Máté 1970.

Methods	Cluster randomised, placebo controlled prevention and treatment trial of amantadine during the 1969 Hong Kong influenza epidemic in 7 military units in Hungary. Treatment was started as soon as the outbreak was recognised and lasted for up to 3 weeks. Randomisation was carried out by dormitory within the same barrack block
Participants	4740 healthy male soldiers aged 18 to 21 from seven different military units
Interventions	Participants were randomised to oral placebo or amantadine 100 mg twice daily as soon as the outbreak became clear. However in the units 1 to 4 the allocation procedure worked well but in the remaining units (5 to 7) the outbreak was only recognised later and some soldiers were partially treated or not treated at all. Surveillance showed that in units 5 and 7 influenza A activity was minimal and mixed with that of other agents, whereas in units 1 to 4 and 6 influenza activity was high. Because of uneven exposure to amantadine, only data for units 1 to 4 were extracted
Outcomes	Serological/laboratory: viral isolation or antibody response Paired sera (from 25 subjects from each arm) Effectiveness: cases of influenza and cases of ILI (no clear definition was given) Fever (duration and peak) Admission to infirmary Complications In the results section outcomes relating to complications and duration of fever are reported by 3 subgroups of soldiers admitted to hospital/infirmary during the trial. These groups are 'V' (amantadine recipients), 'U' (placebo recipients) and 'P' (recipients of no intervention). The complications data are reported in 'U' + 'P' groups with no breakdown, leading to loss of the data As the data relate to soldiers who are already admitted to the infirmary with influenza symptoms, this part of the trial is treatment. ILI outcome data are unclearly reported Safety: the text mentions a pre‐trial test for safety on 50 soldiers in which no adverse effects were detected but no further data are reported
Notes	The authors conclude that amantadine failed to affect morbidity but reduced peak and duration of fever, duration of stay in infirmary and the incidence of complications, especially lower respiratory tract infections. Considering its age the trial is reported reasonably but lack of clarity over allocation schedules, outcome definition and amantadine coverage has led to loss of data. In practice the trial is a mix of prevention and treatment. The analysis presented in the report did not account for the clustering of allocation by dormitory, and inadequate information is provided to allow a post‐analysis adjustment for clustering
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear