Table 3.
Results of the general linear model (GLM) with Poisson error structure and log link function testing the influence of various factors on annual deworming frequency in dogs
| Estimate | SE | 95% CI | z-value | P-value | |
|---|---|---|---|---|---|
| Intercept | 0.719 | 0.307 | 0.087–1.296 | 2.343 | 0.019 |
| Owner gender (ref: male, n = 182) | -0.129 | 0.070 | -0.266–0.006 | -1.855 | 0.064 |
| Owner age | -0.001 | 0.002 | -0.006–0.004 | -0.492 | 0.623 |
| Attitude towards pets | |||||
| Affectionate (n = 269) | Baseline | ||||
| Devoted (n = 126) | -0.100 | 0.077 | -0.253–0.049 | -1.305 | 0.192 |
| Dispassionate (n = 53) | -0.074 | 0.114 | -0.304–0.145 | -0.649 | 0.516 |
| Sceptical (n = 52) | -0.256 | 0.122 | -0.502– -0.023 | -2.101 | 0.036 |
| Veterinary visits | |||||
| Once a year only (n = 227) | Baseline | ||||
| More than once a year (n = 273) | 0.328 | 0.069 | 0.193–0.465 | 4.736 | <0.001 |
| German ESCCAP risk groupa | |||||
| A (n = 10) vs B (n = 24) | -0.176 | 0.303 | -0.932–0.581 | -0.581 | 0.930 |
| A (n = 10) vs C (n = 154) | 0.008 | 0.261 | -0.646–0.662 | 0.031 | 1.000 |
| A (n = 10) vs D (n = 312) | -0.011 | 0.259 | -0.657–0.636 | -0.041 | 1.000 |
| B (n = 24) vs C (n = 154) | 0.184 | 0.178 | -0.260–0.628 | 1.036 | 0.701 |
| B (n = 24) vs D (n = 312) | 0.165 | 0.172 | -0.266–0.596 | 0.958 | 0.749 |
| C (n = 154) vs D (n = 312) | -0.019 | 0.070 | -0.194–0.156 | -0.268 | 0.992 |
| Neighbourhooda | |||||
| Rural (n = 186) vs city (n = 58) | 0.003 | 0.106 | -0.268–0.273 | 0.025 | 1.000 |
| Suburban (n = 79) vs city (n = 58) | -0.051 | 0.124 | -0.367–0.266 | -0.411 | 0.976 |
| Town (n = 177) vs city (n = 58) | -0.012 | 0.107 | -0.286–0.261 | -0.117 | 0.999 |
| Suburban (n = 79) vs rural (n = 186) | -0.054 | 0.099 | -0.306–0.199 | -0.543 | 0.947 |
| Town (n = 177) vs rural (n = 186) | -0.015 | 0.075 | -0.208–0.177 | -0.201 | 0.997 |
| Town (n = 177) vs suburban (n = 79) | 0.038 | 0.100 | -0.218–0.295 | 0.383 | 0.980 |
| Source of information regarding deworming | |||||
| Veterinarian/vet nurse (ref: yes, n = 7) | 0.412 | 0.213 | -0.029–0.811 | 1.928 | 0.054 |
| Non-veterinarian (other pet owners, pet shop staff, etc.) (ref: yes, n = 331) | -0.093 | 0.073 | -0.236–0.050 | -1.278 | 0.201 |
| Books and magazines (ref: yes, n = 93) | 0.079 | 0.089 | -0.098–0.252 | 0.889 | 0.374 |
Note: For this model, three outlier datapoints with a deworming frequency of 12 times/year were removed. The model was significantly different from a null model containing only an intercept term (likelihood ratio test, χ2 = 47.25, df = 15, P < 0.001). Significant P-values are printed in bold
aMultiple comparisons for the levels of ESCCAP risk group and neighbourhood using Tukey contrasts with single-step P-value adjustment were performed using the function glht from the package multcomp in R
Abbreviations: SE, standard error; CI, confidence interval; ref, reference