Weight-related concerns, including eating disorders, disordered eating, and overweight and obesity, are prevalent in the adult population of the USA( Reference Flegal, Carroll and Ogden 1 ). The traditional approach to weight management has been characterized by restriction of energy, specific nutrients or food groups in order to induce weight loss, and generally results in little long-term success. Studies have shown that few participants maintained any weight loss and many participants gained back more weight than was lost during the dieting period( Reference Goodrick and Foreyt 2 – Reference Wadden, Sternberg and Letizia 6 ). Due to these negative outcomes, professionals have expressed ethical concern with the recommendation of restrictive practices for weight loss( Reference Mann, Tomiyama and Westling 4 , Reference Bacon and Aphramor 7 ) with some calling for a paradigm shift in the weight-management field.
One alternative, emerging approach is intuitive eating. Intuitive eating encourages individuals to focus on improving health rather than losing weight. A main focus is on use of internal cues of hunger and fullness to guide eating, with emphasis on acceptance of the body regardless of size( Reference Tribole and Resch 8 – Reference Schaefer and Magnuson 12 ). The approach was developed and endorsed by two registered dietitians/nutritionists (RD/N), Evelyn Tribole and Elyse Resch, in the late 1990s. They observed that overweight clients would lose weight by following a calorie-restricted diet but almost inevitably gain the weight back over time( Reference Tribole and Resch 8 ). The weight gain was often accompanied by psychological distress and feelings of guilt and failure. Through experience, Tribole and Resch found that clients who were able to adopt an intuitive eating lifestyle were able to develop a healthier psychological relationship with food, accept and respect their bodies, and stop the endless cycle of dieting and weight cycling. Over the past two decades, several research studies have investigated this lifestyle.
The intuitive eating lifestyle is associated with lower BMI( Reference Tylka 9 , Reference Hawks, Madanat and Hawks 13 – Reference Tylka and Kroon Van Diest 15 ) and greater psychological well-being( Reference Tylka 9 , Reference Tylka and Kroon Van Diest 15 ), and inversely associated with eating disorder symptomatology( Reference Tylka 9 , Reference Tylka and Kroon Van Diest 15 ). Participants in intuitive eating interventions have generally lost( Reference Ciampolini, Lovell-Smith and Sifone 16 – Reference Roughan, Seddon and Vernon-Roberts 23 ) or maintained( Reference Bacon, Keim and Van Loan 24 – Reference Provencher, Begin and Tremblay 31 ) body weight, improved cardiovascular risk irrespective of weight loss( Reference Bacon, Stern and Van Loan 25 , Reference Carroll, Borkoles and Polman 32 ) and increased body satisfaction( Reference Gagnon-Girouard, Begin and Provencher 18 , Reference Roughan, Seddon and Vernon-Roberts 23 , Reference Bacon, Keim and Van Loan 24 , Reference Ciliska 27 , Reference Higgins and Gray 33 – Reference Tanco, Linden and Earle 35 ). This evidence has been coupled with a divide in the literature on the best approach to weight management( Reference Neumark-Sztainer 36 ) and while intuitive eating research has been positive, additional studies are indicated( Reference Van Dyke and Drinkwater 10 , Reference Schaefer and Magnuson 12 ).
Health professionals have also been encouraging mindful eating, an approach similar to intuitive eating. Mindful eating has been described as the non-judgemental awareness of the physical and emotional sensations associated with eating or environment( Reference Framson, Kristal and Schenk 37 ). Mindful eating encourages individuals to eat according to internal cues of hunger and satiety( Reference Framson, Kristal and Schenk 37 ) and to eat slowly, taste every bite thoroughly and eat without distraction( Reference Singh, Lancioni and Singh 38 ). The originators of intuitive eating have acknowledged that mindful eating is part of intuitive eating( Reference Tribole and Resch 8 ). Regardless of the terminology and similarities or differences between the two approaches, evidence has suggested that many dietitians may be moving away from the weight-loss paradigm and towards a size acceptance, intuitive eating approach( Reference Chapman, Sellaeg and Levy-Milne 39 , Reference Barr, Yarker and Levy-Milne 40 ).
