Net worth of networks: opportunities and potential

Collaborative quality improvement activities between units/regions/countries have an increasing role in improving neonatal care worldwide. Quality improvement activities have been enabled by the formation of collaborative neonatal networks at regional, national and international levels. Such networks act as a conduit for benchmarking 1, trend evaluation 2, quality improvement 3-5, knowledge mobilisation, evaluative research, 3, 4, and integrated comparative effectiveness research 3-5. In this issue of Acta Pediatrica, Kiechl-Kohlendorfer and colleagues 6 report the neonatal outcomes of preterm neonates born at 230–316 weeks’ gestation and admitted to 22 neonatal units in Austria between 2011 and 2016. The authors are to be applauded for developing a robust near-population-based, national cohort of very preterm neonates and obtaining data from a majority of the neonatal units in Austria. Overall, survival of the entire cohort was 91.6% and survival without major neonatal morbidity was 78.2%. Not surprisingly, lower gestational age, lower birth weight, incomplete or missing antenatal steroid administration, male sex and multiples were associated with adverse outcomes. The remarkably low adverse outcome rates were achieved in a nationally funded health care for pregnant women and neonates. While there have been previous reports of near-population-based cohorts with a similar range of outcome rates from various countries, including those from Europe 7, contemporary reports continue to be welcome in this constantly evolving field. In addition to unit-to-unit comparisons, national data can identify time trends in outcomes. Kiechl-Kohlendorfer et al. 6 did not observe any change in mortality or morbidity over the study period; however, they assert that it could be due to the fact that they only studied outcomes over six years. Notwithstanding Austria's small size and fewer than 1000 very preterm infants born annually, it would be of concern that the neonatal outcomes of very preterm neonates did not change over the study period in the country. Some networks have similarly reported no improvement in outcome trends, while others continue to improve. National networks could support outcome improvement by using the collected data to develop, support and implement quality improvement initiatives. The Austrian Preterm Outcome Study Network is part of a governmental quality assessment programme, which could facilitate quality improvement activities and achieve improved outcomes, reduce practice variability, and create opportunities for this to be a learning exercise for everyone involved 3, 8. One problem in comparing data from different networks is a lack of comparable denominators 9. The variations in the reported pregnancy and infant outcomes can partly be attributed to differences in the denominators which can be, for example, the number of pregnancies, number of births, number of live births, number of neonates admitted to neonatal units and number of neonates for whom follow-up data are available. In the current study 6, live births admitted to neonatal units were included in the denominator, but there is lack of clarity regarding delivery room deaths. Furthermore, adverse outcome rates were calculated using all neonates included in the study as the denominator; however, not all neonates were assessed for all outcomes. For example, retinopathy rates could be significantly underestimated if all neonates in a specific gestational age group are used as the denominator which assumes that the neonates who did not survive to get eye examinations are free of retinopathy of prematurity. The challenge calls for efforts to harmonise the denominators used for different outcomes. The Austrian Preterm Outcome Study Network will now collect data on stillbirths to further improve the precision of their reporting. One concern is how much of the reported population-based cohorts are really population based. In the current study, infants were missing from the database especially in the lowest and highest gestational age groups; up to 27% of missing infants at 23 weeks of gestation. Even if the national networks aim for a complete coverage, a large variation exists in the coverage 10. The missing data can create selection bias with the possibility of significant differences in actual versus reported outcomes in the country 9. Networks need to transparently recognise and acknowledge these biases, and work to improve the data coverage. The Austrian Preterm Outcome Study Network has done exemplary work by reporting the completeness of the data by comparing their numbers with the National Birth Registry to evaluate the coverage of the network and to show its weaknesses. Norman et al. 11 have reported their model to validate the data of the Swedish Neonatal Quality register against other national registers for the completeness and agreement of the data. They showed excellent completeness of the Swedish data regarding preterm infants born between 24 and 34 weeks of gestation and high agreement for most diagnoses. The most important step after creating networks with comprehensive databases is to conduct collaborative quality improvement activities that help individual units to improve their outcomes. This can be done by learning from units with better outcomes or by developing joint efforts to target particular morbidities and reduce variability in evidence-based treatments and in outcomes 3-5, 8. For example, Kiechl-Kohlendorfer et al. found that only 64% of very preterm born infants had received a complete course of antenatal steroids in Austria, which they recognised as a potential target for improvement within their network. A bundle of four evidence-based practices, namely birth in the appropriate regional centre, prevention of neonatal hypothermia, and administration of antenatal steroids and early continuous positive airway pressure/surfactant after birth, was estimated to result in an 18% reduction in mortality without an increase in severe morbidity 12. National networks like the Austrian network could play a vital role in studying the implementation of bundled approaches both before and after birth to achieve improved outcomes. Many treatments in neonatology have not been studied by randomised study designs. The obstacles are practical such as the need for large sample sizes and considerable need of resources for data collection, data repository and analyses. Using existing networks with data collection infrastructure could help to navigate these obstacles. The real-world data from networks could be used to generate real-world evidence 13. Additionally, comparative research about treatment effectiveness can also be conducted within a shorter timeframe utilising established networks 14. Finally, networks can act as effective and efficient platforms for knowledge translation. Networks can act as solid platforms providing feedback to the participant units regarding their practices, outcomes and benchmark with similar units. This can facilitate exchanges between units and adoption of potentially better practices. The networks require infrastructure including regulatory approvals, solution of privacy and confidentiality issues, funding and sustainability, and maintaining the high-quality data collection efforts among participants. The key elements for success include acceptance and funding from regulatory and/or governmental agencies, cohesiveness and collaborative spirit of network participants, openness in governance and support for junior members to evolve into leadership positions. Austria has succeeded in this network building. The next step is up to the individual participating units and the government to enable the network to achieve its full net worth. The success of the Austrian neonatal network would provide an example about the ability of a network activity to improve the health of their citizens. With each network success, we learn more about how to overcome barriers and further develop consistent and acceptable safeguards for data access, use and management. The authors gratefully thank the staff at the Maternal-Infant Care Research Centre (MiCare) for organisational support and Sarah Hutchinson, PhD, for editorial assistance in the preparation of this manuscript. MiCare is supported by a team grant from the Canadian Institutes of Health Research (CTP 87518), the Ontario Ministry of Health, and support from participating hospitals. Dr. Shah holds an Applied Research Chair in Reproductive and Child Health Services and Policy Research (APR 126340), a team grant (PBN 150642) and a project grant (PJT 162429) awarded by the Canadian Institutes of Health Research. The authors have no conflicts of interest to disclose.

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.