Skip to main content
. 2017 Mar 31;66(12):339–342. doi: 10.15585/mmwr.mm6612a6

TABLE 2. Difference in proportion of bacteriologically confirmed* TB cases before and after Xpert MTB/RIF installation, and Bootstrap mean difference estimates and 95% CIs for treatment outcomes, by health facility (N = 6) — Kampala, Uganda, 2012 – 2015.

Characteristics Health facility
Alive Medical Services Kisenyi Health Center IV Kisugu Mengo Mulago Ward 5 and 6 Nsambya Hospital
Difference in proportion of bacteriologically confirmed* TB cases %, (95% CI)
8.3 (−3.1 to 29.8)
30.8 (21.3 to 40.2)
14.9 (−3.8 to 33.3)
−10.1 (−26.3 to 6.3)
−1.7 (−12.6 to 9.3)
5.1 (−14.0 to 24.2)
Bootstrap mean difference estimates (95% CI)§ for TB treatment outcomes
TB treatment completed
−0.119 (−0.357 to 0.119)
0.184 (0.059 to 0.307)
−0.153 (−0.364 to 0.056)
−0.012 (−0.130 to 0.097)
−0.064 (−0.169 to 0.040)
−0.728 (−0.839 to −0.598)
Stopped TB treatment before completion
0.000 (−0.214 to 0.214)
−0.179 (-0.292 to 0.063)
0.071 (−0.105 to 0.249)
0.040 (−0.056 to 0.149)
0.030 (−0.063 to 0.125)
−0.018 (−0.134 to 0.098)
Continuing TB treatment 0.119 (−0.048 to 0.298) −0.006 (−0.081 to 0.074) 0.082 (−0.051 to 0.233) −0.028 (−0.079 to 0.031) 0.034 (−0.031 to 0.102) 0.746 (0.608 to 0.866)

Abbreviations: CI = confidence intervals; TB = tuberculosis.

* Bacteriologically confirmed TB includes cases diagnosed using either GenXpert or culture.

Statistically significant (p≤0.05).

§ Bootstrap percentile CIs using 100,000 samples per model.