Background Early exposure to research methodology is essential in medical education, yet many students show limited motivation to engage with nonclinical content. Gamified strategies such as educational escape rooms may help improve engagement, but few studies have explored their feasibility at scale or evaluated their impact beyond student satisfaction. Objective This study aimed to assess the feasibility, engagement, and perceived educational value of a large-scale escape room specifically designed to introduce third-year medical students to the principles of diagnostic test evaluation. Methods We developed a low-cost immersive escape room based on a fictional diagnostic accuracy study with 6 puzzles mapped to five predefined learning objectives: (1) identifying key components of a diagnostic study protocol, (2) selecting an appropriate gold standard test, (3) defining a relevant study population, (4) building and interpreting a contingency table, and (5) critically appraising diagnostic metrics in context. The intervention was deployed to an entire class of third-year medical students across 12 sessions between March 2023 and April 2023. Each session included 60 minutes of gameplay and a 45-minute debriefing. Students completed pre- and postintervention questionnaires assessing their knowledge of diagnostic test evaluation and perceptions of research training. Descriptive statistics and 2-tailed paired t tests were used to evaluate score changes; univariate linear regressions assessed associations with demographics. Free-text comments were analyzed using the hierarchical classification by Reinert. Results Of the 530 participants, 490 (92.5%) completed the full evaluation. Many participants had had limited previous exposure to escape rooms (206/490, 42% had never participated in one), and most (253/490, 51.6%) reported low initial confidence with critical appraisal of scientific articles. Mean overall knowledge scores increased from 62 of 100 (SD 1) before to 82 of 100 (SD 2) after the activity (+32%; P<.001). Gains were observed across all learning objectives and were not influenced by age, sex, or previous experience. Students rated the educational escape room as highly entertaining (mean score 9.1/10, SD 1.1) and educational (mean score 8.2/10, SD 1.5). Following the intervention, 86.9% (393/452) felt more comfortable with critical appraisal of diagnostic test studies, and 79% (357/452) considered the escape room format highly appropriate for an introductory session. Conclusions This study demonstrates the feasibility and enthusiastic reception of a large-scale, reusable escape room aimed at teaching the fundamental principles of diagnostic test evaluation to undergraduate medical students. This approach may serve as a valuable entry point to engage students with evidence-based reasoning and pave the way for deeper exploration of medical research methodology.

Couplet Care ensures zero separation between mother and baby from birth, fostering attachment, particularly for hospitalized premature or full-term infants. Implementation of mother-child paired care was initiated in 2017 at our center. This study aimed to quantify parental presence during hospitalization and identify its determinants. METHOD:This retrospective study analyzed data from September to December 2023, recorded by healthcare staff. We included in the study preterm and full-term infants admitted to our level II neonatal unit during this period. Parental involvement was confirmed by at least one of three factors between 0 and 6 am: breastfeeding, parental care, or skin-to-skin contact. Continuous parental presence was defined as 100 % of nights with parental involvement. RESULTS:Of the 110 children admitted during this period, 37 were excluded, primarily due to admission from home. Among 73 included newborns, 46 (63 %) had parental presence every night. These infants (Permanent Parental Presence-PPP group) had a higher mean gestational age at birth than those with Discontinuous Parental Presence (DPP) (36.3 vs. 31.3 weeks, p < 0.0001). More PPP infants were inborn (80.4 % vs. 51.9 %, p = 0.01). Their hospital stay was shorter (15.9 vs. 45.6 days, p < 0.0001). Breastfeeding rate at admission and at discharge was significantly higher in the PPP group (respectively 91 % vs 74 %, p = 0.05 and 84.7 % vs. 51.8 %, p = 0.002). Distance between home and the neonatal unit was smaller in the PPP group (6.5 km vs 8.8 km, p = 0.03). Multiparity showed no impact on parental presence. CONCLUSION:Couplet Care and single-room accommodations encourage continuous parental presence in neonatology. Different ways to support parents' continuous presence in order to promote parent-child closeness should be promoted.

