Table 2. Final variables used in predictive model.
Variance reduction describes the expected reduction in the probability of above average seroprevalence in village as a result of an observed value of the predictor variable.
| Variable | Categories (states in BN) | Variance reduction |
|---|---|---|
| Urbanisation [27] | • Rural • Urban & peri-urban |
35.30% |
| Population density [27] | • High (> = 1.3) • Low (< 1.3) |
34.30% |
| Subsistence farming (% of population who depend entirely on subsistence crops) [27] |
• High (> = 13%) • Low (<13%) |
28.90% |
| Primary education or less (% of population with primary school education or less) [27] | • High (> = 24%) • Low(<24%) |
23.80% |
| Households with electricity supply (% of households on the Fiji Electric Authority’s supply grid) [27] | • High (> = 89%) • Low(<89%) |
21.50% |
| Percentage of houses with good construction [27] | • High (> = 49%) • Low(<49%) |
19.40% |
| Pigs in village [Questionnaire data, [13] | • Present • Absent |
17.60% |
| Poverty rate (% of population below poverty rate) [32] | • High (> = 39%) • Low(<39%) |
11.70% |
| Dairy farm density (Number of dairy farms/sq km in Tikina) [33] |
• High (> = 0.028 farms/sq km) • Low (<0.028 farms/sq km) |
3.82% |