Model

Study and quality score

Data

Results

Training

Gilbody et al^{w1 w2}

Quality score: 3

17 studies of effectiveness of guideline implementation and other educational strategies in depression (including 10 randomised trials, 5 controlled before and after studies, and 2 interrupted time series)

Training was more effective than usual care in 3/10 randomised trials, 1/ 5 controlled before and after studies, and 2/2 interrupted time series

Huibers et al^w3

Quality score: 3

2 randomised trials of effectiveness of specific psychosocial interventions for depression delivered by primary care clinicians; 2 randomised trials of effectiveness of psychosocial interventions for somatisation

Training primary care clinicians in problem solving in depression was more effective than placebo in 1/1 randomised trial and as effective as drug treatment in 2/2 randomised trials. Training primary care clinicians in somatisation was more effective than usual care in improving patient outcomes and reducing resource use in 2/2 randomised trials

Consultation-liaison

Bower and Sibbald^{w4 w5}

Quality score: 2

3 randomised trials of consultation-liaison similar to the model defined in the present paper. The review examined the effect of consultation-liaison on the behaviour of the primary care clinician before and after studies) also examined the effects of consultation-liaison on the wider practice population

Consultation-liaison was more effective than usual care in reducing primary care consultations in 0/3 randomised trials, improved the adequacy of prescribing in 2/2 randomised trials, and affected referral behaviour in 0/2 randomised trials. Consultation-liaison effected the behaviour of the primary care clinician towards the wider practice population in 1/3 randomised trials and 1/2 controlled before and after studies

Katon and Gonzales^w6

Quality score: 0

2 randomised trials of "second generation" consultation-liaison similar to the model defined in the present paper

Consultation-liaison was more effective than usual care in reducing costs in 1/2 randomised trials, and more effective than usual care in improving patient outcome in 0/2 randomised trials

Model

Study and quality score

Data

Results

Collaborative care

Gilbody et al^{w1 w2}

Quality score: 3

14 randomised trials of collaborative care in depression

Collaborative care was more effective than usual care in improving patient outcome in 11/14 randomised trials

Von Korff and Goldberg^w7 (editorial only)

Quality score: 0

12 randomised trials of collaborative care in major depression

Collaborative care was more effective than usual care in improving patient outcome in 7/12 randomised trials, with effectiveness related to the amount of involvement of the mental health specialist

Badamgarav et al^w8

Quality score: 3

19 randomised trials of collaborative care in depression

Meta-analysis found a standardised effect size of 0.33 on depressive symptoms, 0.51 on patient satisfaction, and 0.36 on compliance with recommended treatment. Collaborative care was associated with an increase in healthcare utilisation and costs

Bijl et al^w9

Quality score: 2

6 randomised trials of collaborative care in depression

Collaborative care was more effective than usual care in improving patient outcome in 4/6 randomised trials

Vergouwen et al^w10

Quality score: 2

11 randomised trials of collaborative care in depression

Collaborative care was more effective than usual care in improving adherence to drug treatment in 9/11 randomised trials and improving patient outcome in 10/11 randomised trials

Model

Study and quality score

Data

Results

Replacement

Bower et al^{w11 w12}

Quality score: 3

6 randomised trials of counselling in common mental health problems

Meta-analysis found a standardised effect size of 0.28 in the short term and 0.07 over the longer term

Churchill et al^w13

Quality score: 1

5 studies of counselling and 5 of cognitive-behaviour therapy in depression

Psychological therapy was more effective than usual care in improving patient outcome in 1/5 randomised trials of counselling, and 4/5 randomised trials of cognitive-behaviour therapy

Brown and Schulberg^w14

Quality score: 1

18 studies of psychosocial treatments in primary care: 10 in patients with distress and 8 in patient with specific psychiatric diagnoses

Psychological therapy was more effective than usual care in improving patient outcome in 5/10 randomised trials in patients with distress and 7/8 randomised trials in patients with specific psychiatric diagnoses

Katon and Gonzales^w6

Quality score: 0

2 studies of psychological therapies

Psychological therapy was more effective than usual care in improving patient outcome in 2/2 randomised trials

Bower and Sibbald^{w4 w15}

Quality score: 2

13 studies of psychological therapies

Psychological therapy was more effective than usual care in reducing primary care consultations in 3/13 randomised trials, affected prescribing in 5/12 randomised trials, and affected referral behaviour in 3/6 randomised trials

Balestrieri et al^w16

Quality score: 2

11 studies of specialist mental health treatment, 9 of which involved counselling or behavioural therapy (8 randomised, 1 controlled before and after)

Meta-analysis found a median standardised effect size across all 11 studies of 0.22. Median effect size for counselling was 0.234, and that for behavioural therapy was 0.224

Schulberg et al^w17

Quality score: 1

12 randomised studies of psychological therapy: 8 with major depression and 5 with minor depression or dysthymia (1 study included both)

In studies of major depression, psychological therapy was more effective than usual care in 4/6 randomised trials and as effective as drug treatment in 4/4 randomised trials. In studies of minor depression, psychological therapy was more effective than usual care in 2/5 randomised trials

Friedli and King^w18

Quality score: 0

19 randomised studies of psychological therapies

Psychological therapy was more effective than usual care in improving patient outcome in 8/19 randomised trials