Table presents the prevalence of the main variables in both initial samples (n = 8,414), after weighting for sex, age, and municipality, according to the demographics of the populations. Type D personality was found in one fifth of both populations. Social inhibition occurred more often than negative affectivity, especially in the West-Brabant region. Low education (highest completed education none or primary school) was less prevalent (6.8-7.5%) than a low income (income below Dutch standard; 38.3-41.5%).
Weighted prevalences of the main variables
Examination of the occurrence of Type D personality over the categories of SES and (determinants of) health showed that the prevalence of Type D personality increased with decreasing education and income (Table ). With regard to lifestyle-related risk factors, the most striking finding was the highest prevalence of Type D personality in the alcohol abstainers. As to health, there seemed to be a dose-response relationship between, on the one hand, self-assessed health and psychological distress, and, on the other hand, the prevalence of Type D personality: the poorer the self-assessed health or the higher the psychological distress, the higher the prevalence of Type D personality (Table ).
Weighted prevalence of Type D personality
Using the total dataset (n = 12,090), adjusted for sex, age, and municipality, Type D personality was significantly associated with both indicators of a low SES: low education (ORadj = 1.7, 95%CI: 1.5-2.0) and low income (ORadj = 1.7, 95%CI: 1.6-1.9) (not tabulated).
Persons with a Type D personality had a small but significantly higher risk of current tobacco smoking (RRadj = 1.1, 95% CI: 1.1-1.2), but not of unsafe alcohol use and obesity (Table ). Furthermore, Type D personalities were at a higher risk of self-assessed poor health (RRadj = 2.8; 95% CI = 2.6-3.1), chronic disease (RRadj = 1.2, 95% CI = 1.1-1.2), cardiovascular disease (RRadj = 1.6, 95% CI = 1.2-2.0), and high psychological distress (RRadj = 8.6, 95% CI = 4.9-15.1). Type D personalities did not have an elevated risk of diabetes. The associations all remained statistically significant when they were analyzed conditionally on low education or on low income, though some relative risks moved slightly towards the null value (Table ).
Results for the (un)conditional associationsa of Type D personality and low SES
Persons with low education as well as those with a low income had significantly higher relative risks for all studied indicators for (determinants of) health, except for unsafe alcohol use (Table ). The risk of unsafe alcohol use was significantly lower for persons with a low education (RRadj = 0.8, 95% CI: 0.6-0.9). All associations remained statistically significant when they were analyzed conditionally on Type D personality, though some relative risks moved slightly towards the null value (Table ).
Interaction between Type D personality and low SES on an additive scale was significant for the effect of low education on high psychological distress (RERI = 12.9, 95% CI: 0.8-32.3), and for the effect of a low income on self-assessed poor health (RERI = 1.4, 95% CI: 0.9-1.9) and on high psychological distress (RERI = 11.4, 95% CI: 3.5-41.0) (Table ). This means, for example, that the relative risk for self-assessed poor health is 1.4 higher in Type D personalities with a low income than if there were no interaction between Type D personality and low income. Because the absolute background risk was 5.8% (the prevalence of a poor self assessed health in the absence of a Type D personality and a low income in the region West-Brabant) this means that the absolute excess risk due to interaction is 8.1% (5.8% × 1.4). Accordingly, the excess risk due to interaction for high psychological distress is 27.1% (2.1% × 12.9) for the interaction between Type D personality and low education and 8.0% (0.7% × 11.4) for the interaction between Type D personality and low income, based on the background prevalence in the region 's-Hertogenbosch.
Results for the modification of effectsa of Type D personality and low SES