeron a pacientes con nefropatías informaron al menos un resultado del cumplimiento de la dieta que mostró una diferencia estadísticamente significativa a favor del grupo de intervención, cualquiera sea la intervención administrada. La mayoría de los estudios que informaron un resultado del cumplimiento de la dieta que favorece al grupo de intervención en comparación con el grupo de control/atención habitual en el corto plazo tampoco informaron efectos significativos en puntos temporales posteriores. Los estudios que investigaron intervenciones como sesiones grupales, sesiones individuales, recordatorios, restricción y técnicas de cambio del comportamiento no informaron ningún resultado del cumplimiento de la dieta que mostrara una diferencia estadísticamente significativa a favor del grupo de intervención. Finalmente, las intervenciones generalmente fueron de corta duración, los estudios utilizaron diferentes métodos para medir el cumplimiento y la calidad de los estudios generalmente fue baja.

Authors' conclusions

Implications for practice

Non‐adherence to dietary advice represents one of the barriers to getting nutrition knowledge into practice, thereby potentially hampering the prevention of the onset or progression of many chronic diseases and ultimately, improved population well‐being and health. This Cochrane review aimed to summarize, categorize and compare the effects of interventions for enhancing adherence to dietary advice for preventing and managing chronic diseases in adults. Some interventions such as telephone follow‐up, video, contract, feedback and nutritional tools demonstrated a mixed effect on diet adherence as they showed some diet adherence outcomes favouring the intervention group compared to the control/usual care group but also no difference in some diet adherence outcomes between groups. Moreover, included studies differed widely according to interventions provided, measures of diet adherence, dietary advice, nature of the chronic diseases and duration of interventions and follow‐up, making assessment of intervention versus intervention rather challenging. Therefore, this systematic review cannot draw firm conclusions from comparisons between interventions, but rather identifies a number of potentially‐beneficial interventions that can be used in practice (telephone follow‐up, video, contract, feedback and nutritional tools). Also, while the majority of multiple interventions have demonstrated a positive effect on diet adherence compared to a control/usual care group, none of the included studies assessed the same combination of interventions, making impossible the identification of the optimal combination of interventions to enhance adherence to dietary advice. Consequently, researchers, decision makers, health professionals and consumers remain with little practical guidance with regard to the best intervention for enhancing adherence to dietary advice. However, it may be argued that in health care, there is often no unique best option for either treatment or process of care, as these options may be influenced by clients’ preferences and values. Although longer‐term, well‐designed RCTs using improved methods for measuring diet adherence are needed, results of this systematic review provide options for both health professionals and consumers that may be used in practice. Interventions shown to be beneficial compared to a control/usual care group could be used depending on clients’ preferences, lifestyle and values, health professionals’ communication skills, and organisational context.

Implications for research

Evidence of the role of a healthy diet and/or specific nutrient intakes on the prevention and management of chronic diseases is well recognized. Further studies are now essential to refine methods for providing dietary advice and improve diet adherence in the context of chronic diseases. Several gaps in knowledge have been identified in this review regarding the effectiveness of interventions to enhance adherence to dietary advice for preventing and managing chronic diseases in adults:

Further good quality studies should be designed to minimize bias and to have an adequate sample size to detect significant differences between groups;
Further studies with a long‐term duration, namely more than 12 months, and a follow‐up evaluation are needed;
Further research should be designed with a comparison between an intervention group and a control/usual care group both providing the same dietary advice to capture the effect of the intervention only, without confounding factors;
Further studies need to define clearly the term 'adherence' and describe the intervention in detail. Moreover, there is a need to develop standardized and validated self‐report tools and robust objective measures (e.g. biomarkers) to assess adherence to dietary advice;
Further studies should investigate the factors contributing to clients’ non‐adherence to dietary advice in order to develop interventions to overcome barriers. These factors include psychosocial and environmental determinants, but also biological factors affecting food intake;
Moreover, perspectives from health professionals and clients about the interventions enhancing adherence to dietary advice should be studied with the aim of identifying those that are most implementable in practice and adaptable to local contexts (Desroches 2011).

Background

Description of the condition

Chronic diseases are defined as diseases of long duration that have generally a slow progression (WHO 2008). The most common chronic diseases include diabetes, cardiovascular diseases (CVD), cancers, asthma, chronic obstructive pulmonary diseases (COPD), arthritis, obesity and renal failure. Considering that chronic diseases are the leading cause of death and disability and account for 60% of all deaths worldwide (WHO 2008), the Department of Chronic Disease and Health Promotion of the World Health Organization (WHO) emphasizes the importance of preventing and managing chronic diseases and their risk factors (WHO 2010). Some health conditions have been found to be risk factors, for example, patients with the metabolic syndrome have an increased risk of developing CVD (Mottillo 2010). Similarly, women with a previous history of gestational diabetes have an increased risk of developing type II diabetes (Bellamy 2009). These risk factors may be targeted in interventions aiming to prevent chronic diseases.

Evidence from epidemiologic, experimental and clinical studies has demonstrated a strong relationship between dietary patterns or nutrient intakes, and prevention and management of chronic diseases including diabetes (Champagne 2009), CVD (Lavie 2009), and obesity (Kennedy 2004). Several authoritative health agencies have recommended the adoption of a healthy diet as the cornerstone in preventing and/or managing chronic diseases such as CVD (Lichtenstein 2006), diabetes (Bantle 2008) and cancer (Kushi 2006). For example, lifestyle interventions including dietary changes were shown to reduce the incidence of diabetes by 58% compared to a control group in individuals at high risk in two large randomized controlled trials (RCTs): the Finnish Diabetes Prevention study (Lindstrom 2003) and the Diabetes Prevention Program (Knowler 2002). In line with this, dietitians and other health professionals provide people with dietary advice designed to improve their nutritional intake (Baldwin 2011).

The concept of 'adherence' recognizes the patient’s right to choose whether or not to follow advice, and implies a patient’s active participation in the treatment regimen (Cohen 2009). For chronic disease management including medication and lifestyle changes, non‐adherence rates are estimated to be between 50% and 80% (WHO 2003). Thus, poor adherence can be a serious threat to patients’ health and wellbeing (DiMatteo 2002), and also carries an economic burden (DiMatteo 2004a). Adherence is particularly important in the context of chronic diseases requiring long‐term therapy and a number of permanent rather than temporary changes in lifestyle behaviours, such as diet, physical activity and smoking (WHO 2003). The extent to which risk‐reduction interventions proved to be as effective in research settings as in individuals' real‐life settings depends on the patient’s adherence to treatment advice. In that regard, results from an RCT assessing adherence to and effectiveness of four popular diets (Atkins, Zone, Weight Watchers, and Ornish) revealed that level of adherence to dietary advice, rather than the type of diet, was the key determinant of greater weight loss and CVD risk factor reductions (Dansinger 2005). Whether the number of intervention goals that an individual has to reach influences adherence was also addressed in a secondary analysis of the PREMIER study (Young 2009). In this RCT that tested the effects of two multicomponent lifestyle interventions on blood pressure control, the authors reported that individuals with the most physical activity and dietary behaviour goals to achieve reached the most goals (Young 2009).

Measurement of adherence to prescribed dietary advice typically involves: 1) assessment of what the client eats through self‐reported methods (e.g. 24‐hour recall, food records, food frequency questionnaires, diet history); and 2) determination of the degree to which the diet approximates the recommended dietary plan (e.g. difference between clients’ recommended macronutrient goals and their self‐reported intake). Although sparsely used, more objective measures of adherence to diets also exist (e.g. 24‐hour urinary sodium excretion to assess adherence to a low sodium diet (Chung 2008)). However, there is no gold standard for the accurate determination of dietary intake. Self‐report of energy intake is a characteristic inherent to nutrition‐related topics and is found to be underestimated compared to objective measures such as resting energy expenditure assessed by indirect calorimetry (Asbeck 2002). Underreporting energy intake has been observed more frequently in women versus men, (Johnson 1994), in older versus young (Huang 2005), and in obese versus normal weight individuals (Briefel 1997). Although self‐report measures are often regarded as susceptible to bias (e.g. over reliance on memory; report error related to meal composition or portion sizes; daily dietary variability; social desirability) (Kumanyika 2000; Wilson 2005) they are a direct, simple and inexpensive method (DiMatteo 2004b), and are readily available for use in practice. Self‐report measures can be improved and validated by using multiple measures of adherence and controlling statistically for bias or by using constructs such as body weight, blood pressure or plasma cholesterol concentrations (Hebert 2001; DiMatteo 2004b).

Description of the intervention

Adherence to dietary advice has been shown to vary according to gender (Chung 2006), socio‐economic status (Reid 1984) and ethnicity (Natarajan 2009). Moreover, numerous barriers to client adherence in health care have been identified. Among them are complexity of treatment plan, and clients’ knowledge of disease and understanding of the importance of treatment in its control and in preventing adverse outcomes (Makaryus 2005 ;Harmon 2006; Robinson 2008). According to a WHO report, "interventions for removing barriers to adherence must become a central component of efforts to improve population health worldwide" (WHO 2003). Although non‐adherence is often attributed to clients who are viewed as "non cooperative", "non compliant" and "unable to follow instructions" (Kapur 2008), it is increasingly recognized that health professionals may help their clients overcome barriers to adherence (Harmon 2006) by improving how they approach their clients' problems, how they provide advice, and how they involve their clients in treatment decision making. Although there is a wide diversity of interventions for enhancing adherence to dietary advice, their underlying aim is to prompt change to facilitate the adoption of recommended dietary behaviours.

How the intervention might work

Behaviour change theories have proved useful for explaining health‐related behaviours, including dietary behaviours. They attempt to identify the determinants that will contribute to predict the adoption of a specific behaviour, and which should be taken into account when developing a behaviour change intervention, such as a method for providing dietary advice. Several models or theories to predict behaviour change can be used in health‐related interventions, such as the Health Belief Model (Rosenstock 1974), the Theory of Planned Behaviour (Ajzen 1991), the Theory of Reasoned Action (Fishbein 1981) and the Social Cognitive Theory (Bandura 1986). More recently, Michie 2011 proposed a framework, the COM‐B system, which includes three principal interrelated components of the determination of a behaviour: 1) the motivation (the direct brain process leading to a behaviour), 2) the capability (the individual’s psychological and physical capacity to engage a behaviour) and 3) the opportunity (the factors that lie outside the individual that make the behaviour possible or not) (Michie 2011).The authors also developed a system for characterizing behaviour change interventions and their components in order to facilitate the identification of the effective behaviour change interventions and the implementation of evidence‐based practice in this area. According to this system, behaviour change interventions can be classified as nine intervention functions: education, persuasion, incentivisation, coercion, training, restriction, environmental restructuring, modelling and enablement (Michie 2011). These theories or models focus on different determinants or combinations of determinants of the behaviours which could be helpful for developing interventions for enhancing adherence to dietary advice.

Why it is important to do this review

As greater adherence to dietary advice is a critical component in preventing and managing chronic diseases, research is needed to identify the characteristics of interventions that will result in a better agreement between health professionals’ evidence‐based dietary advice, and their clients’ eating patterns. Despite growing recognition that non‐adherence to dietary advice is a barrier to getting new nutrition knowledge into practice, previous knowledge syntheses have provided decision makers and knowledge users with little practical guidance on the development of useable interventions for enhancing adherence to dietary advice. Studies have reported on interventions designed to enhance adherence to dietary advice by overcoming barriers to adherence. Although some studies have reported positive effects of interventions to enhance adherence to dietary advice, no systematic review specifically assesses dietary interventions that lead to sustained dietary changes or that refer to a wide array of chronic diseases. Haynes 2008 summarized the results of RCTs of interventions to help clients adhere to prescriptions for medications for medical problems, and excluded interventions targeting dietary advice. Bosch‐Capblanch 2007 systematically reviewed the effects of contracts between clients and health professionals for improving clients' adherence to treatment, prevention and health promotion activities. Although this review is relevant to our review, it reported only the effect of contracts (as opposed to other interventions), and was not specific to dietary advice. Several non‐Cochrane reviews may overlap with our review, but these are not systematic (Brownell 1995b; Brownell 1995a; Burke 1997; Newell 2000; Fappa 2008) and/or are related to only one health condition and not specifically targeting dietary advice (Burke 1997; Newell 2000; Fappa 2008).

This review will improve the knowledge base for adherence to dietary advice; a topic of immense importance for dietetics practice that will also be relevant to clients, and other health professionals.

Objectives

To assess the effects of interventions for enhancing adherence to dietary advice for preventing and managing chronic diseases in adults.

Methods

Criteria for considering studies for this review

Types of studies

Randomized controlled trials (RCTs) including cluster RCTs. Because interventions for enhancing adherence to dietary advice aim to initiate dietary changes, a cross‐over design in which each client received all interventions could induce a carry‐over effect. Therefore, we excluded studies including a cross‐over design.

Types of participants

Clients, aged 18 years and over, in real‐life settings. We define 'client' as an adult participating in a chronic disease prevention or chronic disease management study involving dietary advice. We included clients who had a diet related‐chronic disease (e.g. obesity, cardiovascular disease, renal failure, hypertension) or at least one risk factor for a chronic disease (e.g. overweight, hyperlipidaemia). We included family or non‐family caregivers such as wife/husband or individual living with the client and involved in meal planning and preparation. We also included studies involving health professionals delivering dietary advice.

Types of interventions

We included studies assessing the effects of a single intervention or multiple interventions involving chronic disease prevention and management, on adherence to dietary advice. 'Intervention' was defined as the method used to facilitate changes in dietary habits through dietary advice. To structure the presentation of results, we grouped interventions according to the intervention functions of the behaviour change wheel developed by Michie and colleagues (Michie 2011). Therefore, we classified interventions to enhance adherence to dietary advice as:

Education (increasing knowledge or understanding);
Persuasion (using communication to induce positive or negative feelings or stimulate action);
Incentivisation (creating expectation of reward);
Coercion (creating expectation of punishment or cost);
Training (imparting skills);
Restriction (using rules to reduce the opportunity to engage in the target behaviour);
Environmental restructuring (changing the physical or social context);
Modelling (providing an example for people to aspire to or imitate) ;
Enablement (increasing means/reducing barriers to increase capability or opportunity);
Multiple (combination of two or more different interventions).

We included studies making the following comparisons:

Single intervention for enhancing adherence to dietary advice versus no intervention (control) or a reference standard of care (usual care);
Single intervention for enhancing adherence to dietary advice versus single or multiple interventions with a similar purpose (to enhance adherence to dietary advice);
Multiple interventions for enhancing adherence to dietary advice versus no intervention (control) or a reference standard of care (usual care);
Multiple interventions for enhancing adherence to dietary advice versus single or multiple interventions for enhancing adherence to dietary advice.

The term 'reference standard of care' refers to the usual dietary intervention performed to address a specific health condition. For example, in Amato 1990 two approaches were used with patients who were severely obese using the same dietary advice: 1) weight loss advice versus 2) weight loss advice combined with psychotherapy. The approach with weight loss advice was the reference standard of care while the approach with weight loss advice combined with psychotherapy was the intervention for enhancing adherence to dietary advice. Furthermore, only studies comparing interventions with the same dietary advice component (e.g. increase consumption of fruits and vegetables, decrease fat intake) but differing in terms of the method for changing dietary habits through dietary advice (e.g. education (counselling and follow‐up with health professional, educational tools)) were included. We excluded studies assessing adherence to dietary advice for which interventions were not a method for facilitating changes in dietary habits through dietary advice (e.g. medication for weight loss, exercise, etc.).

