Community-led intensive trapping reduces abundance of key plague reservoir and flea vector

Table 3 Final competitive model(s) for rodent relative abundance inside houses during the intervention period (July 2022-March 2023)

Level	Variable	Rattus rattus relative abundance^a				Mus musculus relative abundance^a
Level	Variable	Captures (model 1^b)	Captures (model 2^b)	Tracking tiles (model 1^c)	Tracking tiles (model 2^c)	Captures (model 1^d)	Captures (model 2^d)	Tracking tiles (model 1^c)
Intervention period^e	Intercept	0.08 (0.06–0.1)	0.09 (0.06–0.12)	1.43 (0.85–2.4)	2.08 (1.09–4.0)	0.17 (0.13–0.22)	0.19 (0.14–0.26)	1.64 (1.16–2.31)
	Treatment (yes)	0.31 (0.20–0.49)	0.31 (0.20–0.49)	0.32 (0.18–0.58)	0.16 (0.06–0.4)
	Month (Nov 2022)		0.7 (0.53–0.92)	0.38 (0.21–0.68)	0.27 (0.12–0.62)		0.94 (0.70–1.26)	2.28 (1.35–3.85)
	Month (March 2023)		0.88 (0.63–1.24)	0.55 (0.32–0.93)	0.27 (0.12–0.6)		0.62 (0.41–0.93)	0.61 (0.39–0.96)
	Treatment (yes)* Month (Nov 2022)				1.72 (0.54–5.47)
	Treatment (yes)* Month (March 2023)				3.99 (1.34–11.86)
Random effects^f	House: Sampling area: Village	0.92 (0.96)	0.91 (0.96)	1.36 (1.17)	1.5 (1.22)	< 0.001 (< 0.001)	< 0.001 (< 0.001)	1.57 (0.4)
	Sampling area: Village	0.02 (0.15)	0.03 (0.17)	0.03 (0.17)	0.05 (0.22)	0.03 (0.2)	0.03 (0.2)	< 0.001 (< 0.001)
	Village	< 0.001 (< 0.001)	< 0.001 (0.006)	< 0.001 (< 0.001)	< 0.001 (< 0.001)	0.11 (0.34)	0.12 (0.34)	< 0.001 (< 0.001)

^aWe present odd ratios and 85% confidence intervals from the final competitive model(s), and variance and standard deviation of random effects. Final competitive models for each analysis included the model with the lowest AIC and any models with an AIC within 2, excluding models which only differed by the inclusion of uninformative parameters based on 85% confidence intervals (Leroux SJ. On the prevalence of uninformative parameters in statistical models applying model selection in applied ecology. PLoS One. 2019;14:p.e0206711. https://doiorg.publicaciones.saludcastillayleon.es/10.1371/journal.pone.0206711). Informative parameters are highlighted in bold
^bModels used a Poisson distribution
^cModels used a binomial distribution
^dModels used a negative binomial distribution
^eIn order, the intervention period row includes odds ratios and 85% confidence intervals (in brackets) for each variable
^f In order, the random effects row displays the variance and standard deviation (in brackets) for each variable

ISSN: 1349-4147