Table 3. Estimates of the sensitivity and positive predictive value of CBS systems.
| Country [ref] | Methodology | Sensitivity calculation | Positive predictive value calculation | Period of interest | Scope | Sensitivity estimate | Positive predictive value estimate |
|---|---|---|---|---|---|---|---|
| Benin [42] | Cross-sectional household survey: in 2011 surveyors visited all households covered by the CBS system to collect the same information as collected by the CBS system in 2010. | Not specified | / | 2010 | Maternal death | 95% | / |
| Infant death | 47% | ||||||
| Under 5 death | 48% | ||||||
| Cambodia [4] | Cross-sectional household survey: surveyors visited households (in 3 out of 7 areas implementing CBS) to collect cases of diseases (preceding month) and vital events (preceding year), using the same case definitions as used by the CBS system. The survey was conducted once. For measles: outbreak investigation data. |
(No. of cases/events identified both by the CBS system and the household survey) / (No. of cases or events identified by the household survey) | (No. of cases or events identified both by the CBS system and household survey) / (No. of cases identified by the CBS system) | One year (2000–2001) | Measles | 93% (n = 86/92) | 90% (n = 86/96) |
| Birth | 82% (n = 28/34) | 100% (n = 28/28) | |||||
| One month (2001) | Severe diarrhoea | 82% (n = 10/12) | 82% (n = 10/12) | ||||
| Chronic cough | 75% (n = 55/73) | 89% (n = 55/62) | |||||
| Malaria | 65% (n = 57/88) | 88% (n = 57/65) | |||||
| Ethiopia [43] | Cross-sectional survey: in randomly selected villages implementing the CBS system, the blood of suspect malaria cases identified by CBS actors were tested to confirm malaria. | / | (No. of malaria cases confirmed with blood test) / (No. of suspected malaria cases identified by the CBS system) | 1995–1996 | Malaria | / | 93% (n = 1453/1562) |
| Nigeria [44,45] | Confirmatory follow-up visits by an investigator in the villages reported having new cases as well as villages reported having zero cases. | (No. of villages reported confirmed as having cases through the follow-up visit) / (Total No. of villages with verified cases of guinea worm) | (No. of villages reported confirmed as having cases through the follow-up visit) / (Total No. of villages with cases reported by the CBS system). | 6 months (1990–1991) | Guinea-worm disease | 79% (n = 50/63) | 93% (n = 50/54) |
| Sierra Leone [46] | Suspected cases were confirmed by laboratory diagnostic test | (No. of confirmed cases detected by the CBS system) / (Total No. of confirmed cases identified in the area). | (No. of confirmed cases detected by the CBS system) / (Total No. of cased detected by CBS system (suspected, probable and confirmed)). | 7 months (2015) | Ebola virus disease | 30% (n = 16/53) | 6% (n = 16/287) |
| Sweden [47] | One week recall survey: a sample of participants in the CBS system was sent a questionnaire to collect the occurrence of influenza like illness in the previous week. Each year each participant went through two-three validation surveys. | / | (No. of participants who reported having influenza like illness in both the CBS system and one-week recall survey) / (No. of participants who reported having influenza like illness in the CBS system) | Two 8-week period (in 2008 and 2009) | Influenza like illness | / | 2008: 79% (n = 73/92); 2009: 88% (n = 70/80) |
| Tanzania [48] | Cross-sectional survey: investigators visited and searched for mosquito larvae habitat in randomly selected housing clusters (consisting of 20–100 houses) covered by CBS system. | (No. of mosquito larvae habitat identified by the CBS system in the areas covered by cross-sectional survey) / (No. of mosquito larvae habitat reported by investigator during the cross-sectional survey) | / | 8 months (2007–2008) | Mosquito larvae habitat | 66.2% (n = 1963/2965) | / |