We excluded studies that aimed primarily to evaluate the effects of an experimental diet or a food plan on health outcomes, and for which adherence was monitored as a secondary outcome to justify, for example, the validity of the results, as these interventions were not designed for enhancing adherence to dietary advice. We only included studies including food‐based dietary advice and representing real‐life conditions. Therefore, we excluded studies involving the provision of meals, food items or dietary supplements (e.g. vitamin, mineral, omega‐3 fatty acid).

Types of outcome measures

Primary outcomes

Client adherence to dietary advice (e.g. biochemical measures within acceptable limits, mean dietary intake, proportion of clients achieving the dietary advice). We included studies reporting adherence to dietary advice as a primary outcome, namely those clearly mentioning a measurement of diet adherence in the title or the objective of the study and/or those reporting the proportion of patients adhering to dietary advice. We excluded studies reporting mean dietary intake without specifically assessing adherence to dietary advice.

Secondary outcomes

Process measures: e.g. attendance at or participation in individual counselling or group sessions, number of completed food records returned to research coordinators, client or family or non‐family caregivers’ satisfaction with the dietary or counselling approaches, health professionals’ skills in performing the experimental interventions or their satisfaction with the counselling approach.
Client‐based health or behaviour outcomes: e.g. blood pressure; plasma cholesterol concentration; plasma glucose concentration; body weight; relief of symptoms; smoking; physical activity; blood glucose monitoring.
Organisational outcomes: e.g. cost; time; resources required by client, family or non‐family caregivers, or healthcare professionals.
Harms or secondary effects: e.g. confusion regarding new eating patterns; feelings of lack of confidence or skills in preparing meals; unhappiness at loss of traditional meals.

Search methods for identification of studies

Electronic searches

We conducted a systematic search, using 29 September 2010 as the cut‐off date, in the following electronic databases:

The Cochrane Library, issue 9 2010 (via Wiley);
PubMed;
EMBASE (Embase.com);
CINAHL (Ebsco);
PsycINFO (PsycNet).

We present detailed search strategies in Appendix 1; Appendix 2; Appendix 3; Appendix 4; Appendix 5. There were no language restrictions and all databases were searched from their start date.

Searching other resources

We conducted additional searches for unpublished studies through grey literature:

Recent years of relevant conference, symposium and colloquium proceedings and abstracts:

- 2009‐10 Scientific sessions of the American Diabetes Association;
- 2009‐10 Scientific sessions of the American Heart Association;
- 2009‐10 Food and Nutrition Conference and Expo of the American Dietetic Association;
- 2010 Canadian Diabetes Association/Canadian Society of Endocrinology and Metabolism Professional Conference and Annual Meeting;
- 2009 International Diabetes Federation World Diabetes Congress North America;
- 2009‐10 Dietitians of Canada National Conference;
- 2009‐10 Obesity Society Annual Scientific Meeting;
- 2009‐10 Experimental Biology Meeting;
- 2009‐10 Canadian Nutrition Society;
Web‐based registries of clinical trials (US National Institutes of Health, The National Library of Medicine, Current Controlled Trials);
Bibliographies of included studies;
Contact with experts in the field to request details of any other known studies.

Data collection and analysis

Selection of studies

Two review authors independently assessed the eligibility of papers identified by the search strategy. All titles and abstracts were screened according to pre‐established inclusion criteria (see Criteria for considering studies for this review). We retrieved full text copies of papers judged to be potentially relevant to the review. Disagreements were resolved by discussion between the two review authors, and when consensus was not reached, with a third review author. We attempted to contact authors to obtain further details of papers containing insufficient information to make a decision about eligibility. If no response was provided, we sent up to two reminders and, when possible, also contacted one co‐author. We contacted 81 authors of whom 67 provided a response.

Data extraction and management

Two review authors performed the data extraction independently from all included studies using a modified version of the Cochrane Consumers and Communication Review Group data extraction template (CCCRG 2010). In addition to the standard form derived from the data extraction template of the Cochrane Consumers and Communication Review Group, other relevant information was extracted including:

Food‐based dietary advice;
Rationale underlying the dietary advice (e.g. clinical practice guidelines, other evidence‐based sources);
Adherence assessment method (proportion of clients achieving the dietary advice, biochemical measures);
Description of the intervention (eg. education, persuasion, training).

Any discrepancies in judgement were resolved by discussion and consensus, or with a third review author. Where information was missing, we contacted the corresponding author. If no answer was provided, we sent up to two reminders and, when possible, also contacted one co‐author. We contacted 38 authors of included studies, of whom 22 provided a response.

Assessment of risk of bias in included studies

Two review authors assessed and reported on the risk of bias of included RCTs in terms of the following individual elements that affect risk of bias:

Random sequence generation;
Allocation concealment;
Blinding ‐ clients, providers and outcome assessors;
Incomplete outcome data;
Selective reporting;
Other bias.

Each of the risk of bias items was assessed as 'low risk of bias', 'high risk of bias' and 'unclear risk' based on the study reports and/or additional information provided by the study authors. Any discrepancies in judgement were resolved by discussion and consensus, or with a third review author.

Measures of treatment effect

The table Characteristics of included studies includes descriptions of study design, setting, country, chronic disease, type of participants (age, sex, ethnicity), sample size, intervention(s) and/or control/usual care, measurement of diet adherence, dietary advice, drop‐out rate and providers. Sample size is presented as the number of randomized clients, or when the authors did not report it, as the number of completers. Drop‐out rate is presented as reported or as calculated when the authors did not report it.

Since the included studies addressed a wide range of interventions, measures of diet adherence, dietary advice, nature of chronic diseases, and duration of interventions and follow‐up, it was impossible to perform meta‐analyses. For this reason, we could not apply all the methods outlined in the protocol (Desroches 2010) but present these in Appendix 6 for application in future updates of the review. To facilitate the presentation of results, two authors independently classified included studies according to the function of the intervention (Michie 2011). Any discrepancies in judgement were resolved by discussion and consensus, or with a third review author. The method(s) for facilitating changes in dietary habits through dietary advice used in the intervention group and differing from the method(s) used in the comparative group (control, usual care or other intervention group) was (were) defined as the intervention and was (were) classified according to different categories of interventions (education, persuasion, incentivisation, coercion, training, restriction, environmental restructuring, modelling, enablement and multiple). Representing each category of interventions, eight additional tables (Additional tables) summarize narratively the number of studies and participants per intervention, the effect on diet adherence and the quality of evidence (GRADE) (Higgins 2011). In case of discrepancies between the results provided by the authors and the risk ratio (RR) or the standardized mean difference (SMD) calculated using Review Manager 5 ('RevMan') software (RevMan 2012), we selected the results provided by RevMan to complete the Additional tables. Some studies assessed and therefore reported multiple diet adherence outcomes (e.g. adherence to fiber intake and adherence to cholesterol‐restricted diet) and/or evaluated diet adherence outcome(s) at different times (e.g. one month, three months, six months). Consequently, we used vote counting, that is we reported the number of diet adherence outcomes favouring the intervention out of the total number of diet adherence outcomes reported, regardless of the statistical significance or size of their results (Higgins 2011), to assess studies that reported diet adherence outcomes between an intervention group and a control/usual care group. Studies are described in more than one category of intervention if they investigated more than one intervention (Baraz 2010; Cummings 1981; Hsueh 2007; Jones 1986; Kendall 1987; Logan 2010; Mahler 1999; McCulloch 1983). Only studies that compared an intervention with a control/usual care group were included in these Additional tables.

We used RevMan to create forest plots when diet adherence outcomes provided raw and complete data (means and standard deviations for continuous data, and number of events and number of total observations for dichotomous data). We analyzed dichotomous data by determining the RR and 95% confidence intervals. We analyzed continuous data by determining the SMD of the intervention and the control groups in each study with 95% confidence intervals. Only studies comparing a single or multiple intervention group with a control/usual care group were included in forest plots. We used mean differences between pre‐post intervention to calculate SMD. When these data were not known, and that baseline data were available for the two groups, we corrected the standard effect size by calculating the difference between pre‐ and post‐intervention values. The pooled estimates standard deviation was used to calculate the standard deviation of this difference. When no baseline data were reported, groups were considered to be similar before the intervention. Outcomes with data including covariate‐adjusted means or imputed means were not analysed with forest plots. For these studies, we presented the qualitative data as reported by the study authors. Some elevated SMDs could represent a high diet adherence (e.g. fruit, vegetable and fiber intakes) whereas some elevated SMDs could represent a low diet adherence (e.g. energy, fat and sodium intakes). Therefore, to correct for difference in the direction of the scale in forest plots, means of the intervention and the control groups were multiplied by ‐1 for outcomes where elevated SMD represented a high diet adherence (e.g. fruit, vegetable and fiber intakes). When authors did not report statistical analyses, we used data to calculate the SMD or the RR in RevMan in order to compare differences in outcomes between groups.

Assessment of heterogeneity

We did not explore heterogeneity due to the wide range of interventions, measures of diet adherence, dietary advice, nature of chronic diseases, and duration of interventions and follow‐up addressed in included studies.

Consumer participation

The Cochrane Consumers and Communication Review Group's editorial process for the protocol (Desroches 2010) and the review involved two anonymous consumer referees. We also sought additional feedback throughout the review process from representatives of the Dietitians of Canada to ensure that important issues for health professionals were addressed.

Results

Description of studies

Results of the search

From the searches, we identified 5183 potentially‐relevant publications after duplicates were removed. From these, we excluded 4786 publications after examining the titles and abstracts, and we retrieved 398 full‐texts of potentially‐relevant publications. From these, 42 publications (describing 38 unique studies) met our inclusion criteria and were considered as eligible. We classified a further 5 publications (describing 6 studies) as ongoing studies (see Characteristics of ongoing studies), and 20 publications as studies awaiting classification (See Characteristics of studies awaiting classification) (see Figure 1, Study Flow Diagram).

Figure 1

Study flow diagram.

Included studies

Three included studies were described in more than one publication. First, Jiang's PhD thesis was published later in an electronic journal (Jiang 2004). Similarly, Chow's PhD thesis was published later in an electronic journal (Wong 2010). Miller 1988, Miller 1989 and Miller 1990 (Miller 1988) all described the same study and reported results for diet adherence at 30 and 60 days, 1 year and 2 years, respectively. We refer to this study as Miller 1988. Therefore we included 38 studies reported in 42 publications (See Characteristics of included studies).

All included studies were RCTs. Only one of them used cluster randomisation (Wood 2008).

Location, setting and duration

Studies were conducted in the following countries:

Country	Number of studies	Studies
United States of America	14	Aldarondo 1999; Beasley 2008; Becker 1998; Cummings 1981; Gans 1994; Gill 2010; Hsueh 2007; Hyman 2007; Kendall 1987; Mahler 1999; Micco 2007; Miller 1988; Racelis 1998; Scisney‐Matlock 2006
United Kingdom	7	Bennett 1986; French 2008; Grace 1996; Jones 1986; Logan 2010; McCulloch 1983; Morey 2008
China	5	Chen 2006; Chiu 2010; Jiang 2004; Wong 2010; Zhao 2004
Canada	4	Arcand 2005; Conrad 2000; Gucciardi 2007; Ryan 2002
Brazil	1	Assuncao 2010
Iran	1	Baraz 2010
The Netherlands	1	Blanson 2009
Finland	1	Laitinen 1993
Norway	1	Meland 1994
South Africa	1	Stewart 2005
Taiwan	1	Tsay 2003
Multiple (France, Italy, Poland, Spain, Sweden, United Kingdom, Denmark and the Netherlands)	1	Wood 2008

All included studies were directed towards clients and none of them was directed towards family or non‐family caregivers or health professionals.

An outpatient setting was reported in the majority of the included studies (n = 31). Four studies were carried out in a research center setting (Beasley 2008; Blanson 2009; Hsueh 2007; Micco 2007) while one study (Gans 1994) included two settings (workplace and community). In two studies, the setting could not be identified (Aldarondo 1999; Bennett 1986).

Nineteen studies evaluated diet adherence to dietary advice over a period of less than 6 months (Aldarondo 1999; Arcand 2005; Baraz 2010; Beasley 2008; Bennett 1986; Blanson 2009; Chen 2006; Chiu 2010; Cummings 1981; Gans 1994; Gill 2010; Grace 1996; Gucciardi 2007; Jones 1986; Mahler 1999; Meland 1994; Scisney‐Matlock 2006; Wong 2010; Zhao 2004), nine studies had a duration between 6 and 12 months (Assuncao 2010; Conrad 2000; Hsueh 2007; Jiang 2004; Kendall 1987; McCulloch 1983; Ryan 2002; Stewart 2005; Tsay 2003), while only 10 studies evaluated diet adherence to dietary advice over a 12‐month period or more (Becker 1998; French 2008; Hyman 2007; Laitinen 1993; Logan 2010; Micco 2007; Miller 1988; Morey 2008; Racelis 1998; Wood 2008).

Clients

The 38 studies included in this review involved 9445 clients. The range in the number of clients in each study varied from 7 to 5405 (median = 83). Only 13 of the 38 studies provided a power calculation (Aldarondo 1999; Assuncao 2010; Beasley 2008; Chiu 2010; French 2008; Hyman 2007; Jiang 2004; Meland 1994; Stewart 2005; Tsay 2003; Wong 2010; Wood 2008; Zhao 2004) and among them, 10 studies recruited the number of clients according to their power analysis (Aldarondo 1999; Assuncao 2010; Beasley 2008; French 2008; Jiang 2004; Meland 1994; Stewart 2005; Tsay 2003; Wong 2010; Zhao 2004).

Prevention of chronic diseases

Five studies included clients receiving dietary advice for the prevention of chronic diseases, such as clients with a high risk of CVD (clients having dyslipidaemia (Gans 1994; Grace 1996), siblings of individuals with coronary heart diseases (Becker 1998)) and overweight clients (Blanson 2009; Jones 1986).

Management of chronic diseases

Twenty‐seven included studies addressed dietary advice for chronic disease management. Eight studies included clients receiving dietary advice for the management of CVD (heart failure (Arcand 2005), coronary heart disease (Logan 2010; Zhao 2004), coronary artery disease (Conrad 2000; Mahler 1999), peripheral artery disease (Racelis 1998), angina pectoris and myocardial infarction (Jiang 2004; Miller 1988)); six studies involved the management of diabetes (French 2008; Gucciardi 2007; Kendall 1987; Laitinen 1993; McCulloch 1983; Ryan 2002); five studies involved the management of hypertension (Chiu 2010; Hyman 2007; Meland 1994; Scisney‐Matlock 2006; Stewart 2005); six studies addressed the management of renal failure (Baraz 2010; Chen 2006; Cummings 1981; Morey 2008; Tsay 2003; Wong 2010); one study addressed the management of obesity (Aldarondo 1999): and one study addressed the management of irritable bowel syndrome (Hsueh 2007).

Prevention and management of chronic diseases

Six studies included clients receiving dietary advice for both the prevention and the management of chronic diseases. One study was conducted with clients with coronary heart disease and clients with a high risk of developing CVD (Wood 2008). The remaining five studies included overweight and obese clients (Assuncao 2010; Beasley 2008; Bennett 1986; Gill 2010; Micco 2007).

Interventions

Included studies assessed interventions in the following categories:

Education

Nine studies offered nutrition counselling and follow‐up with a health professional through telephone follow‐up (Chiu 2010; Cummings 1981; Racelis 1998; Stewart 2005), group sessions (Gill 2010; Jones 1986) or individual sessions with a dietitian (Jones 1986; Micco 2007) or a nurse (Hsueh 2007). Moreover, four studies used educational tools to provide dietary advice such as video (Baraz 2010; Mahler 1999; McCulloch 1983) or booklet (Kendall 1987).

Persuasion

Two studies used reminders (Gans 1994; Ryan 2002).

