Table 2 Detailed information on included studies

Follow-up telephone call

 Biese et al. (2014) [20], USA

 Trial no.: NCT01207180

 Funder: Duke Endowment and the Community Connection for Seniors

 Industry sponsored: no

Study design: RCT

Intervention type: follow-up telephone call

Comparator: TAU and placebo

N: intervention = 39; control = 46; placebo = 35

Age (mean): 75 years

% female: intervention = 59%; control = 61%; placebo = 60%

22% of intervention patients revisited ED compared to 33% of placebo and 27% of control (p = 0.41)

Costs: intervention had a 70% chance of being cost saving and a 3% chance of being cost-effective and 27% chance of resulting in more return visits at higher total costs

 Biese et al. (2018) [21], USA

 Trial no.: NCT01893931z

 Funder: Duke Endowment, the Kenan Family Foundation, and Mr. John A. McNeill, Jr

 Industry sponsored: no

Study design: RCT

Intervention type: follow-up telephone call

Comparator: control group

N: intervention = 974; control = 975

Age (mean): intervention = 74.2 years; control = 73.9 years

% female: intervention = 59.1%; control = 61.1%

Rate of return to ED, hospitalization, or death: 15.5% (95% CI: 13.2–17.8) in intervention; 15.2% (95% CI: 12.9–17.5) in control, p = 0.86

Return to ED: 12.2% (95% CI: 10.1–14.3) in intervention; 12.5% (95% CI: 10.4–14.6) in control

Hospitalization within 30 days: 9% (95% CI: 7.2–10.8) in intervention; 7.4% (95% CI: 5.8–9.0) in control

Outpatient use, seeing a PCP within 30 days: 80.8% in both groups

Stratifying by age did not find a benefit

 van Loon-van Gaalen et al. (2021) [11, 33], Netherlands

 Trial no.: trial NL6598

 Funder: Jacobus Foundation

 Industry sponsored: no

Study design: RCT

Intervention type: follow-up telephone call

Comparator: control group

N: intervention = 1516; control = 1659

Age (median): intervention = median 78 years; control = median 78 years

% female: intervention = 58%; control = 58%

Unplanned hospital admission and/or ED revisit within 30 days: 16% of intervention; 14% of control, OR 1.16, 95% CI: 0.96–1.42; separate rates were not significant

Stratifying by age: patients aged < 78 years: intervention patients had more unplanned hospital admissions and/or ED revisits than control pts. (18% vs 14%, OR 1.33, 95% CI: 1.01–1.75)

Comprehensive geriatric assessment + other

 Foo et al. (2014) [31], Singapore

 Trial no.: National Healthcare Group (NHG) Domain-Specific Review Board (DSRB) C/09/023

 Funder: Ministry of Health’s Healthcare Quality Improvement and Innovation (HQI2) Fund

 Industry sponsored: no

Study design: RCT

Intervention type: geriatric assessment, referral

Comparator: TAU

N: intervention = 234; control = 425

Age (median): intervention = 77 years; control = 77 years

% female: intervention = 53.6%; control = 56.2%

ED reattendance, 3 months: 37.2% in control; 36.8% in intervention, p = 0.97, adjusted OR = 0.91, 95% CI: 0.67–1.24

ED reattendance, 6 months: 50.8% in control; 47.9% in intervention, p = 0.84, OR: 0.82, 95% CI: 0.61–1.11

ED reattendance, 9 months: 59.8% in control; 54.6% in intervention, p = 0.19, OR: 0.74, 95% CI: 0.55–1.01

ED reattendance, 12 months: 66.0% in control; 61.1% in intervention, p = 0.19, OR: 0.75, 95% CI: 0.55–1.03

Hospitalization, 3 months: 28.8% of control; 27.9% of intervention, p = 0.84, adjusted OR: 0.88, 95% CI: 0.63–1.22

Hospitalization, 6 months: 40.4% in control; 38.2% in intervention, p = 0.60, OR: 0.84, 95% CI: 0.62–1.14

