2. Materials and methods
2.1 Target sites and survey population
The provinces of Kampong Cham (KC) and Pursat (PS) were purposively selected for this study. These provinces contribute significantly to cattle production and contain primary access routes for the cattle trade into Vietnam. PS is located in the west of the country, bordering four provinces, the Tonle Sap (Sap River, a tributary of the Mekong River) and Thailand, whereas KC is located in the central lowlands of the Mekong River, bordering six provinces, the Mekong River and Vietnam. Target smallholder farmer households for the study were households participating in the ‘Domestic and International market development for high-value-cattle and beef in South-East Cambodia’ project (ACIAR project AH/2012/046), a collaborative research effort facilitated by the University of New England and the Cambodian General Department of Animal Health and Production (DAHP), with funding by the Australian Centre for International Agricultural Research (ACIAR). The communes and villages targeted for the survey were identified and selected through discussions and consultations between the researcher, representatives of the Provincial Office of Animal Health and Production (POAHP), and the ACIAR-funded project team located in DAHP and POAHP offices. These communes and villages were considered more likely to have an outbreak of FMD than other areas, and village selection was based on the main criteria of “at least 50% of households kept cattle”. Subsequently, five villages were selected in each of the two provinces. A total of 300 smallholder farmers with at least three cattle per household were identified for the survey, comprising 150 in each province.
2.2 Survey questionnaire design and interview technique
A semi-structured questionnaire consisting of open, closed and semi-closed questions was developed to collect both quantitative and qualitative information from the smallholder farmers. The questionnaire was developed in English and translated into Khmer, ensuring all questions were easy to understand and answer. The questionnaires collected information on the background of farmers, household assets, household incomes, cattle management, cattle trading, knowledge and experiences of diseases and vaccination practice, and the financial impact of FMD and perception on the importance of vaccination. In addition to the specific questions, information on opportunities and obstacles in the previous FMD vaccination programs was elicited, and suggestions for improvements to future vaccination programs obtained. Before the survey, the independent survey team received a full day of training to ensure they understood the aims and objectives of the study and were confident in their roles. To ensure best practice, a pilot study was carried out to test the survey questions, delivery technique and the length of time required for the interview. After initial field testing, the survey questionnaires were reviewed and revised according to feedback by the team. To ensure confidentiality and encourage honest responses from the farmers, government officials were excluded from participation in the interviews. At the end of the session, completed questionnaires were checked to ensure that no information was missing.
2.3 Data collection and analysis
The survey data were collected, translated from Khmer into English and transcribed into a database. The statistical analysis was performed using SPSS version 21.0 (SPSS Inc., IL, USA). Outlier values were compared with the questionnaires and corrected where necessary. Standard descriptive analysis of the data was performed to summarize household and livestock information, cattle trading, household assets and incomes, knowledge, experiences of FMD and vaccination practice, plus their perceptions on the importance of vaccination. Statistical analyses were undertaken by the application of both univariable and binary logistic regression methods. The influence of smallholder farmer attitudes and knowledge on FMD outbreak and vaccination practices were classified into three groups: age, education, and the number of household members; livestock husbandry and management; and vaccination knowledge. Variables from each group were subsequently tested against the binary independent variable of whether farmers had undertaken vaccination (i.e., Yes or No to FMD vaccination) using the Pearson’s chi-square test for independence and a P values <0.05 as evidence of significance. Odds ratios (OR) and their 95% confidence interval were also calculated to measure the magnitude of the association between factors and yes/no of the dependent variable. Variables that were significant at P ≤0.25 in the univariate analyses were selected for inclusion in a logistic regression model (Hosmer & Lemeshow, 2000). The model was performed using backwards conditional testing, and odds ratios were calculated. In the final model, variables with P values >0.05 and which had value 1 in the 95% confidence interval of OR were excluded. The model was evaluated by calculating the Hosmer and Lemeshow statistical methodology (Hosmer & Lemeshow, 2000).