Table 3

Significant univariable associations between outcomes related to animal husbandry and infection control practices and risk factors reported by pet-owning households (Ontario, Canada)
Outcome (species)1 All children < 5 yrs (Exposed; N)2 All children ≥ 5 yrs (Unexposed; N)2 Risk Exposed Risk Unexposed Odds Ratio3 95% CI3 P-value3
Any children have access to the litter boxes (cat)
4 40 25.0% 78.4% 0.1 0.02-0.4 0.0004
Any children clean up feces (cat)
0 18 0% 35.3% 0.07* 0-0.5 0.002
Any children touch pet (cat)
13 52 72.2% 100% 0.04* 0-0.3 0.001
Never sleep in children’s bed (cat)
10 14 83.3% 26.9% 12.9 2.4-135.9 0.001
Any children clean-up feces (dog)
0 28 0% 36.8% 0.05* 0-0.3 0.0002
Ever lick children’s face (dog)
20 45 87.0% 61.6% 4.1 1.1-23.5 0.04
Outcome (species)1 Higher Risk Member (Exposed; N)4 No Higher Risk Member (Unexposed; N)4 Risk Exposed Risk Unexposed Odds Ratio3 95% CI3 P-value3
Within past 3 months, any off leash at a dog park (dog)
15 29 12.3% 25.2% 0.4 0.2-0.9 0.02
Household very attached (dog)
106 109 88.3% 97.3% 0.2 0.04-0.8 0.01
Dump aquarium water in bathtub, shower, kitchen sink, or bathroom sink (fish)
6 12 25.0% 60.0% 0.2 0.05-0.9 0.04
Household very attached (bird)
6 0 54.6% 0% 11.5* 1.3-Inf 0.02

1Univariable exact logistic regression used to model the likelihood of each husbandry and infection control outcome as a function of selected risk factors (all children < 5 yrs; higher risk member in household). Each row represents a separate regression model.

2Performed for children-specific outcomes. Comparison between households with all children < 5 yrs (exposed) and households with all children ≥ 5 yrs (unexposed).

3Odds ratio, 95% confidence interval of the odds ratio (95% CI), and P-value computed using exact logistic regression.

4Comparision between households with ≥ 1 higher risk members [i.e., < 5 yrs, ≥ 65 yrs, ever diagnosed with an immunocompromising condition (exposed)] and households without higher risk members (unexposed).

*Median unbiased estimate computed since odds ratio calculated to be infinite.

Stull et al.

Stull et al. BMC Public Health 2013 13:520   doi:10.1186/1471-2458-13-520

Open Data