Table 2. Goodness of fit of three mathematical models to monthly counts of cases of congenital microcephaly, Mexico, 2010–2016.

Altitudea of maternal residence	Modelb	Df	AIC	BIC	LogLik	Deviance	Χ2
							Value	P
≥ 2 200 m	Base model	4	204.8	214.9	−98.38	196.76
	With linear spline	5	201.9	214.6	−95.96	191.91	4.843c	0.027
	With cubic spline	7	201.3	219.0	−93.63	187.27	4.645d	0.098
< 2 200 m	Base model	4	162.4	172.5	−77.2	154.39
	With linear spline	5	156.2	168.9	−73.1	146.19	8.204c	0.004
	With cubic spline	7	147.0	164.7	−66.49	132.97	13.22d	0.001

AIC: Akaike information criterion; BIC: Bayesian information criterion; Df: degrees of freedom; LogLik: log-likelihood.

^a Above sea level.

^b Three different mathematical models were used to see if the incidence of congenital microcephaly in Mexico increased significantly after Zika virus was detected in the country. The base model assumed no change in the trend in incidence with time. The models that incorporated time as a linear or cubic spline allowed for changes in the trend slope at fixed points in time and for departures of that slope from one or more straight lines, respectively.

^c With one degree of freedom.

^d With two degrees of freedom.