Table 1 Tabulation and description of the studies used in the systematic review

1

Abraham et al. (2017)

USA, 1 urban psychiatric hospital

Pharmacist involvement to improve care co-ordination

16 health professionals, 6 patients, 20 patient charts (SMI patients)

Evaluation- interviews and observations of charts

Increased pharmacist involvement in LAI care coordination may contribute to bridging gaps in medication adherence to optimize

treatment outcomes.

To support long-term medication adherence and patient outcomes

2

Attfield et al. (2017)

UK, 2 trusts

Diagnostic-driven Integrated Care Pathways (ICPS)

A random sample of 400 service users

Retrospective case comparison study

The ICP Trust had a 13.5 day shorter average length of stay, (statistically significant). No significant differences in readmission or 7-day follow-up.

Reducing unnecessary tests, interventions and duplication within the care process

3

Bauer et al. (2012)

Germany, 1 hospital

SMS-based

maintenance intervention

165 females. Eating disorders

RCT

Somewhat significant difference in readmission (depending on analysis). Significant difference in treatment utilisation.

Maintain treatment

4

Bennewith et al. (2014)

UK, 3 inpatient wards in southwest England, mixed urban/rural

A contact-based intervention for people recently discharged (letters sent to sus)

102 patients received a letter, 45 received all letters

Pilot case study. Interviews, analysis of outcomes (readmission)

Post-discharge, qualitative interviews with service users showed that most already felt adequately supported and the intervention added little to this.

To reduce suicide post-discharge by providing social connectedness

5

Bonsack et al. (2016)

Switzerland, 1 psychiatric hospital

Transitional case management

51 intervention, 51 control

RCT

Increased short-term rate of engagement with ambulatory care despite no differences

between the two groups after 3 months of follow-up. Intervention did not significantly reduce the rate of readmissions during the first year following discharge.

Improve engagement with care, reduce readmission

6

Botha et al. (2018)

South Africa, 1 hospital

90-day transitional care intervention (four phone calls and one home visit, focusing on maintaining adherence, appointment reminders and psychoeducation.)

60 male patients

Retrospective comparison to matched control group

No effect on readmission rates in this setting.

Bridge gap, reduce readmissions

7

Chen (2014)

USA, all of the community agencies providing CTI in NYC (4)

Community support in critical time intervention (CTI)a time-limited, short-term psychosocial rehabilitation.

Program designed to facilitate the critical transition from

institutional to community settings

12 CTI workers

Interviews

CTI workers self-identified as “extra support” to develop community ties that will help clients sustain stable housing. Propose a transient triangular relationship model, involving three dyadic relationships (worker-client, worker-primary support, primary support client).

To facilitate effective transitional services and enhance continuity of care. Breaking the vicious cycle between institutionalization and homelessness

8

D’Souza (2002)

Australia, rural hospital

Telemedicine (psycho-educational programme and MDT videoconferencing post-discharge)

51 (24 intervention, 27 control) male and female

Controlled study

More side effects in control group, more treatment adherence and satisfaction in intervention group.

Improve treatment adherence

9

De Leo and Heller (2007)

Australia, 1 psychiatric inpatient unit

Intensive case management follow up of high risk people (ICM was weekly face-to-face contact with community case manager and telephone calls from counsellors)

60 male service users with a history of suicide attempts

RCT (TAU or intervention)

People in ICM had lower depression scores, suicidal ideation, QoL, more contact with services, better relationships with therapists and were satisfied with service.

A solution to the reduced care following discharge that is linked to suicide.

10

Exbrayat et al. (2017)

France, single centre

Telephone

follow-up 8,30 and 60 days post attempted suicide

436 patients (387 control patients who were matched from pre-intervention records)

Controlled study

Very early telephone follow-up of our patients effectively reduced recidivism and seemed to be the only protective factor against repeated suicide attempt.

To reduce suicide attempts post-discharge

11

Forchuk et al. (2005)

Canada, 26 wards, 4 hospitals

Transitional discharge model (TDM)

390

Randomised clinical trial using a cluster design

Costs and quality of life were not significantly improved post-discharge compared with the control group. Although not predicted a priori, intervention subjects were discharged an average of 116 days earlier per person.