In an attempt to characterize weight-management practices that Australian dietitians use with clients, Campbell and Crawford( Reference Campbell and Crawford 41 ) developed a questionnaire by compiling a list of practices from the literature that was then reviewed by dietitians with expertise in weight management. In this process, dietitians were asked how frequently they performed each weight-management strategy with their clients (e.g. specific advice to reduce total fat intake)( Reference Campbell and Crawford 41 ). Barr and colleagues( Reference Barr, Yarker and Levy-Milne 40 ) revised this questionnaire to include size acceptance philosophies (e.g. increasing self-acceptance of current weight) in their study to examine Canadian dietitians’ weight-management attitudes and practices. While they have provided valuable insight into the attitudes and practices of dietitians, neither reliability nor validity of these questionnaires was established( Reference Barr, Yarker and Levy-Milne 40 ).
While two valid and reliable measures are available to capture individuals’ intuitive eating behaviour, the more frequently used Tylka and Kroon van Diest’s Intuitive Eating Scale-2( Reference Tylka 9 , Reference Tylka and Kroon Van Diest 15 ) and the less common, Hawks and colleagues’ Intuitive Eating Scale( Reference Hawks, Merrill and Madanat 42 ), no such measure has been validated to gauge health professionals’ knowledge, attitudes and practices regarding this approach. To date, only qualitative research has examined dietitians’ knowledge of non-dieting and size acceptance approaches( Reference Marchessault, Thiele and Armit 43 ). As support for intuitive eating has grown, knowledge, attitudes and intuitive eating practices of RD/N have remained unknown. Surveys are often conducted to understand human behaviour; measuring the knowledge, attitudes and practices can increase insight into a given situation( Reference Vandamme 44 ). Therefore, the purpose of the present study was to develop and validate a tool to measure knowledge, attitudes and practices of RD/N regarding an intuitive eating lifestyle. Based on existing evidence, we hypothesized that this survey will reveal three valid and reliable constructs: (i) knowledge of intuitive eating; (ii) attitudes towards intuitive eating; and (iii) use of intuitive eating practices.
Methods
Participants
Contact information for all RD/N in the USA (n 88 784) was obtained from the Commission on Dietetic Registration. From this list, a 10 % random sample (n 8834) was selected. The survey was distributed by email with a survey link to 8549 individuals for the initial validity testing (excluded from the 10 % sample were 285 RD/N who did not provide an email address or provided an invalid email address). The survey was open for two months, June through July 2014. During this time, 1897 RD/N completed the survey for a response rate of 22·2 %. After initial validation analysis, the survey was distributed to the remaining 90 % of RD/N (n 79 950) to confirm validity and reliability. Those in the 90 % sample who indicated they worked in weight management were included in the present analysis study (n 9249). This provided a response rate of 11·6 % for the validity and reliability testing. There was no incentive given for completing the survey.
Procedures
The development of the survey instrument underwent several phases. Phase I entailed the development of the original instrument by the lead investigator who is an RD/N with training in survey methodology. Existing scales were identified and adapted and original questions were written to assess the knowledge, attitudes and practices of RD/N relevant to intuitive eating (see ‘Survey measures’ for detailed description of the items). In Phase II, two nutrition professionals, who had training in the intuitive eating approach, were consulted to ensure content validity. Based on feedback, three negatively worded practice items were reworded positively to enhance clarity; one item (recommend using a food journal/diary to monitor calories, portions, etc.) was divided into two items to reflect two practices (recommend using a food journal/diary to monitor exact calories, portions, etc. and recommend using a food journal/diary to monitor general calories, portions, etc.); and ‘don’t know’ options were added to each of the three main sections. One of the RD/N consulted also noted that some RD/N may be familiar with the term ‘mindful eating’, but not ‘intuitive eating’, specifically. Thus, the question ‘Have you ever heard of intuitive eating?’ was changed to ‘Have you ever heard of intuitive or mindful eating?’ for those who may use these terms interchangeably; the remainder of the survey referred only to intuitive eating to address the study’s purpose of developing and validating a tool to measure RD/N knowledge, attitudes and practices regarding an intuitive eating lifestyle. After these edits, an online version of the survey was created.
