Table 2.
Results from the cross-classified multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA) for periodontal measures (n = 1936).
| Teeth with PD | Teeth with CAL | |||
|---|---|---|---|---|
| Model 1A a | Model 1B a | Model 2A | Model 2B | |
| RR (95% CI) | RR (95% CI) | RR (95% CI) | RR (95% CI) | |
| Fixed effects (n = 1936 adults) | ||||
| Ethnicity (reference group: White British) | ||||
| White Other | 1.07 (0.81–1.39) | 0.96 (0.63–1.36) | ||
| Black African | 1.19 (0.92–1.61) | 1.12 (0.77–1.67) | ||
| Black Caribbean | 1.11 (0.78–1.47) | 0.80 (0.46–1.27) | ||
| Black Other | 1.03 (0.78–1.35) | 0.83 (0.53–1.20) | ||
| Pakistani | 1.62 (1.22–2.14) | 0.89 (0.61–1.25) | ||
| Indian | 1.70 (1.25–2.30) | 0.94 (0.60–1.49) | ||
| Bangladeshi | 2.17 (1.51–3.04) | 1.11 (0.65–1.79) | ||
| Asian Other | 1.77 (1.40–2.26) | 0.70 (0.50–0.94) | ||
| Nativity status (reference group: Native-born) | ||||
| Foreign-born | 1.16 (1.01–1.30) | 1.60 (1.24–1.92) | ||
| Education (reference group: Low) | ||||
| Medium | 1.14 (0.95–1.36) | 1.03 (0.81–1.35) | ||
| High | 0.98 (0.84–1.17) | 0.74 (0.60–0.91) | ||
| Sex (reference group: Men) | ||||
| Women | 0.93 (0.84–1.02) | 0.81 (0.66–0.96) | ||
| Age groups (reference group: 16–24 years) | ||||
| 25–34 years | 1.29 (1.07–1.46) | 2.19 (1.73–2.78) | ||
| 35–44 years | 1.23 (1.02–1.41) | 2.48 (1.97–3.06) | ||
| 45–54 years | 1.40 (1.12–1.69) | 4.08 (2.91–5.46) | ||
| 55–65 years | 1.39 (1.08–1.75) | 4.90 (3.41–6.57) | ||
| Intercept | 11.39 (10.41–12.47) | 6.47 (5.27–8.02) | 4.01 (3.43–4.62) | 1.76 (1.29–2.29) |
| Random effects (n = 51 intersectional social strata): | ||||
| Between-stratum variance | 0.06 (0.03–0.12) | 0.02 (0.003–0.04) | 0.08 (0.02–0.19) | 0.02 (0.001–0.07) |
| VPC (%) | 4.8% | 1.3% | 2.1% | 0.6% |
| PCV (%) | 73.6% | 74.0% | ||
CAL: clinical attachment loss; PD: probing depth; PCV: percentual change in variance; RR: rate ratio; VPC: variance partition component. a Two-level negative binomial regression models were fitted for each periodontal measure, with participants clustered within intersectional social strata. Model A is the simple intersectional model (variance components model), and Model B is the intersectional interaction model (random intercepts model).