Gurwitz 2014.

Study characteristics
Methods	RT, total study duration not provided
Participants	5077 hospital discharges: 2563 intervention discharges, 2514 control discharges Final sample: 1870 intervention, 1791 control Setting: large multispecialty group practice Diagnostic criteria: ≥ 65 years discharged from hospital to home
Interventions	Professional intervention Intervention: an automated system was developed to facilitate the flow of information to the medical group’s primary care providers about individuals who were discharged to home from the hospital. In addition to notifying providers about an individual’s discharge, the system provided information about new drugs at the time of hospital discharge, warnings about selected drug‐drug interactions, recommendations for consideration of dose changes and laboratory monitoring of high‐risk medications, and alerts to the provider’s support staff to schedule a post hospitalisation office visit within 1 week of discharge. Control: care as usual
Outcomes	Whether discharged individuals had an office visit with a primary care physician in the 7‐, 14‐, and 30‐day periods after hospital discharge was determined, as was whether a participant was rehospitalisation within 30 days. Information related to office visits and hospitalisations was ascertained from the medical group’s electronic health record and from health plan data, which allowed for determination of whether a rehospitalisation had occurred at any hospital and not just the primary hospital that served individuals under the care of the medical group. Analysts blinded to intervention status determined these outcomes at least 6 months after completion of the study.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	A random number generator was used to assign a discharge to the intervention or control group.
Allocation concealment (selection bias)	Low risk	Computer allocated discharges
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Automated system was used. Also participants were not aware of which group they were in
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comparable rates of attrition in intervention and control groups.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Analysts blinded to intervention status determined the outcomes. (The trialist reviewing the data (JHG) was unaware of which type of unit the event had occurred on.)
Selective reporting (reporting bias)	Low risk	No evidence of selective reporting.
Protection against contamination bias	Unclear risk	The study is a cluster design and the authors stated the following, "Efforts were made to limit crossover of prescribers between intervention and control units, however, some prescribers worked simultaneously on both intervention and control units."
Other bias	Unclear risk	Unclear risk