Hospitalization, 9 months: 48.2% in control; 43.9% in intervention; p = 0.28, OR: 0.76, 95% CI: 0.56–1.03

Hospitalization, 12 months: 53.8% in control; 49.6% in intervention, p = 0.30, OR: 0.77, 95% CI: 0.57–1.04

 Heeren et al. (2019) [29], Belgium

 Trial no.: ISRCTN91449949

 Funder: the Flemish government agency for Innovation by Science and Technology

 Industry sponsored: no

Study design: observational study

Intervention type: geriatric assessment tool, care plan, case-manager follow-up, referral to geriatric clinic

Comparator: cohort collected prior to intervention period

N: intervention = 886; control = 794

Age (Q2): intervention = 81 years; control = 80 years

% female: intervention = 52.9%; control = 54.9%

Unplanned ED readmission, 30 days: 12.1% in control; 13.1% in intervention, p = 0.21

Unplanned ED readmission, 90 days: 22.1% in control; 23.9% in intervention, p = 0.11

Median time to unplanned ED readmission within 90 days: 25.1 days (min 0.3, max 88.3) for control; 27.6 (min 0.2, max 88.0) days in intervention, p = 0.66

Median ED LOS: 19.1 h in control; 12.7 h in intervention, p < 0.001

Hospitalization: 70.0% in intervention; 67.0% in control, p = 0.003

Median hospital LOS: 8.7 days in control; 8.6 days in intervention, p = 0.15

 Lin et al. (2021) [34], Taiwan

 Trial no.: IRB no. CE18256

 Funder: Veterans Affairs Council, Taiwan

 Industry sponsored: no

Study design: observational study

Intervention type: comprehensive geriatric assessment, case management

Comparator: TAU

N: intervention = 236; control = 122

Age (median): intervention = 82 years; control = 82 years

% female: intervention = 50.4%; control = 69.7%

ED revisits within three months: 35.3% pre-intervention; 28.4% post-intervention

Not statistically significant

Hospitalization: decreased by 27% (50.8% pre-intervention, 23.1% post-intervention)

Male gender associated with decreased ORs of admission following index ED visit

 Pedersen et al. (2016) [30], Denmark

 Trial no.: none reported

 Funder: none reported

 Industry sponsored: not reported

Study design: RCT

Intervention type: geriatric assessment, home visit by geriatrician and nurse

Comparator: TAU

N: intervention = 693; control = 637

Age (mean): intervention = 86.4 years; control = 86.4 years

% female: intervention = 60%; control = 64%

None reported

Discharges: 56% of intervention patients were discharged directly from the ED compared to 46% of control, p = 0.01

Hospitalization: 12% in intervention; 23% in control, p < 0.001

Of those admitted, LOS for intervention patients (median = 2 days) was significantly shorter than for control (median = 3 days), p = 0.03

Pharmacist-lead intervention

 Santolaya-Perrin et al. (2019) [32], Spain

 Trial no.: none reported

 Funder: RED-FASTER of SEFH (Sociedad Española de Farmacia Hospitalaria)

 Industry sponsored: no

Study design: RCT

Intervention type: medication review programme

Comparator: TAU

N: intervention = 323; control = 342

Age (mean): intervention = 78.99 years; control = 78.2 years

% female: intervention = 51.6%; control = 53.5%

Emergency visits and hospital admissions: no significant differences between groups

RR, 95% CI: 0.857, 0.652–1.126 for 3 months

RR, 95% CI: 0.917, 0.715–1.176 for 6 months

RR, 95% CI: 0.954, 0.766–1.187 for 12 months

None reported

 Shaw et al. (2016) [26], USA

 Trial name: EMBRACE ED

 Funder: Kaiser Permanente Colorado

 Industry sponsored: yes

Study design: observational study

Intervention type: pharmacist intervention

Comparator: non-clinical pharmacist + elder ED, and non-elder ED + clinical pharmacist

