Table 3.

Modelling of the primary and secondary outcomes of the HYPITAT trial as a function of treatment (expectant monitoring versus labour induction), cervical length and the interaction between treatment and cervical length

Variables	Maternal high-risk situation*		Caesarean delivery*		Adverse neonatal outcome*
	Coeff. ± SE	P-value	Coeff. ± SE	P-value	Coeff. ± SE	P-value
Intercept	−0.71 ± 0.32		−2.19 ± 0.45		−2.99 ± 0.73
Treatment (EM versus LI)	−0.37 ± 0.50	0.42	−0.06 ± 0.62	0.91	0.72 ± 0.93	0.44
Cervical length (cm)	−0.03± 0.10	0.78	0.13 ± 0.14	0.32	0.09 ± 0.22	0.67
Treatment × cervical length	0.31± 0.14	0.03	0.13 ± 0.19	0.48	−0.14 ± 0.29	0.64

EM, expectant management (treatment coding = 1), LI, labour induction (treatment coding = 0), Coeff., coefficient, SE, standard error.

When woman is managed expectantly (treatment code = 1) every centimetre longer for cervical length at randomisation is associated with a 32% increase in the risk of maternal high-risk situations (exp [−0.03 + 0.31] = 1.32). When the labour is induced (treatment code = 0), every centimetre longer for cervical length at randomisation is associated with a 3% decrease in the risk of maternal high-risk situations (exp [−0.03] = 0.97).

^*All models were fit based on the Hosmer–Lemeshow goodness-of-fit test (P ≥ 0.75, high P-values indicate good fit).