Skip to main content

Table 2 Primary and secondary study outcomes—comparison of mortality rates between study phases

From: Effect of tuberculosis screening and retention interventions on early antiretroviral therapy mortality in Botswana: a stepped-wedge cluster randomized trial

 

ART enrollees

Deaths (n)a

Rate/100PYb

Crude HRc

(95% CI)

p

AHRcd

(95% CI)

p

Primary outcome: 6-month ART mortality in SOC versus EC+X phase

 SOC

8980

461

11.4

1.00

–

–

1.00

–

–

 EC+X

4215

121

6.3

0.58

(0.48–0.71)

< 0.001

0.77

(0.61–0.97)

0.029

Secondary outcomes: 12-month ART mortality in SOC versus EC+X phase

 SOC

8980

551

7.3

1.00

–

–

1.00

–

–

 EC+X

4215

137

4.6

0.58

(0.48–0.70)

< 0.001

0.76

(0.61–0.95)

0.014

6-month ART mortality in EC versus EC+X phasee

 EC

1768

54

6.5

1.00

  

1.00

  

 EC+X

4215

121

6.3

1.07

(0.62–1.84)

0.800

1.13

(0.63–2.03)

0.690

  1. Abbreviations: SOC standard of care phase, EC enhanced care phase, EC+X enhanced care plus Xpert phase, PY person-years, HR hazard ratio, AHR adjusted hazard ratio, CI confidence interval, XPRES Xpert Package Rollout Evaluation using a Stepped-Wedge design
  2. aRepresents deaths observed among all ART enrollees by the time point specified
  3. bRepresents unadjusted 6- and 12-month ART mortality rates among all ART enrollees in each phase of the study. For mortality rates among ART enrollees included in the adjusted analyses, see Additional file 6
  4. cAll Cox proportional hazards regression models included a random effect for clinic
  5. dAdjusted for the following characteristics at ART initiation: age, sex, pregnancy status, weight, CD4 count, hemoglobin, and ART regimen. Adjusted analysis comparing SOC versus EC+X mortality rates included 7184 SOC enrollees with 350 deaths within 6 months and 424 deaths within 12 months, and 3861 EC+X enrollees with 93 deaths within 6 months and 108 deaths within 12 months
  6. eAnalysis restricted to randomized stepped-wedge portion of the trial, fitting a Cox proportional hazards regression model to the data with the underlying time frame beginning August 2012 (the start of EC enrollment), and including a fixed effect for monthly changes in mortality rates during the first 6 months of ART. Adjusted analysis comparing EC versus EC+X mortality rates included 1653 EC enrollees with 43 deaths within 6 months and 3861 EC+X enrollees with 93 deaths within 6 months