Incentivisation

One study used contracts with rewards (Cummings 1981).

Training

Three studies used feedback (Beasley 2008; French 2008; Meland 1994).

Restriction

Only one study compared an immediate versus an incremental reduction of fat intake (Conrad 2000).

Modelling

Seven studies used nutritional tools such as menus, exchange list and portion size examples in order to enhance diet adherence (Assuncao 2010; Chen 2006; Grace 1996; Kendall 1987; Logan 2010; McCulloch 1983; Scisney‐Matlock 2006).

Enablement

Three studies used one or more behaviour change techniques, including barrier identification/problem solving (Aldarondo 1999; Bennett 1986; Logan 2010), goal setting (Logan 2010), self‐talk (defined as use of self‐instruction and self‐encouragement to support action by Abraham and Michie (Abraham 2008)) (Aldarondo 1999; Bennett 1986) and teaching to use prompts/cues (defined as teaching the person to identify environmental cues that can be used to remind them to perform a dietary behaviour by Abraham and Michie (Abraham 2008)) (Bennett 1986).

Multiple

This category includes 18 studies using a combination of two or more different interventions (Arcand 2005; Baraz 2010; Becker 1998; Blanson 2009; Cummings 1981; Gucciardi 2007; Hsueh 2007; Hyman 2007; Jiang 2004; Jones 1986; Laitinen 1993; Mahler 1999; Miller 1988; Morey 2008; Tsay 2003; Wong 2010; Wood 2008; Zhao 2004).

Outcomes

Twenty‐eight studies compared two groups (Aldarondo 1999; Arcand 2005; Assuncao 2010; Baraz 2010; Beasley 2008; Becker 1998; Blanson 2009; Chen 2006; Chiu 2010; Conrad 2000; Gill 2010; Grace 1996; Gucciardi 2007; Hsueh 2007; Kendall 1987; Jiang 2004; Laitinen 1993; Logan 2010; Meland 1994; Micco 2007; Miller 1988; Morey 2008; Racelis 1998; Scisney‐Matlock 2006; Stewart 2005; Tsay 2003; Wong 2010; Zhao 2004), six studies compared three groups (Bennett 1986; French 2008; Hyman 2007; Mahler 1999; McCulloch 1983; Ryan 2002) and four studies compared four groups (Cummings 1981; Gans 1994; Jones 1986; Wood 2008). Twenty‐five studies assessed a single diet adherence outcome (Arcand 2005; Beasley 2008; Becker 1998; Bennett 1986; Blanson 2009; Chen 2006; Chiu 2010; Conrad 2000; Gans 1994; Gill 2010; Gucciardi 2007; Hyman 2007; Jiang 2004; Jones 1986; Logan 2010; Mahler 1999; McCulloch 1983; Meland 1994; Micco 2007; Miller 1988; Morey 2008; Racelis 1998; Scisney‐Matlock 2006; Tsay 2003; Zhao 2004) while 13 studies assessed multiple diet adherence outcomes (Aldarondo 1999; Assuncao 2010; Baraz 2010; Cummings 1981; French 2008; Grace 1996; Hsueh 2007; Kendall 1987; Laitinen 1993; Ryan 2002; Stewart 2005; Wong 2010; Wood 2008). Twenty studies assessed diet adherence outcome(s) once (Aldarondo 1999; Arcand 2005; Assuncao 2010; Baraz 2010; Beasley 2008; Becker 1998; Bennett 1986; Blanson 2009; Chen 2006; Chiu 2010; Conrad 2000; French 2008; Gans 1994; Gill 2010; Grace 1996; Gucciardi 2007; Jones 1986; McCulloch 1983; Racelis 1998; Wood 2008), 13 studies assessed diet adherence outcome (s) twice (Cummings 1981; Hsueh 2007; Hyman 2007; Jiang 2004; Kendall 1987; Laitinen 1993; Logan 2010; Mahler 1999; Meland 1994; Micco 2007; Stewart 2005; Wong 2010; Zhao 2004) while 5 studies assessed diet adherence outcome (s) 3 or more times (Miller 1988; Morey 2008; Ryan 2002; Scisney‐Matlock 2006; Tsay 2003). Consequently, 32 studies compared diet adherence outcomes between an intervention group and a control/usual care group, and 9 studies compared two intervention groups.

Excluded studies

As described in the Characteristics of excluded studies table, reasons for exclusion included: no measure of adherence outcome; not the same dietary advice component in groups; not a randomized controlled trial; provision of meals, food, items or dietary supplements; not involving clients with or at risk of chronic diseases; intervention not intended to improve diet adherence; not a real‐life setting; clients were under the age of 18; and study did not involve a nutritional intervention.

Risk of bias in included studies

As described in the Characteristics of included studies, eight risk of bias criteria were applied to each study (random sequence generation, allocation concealment, blinding: clients, providers and outcome assessors, incomplete outcome data, selective reporting and other bias). Two studies were rated as low risk on 4 of the 8 criteria (Gucciardi 2007; Zhao 2004), 8 studies were low risk on 3 criteria (Aldarondo 1999; French 2008; Jiang 2004; Meland 1994; Morey 2008; Scisney‐Matlock 2006; Stewart 2005; Tsay 2003), 11 studies were rated as low risk on 2 criteria (Arcand 2005; Assuncao 2010; Baraz 2010; Chen 2006; Cummings 1981; Kendall 1987; Laitinen 1993; Logan 2010; Mahler 1999; Ryan 2002; Wong 2010), 11 studies were rated as low risk on one criterion (Beasley 2008; Becker 1998; Bennett 1986; Blanson 2009; Chiu 2010; Conrad 2000; Gill 2010, Hsueh 2007; McCulloch 1983; Miller 1988; Racelis 1998) and six studies were not rated low risk for any criteria (Gans 1994; Grace 1996; Hyman 2007; Jones 1986; Micco 2007; Wood 2008) (see Figure 2).

Figure 2

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

The allocation sequence was adequately generated in the majority of studies (n = 26). Twelve studies did not report sufficient information to determine this risk of bias (Blanson 2009; Conrad 2000; Gans 1994; Grace 1996; Hyman 2007; Jones 1986; McCulloch 1983; Micco 2007; Miller 1988; Racelis 1998; Tsay 2003; Wood 2008).

The allocation was adequately concealed only in seven studies (Assuncao 2010; Jiang 2004; Laitinen 1993; Mahler 1999; Meland 1994; Stewart 2005; Zhao 2004) while nine studies reported an inadequate allocation (Aldarondo 1999; Baraz 2010; Beasley 2008; Bennett 1986; Cummings 1981; Gucciardi 2007; Kendall 1987; Logan 2010; Miller 1988). The majority of the studies (n = 22) did not describe the allocation concealment in sufficient detail to permit evaluation.

Blinding

The majority of the interventions provided to clients were difficult to blind for clients, providers and outcomes assessors. Therefore, only three studies (Aldarondo 1999; Blanson 2009; Scisney‐Matlock 2006), two studies (Gucciardi 2007; Tsay 2003) and six studies (French 2008; Gucciardi 2007; Jiang 2004; Kendall 1987; Stewart 2005; Zhao 2004) respectively blinded clients, providers and outcome assessors.

Incomplete outcome data

Twelve studies adequately addressed incomplete outcome data (Aldarondo 1999; Arcand 2005; Baraz 2010; Chen 2006; Conrad 2000; McCulloch 1983; Meland 1994; Morey 2008; Racelis 1998; Ryan 2002; Tsay 2003; Zhao 2004) whereas 17 studies did not (Assuncao 2010; Beasley 2008; Becker 1998; Bennett 1986; Blanson 2009; Cummings 1981; Gucciardi 2007; Hyman 2007; Jiang 2004; Jones 1986; Kendall 1987; Laitinen 1993; Logan 2010; Miller 1988; Stewart 2005; Wong 2010; Wood 2008). The principal reason for the incomplete outcome data bias was that missing outcomes are enough to induce clinically‐relevant bias in the observed effect estimate. Nine studies reported insufficient information to permit an evaluation of this criterion (Chiu 2010; French 2008; Gans 1994; Gill 2010, Grace 1996; Hsueh 2007; Mahler 1999; Micco 2007; Scisney‐Matlock 2006).

Selective reporting

Study protocols were available for only one study and all of the study's pre‐specified outcomes that were of interest in the study were reported in the pre‐specified way. Therefore, only this study (French 2008) was free of suggestion of selective outcome reporting. Eighteen studies incompletely reported some outcomes of interest (Aldarondo 1999; Assuncao 2010; Becker 1998; Conrad 2000; Grace 1996; Hyman 2007; Jiang 2004; Kendall 1987; Laitinen 1993; Meland 1994; Micco 2007; Miller 1988; Morey 2008; Racelis 1998; Ryan 2002; Stewart 2005; Wong 2010; Wood 2008) whereas others provided insufficient information to address this criterion (n = 19).

Other potential sources of bias

Eight studies (Cummings 1981; Gucciardi 2007; Logan 2010; Miller 1988; Morey 2008; Scisney‐Matlock 2006; Tsay 2003; Wong 2010) appeared free of other potential sources of bias, whereas 13 studies had at least one important risk of bias such as a baseline imbalance between groups which was not taken into consideration in statistical analyses, a diet adherence not clearly defined, a diet adherence assessed by a non‐validated self‐reporting method, a potential conflict of interest or a potential intervener effect (Assuncao 2010; Beasley 2008; Becker 1998; Chiu 2010; Conrad 2000; French 2008; Gans 1994; Grace 1996; Hsueh 2007; Hyman 2007; Ryan 2002; Stewart 2005; Zhao 2004). Other studies did not report sufficient information to assess other potential sources of bias (n = 17).

Effects of interventions

Included studies differed widely according to interventions provided, measures of diet adherence, dietary advice, nature of the chronic diseases and duration of interventions and follow‐up. Therefore, data were not pooled statistically. Instead, we present a qualitative analysis described in a narrative table using vote counting for each category of interventions (see Additional tables). We also created forest plots for outcomes from studies comparing a single or multiple intervention group with a control/usual care group (see Figure 3; Figure 4; Figure 5). Among the 32 studies that measured diet adherence outcomes between an intervention group and a control/usual care group, 32 out of 123 diet adherence outcomes favoured the intervention group, 4 favoured the control group whereas 62 had no significant difference between groups. This result was impossible to assess for 25 diet adherence outcomes as data and/or statistical analyses needed for comparison between groups were not provided (Additional tables).

Figure 3

Forest plot of comparison: 6 Nutritional tools versus control in diet adherence, outcome: 6.1 Continuous data. *Means represent the difference between pre‐and post‐ intervention.

Figure 4

Forest plot of comparison: 8 Multiple interventions versus control in diet adherence, outcome: 8.1 Continuous data. *Means represent the difference between pre‐and post‐ intervention.

Figure 5

Forest plot of comparison: 8 Multiple interventions versus control in diet adherence, outcome: 8.2 Dichotomous data. *Means represent the difference between pre‐and post‐ intervention.

Education

See Table 1.

Open in table viewer

Table 1. Summary of results: education

Study	Intervention group (description)	Comparative group(s) (description)		Effects on adherence				No of studies (no of participants	Quality of the evidence (GRADE)
				Favours intervention group	Favours comparative group	No difference
Telephone follow‐up								4 (283)	⊝⊝⊝ Very low ^2,4
In summary, among studies using a control/usual care group, three out of ten diet adherence outcomes favoured the intervention group compared to control group and seven diet adherence outcomes had no significant difference between groups. However, these three diet adherence outcomes favouring the intervention group were no longer significant at a later time point.
Chiu 2010	Telephone follow‐up	Control				Adherence to sodium‐restricted diet, fat, fruit and vegetable intakes at 8 weeks
Cummings 1981	Telephone follow‐up	Control; Interventions: (1) contract; (2) contract with the involvement of a family member or friend		vs control: Adherence to potassium‐restricted diet and fluid‐restricted diet at 6 weeks		vs control: Adherence to potassium‐restricted diet and fluid‐restricted diet at 3 months; vs (1) (2): Adherence to potassium‐restricted diet and fluid‐restricted diet at 6 weeks and 3 months
Racelis 1998	Telephone follow‐up	Control				Adherence to diet
Stewart 2005	Telephone follow‐up	Control		Adherence to sodium‐restricted diet at 24 weeks		Adherence to sodium‐restricted diet at 36 weeks; Non‐adherence to alcohol intake at 24 and 36 weeks
Group sessions								2 (144)	⊝⊝ Low ^1,2
In summary, these studies did not allow us to draw conclusions on the effect of group sessions on diet adherence outcomes.
Gill 2010 *	Group sessions	Control
Jones 1986	Group sessions	Interventions: (1) Group sessions and teach to use prompts/cues; (2) Individual sessions with a dietitian; (3) Individual sessions with a dietitian and teach to use prompts/cues				vs (1) (2) (3): Adherence to diet at 16 weeks
Individual sessions with a dietitian								2 (203)	⊝⊝ Low ^1,2
In summary, these studies did not allow us to draw conclusions on the effect of individual sessions with a dietitian on diet adherence outcomes.
Jones 1986	Individual sessions with a dietitian	Interventions: (1) Group sessions and teach to use prompts/cues; (2) Individual sessions with a dietitian; (3) Individual sessions with a dietitian and teach to use prompts/cues				vs (1) (2) (3): Adherence to diet at 16 weeks
Micco 2007 *	Individual sessions with a dietitian	Control
Individual sessions with a nurse								1 (81)	⊝⊝⊝ Very low ^1,2,3
In summary, this study did not allow us to draw conclusions on the effect of individual sessions with a nurse on diet adherence outcomes.
Hsueh 2007	Individual sessions with a nurse		Intervention: telephone follow‐up and individual sessions with a nurse				Adherence to fiber, vegetable and fruit intakes at 3 and 6 months
Educational tools‐video								3 (318)	⊝ Moderate ¹
In summary, among studies using a control/usual care group, two out of three diet adherence outcomes favoured the intervention group compared to the control/usual care group and one diet adherence outcome had no significant difference between groups. However, one out of two diet adherence outcomes favouring the intervention group was no longer significant at a later time point.
Baraz 2010	Educational tools ‐ video	Intervention: group sessions and educational tools – booklet				Adherence to diet and fluid‐restricted diet at 2 months
Mahler 1999	Educational tools ‐ video	Control; Intervention: (1) video and relapse prevention/coping planning		vs control: Adherence to cholesterol and saturated fat‐restricted diet at 1 month		vs control: Adherence to cholesterol and saturated fat‐restricted diet at 3 months; vs (1): Adherence to cholesterol and saturated fat‐restricted diet at 1 and 3 months;
McCulloch 1983	Educational tools ‐ video	Usual care; Intervention: (1) nutritional tool		vs usual care: Adherence to day to day consistency in carbohydrate intake at 6 months		vs (1): Adherence to day to day consistency in carbohydrate intake at 6 months
Educational tools‐booklet								1 (83)	⊝⊝⊝ Very low^,2,3,4
In summary, this study did not allow us to draw conclusions on the effect of booklet on diet adherence outcomes.
Kendall 1987	Educational tools ‐ booklet	Intervention: nutritional tool				Adherence to energy, protein, vitamin A, vitamin C, thiamin, riboflavin, niacin, calcium, phosphorus, iron, zinc intakes at 3 and 6 months

*The authors did not report measures of adherence for both groups, making comparison between groups impossible.

GRADE ‐ Factors decreasing the quality level of a body of evidence:

¹ Limitations in the design and implementation of available studies suggesting high likelihood of bias.

² Indirectness of evidence

³ Imprecision of results

⁴ Downgraded by two levels due to important limitations in the design and implementation of available studies suggesting high likelihood of bias.

