This quasi-experimental study was conducted in 2019. It was decided to determine the effects of an educational intervention, based on the PRECEDE-PROCEED model, on changing the awareness, attitude and practice of livestock breeders in prevention of brucellosis and in vaccinating their livestock against brucellosis. The presence of brucellosis antibodies in animals’ blood was examined as the outcome measure of the intervention.
Specific objectives of the study were:
- Comparing the livestock breeders’ awareness level in intervention group with control group, before and after the intervention
- Comparing the livestock breeders’ attitude level in intervention group with control group, before and after the intervention
- Comparing the livestock breeders’ practice level in intervention group with control group, before and after the intervention
- Comparing the presence of anti-brucellosis antibody among livestock in the intervention group with the control group, before and after the intervention
The study was conducted in Lighvan, a village, located in the northern slopes of the Sahand Mountain, in the suburb of Tabriz metropolitan city, Iran. In this region, livestock breeders breed mainly sheep and goats and rarely cattle. In addition to animal breeding, they produce most of the country’s dairy products. Animal breeding and dairy production is the main job of Lighvan’s residents. So, Lighvan plays a significant role in cheese production in the country. Lighvan has a population of more than ten thousand people. Each year, one hundred thousand sheep and goats and about 15 thousand lambs and yeanling are bred in lighvan and at least 2 veterinarians and 2 livestock vaccinators vaccinate the livestock. They observe adherence to health protocols in more than 100 cheese production workplaces in this region.
Participants of this study were livestock breeders living in Lighvan. Livestock breeders, who were willing to participate; did not have a history of brucellosis in the past and had not participated in any similar training courses before, were eligible to be included in this study. Participants with the inability to use the educational package of this study, those who had physical disabilities such as visual or auditory problems, were not included. Being absent in more than 20% of training sessions was an exclusion criterion in this study.
Sample size calculation
Power & Sample Size Calculator software, version 3.0, was employed to calculate the sample size. The sample size was calculated applying confidence interval of 95%, power of 90%. The findings of the previous studies were taken into account too. The final sample size was estimated at 110, considering a 20% drop rate.
Study population and sampling method
Nearly 1700 livestock breeders work in about 100 dairying centers in Lighvan and 10 to 50 breeders work in every center. Eight centers were included in this study upon their manager’s willingness to participate in this study. In order to decrease the diffusion effect, breeders in the intervention and control groups were selected from different centers. So, those volunteer centers were randomly assigned to intervention or control groups by the research randomizer software. A stratified quota sampling method was employed to determine the exact number of participants from each center.
The steps to design educational intervention
Applying PRECEDE-PROCEED planning model, an educational package was designed to increase awareness, attitude and practice of livestock breeders in preventing brucellosis by animal vaccination. The steps to design the intervention are summarized in Figure1.
To design the educational package, based on the PRECEDE phase of the model, social, epidemiological, educational and ecological assessments were done in Lighvan. By doing so, predisposing, enabling and reinforcing factors of livestock breeders' vaccination behavior were identified. In this phase, data were collected through semi-structured interviews with all stakeholders, including livestock breeders, veterinarians, vaccinators and health care providers in Lighvan. The data was completed through direct observations of the main researcher (a veterinarian and health education specialist) during one month.
The findings from the interviews and observations in the PRECEDE phase were pooled with the reports of some worldwide evidence-based best practices. In this way, the first draft of the intervention effectiveness assessment Brucellosis Prevention Questionnaire (BPQ) was developed and validated. The psychometric properties of the BPQ and the most important predictors of the vaccination behavior of the livestock breeders were identified. The findings of PRECEDE phase and the process of designing and validating BPQ, as a valid and reliable questionnaire for assessment of the educational intervention in this study, has been published in the first article of this project. BPQ has been attached to this text as Additional file 1.
To design the package for the educational intervention, according to Harden’s six-level program development model, first, learning objectives and then the content of the training package for animal breeders were determined.
