Development and validity of a questionnaire to test the knowledge of primary care personnel regarding nutrition in obese adolescents

Background In light of its epidemic proportions in developed and developing countries, obesity is considered a serious public health issue. In order to increase knowledge concerning the ability of health care professionals in caring for obese adolescents and adopt more efficient preventive and control measures, a questionnaire was developed and validated to assess non-dietitian health professionals regarding their Knowledge of Nutrition in Obese Adolescents (KNOA). Methods The development and evaluation of a questionnaire to assess the knowledge of primary care practitioners with respect to nutrition in obese adolescents was carried out in five phases, as follows: 1) definition of study dimensions 2) development of 42 questions and preliminary evaluation of the questionnaire by a panel of experts; 3) characterization and selection of primary care practitioners (35 dietitians and 265 non-dietitians) and measurement of questionnaire criteria by contrasting the responses of dietitians and non-dietitians; 4) reliability assessment by question exclusion based on item difficulty (too easy and too difficult for non-dietitian practitioners), item discrimination, internal consistency and reproducibility index determination; and 5) scoring the completed questionnaires. Results Dietitians obtained higher scores than non-dietitians (Mann–Whitney U test, P < 0.05), confirming the validity of the questionnaire criteria. Items were discriminated by correlating the score for each item with the total score, using a minimum of 0.2 as a correlation coefficient cutoff value. Item difficulty was controlled by excluding questions answered correctly by more than 90% of the non-dietitian subjects (too easy) or by less than 10% of them (too difficult). The final questionnaire contained 26 of the original 42 questions, increasing Cronbach’s α value from 0.788 to 0.807. Test-retest agreement between respondents was classified as good to very good (Kappa test, >0.60). Conclusion The KNOA questionnaire developed for primary care practitioners is a valid, consistent and suitable instrument that can be applied over time, making it a promising tool for developing and guiding public health policies.


Background
In recent years, the worldwide prevalence of overweight and obesity in children and adolescents has increased significantly [1]. The high overweight incidence is considered a serious public health issue in the United States [2] and developing countries [3,4], highlighting its importance for healthcare providers and administrators [5].
Obesity has become an epidemic and, although this situation must be remedied, current public healthcare models have been inefficient in doing so. In order to revert this situation, a number of approaches have been suggested to prevent, evaluate and treat overweight and obesity in children and adolescents [6], particularly in primary care [7]. However, the success of any action plan depends on primary care staff's having the knowledge and ability to prevent overweight and obesity in children and adolescents. Obese children and teenagers and their families require extensive guidance and assistance to increase their self-management capabilities and promote a healthy lifestyle [4].
The role of primary care practitioners is to provide obese patients with basic dietary counseling and, if necessary, refer them to a dietitian for individual care [8]. However, primary healthcare providers are rarely capable of providing dietary counseling, [8,9]. This is a serious problem since nutrition is essential as auxiliary therapy in the treatment of chronic illnesses [10]. These diseases have increased drastically in recent years and currently affect a significant portion of the population [10].
Against this backdrop, the nutritional counseling provided to obese patients must be improved and standard protocols established [8]. Primary care practitioners can provide effective nutritional counseling if they: 1) receive appropriate training, 2) are given effective tools, 3) operate in an organized system and, 4) are assisted by qualified dieticians [11]. For family physicians, successful implementation of Clinical Practice Guidelines on the Management and Prevention of Obesity requires adequate training, and evidence-based interventions related to nutrition [9].
Primary care practitioners should be trained not only to evaluate overweight and obesity in children and adolescents, but also to provide these patients with the care required for a healthier life. In order to adequately treat these conditions, care providers should participate in specific and continuous training programs [5].
In light of the above, it is important to evaluate the knowledge of primary care practitioners regarding adolescent nutrition and obesity. However, the tools used to assess these disorders in teenagers were developed for patients, and protocols to evaluate the knowledge of nondietitian primary care practitioners are as yet unavailable. Though scarce, these studies are essential in guiding public health policies developed to combat childhood and adolescent obesity, and in creating training programs that enable primary care practitioners to provide adequate nutritional counseling to overweight and obese children and adolescents.
The present study was conducted to compile and validate a specific questionnaire to test the Knowledge of Nutrition in Obese Adolescents (KNOA) of non-dietitian primary care practitioners.

