Premium
Estudios con muestreo por racimos (clusters) y con muestreo por lotes para garantizar la calidad para evaluar la cobertura de la inmunización contra la Fiebre Amarilla tras una campaña nacional en Bolivia en el 2007
Author(s) -
Pezzoli Lorenzo,
Pineda Silvia,
Halkyer Percy,
Crespo Gladys,
Andrews Nick,
Ronveaux Olivier
Publication year - 2009
Publication title -
tropical medicine and international health
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.056
H-Index - 114
eISSN - 1365-3156
pISSN - 1360-2276
DOI - 10.1111/j.1365-3156.2009.02231.x
Subject(s) - lot quality assurance sampling , cluster (spacecraft) , cluster sampling , medicine , statistics , environmental health , sample size determination , population , sample (material) , universal coverage , public health , geography , mathematics , health policy , computer science , chemistry , chromatography , programming language , nursing
Summary Objective To estimate the yellow fever (YF) vaccine coverage for the endemic and non‐endemic areas of Bolivia and to determine whether selected districts had acceptable levels of coverage (>70%). Methods We conducted two surveys of 600 individuals (25 × 12 clusters) to estimate coverage in the endemic and non‐endemic areas. We assessed 11 districts using lot quality assurance sampling (LQAS). The lot (district) sample was 35 individuals with six as decision value (alpha error 6% if true coverage 70%; beta error 6% if true coverage 90%). To increase feasibility, we divided the lots into five clusters of seven individuals; to investigate the effect of clustering, we calculated alpha and beta by conducting simulations where each cluster’s true coverage was sampled from a normal distribution with a mean of 70% or 90% and standard deviations of 5% or 10%. Results Estimated coverage was 84.3% (95% CI: 78.9–89.7) in endemic areas, 86.8% (82.5–91.0) in non‐endemic and 86.0% (82.8–89.1) nationally. LQAS showed that four lots had unacceptable coverage levels. In six lots, results were inconsistent with the estimated administrative coverage. The simulations suggested that the effect of clustering the lots is unlikely to have significantly increased the risk of making incorrect accept/reject decisions. Conclusions Estimated YF coverage was high. Discrepancies between administrative coverage and LQAS results may be due to incorrect population data. Even allowing for clustering in LQAS, the statistical errors would remain low. Catch‐up campaigns are recommended in districts with unacceptable coverage.