Reduce bed occupancy, improve quality of life

12

Forchuk et al. (2008)

Canada, 1 hospital

Intervention to prevent homelessness- immediate assistance in accessing housing and assistance in paying their first and last month’s rent

14 participants at risk of being discharged without housing (7 in intervention group)

RCT, incl. interviews

All intervention group maintained housing after 3 and 6 months. All but one individual in the control group remained homeless after 3 and 6 months. Results of this pilot were so dramatic that randomizing to the control group was discontinued

To reduce discharge from inpatient wards to shelters or the street

13

Forchuk et al. (2012)

Canada, 6 hospitals

Transitional relationship model (TRM) (providing hospital staff involvement until a therapeutic relationship has been established with a community care provider as well as peer support.)

No participant numbers as ethnographic analysis. 14 A wards, 12 B wards and 10 C wards.

A quasi-experimental, action-oriented

research design

Staged large-scale implementation allowed for iterative improvements to the

model leading to positive outcomes. This study highlights the need to address work environment issues, particularly inter-professional teams.

To improve staff uptake of interventions

14

Forchuk et al. (2013)

Canada, all patients in Ontario at risk of homelessness, 1 acute care hospital, 1 territory

Intervention to prevent homelessness -

Pre-discharge assistance in securing housing

112 men and 107 women at risk of homelessness post-discharge

Programme evaluation design- interviews, focus groups

The results highlight several benefits of the intervention and show that homelessness can be reduced by connecting housing support, income support, and psychiatric care.

To stop people being discharged to street or shelters

15

Ghadiri Vasfi et al. (2015)

Iran, 1 hospital

Aftercare Services (three components: follow-up

Care (home visits or telephone follow-up), family psychoeducation, And social skills training for patients.)

120 patients (schizophrenia and bipolar) ages 15–65. 60 control

RCT

The cumulative number of hospitalizations during the follow-up period was 55 for the control group and 26 for the intervention group. Length of stay was significantly greater in the control group. Psychopathology was significantly less severe in intervention group compared with the control

Reducing readmissions and length of stay

16

Hampson et al. (2000)

UK, 1 trust (North Nottingham and Hucknall)

Community Link Team (CLT) to facilitate early discharge- team-based service offering intensive support during the day

142 (all admissions to team in 12 month period)

Retrospective comparison

Median length of stay during CLT project was 19 days, a highly significant reduction from 36 days in the NABUS study. Cannot be attributed to team but justifies a RCT to test this hypothesis,

To speed up discharge due to costs to provider and patients

17

Hanrahan et al. (2014)

USA, 1 hospital

Transitional care model (TCM)

40 (20 control)

RCT

The intervention group showed higher medical and psychiatric rehospitalisation than the control group. Emergency room use lower for intervention group but not statistically significant. Continuity of care with primary care appointments were significantly higher for the intervention group. The intervention group’s general health improved but was not significant

Reduce transition failures

18

Hegedus et al. (2018)

Switzerland, 2 wards, 1 hospital

Short transitional intervention in psychiatry (step)

14 control, 15 intervention

Quasi-experimental pilot study to determine the feasibility of the intervention,

The intervention did not affect primary or secondary outcomes; however, it was shown to be feasible, and patients’ feedback highlighted the importance of post-discharge contact sessions.

Prepare patients for situation outside of hospital

19

Hengartner et al. (2015)

Switzerland, 1 catchment area, which is an urban/suburban area of high-level resources near the city of Zurich

Post-discharge network coordination

3 patients

Case studies- narrative review and qualitative analysis of three patients who participated in the program

Case reports revealed that patients’ social networks are small and their

relationships are commonly conflictual and unstable.

Reducing readmission, improving mental health and psychosocial functioning. Improve hospital discharge planning and to ease the transition

20

Hengartner et al. (2016)

Switzerland, 1 catchment area, which is an urban/suburban area of high-level resources near the city of Zurich

Post-discharge network coordination

151 patients

RCT using parallel group blocking

In the short-term, no significant effect emerged in any outcome. In the long term the two groups did not differ significantly with rate and duration of rehospitalisation. The intervention did not reduce psychiatric symptoms, did not improve social support, and did not improve quality of life.

Reducing readmission, improving mental health and psychosocial functioning. Improve hospital discharge planning and to ease the transition

21

Herman et al. (2011)

USA, 2 transitional residences in hospital grounds metropolitan area

Critical Time Intervention to prevent homelessness

150 previously homeless men and women with SMI

RCT

CTI group had less homelessness than TAU

Reduce homelessness following discharge

22

Jenson et al. (2010)

Canada, poor city, high unemployment, 1 acute ward and 1 community service provider within same region

Community-Based Discharge Planning (in-reach model- discharge planner based in community visits ward daily)

36 service users

Single group programme evaluation, analysis of admin data and interview with clients

Readmission rates were 40% lower in the year following

the change in service delivery model. This change was statistically significant.

