Introduction

Safe animal feeding is important for animal health, animal food consumer safety, and the environment. There is a close link between the safety of animal feed and derived foods such as eggs. However, additives are deliberately added to animal feed or to the animal directly¹. While eggs are a high-demand product that has promoted the growth of the poultry industry and intensive agriculture, they have also increased the morbidity and mortality of farm poultry, which in turn can lead to diseases in the population such as fowl cholera, avian flu, spotted liver disease, avian salmonellosis, infectious bronchitis, Marek's disease, Gumboro disease, and parasitic diseases² due to bacteria, viruses, fungi, internal and external parasites, and other handling-related diseases³. Veterinary antibiotics are one of the most viable solutions to combat them.

There is evidence that some poultry producers administer human antibiotics or antibiotics prescribed for other animal species⁴; this may be legal, but their residues may be present in eggs, egg byproducts, and biowaste, including eggshells⁵, leading to the development of antimicrobial resistance⁶. Thus, the presence of antibiotics in egg yolk and albumin is related to the active ingredient or to the pharmacokinetic properties of the antibiotic in question. which will follow different distribution routes within the organism or in the animal tissues, leaving residues that depend on the type of antimicrobial⁷. Consequently, this instills a suspicion of unsafety in egg consumers due to the potential risks involved, mainly because of the constant health crises and alarms, such as the avian flu virus and others that attack poultry.

Food safety is highly compromised when the Maximum Residue Limits, the withdrawal periods for the antibiotics administered, the effects of antibiotics on animals, and the regulatory standards for the use of veterinary antibiotics are not respected. At the international level, the norm CX/MRL 2-2021 of the Codex Alimentarius of the United Nations’ World Health Organization and Food and Agriculture Organization (FAO), which deals with Maximum Residue Limits (MRLs) and Risk Management Recommendations (RMR) for veterinary antibiotic residues in food, is taken as a reference⁸. In the case of the Dominican Republic, it is regulated by Decree No. 354-10, which establishes the technical regulation of the MRLs of veterinary antibiotics and related substances in food of animal origin⁹. This implies that consumers are constantly exposed to this type of antibiotics and any other additives used in animal feed and disease control that can put their health at risk.

Therefore, it is of great interest to have instruments to determine the reliability and validity of the use of veterinary antibiotics in poultry, in addition to the consumer’s knowledge, attitude, and practice (KAP) for the consumption of food of animal origin, as all these play a role in ensuring food safety.

The objective of this research was to design and validate a questionnaire to assess the risk of veterinary antibiotics used in poultry production in egg-laying hens and their perceived impact on the consumers in terms of food safety in the Dominican Republic.

Material and methods

Results

A total of 429 people responded to the questionnaire. The sample of this research was made up of 93.2 % men and 6.8 % women in the case of farms and veterinaries. For consumers, it consisted of 50.9 % women and 49.1 % men.

Cronbach's alpha score measuring the internal consistency of the questions was satisfactory (α= 0.799 and 0.771). Tables 1a and 1b show their values for each questionnaire. Internal consistency was satisfactory in all the domains. However, 7 items were eliminated from the initial 29 items of the first questionnaire (Table 1a), and 9 items, from the original 17 items of the second questionnaire (Table 1b), considering the analysis of the corrected item total, whose correction was deemed necessary because it exhibited a negative correlation and very low representativeness among the questions, which affected the subsequent analysis.

The normality test for both questionnaires showed that there is no significant correlation (α= 0.05) for the Kolmogorov-Smirnow and Shapiro-Wilk methods, as all variables show significance results of P<0.000, i.e., below alpha.

The exploratory factor analysis identified a Kaiser - Meyer - Olkin sampling adequacy measure for egg producers and veterinarians of 0.558, while for consumers it was 0.79. Bartlett's test of sphericity is significant at P=0.000 <α<0.0. The degree of significance has a value of 0.000, i.e., the hypothesis of the identity matrix is rejected, and there is a correlation between the variables (Tables 2 and 3).

The total variance explained test for egg producers and veterinarians found that the first 8 components were able to account for 72.035 % of the cumulative variance representativeness of the selected items (Table 2). For consumers, the amount of total variance that is explained by each extracted factor is 3 factors, with a cumulative variance representativeness of 74.807 % (Table 3).

In the test of unidimensionality of the construct, as established by Kaiser's rule in the sedimentation graph, 8 factors were obtained according to the line drawn at the eigenvalue level for egg producers and veterinarians and 3 for consumers, which explain most of the total variability (Figures 1 and 2).

Figure 1 Questionnaire survey from egg producers and veterinarians 

Figure 2 Questionnaire for egg consumers 

Tables 4 and 5 of the rotated component matrix show the component data that were extracted, using the Varimax orthogonal rotation with Kaiser normalization, for eight components for egg producers and veterinarians and three components for consumers. The cut-off point as coefficient of factor loadings of the weights and weightings started at 0.5 within each factor, and the communality value was equal to or greater than 0.5.

The instrument or model studied for egg producers and veterinarians was structured with 22 items grouped into 9 factors or dimensions. For consumers, it was made up of 8 items and 3 factors or components.

