Conception cohorts, birth cohorts and gestational age–period–cohort effects: Study design and interpretation

Most epidemiologic studies that have attempted to quantify the population effects of the Coronavirus disease 2019 (COVID-19) pandemic have been based on birth cohorts, while a few studies have utilised conception cohorts.1-4 This issue of Paediatric and Perinatal Epidemiology includes a study by Funk and colleagues,5 which examined perinatal mortality before and during the COVID-19 pandemic using the latter approach. Conception cohorts and birth cohorts quantify different aspects of multi-dimensional environmental phenomena, such as the COVID-19 pandemic, and recognising how these design options differ can aid in the interpretation of findings. In this Commentary, we discuss the utility of conception and birth cohorts for quantifying period and cohort effects and make a case for simultaneously addressing gestational age, period and cohort effects to isolate the independent effects of each (to the extent possible). By definition, conception and birth cohorts are restricted to a singular period in time with regard to cohort inception: conception cohorts are constrained by specific conception dates. In contrast, birth cohorts are limited to a specific range of birth dates. This difference, and variations in gestational duration, ensures that members of conception cohorts will have a wider range of birth dates (relative to conception dates) members of birth cohorts will have a relatively wide range of conception dates (relative to birth dates). Also of note, pandemic studies based on reconstructed conception and birth cohorts have typically not included information on early pregnancy losses because pregnancy-related events before 20 weeks gestation are not typically captured in routinely collected data and information systems. The COVID-19 pandemic affected perinatal outcomes through a variety of mechanisms, including disrupted obstetrical services and changes in the health-related behaviour of reproductive age women and pregnant women.2-4, 6, 7 Clinic closures and social distancing characterised the early months of the pandemic, while changes regarding pregnancy intention and family planning marked both the early and subsequent waves of the pandemic.2-4, 6, 7 The pandemic significantly impacted antenatal care and foetal surveillance because of reduced interaction between pregnant women and clinicians.3, 4 However, the most consequential aspects of this disruption in clinical care were typically short-lived, and obstetrical services were progressively restored in the months following pandemic onset. Another pandemic-related change involved behavioural alterations, with attendant consequences such as improved hygiene (due to an increased frequency of handwashing, for example), and reduced work-related stress among pregnant women. These changes were proposed as an explanation for the sudden decrease in pre-term birth rates observed in the early months of the pandemic. However, the hypothesis lost credence when studies showed that the changes in pre-term birth frequency were due to reductions in clinician-initiated pre-term birth and not spontaneous pre-term birth.3, 4, 8 Contraceptive use, pregnancy intention and family planning (including abortion services) were also affected by the pandemic, with countries and subpopulations (e.g. young women) affected to varying extents.7, 9, 10 Work-from-home restrictions also differentially altered fertility rates in subpopulations, increasing the fertility of some population segments and depressing the fertility of others. The study by Funk et al.5 showed pandemic-related changes in the proportion of conceptions by maternal age and socio-economic status across pre-pandemic and pandemic periods. These differences in conception cohorts, in terms of risk factors for perinatal outcomes, likely confound pre-pandemic versus pandemic comparisons of perinatal outcomes. The disruption in obstetric services, which affected antenatal care and foetal surveillance, resulted in fewer clinician-initiated early deliveries in March 2020 and in subsequent months.3, 4 Pregnancies most acutely affected by compromised prenatal care were those in the third trimester in the early stages of the pandemic when foetal surveillance and clinician-initiated early delivery could have prevented adverse perinatal outcomes. Pregnancies in the first or second trimester during the pandemic-related obstetrical services disruption were less acutely affected if foetal surveillance was restored prior to the third trimester. Thus, disruption of obstetrical care was a ‘period’ phenomenon that primarily affected pregnancies that had reached viability in the early months of the pandemic.3, 4 A birth cohort approach, with each birth cohort restricted to a reasonably small time window (e.g. a month), is thus ideal for quantifying birth-associated phenomena, such as pre-term labour induction and/or pre-term caesarean delivery. Figure 1A shows rates of pre-term labour induction and/or pre-term caesarean delivery in monthly birth cohorts in the United States from January 2016 to December 2022. An abrupt decline in rates is evident at the onset of the pandemic in March 2020, followed by a recovery from January 2021 onwards. A seasonal pattern is also highlighted in Figure 1A, with the lowest rates of pre-term labour induction and/or pre-term caesarean delivery occurring in September (see explanation below). Utilising a conception cohort approach for quantifying the disruption in obstetrical care provides a less clear picture of this period phenomenon. Figure 1B presents the monthly conception cohort-based rates of pre-term labour induction and/or pre-term caesarean delivery in the United States from May 2015 