Measurement error, microcephaly prevalence and implications for Zika: an analysis of Uruguay perinatal data

Abstract

Background and objectiveThe Zika virus outbreak has drawn attention to microcephaly, whose definition is based on head circumference measuring below a percentile or number of SDs below the mean. The objective of this analysis was to assess how differences in measurement precision might affect prevalence and trends of microcephaly.MethodsData from all births in Uruguay during 2010–2015 were obtained from the Perinatal Information System. The prevalence of births with microcephaly was calculated based on head circumference measurement at birth applying the INTERGROWTH-21st standards for sex and gestational age, and compared by method of ascertaining gestational age.ResultsRounding and digit preference was observed: 74% of head circumference measurements were reported as a whole centimetre value. The prevalence of births varied substantially by the criterion used to define microcephaly (<3 SD, <2 SD, <3rd percentile for gestational age) and could be halved or doubled based on adding or subtracting a half-centimetre from all reported head circumference measurements. If 4 days were added to gestational age calculations, rather than using completed gestational weeks (without days) for gestational age reporting, the prevalence was 1.7–2 times higher.DiscussionRounding in measurement of head circumference and reporting preferences of gestational age may have contributed to a lower prevalence of microcephaly than expected in this population. Differences in head circumference measurement protocols and gestational age dating have the potential to affect the prevalence of babies reported with microcephaly, and this limitation should be acknowledged when interpreting head circumference data collected for surveillance.