To shift mental health services from institution to community

23

Juven-Wetzler et al. (2012)

Israel, 1 ward

“Continuation of Care” model (continuing follow-up in the ward, by the same staff, instead of being referred to the outpatient department.)

35 service users

Pre and post within participant design

The number of hospitalizations in the 18 months following the index hospitalization was 1.79 _ 3.51 as compared to 4.67 _ 1.79 before the index hospitalization (p = 0.0002), and the number of days of hospitalization

18 months after was 24 _ 41.65 as compared to 119.71.

To reduce length of stay and readmission

24

Kariel-Lauer (2000)

Israel, 1 hospital

Re-entry group (short-term group meetings- psychoeducational approach)

75 participants (42 in intervention) men and women

A controlled study

Intervention group had less readmissions, high rates of absorption into therapy and remaining in therapy

Reduce hospitalisations, increase compliance with outpatient appointments

25

Kaspow and Rosenheck (2007)

USA, 8 veteran medical centres

Critical time intervention

Case management (a modification of the critical time intervention

(CTI) community case management model)

278 control cohort, 206 intervention cohort

Nonrandomized pre–post cohort design

19% more days housed in each 90-day reporting period over the one-year follow-up and 14% fewer days in institutional settings. Veterans

In phase 2 also had 19% lower addiction severity index alcohol use scores,14% lower drug use scores

And 8% lower psychiatric problem scores

Reduce homelessness,

26

Khaleghparast et al. (2013)

Iran, 2 hospitals

Discharge planning (self-care training programme/nursing process model)

46 service users

Longitudinal clinical trial

The intervention group had improved clinical symptoms and higher knowledge levels compared with control group.

Statistically significantly lower readmissions in the intervention group.

To increase patient knowledge, reduce clinical symptoms and rehospitalisation.

27

Khanbhai et al. (2018)

Australia, 1 medical centre

Discharge checklist

230 checklists

Quasi-experimental, pre–post intervention design

There was a small, but statistically non-significant, reduction in readmission rates.

Reduce readmission

28

Kidd et al. (2016)

Canada, 1 large hospital in city

‘Welcome Basket.’ (6 week peer support, contact on wards, basket of items, environmental support)

23

Evaluation- a mixed methods design, pre-post for quantitative outcomes, interviews and readmission rates

Pre–post analysis indicated no change in psychiatric symptoms but improvement in community functioning, community integration, and quality of life. No difference in readmission

Reduce suboptimal outcomes in first month, bridge gap

29

Kisely et al. (2017)

Australia, 1 hospital- intervention and control wards within it

Motivational aftercare planning (motivational interviewing with advance directives)

100 intervention plans, 197 control, 20 service user interviews

Controlled before-and-after design, interviews

Intervention ward improved significantly (e.g. identification of triggers significantly increased from 52 to 94%, This did not occur in the control wards. Interviews showed improvements in experiences of discharge planning.

To increase patient input into discharge planning, increase treatment plan following

30

Lawn et al. (2008)

Australia, 3 hospitals

Peer support

No participant numbers in evaluation

Evaluation methodology.

Intervention at this stage of their recovery seems highly effective as an adjunct to mainstream mental health services. It has personal benefit to consumers and peers, substantial savings to systems, as well as much potential for encouraging mental health service culture and practice towards a greater recovery focus and improved collaboration with GPs

To reduce hospital avoidance and facilitate early discharge

31

Lin et al. (2018)

Taiwan, 1 hospital

Needs-oriented hospital discharge planning for caregivers

114 caregivers (of people with schizophrenia) 57 in each group

A quasi-experimental research design

The caregiver burden and health status of the experimental group improved more significantly compared with control group.

Reducing readmission and improving medication adherence, reducing care giver burden

32

Puschner et al. (2011)

Germany, 5 hospitals

Needs-oriented discharge planning intervention

491

Multicentre RCT

No effect of the intervention on outcomes.

Reduce high utilisation of inpatient care

33

Reynolds et al. (2004)

Scotland, 1 unit, 3 wards

Transitional Discharge Model (ward nurse worked with SU until relationship built with community nurse, then support from service users)

25 services user (14 control, 11 experimental)

Randomised experimental design

Both control and experimental group demonstrated significant improvements in symptom severity and functional ability after 5 months. Usual treatment subjects in the control group were more than twice as likely to be re-admitted to hospital.

