Eleven studies (n = 6,611) were included in the review: five were RCTs (n = 3,557); four were CCTs (n = 922); and two were case-control studies (n = 2,127). Sample sizes ranged from 61 to 1,531. Study validity was described as variable. Follow-up ranged from one to 12 months.
Functional decline (two RCTs). The pooled OR indicated a statistically significantly slower rate of functional decline at discharge for patients in geriatric units than those in conventional units (OR: 0.82, 95% CI: 0.68, 0.99, p = 0.04, two RCTs). Inclusion of a non-randomised study produced a similar result, but with increased statistical heterogeneity. There were no differences between the groups reported at three-month follow up.
Living at home (five RCTs). There was a statistically significantly higher likelihood of patients living at home after discharge (OR 1.30, 95% CI: 1.11, 1.52, p =0.001, five RCTs), which was maintained with the inclusion of non-randomised studies (number unclear). There was a trend towards a significant difference at three-month follow-up (OR 1.16, 95% CI: 0.99, 1.37, p = 0.07, four RCTs).
Case fatality (five RCTs). There were no significant differences between the groups in case fatality incidence either at discharge (five RCTs) or at three-month follow-up (four RCTs).
Hospital stay (11 studies, five RCTs). Trends towards reduced length of stay were observed in nine studies, but no statistically significant differences were observed and heterogeneity was statistically significant.
Readmission to hospital (two RCTs). There were no significant differences between groups in readmission within three months.
The inclusion of non-RCTs did not significantly alter the results.