Skip to main content

Table 4 Change in effect estimates using varying approaches to handle censored observations

From: Use of predicted vital status to improve survival analysis of multidrug-resistant tuberculosis cohorts

Model #

Covariate

Approach 1: Using initial treatment outcomes

Approach 2: Incorporating predicted vital status

Approach 3: Using actual vital status

Relative Change (Approach 1 & 3)

Relative Change (Approach 2 & 3)

Reduction in bias from using Approach 2 instead if Approach 1

HR (95% CI)

HR (95% CI)

HR (95% CI)

1

AR

0.32 (0.15, 0.69)*

0.31 (0.14, 0.66)*

0.26 (0.17, 0.41)**

22.1%

16.7%

5.4%

2

0.24 (0.10, 0.54)*

0.23 (0.10, 0.52)*

0.22 (0.14, 0.36)**

6.3%

3.2%

3.1%

  1. AR Aggressive Regimen; C: Confidence Interval; HR: Hazard Ratio; *p-value: < 0.05; **p-value: < 0.0001
  2. Model 1: Univariate
  3. Model 2: Multivariable: receipt of an aggressive regimen, age, sex, alcohol abuse/dependence, baseline comorbidities, severe clinical status, XDR-TB [used covariates found significant in previous studies for which no data were missing as to not introduce imputation or missing data problems]
Back to article page

BMC Medical Research Methodology

ISSN: 1471-2288

Contact us