Table 3. Incidence rate ratios of tuberculosis, in the UK born and non-UK born cohorts relevant to the targeted neonatal BCG vaccination scheme, using the best fitting modelsa as determined by comparison of the LOOIC, England, 2000–2015.
| Variable | IRR (95% CrI) | |
|---|---|---|
| UK born | Non-UK born | |
| Policy changeb | ||
| Pre-change | Reference | Reference |
| Post-change | 0.96 (0.82 to 1.14) | 0.62 (0.44 to 0.88) |
| Age | ||
| 0 | Reference | Reference |
| 1 | 1.39 (1.20 to 1.61) | 0.49 (0.30 to 0.83) |
| 2 | 1.24 (1.06 to 1.44) | 0.49 (0.30 to 0.80) |
| 3 | 1.21 (1.03 to 1.41) | 0.42 (0.26 to 0.68) |
| 4 | 0.90 (0.76 to 1.06) | 0.41 (0.25 to 0.66) |
| 5 | 0.89 (0.75 to 1.06) | 0.27 (0.16 to 0.45) |
| Incidence | ||
| UK born incidence rate (per standard deviation) | 1.12 (1.06 to 1.18) | NA |
| Non-UK born incidence rate (per standard deviation) | NA | 1.25 (1.04 to 1.51) |
| Year of study eligibility, group level | ||
| Intercept (standard deviation) | 1.13 (1.04 to 1.26) | NA |
| Year of study eligibility, individual level | ||
| 2000 | 0.83 (0.68 to 0.99) | NA |
| 2001 | 0.93 (0.79 to 1.07) | NA |
| 2002 | 1.08 (0.95 to 1.28) | NA |
| 2003 | 1.07 (0.93 to 1.26) | NA |
| 2004 | 1.12 (0.97 to 1.32) | NA |
| 2005 | 1.02 (0.89 to 1.17) | NA |
| 2006 | 1.02 (0.89 to 1.17) | NA |
| 2007 | 0.97 (0.83 to 1.11) | NA |
| 2008 | 1.01 (0.88 to 1.15) | NA |
| 2009 | 1.01 (0.88 to 1.16) | NA |
| 2010 | 0.98 (0.85 to 1.13) | NA |
BCG: Bacillus Calmette–Guérin; CrI: credible interval; IRR: incidence rate ratio; LOOIC: leave one out cross validation information criterion; SE: standard error; NA: not applicable (i.e. model terms were not included in the given cohort); UK: United Kingdom.
a The best fitting model for the UK born was a Poisson model with a random intercept for year of study entry, adjusting with fixed effects for the change in policy, age and incidence rates in the UK born (Model 16; Supplement Table S1). For the non-UK born a negative binomial model was the best fit, adjusting with fixed effects for the change in policy, age and incidence rates in the non-UK born (Model 8; Supplement Table S1).
b There was an improvement in the LOOIC score of 0.92 (SE: 1.07) from dropping the change in policy from the model in the UK born cohort and a −3.45 (SE: 4.63) improvement in the non-UK born cohort.