BACKGROUND:Script Concordance Tests (SCTs) represent an innovative assessment method which have been introduced in the 2024 French National Ranking Examinations (EDN). These tests compare a student's clinical reasoning with that of a panel of experts under conditions of uncertainty. Typically, the question involves the impact of new information on an initially proposed hypothesis, with answers given on a Likert scale. MAIN FINDINGS:This article aims to didactically illustrate how SCTs are consistent with probabilistic reasoning as modeled by Bayes' theorem. In addition, by comparing SCT writing guidelines with Bayesian reasoning concepts, several ambiguities were identified: (1) What stage of clinical reasoning do SCTs evaluate? (2) What are the appropriate labels for Likert scale responses? (3) Does the expert panel provide a relevant reference for SCTs? CONCLUSIONS:Currently, many of these questions remain unanswered in the literature, with recent data suggesting that experienced physicians' responses to SCTs are often biased. Beyond their use as an assessment tool in the EDN, SCTs offer a valuable opportunity to develop and deepen the teaching of probabilistic reasoning in medical education and serve as a potential area of research to improve clinical practice.

BackgroundMachine-learning methods are gaining in popularity to predict medical events but their added value to other methods is still to be determined. We compared performances of clinical prediction models for bronchopulmonary dysplasia (BPD) or death in very preterm infants using logistic regression and random forests methods.MethodsTwo population-based cohorts of very preterm infants were used: EPIPAGE-2 (France, 2011) for development and internal validation and EPICE (Europe, 2011) for external validation. Eligible infants were born before 30 weeks' gestation and admitted in neonatal units. BPD was defined as any respiratory support at 36 weeks postmenstrual age. Candidate predictors were available shortly after birth or at day 3. Logistic regression and random forest models performance was assessed in terms of discrimination (c-statistic) and calibration plots.ResultsPrevalence of BPD/death was 32.1% (668/1923) in EPIPAGE-2 and 41.0% (1368/3335) in EPICE. At both time points, logistic regression and random forest models showed similar performance during internal validation. At birth, external validation in EPICE showed good discrimination (logistic regression model: c-statistics 0.81, 95% CI 0.80-0.83; random forest: 0.80, 95% CI 0.79-0.81) but both models underestimated the probability of BPD/death. Model performances were heterogeneous throughout European regions.ConclusionsBoth modelling methods performed similarly to predict BPD/death shortly after birth in very preterm children.ImpactWhether machine-learning methods predict better short-term respiratory outcomes in very preterm infants than logistic regression models is debated.Random forest-based prediction models did not perform better than logistic regression to predict bronchopulmonary dysplasia or death shortly after birth in very preterm infants.Calibration performances varied among European countries.While offering the same performance, regression models are easier to understand, to disseminate and to apply to different populations.

ObjectiveWe aimed to study neurodevelopmental outcomes and healthcare utilisation at age 5-6 years in very preterm children with bronchopulmonary dysplasia (BPD). DesignProspective and national population-based study. SettingAll the neonatal units in 25 French regions (21 of the 22 metropolitan regions and 4 overseas regions). PatientsChildren born before 32 weeks' gestation in 2011. InterventionsBlind, comprehensive and standardised assessment by trained neuropsychologists and paediatricians at age 5-6 years. Main outcome measuresOverall neurodevelopmental disabilities, behavioural difficulties, developmental coordination disorders, full-scale IQ, cerebral palsy, social interaction disorders, rehospitalisation in the previous 12 months and detailed developmental support. ResultsOf the 3186 children included, 413 (11.7%) had BPD. The median gestational age of children with BPD was 27 weeks (IQR 26.0-28.0) and without BPD was 30 weeks (28.0-31.0). At age 5-6 years, 3150 children were alive; 1914 (60.8%) had a complete assessment. BPD was strongly associated with mild, moderate and severe overall neurodevelopmental disabilities (OR 1.49, 95% CI 1.05 to 2.20; 2.20, 1.41 to 3.42 and 2.71, 1.67 to 4.40). BPD was associated with developmental coordination disorders, behavioural difficulties, lower IQ score as well as rehospitalisation in the last 12 months and developmental support. The association between BPD and cerebral palsy was statistically significant before adjustment but not in adjusted analyses. ConclusionsBPD was strongly and independently associated with many neurodevelopmental disabilities. Improving medical and neurodevelopmental management of BPD in very preterm children should be a priority to reduce its long-term consequences.

