Table 2. TB infection prevalence estimates and Cohen's kappa coefficients per community.
Community code | Geography | Urban/rural | 1HIV prevalence | Number with both TST & QFT results | QFT-GIT positive (%) | TST positive (10 mm) (%) | Kappa |
SA1 | Province | Urban | Moderate | 109 | 78 | 42 | 0.17 |
SA 3 | Metropole | Urban | High | 98 | 77 | 77 | 0.49 |
SA 6 | Province | Rural | Moderate | 70 | 77 | 27 | 0.15 |
Z 4 | Lusaka | Urban | High | 69 | 72 | 41 | 0.25 |
SA 5 | Metropole | Urban | Moderate | 52 | 71 | 62 | 0.28 |
Z 5 | Copperbelt | Urban | High | 45 | 71 | 13 | −0.09 |
SA 2 | Province | Rural | Moderate | 87 | 70 | 28 | 0.16 |
SA 4 | Metropole | Urban | High | 140 | 68 | 61 | 0.18 |
Z1 | Lusaka | Urban | High | 93 | 67 | 49 | 0.36 |
Z 15 | Luapula | Rural | Moderate | 24 | 67 | 33 | 0.25 |
SA 7 | Metropole | Urban | High | 74 | 66 | 76 | 0.38 |
Z 11 | Luapula | Rural | Moderate | 40 | 65 | 73 | 0.02 |
SA 8 | Metropole | Urban | High | 124 | 65 | 24 | −0.01 |
Z 7 | Lusaka | Urban | High | 92 | 61 | 38 | 0.19 |
Z 6 | Lusaka | Urban | High | 84 | 61 | 61 | 0.48 |
Z 8 | Southern | Urban | High | 121 | 60 | 42 | 0.41 |
Z 3 | Copperbelt | Urban | High | 48 | 60 | 17 | 0.16 |
Z 10 | Central | Urban | High | 68 | 59 | 34 | 0.08 |
Z 12 | Copperbelt | Urban | High | 97 | 59 | 15 | 0.19 |
Z 13 | Central | Urban | High | 103 | 50 | 23 | 0.16 |
Z 2 | Copperbelt | Urban | High | 90 | 48 | 28 | 0.32 |
Z 16 | Southern | Rural | Moderate | 18 | 44 | 17 | 0.16 |
Z 9 | Southern | Urban | High | 29 | 38 | 7 | 0.22 |
Z 14 | Southern | Rural | Moderate | 28 | 29 | 11 | 0.25 |
Communities arranged from highest to lowest TB infection prevalence estimates as defined by Quantiferon-TB Gold In-Tube (QFT-GIT) test. SA: South African community; Z: Zambian community; TST (Tuberculin skin test). Geography, urban/rural and HIV prevalence as described elsewhere [22], [25].
A panel of eight experts critically examined data from ante-natal clinic surveillance, prevention of mother-to-child transmission programmes, voluntary counseling and testing clinics and provincial demographic and health survey data and made an informed decision whether to categorize HIV prevalence as ‘high’ or ‘moderate’ for each community [25].