Causes of ART‐related outcomes in the COVID‐19 era

The challenges of time for studies on the population effects of the COVID‐19 pandemic on perinatal outcomes

“…Both methodological approaches differ from those in the usual perinatal epidemiology toolbox, and their increasing use in COVID‐19 research holds more general lessons for research in our field. For instance, an assumption in before‐after designs (termed the ‘stacked calendar’ approach by Bruckner and Gemmill 4 ) is that it is possible to control for time trends by adjusting for changes in important population characteristics such as maternal age, body mass index, educational level or smoking. However, for preterm birth rates, this approach does not explain all changes over time, as seen by widely contrasting contemporaneous preterm birth trends in high‐income countries despite similar evolution of these sociodemographic characteristics 5 .…”

“…Inaccurate modelling of time trends may be one explanation for the contradictory results that fuel continued debate about the existence, magnitude, and reasons for preterm birth rate decreases during the pandemic. Other issues, raised in the commentaries, include the management of outliers and derivation of confidence intervals when evaluating time series 4 …”

“…Other issues, raised in the commentaries, include the management of outliers and derivation of confidence intervals when evaluating time series. 4 The use of conception versus birth cohorts, a topic addressed by all the studies in this issue, also challenges perinatal epidemiologists to refine our concepts of the population at risk. A pandemic is a major disruption that calls into question the underlying assumptions of stable fertility trends and childbearing patterns on which we silently rely when using birth cohort denominators.…”

“…The temporal dimension of research on the COVID‐19 pandemic and pregnancy outcomes poses multiple challenges for perinatal epidemiologists. This issue of Paediatric and Perinatal Epidemiology features two studies and associated commentaries 1 , 2 , 3 , 4 that investigate population‐level changes in preterm birth and in vitro fertilisation (IVF) births related to the COVID‐19 pandemic and speak to two of these challenges: using appropriate methods for time trend analyses, that is, time series as opposed to before‐after designs, and appropriate denominators to capture time‐varying exposures and potential changes in the characteristics of the childbearing population, that is, conception versus birth cohort designs.…”

“…Other issues, raised in the commentaries, include the management of outliers and derivation of confidence intervals when evaluating time series. 4 …”