Table 1 Comparison studies of different CRT models - study characteristics and outcomes
Study Reference | MMAT score | Study characteristics | CRT models compared | Results |
|---|---|---|---|---|
Allen (2009) [31] | n/a | Natural experiment with pre- and post-comparison; Buckinghamshire, UK | CRT team pre and post several organisational changes: “patient typing” system categorising service user needs introduced; greater use of leave from hospital to promote early discharge; referrals to CRT from other mental health services accepted without reassessment; structured screening tools introduced for acute assessments; closer links between CRT and day hospital introduced with single key worker system | Reduction in inpatient bed provision and greater service user satisfaction reported following changes. No statistical tests or numerical results reported |
Doyle (1994) [32] | 3 | Natural experiment with parallel groups: 1 Team in Folkestone, UK 1 Team in Barnet, UK | Folkestone CRT with 9 am-5 pm opening hours | No clear difference between CRT models reported and no statistical tests reported. Over the follow-up period |
Barnet CRT with 24 hour opening | ||||
9% of the 9-5 (Folkestone) CRT service users were admitted to hospital vs 5% Barnet | ||||
Happell (2009) [33] | 3 | Natural experiment with pre- and post-comparison of parallel groups; Melbourne, Australia | Control Group: Day after initial assessment, full assessment given by trainee psychiatrist, care management plan formulated. | Mean HoNOS scores not significantly different between the groups at baseline – no p-values reported |
Treatment Group: Day after initial assessment, full assessment given by nurse practitioner, who took role of trainee psychiatrist (After 7 days service users in nurse-initiated care group reverted to treatment as usual) | HoNoS scores for both groups significantly improved with treatment (difference for control group: t = 7.90, df = 51, p < .001; difference for treatment group: t = 6.90, df = 50, p < .001) No information given as to whether there was a significant difference between group HoNOS scores after treatment. | |||
Service user and carer satisfaction scores were reported as not significantly different between groups – no p-values given | ||||
Harrison (2011) [34] | 3 | Natural experiment with pre- and post-comparison of a single CRT; Manchester, UK | • In 2005 (‘pre’), referrals were only taken from secondary services. | • 301 people treated in a six-month period in 2008/09, 128 in a comparable period in 2005 |
• In 2008-2009 (‘post’), referral routes extended to primary care. | ||||
• Mean duration of contact in 2008/09 – 24 days, in 2005 – 69 days | ||||
• 39% in 08/09 already known to services and in receipt of Care Plan Assessment (CPA), 70% in 05 (P < 0.005) | ||||
• Increase in proportion treated for less severe illnesses (less severe depression and other diagnoses) in 2008/09 compared with 2005 (increase from 25 to 50%, P < 0.0001) | ||||
• Fewer treated with severe mental illness (schizophrenia and related disorders, bipolar disorder and psychotic depression); 50% in 08/09, 75% in 2005, P < 0.0001 | ||||
Reding (1995) [35] | 4 | Retrospective natural experiment with pre- and post-comparison; Kalamazoo County, Michigan, USA | • Comparison of before and after the introduction of a psychiatrist to the team | There were significantly fewer state hospital admissions in the team with a psychiatrist (p < 0.001). (The decrease in state hospital admissions was not offset by a corresponding increase in admissions to the local private psychiatric hospital.) |