The Script Concordance Tests (SCTs) are an examination modality introduced by decree in the French National Ranking Exam for medical students in 2024. Their objective is to evaluate clinical reasoning in situations of uncertainty. In practice, SCTs assess the impact of new information on the probability of a hypothesis formulated a priori based on an authentic clinical scenario. This approach resembles probabilistic (or Bayesian) reasoning. Due to the uncertainty associated with the explored clinical situation, SCTs do not compare the student's response to an expected one in a theoretical knowledge reference. Instead, the distribution of responses from a panel of experienced physicians is used to establish the question's scoring scale. Literature data suggest that physicians, even experienced ones, like most humans, often exhibit biased intuitive probabilistic reasoning. These biases raise questions about the relevance of using expert panel responses as scoring scales for SCTs.

Les tests de concordance de script (TCS) constituent une modalité docimologique introduite par arrêté dans les épreuves dématérialisées en 2024. Leur objectif serait d’évaluer le raisonnement clinique en situation d’incertitude. En pratique, les TCS interrogent l’étudiant sur l’impact d’une nouvelle information sur la probabilité d’une hypothèse formulée a priori à partir d’une situation clinique authentique. Cette démarche s’apparente au raisonnement probabiliste (ou bayésien). En raison de l’incertitude liée à la situation clinique explorée, les TCS ne confrontent pas la réponse de l’étudiant à celle attendue dans un référentiel de connaissances théoriques. C’est la distribution des réponses d’un panel de médecins expérimentés qui sert à établir le barème de la question. Les données de la littérature suggèrent que les médecins, même expérimentés, comme la plupart des humains ont un raisonnement probabiliste intuitif le plus souvent biaisé. Ces biais posent la question de la pertinence d’utiliser les réponses du panel d’experts comme barème de notation pour les TCS.

To describe the circumstances, causes and timing of death in extremely preterm infants. We included from the EPIPAGE-2 study infants born at 24–26 weeks in 2011 admitted to neonatal intensive care units (NICU). Vital status and circumstances of death were used to define three groups of infants: alive at discharge, death with or without withholding or withdrawing life-sustaining treatment (WWLST). The main cause of death was classified as respiratory disease, necrotizing enterocolitis, infection, central nervous system (CNS) injury, other or unknown. Among 768 infants admitted to NICU, 224 died among which 89 died without WWLST and 135 with WWLST. The main causes of death were respiratory disease (38%), CNS injury (30%) and infection (12%). Among the infants who died with WWLST, CNS injury was the main cause of death (47%), whereas respiratory disease (56%) and infection (20%) were the main causes in case of death without WWLST. Half (51%) of all deaths occurred within the first 7 days of life, and 35% occurred within 8 and 28 days. The death of extremely preterm infants in NICU is a complex phenomenon in which the circumstances and causes of death are intertwined.

Les bêta-lactamines, en particulier l’amoxicilline, sont les antibiotiques les plus prescrits en pédiatrie. Le médecin de l’enfant ou l’urgentiste sont parfois conduits à suspecter une allergie à l’amoxicilline lorsqu’une éruption cutanée survient en cours de traitement. Cependant, cette allergie n’est confirmée que dans 1 cas sur 10. On classe les réactions d’hypersensibilité selon leur chronologie et leur sévérité. L’analyse sémiologique permet d’argumenter l’imputabilité de l’antibiotique, complétée de tests cutanés allergologiques dans les formes immédiates et/ou sévères. Selon les cas, une réintroduction de l’antibiotique peut être proposée afin d’infirmer ou de confirmer le diagnostic d’hypersensibilité. Cette démarche rigoureuse est indispensable afin de retirer l’étiquette « allergique » à la majeure partie des enfants suspects, et ne retenir que les rares cas d’hypersensibilité justifiant la contre-indication de tout ou d’une partie des bêta-lactamines.Beta-lactams, particularly amoxicillin, are the most commonly prescribed antibiotics in pediatrics. The pediatrician or emergency doctor may suspect an allergy to amoxicillin when a skin rash occurs during treatment. However, this allergy is only confirmed in 1 out of 10 cases. Hypersensitivity reactions are classified according to their timing and severity. The semiological analysis argues for the imputability of the antibiotic and is completed by allergological skin tests in immediate and/or severe forms. Depending on the case, reintroduction of the antibiotic may be proposed in order to invalidate or confirm the diagnosis of hypersensitivity. This rigorous approach is essential in order to remove the label of “allergic” from the majority of suspected cases, and to select only rare cases of hypersensitivity that justify a contraindication to all or some of the beta-lactams.