In Phase III, the instrument was sent to dietetic interns at Kent State University (n 13) to pilot-test the survey. Most participants were female (n 11), between 23 and 26 years of age, and all but one were non-Hispanic Caucasian. The interns completed the survey in its near final form. At the end of each section the interns were asked to provide feedback about the section they just completed and specifically to state if any items were unclear or if there were any errors. The interns indicated that all the items were clear. There was one spelling error and one error in the layout of the online survey. After these corrections, the development of the instrument was complete.
All RD/N in the sample described above were sent an email message requesting their participation in the survey. The email asked the RD/N to follow a link to the survey website where they were first prompted to read and agree to informed consent. The survey was open for two months during which the RD/N received the original email and two reminders to complete the survey. All procedures were reviewed and approved by the Kent State University Institutional Review Board.
Survey measures
Descriptive characteristics
Participants were asked to report their gender, age, race, highest level of education, main practice setting, state of practice, and if they had completed a certificate in paediatric or adult weight management. Participants were also asked if they had ever heard of intuitive or mindful eating and if they currently counsel overweight and/or obese clients for weight management; if they did, the number of years’ experience in this practice was requested.
Practices
Participants completed this section if they reported that they do currently counsel overweight and/or obese clients for weight management. Participants were asked to report on a Likert scale (0=‘never’, 1=‘rarely’, 2=‘sometimes’, 3=‘often’, 4=‘usually’) how often they use various practices when counselling overweight and/or obese clients. This section of the survey was adapted from a tool used by Barr and colleagues( Reference Barr, Yarker and Levy-Milne 40 ) to describe how often Canadian dietitians utilize several specific practices with their weight-management clients.
Knowledge
All participants completed this section. The first ten questions were adapted from the Intuitive Eating Scale-2( Reference Tylka and Kroon Van Diest 15 ) and described behaviours that are and are not consistent with the intuitive eating lifestyle. Questions were chosen to represent the four factors of intuitive eating (unconditional permission to eat, eating based on internal cues, eating for physical rather than emotional reasons, body–food congruence) and were reworded to reflect general knowledge about the intuitive eating lifestyle as opposed to personal behaviour. Additionally, four questions were developed to assess knowledge on current research regarding intuitive eating. These questions were developed based on findings from a recent review paper( Reference Schaefer and Magnuson 12 ). Participants were asked to report if each statement was characteristic of an intuitive eater or if they did not know.
Attitudes
All participants completed this section. This section gauged the attitudes of RD/N towards various health behaviours and health attitudes. These items were developed by the researcher. Items assessed attitudes towards key aspects of intuitive eating (e.g. ‘It is important for individuals to learn to eat based on internal cues of hunger, fullness and satisfaction’) and towards the traditional weight-loss approach (e.g. ‘Weight loss should be the primary focus to improve health in overweight and/or obese individuals’). The items favourable towards intuitive eating reflected the four factors of intuitive eating. Participants were asked to rate the degree to which they agreed or disagreed with each statement on a Likert scale (1=‘strongly disagree’, 2=‘disagree’, 3=‘neutral’, 4=‘agree’, 5=‘strongly agree’, and a ‘don’t know’ option). Nine items were consistent with and four items were inconsistent with the intuitive eating lifestyle.
Data analysis
Since only RD/N who worked in weight-management counselling completed the whole survey, data from this sample were used to conduct initial validity analysis. Construct validity was examined first with exploratory factor analysis (EFA) using principal axis factoring to extract factors by estimating the shared variance between items and oblique rotation of factors with promax rotation in order to allow the factors to be correlated. The correlation matrix was explored to ensure there was no singularity or multicollinearity. The number of factors was determined by examining eigenvalues and the scree plot. Items with a factor loading of at least 0·35 and a cross-loading difference of at least 0·2 were retained to create the final instrument. This analysis was repeated using data from RD/N who did not work in weight management to ensure validity of the knowledge and attitudes factors of all RD/N.
Construct validity was further assessed through confirmatory factor analysis (CFA) with maximum likelihood estimation. Hu and Bentler( Reference Hu and Bentler 45 ) have recommended several two-index strategies to assess fit, including one recommended combination of the standardized root-mean-square residual (SRMR) with recommended value ≤0·08 and the root-mean-square error of approximation (RMSEA) with recommended value ≤0·06. Factor loadings were examined to ensure each had a loading of at least 0·35. Cronbach’s α was reported to assess reliability.
