Women’s experience of unintended pregnancy and changes in contraceptive methods: evidence from a nationally representative survey

Table 3 Multilevel multinomial logistic regression model with pregnancy intention as the sole correlate of the change in contraceptive methods use, from before to after the most recent pregnancy

	Changes in the contraceptive methods before to after pregnancy (reference: switched to a higher effective method)
	No change in methods		Switched to a less effective methods
	Odds ratio (95% CI)	p	Odds ratio (95% CI)	p
Most recent pregnancy intention at conception: wanted conception (ref)	1.00	–	1.00	–
Mistimed conception	2.09 (1.65–2.65)	< 0.01	1.92 (1.38–2.68)	< 0.01
Unwanted conception	2.12 (1.54–2.91)	< 0.01	2.71 (1.80–4.06)	< 0.01
Constant	2.17 (1.99–2.36)	< 0.01	0.44 (0.39–0.50)	< 0.01
Random effects^a
Cluster-level variance (SE)^b	0.08 (0.05)***
Log-likelihood for fixed effects to random effects model	1423.29***
Log-likelihood ratio test for the null model to the random effects model (Chi-square)^c	62.97***

^aWe assume that the within cluster-level random effects are equal for the ‘no change of contraceptive method’ and ‘switch to a less effective contraceptive methods’; therefore, only between cluster-level variance estimates are reported
^bSignificance of random effects evaluated by comparing the model with a similar one in which random effects were constrained to zero
^cCompared to the null model with no-covariates. ***p < 0.01

ISSN: 1742-4755