Knowledge, Attitudes and Practices on Fad Diets of a Panel of Spanish Consumers: Protocol of a Cross-sectional Study

Sample size: The sample size is estimated at 2 600 for a confidence level of 95%, a power of 80% and taking into account a response rate of 60-70% and a percentage of losses (people who accept the invitation but do not complete the survey) of 10%.

Aditional processess: the preselected subjects will be sent an invitation to participate in the study and will require a web record where basic sociodemographic data will be collected. The subjects will receive an incentive to join the study that will consist of points redeemable for gifts, the number of points received being proportional to the length and complexity of the final survey. The initial invitation does not contain information on the type of survey or incentive, but the estimated time for completion. During the recruitment process, basic socio-demographic data will be reviewed to ensure that stratification criteria are met and duplicate cases will be eliminated.

Intruments (KAP survey): survey of Knowledge, Attitudes and Practices , built ad hoc, not validated, online and self-administered. The survey will include the 3 domains that should be evaluated: knowledge, attitudes and practices, and will have a total of no more than 18 questions, according to recommendations established in literature. The recommended stages will be followed for the creation of the KAP survey:

• definition of the objective, target population and survey modules
• determine the topics to be covered that will be translated into necessary modules
• set questions that cover the topics
• select and prioritize the questions that will finally be part. To do so, each question will be categorized into: (a) essential questions; (b) optional questions; (c) specific questions, must have a specific objective assigned and explain the importance of it.
• adapt the questions to be a self-administered online questionnaire
• There will be a pilot period for both the content and its online functional viability of the survey.

To assess knowledge, categorized questions will be created, which will consist of a list of response options considered correct and not correct, which will include the answers "others" and "I don't know". The evaluators will determine whether the topic is "Know" or "Don't Know", as well as the number of correct/incorrect answers.

To assess attitudes, 3-point likert scales will be created in which subjects will be asked to judge to what degree they are positively or negatively inclined towards: a problem related to fad diets, a practice related to fad diets, and follow a dietary recommendation, whatever. Faces will be used to display all three categories. It is recommended to create the questions of this attitude module following the Health Belief Model.

To assess practices, questions will be asked about whether or not some behaviors have been carried out for long periods of time. Yes / no questions will be asked on a general practices + and "no answer", and if so, it will be deepened with categorized questions with mutiple choice + "other". Participants will be asked about their practices in the last year and its typology, as well as those followed so far.

Other instruments: in addition to the KAP survey, an informed consent and a questionnaire will be passed to obtain basic sociodemographic data according to recommendations in this type of study.

Potential biases: it is assumend that the results will be affected by the typical biases of similar cross-sectional studies:

• selection bias: (a) volunteer bias: given the nature of the databases used as well as the professionalization, predisposition and incentives of the respondents, it is possible that the results have been affected by this bias, and that the sample finally selected will be systematically different from the Spanish general population; (b) Non-response bias: will exist if many contactees did not want to participate. In other similar investigations, it was estimated at 30-40%, so the sample size will be increased to ensure a quantity of sample appropriate to the research objectives.
• attrition bias: it will occur or not, as long as the subjects who accept the survey do not complete it. To ensure an adequate sample size for the objectives of the investigation, the sample size will be increased and the factors related to the survey that, according to similar research, could lead to dropouts, will be avoided (mainly: large surveys, open ended answers ).
• recall bias: being a typical bias in cross-sectional studies and surveys, it is possible that subjects do not remember what really succeed. It will be avoided questions that implies having to remember events that happened for a long time ago (limiting to one year ago).

Data analysis: tables with the characteristics of the sample that include the variables of age, sex, educational level, income level and autonomous community will be included. Likewise, the response and drop-out ratio will be included. The analysis will be stratified by sex, age ranges 18-24a, 25-44a, 18-44a, 45-64.

• analysis of the knowledge domain the following statistics will be estimated: number and frequency of subjects that give correct response (95% CI); number and frequency of subjects who do NOT give the correct answer (95% CI); number and frequency of subjects with all correct answers (95% CI); number and frequency of subjects partially correct answers (95% CI); Knowledge score per question = sum of correct answers given by the subjects, divided by the total subjects consulted (95% CI).
• analysis of the attitudes domain the following statistics will be estimated: number and frequency of responses considered positively inclined towards. (95% CI); number and frequency of responses considered inclined negatively towards. (95% CI); number and frequency of responses considered uncertainty of inclination towards. (95% CI); Attitude score per question: a number is assigned to each category of the scale and the score will be the sum of the scores of all the subjects in that question, divided by the total number of subjects (95% CI)
• analysis of the practices domain the following statistics will be estimated: number and frequency of subjects performing a practice (95% CI); number and frequency of subjects who do not perform a practice (95% CI); number and frequency of subjects who do not know if they perform a practice (95% CI); number and frequency of subjects who perform a specific practice indicated in more depth (95% CI).

Inclusion criteria

• age between 18-65 years old
• male/Female
• spanish Nationality
• be a consumer of an audited and verified internet portal

Exclusion criteria

• age out of the range
• be of a different nationality than the Spanish
• be an internet consumer of an unaudited portal