Table 3.
Effect estimates for the interaction between age and education from complete and imputed data models
| Associations between age group and cervical cancer mortality for each level of education | ||||||||
|---|---|---|---|---|---|---|---|---|
| No education | Primary education | Secondary education | Higher education | |||||
| Age groups (years) | IRR | (95 % CI) | IRR | (95 % CI) | IRR | (95 % CI) | IRR | (95 % CI) |
| Complete data analysis (n = 4247) | ||||||||
| 25–29 vs. 20–24 | 1.90 | (0.50–7.21) | 3.78 | (2.32–6.16) | 4.03 | (2.56–6.34) | 4.31 | (1.58–11.8) |
| 30–34 vs. 20–24 | 6.83 | (2.09–22.3) | 6.16 | (3.86–9.82) | 11.6 | (7.57–17.8) | 15.1 | (5.93–38.5) |
| 35–39 vs. 20–24 | 7.90 | (2.46–25.4) | 9.77 | (6.18–15.4) | 19.0 | (12.4–28.9) | 27.7 | (11.0–69.9) |
| 40–44 vs. 20–24 | 10.4 | (3.28–33.1) | 12.6 | (7.97–19.8) | 22.2 | (14.6–33.8) | 29.8 | (11.8–75.5) |
| 45–49 vs. 20–24 | 11.3 | (3.58–35.9) | 16.5 | (10.5–25.9) | 26.4 | (17.3–40.4) | 49.7 | (19.9–124) |
| Imputed data analysis (n = 5098) | ||||||||
| 25–29 vs. 20–24 | 2.61 | (0.72–9.40) | 3.85 | (2.47–6.02) | 4.46 | (2.90–6.87) | 4.19 | (1.67–10.5) |
| 30–34 vs. 20–24 | 7.55 | (2.32–24.6) | 6.28 | (4.09–9.63) | 12.5 | (8.33–18.9) | 13.8 | (5.85–32.6) |
| 35–39 vs. 20–24 | 10.0 | (3.14–32.0) | 10.1 | (6.65–15.4) | 20.6 | (13.7–30.8) | 25.4 | (10.9–59.2) |
| 40–44 vs. 20–24 | 13.3 | (4.21–42.1) | 12.9 | (8.52–19.6) | 24.1 | (16.1–36.0) | 28.9 | (12.4–67.6) |
| 45–49 vs. 20–24 | 15.0 | (4.76–47.3) | 16.8 | (11.1–25.4) | 29.4 | (19.6–44.0) | 47.4 | (20.5–110) |
IRR Incidence rate ratios, 95 % CI 95 % confidence intervals
Results summarise both the analysis of data for cases with complete information on all risk factors of interest and the imputed data analysis for models examining the association between educational level and cervical cancer mortality (Model 1). Both multivariable models included fixed effects for age group, urban or rural residence, and region of residence, as well as interactions between educational level and age