To the Editor In the retrospective study of the association of repeat tetanus, diphtheria, and acellular pertussis (Tdap) vaccination in pregnancy and adverse birth outcomes by Dr Sukumaran and colleagues,1 we noticed potential measurement bias and confounding, which require careful interpretation of the results.
To identify gestational age, Sukumaran and colleagues1 used a validated algorithm with the accuracy within 30 days for 98% of live births.2 This 30-day margin has the potential to misclassify preterm and full-term deliveries nondifferentially, which might have biased the study results toward no significance. Furthermore, because of the low prevalence of preterm delivery,3 the validated algorithm can still accurately estimate gestational age for a high percentage of live births even if the accuracy for the estimates of preterm delivery is low. To illustrate, if the algorithm estimated the gestational age of all the full-term deliveries correctly within 30 days in this study, it would have already achieved 93% accuracy because preterm deliveries only accounted for approximately 7% of all live births.1 Therefore, it is important to report the sensitivity and specificity of the algorithm for gestational age estimate of preterm delivery specifically to address concerns about measurement bias.
Zhu Y, van Boemmel-Wegmann S, Albogami Y. Repeat Tdap Vaccination and Adverse Birth Outcomes. JAMA. 2016;315(12):1285-1286. doi:10.1001/jama.2015.19078