| Quality assessment | No of patients | Effect | Quality | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| No of studies | Design | Risk of bias | Inconsistency | Indirectness | Imprecision | Other considerations | Shorter duration empiric antibiotics | Longer duration empiric antibiotics | Relative (95% CI) | Absolute | |
| Death (within 30 days) | |||||||||||
| 4 | randomised trials | very serious1 | no serious inconsistency | serious2 | very serious3 | none | 5/95 (5.3%) | 2/103 (1.9%) | OR 5.18 (0.95 to 28.16) | 74 more per 1000 (from 1 fewer to 339 more) | VERY LOW |
| Length of stay (Better indicated by lower values) | |||||||||||
| 1 | randomised trials | serious | no serious inconsistency | no serious indirectness | serious3 | none | 36 | 39 | - | mean 0.7 days lower (5.54 lower to 4.41 higher) | LOW |
| Duration of fever (Better indicated by lower values) | |||||||||||
| 1 | randomised trials | serious4 | no serious inconsistency | no serious indirectness | serious5 | none | 36 | 39 | - | mean 0.8 days lower (2.08 lower to 0.48 higher) | LOW |
3 of the 4 studies were not placebo-controlled and reported no detail about the method of randomisation employed, whether there was allocation concealment and no power analysis.
2 of the 4 studies were from the 1970s and used first generation antibiotic agents and all the deaths occurred in these two older trials.
Very low event rate.
Unclear allocation concealment, insufficient details about randomisation and not placebo controlled
Uncertainty in the estimate of effect, the confidence interval spans both appreciable benefit and harm.
From: Subsequent Treatment: guideline chapter seven
NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.