Methods
The dimensions of knowledge and question were based on the studies listed in Appendix 1. They consist of didactic materials and guidelines for healthy adolescent diets available to primary care practitioners, some of which are recommended by the Brazilian Ministry of Health. The questionnaire development and validation procedure is illustrated in Figure 1 and described below.
Step 1definition of knowledge dimensions KNOA was considered to be the nutritional knowledge that any primary care practitioner must have in order to treat obese adolescents. After an extensive literature review (Appendix 1) we didactically selected dimensions for KNOA (Table 1), as follows: epidemiology of obesity in adolescence; clinical implications of obesity in adolescence; mapping obesity in adolescence; nutritional management for obese adolescents, and knowledge of fruit, vegetable, fat and sugar consumption, as described in Table 1. Albeit somewhat arbitrary, this categorization was useful in obtaining better balance in the next step of the study.
Step 2 -Development of the items Considering the studies listed in Appendix 1, we selected 81 sentences related to the KNOA dimensions described in Table 1 (more than 10 questions per dimension). From this database we selected 42 questions based on two criteria: the frequency with which the issues arose in the different studies and their importance and prevalence in clinical practice. The initial questionnaire was elaborated and applied in Portuguese.
This initial questionnaire was submitted to a panel of five experts: two dietitians, two pediatricians and one nurse, who evaluated the appropriateness, relevance, accuracy, and formulation of each question. The adequacy and clarity of the questionnaire was also assessed in a pilot study conducted with 20 primary care practitioners (2 physicians, 13 nurses and 5 dentists) belonging to a private health group in a city in northern Minas Gerais state. The pilot study was performed to check for clarity and readability of individual questionnaire items and the adequacy of this tool in assessing the knowledge of health practitioners, thereby contributing to the face validity of the questionnaire. Problems raised by the health professionals were noted, and suitable changes were made accordingly. Unclear and dubious questions were excluded.
Step 3 -Ensuring criterion validity of the questionnaire To determine criterion validity, all 402 primary care practitioners registered with the Secretary of Health in a city in northern Minas Gerais and dietitians from public and private services were invited to complete the 42-item questionnaire. All the subjects had at least two years of primary care experience. They provided personal information (demographic and educational) and were also questioned regarding the care of obese adolescents, who took the initiative to talk about nutrition (practitioner or patient), and the main barriers to discussing this subject.
The professionals were expected to achieve a high or low rate of correct answers in accordance with their level of nutritional education [12]. Thus, dietitians were expected to achieve a higher mean score than primary care practitioners, with a statistically significant difference to ensure criterion validity. The questionnaire was applied at their workplace in the first half of 2012.
Step 4 -Questionnaire reliability Several procedures were implemented to increase questionnaire reliability: a) in validity testing, items correctly answered by more than 90% of the 265 non-dietitians (considered too Pilot study (20 primary care practitioners) Questionnaire adjustments (2 questions excluded) Step 2 -Development of the items Step 3 -Ensuring criterion validity of the questionnaire 40-item questionnaire application to 300 primary care practitioners Contrast between dietitian (N=35) and nondietitia ns (N=265) ra ting Step 4-Questionnaire reliability Item difficulty (7 questions excluded) Item discrimination (7 questions excluded) Internal consistency Reproducibility index 42-item questionnaire

Knowledge of Nutrition in Obese Adolescents (KNOA)
Question analysis by specialists easy) or less than 10% (considered too difficult) were excluded [13]; b) items were discriminated by correlating the score of each item with the mean score of the complete questionnaire. A minimum cutoff point of 0.2 was used to determine the correlation coefficient between the total mean score and each questionnaire item, and items with lower values were excluded [14]; c) the internal consistency (reliability) of questionnaire results was evaluated using Cronbach α values [15]; d) the reproducibility index was assessed over two weeks, applying the reformulated version to nearly 20% (60 subjects) of the initial sample [16].
Step 5 -Score determination A score ranging from −1 to +1 was determined for each dimension. Items were scored −1 for an incorrect answer, 0 when the response was "I don't know" and +1 for correct answers. The mean values of the items were calculated to obtain the dimension scores for each respondent, followed by calculation of the mean score for each group.