Counselling and follow‐up with health professional

Telephone follow‐up

Chiu 2010 assessed the effects of telephone follow‐up on: adherence to a sodium‐restricted diet; fat intake and fruit and vegetable intake, in clients with hypertension. The authors reported no differences in diet adherence between the intervention group and the control group at eight weeks. However, a greater decrease in systolic and diastolic blood pressure was observed in the intervention group compared to the control group as well as a greater increase in exercise adherence.

One study (Cummings 1981) reported significantly higher adherence to a potassium‐restricted diet and fluid‐restricted diet at six weeks in clients with renal failure who received telephone follow‐up, compared to clients in the control group. However, these differences were no longer significant at three months. This study also compared clients with renal failure receiving telephone follow‐up with clients writing a formal agreement (contract) and with clients writing a contract with the involvement of a family member or friend, but found no differences in adherence to a potassium‐ and fluid‐restricted diet at three months between groups.

Racelis 1998 assessed the effects of telephone follow‐up on adherence to diet in clients with peripheral artery disease. The authors indicated that no significant difference was noted between the intervention and the control groups.

Stewart 2005 also evaluated the effects of telephone follow‐up on adherence to a sodium‐restricted diet in clients with hypertension. The authors reported that a higher proportion of clients adhered to the sodium‐restricted diet at 24 weeks in the intervention group compared to the control group, but the difference was no longer significant at 36 weeks. No differences were found in systolic and diastolic blood pressure between groups. The authors also noted no difference in non‐adherence to alcohol intake at 24 and 36 weeks between groups.

Among studies using a control/usual care group, three out of ten diet adherence outcomes favoured the intervention group compared to control group and seven diet adherence outcomes had no significant difference between groups (see Table 1). However, these three diet adherence outcomes favouring the intervention group were no longer significant at a later time point.

Group sessions

Gill 2010 evaluated the effects of group sessions in overweight‐obese college women on adherence to the Dietary Approaches to Stop Hypertension (DASH) diet. However, the authors did not report measures of diet adherence for the intervention and the control groups, making comparison between groups impossible.

Jones 1986 compared an intervention using group sessions (GS) with three other groups for overweight clients: group sessions with a dietitian combined with a leaflet providing advice to reduce exposure to food cues (GS + cues); individual sessions with a dietitian (IS); individual sessions with a dietitian combined with a leaflet providing advice to reduce exposure to food cues (IS + cues). Adherence to diet at 16 weeks was assessed but no significant difference between groups was found. The SMD for weight loss was calculated using RevMan software (RevMan 2012) and no significant difference was found between groups at 16 weeks (vs 1 SMD ‐0.24 (95% CI ‐1.22 to 0.75); vs 2 SMD ‐0.03 (95% CI ‐ 0.94 to 0.88); vs 3 SMD ‐0.55 (95% CI ‐1.55 to 0.46).

Overall, these studies did not allow us to draw conclusions on the effect of group sessions on diet adherence outcomes (see Table 1).

Individual sessions with a dietitian

To assess the effects of a 16‐week intervention promoting individual sessions with a dietitian (IS), Jones 1986 compared this intervention in overweight clients with three others: group sessions with a dietitian (GS); 2) group sessions with a dietitian combined with a leaflet providing advice to reduce exposure to food cues (GS + cues); 3) individual sessions with a dietitian combined with a leaflet providing advice to reduce exposure to food cues (IS + cues). Adherence to diet at 16 weeks was assessed but no significant difference was found between groups. The SMD for weight loss was calculated using RevMan 2012 and no significant difference between groups was found at 16 weeks (vs 1 SMD 0.23 (95% CI ‐0.46 to 0.93); vs 2 = SMD 0.30 (95% CI‐0.69 to 1.08); vs 3 SMD 0.59 (95% CI ‐0.35 to 1.52).

Another study (Micco 2007) evaluated the effects of individual sessions with a dietitian in overweight‐obese clients on diet adherence. The authors assessed diet adherence but they did not report measures for the intervention and the control groups, making the comparison between groups impossible. The authors reported no weight loss difference between groups at 12 months.

Overall, these studies did not allow us to draw conclusions on the effect of individual sessions with a dietitian on diet adherence outcomes (see Table 1).

Individual sessions with a nurse

Hsueh 2007 compared a single intervention comprising individual sessions with a nurse, on adherence to dietary advice to increase fiber, vegetable and fruit intakes in clients with irritable bowel syndrome, with a multiple intervention comprised of individual sessions with a nurse alternating with telephone follow‐up. The authors reported no difference in the proportion of high‐compliant clients for fiber, vegetable and fruit intakes between groups at three months and six months.

Educational tools

Video

One study (Baraz 2010) compared a single intervention using a video as an educational tool with a multiple intervention using a booklet as educational tool, combined with group sessions in clients with chronic end‐stage renal disease. The authors did not report the proportion of clients classified as adherent to diet for both groups, making a comparison between groups impossible. The risk ratio (RR) for the proportion of clients who adhered to the diet and fluid‐restricted diet was calculated using RevMan and no difference was found between groups at two months for diet (RR 0.48 (95% CI 0.17 to 1.35)) and fluid‐restricted diet (RR 0.81 (95% CI 0.25 to 2.57)).

Mahler 1999 evaluated the effects of a video as an educational tool on adherence to a cholesterol and saturated fat‐restricted diet in clients with coronary artery disease. Adherence to a cholesterol and saturated fat‐restricted diet was significantly higher in the intervention group compared to the control group at one month but this difference was no longer significant at three months. The authors also compared the intervention with another intervention using a video as an educational tool combined with relapse prevention/coping planning, and found no difference between groups.

Another study (McCulloch 1983) reported a significant difference in day‐to‐day consistency in carbohydrate intake in clients with insulin dependent diabetes receiving an intervention using a video as an educational tool, compared to the usual care group at six months. Moreover, glycated haemoglobin (HbA_1c) was significantly lower in the intervention group than in the usual care group at six months. The authors also compared the intervention with another intervention using nutritional tool and no difference between groups was noted.

Among studies using a control/usual care group, two out of three diet adherence outcomes favoured the intervention group compared to the control/usual care group and one diet adherence outcome had no significant difference between groups (see Table 1). However, one out of two diet adherence outcomes favouring the intervention group was no longer significant at a later time point.

Booklet

Kendall 1987 compared an intervention using a booklet as an educational tool with an intervention using exchange lists as a nutritional tool in clients with non‐insulin‐dependent diabetes. No difference between groups was reported for adherence to energy, protein, vitamin A, vitamin C, thiamine, riboflavin, niacin, calcium, phosphorus, iron and zinc intakes at three and six months. Moreover, there was no difference between groups for health outcomes such as systolic and diastolic blood pressure, weight, plasma glucose, HbA_1c, serum cholesterol, low‐density lipoprotein (LDL)‐cholesterol, high‐density lipoprotein (HDL)‐cholesterol and serum triglycerides at six months.

Persuasion

See Table 2.

Open in table viewer

Table 2. Summary of results: persuasion

Study	Intervention group (description)	Comparative group(s) (description)	Effects on adherence			No of studies (no of participants)	Quality of the evidence (GRADE)
Study	Intervention group (description)	Comparative group(s) (description)	Favours intervention group	Favours comparative group	No difference	No of studies (no of participants)	Quality of the evidence (GRADE)
Reminders						2 (248)	⊝ Moderate ¹
In summary, among studies using a control/usual care group, 3 out of 19 diet adherence outcomes had no significant difference between groups. It was impossible to assess this result for 16 diet adherence outcomes since data and/or statistical analyses needed for comparison between groups were not provided.
Gans 1994	Reminder ‐ client	Usual care			Adherence to diet at 3 months
Gans 1994	Reminder ‐ physician	Usual care			Adherence to diet at 3 months
Gans 1994	Reminder ‐ client and physician	Usual care			Adherence to diet at 3 months
Ryan 2002 *	Reminder ‐ 2 weeks, 3 and 6 months	Control
Ryan 2002 *	Reminder ‐ 3 and 6 months	Control

*The authors did not report measures of adherence for both groups, making comparison between groups impossible.

GRADE ‐ Factors decreasing the quality level of a body of evidence:

¹ Limitations in the design and implementation of available studies suggesting high likelihood of bias.

² Indirectness of evidence

³ Imprecision of results

⁴ Downgraded by two levels due to important limitations in the design and implementation of available studies suggesting high likelihood of bias.

Reminders

Gans 1994 compared three interventions using reminders with a usual care group in clients with elevated blood cholesterol: 1) clients received the reminder, 2) physicians received a reminder postcard which they could mail to the clients, 3) clients received the reminder in addition to the physicians who received a reminder postcard which they could mail to the clients. The authors reported no difference in the proportion of clients that adhered to diet in any of these groups compared to the usual care group at three months, and no difference between groups for the compliance to lifestyle recommendations at three months.

Another study (Ryan 2002) compared two interventions using knowledge and self‐care practices as reminders with a control group in clients with type II diabetes: 1) reminders provided to clients at two weeks, three months and six months, 2) reminders provided to clients at three months and six months. The authors reported adherence to frequency of meals and snacks combined for all three groups, making comparison between groups impossible.

Overall, the studies used reminders for patients and physicians (Gans 1994) or for patients (Ryan 2002) to enhance adherence to dietary advice. Among studies using a control/usual care group, three out of 19 diet adherence outcomes had no significant difference between groups. It was impossible to assess this result for 16 diet adherence outcomes since data and/or statistical analyses needed for comparison between groups were not provided (see Table 2).

Incentivisation

See Table 3.

Open in table viewer

Table 3. Summary of results: incentivisation

Study	Intervention group (description)	Comparative group(s) (description)	Effects on adherence			No of studies (no of participants)	Quality of the evidence (GRADE)
			Favours intervention group	Favours comparative group	No difference
Contracts with rewards						1 (116)	⊝⊝ Low ^1,3
In summary, among studies using a control/usual care group, two out of four diet adherence outcomes favoured the intervention group compared to the control group and two diet adherence outcomes had no significant difference between groups. However, these two diet adherence outcomes favouring the intervention group were no longer significant at three months.
Cummings 1981	Contract	Control; Interventions: (1) telephone follow‐up; (2) contract with the involvement of a family member or friend	vs control: Adherence to potassium‐restricted diet and fluid‐restricted diet at 6 weeks		vs control: Adherence to potassium‐restricted diet and fluid‐restricted diet at 3 months; vs (1) (2): Adherence to potassium‐restricted diet and fluid‐restricted diet at 6 weeks and 3 months

GRADE ‐ Factors decreasing the quality level of a body of evidence:

¹ Limitations in the design and implementation of available studies suggesting high likelihood of bias.

² Indirectness of evidence

³ Imprecision of results

⁴ Downgraded by two levels due to important limitations in the design and implementation of available studies suggesting high likelihood of bias.

Contracts with rewards

One study (Cummings 1981) reported significantly higher adherence to a potassium‐restricted diet and to a fluid‐restricted diet at six weeks in clients with renal failure who wrote a behavioural contract, compared to clients in the control group (see Table 3). However, these differences were no longer significant at three months (see Table 3). This study also compared clients with renal failure writing a contract with clients receiving telephone follow‐up and with clients writing a contract with the involvement of a family member or friend but no difference was noted in adherence to the potassium‐restricted diet and fluid‐restricted diet at six weeks and three months between groups.

Training

See Table 4.

Open in table viewer

Table 4. Summary of results: training

Study	Intervention group (description)	Comparative group(s) (description)	Effects on adherence			No of studies (no of participants)	Quality of the evidence (GRADE)
Study	Intervention group (description)	Comparative group(s) (description)	Favours intervention group	Favours comparative group	No difference	No of studies (no of participants)	Quality of the evidence (GRADE)
Feedback						3 (661)	⊝⊝ Low ⁴
In summary, among studies using a control/usual care group, one out of seven diet adherence outcomes favoured the intervention group compared to the control/usual care group, four favoured the control group whereas two had no significant difference between groups.
Beasley 2008	Feedback	Control	Adherence to energy, fat, saturated fat and cholesterol intakes at 4 weeks
French 2008	Feedback – less intensive	Usual care		Adherence to general diet and specific diet at 12 months
French 2008	Feedback – most intensive	Usual care		Adherence to general diet and specific diet at 12 months
Meland 1994	Feedback	Control			Adherence to sodium‐restricted diet at 1 and 3 months

GRADE ‐ Factors decreasing the quality level of a body of evidence:

¹ Limitations in the design and implementation of available studies suggesting high likelihood of bias.

² Indirectness of evidence

³ Imprecision of results

⁴ Downgraded by two levels due to important limitations in the design and implementation of available studies suggesting high likelihood of bias.

Feedback

Beasley 2008 reported a higher adherence to energy, fat, saturated fat and cholesterol intakes in overweight‐obese clients in the intervention group using feedback based on self‐monitoring using an electronic food diary compared to the control group. However, no difference in weight loss was observed between groups.

French 2008 compared two interventions using feedback based on self‐monitoring of blood glucose with a usual care group in clients with type II diabetes: 1) less intensive intervention, 2) most intensive intervention. Adherence to general and specific diet at 12 months was greater in the control group compared to both intervention groups.

Another study (Meland 1994) assessed the effects of feedback using self‐monitoring of urine chloride concentration on adherence to a sodium‐restricted diet in clients with hypertension. No difference was reported in adherence to the sodium‐restricted diet or in blood pressure between the intervention group and the control group at one and three months.

In this category, three studies used feedback based on self‐monitoring using an electronic food diary (Beasley 2008), blood glucose (French 2008) and urine chloride concentration (Meland 1994). Among studies using a control/usual care group, one out of seven diet adherence outcomes favoured the intervention group compared to the control/usual care group, four favoured the control group whereas two had no significant difference between groups (see Table 4).

Restriction

See Table 5.

Open in table viewer

Table 5. Summary of results: restriction

Study	Intervention group (description)	Comparative group(s) (description)	Effects on adherence			No of studies (no of participants)	Quality of the evidence (GRADE)
			Favours intervention group	Favours comparative group	No difference
Restriction						1 (7)	⊝⊝⊝ Very low^1,2,3
In summary, this study did not allow us to draw conclusions on the effect of restriction on diet adherence outcomes.
Conrad 2000*	Restriction	Control			Adherence to very low fat diet at 7 months

*The authors did not report measures of adherence for both groups, making comparison between groups impossible.

GRADE ‐ Factors decreasing the quality level of a body of evidence:

¹ Limitations in the design and implementation of available studies suggesting high likelihood of bias.

² Indirectness of evidence

³ Imprecision of results

⁴ Downgraded by two levels due to important limitations in the design and implementation of available studies suggesting high likelihood of bias.

Conrad 2000 assessed the effects of an intervention proposing an incremental reduction in fat to a goal of 10% of energy intake compared to an intervention proposing an immediate reduction in fat to a goal of 10% of energy intake in clients with coronary artery disease. The authors did not compare adherence to fat intake advice between groups. Therefore, we calculated the SMD for adherence to the very low fat diet using RevMan 2012 and found no differences between groups at seven months (SMD ‐1.88 (95% CI ‐4.00 to 0.23)) (see also Table 5).

Modelling

See Table 6.

