Table 1.
Study characteristics
| Study (Yr) | Country | Setting | Number of subjects in the study | Intervention period | Comparator |
|---|---|---|---|---|---|
| Crane (2012) | USA | 200 bed not for profit hospital. | 36 (from initial group of 255) | July 2009–June 2010 | Pre and post |
| DeHaven (2012) | USA | 1 ED in Dallas | 574 (265 intervention and 309 controls) | April 2003–July 2004 | Controls |
| Edgren (2016) | Sweden | 5 counties in Sweden. | Group One used Zelen’s design and had 7280 in intervention group and 3508 in control group. Group two used RCT design and had 934 in intervention group and 459 in control group. | 2010- March2014. | Controls |
| Enard (2013) | USA | Nine Memorial Hermann EDs in the Houston area | 13,642 participants (1905 in intervention and 11,373 in control group). | Nov 2008- April 2011 | Pre and post |
| Grimmer-Somers (2010) | Australia | One metropolitan health region | 37 patients | 18 months from 2007 to 2009 | Pre and post |
| Grover (2018) | USA | ED community hospital with 225 bed hospital in suburban area. | 158 in intervention. | Oct 2013- June 2015 | Pre and post. |
| Hardin (2017) | USA | Inner city tertiary hospital. 80,000 annual ED visits. | 339 in intervention | Nov 2012-Dec 2015 | Pre and post |
| Lin (2017) | USA | Brigham and Women’s hospital. Large urban acaedemic medical centre | 72 (36 in intervention and 36 in control) | Oct 2014-April 2015 | Controls. |
| Murphy (2013) | USA | Regional medical and trauma centre. 644 beds and 80,000 ED visits per yr. | 141 (76 extreme and 65 frequent ED users) | Jan 2008- Dec 2010 | Pre and post. |
| Navratil-Strawn (2014) | USA | Patients enrolled in an insurance scheme. | 14,140 (7070 participants and an equal number of controls) | June-Nov 2011 | Controls. |
| Okin (2000) | USA | San Francisco General Hospital. | 53 patients. | June 1995–June 1996 | Pre and post . |
| Reinius (2012) | Sweden | Karolinska University Hospital with 90,000 visits per yr. | 268 patients (211 in intervention and 57 in control group). | Sep 2010- Sep 2011 | Controls |
| Seaberg (2017) | USA | Urban ED with 57,000 presentations per yr. | 318 (184 in intervention and 134 controls). | Jan- June 2015 | Controls. |
| Shumway (2008) | USA | San Francisco General Hospital. Sole Level 1 trauma hospital in the county. | 252 (167 in intervention and 85 to usual care) | March 1997 to Feb 1999 | Controls. |
| Stokes-Buzzelli (2010) | USA | Urban hospital with 95,000 ED presentations per yr. | 36 patients | June 2005–July 2007 | Pre and post |
| Tadros (2012) | USA | One urban hospital. | 51 patients | Dec 2006-June 2009 | Pre and post |
The Evidence Project risk of bias tool [26] showed that the majority of the papers were cohort studies. Five of the studies randomly assigned participants. One study reported on follow up rates, in the other papers follow up was either not applicable because of study design or not reported. The table is included in the Additional file 1