Table 4. Subgroup analyses of studies of hospital use in the US in people with underlying diabetes: Comparing those with and without SMI.
| No. of studies | Pooled effect size (95%CI) of hospital use in people with SMI compared to those without | I2 (%) | p-value for between group differences | |||
|---|---|---|---|---|---|---|
| The effect of SMI on hospital use in people with diabetes (OR) | ||||||
| Study population | 0.0365 | |||||
| Medicaid/Medicare | 4 | 1.03 (0.86–1.22) | 95.4 | |||
| Veterans’ health | 1 | 1.00 (0.94–1.07) | -- | |||
| Insured | 3 | 1.22 (0.96–1.56) | 80.1 | |||
| Complete | 1 | 1.24 (1.07–1.44) | -- | |||
| The effect of SMI on hospital use in people with diabetes (HR) | ||||||
| Study population | 0.005 | |||||
| Veterans’ health | 1 | 1.01 (0.98–1.04) | ||||
| Insured | 1 | 1.14 (1.05–1.23) | ||||