Open in table viewer

Table 6. Summary of results: modelling

Study	Intervention group (description)	Comparative group(s) (description)	Effects on adherence			No of studies (no of participants)	Quality of the evidence (GRADE)
Study	Intervention group (description)	Comparative group(s) (description)	Favours intervention group	Favours comparative group	No difference	No of studies (no of participants)	Quality of the evidence (GRADE)
Nutritional tools						7 (514)	⊝⊝⊝ Very low^2,4
In summary, among studies using a control/usual care group, 3 out of 17 diet adherence outcomes favoured the intervention group and 11 diet adherence outcomes had no significant difference between groups. It was impossible to assess this result for three diet adherence outcomes as data and/or statistical analyses needed for comparison between groups were not provided.
Assuncao 2010	Nutritional tools	Usual care	Adherence to sodium at 6 months		Adherence to energy, protein, fat, carbohydrate, cholesterol, fiber, fruit, vegetable and sweet food intakes at 6 months
Chen 2006	Nutritional tools	Control	Adherence to protein intake at 1 month
Grace 1996	Nutritional tools	Control	Adherence to fat intakes at 12 weeks		Adherence to energy at 12 weeks
Kendall 1987	Nutritional tools	Intervention: educational tool ‐ booklet			Adherence to energy, protein, vitamin and mineral intakes at 3 and 6 months
Logan 2010	Nutritional tools	Intervention: Barrier identification/problem solving and goal setting			Adherence to Mediterranean diet at 6 and 12 months
McCulloch 1983	Nutritional tools	Usual care; Intervention: (1) educational tool ‐ video			vs control and (1): Adherence to day to day consistency in carbohydrate intake at 6 months
Scisney‐Matlock 2006*	Nutritional tools	Control

*The authors did not report measures of adherence for both groups, making comparison between groups impossible.

GRADE ‐ Factors decreasing the quality level of a body of evidence:

¹ Limitations in the design and implementation of available studies suggesting high likelihood of bias.

² Indirectness of evidence

³ Imprecision of results

⁴ Downgraded by two levels due to important limitations in the design and implementation of available studies suggesting high likelihood of bias.

Nutritional tools

Assuncao 2010 assessed the effects of nutritional tools such as portion size examples and food lists on diet adherence in overweight‐obese clients using an intention‐to‐treat analysis. Authors reported a significant enhancement of adherence to sodium and sweet food intake goals at six months in clients in the intervention group compared with those receiving usual care. However, a discrepancy was found between the results provided by the authors and the SMD calculated using RevMan which showed no difference for adherence to sweet food intake at six months between groups. No difference was found between groups for adherence to recommended energy, protein, fat, carbohydrate, cholesterol, fiber, fruit and vegetable intakes (see Analysis 1.1). An increase in physical leisure activity as well as a decrease in fasting glucose were reported in the intervention compared to the usual care group at six months, whereas no difference between groups was observed for weight loss, blood pressure and lipid profile.

Chen 2006 reported a higher proportion of intervention‐group clients with renal failure reaching the target for protein intake at one month using menu suggestions, exchange lists and portion sizes as nutritional tools compared to the control group.

Grace 1996 evaluated the effects of nutritional tools such as an additional package containing low‐fat cooking methods and low‐fat recipe adaptation on adherence to energy and fat intakes in clients with hyperlipidaemia. The authors reported a higher reduction in percentage of fat intake in the intervention group compared to the control group. However, they reported no difference for energy change between the intervention and the control groups at 12 weeks (see also Analysis 1.1).

Kendall 1987 compared an intervention using exchange lists as a nutritional tool with an intervention using a booklet as an educational tool in clients with non‐insulin‐dependent diabetes. No difference between groups was reported for adherence to energy, protein, vitamin A, vitamin C, thiamine, riboflavin, niacin, calcium, phosphorus, iron and zinc intakes at three and six months. Moreover, there was no difference between groups for health outcomes such as systolic and diastolic blood pressure, weight, plasma glucose, HbA_1c, serum cholesterol, LDL‐cholesterol, HDL‐cholesterol and serum triglycerides at six months.

One study (Logan 2010) compared an intervention using recipes and meal plans with an intervention using barrier identification/problem solving and goal setting in clients with coronary heart disease. The authors reported no difference between groups for adherence to the Mediterranean diet at 6 and 12 months.

Another study (McCulloch 1983) reported no difference in day‐to‐day consistency in carbohydrate intake in clients with insulin‐dependent diabetes following an intervention using exchange lists and lunch time with health professionals as nutritional tools, compared to usual care group. However, HbA_1c was significantly lower in the intervention group at 9 months compared to the control group. The authors also compared the intervention with another intervention using a video as an educational tool and found no difference between groups.

Scisney‐Matlock 2006 evaluated the effects of wheels and bar charts displaying Cognitive Representations of the DASH diet as a nutritional tool on adherence to the DASH diet in clients with hypertension compared to a control group. The authors reported results grouped for both groups, making comparison between groups impossible.

To summarize the interventions in this category: two studies included portion sizes (Assuncao 2010; Chen 2006), three studies used menu suggestions and recipes (Chen 2006; Grace 1996; Logan 2010), three studies included exchange lists (Chen 2006; Kendall 1987; McCulloch 1983), one study used an additional package containing low‐fat cooking methods (Grace 1996), one study used lunch time with health professionals (McCulloch 1983), and one study used wheels and bar charts displaying Cognitive Representations of the DASH diet (Scisney‐Matlock 2006) as nutritional tools in their intervention.

Among studies using a control/usual care group, 3 out of 17 diet adherence outcomes favoured the intervention group and 11 diet adherence outcomes had no significant difference between groups. It was impossible to assess this result for three diet adherence outcomes as data and/or statistical analyses needed for comparison between groups were not provided (Table 6).

Enablement

See Table 7.

Open in table viewer

Table 7. Summary of results: enablement

Study	Intervention group (description)	Comparative group(s) (description)	Effects on adherence			No of studies (no of participants)	Quality of the evidence (GRADE)
Study	Intervention group (description)	Comparative group(s) (description)	Favours intervention group	Favours comparative group	No difference	No of studies (no of participants)	Quality of the evidence (GRADE)
Behaviour change techniques						3 (136)	⊝⊝⊝ Very low^2,4
In summary, only one study used a control group and three out of three diet adherence outcomes had no difference between groups.
Aldarondo 1999	Barrier identification/problem solving and self‐talk	Control			Adherence to energy, fat and saturated fat intakes at 14 weeks
Bennett 1986	Teach to use prompts/cues	Interventions: (1) self‐talk; (2) barrier identification/problem solving	vs (1) (2): Adherence to energy intake between baseline and 15 weeks
Bennett 1986	Self‐talk	Interventions: (1) teach to use prompts/cue; (2) barrier identification/problem solving		vs (1): Adherence to energy intake between baseline and 15 weeks	vs (2): Adherence to energy intake between baseline and 15 weeks
Bennett 1986	Barrier identification/problem solving	Interventions: (1) teach to use prompts/cue; (2) self‐talk		vs (1): Adherence to energy intake between baseline and 15 weeks	vs (2): Adherence to energy intake between baseline and 15 weeks
Logan 2010	Barrier identification/problem solving and goal setting	Intervention: Nutritional tools			Adherence to Mediterranean diet at 6 and 12 months

GRADE ‐ Factors decreasing the quality level of a body of evidence:

¹ Limitations in the design and implementation of available studies suggesting high likelihood of bias.

² Indirectness of evidence

³ Imprecision of results

⁴ Downgraded by two levels due to important limitations in the design and implementation of available studies suggesting high likelihood of bias.

Behaviour change techniques

Aldarondo 1999 reported no difference in adherence to energy, fat and saturated fat intake at 14 weeks between the intervention group using barrier identification/problem solving and self‐talk compared to the control group in obese clients.

Another study (Bennett 1986) compared three interventions using behavioural change techniques in overweight‐obese clients: 1) teaching clients to use prompts/cues, 2) self‐talk, 3) barrier identification/problem solving. The authors reported that clients in the intervention group using food cues adhered more closely to energy intake goals than those in the two other groups between baseline and 15 weeks.

One study (Logan 2010) compared an intervention using barrier identification/problem solving and goal setting with an intervention using recipes and meal plans as nutritional tools in clients with coronary heart disease. The authors reported no difference between groups for adherence to the Mediterranean diet at 6 and 12 months.

Overall, in this category: three studies used behavioural change techniques such as barrier identification/problem solving and self‐talk (Aldarondo 1999), teaching clients to use prompts/cues, self‐talk and barrier identification/problem solving (Bennett 1986) and barrier identification/problem solving and goal setting (Logan 2010).

Only one study used a control group and three out of three diet adherence outcomes had no difference between groups (see Table 7).

Multiple interventions

See Table 8.

Open in table viewer

Table 8. Summary of results: multiple interventions

Study	Intervention group (description)	Comparative group(s) (description)	Effects on adherence			No of studies (no of participants)	Quality of the evidence (GRADE)
Study	Intervention group (description)	Comparative group(s) (description)	Favours intervention group	Favours comparative group	No difference	No of studies (no of participants)	Quality of the evidence (GRADE)
Multiple interventions						18 (7700)	**
In summary, among studies using a control/usual care group, 21 out of 56 diet adherence outcomes favoured the intervention group whereas 32 diet adherence outcomes had no significant difference between groups. It was impossible to assess this result for three diet adherence outcomes as data and/or statistical analyses needed for comparison between groups were not provided. However, 4 out of 21 diet adherence outcomes favouring the intervention group was no longer significant at a later time point.
Arcand 2005	Individual sessions with a dietitian and goal setting	Usual care			Adherence to sodium‐restricted diet at 3 months
Baraz 2010	Intervention: group sessions and educational tools – booklet	Intervention: educational tools ‐ video			Adherence to diet and fluid‐restricted diet at 2 months
Becker 1998	Telephone follow‐up and barrier identification/problem solving	Usual care			Adherence to fat‐restricted diet at 2 years
Blanson 2009	Motivational interviewing and self‐monitoring and feedback‐diary	Control			Adherence to diet at 28 days
Cummings 1981	Contract with the involvement of a family member or friend	Control; Interventions: (1) telephone follow‐up; (2) contract	vs control: Adherence to potassium‐restricted diet and fluid‐restricted diet at 6 weeks		vs control: Adherence to potassium‐restricted diet and fluid‐restricted diet at 3 months; vs (1) (2): Adherence to potassium‐restricted diet and fluid‐restricted diet at 3 months
Gucciardi 2007	Group sessions, nutritional tools and barrier identification/problem solving	Control	Adherence to diet at 3 months
Hsueh 2007	Telephone follow‐up and individual sessions with a nurse	Intervention: Individual sessions with a nurse			Adherence to fiber, vegetable and fruit intake at 3 and 6 months
Hyman 2007	Telephone follow‐up and motivational interviewing ‐ simultaneous	Usual care; Intervention: (1) telephone follow‐up and motivational interviewing ‐ sequential	vs usual care and (1): Adherence to sodium‐restricted diet at 6 months		vs usual care and (1): Adherence to sodium‐restricted diet at 18 months
Hyman 2007	Telephone follow‐up and motivational interviewing ‐ sequential	Usual care; Intervention: (1) telephone follow‐up and motivational interviewing ‐ simultaneous		vs (1): Adherence to sodium‐restricted diet at 6 months	vs control: Adherence to sodium‐restricted diet at 6 months and 18 months; vs (1): Adherence to sodium‐restricted diet at 18 months
Jiang 2004	Individual sessions with a nurse, telephone follow‐up and goal setting	Usual care	Adherence to ATP step II diet at 3 and 6 months.
Jones 1986	Group sessions and teach to use prompts/cues	Interventions: (1) group sessions; (2) individual sessions with a dietitian; (3) individual sessions with a dietitian and teach to use prompts/cues			vs (1) (2) (3): adherence to diet at 16 weeks
Jones 1986	Individual sessions with a dietitian and teach to use prompts/cues	Interventions: (1) group sessions; (2) individual sessions with a dietitian; (3) group sessions and teach to use prompts/cues			vs (1) (2) (3): adherence to diet at 16 weeks
Laitinen 1993	Individual sessions with a dietitian, nutritional tools and goal setting	Usual care	Adherence to saturated fat intakes at 15 months		Adherence to saturated fat intakes at 3 months; Adherence to total fat, unsaturated fat, carbohydrate, fiber and cholesterol intakes at 3 and 15 months
Mahler 1999	Educational tools – video and relapse prevention/coping planning	Control; Intervention: (1) educational tools ‐ video	vs control: Adherence to cholesterol and saturated fat‐restricted diet at 1 month.		vs control: Adherence to cholesterol and saturated fat‐restricted diet at 3 months; vs (1): Adherence to cholesterol and saturated fat ‐restricted diet at 1 and 3 months
Miller 1988	Individual sessions with a nurse and barrier identification/problem solving and goal setting	Control	Adherence to diet at 2 years		Adherence to diet at 30 days, 60 days and 1 year
Morey 2008	Individual sessions with a dietitian, educational tools‐booklet, reminders, motivational interviewing	Control	Adherence to phosphate‐restricted diet at 3 months
Tsay 2003	Self‐monitoring and feedback‐diary, stress management and goal setting	Usual care	Adherence to fluid‐restricted diet at 3 months and 6 months		Adherence to fluid‐restricted diet at 1 month
Wong 2010	Telephone follow‐up and goal setting	Control			Non‐adherence to diet (days and degree) at 7 weeks and 13 weeks Non‐adherence to fluid‐restricted diet (days and degree) at 7 weeks and 13 weeks
Wood 2008 – coronary heart disease	Individual sessions with a nurse and motivational interviewing	Usual care	Adherence to saturated fat, oily fish, fish and fruit and vegetable intakes at 1 year
Wood 2008 – high risk of coronary heart disease	Individual sessions with a nurse and motivational interviewing	Usual care	Adherence to oily fish, fish and fruit and vegetables intakes at 1 year
Zhao 2004	Telephone follow‐up, individual sessions with a dietitian and goal setting	Usual care	High adherence to diet at 4 and 12 weeks

**Multiple interventions included a variety of interventions, which did not allow the use of GRADE.

Arcand 2005 evaluated the effects of individual sessions with a dietitian combined with goal setting, on adherence to a sodium‐restricted diet in clients with heart failure. The authors did not compare adherence to the sodium‐restricted diet nor blood pressure between groups. Therefore, we calculated the SMD for adherence to the sodium‐restricted diet and blood pressure using RevMan, and found no difference between groups for sodium‐restricted diet (see also Analysis 2.1), systolic blood pressure (SMD‐0.30 (95% CI ‐0.88 to 0.27)) and diastolic blood pressure (SMD‐0.53 (95% CI ‐1.11 to 0.05)).

One study (Baraz 2010) compared a multiple intervention using a booklet as educational tool combined with group sessions, with a single intervention using a video as an educational tool, in clients with chronic end‐stage renal disease. The authors did not report the proportion of clients classified as adherent to diet for both groups, making comparison between groups impossible. Therefore, we calculated the RR for the proportion of clients who adhered to the diet and fluid‐restricted diet at two months, using RevMan, and found no difference between groups.

Using an intention‐to‐treat analysis, Becker 1998 reported no difference in the proportion of clients at risk of coronary heart disease who received telephone follow‐up combined with a barrier identification/problem solving intervention for adherence to a fat‐restricted diet at two years, compared to clients in the usual care group. Moreover, no difference was found for LDL‐cholesterol, HDL‐cholesterol and triglyceride levels at two years between groups.

Blanson 2009 evaluated the effects of self‐monitoring using a computer assistant combined with feedback using motivational interviewing in overweight clients. They reported no significant difference in adherence to diet at 28 days between the intervention and the control groups.

Cummings 1981 reported a significantly higher adherence to a fluid‐restricted diet at six weeks in clients with renal failure asked to write a formal agreement (contract) with the involvement of a family member or friend, compared to clients in the control group. However, these differences were no longer significant at three months. This study also compared clients writing a formal agreement (contract) with the involvement of a family member or friend, with clients writing a contract, and with clients who received telephone follow‐up, but no differences in adherence to a potassium‐ and fluid‐restricted diet at three months were found between groups.

