This data set has been originally published in:
Hedeker D. and Gibbons R.D. (1996) A computer program for mixed-effects ordinal regression analysis. Computer Methods and Programs in Biomedicine 49, 157-176.
These data are from the National Institute of Mental Health Schizophrenia Collaborative Study and are available here. Patients were randomized to receive one of four medications, either placebo or one of three different anti-psychotic drugs. The protocol indicated subjects were to then be evaluated at weeks 0, 1, 3, 6 to assess severity of illness; additionally some measurements were made at weeks 2, 4, and 5. The primary outcome is item 79 on the Inpatient Multidimensional Psychiatric.
Scale which indicates severity of illness. We will analyze imps79o which is an ordinally scaled version of the original variable imps79 which has the following interpretation
|1 and 2||1 (not ill or borderline)|
|3 and 4||2 (not ill or borderline)|
|6 and 7||4 (severely or most extremely ill)|
Predictor variables of interest are TxDrug a dummy coded variable indicating treatment with drug or placebo. Below is an extract of the data set:
The columns have the following meaning:
- ID: patient identification number. It is the identifier (ID).
- IMPS: original variable imps79. This column is ignored as we use an ordinally scaled version of this variable.
- IMPSo: ordinally version of imps79. It is the measurement (Y).
- TxDrug indicates if there is a treatment or the placebo. This is a categorical covariate (CAT).
- Week: measurement period. It is the time (TIME).
One can see on the following figure the evolution of the IMPSo with respect to time on both treated patients and patients with placebo.