Ambiente alimentar hospitalar: desenvolvimento e avaliação de confiabilidade de um instrumento na rede pública de uma metrópole brasileira

Abstract

Although many studies have assessed food environments, little research has been done to evaluate the hospital food environment. Only one instrument, developed to evaluate this setting in the United States, has been assessed for reliability. Despite the significant number of employees that work in hospitals in Brazil, there is no instrument suitable for evaluating the food environment of these institutions. This study aimed to develop a theoretical model of the hospital food environment, a tool to assess this environment, and its reliability. Subsidizing the theoretical model and instrument development, two sets containing definitions of the dimensions and items adopted were elaborated following a broad bibliographical research. In all, 11 experts gave an opinion on these materials in one face-to-face workshop. Adjustments on both materials were made based on the suggestions of the experts. Afterward, a pretest of the tool was conducted. A total of 24 public hospitals run by the municipality of Rio de Janeiro, a convenience sample of hospitals´ workers, and a chief nutritionist of each hospital were studied. The instrument was tested for inter-and intra-rater reliability. Its stability was estimated by calculating the percentage agreement, kappa index, adjusted by the prevalence and interviewer bias kappa and weighted kappa for categorical and count variables. The intraclass correlation index was calculated for continuous variables. A final version of the theoretical model contemplates the institutional and food services scopes as well as the decisional level and the hospitals´ surroundings. These scopes include the dimensions: availability, accessibility, affordability, accommodation, nutritional information, promotion, ambiance, and infrastructure for food and drink consumption. In the final version of the instrument, the first section consisted of a checklist to evaluate the characteristics of the hospital, the infrastructure for food and drink consumption and break rooms for employees. It evaluates the existing establishments, establishments setting, availability of food, information, propaganda, prices, and hospital surrounding. The second section consisted of an interview with the hospital´s workers to assess their opinion on the hospital food environment and of an interview with one chief nutritionist of each hospital about the decisions related to the offer and selling of food and drinks within these places. Regarding the inter-rater reliability, 30.0% of all items of the instrument showed very good or excellent agreement concerning kappa. Regarding the items calculated for adjusted by the prevalence and interviewer bias kappa, 69.0% showed very good or excellent agreement. Concerning the percentage agreement, 80.0% showed very good or excellent agreement. Regarding the intra-rater assessment, the items that showed very good or excellent agreement: 25.0% when kappa was analyzed; 56.0% when the adjusted by the prevalence and interviewer bias kappa was analyzed and 77.0% when the percentage agreement was analyzed. Concerning weighted kappa, 67.0% showed very good or good agreement. It is assumed that the instrument developed showed satisfactory reliability in the context in which it was used. The instrument and the theoretical model designed are expected to support the characterization of hospital food environments in other settings to guide actions on improvement of the eating conditions of hospital workers. Other researchers are expected to keep on analyzing the performance of the instrument.