N: intervention = 4103 patients in total; 342 treated by EMBRACE clinical pharmacy specialist; control = 530 treated with non-clinical pharmacy specialist but in EMBRACE; 3231 non-EMBRACE

Age (mean): 77 years overall

% female: 58% overall

Patients in CPS + EMBRACE group more likely to experience a 30-day return visit than those in non-EMBRACE group (unadjusted OR 1.42, 95% CI: 1.09–1.85), and a 90-day return visit than those in the non-EMBRACE group (unadjusted OR 1.34, 95% CI: 1.06–1.69)

30-day ED return visit: 24% in intervention, 18.2% in non-EMBRACE, 18.9% in non-CPS

90-day ED return visit: 36.3% in intervention, 29.8% in non-EMBRACE, 30.6% in non-CPS

Admitted from index ED visit: 42.4% in intervention, 42.1% in non-EMBRACE, 42.5% in non-CPS; no differences

Care transition intervention

 Schumacher et al. (2021) [35], USA

 Trial no.: NCT02079987

 Funder: Patient-Centered Outcomes Research Institute award

 Industry sponsored: no

Study design: RCT

Intervention type: care transition intervention

Comparator: TAU

Definition of frequent user: 3 or more visits in prior year

N: intervention = 557; control = 544

Age (mean): intervention = 72.4 years; control = 72.8 years

% female: intervention = 60%; control = 63%

Post-intervention: 627 return ED visits made within 60 days

ED visit: OR 1.08 (95% CI: 0.83–1.39), not statistically significant

Intervention did not significantly affect hospital-based acute care, but ED return visits were less likely to result in hospitalization

Hospital admission at index ED visit: intervention participants had 36% lower odds of hospitalization upon ED return (OR 0.64, 95% CI: 0.45–0.91)

Outpatient visit: OR 1.13 (95% CI: 0.77–1.67), not statistically significant

Reasons for ED return visits: patients were confident they would get needed care in the ED, PCP often encourage patients to seek emergency care

Reasons for not visiting outpatient clinics: barriers to timely outpatient care, difficulty scheduling appointments, office-based outpatient visits discouraged if diagnostic tests were required

Physician perspective on hospital admissions: ED physicians often hospitalize patients if outpatient follow-up or social support are in question

Early assessment and intervention

 Cassarino et al. (2021) [22], USA

 Trial no.: NCT03739515

 Funder: Health Research Board of Ireland through the Research Collaborative for Quality and Patient Safety

 Industry sponsored: no

Study design: RCT

Intervention type: early assessment and intervention

Comparator: TAU

N: intervention = 176; control = 177

Age (mean): intervention = 78.6 years; control = 80.6 years

% female: intervention = 61.4%; control = 57.1%

30-day ED revisit: 18.8% for intervention; 13.6% for control, OR 1.42 (95% CI: 0.79–2.55), p = 0.23

6-month ED revisit: 31.4% for intervention; 43.3% for control, OR = 0.65 (95% CI: 0.42–1.02), p = 0.06

Time spent in ED, index visit, median (IQR), hours: 6.43 (4.05–14.87) for intervention; 12.1 (6.18–22.14) for control, p < 0.001

Hospitalization at index visit: 19.3% for intervention; 55.9% for control, p < 0.001

Unscheduled hospital admission at 30 days: 11.9% for intervention; 12.4% for control, OR 0.96 (95% CI: 0.51–1.84), p = 0.92

Unscheduled hospital admission at 6 months: 19.4% for intervention; 33.3% for control, OR = 0.52 (95% CI: 0.32–0.88), p = 0.02

Hospital LOS at index visit, median (IQR), days: 9 (3–13) for intervention; 9 (5–24) for control, p = 0.32

Satisfaction at index visit, mean (SD): 25.8 (3.03) for intervention; 24.8 (3.74) for control, p = 0.008

