Twenty-six studies (n=11,273) were included: 17 RCTs (n=5,596), 3 non-randomised trials (n=2,556), 1 quasi-experimental study (n=482), 4 before-and-after studies (n=2,431) and 1 cross-sectional study (n=208).
Studies conducted in ED populations (7 studies including 2 RCTs, 1 non-randomised trial and 2 before-and-after studies).
Two of the 7 studies (1 non-randomised trial and 1 before-and-after study) reported that the interventions reduced return ED visits. One RCT reported that a long-term case-management intervention significantly increased ED visits, while 2 RCTs reported a non statistically significant short-term increase in ED visits. One RCT reported no statistically significant difference in ED visits between the intervention and control. The statistical significance of results from the seventh study (a before-and-after study) were unclear.
Studies conducted in hospital in-patients (2 RCTs, 1 non-randomised trial and 1 before-and-after study). None of the studies reported any statistically significant effect on return ED visits with the interventions.
Studies conducted in out-patients and/or primary care (9 RCTs and 1 cross-sectional study). Five of the 7 longer term (3 to 24 months) RCTs of geriatric evaluation and management reported a significant reduction in ED utilisation with the intervention compared with the control. The cross-sectional study reported a significant reduction in ED visits with the intervention. Two RCTs reported no significant reduction in ED utilisation.
Studies conducted in home care settings (2 RCTs, 1 quasi-experimental study and 1 non-randomised trial).
One RCT reported a significant reduction in the time to first ED visit with the intervention. The quasi-experimental study reported a significant reduction in ED utilisation in the control group compared with the intervention group. One RCT reported no significant difference between two models of case management. The short-term non-randomised trial reported a significant reduction in ED utilisation with the intervention.
Studies conducted in the community (1 before-and-after study). This study reported a significant reduction in ED visits after the intervention was implemented.
Comparison of ED utilisation between studies in control groups.
In general, higher rates of ED utilisation were reported in control groups in studies set in the ED and hospitals compared with non-hospital-based settings.