Results
Participants
With regard to the initial survey distributed to the random 10 % of RD/N, most participants were female (96·8 %), non-Hispanic (96·2 %) and Caucasian (91·0 %; Table 1). Nearly 44 % had a Bachelor’s degree while an additional 50·2 % had completed a Master’s degree. Most worked in a clinical setting (40·9 %) while others worked in the community (15·3 %), research (6·0 %), private practice (7·8 %) or other settings (19·7 %). Several were not practising in a dietetics-related field (10·2 %). Roughly half of the respondents reported that they work in the weight management field (53·7 %). With regard to the second distribution to the remaining 90 % of RD/N, demographics were similar. Most participants were female (97·0 %), non-Hispanic (96·4 %) and Caucasian (91·9 %). About half of the RD/N had at least a Master’s degree (49·2 %). Less than half of all respondents reported they worked in a clinical setting while about half (50·3 %) reported that they work in weight management.
Table 1 Sample characteristics of registered dietitians/nutritionists who completed the survey for exploratory factor analysis (n 1895)
Validity and reliability
The EFA with principal axis factoring and oblique rotation was conducted to assess construct validity. The correlation matrix was examined to ensure there was no singularity or multicollinearity. The sample size met the recommended 20:1 sample size to parameters ratio( Reference Kline 46 ). The overall KMO (Kaiser–Meyer–Olkin) was 0·88; values ranged from 0·67 to 0·96, which indicated acceptable sampling adequacy. The communality values were assessed to assure there was shared variance between the items. The number of factors was determined by examining eigenvalues, the scree plot and the factor solution. There were five eigenvalues greater than 1. The scree plot inflection was between four and five factors (Fig. 1). The five-factor solution was examined first. Few items loaded on the fifth factor and of those that did, several cross-loaded with another factor. Thus, the four-factor solution was examined. The overall KMO value (0·88) and the communality values were still adequate.
Fig. 1 Scree plot of eigenvalues obtained from exploratory factor analysis among dietitians/nutritionists (n 1895). The numbers on the figure represent the number of factors proposed. The inflection point of the graph is between four and five, suggesting that either a four- or five-factor solution would be recommended based on the results of the scree plot
In this four-factor solution, all of the knowledge items loaded strongly with each other and were retained. Five attitude items (items 1, 6, 8, 9 and 11) that did not load on any factor (factor loading <0·35) and three attitude items (items 2, 5 and 7) that loaded with the practice items were removed. After examination of these three items, it was apparent that the wording of these items assessed preference of a particular practice (i.e. ‘To lose weight, overweight and/or obese individuals should consciously restrict calories, fat and/or carbohydrates’); thus, these items were assessed in the practices section.
The practice items loaded on to two distinct factors. Six practice items (items 4, 11, 12, 23, 24 and 25) were removed that did not load at a value of at least 0·35 on any factor. When the EFA was conducted again without these items, only two items were problematic. One practice item (item 10) did not load on the factor (factor loading <0·35) and one practice item (item 16) was cross-loading with two factors. Seven items loaded on one factor that included traditional and restrictive weight-management practices that recommend limiting calories, nutrients or eating in general, or monitoring intake and/or weight. The ten remaining practices included strategies that did not directly imply restriction (i.e. ‘Work with clients using behaviour modification techniques’) and strategies that promoted intuitive eating (i.e. ‘Recommend keeping a hunger awareness journal/diary’). The practice items that were removed were either not specifically about eating (e.g. ‘How often do you give general advice about exercise?’) or were not related to one of the two approaches (non-restrictive/intuitive eating or restrictive/traditional); for example, ‘How often do you recommend herbs or botanicals for weight loss?’ or ‘How often do you recommend a commercial or community-based weight-loss programme?’ These items could have been consistent with a non-restrictive/intuitive eating approach, a traditional/restrictive approach, or not consistent with either depending how these strategies were implemented.
After removal of these two items, the results indicated four distinct factors with strong factor loadings (≥0·35; Table 2) and no cross-loading (difference >0·20). The first factor consisted of fourteen items that represent knowledge of intuitive eating. The second factor consisted of seven items that represent attitudes towards intuitive eating. The third factor consisted of ten items that represent practices consistent with a non-restrictive, intuitive eating approach, while the fourth factor consisted of seven items that represented practices consistent with a restrictive, traditional approach to weight management.
