Correction to: Nature Mental Health https://doi.org/10.1038/s44220-024-00249-7, published online 30 April 2024.

In the version of article initially published, some data underlying the models presented in the paper were incorrectly transposed due to a file conversion error. This affected 22 participants’ 16-year post-Trier cortisol values, and thus their reactivity scores. All other data, including 12-year cortisol reactivity values, 12- and 16-year psychopathology values and 12- and 16-year pre-ejection period (PEP) values were unaffected. The analyses were repeated with the original code and updated data.

The primary findings of the indirect effect of institutional deprivation on P at 16 years through blunted cortisol reactivity at 12 years, and of institutional deprivation on EXT at 16 years through blunted PEP reactivity at 12 years, were substantively no different from the original results. However, some within-time and longitudinal associations differed from the original analyses. First, in the deprivation models, the within-time associations between cortisol reactivity and INT and EXT at 16 years were no longer significant; instead, a within-time association between higher P and more blunted cortisol reactivity at 16 years was observed, similar to what was observed at 12 years. Second, in the cortisol models, there was a new longitudinal association between higher INT at 12 years and elevated cortisol reactivity at 16 years which was not previously observed. In the intervention models, a new single parameter emerged as significant, which was the within-time association between higher P and more blunted cortisol reactivity at 16 years (similar to the deprivation model above). Minor differences to the results from the sensitivity analyses were noted, with the overall pattern of indirect effects being consistent with the original models (see updated Supplementary Information in the article).

The first paragraph of the Results “Cortisol reactivity as a mediator of early deprivation” and “Intervention effects” sections have been revised, the fourth paragraph of the Discussion has been revised, Figs. 1–3 and Supplementary Figs. 1, 2 were revised. Separately, the Supplementary Methods have been revised to update reporting of cortisol values from µg/dl to nmol/L. For comparison, the unaltered figures, text and Supplementary Information are available as Supplementary Information alongside the online version of this amendment.