Table 2

Multivariable logistic regression analysis of factors associated with Flag-4 registration and unpreventable tuberculosis (TB) if Flag-4 registered

Flag-4 status if registered with primary care*Unpreventable TB with Flag-4 registration†
VariablesnAOR (95% CI)‡Sig‡nAOR (95% CI)‡Sig‡
Gender0.0750.803
 Male§212/3951.094/2121.0
 Female246/4151.30 (0.98–1.72)0.075114/2461.07 (0.64–1.80)0.803
Age at primary care registration0.0260.196
 <16 years§17/481.02/171.0
 16–35 years307/5422.32 (1.22–4.40)0.043132/3078.88 (0.81–96.8)0.071
 >35 years134/2202.48 (1.27–4.83)0.02174/1349.29 (0.81–106)0.070
WHO TB incidence region0.0010.063
 ≥500/106§38/871.024/381.0
 150–499/106402/6752.41 (1.30–4.45)<0.001175/4021.31 (0.41–4.15)0.190
 <150/10618/480.77 (0.37–1.60)0.5059/185.29 (1.03–27.1)0.022
Disease type0.7220.182
 Non-pulmonary§271/4801.0124/2711.0
 Pulmonary smear negative136/2371.14 (0.83–1.58)0.57758/1361.62 (0.90–2.94)0.112
 Pulmonary smear positive51/931.07 (0.67–1.70)0.99926/511.81 (0.73–4.46)0.197
Interval between UK entry and GP registration, months1.03 (1.02–1.03)<0.001
  • The data presented are for the enter method; backward stepwise modelling retained all significant variables presented for the enter method at a significance level of 5% (see Methods).

  • *Analyses are confined to cases with a history of primary care registration (N=810). Model performance: Somers’ D=0.223; c statistic=0.612.

  • †Analyses are confined to cases with Flag-4 registration (N=458). Model performance: Somers’ D=0.793; c statistic=0.896.

  • ‡ORs are adjusted (AOR) for all input variables to the model with Wald 95% CIs.

  • §Reference subgroup for comparison.