In type II diabetes clients, the comparison of an intervention using group sessions and nutritional tools combined with barrier identification/problem solving versus control (Gucciardi 2007) showed a higher adherence to dietary advice in the intervention group at three months. However, the authors reported no difference in HbA_1c between the groups at three months (see also Analysis 2.1).

Hsueh 2007 compared a multiple intervention comprising individual sessions with a nurse alternating with telephone follow‐up on adherence to fiber, vegetable and fruit intakes in clients with irritable bowel syndrome, with a single intervention comprising individual sessions with a nurse. The authors reported no difference in the proportion of high‐compliant clients for fiber, vegetable and fruit intakes between groups at three and six months.

To assess the effectiveness of an intervention using telephone follow‐up combined with motivational interviewing, Hyman 2007 compared two interventions in clients with hypertension with a usual care group: 1) simultaneous behaviour change (stop smoking, reduce dietary sodium level and increase physical activity); 2) sequential behaviour change (stop smoking, then reduce dietary sodium levels and finally increase physical activity). A higher proportion of clients adhered to the sodium‐restricted diet in the simultaneous group, compared to the sequential intervention and the usual care group at six months, but no difference was observed at 18 months. No difference was reported for blood pressure between groups (see also Analysis 2.2).

Jiang 2004 assessed the effects of an intervention using individual sessions with a nurse and telephone follow‐up combined with goal setting, on adherence to the Adult Treatment Panel (ATP) step II diet (hypocholesteraemic diet) in clients with angina pectoris or myocardial infarction. Using an intention‐to‐treat analysis, the authors reported better adherence to the step II diet in the intervention group compared to the usual care group at three and six months. At three months, triglyceride, total cholesterol, LDL‐cholesterol levels and blood pressure decreased significantly more in the intervention group than the usual care group, while no difference was noted for HDL‐cholesterol and body weight. At six months, only the differences in triglyceride, total cholesterol and LDL‐cholesterol levels remained significant.

Jones 1986 compared four interventions in overweight clients: group sessions with a dietitian (GS); group sessions with a dietitian combined with a leaflet providing advice to reduce exposure to food cues (GS + cues); individual sessions with a dietitian (IS); individual sessions with a dietitian combined with a leaflet providing advice to reduce exposure to food cues (IS + cues). The authors found no significant difference between groups for adherence to diet, as well as for weight loss, at 16 weeks.

Laitinen 1993 evaluated the effects of individual sessions with a dietitian and nutritional tools combined with goal setting, on adherence to total fat, saturated fat, unsaturated fat, carbohydrate, fiber and cholesterol advice in clients with non‐insulin‐dependent diabetes. Although the authors reported no differences for total, saturated and unsaturated fat intake at three months, there was a higher proportion of clients who adhered to total and saturated fat intake recommendations in the intervention group compared to the usual care group at 15 months, whereas a higher proportion of clients adhered to unsaturated fat in the usual care group at 15 months. However, a discrepancy was found between the results provided by the authors and the RR calculated using RevMan which revealed no difference for adherence in total fat and unsaturated fat at 15 months between groups. Fasting blood glucose and HbA_1c decreased significantly more in the intervention group at 15 months than in the control group, while no difference was noted for body weight, total cholesterol and HDL‐cholesterol levels. From data provided by the authors, we used RevMan to calculate the SMD for the proportion of clients who adhered to carbohydrate, fiber and cholesterol intakes, and found no differences between groups at 3 and 15 months (see also Analysis 2.2).

Mahler 1999 assessed the effects of a video as educational tool combined with relapse prevention/coping planning on adherence to a cholesterol‐ and saturated fat‐restricted diet in clients with coronary artery disease. Adherence to a cholesterol‐ and saturated fat‐restricted diet was significantly higher in the intervention group compared to the control group at one month, but this difference was no longer significant at three months. The authors also compared the intervention with another intervention using a video as an educational tool, and found no difference between groups.

Miller 1988 evaluated the effects of individual sessions with a dietitian combined with barrier identification/problem solving and goal setting in clients with myocardial infarction. While no difference was found at 30 days, 60 days and 1 year, the authors reported a significant difference in adherence to diet at 2 years between the intervention and the control groups.

Morey 2008 compared an intervention including individual sessions with a nurse, a booklet as educational tool and reminders combined with motivational interviewing intervention with a control group. They reported a higher proportion of clients with end‐stage kidney disease adhering to a phosphate‐restricted diet at three months in the intervention group compared to the control group. Data for adherence to the phosphate‐restricted diet at 6 and 12 months were not reported (see also Analysis 2.2).

A multiple intervention (Tsay 2003) including self‐monitoring in a diary and feedback combined with stress management and goal setting in clients with end‐stage renal disease showed a significant group main effect in adherence to a fluid‐restricted diet when baseline mean weight gains were applied as covariate. From data provided by the authors, we used RevMan to calculate the SMD for adherence to a fluid‐restricted diet at 1 month, 3 months and 6 months, respectively. No difference was found between groups at one month but adherence to a fluid‐restricted diet at three months and six months was significantly higher in the intervention group compared to the control group (see also Analysis 2.1).

Wong 2010 reported a difference in the degree of non‐adherence to diet at seven weeks in clients with renal failure who received telephone follow‐up combined with goal setting compared to clients in the control group. However, a discrepancy was found between the results provided by the authors and the SMD calculated using RevMan which revealed no difference for the degree of non‐adherence to diet at seven weeks between groups. No difference was found between groups for the degree of non‐adherence to diet at 13 weeks and for the number of days of non‐adherence to diet, as well as non‐adherence to fluid restriction (degree and days) at 7 and 13 weeks (see also Analysis 2.1).

Wood 2008 studied two populations: clients with coronary heart disease and clients at high risk of the disease. In clients with coronary heart disease, the authors reported a higher proportion of clients achieving the target for saturated fat, oily fish and fruit and vegetable intakes at one year in the intervention group (individual sessions with a nurse combined with motivational interviewing) compared to the usual care group. No difference was observed in adherence to fish consumption advice between groups. However, a discrepancy was found between the results provided by the authors and the RR calculated using RevMan which revealed a higher proportion of clients achieving the target for fish intake in the intervention group. A higher proportion of clients achieved the target for blood pressure in the intervention group compared to the usual care group, while no difference was found for body weight, and total and LDL‐cholesterol levels. In clients at high risk of coronary heart disease, a higher proportion of clients achieving the target of fruit and vegetable intakes was reported at one year in the intervention group, while no difference was observed in adherence to recommended fish and oily fish intakes between groups. However, a discrepancy was found between the results provided by the authors and the RR calculated using RevMan which revealed a higher proportion of clients achieving the target for oily fish and fish intake in the intervention group. A higher proportion of clients also achieved the target for blood pressure and body weight in the intervention group compared to the usual care group while no difference was found for total and LDL‐cholesterol levels (see also Analysis 2.2).

One study (Zhao 2004) evaluating the effects of telephone follow‐up as well as individual sessions with a dietitian combined with goal setting in clients with coronary heart disease reported a higher proportion of clients with high adherence to diet in the intervention group compared to the usual care group at 4 and 12 weeks (see also Analysis 2.2).

Overall, in this category, 13 studies combined an educational intervention with another intervention such as an enablement intervention (Arcand 2005; Becker 1998; Gucciardi 2007; Hyman 2007; Jones 1986, Mahler 1999; Miller 1988; Wong 2010; Wood 2008; Zhao 2004), modelling and enablement interventions (Laitinen 1993), persuasion and enablement interventions (Morey 2008), and two educational interventions with enablement interventions (Jiang 2004). Two studies combined two different educational interventions (Baraz 2010; Hsueh 2007). One study combined a training intervention with an enablement intervention (Blanson 2009) and one study combined two enablement interventions and a training intervention (Tsay 2003). One study combined an incentivisation with a persuasion intervention (Cummings 1981).

In this category, among studies using a control/usual care group, 21 out of 56 diet adherence outcomes favoured the intervention group whereas 32 diet adherence outcomes had no significant difference between groups. It was impossible to assess this result for three diet adherence outcomes as data and/or statistical analyses needed for comparison between groups were not provided (Table 8). However, 4 out of 21 diet adherence outcomes favouring the intervention group was no longer significant at a later time point.

Discussion

Summary of main results

This review included 38 studies investigating the effects of interventions enhancing adherence to dietary advice for preventing and managing chronic diseases in adults. Studies reporting at least one diet adherence outcome showing statistically significant differences favouring the intervention group included the following interventions: telephone follow‐up, video, contract, feedback, nutritional tools and multiple interventions. However, these interventions also showed no difference in some diet adherence outcomes compared to a control/usual care group. Moreover, the included studies differed widely according to interventions provided, measures of diet adherence, dietary advice, nature of the chronic diseases and duration of interventions and follow‐up.

The majority of these studies were conducted in United States of America. Cardiovascular disease, diabetes, hypertension, and renal diseases were the most frequently studied chronic diseases. The adoption of a healthy diet is recommended as a prevention or management strategy for each of these chronic diseases (Lichtenstein 2006; Bantle 2008; Kopple 2001). Interestingly, all studies including clients with renal diseases reported at least one diet adherence outcome showing a statistically significant difference favouring the intervention group, no matter which intervention was provided.

Only 10 of the 38 included studies evaluated diet adherence to dietary advice over a 12‐month period (Becker 1998; French 2008; Hyman 2007; Laitinen 1993; Logan 2010; Micco 2007; Miller 1988; Morey 2008; Racelis 1998; Wood 2008). Among those 10 studies, only three studies showed at least one statistically significant difference in diet adherence outcomes favouring the intervention group over a 12‐month period.

A broad range of interventions, all related to the method for changing dietary habits through dietary advice, was covered in this review, including education (telephone follow‐up, group sessions, individual sessions with a dietitian or a nurse, and educational tools (video or booklet)), persuasion (reminders), incentivisation (contracts with rewards), training (feedback), restriction, modelling (nutritional tools) and enablement (behaviour change techniques). However, the majority of studies included a combination of two or more different interventions.

This review included studies comparing one or more intervention group(s) with one control/usual care group, but also studies comparing two or more intervention groups to each other. However, only comparisons made between an intervention group and a control/usual care group allowed the evaluation of the effect of the intervention alone on adherence to dietary advice. Therefore, among studies that measured diet adherence outcomes between an intervention group and a control/usual care group, 32 out of 123 diet adherence outcomes favoured the intervention group. More specifically, studies reporting at least one diet adherence outcome showing statistically significant differences favouring the intervention group included the following interventions: telephone follow‐up (3 out of 10 diet adherence outcomes), video (2 out of 3 diet adherence outcomes), contract (2 out of 4 diet adherence outcomes), feedback (1 out of 7 diet adherence outcomes), nutritional tools (3 out of 17 diet adherence outcomes) and multiple interventions (21 out of 56 diet adherence outcomes). Studies investigating interventions such as a group session, individual session, reminders, restriction and behaviour change techniques reported no diet adherence outcome showing a statistically significant difference favouring the intervention group. However, these results should be interpreted with caution as several studies evaluated two or more diet adherence outcomes. Among those, most of the studies showing a statistically significant difference favouring the intervention group for diet adherence outcome(s) also showed no significant differences between groups for other diet adherence outcome(s) (Assuncao 2010; Cummings 1981; Grace 1996;Hyman 2007; Laitinen 1993; Mahler 1999; McCulloch 1983; Miller 1988; Stewart 2005; Tsay 2003). For example, Laitinen 1993 assessed the effects of a multiple intervention and reported better adherence to saturated fat intake at 15 months in the intervention group whereas no differences were observed for adherence to intake of total, saturated or unsaturated fat carbohydrate, fiber or cholesterol between the intervention group and the control group at either 3 or 15 months. In addition, where studies measured outcomes at multiple time points, the majority of studies reporting a diet adherence outcome favouring the intervention group compared to the control/usual care group in the short‐term also reported no significant effect at later time points. Interestingly, the majority of studies involving multiple interventions reported positive results on adherence to dietary advice. However, because multiple components within these interventions acted as co‐interventions, it may have introduced confounding effects. Therefore, drawing conclusions about whether the interventions enhanced adherence to dietary advice is very difficult.

Overall completeness and applicability of evidence

Although we included a substantial number of studies covering a broad range of chronic diseases and interventions, very few studies assessing a specific chronic disease condition evaluated the same intervention. In addition, measures of adherence and dietary advice varied widely across studies.

This review assessed the effects on adherence related to the intervention alone since only the intervention, related to the method for facilitating changes in dietary habits through dietary advice, differed between the intervention group and the control/usual care group. Comparisons between two or more intervention groups were also reported. However, comparisons between multiple interventions were all different. In order to isolate the effects of the intervention, both clients in the intervention group and the control/usual care group received the same dietary advice related to their chronic disease condition. This situation could explain why adherence to dietary advice in the control/usual care group increased in some studies. However, factors other than the intervention provided could have affected adherence to dietary advice. For example, clients' intrinsic characteristics such as an elevated level of self‐efficacy (Mishali 2011; Aljasem 2001) as well as few perceived barriers (Walsh 2011) are associated with better dietary adherence in clients with chronic diseases. Some studies also reported that the client’s stage of change based on the Transtheoretical Model predicted long‐term changes in dietary behaviours (Mochari 2010; Blissmer 2010). Therefore, confounding factors should be taken into consideration in studies evaluating adherence to dietary advice.

In this review, secondary outcomes related directly to the chronic disease condition (e.g. HbA_1c and/or blood glucose in clients with diabetes, weight for clients with obesity) were reported. Few studies reported other secondary outcomes such as process measures, services outcomes and harms or secondary effects, making interpretation about these secondary outcomes impossible. Fourteen studies comparing an intervention group with a control/usual care group also reported clinical and/or biochemical outcome(s) in addition to adherence to dietary advice. Among those, six reported improvement in at least one chronic disease‐related clinical or biochemical outcome in the intervention group. As mentioned earlier, these results should be interpreted with caution as several studies evaluated two or more clinical and/or biochemical outcomes.

Seventeen studies provided advice in order to induce changes other than diet such as physical activity, medication compliance, smoking cessation and blood glucose monitoring. All of these studies independently assessed adherence to dietary advice, but because those studies varied widely according to interventions provided and nature of the chronic diseases, we cannot conclude that adherence to dietary advice is improved when multifaceted interventions are provided.

Quality of the evidence

Despite a high number of included studies (n = 38), these studies varied widely according to interventions provided, measures of diet adherence, dietary advice, nature of the chronic diseases and duration of interventions and follow‐up. The numbers of clients included in the review is impressive (9445), but the range of number of clients in each study was wide, varying from 7 to 5405 clients. Only 13 of the 38 included studies provided a power calculation (Aldarondo 1999; Assuncao 2010; Beasley 2008; Chiu 2010; French 2008; Hyman 2007; Jiang 2004; Meland 1994; Stewart 2005; Tsay 2003; Wong 2010; Wood 2008; Zhao 2004) and among them, 10 studies recruited the number of clients according to their power analysis (Aldarondo 1999; Assuncao 2010; Beasley 2008; French 2008; Jiang 2004; Meland 1994; Stewart 2005; Tsay 2003; Wong 2010; Zhao 2004).

While an elevated drop‐out rate could be considered as an indirect measure of non‐adherence, such as in studies of pharmaceutical interventions where participants who withdraw no longer have access to medication, it cannot be assumed that clients dropping out of dietary intervention studies are non‐adherent to dietary advice. Most studies included in this review had a low drop‐out rate. In fact, 19 studies reported a drop‐out rate lower than 20% (9 of those had no drop‐out). Nine studies had a drop‐out rate of between 20% and 30% and only five studies had a drop‐out rate over 30%. It was impossible to calculate the drop‐out rate for six studies (Gans 1994; Gill 2010; Hsueh 2007; Jones 1986; Mahler 1999; McCulloch 1983). One study (Wood 2008) reported adherence for two populations, which explains why the total number of included studies adds up to 39, and not 38.