The components according to the group of items, and to their internal consistency, that the model incorporates for egg producers and veterinarians are the following:

Factor 1, technical characteristics of the control and prescription of veterinary antibiotics, and knowledge and compliance with regulations for their use in poultry production. It comprises 9 items and accounts for 23.283 % of the total variance. For Factor 2, characteristics of veterinary antibiotics used in poultry production, accounts for 10.435 % of the total variance and contains 4 items. Factor 3, technical factors of poultry sanitary management and use of veterinary antibiotics, produces 9.140 % of the total variance and includes 4 items. Factor 4, time of duration of veterinary treatments when administered to laying hens, is the cause of 7.769 % of the total variance and consists of 2 items. Factor 5, technical characteristics in the frequent application of veterinary antibiotics to laying hens, results in 5.948 % of the total variance and is grouped into 2 items. Factor 6, regular use of veterinary antibiotics in egg production and food safety, accounts for 5.444 % total variance, with a single item. Factor 7, technical characteristics of bird handling in the administration routes of veterinary antibiotics, has an explanatory value of 5.060 % of the total variance and is represented by 2 items. Factor 8, administrative management of the farms/veterinaries with respect to keeping records to establish traceability systems in the production, amounts to 4.955 % of the total variance and is represented by a single item.

The components extracted for the consumers incorporated into the model by group of items and their internal consistency are the following:

Factor 1, characteristic of egg consumption and frequency of consumption. It covers 42.884 % of the total variance, includes 4 items that have a positive correlation with the consumption of hen’s eggs (table eggs) as food; the number of eggs consumed according to the frequency; the manner of egg consumption, and the frequency of consumption of eggs. Factor 2, purchase and verification of the quality and hygiene conditions of eggs at the point of sale (supermarket, market, other). This component accounts for 19.151 % of total variance and comprises 2 items and groups: when you buy the egg products, do you take notice of the packing date and expiration date and the type of commercial brand? Do you consider egg consumption to be safer than that of other foods? The latter corresponds to section 3. While, Factor 3, consumer perception of the presence of veterinary antibiotic residues in eggs and compliance with the regulations by poultry producers, represents the variables considered as consumer perception of compliance by poultry farmers with the sanitary measures in egg production. This component represents 12.772 % of the total variance, includes 2 items, and covers compliance of egg producers in the application of the Dominican animal health legislation to treat diseases in laying hens, as well as compliance with the withdrawal period of veterinary antibiotics in laying hens according to the antibiotic label when placing eggs on the market for consumption.

Discussion

The assessment of the potential risks of veterinary antibiotics in egg consumption and their impact on food safety is not easy to analyze due to various factors associated with the use of veterinary antibiotics. In this sense, this paper provides a practical tool to evaluate aspects related to the use of antibiotics and its role in ensuring food safety. In the present study, construct validity was assessed by means of an exploratory principal-component factor analysis, and the internal consistency of the questionnaires was evaluated by means of Cronbach's alpha coefficient. Cronbach's alpha model in veterinary epidemiology has been applied very scarcely for the development, evaluation, and validation of questionnaires¹³; even so, it has been used in preventive veterinary medicine¹⁴, being useful for this research.

The content and logical validity assessments of the questionnaires by a group of experts were favorable. The majority of the surveyed professionals responded with a maximum score, indicating that they agreed with the format, wording, and usefulness of the questionnaire and that Cronbach's alpha model in veterinary epidemiology has been applied very scarcely for the development, evaluation, and validation of questionnaires.

The internal consistency of the questionnaires obtained a Cronbach's alpha of 0.799 for the egg producer/veterinarian questionnaire, and 0.771 for the consumers’ questionnaire, indicating that the instruments have adequate reliability for the measurement of veterinary antibiotic use and the perception of food safety-related impact, respectively.

Regarding the construct validation, it was observed that the principal component analysis yielded 8 factors for the questionnaire for egg producers in poultry farms or veterinary establishments, and 3 factors for the consumer questionnaire, which associates the similarity of correction between the variables of the evaluated study. This suggests that what has been described above constitutes a first insight into the perception by the farm owners, veterinarians, and consumers of the association that exists in the use and management of veterinary antibiotics or the use of antimicrobials in food production for human consumption.

The instrument applied to producers in egg farms and veterinary establishments and consumers was designed to evaluate the use of veterinary antibiotics in laying hens and the consumers' perception of the risk associated with table egg consumption and food safety. The results show that the hypothesis in the correlation matrix was positive between the variables with Bartlett's test of sphericity. The Kaiser-Meyer-Olkin (KMO) measures of sampling adequacy for egg producers and veterinarians (0.558) and consumers (0.797) have a high positive correlation, which indicates that their values are adequate because they range between 0 and 1, i.e., they are close to unity. These results coincide with those found by Salazar(¹⁵, who obtained a relatively high KMO (of 0.725). A KMO of over 0.80 in the data matrix is appropriate for running the factorization¹⁶. In another study¹⁷, a high KMO value (0.94) was observed for the food estimation and food frequency section, where Bartlett's test of sphericity proved significant (P<0.001), converging in 10 iterations and a six-factor structure.