Readmissions and not able to adapt to community, focus on need for relationships

34

Rose et al. (2007)

USA, 1 large urban medical centre, mostly African- American patients

Transitional care model a nurse-based in-home transitional care intervention

10 service users (schizophrenia, bipolar)

Evaluation- analysis of nurse logs

Offers an alternative to patients who might otherwise be left poorly treated or untreated in the community setting.

Lack of continuity of care and meet immediate post discharge needs of SU

35

Sato et al. (2012)

Japan, 5 hospitals

Community re-entry program. Discharge preparation program (psychosocial program for preparing long-term

hospitalized patients)

26 intervention, 23 control (schizophrenia)

RCT

The program may be capable of promoting discharge of long-term hospitalized psychiatric patients.

There was no significant difference between both groups for number of patients discharged 6 months after end of program.

To reduce length of stay

36

Scanlan et al. (2017)

Australia, 3 geographical areas, large non-governmental mental health service

Peer-delivered, transitional and post-discharge support program

38 service users

Evaluation, outcome measures, interviews

Participants reported improvements in functional and clinical recovery and in the areas of intellectual, social and psychological wellness. Self-report of hospital readmissions suggested that there was a reduction in hospital bed days following the program

Reduce readmission, increase wellbeing

37

Shaffer et al. (2015)

USA, 6 community-based provider organizations within network of a not-for-profit, managed behavioural health care organization

Brief critical time intervention (a brief, three-month version of CTI)

149 adults with readmission within 30 days, 224 control

A quasi-experimental investigation

BCTI was associated with decreased early readmission rates,

Reduce readmission

38

Shaw et al. (2000)

Scotland, 3 acute wards, 1 hospital

Pharmacy discharge planning (receiving a baseline Pharmaceutical needs assessment, information about medicines and then a Pharmacy discharge plan sent to their community pharmacy)

97 service users

Controlled study

No significant difference between the groups in baseline medicine knowledge. One week post-discharge, both groups showed

Significant improvement in knowledge of medication from baseline and was maintained at 12 weeks.

Fewer medication problems for the intervention group.

To reduce medicine-related problems that cause readmission

39

Simpson et al. (2014)

UK, 4 wards, inner city (London)

Peer support

46 service users 23 peer support 23 control

Pilot randomised controlled trial with economic evaluation

No statistically significant benefits for peer support for hope or QoL, there is an indication that hope may be further increased in those in receipt of peer support. The total cost per case for the peer support was £2154 compared to £1922 for control.

To increase hope and quality of life

40

Smelson et al. (2010)

USA, 1 acute inpatient psychiatric unit

Brief Treatment Engagement (5 h per week of services- assertive community treatment using BCTI, peer support, dual recovery therapy)

102 veterans, (56 control)

Prospective randomized trial

69%

Of intervention participants attended an outpatient appointment within 14 days of discharge, compared to only 33% of control. Intervention participants were also significantly more likely to be engaged in outpatient services at the end of the intervention period.

Treatment engagement

41

Taylor et al. (2014)

USA, 1 large psychiatric hospital

Brief care management

Intervention (brief interview prior to discharge)

87 intervention, 108 control, 195 total

Controlled study

Individuals in the control group were more likely to be readmitted within 30 days of an index readmission than individuals in the intervention group.

Increase aftercare engagement, reduce readmissions

42

Tomita et al. (2012)

USA, 2 New York City hospitals

Critical time intervention (CTI)

150 total previously homeless men and women

RCT

At the end of the follow-up period, psychiatric re-hospitalization was significantly lower for the group assigned to CTI compared with the usual services group.

Reducing readmission

43

Virgolesi et al. (2017)

Italy, 3 hospitals in Rome

Nursing discharge programme (a short-term nursing discharge programme with follow-up phone calls 7–10 days)

135 patients

Prospective correlational design

The interpersonal and educational nursing intervention improves adherence to a treatment plan.

Medication adherence and patient satisfaction

44

Walker et al. (2000)

UK, 3 wards (2 control)

Discharge co-ordinators

343, 119 intervention, 224 control

Controlled cohort study

No differences in outcomes (readmission, LoS, mental health status, satisfaction). More satisfaction for those without intervention

Improve communication between primary and secondary care

45

Zheng and Arthur (2005)

China, 1 large hospital in Beijing

Family education

101 patients (schizophrenia)

RCT pre-test, post-test

Significant improvement in knowledge about Schizophrenia in the experimental group. Significant difference in symptom scores and functioning at 9 months after discharge.

Knowledge about condition and rehospitalisation.

There is a need for culturally sensitive family treatments offered by nurses