Six RCTs compared outpatient versus in-patient treatment (738 participants, range 80 to 170) and eight compared oral versus intravenous administration of outpatient antibiotics (857 participants, 41 to 177). Six studies reported allocation generation and five reported allocation concealment. No studies reported blinding. Information on withdrawals was available from nine studies. Intention-to-treat analysis was reported in four studies.
There were no significant differences between in-patient and outpatient treatment regimens for treatment failure (RR 0.81, 95% CI 0.55 to 1.19; six studies). The were no significant differences between groups for absolute risk reduction. There were also no significant differences in treatment failure between oral compared with intravenous administration of antibiotics (RR 0.93, 95% CI 0.65 to 1.32; eight studies). There were no significant differences between groups for absolute risk reduction. Neither analysis contained substantial heterogeneity.
There were no significant differences between in-patient and outpatient management for all secondary outcomes (mortality, toxicity and readmission). No differences were observed between paediatric and adult studies. Sensitivity analyses that accounted for differences in study design did not change the direction or significance of the results. Publication bias was not assessed due to the small number of included trials.