Aim: To describe the circumstances, causes and timing of death in extremely preterm infants.Methods: We included from the EPIPAGE-2 study infants born at 24-26 weeks in 2011 admitted to neonatal intensive care units (NICU). Vital status and circumstances of death were used to define three groups of infants: alive at discharge, death with or without withholding or withdrawing life-sustaining treatment (WWLST). The main cause of death was classified as respiratory disease, necrotizing enterocolitis, infection, central nervous system (CNS) injury, other or unknown.Results: Among 768 infants admitted to NICU, 224 died among which 89 died without WWLST and 135 with WWLST. The main causes of death were respiratory disease (38%), CNS injury (30%) and infection (12%). Among the infants who died with WWLST, CNS injury was the main cause of death (47%), whereas respiratory disease (56%) and infection (20%) were the main causes in case of death without WWLST. Half (51%) of all deaths occurred within the first 7 days of life, and 35% occurred within 8 and 28 days.Conclusion: The death of extremely preterm infants in NICU is a complex phenomenon in which the circumstances and causes of death are intertwined.

Introduction: The choice of diagnostic tests in front of a given clinical case is a major part of medical reasoning. Failure to prescribe the right test can lead to serious diagnostic errors. Furthermore, unnecessary medical tests are a waste of money and could possibly generate injuries to patients, especially in family medicine. Methods: In an effort to improve the training of our students to the choice of laboratory and imaging studies, we implemented a specific multiple-choice questions (MCQ), called comprehensive MCQ (cMCQ), with a fixed and high number of options matching various basic medical tests, followed by a certainty-based mark (CBM). This tool was used in the assessment of diagnostic test choice in various clinical cases of general practice in 456 sixth-year medical students. Results: The scores were significantly correlated with the traditional exams (standard MCQ), with matched themes. The proportion of “cMCQ/CBM score” variance explained by “standard MCQ score” was 21.3%. The cMCQ placed students in a situation closer to practice reality than standard MCQ. In addition to its usefulness as an assessment tool, those tests had a formative value and allowed students to work on their ability to measure their doubt/certainty in order to develop a reflexive approach, required for their future professional practice. Conclusion: cMCQ followed by CBM is a feasible and reliable evaluation method for the assessment of diagnostic testing.

Background: Teaching about craniofacial traumas is challenging given the complexity of the craniofacial anatomy and the necessity for good spatial representation skills. To solve these problems, three-dimensional printing seems to be an appropriate educative material. In this study, the authors conducted a randomized controlled trial. The authors' main objective was to compare the performance of the undergraduate medical students in an examination based on the teaching support: three-dimensionally printed models versus two-dimensional pictures. Methods: All participants were randomly assigned to one of two groups using a random number table: the three-dimensionally-printed support group (three-dimensional group) or the two-dimensionally-displayed support group (two-dimensional group). All participants completed a multiple-choice question evaluation questionnaire on facial traumatology (first, a zygomatic bone fracture; then, a double mandible fracture). Sex and potential confounding factors were evaluated. Results: Four hundred thirty-two fifth-year undergraduate medical students were enrolled in this study. Two hundred six students were allocated to the three-dimensional group, and 226 were allocated to the two-dimensional group. The three-dimensionally printed model was considered to be a better teaching material compared with two-dimensional support. The global mean score was 2.36 in the three-dimensional group versus 1.99 in the two-dimensional group (p = 0.008). Regarding teaching of biomechanical aspects, three-dimensionally-printed models provide better understanding (p = 0.015). Participants in both groups exhibited similar previous student educational achievements and visuospatial skills. Conclusions: This prospective, randomized, controlled educational trial demonstrated that incorporation of three-dimensionally-printed models improves medical students' understanding. This trial reinforces previous studies highlighting academic benefits in using three-dimensionally-printed models mostly in the field of understanding complex structures.