The majority of included studies were of poor methodological quality and/or poorly reported risk of bias elements. All included studies met less than five of the eight criteria of risk of bias (see Assessment of risk of bias in included studies). Among those eight criteria, three of them evaluated respectively the blinding of clients, providers and outcome assessors. Very few included studies met these criteria because blinding in the context of delivering a nutritional intervention is very difficult to achieve, even impossible in some designs. Unlike most pharmaceutical designs using placebo, both clients and providers from nutritional studies usually know which intervention is delivered.

A major challenge in the measurement of diet adherence is the correct estimation of dietary intake, as no method for accurate determination of dietary intake has been developed yet. In this review, 31 studies used self‐reported measures of diet adherence while 6 studies assessed diet adherence using objective measures. Objective measures included serum micronutrients (e.g. potassium, sodium, phosphate) and interdialytic weight gain to evaluate respectively adherence to diet and to fluid‐restricted diet in clients with renal diseases, and urinary electrolytes excretion (sodium, chloride) to evaluate adherence to a sodium‐restricted diet in clients with hypertension. Those methods have been validated and are usually more reliable than self‐reported measures. However, the assessment of many food and nutrient intakes cannot always be performed by objective measures, especially when dietary advice targets food groups (e.g. fruit and vegetables) rather than a specific nutrient (e.g. sodium). The Academy of Nutrition and Dietetics states that "total diet or overall pattern of food eaten is the most important focus of a healthful eating style" (JADA 2007). Consequently, most studies providing dietary advice focusing on a global healthy diet rather than a specific nutrient used self‐reported methods such as dietary tools (e.g. food records, food frequency questionnaires and validated diet questionnaires or scales). Misreporting of dietary intake is a major issue and has been related to body mass index, age, sex, socio‐economic status and education (Poslusna 2009). In addition, other sources of misreporting have been identified such as memory relapses, misrepresentation of portion size consumed, social desirability and daily dietary variability (Kumanyika 2000; Wilson 2005). Therefore, establishing validity and reliability of dietary tools is crucial in order to avoid inconsistent estimates of dietary intake leading to a high risk of bias. In this review, only 14 studies of 32 stated that the self‐reported measures of diet adherence had been validated and/or shown to be reliable, suggesting that adherence to dietary advice in those studies could be biased. To gain a thorough understanding of adherence to dietary advice, both self‐report and objective measures of adherence are needed. While objective measures provide information on food intake only, self‐report measures also provide useful information on the circumstances of non‐adherence. The latter is important for clinicians to understand the reasons why the client is non‐adherent (which may include the clinicians’ lack of behavioral skills) and to promote a collaborative relationship that considers clients' values and preferences. More research is therefore needed to both develop standardized and validated self‐report adherence measures and to identify more robust and objective measures of adherence to dietary advice.

Potential biases in the review process

Strengths of this review include the fact that we contacted many study authors during the data extraction process to gather additional information. The main reason was that some authors did not adequately describe the intervention provided in the intervention group and/or in the control/usual care group, in the published report. Additional information we received allowed us to better classify the included studies according to the intervention provided.

As expected, a limitation of this review is the definition of adherence to dietary advice. Adherence to dietary advice is a wide concept and includes many different measures including self‐reported measures which are not always comparable. Accordingly, in this review, some included studies assessed adherence to dietary advice by reporting the proportion of clients achieving the dietary recommendations. However, the majority of included studies evaluated adherence to dietary advice by comparing the mean dietary intake between groups. These different ways to measure adherence to dietary advice suggest that there is a need to develop standardized and validated tools to assess adherence to dietary advice.

In this review, we only included studies clearly mentioning a measure of adherence to dietary advice in the title or the objective of the study and/or those reporting the proportion of clients adhering to dietary advice. Therefore, we excluded all studies reporting mean dietary intake between groups without specifically assessing adherence to dietary advice as a primary outcome. Despite an extensive search in standard databases as well as in the grey literature, we cannot exclude the possibility that we missed some studies measuring adherence to dietary advice if those studies were not indexed in bibliographic databases as reporting adherence or compliance.

We categorized interventions according to Michie et al (Michie 2011) intervention functions to simplify and structure the presentation of results and not to provide insights about which intervention function was most effective for enhancing adherence to dietary advice. Although two review authors assigned the interventions to the categories through consensus, the assignment was arbitrary and we cannot exclude the fact that others may have assigned interventions to other categories. However, it must be emphasized that the process did not interfere with the interpretation of results.

Agreements and disagreements with other studies or reviews

Few systematic reviews evaluated clients' adherence to recommendations in the context of preventing and/or managing chronic diseases. Among systematic reviews reporting the effectiveness of interventions to enhance adherence to dietary advice, none assessed the same criteria as this review, making comparisons difficult. For example, two systematic reviews included other components in the assessment of adherence in addition to diet, such as physical activity and medication (Matteson 2010; Greaves 2011). The evaluation of diet adherence alone for those studies was therefore impossible. Fappa et al (Fappa 2008) performed a non‐systematic review on lifestyle interventions for enhancing adherence to diet and exercise in the management of the metabolic syndrome. However, dietary advice provided in the majority of included studies differed between the intervention and the control groups. Consequently, the effects of the intervention could not be isolated.

Burke 1997 conducted a non‐systematic review of successful strategies to increase adherence to dietary advice in the context of CVD prevention. Among eleven included studies, interventions found to be effective to improve adherence to nutritional therapy were behavioural skill training, spouse support and self‐efficacy enhancement.

Our results are consistent with those of Brownell and colleagues (Brownell 1995b) who performed an overview of studies with diet adherence data. They reported inconsistencies in methods and had difficulty interpreting results because of the broad variation of diseases covered and interventions provided. Similarly, Newell et al (Newell 2000) performed a non‐systematic review of strategies for improving cardiovascular client compliance to non‐pharmacologic treatments. No strong evidence was reported for the enhancement of dietary regime, and studies included were assessed as fair quality in term of study design. Those conclusions underline the fact that further good‐quality studies assessing adherence to dietary advice for preventing and managing chronic diseases should be performed.

Figure 1

Study flow diagram.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Figure 2

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Figure 3

Forest plot of comparison: 6 Nutritional tools versus control in diet adherence, outcome: 6.1 Continuous data. *Means represent the difference between pre‐and post‐ intervention.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Figure 4

Forest plot of comparison: 8 Multiple interventions versus control in diet adherence, outcome: 8.1 Continuous data. *Means represent the difference between pre‐and post‐ intervention.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Figure 5

Forest plot of comparison: 8 Multiple interventions versus control in diet adherence, outcome: 8.2 Dichotomous data. *Means represent the difference between pre‐and post‐ intervention.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 1.1

Comparison 1 Nutritional tools versus control in diet adherence, Outcome 1 Continuous data.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 2.1

Comparison 2 Multiple interventions versus control in diet adherence, Outcome 1 Continuous data.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 2.2

Comparison 2 Multiple interventions versus control in diet adherence, Outcome 2 Dichotomous data.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Table 1. Summary of results: education

Study	Intervention group (description)	Comparative group(s) (description)		Effects on adherence				No of studies (no of participants	Quality of the evidence (GRADE)
				Favours intervention group	Favours comparative group	No difference
Telephone follow‐up								4 (283)	⊝⊝⊝ Very low ^2,4
In summary, among studies using a control/usual care group, three out of ten diet adherence outcomes favoured the intervention group compared to control group and seven diet adherence outcomes had no significant difference between groups. However, these three diet adherence outcomes favouring the intervention group were no longer significant at a later time point.
Chiu 2010	Telephone follow‐up	Control				Adherence to sodium‐restricted diet, fat, fruit and vegetable intakes at 8 weeks
Cummings 1981	Telephone follow‐up	Control; Interventions: (1) contract; (2) contract with the involvement of a family member or friend		vs control: Adherence to potassium‐restricted diet and fluid‐restricted diet at 6 weeks		vs control: Adherence to potassium‐restricted diet and fluid‐restricted diet at 3 months; vs (1) (2): Adherence to potassium‐restricted diet and fluid‐restricted diet at 6 weeks and 3 months
Racelis 1998	Telephone follow‐up	Control				Adherence to diet
Stewart 2005	Telephone follow‐up	Control		Adherence to sodium‐restricted diet at 24 weeks		Adherence to sodium‐restricted diet at 36 weeks; Non‐adherence to alcohol intake at 24 and 36 weeks
Group sessions								2 (144)	⊝⊝ Low ^1,2
In summary, these studies did not allow us to draw conclusions on the effect of group sessions on diet adherence outcomes.
Gill 2010 *	Group sessions	Control
Jones 1986	Group sessions	Interventions: (1) Group sessions and teach to use prompts/cues; (2) Individual sessions with a dietitian; (3) Individual sessions with a dietitian and teach to use prompts/cues				vs (1) (2) (3): Adherence to diet at 16 weeks
Individual sessions with a dietitian								2 (203)	⊝⊝ Low ^1,2
In summary, these studies did not allow us to draw conclusions on the effect of individual sessions with a dietitian on diet adherence outcomes.
Jones 1986	Individual sessions with a dietitian	Interventions: (1) Group sessions and teach to use prompts/cues; (2) Individual sessions with a dietitian; (3) Individual sessions with a dietitian and teach to use prompts/cues				vs (1) (2) (3): Adherence to diet at 16 weeks
Micco 2007 *	Individual sessions with a dietitian	Control
Individual sessions with a nurse								1 (81)	⊝⊝⊝ Very low ^1,2,3
In summary, this study did not allow us to draw conclusions on the effect of individual sessions with a nurse on diet adherence outcomes.
Hsueh 2007	Individual sessions with a nurse		Intervention: telephone follow‐up and individual sessions with a nurse				Adherence to fiber, vegetable and fruit intakes at 3 and 6 months
Educational tools‐video								3 (318)	⊝ Moderate ¹
In summary, among studies using a control/usual care group, two out of three diet adherence outcomes favoured the intervention group compared to the control/usual care group and one diet adherence outcome had no significant difference between groups. However, one out of two diet adherence outcomes favouring the intervention group was no longer significant at a later time point.
Baraz 2010	Educational tools ‐ video	Intervention: group sessions and educational tools – booklet				Adherence to diet and fluid‐restricted diet at 2 months
Mahler 1999	Educational tools ‐ video	Control; Intervention: (1) video and relapse prevention/coping planning		vs control: Adherence to cholesterol and saturated fat‐restricted diet at 1 month		vs control: Adherence to cholesterol and saturated fat‐restricted diet at 3 months; vs (1): Adherence to cholesterol and saturated fat‐restricted diet at 1 and 3 months;
McCulloch 1983	Educational tools ‐ video	Usual care; Intervention: (1) nutritional tool		vs usual care: Adherence to day to day consistency in carbohydrate intake at 6 months		vs (1): Adherence to day to day consistency in carbohydrate intake at 6 months
Educational tools‐booklet								1 (83)	⊝⊝⊝ Very low^,2,3,4
In summary, this study did not allow us to draw conclusions on the effect of booklet on diet adherence outcomes.
Kendall 1987	Educational tools ‐ booklet	Intervention: nutritional tool				Adherence to energy, protein, vitamin A, vitamin C, thiamin, riboflavin, niacin, calcium, phosphorus, iron, zinc intakes at 3 and 6 months
The authors did not report measures of adherence for both groups, making comparison between groups impossible. GRADE ‐ Factors decreasing the quality level of a body of evidence:* ¹ Limitations in the design and implementation of available studies suggesting high likelihood of bias. ² Indirectness of evidence ³ Imprecision of results ⁴ Downgraded by two levels due to important limitations in the design and implementation of available studies suggesting high likelihood of bias.

Table 1. Summary of results: education

Ir a la tabla de la revisión

Table 2. Summary of results: persuasion

Study	Intervention group (description)	Comparative group(s) (description)	Effects on adherence			No of studies (no of participants)	Quality of the evidence (GRADE)
Study	Intervention group (description)	Comparative group(s) (description)	Favours intervention group	Favours comparative group	No difference	No of studies (no of participants)	Quality of the evidence (GRADE)
Reminders						2 (248)	⊝ Moderate ¹
In summary, among studies using a control/usual care group, 3 out of 19 diet adherence outcomes had no significant difference between groups. It was impossible to assess this result for 16 diet adherence outcomes since data and/or statistical analyses needed for comparison between groups were not provided.
Gans 1994	Reminder ‐ client	Usual care			Adherence to diet at 3 months
Gans 1994	Reminder ‐ physician	Usual care			Adherence to diet at 3 months
Gans 1994	Reminder ‐ client and physician	Usual care			Adherence to diet at 3 months
Ryan 2002 *	Reminder ‐ 2 weeks, 3 and 6 months	Control
Ryan 2002 *	Reminder ‐ 3 and 6 months	Control
The authors did not report measures of adherence for both groups, making comparison between groups impossible. GRADE ‐ Factors decreasing the quality level of a body of evidence:* ¹ Limitations in the design and implementation of available studies suggesting high likelihood of bias. ² Indirectness of evidence ³ Imprecision of results ⁴ Downgraded by two levels due to important limitations in the design and implementation of available studies suggesting high likelihood of bias.

Table 2. Summary of results: persuasion

Ir a la tabla de la revisión

Table 3. Summary of results: incentivisation

Study	Intervention group (description)	Comparative group(s) (description)	Effects on adherence			No of studies (no of participants)	Quality of the evidence (GRADE)
			Favours intervention group	Favours comparative group	No difference
Contracts with rewards						1 (116)	⊝⊝ Low ^1,3
In summary, among studies using a control/usual care group, two out of four diet adherence outcomes favoured the intervention group compared to the control group and two diet adherence outcomes had no significant difference between groups. However, these two diet adherence outcomes favouring the intervention group were no longer significant at three months.
Cummings 1981	Contract	Control; Interventions: (1) telephone follow‐up; (2) contract with the involvement of a family member or friend	vs control: Adherence to potassium‐restricted diet and fluid‐restricted diet at 6 weeks		vs control: Adherence to potassium‐restricted diet and fluid‐restricted diet at 3 months; vs (1) (2): Adherence to potassium‐restricted diet and fluid‐restricted diet at 6 weeks and 3 months
GRADE ‐ Factors decreasing the quality level of a body of evidence: ¹ Limitations in the design and implementation of available studies suggesting high likelihood of bias. ² Indirectness of evidence ³ Imprecision of results ⁴ Downgraded by two levels due to important limitations in the design and implementation of available studies suggesting high likelihood of bias.

Table 3. Summary of results: incentivisation

Ir a la tabla de la revisión

Table 4. Summary of results: training

Study	Intervention group (description)	Comparative group(s) (description)	Effects on adherence			No of studies (no of participants)	Quality of the evidence (GRADE)
Study	Intervention group (description)	Comparative group(s) (description)	Favours intervention group	Favours comparative group	No difference	No of studies (no of participants)	Quality of the evidence (GRADE)
Feedback						3 (661)	⊝⊝ Low ⁴
In summary, among studies using a control/usual care group, one out of seven diet adherence outcomes favoured the intervention group compared to the control/usual care group, four favoured the control group whereas two had no significant difference between groups.
Beasley 2008	Feedback	Control	Adherence to energy, fat, saturated fat and cholesterol intakes at 4 weeks
French 2008	Feedback – less intensive	Usual care		Adherence to general diet and specific diet at 12 months
French 2008	Feedback – most intensive	Usual care		Adherence to general diet and specific diet at 12 months
Meland 1994	Feedback	Control			Adherence to sodium‐restricted diet at 1 and 3 months
GRADE ‐ Factors decreasing the quality level of a body of evidence: ¹ Limitations in the design and implementation of available studies suggesting high likelihood of bias. ² Indirectness of evidence ³ Imprecision of results ⁴ Downgraded by two levels due to important limitations in the design and implementation of available studies suggesting high likelihood of bias.

