Exploring interaction effects of social determinants of health with hospital admission type on academic performance: a data linkage study

Objective To investigate the moderating effects of socio-demographic social determinants of health (SDH) in the relationship between types of childhood hospitalisation (ie, none, injury, non-injury, injury+non-injury) and academic performance.Design, setting and patients Children residing in Wales 2009-2016 (N=369 310). Secure Anonymised Information Linkage databank linked Tagged Electronic Cohort Cymru (five data sources) from the Wales Electronic Cohort for Children.Main outcome measure Binary educational achievement (EA) measured across three key educational stage time points: grade 6 (mean age 11 years, SD 0.3), 9 (mean age 14 years, SD 0.3) and 11 (mean age 16 years, SD 0.3).Results Of the 369 310 children, 51% were males, 25.4% of children were born in the lowest two Townsend deciles. Females were more likely to meet EA than males (adjusted risk ratio (aRR) (95% CI): 1.047 (1.039, 1.055)). EA was lower for injury admissions in males and any admission type in females (interactions: femalexnon-injury 0.982 (0.975, 0.989); femalexinjury+non-injury 0.980 (0.966, 0.994)). Children born into a more deprived decile were less likely to achieve EA (0.979 (0.977, 0.980)) and worsened by an injury admission (interactions: townsendxinjury 0.991 (0.988, 0.994); Townsendxinjury+non-injury 0.997 (0.994, 1.000)). Children with special educational needs (SEN) were less likely to meet EA (0.471 (0.459, 0.484) especially for an injury admission (interactions: SENxinjury 0.932 (0.892, 0.974)).Conclusion SDH moderated the impact of hospital admission type on educational outcomes prompting future investigation into the viability of in-hospital routine screening of families for SDH and relevant post-hospital interventions to help reduce the impact of SDH on educational outcomes post-hospitalisation.