The total variance tests of this study confirm that the variance matrix values, covariance, and percentage of each of the items, and the eigenvalues of the quantities of poultry production farms, veterinarians, and egg consumers are accounted for by each extracted factor and by the related percentages in the equation model. The residual analysis for checking the goodness of fit of the utilized factorial model shows that the results of the differences between the initial observed correlation matrix and those reproduced by the model indicate that this value is considered an indicator of good fit as it is close to absolute zero.

The analysis of the resulting principal components that were above 0.5 according to the groups of items, both for egg producers/veterinarians and consumers, made it possible to determine the magnitude of the samples of the effect that the variables had on each one of the components that provide the best exposure of the initial variables obtained in each component, with their respective positive or negative factor loadings. Loadings of 0.50 can generally be considered strong and allow the magnitude of factor loadings to be evaluated as a function of sample size^18,19. This allows the interpretation of the factor loadings that have an absolute value above 0.4 with their variance of the variables evaluated²⁰. In this sense, another validation study of the questionnaire on food estimation and frequency of food consumption¹⁷ found a correlation ≥ 0.40 with a reliability index of 0.92 for section estimation and of 0.90 for food frequency. The data found in this study allowed us to discriminate the variables with positive or negative factor loadings below 0.5, so that each of the dimensions of the instrument had acceptable values (≥ 0.5) and made it possible to perform the global scale analysis.

The domains or dimensions used for the evaluation of the use of veterinary antibiotic use and fowl management factors in poultry production are related to the items studied by Chah et al²¹ mainly in regard to the characteristics of antibiotic use in small-scale poultry farming, the knowledge, and the kinds and frequency of the antibiotics utilized in poultry farms. Speksnijder et al²² also evaluated dimensions related to having a lower threshold for applying antibiotics to animals; their results resemble the ones obtained for the items evaluated in this research both in sections two and eight, on the characteristics of veterinary antibiotics used in poultry production and farm/veterinary administrative management, respectively.

Principal component analysis for the egg producers’ and veterinarians’ questionnaire confirmed that positive scores above 0.8 were related to such items as whether or not to vaccinate the birds regularly, the route of application of veterinary antibiotics, knowledge of the withdrawal periods of the antibiotics, keeping records of veterinary antibiotic applications, and treatment with antibiotics prohibited for birds. These findings prove that poultry producers adequately manage the following poultry components: 2) characteristics of veterinary antibiotics used in poultry production; 4) duration time of veterinary treatments administered to laying hens; 6) regular use of veterinary antibiotics in egg production and food safety; 7) technical characteristics of poultry handling in veterinary antibiotic administration routes, and 8) administrative management of farms/veterinaries with respect to keeping records to establish a production traceability system, respectively.

The highest score (0.9) for the consumer questionnaire was achieved in section or domain, characteristics of egg consumption, and frequency of egg consumption. The other assessed sections ―including the consumers' perception of the presence of veterinary antibiotic residues in eggs and the verification of egg quality and hygiene conditions in sales outlets (supermarket, market, others)― have lower scores (≥ 0.6 and ≥0.8), which agree with the results obtained by other authors¹⁷. Studies carried out by various researchers^23,24 assessed methods for developing food safety, knowledge, and attitude scales to determine criteria for reliability and validity. According to the study by Al-Makhroumi et al²⁵, in the three evaluated sections, the respondents had low food safety knowledge, with a value of 44 %, compared to the other sections such as good practices, with 70 %, and positive attitudes, with 77 %. Other researchers²⁶ found a moderately positive correlation between the mean scores of antibiotic knowledge and antibiotic use (0.55 P<0.001), and a moderately positive correlation between the participants' mean scores on antibiotic resistance knowledge and their scores for knowledge of antibiotic use (0.41P<0.001). The results obtained from the consumers surveyed in this study show that the second domain, on the consumers' perception of the presence of veterinary antibiotic residues, and the third domain, of egg quality and hygiene verification at points of sale, have low scores compared to the first domain, on the egg consumption and frequency of consumption characteristics, indicating a lack of independent consumer awareness of poultry management practices and little knowledge of the antibiotics administered to laying hens.

Finally, in this study, in order to establish the model with a Cronbach's Alpha coefficient over 0.5, it was necessary to adjust the items, i.e., to eliminate some variables that could be important for future studies and discussions of the original model. Furthermore, as mentioned by Hernández and Amador²⁷, a confirmatory factor analysis should be performed to confirm the theory, as the purpose of the utilized factor analysis was to construct the theory.

Conclusions and implications

The results of the present study confirm the reliability and validity of the questionnaire items, finding a satisfactory fit between the use of veterinary antibiotics in egg production and egg consumption. The value of over 0.7 obtained for both questionnaires in the assessment of the validity and reliability of the results of Cronbach's Alpha coefficient shows that the established model fits the extracted components with their variance of over 50 %; this represents a strength of the research because the scale of competence used for the construct produces fast and reliable results that serve to measure the incidence or risks in the health of people due to the consumption of food contaminated through the use of veterinary antibiotics in laying hens as antimicrobials or as growth promoters for egg production.