Table 2 Exploratory factor analysis factor loadings among registered dietitians/nutritionists who work in weight management (n 1018)
Items with factor loadings ≥0·35 are shown.
To further explore the validation of factors, the EFA was re-run in the complete sample (all RD/N, not just those who work in weight management), without the practices section of the survey, to ensure that the knowledge and attitudes factors were valid in all RD/N, not just those who work in weight management. The factor structure for knowledge and attitudes was upheld.
Next, CFA was conducted to ensure validity of the factors. Data consisted of the responses from the 9249 RD/N who completed the instrument distributed after the EFA was complete. The CFA model was specified with four factors. The RMSEA value was 0·07, close to the recommended value around 0·06 and less than the critical value of 0·10 that would have suggested poor fit. The SRMSR value was 0·07, indicating acceptable fit. All items loaded on their respective factor with a factor loading of at least 0·35 except for two attitude items: ‘How strongly do you support the use of intuitive eating to promote a healthy lifestyle?’ and ‘Intuitive eating is more effective than calorie-restricted dieting for long-term weight loss and/or maintenance’. With regard to the former item, it is possible that support does not necessarily align with attitude; the latter item is a matter of evidence rather than attitude. After these two items were removed, all items loaded on their respective factors with a loading for at least 0·35 (Table 3). The correlations between constructs were low (Table 3), indicating little overlap between factors.
Table 3 Results of confirmatory factor analysis factor loadings, correlation between factors and reliability coefficients among registered dietitians/nutritionists who completed the instrument distributed after the exploratory factor analysis was complete (n 9249)
* Reliability values presented are Cronbach’s α values for registered dietitians/nutritionists who work in weight management.
Finally, Cronbach’s α was calculated to assess reliability of each factor (Table 3). The Cronbach’s α value for the traditional/restrictive practices factor was 0·74. The Cronbach’s α value for the non-restrictive/intuitive eating practices factor was 0·84. The Cronbach’s α value for the knowledge factor was 0·88 for those who work in weight management and 0·91 for the complete sample. The Cronbach’s α value for the attitudes factor was 0·75 for those who work in weight management and 0·79 for the complete sample. While reliability was on the lower end for the traditional/restrictive practices and attitudes factors, all values indicated adequate internal reliability of the factors( Reference Nunnally and Bernstein 47 ).
Discussion
The results indicated that, contrary to the hypothesized three factors (knowledge, attitudes and practices), the proposed instrument actually measured four distinct factors: (i) knowledge of intuitive eating; (ii) attitudes towards intuitive eating; (iii) use of restrictive and traditional weight-management practices; and (iv) use of non-restrictive and intuitive eating practices. All fourteen proposed knowledge items loaded strongly together. This factor measured RD/N knowledge of intuitive eating and the research regarding intuitive eating. These items were expected to load strongly together since most were adapted from a validated measure of intuitive eating behaviour( Reference Tylka 9 , Reference Tylka and Kroon Van Diest 15 ).
Originally, the authors expected the two practice factors to load on one factor in opposite directions. However, the items represented two distinct factors. One potential explanation is that RD/N do not distinctly use one approach or the other, which would have caused the two groups of questions to load on one factor in opposite directions. RD/N could use both approaches depending on the client. Further investigation into RD/N practices would be needed to investigate this finding. These two factors could be used to gauge how frequently RD/N who work in weight management use practices from the traditional weight-loss paradigm as well as practices that are non-restrictive and consistent with the intuitive eating approach. Measuring the two factors separately could prove to be more valuable in studying weight-management practices among health professionals than clustering all practices together.
According to the Academy of Nutrition and Dietetics, RD/N are nutrition experts who translate the science of nutrition into practical solutions to help individuals make positive lifestyle changes. While support for the intuitive eating approach has grown, our understanding of RD/N knowledge and use of this approach has not been explored until now. The present study is the first to validate a measure to assess the concept of intuitive eating. Given the current divide in the weight-management philosophies (traditional v. intuitive eating), it has also been unknown how favourably RD/N view the intuitive eating lifestyle. With the validation of this survey, these gaps in the literature can now be examined.
