RESUMEN
BACKGROUND: Measles is often underreported. We evaluated the sensitivity of the measles surveillance in 2 districts of West Bengal in 2005-2006. METHODS: We sampled households with children aged <5 years in village clusters selected with probability proportional to size. We searched households door to door to identify World Health Organization-defined suspected measles cases that had occurred during 12 months in 2004-2005 in Howrah and in 2006 in Purulia. We interviewed mothers about use of health care services during episodes and calculated the proportion of patients seen in the public sector. We reviewed surveillance records at all levels to estimate the proportion of cases seen in public health care facilities that had been reported to the district. We calculated the overall sensitivity of measles surveillance by multiplying these 2 proportions. RESULTS: In Howrah, we identified 240 cases of measles. Of these, 8 (3.3% [95% confidence interval {CI}, 1.5%-6.5%]) had been seen in public facilities and recorded. Of 980 cases identified in 448 public facilities in the periphery, 962 (98%) had been transmitted to the district (overall sensitivity of surveillance, 3.2%). In Purulia, we identified 167 measles cases. Of these, 39 (23.4% [95% CI, = 17.2%-30.5%]) had been seen in public facilities and recorded. Of 418 cases identified in public facilities in the periphery, 414 (99%) had been transmitted to the district (overall sensitivity of surveillance, 23.1%). CONCLUSIONS: Measles surveillance captured a minority of measles cases, but cases captured were transmitted well to the district. Surveillance must engage the private sector. Health education focusing on vitamin A treatment for measles might provide an incentive to seek care, which could increase the sensitivity of surveillance.