Background: Postnatal steroids (PNS) have been used to prevent bronchopulmonary dysplasia (BPD) in preterm infants but have potential adverse effects on neurodevelopment. These effects might be modulated by their risk of BPD. We aimed to compare patients’ neurodevelopment with PNS treatment according to their risk of BPD in a European cohort. Methods: We developed a prediction model for BPD to classify infants born between 24 + 0 and 29 + 6 weeks of gestation in three groups and compared patients’ neurological outcome at two years of corrected age using the propensity score (PS) method. Results: Of 3662 neonates included in the analysis, 901 (24.6%) were diagnosed with BPD. Our prediction model for BPD had an area under the ROC curve of 0.82. In the group with the highest risk of developing BPD, PNS were associated with an increased risk of gross motor impairment: OR of 1.95 after IPTW adjustment (95% CI 1.18 to 3.24, p = 0.010). This difference existed regardless of the type of steroid used. However, there was an increased risk of cognitive anomalies for patients treated with dexa/betamethasone that was no longer observed with hydrocortisone. Conclusions: This study suggests that PNS might be associated with an increased risk of gross motor impairment regardless of the group risk for BPD. Further randomised controlled trials exploring the use of PNS to prevent BPD should include a risk-based evaluation of neurodevelopmental outcomes. This observation still needs to be confirmed in a randomised controlled trial.

Although well documented in randomised trials, the efficacy of prophylaxis against respiratory syncytial virus (RSV) in real-word conditions is less studied. The objective was to assess the impact of partial versus full RSV prophylaxis for acute respiratory infections (ARIs) and ARI-related hospital admissions in preterm children. This study included children born preterm in 2011 in France who were eligible for RSV prophylaxis and received at least one palivizumab dose from October 2011 to March 2012. Full prophylaxis was defined as receiving at least one palivizumab dose for each month of RSV exposure in the community. Children with full and partial prophylaxis were matched, and odds of ARIs and ARI-related hospital admission were compared by logistic regression. Full prophylaxis concerned 861/1083 (80%) children. As compared with full prophylaxis, partial prophylaxis was not associated with ARI occurrence (odds ratio OR 1.3, 95% confidence interval CI 0.9-1.9) but was significantly associated with ARI-related hospital admission during the RSV epidemic (OR 1.9, 95% CI 1.2-2.9). During the 2011-2012 RSV epidemic, hospital admission rates were higher for preterm children with partial than full RSV prophylaxis. Improving compliance could help alleviate the burden of RSV on healthcare systems.