Qualitative evidence has demonstrated that while some RD/N maintain a focus on weight loss, many have moved towards the new paradigm that promotes concepts consistent with intuitive eating( Reference Chapman, Sellaeg and Levy-Milne 39 , Reference Barr, Yarker and Levy-Milne 40 ). Other studies have assessed dietitians’ attitudes towards overweight and obesity( Reference Chapman, Sellaeg and Levy-Milne 39 – Reference Campbell and Crawford 41 , Reference Harvey, Summerbell and Kirk 48 – Reference Oberrieder, Walker and Monroe 50 ). The present research is the first to develop a validated measure of attitudes towards intuitive eating. Similarly, researchers have investigated dietitians’ use of different weight-management practices with clients using qualitative( Reference Chapman, Sellaeg and Levy-Milne 39 ) and quantitative( Reference Barr, Yarker and Levy-Milne 40 , Reference Campbell and Crawford 41 , Reference Harvey, Summerbell and Kirk 48 ) methods. These previous studies have been limited in that reliability and validity were not established( Reference Barr, Yarker and Levy-Milne 40 ).
One strength of the present study is the large sample size as this is required for accuracy in EFA( Reference Kline 46 ). The provision of contact information by the Commission on Dietetic Registration enabled the researchers to collect adequate data to be able to conduct this analysis. One limitation of the study is the potential for selection bias. Participants self-selected into the study. Those who chose to participate may differ from those who chose not to participate. Since there are no population statistics on RD/N in the USA, the degree of potential selection bias was unknown.
There are several important next steps that should follow the present study. Future research should measure convergent and discriminant validity to ensure validity of this measure. In addition, as eating- and weight-related issues continue to challenge health professionals and individuals, and as research continues to grow in favour of the intuitive eating approach, future research could use this tool to assess the knowledge, attitudes and practices regarding intuitive eating in RD/N and other health professionals both in the USA and globally. Conducting this survey with RD/N and other health professionals could lend insight into the current state of practice in the weight-management field. A link between research and practice is important to advance this challenging field. Further, evidence suggests that nutrition practitioners should use theoretical frameworks to enhance the effectiveness of programmes designed to address weight concerns( Reference Sharma 51 , Reference Spahn, Reeves and Keim 52 ). This survey could be expanded to assess RD/N application of theoretical constructs to both non-restrictive/intuitive eating and restrictive/traditional weight-management practices with clients.
Conclusion
In conclusion, the present study developed and validated an instrument to measure RD/N knowledge of and attitudes towards intuitive eating, as well as use of traditional/restrictive and non-restrictive/intuitive eating practices. As it has become apparent that the traditional, restrictive approach to promote weight loss is ineffective and as the support for an intuitive eating approach has grown, researchers have begun to discuss ethical issues associated with continuing to promote the use of traditional, restrictive practices for weight management( Reference Mann, Tomiyama and Westling 4 , Reference Bacon and Aphramor 7 , Reference Aphramor 11 ). In particular, Aphramor( Reference Aphramor 11 ) has asserted that the ineffectiveness of the traditional energy-deficit approach to weight management has not only failed to meet standards of evidence-based practice, but has also failed to ignite a conversation about the ethical implications of continuing to use these practices and yet it continues to dominate research in the field. The tool developed and validated in the present study could help spark such a debate, by examining the current state of practice, in the hope of moving the field forward.
Acknowledgements
Acknowledgements: The authors would like to thank Amy Magnuson, PhD, RD, LD/N and Natalie Caine-Bish, PhD, RD, LD/N for reviewing the initial survey and providing feedback. Their time and efforts are greatly appreciated. Financial support: This research received no specific grant from any funding agency in the public, commercial or not-for profit sectors. Conflict of interest: None. Authorship: J.T.S. developed and distributed the survey, conducted all data analyses, and drafted the main sections of the manuscript. M.D.Z. assisted in the conception and design of the study, supervised data collection and analysis, and contributed to and edited the introduction, discussion and conclusions. Both authors have read and approved the manuscript. Ethics of human subject participation: This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects/patients were approved by the Institutional Review Board at Kent State University. Written informed consent was obtained from all subjects/patients.