Physical therapy

 Lesser et al. (2018) [24], USA

 Trial no.: none reported

 Funder: none reported

 Industry sponsored: not reported

Study design: observational study

Intervention type: physical therapy services

Comparator: TAU

N: intervention = 17,791; control = 173,651

Age (mean): intervention = 82.4 years; control = 80.6 years

% female: intervention = 70.4%; control = 66.2%

30-day follow-up, all-cause ED: 20.4% in intervention; 21.7% in control

Fall-related ED revisits at 30 days: 1.7% in intervention; 2.6% for control

Fall-related ED revisits at 60 days: 2.5% in intervention; 3.6% in control

p < 0.01 for both 30 and 60 days

None reported

Multi-faceted interventions

 Arendts et al. (2018) [27], Australia

 Trial no.: ACTRN12612000798864

 Funder: State Health Research Advisory Council of Western Australia

 Industry sponsored: no

Study design: RCT

Intervention type: education, follow-up telephone call, referral

Comparator: TAU

N: intervention = 81; control = 80

Age (mean): intervention = 78 years; control = 78 years

% female: intervention = 39%; control = 37%

8% absolute (95% CI: 7–20) and 20% relative risk reduction for an intervention patient making an unplanned ED reattendance within 28 days

Hospitalization: no significant difference in 28-day hospitalization rates or hospital bed day usage

 Goldberg et al. (2020) [23], USA

 Trial no.: NCT03360305

 Funder: National Institute on Aging and Society for Academic Emergency Medicine Foundation/Emergency Medicine Foundation GEMSSTAR for Emergency Medicine Supplemental Funding

 Industry sponsored: no

Study design: RCT

Intervention type: pharmacist-led medication review; physical therapy consultation

Comparator: TAU

N: intervention = 55; control = 55

Age (median): intervention = 81.9 years; control = 80.1 years

% female: intervention = 67%; control = 67%

Total visits: control = 66; intervention = 30

Adjusted rate of all ED visits: control = 1.54 (95% CI: 1.04–2.30); intervention = 0.73 (95% CI: 0.45–1.17), IRR: 0.47 (95% CI: 0.29–0.74)

ED LOS: control = 5.3 h; intervention = 5 h, p < 0.94

Hospitalization: control = 34; intervention = 19; adjusted rate: control = 0.77 (95% CI: 0.46–1.31); intervention = 0.44 (95% CI: 0.24–0.82); IRR: 0.57 (95% CI: 0.31–1.04)

 Liberman et al. (2020) [25], USA

 Trial name: the GAP-ED project (Geriatric and Palliative Emergency Department)

 Funder: Fan Fox and Leslie R. Samuels Foundation

 Industry sponsored: no

Study design: observational study

Intervention type: care plan, education, referral

Comparator: historical usual-care group

N: intervention = 283; control = 283

Age: intervention = 11% 65–75 years, 40% 76–85 years, 44% 86–94 years, 5% 95 + ; control = 11% 65–75 years, 40% 76–85 years, 44% 86–94 years, 5% 95 + 

% female: intervention = 78%; control = 78%

Average revisits within 30 days: 0.22 in control, 0.20 in intervention, p = 0.34

In both groups, over 80% of patients had no revisits

Hospitalization: of those who revisited within 30 days, 40% of intervention were admitted, and 57% of control were admitted, p = 0.001

 Shrapnel et al. (2019) [28], Australia

 Trial no.: none reported

 Funder: the Mater Hospital Brisbane funded one full-time nursing position to implement the MACIAE study; co-author ED is funded by an Australian National Health and Medical Research Council Early Career Fellowship

 Industry sponsored: no

Study design: observational study

Intervention type: care coordination, support

Comparator: TAU

N: intervention = 391; control = 730

Age (not specified): intervention = 83.1 years; control = 84.8 years

% female: not reported

Revisit within 28 days: 4.6% for intervention; 17.8% for control, p < 0.001

Admission after ED presentation: 40.6% for intervention; 71.9% for control, p < 0.001

LOS, days, mean (SD): 1.0 (3.5) for intervention; 2.0 (3.5) for control, p = 0.840