Model | No | Akaike information criterion* | Bayesian information criterion† | Likelihood ratio tests P value |
PCE‡ | 9728 | 2371 | 2386 | |
i.PCE + i.Education§ | 9717 | 2366 | 2395 | 0.004 |
(i.PCE)x(i.Education) | 9717 | 2331 | 2374 | <0.0001 |
i.PCE + i.ADI¶ | 9728 | 2371 | 2400 | 0.14 |
(i.PCE) x (i.ADI) | 9728 | 2346 | 2389 | <0.0001 |
i.PCE + i.Education + i.ADI | 9717 | 2366 | 2409 | 0.002 |
(i.PCE) x (i.Education)x(i.ADI) | 9717 | 2328 | 2458 | <0.0001 |
All models that added in the social deprivation factor as a risk factor was compared with the PCE without a social deprivation factor.
All models that added in social deprivation as an interaction term was compared with the PCE model with social deprivation added as a risk factor.
*Akaike information criterion measures goodness-of-fit between observed values and expected values; lower scores compared with referent indicate an improvement in prediction.
†Bayesian information criterion measures goodness-of-fit between observed values and expected values; lower scores compared with a referent model indicate an improvement in prediction.
‡PCE predicted risk was stratified into four categories of risk: 0%–5%; >5%–10%; >10%–15%; >15%.
§Education was stratified into three categories: no high school; high school/some college; college or above (referent).
¶Higher ADI indicates higher neighbourhood deprivation and was stratified into three categories according to the IQR: top ADI quartile; middle two ADI; lowest ADI quartile (referent).
ADI, Area Deprivation Index; ASCVD, atherosclerotic cardiovascular disease; PCE, pooled cohort equations.