Data analysis
Descriptive statistics with distribution frequency was used to characterize the sample.
The Mann-Whitney U test was applied to compare dietitians with non-dietitians in questionnaire validity testing (Phase 3), using a significance level of P < 0.05.
In Step 4, the Kappa test was used to evaluate agreement between mean test-retest scores for questionnaire items [16].
Statistical tests were carried out using Statistical Package for the Social Science (SPSS) software, version 15.0.

Ethical questions
The investigation respected ethical guidelines for research with human beings. Participation was voluntary and subjects gave their informed consent. The study was approved by the Research Ethics Committee of the Universidade Estadual de Montes Claros, under protocol number 3016/2011.

Questionnaire development and validity Subject characterization
Thirty five dietitians and 265 non-dietitian primary care practitioners (76 physicians, 139 nurses and 50 dentists) participated in the study. Most participants were women (65%), aged 32.7 ± 8.6 years, who had graduated less than 5 years before (54%), and were specialized in primary health care (67%). A total of 84.3% of participating practitioners treated obese adolescents. Among the main difficulties in caring for these patients were poor patient adherence to recommendations (41.3%) and lack of knowledge regarding adolescent nutrition (25.4%) ( Table 2).

Final questionnaire
Seven of the initial 42 questions (Step 4) were excluded by the difficulty criterion: 7 based on the low correlation obtained in item discrimination and 2 after the pilot study because they were considered too specific for a questionnaire involving non-dietitian professionals. The final questionnaire contained 26 items (Table 3).

Criterion validity of the questionnaire
As expected, the mean score obtained in the dimensions evaluated was higher for dietitians than non-dietitians, except for dimension I -Epidemiology of obesity in adolescence (Table 4).

Table 1 Dimensions of knowledge on KNOA and its objectives
Dimension Objective

Epidemiology of obesity in adolescence
Understanding the distribution and determinants of obesity in adolescence.

Clinical implications of obesity in adolescence
Association between obesity in adolescence and health loss, diseases and associated complications.

Mapping obesity in adolescence
How obesity is diagnosed in adolescents and how this issue is discussed with the patient.

Nutritional management for obese adolescents
Knowledge on general nutritional measures for treating obese adolescents.

Fruit and vegetable consumption
Knowledge of the nutritional properties of fruits and vegetables, their necessity and importance for obese adolescents.
Fat consumption Knowledge of the nutritional properties of fats, their requirement in adolescents and adjustments for obese individuals.

Sugar consumption
Knowledge of the nutritional needs of adolescents in terms of carbohydrates and sugars and adjustments for obese individuals, considering their eating habits (food, candy, desserts and sweeteners).

Internal consistency
Principal Component Analysis was applied to the knowledge dimensions (Table 1), and since the results did not support a factor structure, the questionnaire was treated as a single scale. Accordingly, a single Chronbach α value of 0.788 was calculated for the original 42-item questionnaire. Following the exclusion of 16 inadequate items, the α value increased to 0.807, indicating satisfactory internal consistency. Test-retest agreement between answers (Phase 4) was classified as good and very good (Table 6).

Discussion
The present study met the goal of developing a questionnaire to test the Knowledge of Nutrition in Obese Adolescents (KNOA) among non-dietitian primary care practitioners. Other studies have validated tools assessing the eating behavior of adolescents, but these were not designed for healthcare practitioners.

-
Most studies indicate that the prevalence of overweight or obesity in adolescents ranges from 10% to 15%.

-
Changes in nutrition habits such as increase in carbohydrate and fat consumption are directly associated to the current prevalence of adolescent obesity.
3 -Obese adolescents have potential for becoming obese adults.

-
The chance of obese adolescents developing type 2 diabetes is 2 to 3 times greater than that of non-obese adolescents.

-
At least 10% of obese adolescents present with arterial hypertension.
6 -Obesity in adolescents is positively related to dyslipidemia.
7 -Body mass index (BMI) is considered a sufficient indicator of the nutritional status of adolescents.

-
Preceding the pubertal growth spurt, adolescents may exhibit an overweight appearance that is not diagnosed as obesity.

-
Adolescents with a weight-for-age percentile greater than 85% are diagnosed as obese.
10 -Adolescents undergoing treatment must be evaluated every 6 months.

