The Relationship of Multimorbidity, Socio-economic Deprivation and Mortality: Findings from UK Biobank Cohort

Problem

Multimorbidity, presence of ≥2 long-term conditions (LTCs), is the norm rather than the exception. Multimorbidity is socially patterned, with people in socio-economically deprived areas experiencing multimorbidity a decade earlier. The relationship between multimorbidity, socio-economic deprivation and mortality in general population remains unclear.

Approach

To examine the number of LTCs and their associations, if any, with mortality, as well as to study the effects of socio-economic status on this relationship in UK Biobank participants. Data from 500792 participants in UK Biobank, an anonymised community research cohort, aged between 40-73 years, recruited between 2006-2010 from across the UK. Self-reported LTCs (n=42) were identified in all participants at baseline; multimorbidity was classified into three categories (0 LTCs, 1-3 LTCs, 4 or more LTCs). All-cause mortality, vascular and cancer mortality was available for a median follow-up period of 7 years (Interquartile range 76 - 93 months) by linking UK Biobank records with national mortality records. Hazard Ratios (HRs) examined associations between number of LTCs and all-cause mortality with 95% confidence intervals (CI). Results were adjusted for age, sex, socio-economic status, smoking and alcohol status, body mass index (BMI) and levels of physical activity reported at baseline. A sub-group analysis was performed by repeating the analysis described above after dividing the participants based on socio-economic status. The analysis was repeated using three different classifications of socio-economic status based on Townsend score, annual family income and education qualification reported by participants.

Findings

The rate of all-cause mortality was 2.9% (14348 participants) at 7 years; 8350 participants died due to cancer and 2985 participants died due to vascular causes. Presence of 1-3 LTCs was associated with the risk of all-cause mortality (HR 1.76; CI 1.68-1.85), cancer mortality (HR 1.69; CI 1.59-1.79) and vascular mortality (HR 1.84; CI 1.65-2.06) when compared to participants without any LTCs. A stronger association was observed between the presence of ≥4 LTCs with all-cause mortality (HR 3.04; CI 2.85-3.24), cancer mortality (HR 2.21; CI 2.02-2.42) and vascular mortality (HR 3.97; CI 3.45-4.56). Results adjusted for the confounders described above. In the sub-group analysis based on socio-economic status, the adjusted effect size of multimorbidity on all-cause mortality was significantly higher for the deprived group as compared to the affluent group (e.g. deprived participants with ≥4 LTCs were 80% more likely to die as compared to their affluent counterparts). Results were consistent across all three classifications of socio-economic status examined.

Consequences

In a large community cohort of half a million people, a dose response relationship was observed between multimorbidity and mortality. The excess mortality of socieoeconomically deprived people with multimorbidity was not explained by lifestyle factors alone. Future research should explore mechanisms underpinning this finding and consideration of new models of care.