Background
Methods
Literature Search
Inclusion criteria
Data extraction
Study Quality
Criterion | Score and rating criteria |
---|---|
(1) Objectives and specification main outcomes a priori | 0 = objectives unclear 1 = objectives clear but main outcomes not specified a priori 2 = objectives clear with a priori specification of main method for assessment of outcome |
(2) Adequate sample size (n per group) | 0 = inadequate (< 50/group) 1 = moderate (50-100/group) 2 = large (> 100/group or justified by power calculations) |
(3) Appropriate duration of trial including follow up | 0 = too short (< 3 months) 1 = reasonable length (3-6 months) 2 = long enough for assessment of long term outcomes (6-12 months) |
(4) Power calculation | 0 = not reported 1 = mentioned without details 2 = details of calculations provided |
(5) Method of allocation | 0 = unrandomized and likely to be biased 1 = partially or quasi randomized with some bias possible 2 = randomized allocation |
(6) Concealment of allocation | 0 = not done or not reported 2 = concealment of allocation code detailed |
(7) Clear description of treatments (including doses of drugs used) and adjunctive treatments | 0 = main treatments not clearly described 1 = inadequate details of main or adjunctive treatments 2 = full details of main and adjunctive treatments |
(8) Blinding of subjects | 0 = not done 1 = done but no test of blind 2 = done and integrity of blind tested |
(9) Source of subjects described and representative sample recruitment | 0 = source of subjects not described 1 = source of subjects given but no information on sampling or use of unrepresentative sample (for example, volunteers) 2 = source of subjects described plus representative sample taken (for example, all consecutive admissions or referrals, or random sample taken) |
(10) Use of diagnostic criteria (or clear specification of inclusion criteria) | 0 = none 1 = diagnostic criteria or clear inclusion criteria 2 = diagnostic criteria plus specification of severity |
(11) Record of exclusion criteria and number of exclusions and refusals reported | 0 = criteria and number not reported 1 = criteria or number of exclusions and refusals not reported 2 = criteria and number of exclusions and refusals reported |
(12) Description of sample demographics | 0 = little/no information (only age/sex) 1 = basic details (for example, marital status/ethnicity) 2 = full description (for example, socioeconomic status, clinical history) |
(13) Blinding of assessor | 0 = not done 1 = done but no test of blind 2 = done and integrity of blind tested |
(14) Record of number and reasons for withdrawal by group | 0 = no info on withdrawals by group 1 = withdrawals by group reported without reason 2 = withdrawals and reason by group |
(15) Outcome measures described clearly (and therefore replicable) or use of validated (or referenced) instruments | 0 = main outcomes not described clearly 1 = some of main outcomes not clearly described 2 = main outcomes clearly described or valid and reliable instruments used |
(16) Information on comparability and adjustment for differences in analysis | 0 = no information on comparability 1 = some information on comparability with appropriate adjustment 2 = sufficient information on comparability with appropriate adjustment |
(17) Inclusion of all subjects in analyses (Intention to treat analysis) | 0 = no 2 = yes |
(18) Presentation of results with inclusion of data forre-analysis of main outcomes (for example, SDs) | 0 = little information presented 1 = adequate information 2 = comprehensive |
(19) Appropriate statistical analysis (including correction for multiple tests where applicable) | 0 = inadequate 1 = adequate 2 = comprehensive and appropriate |
(20) Conclusions justified | 0 = no 1 = partially 2 = yes |
(21) Declaration of interests (for example, 0 = no source of funding) | 0 = no 2 = yes |
Effect Size and meta-analysis
Results
Search results
Study | Country | Recruitment, Inclusion Criteria | Randomization | Na
| Intervention, Role of GP Training | Control Group | Comparison | Quality |
---|---|---|---|---|---|---|---|---|
Baker (2001) [36] | GB | Consecutive patients; ≥ 18 yrs Patients seeking consultation for new-onset depression | Practices | 402 | Tailored intervention to promote guideline implementation (additional feedback, educational visits, group discussions) Additional to guideline | UCb
| Patients of experimental group vs. control group | 30 |
