Author reply

We thank Dr Anita Matai and Dr Avir Sarkar for their interest in our article and for their comments.1 Their letter gives us the opportunity to clarify several issues in relation to our work.2 The first issue is the definition of premature mortality, which in public health generally refers to deaths occurring before the average age of death in a given population.3 By analogy, we felt that deaths among adolescent girls could be considered as premature mortality. The second issue relates to the lack of information on parity. Indeed, parity is not available for all pregnancies in our database. However, for girls aged 12–18 years, first pregnancies were selected by checking whether there had been no previous pregnancy in the five previous years. With regards to sampling, no sampling was carried out for the ‘pregnant adolescents’ group, which included almost all adolescent deliveries, as almost all deliveries (99.6%) are recorded in the national hospital database.4 No sampling was carried out for the second control group either (young pregnant women aged 19–25 years). The first control group (non-pregnant adolescents) was drawn from a sample of 5% of the total French adolescent population, recorded on the French national health data system. We adopted a 1:2 matched control strategy to increase the number of subjects and to gain statistical power. An important issue is that of possible recruitment bias as a result of incomplete data. In our hospital data, information on births can be considered exhaustive.4 Information on deaths can also be considered exhaustive (provided by the national death registry). However, the causes of death were not available for the years 2016–2017. Nevertheless, the statistical models were based solely on the dates of death, and not on the causes of death, and were therefore not affected. The analyses carried out on the causes of death were descriptive and concerned only the years for which the data were available (2014–2015). To the best of our knowledge, there is therefore no major bias linked to missing data in our multivariate analyses. We are grateful to Drs Matai and Sarkar for pointing out the inconsistency between tables 1 and 3. In table 1, we can confirm that there were indeed 12 703 deliveries among adolescents aged 12–18 years. Similarly, there were 383 hospitalisations for intoxication during the 3-year follow-up for these adolescents. As requested by Drs Matai and Sarkar, we carried out a Kaplan–Meier survival analysis to better show how the differences between the groups evolve over time (Figure 1). Finally, with regards to the difficulty of drawing conclusions given the retrospective case–control design, we fully agree that the data were collected retrospectively, as is usually the case in any medico-administrative database. However, the design of the study corresponds to a cohort study, with delivery as the starting point and self-harm as the outcome. The women were followed for 3 years, until the outcome occurred or the study ended. Under these conditions, we were able to take account of the temporality of the association, as is the case in a prospective study. CQ and FJ were the coordinators of the study. KG, FJ and CQ conceived and designed the study. KG and CQ were responsible for data collection. KG and CQ accessed and verified the data. KG and JC were in charge of analyses. KG, FJ and CQ wrote the first draft. All authors were involved in the interpretation of findings, critically reviewed the first draft, and approved the final version. The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted. The project was funded by the French Ministry of Health, Direction de la recherche, des études, de l'évaluation et des statistiques (DREES), 2018. The authors thank Suzanne Rankin for editing the English language and Gwenaëlle Periard for her help with the layout and management of this article. The authors have no conflicts of interest relevant to this letter to disclose. This study follows the World Medical Association's Declaration of Helsinki. Our department’s use of these data was approved by the “Expert Committee for research, studies and evaluations in the health field” (CEREES) and the French “National Committee for data protection” (CNIL) (registration number DR-2019-021). Individual written consent was not required. Data described in the manuscript will not be made available. The database is made available by the National Health Insurance Fund (CNAM, Caisse Nationale de l'Assurance Maladie) which is responsible for the storage and extraction of the data from the French national health data system. Data used in this study are only available for researchers who meet specific criteria including training that provides personal accreditation, and approval of the protocol by required authorities (CEREES and CNIL) according to the law “Décret n° 2016-1872 du 26 décembre 2016 modifiant le décret n° 2005-1309 du 20 octobre 2005 pris pour l’application de la loi n° 78-17 du 6 janvier 1978 relative à l’informatique, aux fichiers et aux libertés”, https://www.legifrance.gouv.fr/eli/decret/2016/12/26/2016-1872/jo/texte. Contact for more information: Caisse Nationale de l’Assurance Maladie 50 Avenue du Professeur André Lemierre, 75020 Paris https://www.ameli.fr/assure/adresses-et-contacts

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.