| Methods |
Triple‐blind, placebo‐controlled RCT |
| Participants |
660 participants, mean age 31; 55% female. Productive cough for < 2 weeks, no serious medical comorbidity, and no antibiotic treatment in previous 2 weeks. All participants had HIV test and chest X‐ray at baseline. Excluded if chest X‐ray showed pneumonia or tuberculosis |
| Interventions |
Amoxicillin 500 mg 3 times a day for 7 days versus identical placebo tablet |
| Outcomes |
Clinical cure at 14 days as measured by > 75% reduction in Acute Bronchitis Severity Score |
| Notes |
Reported as first study of acute bronchitis treatment that used an equivalence design. Data available for HIV‐positive patients but not included in the review. |
| Risk of bias |
| Bias |
Authors' judgement |
Support for judgement |
| Random sequence generation (selection bias) |
Low risk |
Participants were randomised independently using a random number generator. |
| Allocation concealment (selection bias) |
Low risk |
Antibiotic or placebo tablets identical in appearance, taste, and smell were placed in identical sealed, opaque containers identifiable only with a unique study identifier. |
| Blinding (performance bias and detection bias)
All outcomes |
Low risk |
All clinical and research staff were blinded to the allocation of participants, and the allocation schedule was kept in the office of the Chief Research Pharmacist in the host institution. |
| Blinding (performance bias and detection bias) |
Low risk |
|
| Incomplete outcome data (attrition bias)
All outcomes |
Low risk |
> 85% follow‐up for outcome data |
| Selective reporting (reporting bias) |
Unclear risk |
No access to original protocol, though selective reporting not apparent from trial description. |
| Other bias |
Low risk |
|
