Table 3.
Binomial logistic regression analysis to determine whether tri-morbidity acts as a predictive risk for unplanned hospital admissionsa
| P-value | OR | 95% CI | |
|---|---|---|---|
| Tri-morbidity | 0.00 | 4.19 | 1.90 to 9.27 |
Cases and controls were matched for age, sex, type of homelessness, and duration of registration with HGPS and so these variables were not added to the model. There was a matching error with the sex of the cases and controls. CI = confidence interval. HGPS = homeless GP surgery. OR = odds ratio.