to February 2022. A larger variability in rates is evident, which reflects the conception cohort's wider range of birth dates. An abrupt decline in pre-term labour induction and/or pre-term caesarean delivery is also observed in this analysis, although as expected, the phenomenon first manifests with the conception cohort of June 2019. The COVID-19 pandemic led to a brief closure of many fertility clinics and affected women planning to conceive through assisted reproductive technologies (ART).2 These clinic closures occurred in March, April and May 2020 and resulted in a sharp reduction in the proportion of ART-conceived pregnancies during these months. A conception cohort approach, with each conception cohort restricted to a reasonably small time window (e.g. a month), is thus ideal for quantifying the effects of this conception-associated phenomenon. Figure 2A shows rates of ART-conceived pregnancies in monthly conception cohorts in the United States from mid-2015 to early 2022. There is a seasonality in rate patterns, with low rates in December and an abrupt, large decline in rates in March, April and May 2020. The low ART rates in the December conception cohorts are attributable to ART clinic closures associated with the holiday season, and the even lower ART rates in the conception cohorts of March–May 2020 are due to the pandemic-associated ART clinic closures. Utilising a birth cohort approach for quantifying the effects of ART clinic closures reveals a less clear picture with regard to ART-conceived pregnancies (Figure 2B). The patterns highlighted by the conception cohort approach are also seen in this birth cohort-based analysis, although the effects are more dispersed. Thus, the seasonal clinic closures in December manifest in lower rates of ART-conceived pregnancies in September and, to a lesser extent, in August (of the subsequent year), while effects of pandemic-associated clinic closures are evident over an extended period between August 2020 and February 2021. Figure S1 highlights how ART clinic closures altered the proportion of twin births in monthly birth and conception cohorts in the United States. Note also that the pre-term labour induction and/or pre-term caesarean delivery rates in September (Figure 1A) reflect the closure of ART clinics in the previous December holiday season. Some aspects of the health services disruption during the early pandemic, such as the reductions in clinician-initiated early delivery, constituted a period effect primarily.3, 4, 8 Other pandemic-related changes, such as the closure of ART clinics (which altered the plurality composition of conception cohorts in March–May 2020), and changes in the pregnancy intentions of specific subpopulations (which altered the maternal age, socio-economic status and other features of conception cohorts during the pandemic, as shown by Funk et al.5) resulted in cohort effects. Additionally, a potential effect of changes in the gestational age distribution at birth cannot be excluded. Outcomes such as perinatal mortality were likely affected by gestational age, period and cohort effects during the COVID-19 pandemic. While the conception cohort study by Funk et al.,5 and previous birth cohort studies3, 4, 8 offer valuable insights into COVID-19 pandemic effects, there may be value in simultaneously examining the multi-dimensional gestational age, period and cohort effects of the pandemic on perinatal outcomes. Pandemic studies, which have employed either conception or birth cohort approaches, highlight the general absence of perinatal research that has simultaneously addressed gestational age, period and cohort effects. Although some perinatal phenomena are likely to be primarily influenced by a singular effect, many perinatal outcomes in the population will likely result from the action of more than one effect. Pre-term birth rates are a case in point: clinician-initiated preterm birth is likely to be primarily a period effect, whereas spontaneous pre-term birth is likely to follow a cohort pattern. The overlap between these effects can alter findings and inferences, and attempts to isolate each distinct effect (to the extent possible) may be revealing in specific circumstances. We look forward to future studies that comprehensively examine the gestational age–period–cohort effects of the COVID-19 pandemic, and other determinants of relevant perinatal outcomes. Sarka Lisonkova is an Associate Professor at the University of British Columbia. Her research interests include maternal, foetal, and neonatal health and health services, and her research focuses on adverse perinatal outcomes among high-risk and vulnerable women. Bahi Fayek is a Research Fellow in the Division of Reproductive Endocrinology and Infertility at the University of British Columbia. His research interests include recurrent pregnancy loss and endometriosis. K. S. Joseph is a Professor at the University of British Columbia. His research interests include pregnancy complications, preterm birth, foetal growth, perinatal mortality, serious neonatal morbidity, severe maternal morbidity, and maternal mortality. He serves on the Editorial Board of Paediatric and Perinatal Epidemiology. SL: Conceptualisation, Manuscript Writing, Editing, Funding Acquisition. BF: Conceptualisation, Manuscript Writing, Editing. KSJ: Conceptualisation, Data Analysis, Manuscript Writing, Editing. SL was supported by the Canadian Institute of Health Research. KSJ is supported by an Investigator award from the BC Children’s Hospital Research Institute. None declared. Data used in this study and information on data elements are publicly available at https://www.cdc.gov/nchs/data_access/vitalstatsonline.html. Figure S1. Data S1. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.