BACKGROUND:Regionalisation programmes aim to ensure that very preterm infants are born in level III units (inborn) through antenatal referral or transfer. Despite widespread knowledge about better survival without disability for inborn babies, 10%-30% of women deliver outside these units (outborn).OBJECTIVE:To investigate risk factors associated with outborn deliveries and to estimate the proportion that were probably or possibly avoidable.METHODS:We used a national French population-based cohort including 2205 women who delivered between 24 and 30+6  weeks in 2011. We examined risk factors for outborn delivery related to medical complications, antenatal care, sociodemographic characteristics and living far from a level III unit using multivariable binomial regression. Avoidable outborn deliveries were defined by pregnancy risk (obstetric history, antenatal hospitalisation) and time available for transfer.RESULTS:25.0% of women were initially booked in level III, 9.1% were referred, 49.8% were transferred, and 16.1% had outborn delivery. Risk factors for outborn delivery were gestational age <26 weeks (adjusted relative risk (aRR) 1.37, 95% confidence interval (CI) 1.13, 1.66), inadequate antenatal care (aRR 1.39, 95% CI 1.10, 1.81), placental abruption (aRR 1.66, 95% CI 1.27, 2.17), and increased distance to the closest level III unit ((aRR 2.79, 95% CI 2.00, 3.92) in the 4th versus 1st distance quartile). Among outborn deliveries, 16.7% were probably avoidable, and 25.6% possibly avoidable, which could increase the proportion of inborn deliveries between 85.9% and 92.9%. Avoidable outborn deliveries were mainly associated with gestational age, intrauterine growth restriction, preterm premature rupture of membranes, and haemorrhage, but not distance.CONCLUSIONS:Our study identified some modifiable risk factors for outborn delivery; however, when regionalised care relies heavily on antenatal transfer, as it does in France, only some outborn deliveries may be prevented. Earlier referral of high-risk women will be needed to achieve full access to tertiary care.

AIM:Although well documented in randomised trials, the efficacy of prophylaxis against respiratory syncytial virus (RSV) in real-word conditions is less studied. The objective was to assess the impact of partial versus full RSV prophylaxis for acute respiratory infections (ARIs) and ARI-related hospital admissions in preterm children.METHODS:This study included children born preterm in 2011 in France who were eligible for RSV prophylaxis and received at least one palivizumab dose from October 2011 to March 2012. Full prophylaxis was defined as receiving at least one palivizumab dose for each month of RSV exposure in the community. Children with full and partial prophylaxis were matched, and odds of ARIs and ARI-related hospital admission were compared by logistic regression.RESULTS:Full prophylaxis concerned 861/1083 (80%) children. As compared with full prophylaxis, partial prophylaxis was not associated with ARI occurrence (odds ratio OR 1.3, 95% confidence interval CI 0.9-1.9) but was significantly associated with ARI-related hospital admission during the RSV epidemic (OR 1.9, 95% CI 1.2-2.9).CONCLUSION:During the 2011-2012 RSV epidemic, hospital admission rates were higher for preterm children with partial than full RSV prophylaxis. Improving compliance could help alleviate the burden of RSV on healthcare systems.

La corticothérapie postnatale (CPN) est utilisée chez le grand prématuré pour réduire l’évolution vers la dysplasie bronchopulmonaire (DBP) malgré les effets secondaires auxquels ce traitement expose, notamment neurologiques. Certains auteurs ont suggéré que la balance bénéfice/risque de la CPN pourrait être favorable chez les enfants ayant un risque initial élevé de développer une DBP. Notre objectif était de comparer le développement neurologique à deux ans d’âge corrigé des patients traités par CPN avec les patients non traités selon leur risque initial de DBP. Les patients inclus faisaient partie de la cohorte européenne EPICE et étaient nés entre 24 et 29 semaines d’aménorrhée (SA). Nous avons élaboré un modèle de prédiction du risque de DBP qui était appliqué à 14 jours de vie. Nous avons ensuite évalué la capacité discriminante et la bonne calibration de ce modèle. Trois groupes de risque de DBP ont ainsi pu être définis. Nous avons comparé le développement neurologique moteur et cognitif à deux ans d’âge corrigé des patients traités par CPN par rapport à celui des patients non traités dans la population globale puis en les distinguant selon leur groupe de risque de DBP. Les facteurs de confusion liés à la gravité respiratoire et neurologique des patients étaient pris en compte dans un score de propension. Un total de 3662 patients nés entre 24 + 0 et 29 + 6SA et vivants à 14 jours de vie étaient inclus dans notre travail. Parmi eux, 25% ont développé une DBP. Huit variables étaient retenues dans notre score de prédiction de la DBP. L’AUC de notre modèle de prédiction était de 0.82 avec une bonne calibration. Dans la population globale, à deux ans d’âge corrigé, les patients traités par CPN présentaient plus d’anomalies de la motricité globale après pondération inverse sur le score de propension (OR = 1.77, IC95% [1.13;2.77], p = 0.013). Cette différence était significative, même dans le groupe à haut risque de DBP (OR = 1.89, IC95% [1.07;3.33], p = 0 .028). Nous n’avons pas observé de différence de développement cognitif chez les patients traités par CPN par rapport aux non traités. Nos résultats vont à l’encontre de l’hypothèse d’une balance bénéfice/risque positive de la CPN dans le groupe à risque plus élevé de DBP. Approbation par le comité d’éthique de chaque centre participant à la cohorte EPICE, avis favorable du CCTIRS et autorisation de la CNIL en France.

