Skip to main content

Table 1 Simulation results, data generated with a log-normal backward recurrence density

From: A new likelihood model for analyses of pharmacoepidemiologic case–control studies which avoids decision rules for determining latent exposure status

   σ = exp(-0.35) σ = exp(-0.25)
Analysis method(*) Relative bias (%) SE log(OR) Coverage 95% CI (%) VIF Relative bias (%) SE log(OR) Coverage 95% CI (%) VIF
μ = log(1.5/12) 1:1 CC         
True expo 0.0 0.027 95.6 1.00 0.0 0.027 95.3 1.00
CC WTD 0.4 0.029 95.1 1.15 0.8 0.029 93.5 1.19
WTD prob 17.0 0.037 0.0 1.91 17.1 0.038 0.0 2.00
90 days -6.7 0.028 24.3 1.08 -7.8 0.028 14.2 1.12
30 days -12.1 0.037 5.4 1.90 -12.9 0.038 3.8 2.01
1:10 CC         
True expo 0.0 0.036 95.1 1.00 0.0 0.036 94.5 1.00
CC WTD 0.3 0.038 95.0 1.09 0.7 0.038 94.2 1.11
WTD prob 8.7 0.046 46.4 1.61 8.3 0.047 51.4 1.66
90 days -7.5 0.038 40.0 1.06 -9.1 0.038 25.6 1.08
30 days -18.2 0.046 1.0 1.60 -19.2 0.047 0.7 1.69
μ = log(2/12) 1:1 CC         
True expo 0.0 0.027 94.4 1.00 0.0 0.027 94.4 1.00
CC WTD 1.0 0.029 93.6 1.19 1.7 0.030 90.2 1.24
WTD prob 14.4 0.036 0.8 1.84 14.1 0.037 1.3 1.91
90 days -8.1 0.029 14.1 1.18 -9.2 0.030 7.6 1.23
30 days -13.9 0.041 3.8 2.35 -14.5 0.042 3.9 2.48
1:10 CC         
True expo 0.0 0.036 94.8 1.00 0.1 0.036 94.7 1.00
CC WTD 0.8 0.038 94.6 1.11 1.4 0.039 93.6 1.14
WTD prob 6.7 0.045 62.6 1.56 6.1 0.046 68.9 1.60
90 days -10.2 0.038 17.2 1.12 -11.7 0.039 8.2 1.15
30 days -21.1 0.051 0.5 1.94 -21.5 0.052 0.3 2.03
  1. The datasets had a sample size of 39,600, and on average 80% of patients continue treatment at the index date, 25% of patients have a prescription redemption in the year before the index date and the true OR is 3. For each setting 2,500 datasets were generated and analyzed. \(\mu\) and \(\sigma\) are parameters of the assumed Log-Normal Backward Recurrence Density used for generating data, see text for details
  2. (*) Analysis methods:
  3. 1:1 CC indicates 1 control per case, 1:10 CC indicates 10 controls per case
  4. True expo – logistic regression with the actual exposure status as covariate (the reference analysis)
  5. CC WTD – estimation based on joint likelihood for case–control status and the reverse WTD, Log-Normal Backward Recurrence Density
  6. WTD prob – a reverse WTD with Log-Normal Backward Recurrence Density is estimated to predict the probability of an individual being exposed and this exposure probability is used as covariate in logistic regression
  7. 90 days – individuals are considered exposed if they have a redemption < 90 days before index date
  8. 30 days – individuals are considered exposed if they have a redemption < 30 days before index date, logistic regression