Table 5. Logistic regression for prediction of remission* (CGI ≤ 2 upon discharge) during inpatient treatment (n=163).
| 95%CI for OR | p-value | |||||
|---|---|---|---|---|---|---|
| Independent variables† | B | SE | OR | Lower | Upper | |
| Diagnosis, depressive episode | - | - | 1.00 | - | - | - |
| Diagnosis, schizophrenia | -0.40 | 0.81 | 0.66 | 0.13 | 3.32 | 0.62 |
| Diagnosis, manic episode | 1.01 | 0.64 | 4.03 | 1.14 | 14.30 | 0.03‡ |
| Constant | 0.99 | 0.93 | 2.70 | - | - | 0.29 |
Adjusted for Clinical Global Impression Scale (CGI) scores at admission. R2 = 0.049 (Hosmer & Lemeshow); 0.072 (Cox & Snell); 0.125 (Nagelkerke). χ2 of the model: 12.16, p = 0.007.
95%CI = 95% confidence interval; OR = odds ratio; SE = standard error.
*
Remission (CGI ≤ 2 upon discharge) is the dependent variable, coded as 0 = no; 1 = yes (reference category).
†
Diagnosis: 0 = depressive episode (reference category); 1 = schizophrenia; 2 = manic episode.
‡
p < 0.05.