Skip to main content

Table 8 Analysis of changes in ANC HIV prevalence trend and the impact of program intervention using various imputation approaches and analysis model: logit(p st ) = β 0 + β 1 t + β 2 I + β 3 I × t + α 0s  + α 1s t

From: A comparison of missing data procedures for addressing selection bias in HIV sentinel surveillance data

Methods

β ^ 0

β ^ 1

β ^ 2

β ^ 3

Complete-case (mixed-effects)

165.09 (117.21, 212.98)

−0.085 (−0.11, -0.06)

56.21 (−4.43, 116.86)

−0.027 (−0.058, 0.003)

Model 1

176.10 (136.91, 215.29)

−0.091 (−0.13, -0.05)

46.95 (−3.66, 97.55)

−0.023 (−1.16, 0.03)

Model 2

448.23 (−1932.11, 2828.57)

−0.229 (−2.77, 2.31)

550.47 (−3602.30, 4703.24)

−0.274 (−4.75, 4.20)

Model 3

402.2 (358.89, 445.42)

−0.203 (−0.25, -0.16)

−148.32 (−206.24, -90.40)

0.074 (0.01, 0.14)

  1. 1. National Aids Control Organisation and Government of India Ministry of Health Family Welfare: ANNUAL REPORT 2008–2009. In. New Delhi; 2009.