{
  "abstract": "Background Dementia is a significant public health concern. Evidence of falling age-specific incidence, and identification of potentially modifiable risk factors, suggests that prevention is possible, but more evidence is needed to understand the social determinants of dementia. The polysocial risk score (PRS) is a new methodology that considers the overall contribution of multiple social risk factors to disease risk, recognising their overlapping and interactive effects.Methods In EPIC-Norfolk, a population-based cohort in the UK with 30 years of follow-up through linked healthcare and mortality records, a PRS was constructed using social factors significantly associated with dementia (level of education, number of childhood difficulties, marital status and financial stability). Sex-stratified survival models examined the relationship between PRS tertiles and incident dementia, adjusted for age, ethnicity, and APOE e4 genotype. Medical risk was estimated using the Lifestyle for Brain Health (LIBRA) risk score. Additional models examined the interplay between PRS and medical factors and dementia risk, and the association between PRS and different diagnostic subtypes of dementia.Results Polysocial risk was associated with higher dementia risk (medium risk tertile HR 1.14 (95% CI 1.03,1.26); high risk tertile 1.38 (1.18,1.61); reference = low risk tertile). Social and medical factor associations with increased dementia risk appeared to be largely independent. There was some evidence of a smaller effect for Alzheimer’s disease (high risk tertile HR 1.06 (95% CI 0.71,1.58); reference = low risk tertile) compared to other diagnostic subtypes. Results were similar for men and women. Models were robust to exclusion of incident cases in the first 10 years of follow-up.Conclusion Higher polysocial risk was associated with significantly increased all-cause dementia risk, which appeared to be largely independent of medical factors. These findings support a holistic approach to dementia risk reduction that incorporates both individual-level health behaviour support, and population-level action on social determinants of health.",
  "authors": [
    {
      "affiliations": [
        "Department of Public Health and Primary Care, University of Cambridge, Cambridge, UK"
      ],
      "name": "Emma Lockwood"
    },
    {
      "affiliations": [
        "Department of Public Health and Primary Care, University of Cambridge, Cambridge, UK"
      ],
      "name": "Sebastian Walsh"
    },
    {
      "affiliations": [
        "Department of Public Health and Primary Care, University of Cambridge, Cambridge, UK"
      ],
      "name": "Carol Brayne"
    }
  ],
  "title": "P06-2 Long-term associations between a polysocial risk score and dementia: a retrospective cohort study",
  "uid": "767d14a8-b67d-5502-b192-aec09d5f7adc"
}