Introduction: Postnatal corticosteroids (PNC) are effective for reducing bronchopulmonary dysplasia (BPD) in very preterm neonates but are associated with adverse effects including an increased risk of cerebral palsy. PNC use in Europe is heterogeneous across regions. This study aimed to assess whether European neonatal intensive care units (NICUs) with a low use of PNC or an explicit policy to reduce PNC use had higher risks of mortality or BPD. Methods: We included 3,126 infants in 105 NICUs born between 24 + 0 and 29 + 6 weeks’ gestational age in 19 regions in 11 countries in the EPICE cohort. First, we identified clusters of NICUs using hierarchical clustering based on PNC use and BPD prevalence and compared case mix and mortality between the clusters. Second, a multilevel analysis was performed to evaluate the association between a restrictive PNC policy and BPD occurrence. Results: There were 3 clusters of NICUs: 52 with low PNC use and a low BPD rate, 37 with low PNC use and a high BPD rate, and 16 with high PNC use and a medium BPD rate. Neonatal mortality did not differ between clusters (p = 0.88). A unit policy of restricted PNC use was not associated with a higher risk of BPD (odds ratio 0.68; 95% confidence interval: 0.45–1.03) after adjustment. Conclusion: Up to 49% of NICUs had low PNC use and low BPD rates, without a difference in mortality. Infants hospitalized in NICUs with a stated policy of low PNC use did not have an increased risk of BPD.

Objective: Caesarean section (CS) may reduce mortality and morbidity for very preterm breech infants, but evidence is inconclusive. We evaluated neonatal outcomes for singleton breech infants by mode of delivery in a European cohort. Study design: Data come from the EPICE population-based cohort of very preterm births in 19 regions in 11 European countries (7770 live births). The study population was singleton spontaneous-onset breech births at 24-31 weeks gestational age (GA) without antenatal medical complications requiring caesarean delivery (N = 572). Mixed-effects regression models adjusting for maternal and pregnancy covariates and propensity score matching was used to examine the effect of (1) CS and (2) a unit policy of systematic CS for breech presentation by GA. The primary outcome was a composite of in-hospital mortality, intraventricular haemorrhage grades III & IV or cystic periventricular leukomalacia. Secondary outcomes were each component separately, five minute Apgar score below seven and mortality within six hours of delivery. Results: 64.4% of infants were delivered by CS with a range across regions from 41% to 100%; these infants had higher GA and were more likely to be small for gestational age, receive antenatal steroids, and have mothers who were hospitalised for more than one day before delivery compared to those delivered vaginally. CS was associated with lower risks of all outcomes in mixed-effects adjusted models (odds ratio (OR) for the composite outcome: 0.50, 95% confidence interval (Cl): 0.30-0.81), but not in propensity score matched models (OR: 0.72, 95% CI: 0.41; 1.29). A systematic CS policy was associated with lower mortality and morbidity in unadjusted, but not adjusted models (OR for composite outcome: 0.76, 95% CI: 0.44; 1.28). 35% of births 24-25 weeks were delivered by CS and protective effects were consistently stronger, but not statistically significant. Conclusions: Point estimates indicated protective effects of caesarean delivery for very preterm breech infants in conventional statistical models. However, analyses using propensity scores and based on unit policies did not confirm statistically significant associations. Prospective large-scale studies are needed to establish best practice and could be implemented in European regions where vaginal delivery remains an option. (C) 2019 Published by Elsevier B.V.