RESUMEN
BACKGROUND: First dose oral co-trimoxazole and referral are recommended for WHO-defined severe pneumonia. Difficulties with referral compliance are reported in many low-resource settings, resulting in low access to appropriate treatment. The objective in this study was to assess whether community case management by lady health workers (LHWs) with oral amoxicillin in children with severe pneumonia was equivalent to current standard of care. METHODS: In Haripur district, Pakistan, 28 clusters were randomly assigned with stratification in a 1:1 ratio to intervention and control clusters by use of a computer-generated randomisation sequence. Children were included in the study if they were aged 2-59 months with WHO-defined severe pneumonia and living in the study area. In the intervention clusters, community-based LHWs provided mothers with oral amoxicillin (80-90 mg/kg per day or 375 mg twice a day for infants aged 2-11 months and 625 mg twice a day for those aged 12-59 months) with specific guidance on its use. In control clusters, LHWs gave the first dose of oral co-trimoxazole (age 2-11 months, sulfamethoxazole 200 mg plus trimethoprim 40 mg; age 12 months to 5 years, sulfamethoxazole 300 mg plus trimethoprim 60 mg) and referred the children to a health facility for standard of care. Participants, carers, and assessors were not masked to treatment assignment. The primary outcome was treatment failure by day 6. Analysis was per protocol with adjustment for clustering within groups by use of generalised estimating equations. This study is registered, number ISRCTN10618300. FINDINGS: We assigned 1995 children to treatment in 14 intervention clusters and 1477 in 14 control clusters, and we analysed 1857 and 1354 children, respectively. Cluster-adjusted treatment failure rates by day 6 were significantly reduced in the intervention clusters (165 [9%] vs 241 [18%], risk difference -8·9%, 95% CI -12·4 to -5·4). Further adjustment for baseline covariates made little difference (-7·3%, -10·1 to -4·5). Two deaths were reported in the control clusters and one in the intervention cluster. Most of the risk reduction was in the occurrence of fever and lower chest indrawing on day 3 (-6·7%, -10·0 to -3·3). Adverse events were diarrhoea (n=4) and skin rash (n=1) in the intervention clusters and diarrhoea (n=3) in the control clusters. INTERPRETATION: Community case management could result in a standardised treatment for children with severe pneumonia, reduce delay in treatment initiation, and reduce the costs for families and health-care systems. FUNDING: United States Agency for International Development (USAID).