Table 4. Summary of results: training

Ir a la tabla de la revisión

Table 5. Summary of results: restriction

Study	Intervention group (description)	Comparative group(s) (description)	Effects on adherence			No of studies (no of participants)	Quality of the evidence (GRADE)
			Favours intervention group	Favours comparative group	No difference
Restriction						1 (7)	⊝⊝⊝ Very low^1,2,3
In summary, this study did not allow us to draw conclusions on the effect of restriction on diet adherence outcomes.
Conrad 2000*	Restriction	Control			Adherence to very low fat diet at 7 months
The authors did not report measures of adherence for both groups, making comparison between groups impossible. GRADE ‐ Factors decreasing the quality level of a body of evidence:* ¹ Limitations in the design and implementation of available studies suggesting high likelihood of bias. ² Indirectness of evidence ³ Imprecision of results ⁴ Downgraded by two levels due to important limitations in the design and implementation of available studies suggesting high likelihood of bias.

Table 5. Summary of results: restriction

Ir a la tabla de la revisión

Table 6. Summary of results: modelling

Study	Intervention group (description)	Comparative group(s) (description)	Effects on adherence			No of studies (no of participants)	Quality of the evidence (GRADE)
Study	Intervention group (description)	Comparative group(s) (description)	Favours intervention group	Favours comparative group	No difference	No of studies (no of participants)	Quality of the evidence (GRADE)
Nutritional tools						7 (514)	⊝⊝⊝ Very low^2,4
In summary, among studies using a control/usual care group, 3 out of 17 diet adherence outcomes favoured the intervention group and 11 diet adherence outcomes had no significant difference between groups. It was impossible to assess this result for three diet adherence outcomes as data and/or statistical analyses needed for comparison between groups were not provided.
Assuncao 2010	Nutritional tools	Usual care	Adherence to sodium at 6 months		Adherence to energy, protein, fat, carbohydrate, cholesterol, fiber, fruit, vegetable and sweet food intakes at 6 months
Chen 2006	Nutritional tools	Control	Adherence to protein intake at 1 month
Grace 1996	Nutritional tools	Control	Adherence to fat intakes at 12 weeks		Adherence to energy at 12 weeks
Kendall 1987	Nutritional tools	Intervention: educational tool ‐ booklet			Adherence to energy, protein, vitamin and mineral intakes at 3 and 6 months
Logan 2010	Nutritional tools	Intervention: Barrier identification/problem solving and goal setting			Adherence to Mediterranean diet at 6 and 12 months
McCulloch 1983	Nutritional tools	Usual care; Intervention: (1) educational tool ‐ video			vs control and (1): Adherence to day to day consistency in carbohydrate intake at 6 months
Scisney‐Matlock 2006*	Nutritional tools	Control
The authors did not report measures of adherence for both groups, making comparison between groups impossible. GRADE ‐ Factors decreasing the quality level of a body of evidence:* ¹ Limitations in the design and implementation of available studies suggesting high likelihood of bias. ² Indirectness of evidence ³ Imprecision of results ⁴ Downgraded by two levels due to important limitations in the design and implementation of available studies suggesting high likelihood of bias.

Table 6. Summary of results: modelling

Ir a la tabla de la revisión

Table 7. Summary of results: enablement

Study	Intervention group (description)	Comparative group(s) (description)	Effects on adherence			No of studies (no of participants)	Quality of the evidence (GRADE)
Study	Intervention group (description)	Comparative group(s) (description)	Favours intervention group	Favours comparative group	No difference	No of studies (no of participants)	Quality of the evidence (GRADE)
Behaviour change techniques						3 (136)	⊝⊝⊝ Very low^2,4
In summary, only one study used a control group and three out of three diet adherence outcomes had no difference between groups.
Aldarondo 1999	Barrier identification/problem solving and self‐talk	Control			Adherence to energy, fat and saturated fat intakes at 14 weeks
Bennett 1986	Teach to use prompts/cues	Interventions: (1) self‐talk; (2) barrier identification/problem solving	vs (1) (2): Adherence to energy intake between baseline and 15 weeks
Bennett 1986	Self‐talk	Interventions: (1) teach to use prompts/cue; (2) barrier identification/problem solving		vs (1): Adherence to energy intake between baseline and 15 weeks	vs (2): Adherence to energy intake between baseline and 15 weeks
Bennett 1986	Barrier identification/problem solving	Interventions: (1) teach to use prompts/cue; (2) self‐talk		vs (1): Adherence to energy intake between baseline and 15 weeks	vs (2): Adherence to energy intake between baseline and 15 weeks
Logan 2010	Barrier identification/problem solving and goal setting	Intervention: Nutritional tools			Adherence to Mediterranean diet at 6 and 12 months
GRADE ‐ Factors decreasing the quality level of a body of evidence: ¹ Limitations in the design and implementation of available studies suggesting high likelihood of bias. ² Indirectness of evidence ³ Imprecision of results ⁴ Downgraded by two levels due to important limitations in the design and implementation of available studies suggesting high likelihood of bias.

Table 7. Summary of results: enablement

Ir a la tabla de la revisión

Table 8. Summary of results: multiple interventions

Study	Intervention group (description)	Comparative group(s) (description)	Effects on adherence			No of studies (no of participants)	Quality of the evidence (GRADE)
Study	Intervention group (description)	Comparative group(s) (description)	Favours intervention group	Favours comparative group	No difference	No of studies (no of participants)	Quality of the evidence (GRADE)
Multiple interventions						18 (7700)	**
In summary, among studies using a control/usual care group, 21 out of 56 diet adherence outcomes favoured the intervention group whereas 32 diet adherence outcomes had no significant difference between groups. It was impossible to assess this result for three diet adherence outcomes as data and/or statistical analyses needed for comparison between groups were not provided. However, 4 out of 21 diet adherence outcomes favouring the intervention group was no longer significant at a later time point.
Arcand 2005	Individual sessions with a dietitian and goal setting	Usual care			Adherence to sodium‐restricted diet at 3 months
Baraz 2010	Intervention: group sessions and educational tools – booklet	Intervention: educational tools ‐ video			Adherence to diet and fluid‐restricted diet at 2 months
Becker 1998	Telephone follow‐up and barrier identification/problem solving	Usual care			Adherence to fat‐restricted diet at 2 years
Blanson 2009	Motivational interviewing and self‐monitoring and feedback‐diary	Control			Adherence to diet at 28 days
Cummings 1981	Contract with the involvement of a family member or friend	Control; Interventions: (1) telephone follow‐up; (2) contract	vs control: Adherence to potassium‐restricted diet and fluid‐restricted diet at 6 weeks		vs control: Adherence to potassium‐restricted diet and fluid‐restricted diet at 3 months; vs (1) (2): Adherence to potassium‐restricted diet and fluid‐restricted diet at 3 months
Gucciardi 2007	Group sessions, nutritional tools and barrier identification/problem solving	Control	Adherence to diet at 3 months
Hsueh 2007	Telephone follow‐up and individual sessions with a nurse	Intervention: Individual sessions with a nurse			Adherence to fiber, vegetable and fruit intake at 3 and 6 months
Hyman 2007	Telephone follow‐up and motivational interviewing ‐ simultaneous	Usual care; Intervention: (1) telephone follow‐up and motivational interviewing ‐ sequential	vs usual care and (1): Adherence to sodium‐restricted diet at 6 months		vs usual care and (1): Adherence to sodium‐restricted diet at 18 months
Hyman 2007	Telephone follow‐up and motivational interviewing ‐ sequential	Usual care; Intervention: (1) telephone follow‐up and motivational interviewing ‐ simultaneous		vs (1): Adherence to sodium‐restricted diet at 6 months	vs control: Adherence to sodium‐restricted diet at 6 months and 18 months; vs (1): Adherence to sodium‐restricted diet at 18 months
Jiang 2004	Individual sessions with a nurse, telephone follow‐up and goal setting	Usual care	Adherence to ATP step II diet at 3 and 6 months.
Jones 1986	Group sessions and teach to use prompts/cues	Interventions: (1) group sessions; (2) individual sessions with a dietitian; (3) individual sessions with a dietitian and teach to use prompts/cues			vs (1) (2) (3): adherence to diet at 16 weeks
Jones 1986	Individual sessions with a dietitian and teach to use prompts/cues	Interventions: (1) group sessions; (2) individual sessions with a dietitian; (3) group sessions and teach to use prompts/cues			vs (1) (2) (3): adherence to diet at 16 weeks
Laitinen 1993	Individual sessions with a dietitian, nutritional tools and goal setting	Usual care	Adherence to saturated fat intakes at 15 months		Adherence to saturated fat intakes at 3 months; Adherence to total fat, unsaturated fat, carbohydrate, fiber and cholesterol intakes at 3 and 15 months
Mahler 1999	Educational tools – video and relapse prevention/coping planning	Control; Intervention: (1) educational tools ‐ video	vs control: Adherence to cholesterol and saturated fat‐restricted diet at 1 month.		vs control: Adherence to cholesterol and saturated fat‐restricted diet at 3 months; vs (1): Adherence to cholesterol and saturated fat ‐restricted diet at 1 and 3 months
Miller 1988	Individual sessions with a nurse and barrier identification/problem solving and goal setting	Control	Adherence to diet at 2 years		Adherence to diet at 30 days, 60 days and 1 year
Morey 2008	Individual sessions with a dietitian, educational tools‐booklet, reminders, motivational interviewing	Control	Adherence to phosphate‐restricted diet at 3 months
Tsay 2003	Self‐monitoring and feedback‐diary, stress management and goal setting	Usual care	Adherence to fluid‐restricted diet at 3 months and 6 months		Adherence to fluid‐restricted diet at 1 month
Wong 2010	Telephone follow‐up and goal setting	Control			Non‐adherence to diet (days and degree) at 7 weeks and 13 weeks Non‐adherence to fluid‐restricted diet (days and degree) at 7 weeks and 13 weeks
Wood 2008 – coronary heart disease	Individual sessions with a nurse and motivational interviewing	Usual care	Adherence to saturated fat, oily fish, fish and fruit and vegetable intakes at 1 year
Wood 2008 – high risk of coronary heart disease	Individual sessions with a nurse and motivational interviewing	Usual care	Adherence to oily fish, fish and fruit and vegetables intakes at 1 year
Zhao 2004	Telephone follow‐up, individual sessions with a dietitian and goal setting	Usual care	High adherence to diet at 4 and 12 weeks
**Multiple interventions included a variety of interventions, which did not allow the use of GRADE.

Table 8. Summary of results: multiple interventions

Ir a la tabla de la revisión

Comparison 1. Nutritional tools versus control in diet adherence

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Continuous data Show forest plot	2		Std. Mean Difference (IV, Random, 95% CI)	Totals not selected

1.1 Adherence to energy intake at 6 months	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.2 Adherence to protein intake at 6 months	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.3 Adherence to fat intake at 6 months	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.4 Adherence to carbohydrate intake at 6 months	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.5 Adherence to cholesterol intake at 6 months	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.6 Adherence to fiber intake at 6 months	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.7 Adherence to sodium intake at 6 months	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.8 Adherence to fruit intake at 6 months	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.9 Adherence to vegetable intake at 6 months	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.10 Adherence to sweet food intake at 6 months	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.11 Adherence to energy intake at 12 weeks	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.12 Adherence to fat intake at 12 weeks	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]

Comparison 1. Nutritional tools versus control in diet adherence

Ir a la tabla de la revisión

Comparison 2. Multiple interventions versus control in diet adherence

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Continuous data Show forest plot	4		Std. Mean Difference (IV, Random, 95% CI)	Totals not selected

1.1 Adherence to sodium‐restricted diet at 3 months	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.2 Adherence to diet at 3 months	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.3 Adherence to fluid‐restricted diet at 1 month	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.4 Adherence to fluid‐restricted diet at 3 months	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.5 Adherence to fluid‐restricted diet at 6 months	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.6 Non‐adherence to diet (days) at 7 weeks	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.7 Non‐adherence to diet (days) at 13 weeks	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.8 Non‐adherence to diet (degree) at 7 weeks	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.9 Non‐adherence to diet (degree) at 13 weeks	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.10 Non‐adherence to fluid‐restricted diet (days) at 7 weeks	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.11 Non‐adherence to fluid‐restricted diet (days) at 13 weeks	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.12 Non‐adherence to fluid‐restricted diet (degree) at 7 weeks	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.13 Non‐adherence to fluid‐restricted diet (degree) at 13 weeks	1		Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
2 Dichotomous data Show forest plot	5		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

2.1 Adherence to sodium‐restricted diet at 18 months	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.2 Adherence to fat intake at 3 months	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.3 Adherence to saturated fat intake at 3 months	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.4 Adherence to unsaturated fat intake at 3 months	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.5 Adherence to carbohydrate intake at 3 months	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.6 Adherence cholesterol intake at 3 months	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.7 Adherence to saturated fat intake at 15 months	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.8 Adherence to fat intake at 15 months	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.9 Adherence to unsaturated fat intake at 15 months	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.10 Adherence to carbohydrate intake at 15 months	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.11 Adherence to fiber intake at 15 months	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.12 Adherence to cholesterol intake at 15 months	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.13 Adherence to phosphate‐restricted diet at 3 months	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.14 Adherence to saturated fat intake at 1 year ‐ CHD patients	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.15 Adherence to oily fish intake at 1 year ‐ CHD patients	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.16 Adherence to fish intake at 1 year ‐ CHD patients	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.17 Adherence to fruit and vegetable intake at 1 year ‐ CHD patients	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.18 Adherence to oily fish intake at 1 year ‐ high risk CHD patients	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.19 Adherence to fish intake at 1 year ‐ high‐risk CHD patients	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.20 Adherence to fruit and vegetable intake at 1 year ‐ high‐risk CHD patients	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.21 Adherence to diet at 4 weeks	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
2.22 Adherence to diet at 12 weeks	1		Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]

Comparison 2. Multiple interventions versus control in diet adherence

Ir a la tabla de la revisión

	Idioma de la Revisión Cochrane Escoja su idioma de preferencia para las revisiones Cochrane y otros contenidos. Las secciones sin una traducción aparecerán en inglés.

	Idioma de la web Escoja su idioma de preferencia para la web de la Biblioteca Cochrane.

Idioma de la Revisión Cochrane

Idioma de la web

Resumen

Antecedentes

Objetivos

Métodos de búsqueda

Criterios de selección

Obtención y análisis de los datos

Resultados principales

Conclusiones de los autores

Resumen en términos sencillos

Intervenciones para mejorar el cumplimiento del asesoramiento dietético para la prevención y el control de las enfermedades crónicas en adultos

Resumen visual

Authors' conclusions

Implications for practice

Implications for research

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Assessment of heterogeneity

Consumer participation

Results

Description of studies

Results of the search

Included studies

Location, setting and duration

Clients

Prevention of chronic diseases

Management of chronic diseases

Prevention and management of chronic diseases

Interventions

Education

Persuasion

Incentivisation

Training

Restriction

Modelling

Enablement

Multiple

Outcomes

Excluded studies

Risk of bias in included studies

Allocation

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

Education

Counselling and follow‐up with health professional

Telephone follow‐up

Group sessions

Individual sessions with a dietitian

Individual sessions with a nurse

Educational tools

Video

Booklet

Persuasion

Reminders

Incentivisation

Contracts with rewards