Background Socio-economic factors are important determinants of health outcomes and mortality for CYP. Individual level data on parental occupation is collected for deceased CYP <16 years old during death registration. Non-trivial proportions of records are not assigned a SEC; a validated adjustment technique has not been established. Area based measures of deprivation are relied on as imperfect proxies for the socio-economic situation of individuals introducing the problem of ecological fallacy. This analysis describes the relationship between area based and individual level SEC measures contained in death records.

Methods Retrospective population based observational study of all CYP deaths <16 years old in England, 2001–2010 (N=19,076). Area based deprivation was measured using Index of Multiple Deprivation (IMD) quintiles (1 most deprived; 5 least deprived). Parental occupation was coded in line with National Statistics SEC (NS-SEC), used in death registry data since 2001 (from 1: managerial and professional, to 7: routine occupations). Those with unassigned SEC include; full-time students; never worked; long-term unemployed; inadequately described, not classifiable and missing. The relationship between IMD and SEC was described. Unassigned SEC cases were described by; IMD quintile, UK region and UK origin.

Results IMD and SEC measures have a broadly linear association; the proportion of those living in the most deprived areas increases as SEC gets lower from professional to routine occupations, the inverse is true for those living in the least deprived areas. However, 30% of CYP with parents in the highest SEC group (professional) live in the most deprived 2 quintiles, and 17% of those in the lowest SEC group (routine) live in the least deprived 2 quintiles. Over the study period 16% of CYP deaths have unassigned SEC data. Over half (52%) of CYP with unassigned SEC lived in the most deprived quintile. North East had the highest proportion of unassigned SEC (23%), South East had the lowest (11%). CYP born in the UK were less likely than those born outside the UK to have unassigned SEC (16% and 24% respectively).

Conclusion Area based measures are important tools but are not necessarily good proxies for measuring individual socio economic factors. Whilst being broadly congruent, substantial disparities exist between IMD and SEC. Bias associated with unassigned SEC cases severely limits usability of this measure. Further research is needed to develop methods for adjusting unassigned SEC data to a validated distribution tested with additional data sources, as has previously been done for adult data.