-
In the treatment of obese adolescents, weight gain interruption, with weight stabilization within a growth chart percentile which represents obesity can be assumed to be a satisfactory preliminary therapeutic result.
12 -Food guide pyramids should be shown to patients to explain nutrient variety, moderate consumption and proportion of food items.

-
It is recommended that obese adolescents have 4 daily meals: breakfast, lunch, afternoon snack and dinner.
14 -Fruit can be replaced by fruit juice.

-
The health benefits of fruits and vegetables are that of providing vitamins and mineral salts.

-
Consuming one apple and one banana every day meets daily fruit consumption recommendations.

-
Consuming fiber-rich fruits promotes a feeling of satiety, contributing to weight control.
18 -Fats must be excluded from the diets of obese adolescents.

-
Obese adolescents should consume "diet food" to limit dietary fat.

-
Obese adolescents should avoid drinking milk due to its high fat content.

-
Obese adolescents may include low-fat sandwiches (containing turkey breast, ricotta cheese and green leaves) in their diet.

-
In contrast to saturated fats, unsaturated fats do not cause health problems unless they are consumed excessively.

-
The amount of sweets or sugary food items recommended for obese adolescents is limited to a maximum of one daily portion.

-
The use of artificial sweeteners as a substitute for sugar is indicated in the treatment of obese adolescents.

-
Carbohydrates with a low-glycemic index are known to play a positive role in dietotherapy of obesity.

-
Fruit-based rather than creamy desserts should be adopted when treating obese adolescents. Johnson et al., (2002) compiled and validated a 23item checklist to measure eating behavior in adolescents by analyzing fat, fiber, fruit and vegetable consumption, dietary restrictions, knowledge on nutrition and family income. Turconi et al. [17] developed a questionnaire on the food habits, eating behavior and nutritional knowledge of teenagers, with potential for measuring the effects of nutritional interventions. Whati et al. [13] designed and validated a survey to assess the nutritional knowledge of South African adolescents aged 13 to18 years. Kruseman et al. [18] proposed a self-administered questionnaire for 9 to 15-year-olds to assess their knowledge of nutrition, and which could also be used by primary care practitioners involved in nutritional counseling. To complement these investigations, the present study proposes and validates a tool to assess KNOA in primary care practitioners. According to the data recorded, 84.3% of the primary care practitioners participating in the study had previously treated obese adolescents, reinforcing the presence of these patients in primary health centers as well as the importance of applying the questionnaire to establish strategies that improve their nutritional counseling.
In regard to questionnaire validity, dietitians obtained higher KNOA scores than other practitioners, which was expected considering their educational level. The poor performance exhibited by the other professionals assessed suggests limited KNOA [7,8,10,11]. In practice, this result is relevant and alarming since such a lack of preparation can compromise the nutritional counseling of obese adolescents.
The final version of the questionnaire demonstrated satisfactory internal consistency and reliability, as shown by the Cronbach statistics and test-retest procedure. In addition, questionnaire dimensions must be evaluated concurrently, given that internal consistency decreases when these are analyzed individually. Since the Cronbach α value indicates a positive association with the number of items on a scale, a higher number of questions is associated to greater consistency (Hair et al., 2005). Similar results were obtained in research performed by Steyn et al. [14].
The test-retest results indicate temporal stability, allowing comparisons to be drawn over time [17]. Thus, the questionnaire developed can be potentially applied usjng a temporal approach to evaluate, for example, the effects of interventions such as educational programs and training of primary care practitioners.
This study was conducted using intentional sampling of primary care practitioners from a single municipality, and given the cultural particularities of this group, extrapolations of the findings are limited. Thus, questionnaire items related to epidemiology should be adapted or replaced if applied to other localities. Despite this, the present study can gives rise to further investigations on KNOA in primary care practitioners in light of the scarcity of national and international research on this topic. Future studies will explore the potentialities of the questionnaire, considering its use in assessing the effectiveness of educational programs, and diagnosing KNOA problems in primary care practitioners.

Conclusions
The KNOA questionnaire developed specifically for primary care practitioners is valid, consistent and suitable for application over time in the study area. Although it must be adjusted to be applied to localities of different epidemiological contexts, it is a promising tool in the development and guidance of public health actions related to obesity in adolescence.