Exposure Variable | Univariable Model Odds Ratio (95% Confidence Interval) | Multivariable Total Effect Model^{a} Odds Ratio (95% Confidence Interval) | Multivariable Direct Effect Model^{b} Odds Ratio (95% Confidence Interval) |
---|---|---|---|

Late presentation (n=2001) | |||

Socioeconomic deprivation (per quintile higher)^{c} | 1.00 (0.93 to 1.08) | 1.05 (0.96 to 1.15) | 1.04 (0.95 to 1.14) |

Distance to center (per 10 km) | 1.01 (0.98 to 1.03) | — | 1.00 (0.98 to 1.03) |

Urban location^{d}^{,}^{e} | 0.91 (0.69 to 1.20) | — | 1.04 (0.75 to 1.45) |

Pre-emptive transplantation (n=1529) | |||

Socioeconomic deprivation (per quintile higher)^{c} | 1.21 (1.12 to 1.31) | 1.21 (1.10 to 1.32) | 1.20 (1.10 to 1.31) |

Distance to center (per 10 km) | 1.04 (1.01 to 1.06) | — | 1.02 (0.99 to 1.05) |

Urban location^{d}^{,}^{f} | 0.88 (0.66 to 1.17) | — | 1.09 (0.78 to 1.53) |

↵a Model represents total effect of socioeconomic deprivation on outcome; because distance to center and whether base nephrology unit is a transplanting center are thought to mediate any deprivation-outcome association, these are omitted from the multivariable model.

↵b Multivariable model looks at direct effect of exposure on outcome: variables included in model are distance (per 10 km) or rural/urban location, socioeconomic deprivation, sex, age group, ethnic group, primary kidney disease, period of KRT start, and whether base nephrology unit is a transplanting center.

↵c Socioeconomic deprivation is parameterized as an ordinal variable because this offered the best goodness of fit. Odds ratios represent unit change in odds for each higher quintile of deprivation (or higher area affluence).

↵d Distance is omitted in the urban location multivariable model.

↵e Analysis excludes children from Northern Ireland (

*n*=1942).↵f Analysis excludes children from Northern Ireland (

*n*=1480).