Socio–economic achievements of individuals born very preterm at the age of 27 to 29 years

Follow-up studies on outcomes of preterm infants have received increasing attention owing to the remarkable improvements in the survival of very preterm (VPT) infants. These studies have shown that children born preterm have high prevalences of neurodevelopmental disabilities, behavioural and emotional difficulties, higher rates of dysfunction in cognition and executive functioning, poorer academic achievement, grade failure, and increased utilization of remedial education that persists to adolescence. With these pessimistic observations, it is not surprising that some investigators predicted that nearly half of children born very preterm would not become fully independent adults. However, over the last decade there are a number of outcome studies of preterm infants at adulthood that are fairly positive, using two different designs. The most common design is descriptive cohort studies with matched controls followed longitudinally, with meticulous efforts to gain compliance, administration of standardized tests and validated self-completed questionnaires, and access to databases collected prospectively.1, 2 These studies are expensive and time-consuming, but provide accurate and valuable information on the severity of disabilities, functional abilities, emotional and behavioural concerns, and self-perception of quality of life. Although most studies show a somewhat lower rate of educational achievement, employment, independent living, dating, and sexual activity, the preterm group has also been reported to have decreased rates of risk-seeking behaviours than term controls.1 Recently, researchers in Europe have taken an innovative approach and published large epidemiological studies to adulthood with data from national registers.3, 4 The databases have unique identifiers linking birth data to subsequent vital statistics of the individual, including education, employment, income, marriage, children, and even contact with enforcement agents law. So far, such studies have been reported in Sweden3 and Norway,4 and now in this study by Mathiasen et al. in Denmark.5 The message is the same: the majority of young adult survivors born VPT appear to do reasonably well in terms of education, employment, and independent living. However, in all these studies, with some variations, statistically significant differences were observed, with the VPT group having a higher prevalence of impairments, more young adults living at home, lower levels of education, lower income, and a higher level of unemployment and dependence on social benefits. These disadvantages were greater with decreasing gestational age. In another Norwegian study,6 stillbirth rates were higher and reproductive rates significantly lower among the preterm cohort. Further, females born preterm showed a gestational age dose response, with the more immature females having a higher risk of delivering a preterm offspring. These large epidemiological studies provide a wealth of information and are extremely cost-effective. Other advantages are that the data are readily available, with minimal losses due to missing data. However, there are some limitations: lack of information regarding severity of impairments and functional abilities, and the inability to collect data on emotional issues and depression. For two reasons, we prefer to have the largest feasible sample size: more accurate parameter estimation, and to allow for sub-group analyses. However, the downside to large sample sizes is that it dooms us to statistical significance. With a large sample size, even trivial differences or relationships are unlikely to have arisen by chance. For example, with a sample size of 1400 in this study,5 a correlation of 0.053 (accounting for ¼ of 1% of the variance) would be significant; and group differences smaller that 1/10th of a standard deviation are significant. Consequently, it is necessary to temper interpretations of statistical significance with questions about clinical importance. As the authors state, despite the statistical differences, the public health impact is small.5 Similarly, many analyses lead to the problem of multiplicity; inflating the probability of significant results by chance. If 10 analyses are done, the probability of at least one significant chance finding is 40%; and it is over 78% if 30 analyses are run. This assumes that the analyses are independent. When they are not, as with education or income in this paper, the probability of finding significance by chance escalates considerably. Usually, we correct for this by adopting a more conservative alpha level, using either the Bonferroni correction, or a sequential testing procedure. However, this is rarely done.2 Other differences between European and North American studies are that the majority of survivors, as in this study, are between 30 and 32 weeks’ gestation, and therefore do not represent the most immature infants who are at highest risk. Unlike the Cleveland study,1 the European population is also more racially homogeneous, socioeconomically more advantaged, and with access to national health services. Despite the greater immaturity and higher rates of impairments in the US1 and Canadian cohorts,2 a significant proportion was reported to doing reasonably well. Ultimately, had these studies to adulthood not been undertaken we would never have known the extent of recovery. It appears that despite disabilities, a large proportion of young adults are doing better than expected, and that the earlier dire predictions are not borne out.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.