Table 5.

Pneumonia case management: characteristics of included studies

Study identifiers, context, and population				Study design and limitations
Study (first-named author, year of publication)	Country	Study years	Age of study population	Study design	Sample size and total number of clusters	Ascertainment of deaths	Co- interventions	Intervention coverage
Studies with cause-specific data
Mtango, 1986 [28]	Tanzania	1983–85	0–59 mo.	Cluster RCT	16 126 children, 16 clusters	Reporting by VHWs, yearly census, VA (2 wk recall)	None	NR
Pandey, 1991 [29]	Nepal	1984–87	0–59 mo.	Step wedge	2393 children, number of clusters unclear	Village-based registration with vital events confirmed by interviewers, VA (1 month recall)	None	0·85 antibiotic treatments per child-yr
Fauveau, 1992 [25]	Bangladesh	1988–89	0–59 mo.	Quasi- experimental	31 632 children, 2 units of analysis	DSS, reporting by CHWs, VA	Diarrhoea management, EPI	NR
Bang, 1990 [24]	India	1988–89	0–59 mo.	Quasi- experimental	10 122 children, 2 units of analysis	Reporting by VHWs, bi-yearly census, VA	None	76%, calculated based on expected number of pneumonia cases
Khan, 1990 [26]	Pakistan	1985–86	0–59 mo.	Quasi- experimental	4978 children, 38 clusters	Yearly census, active surveillance by CHWs in intervention area, quarterly census in comparison area, VA	Maternal health education	NR
Studies with all-cause data
Kielmann, 1978 [27]	India	1970–72	0–59 mo.	Quasi- experimental	2735 children, 10 clusters	Vital registration via CHW reports, bi-monthly visits to key informants, registration by key informants	None	NR

CHWs, Community health workers; DSS, demographic surveillance site; EPI, expanded programme on immunizations; NR, not reported; VA, verbal autopsy; VHWs, village health workers.