The learning objectives and livestock breeders’ educational needs were derived from the assessment results in the PRECEDE phase. The most important predictors of livestock breeders’ vaccination-related prevention behavior were identified based on the results of the factor analysis of the research questionnaire. The most important predictors, which had a greater contribution in determining the educational content, were livestock breeders’ awareness, attitude and practice. Items of livestock breeders’ awareness in BPQ were subcategorized in three groups of direct awareness, indirect awareness and vaccine awareness. All items are presented in Additional file 1. More training hours were considered to cover more important predictors (those which had a higher predictability power). The learning objectives and the educational content of the package were discussed and matched with educational needs of participants by the research team. Later, the content was systematically organized into different training sessions.
The educational content was mainly delivered through the basket method . In this method, livestock breeders faced many pictures on a board. Pictures included a range of some proper or improper issues about livestock immunization against brucellosis. Livestock breeders were expected to pick up suitable pictures after they heard a scenario. Those scenarios had been written based on the situational assessment data, which was gathered from the PRECEDE phase of the study. Livestock breeders' choices in selecting the pictures reflected their real-time decisions. Participants received constructive feedback after each individual session. After the individual training was completed in two different sessions, livestock breeders were asked to participate in focus group discussions in order to share their learning and experiences with their peers.
The method for teaching the content was not limited to the basket method and feedback from the researchers and peers. The other training methods and techniques such as questions and answers (Q&A) or brainstorming were flexibly employed when they were necessary. The educational intervention was completed through a final mini-lecture at a proper public place and time by the main researcher. Proper educational strategies such as problem-based and integrated learning were adopted to increase the quality of the education as well. Lesson plans for all training sessions and educational scenarios were written and finalized by the research team members.
The steps to Assessment of vaccination rate
Rose Bengal rapid screening test was used to evaluate and compare the vaccination rate of sheep in the two groups. For this purpose, in two intervention and control groups, three sheep were sampled from each participating livestock breeder in the study. In the time before vaccination, 135 Rose Bengal tests were performed in each group. The sheep were randomly selected with software (https://www.randomizer.org). The sampling and the Rose Bengal test were repeated in 2 to 3 weeks after Brucella vaccination, when the antibody level was at the highest rate. Vaccination was performed for free by the Veterinary Organization. The positive result of the Rose Bengal test and the presence of antibodies in the animal's blood was considered as the success of a livestock breeder in vaccinating his animal[20–22].
Data were summarized and expressed as with frequency and percentage for categorical variables and mean (standard deviation (SD)) for numeric variables. The normal distribution of the numeric variables was assessed by Kolmogorov–Smirnov test. Skewness (within ±1.5 as normal) and kurtosis (within ±2.0 as normal) of the data were assessed too. To compare the baseline variables between intervention and control groups, independent, Mann-Whitney and Chi-square (utilizing an exact procedure) tests were used.
To assess the within group changes over measurements done, the repeated measure analysis of variance (RMANOVA) were used and the group by time interaction effect and the group and time main effects were investigated. The sphericity as an assumption in this analysis was assessed by Mauchly’s test and deviation from the assumption was corrected through Greenhouse-Geiser procedure. To assess the intervention effect i. e. the between group comparisons of changes (immediately after intervention, one month after intervention and six months after intervention), the analysis of covariance was conducted in two models. In the first model, the baseline measurements were analyzed alone and in the second model, the baseline measures and the other potential confounders were adjusted for variables ,including age, occupation, family dimension and education level of the livestock breeders.
The McNemar tests utilizing the exact procedure were utilized to compare the results of the binary outcome i.e. positive/negative Rose Bengal tests results. The percent of changes were computed too. Fisher’s exact test was conducted to compare the binary outcome at the baseline and logistic regression model was used to model the difference of the binary outcome at after intervention adjusting for baseline measures. Moreover, to assess the effect of the intervention, the absolute risk reduction (ARR) and number needed to harm (NNH) were estimated along with their 95% confidence interval by Newcomb’s method and Bender's methods for ARR and NNH, respectively[23, 24]. In all analyses, P<0.05 was considered as statistically significant. The analyses were conducted using IBM SPSS Statistics version 25 (IBM corporation, Armonk, USA), STATA 16 (Stata Corp, College Station, Texas, USA) and the graphs were drawn by Graph Pad version 8.3 (www.graphpad.com).
All methods were performed in accordance with the relevant guidelines, regulations and ethical standards of the responsible committee approving the research at Tarbiat Modares University and Iranian Registry of Clinical Trials and with the Declaration of Helsinki, as revised in 2000.