Bosmans (2006) [33] | NL | Consecutive patients; ≥ 55 yrs PRIME-MD = MD | Practices | 145 | 4 hrs training session on screening, diagnosis and treatment as in Dutch guidelines | UC | Patients of experimental group vs. control group | 39 |
Gask (2004) [38] | GB | GP referral; 16-65 yrs Intention or current treatment of depression (symptoms < 6 mo) HAM-D ≥ 13 | Practice | 189 | Acquisition of clinical skills, 5 lectures à 2 hrs on assessment and treatment; Sole intervention | WL | Patients of experimental group vs. control group | 36 |
Kendrick (2001) [34] Thompson (2000) [35] | GB | Consecutive patients; ≥ 16 yrs HADS-D ≥ 8 | Practice | 733 | Guideline implementation & GP training (4 h seminars, educators available for 9 more mo); Additional to guideline | WL | Patients of experimental group vs. control group Sensitivity of recognition rates of experimental group vs. control group | 36 |
King (2002) [41] | GB | Consecutive patients; ≥ 18 yrs HADS-D/A ≥ 11 | Practice | 272 | Training of GPs in brief cognitive behaviour therapy (4 half day workshops); Sole intervention | WL | Patients of experimental group vs. control group | 34 |
Llewellyn-Jones (1999) [40] | AUS | Residential facility; ≥ 65 yrs GDS ≥ 10 MMSE ≥ 18 | Patient | 220 | Shared Care Intervention, including GP training & education, health education and promotion, psychoeducation; Central part of complex intervention | WLc
| Experimental group vs. control group | 33 |
Rost (2001) [44] Pyne (2003) [43] Rost (2005) [45] | USA | GP referral; DSM III-R MD (latter two studies exclude patients currently in treatment) | Practice | 479 | QuEST intervention, 4 academic telephone calls to implement guidelines, nurse w/8-hour face-to-face training; Guidelines implementation | UC | Patients of experimental group vs. control groupd
| 38 |
Worrall (1999) [39] | CAN | GP referral; GP diagnosis and severity rating, later CES-D ≥ 16 | Practice | 147 | 3-hour sessions on guideline implementation + possible consultation of psychiatrist; Guidelines implementation | UCb
| Patients of experimental group vs. control group | 26 |
Study characteristics
Interventions
Effectiveness of provider training
Study | Follow Up | Attrition Rate %a
| Outcome | Results | Limitations | Effect Sizeb
|
---|---|---|---|---|---|---|
Baker (2001) [36] | 16 weeks | 6 | Proportion of patients w/BDI < 11 | Sign. diff in proportion of patients w/BDI > 11 (OR = 2.5) | Tailored intervention that makes GP comparison impossible since they all received diff kinds of intervention | / |
Bosmans (2006) [33] | 12 mo | 21 | PRIME-MD | No sign. diff in MD recovery | Possible Hawthorne effect, less severe episodes of MD in primary care, no blinding of patients | -0.07 |
Gask (2004) [38] | 3, 12 mo | 37 | HAM-D | No sign. diff in scores at both follow up points | Use of a new-onset (depressed less than 6 mo) sample Attrition rate rather high | -0.24 |
Kendrick (2001) [34] | 12 mo | 19 | Hospital Admittance | No sign. difference | Implemented guidelines had been tested in highly selected samples Dimensional diagnosis Potential conservative bias (chronic depressed patients) | / |
Thompson (2000) [35] | 6 weeks/6 mo | 50 | HAD | No diff in improvement, no diff in caseness rating at both points Only patients recognized as cases at baseline improve sign. during first 6 weeks (p = 0.044), no diff at 6 mo | See Kendrick (2001) | / |
Diagnosis sensitivity | No diff in sensitivity nor specificity | See Kendrick (2001) | / | |||
King (2002) [41] | 6 mo | 10 | BDI | No sign. diff in BDI scores (p = 0.84) | Smaller sample than anticipated Cut off score for inclusion rather high (makes intervention effect of CBT by lay GPs less likely) | 0.08 |
Llewellyn-Jones (1999) [40] | 9.5 mo | 23 | GDS | Sign. change in GDS scores | Serial mono-centered design Control group assessment before implementation of intervention | -0.17 |
Rost (2001) [44] | 6 mo | 10 | mCES-D | Sign. reduction in score in patients beginning new treatment episode | GP training effect unclear, feedback of diagnosis could be responsible for treatment effect Homogenous sample | -0.29 |
Pyne (2003) [43] | 12 mo | 65 | Depression severity | Sign. decrease (7.7 units) in experimental group | See Rost (2001) | / |
Rost (2005) [45] | 24 mo | 70 | Depression Free Days | Sign. more depression free days in experimental group (647.6 vs. 558.2) | See Rost (2001) | / |
Worrall (1999) [39] | 6 mo | ? | Gain score CES-D | Sign. improvement in experimental group | Possible Hawthorne effect | -0.22 |