Table 3– Outcome analysis of acceptance, offer and coverage of HIV testing among tuberculosis clinics in London from September 2009 to March 2010
OutcomeControl/interventionOutcomeMultivariable analysisChi-squared test for equality of proportionsAdjusted OR (95% CI)
Group A
 Test acceptedIntervention n = 534Test accepted462 (86.5)0.761.53 (0.84–2.81)#,¶
Test not accepted69 (12.9)
Test information missing3 (0.6)
Control n = 217Test accepted183 (84.3)1
Test not accepted25 (11.5)
Test information missing9 (4.1)
 Test offeredIntervention n = 655Test offered568 (86.7)0.0021.67 (1.07–2.60)#,+
Test not offered86 (13.1)
Offer information missing1 (0.2)
Control n = 308Test offered235 (76.3)1
Test not offered73 (23.7)
Offer information missing0 (0)
 CoverageIntervention n = 655Test coverage yes534 (81.5)0.0041.83 (1.3–2.7)#,+
Test coverage no121 (12.1)
Test coverage information missing0 (0)
Control n = 308Test coverage yes221 (71.7)1
Test coverage no87 (28.2)
Test coverage information missing0 (0)
Group B
 Test acceptedIntervention n = 197Test accepted172 (87.3)0.41.40 (0.67–2.91)#,¶
Test not accepted21 (10.7)
Test information missing4 (1.5)
Control n = 112Test accepted91 (81.3)1
Test not accepted18 (16.1)
Test information missing3 (2.7)
 Test offeredIntervention n = 211Test offered202 (95.7)0.023.76 (1.31–12.25)#,+,§
Test not offered6 (2.8)
Offer information missing3 (1.4)
Control n = 141Test offered125 (88.7)1
Test not offered14 (9.9)
Offer information missing2 (1.4)
 Test coverageIntervention n = 211Test coverage yes180 (85.3)0.041.84 (1.03–3.29)#,+
Test coverage no31 (14.7)
Test coverage missing0 (0)
Control n = 141Test coverage yes107 (75.9)1
Test coverage no34 (24.1)
Test coverage missing0 (0)
  • Data are presented as n (%), unless otherwise stated. Group A: clinics introducing opt-out testing, training and resources. Group B: clinics introducing training and resources only. #: adjusted for age group; : model included only patients who had a test offered to them and who had not previously had a test; +: model included all patients having had a prior test; §: adjusted odds ratio obtained from exact logistic regression.