Question 7: Does the adjustment for possible confounding variables change the estimate of the effect on readmission of the crisis plan?
We need the hazard ratio for the crisis plan without any adjustment. In Stata, the way to do this is to run the Cox regression with crisis plan as the only predictor. We do this exactly as in the last question, except for the covariate list, which contains only "Crisis plan".
We get the following output:
. stcox crisis
failure _d: readmit
analysis time _t: time
Iteration 0: log likelihood = -3557.9512
Iteration 1: log likelihood = -3544.9427
Iteration 2: log likelihood = -3544.0134
Iteration 3: log likelihood = -3544.0121
Refining estimates:
Iteration 0: log likelihood = -3544.0121
Cox regression -- Breslow method for ties
No. of subjects = 1030 Number of obs = 1030
No. of failures = 539
Time at risk = 526649
LR chi2(1) = 27.88
Log likelihood = -3544.0121 Prob > chi2 = 0.0000
------------------------------------------------------------------------------
_t | Haz. Ratio Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
crisis | 1.820508 .1944382 5.61 0.000 1.47666 2.244422
------------------------------------------------------------------------------
Without any adjustment, we estimate that the effect of the crisis plan is to increase readmission by a factor of 1.82 (95% confidence interval 1.48 to 2.24).
The adjusted estimate of the effect of the crisis plan was an increased risk of readmission by a factor estimated to be 1.76 (95% confidence interval 1.41 to 2.19). Hence the adjustment has made very little difference to the estimate.
Back to Exercise: Readmission to hospital for asthmatic children.
To Biostatistics in Research Practice index.
To Martin Bland's M.Sc. index.
This page maintained by Martin Bland.
Last updated: 6 February, 2008.