Table 4.
Individual belief models andremission- multivariate analysis
| Model and Variables1 |
Multivariate statistics2 |
|
|---|---|---|
| OR | 95% CI | |
|
Model for beliefs atbaseline:3 |
||
| - Explanatory model- evil spirits as causal |
0.18 |
0.04-0.75 |
| - Residence –urban |
34.07 |
2.96-391.95 |
| -Fluctuating course of initial illness |
4.18 |
1.30-13.43 |
|
Model for beliefs at12-months:3 |
||
| - Explanatory model- disease as causal |
2.29 |
1.06-4.91 |
| - Residence –urban |
2.98 |
1.13-7.66 |
| -Fluctuating course of initial illness |
3.04 |
1.27-7.24 |
|
Models for beliefs at60-months:3 |
||
| I - Explanatory model- disease as causal |
31.65 |
7.34-136.55 |
| - Residence –urban |
26.22 |
2.16-317.84 |
| - Fluctuating course of initial illness |
5.94 |
1.30-27.18 |
| II- Explanatory model- black magic as causal |
0.14 |
0.04-0.46 |
| - Residence –urban |
16.22 |
1.56-168.75 |
| - Fluctuating course of initial illness |
3.54 |
1.05-11.99 |
| III- Preferred causal explanatory model – disease model |
55.62 |
9.69-319.31 |
| - Residence –urban |
29.11 |
2.00-424.25 |
| - Fluctuating course of initial illness |
7.41 |
1.33-41.45 |
| IV - Preferred causal explanatory model – non medical |
0.02 |
0.003-0.10 |
| - Residence –urban |
29.14 |
2.00-424.25 |
| V – Belief in at least one non-medical explanatory model |
11.96 |
3.52-40.60 |
| VI - Belief in disease model plus at least one non-medical model |
61.45 |
16.27-232.04 |
| - Residence –urban |
3.63 |
1.03-12.77 |
| VII - Interaction between disease and non-medical model |
3.70 |
2.15-6.37 |
| - Residence –urban | 24.57 | 1.79-337.12 |
OR- Odds Ratio; 95%CI- 95% Confidence Interval for Odds Ratio; 1- No explanatory models significant at 6 months; 2- logistic regression; 3- model includes belief plus factors known to be associated with knowledge (age, sex, education, residence) and baseline clinical variables significantly associated with outcome (no. of meals/day, duration of untreated psychosis, fluctuating nature of initial illness).