Incidence rate ratio of hospital admissions between individuals with type I diabetes compared to normal population controls at different levels of exposures of interest
| Model (DIC) | Category | Parameter estimates | Variances | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| IRR | (SE) | (95% CI) | p≤ | ESS | Level | Variance (SE) | (95% CI) | ESS | |||
| 1 (22377) | Sex | ||||||||||
| Male (ref) | 5.619 | (1.100) | (5.039 to 6.258) | 0.001 | 3075 | 1 | – | – | – | – | |
| Female (ref) | 1.058 | (1.005) | (0.905 to 1.238) | 0.237 | 2953 | 2 | 1.260 | (0.046) | (1.173 to 1.354) | 6287 | |
| 3 | 0.092 | (0.023) | (0.048 to 0.138) | 1290 | |||||||
| 2 (22350) | SES | ||||||||||
| 1 (ref) | 4.794 | (1.161) | (3.973 to 5.771) | 0.001 | 996 | 1 | – | – | – | – | |
| 2 | 1.182 | (1.022) | (0.916 to 1.531) | 0.100 | 1594 | 2 | 1.269 | (0.046) | (1.181 to 1.363) | 5185 | |
| 3 | 1.308 | (1.036) | (1.013 to 1.697) | 0.020 | 1516 | 3 | 0.077 | (0.024) | (0.030 to 0.124) | 814 | |
| 4 | 1.139 | (1.017) | (0.886 to 1.465) | 0.156 | 1416 | ||||||
| 5 | 1.393 | (1.043) | (1.088 to 1.786) | 0.004 | 1376 | ||||||
| 3 (22370) | Centre size | ||||||||||
| Small (ref) | 6.231 | (1.098) | (5.632 to 6.892) | 0.001 | 3898 | 1 | – | – | – | – | |
| Large | 0.821 | (0.984) | (0.700 to 0.965) | 0.009 | 3979 | 2 | 1.264 | (0.046) | (1.178 to 1.357) | 7327 | |
| 3 | 0.109 | (0.023) | (0.065 to 0.157) | 1501 | |||||||
| 4 (22354) | Age at diagnosis (years) | ||||||||||
| <5 (ref) | 7.843 | (1.185) | (6.651 to 9.202) | 0.001 | 1336 | 1 | – | – | – | – | |
| 5–10 | 0.742 | (0.969) | (0.605 to 0.908) | 0.002 | 1745 | 2 | 1.260 | (0.046) | (1.174 to 1.353) | 7068 | |
| 10+ | 0.599 | (0.948) | (0.490 to 0.737) | 0.001 | 1629 | 3 | 0.104 | (0.023) | (0.060 to 0.150) | 1396 | |
| 5 (22372) | Urban rural | ||||||||||
| Rural (ref) | 5.523 | (1.128) | (4.801 to 6.336) | 0.001 | 1766 | 1 | – | – | – | – | |
| Urban | 1.073 | (1.006) | (0.907 to 1.277) | 0.206 | 1760 | 2 | 1.267 | (0.047) | (1.178 to 1.361) | 6460 | |
| 3 | 0.108 | (0.024) | (0.062 to 0.157) | 1413 | |||||||
| 6 (22359) | Re county | ||||||||||
| Case | 5.785 | (1.075) | (5.338 to 6.268) | 0.001 | 7049 | 1 | – | – | – | – | |
| 2 | 1.269 | (0.046) | (1.182 to 1.362) | 6618 | |||||||
| 3 | 0.091 | (0.023) | (0.047 to 0.139) | 1163 | |||||||
| 4 | 0.020 | (0.011) | (0.007 to 0.050) | 18 169 | |||||||
| 7 (22335) | Date of birth | ||||||||||
| <90 (ref) | 4.737 | (1.152) | (3.973 to 5.653) | 0.001 | 1033 | 1 | – | – | – | – | |
| 90–94 | 0.992 | (0.999) | (0.795 to 1.242) | 0.474 | 1395 | 2 | 1.265 | (0.046) | (1.179 to 1.359) | 7334 | |
| 95–99 | 1.452 | (1.044) | (1.158 to 1.821) | 0.001 | 1445 | 3 | 0.093 | (0.023) | (0.051 to 0.141) | 1245 | |
| 00–04 | 1.692 | (1.076) | (1.288 to 2.233) | 0.001 | 2000 | ||||||
| >05 | 1.497 | (1.103) | (0.930 to 2.408) | 0.049 | 6364 | ||||||
All models use a multilevel Poisson model. Results are reported using the mean of the posterior distribution to indicate IRR, the SD of the posterior chain is used to indicate the parameter SE, 95% posterior probability intervals (95% CI) represent the 2.5th and 97.5th centiles of the posterior distribution, and directional posterior probabilities (p≤). ESS indicates the effectiveness of MCMC chain mixing. Bayesian DIC is used to indicate model fit. The reference category comparing the IRR between cases and controls is indicated, and the IRR of the exposure of interest is represented by the interaction between the exposure of interest and case–control status. Level 2 variance indicates the overdispersion parameter, level 3 variance indicates the matching criteria, and level 4 (model 6) variance indicates the county level variance.
DIC, Deviance Information Criterion; ESS, effective sample size; IRR, incidence rate ratio; SES, socioeconomic status.