Table 2 Summary of findings
Intervention type and outcome measure | Number of RCTs and participants | Reasons for downgrade | Direction and magnitude of relative effect | Grade |
|---|---|---|---|---|
Effects of provider intervention on healthcare professional behavior | ||||
 Provider intervention vs UCP | ||||
  Odds of achieved provider adherence (main indication) | 13 RCTs [51, 52, 57, 58, 60,61,62, 64, 78,79,80,81,82,83] N = 3158 | H | Provider interventions not statistically significantly different from comparator groups (OR 1.60; CI 0.76, 3.37) | Moderate |
  Mean difference in achieved provider adherence (main indication) | 9 RCTs [52, 60, 62, 63, 83,84,85,86,87, 92, 93] N = 1236 | H, DE | Provider interventions not statistically significantly different from comparator groups (SMD 0.17; CI − 0.16, 0.50) | Low |
  Incidence rate of achieved provider adherence (main indication) | N = 63,588 | H, IMP | Provider interventions not statistically significantly different from comparator groups (IRR 1.16; CI 0.63, 2.14) | Low |
  Odds of improved medication prescribing | 11 RCTs [51, 52, 57, 58, 60,61,62, 64, 78, 81, 82] N = 4116 | H, IMP | Provider interventions statistically significantly different from comparator groups (OR 1.42; CI 1.04, 1.92) favoring the intervention | Low |
  Mean difference in improved medication prescribing | 3 RCTs [86] N = 414 | DE, IMP | Provider interventions not statistically significantly different from comparator groups (SMD 0.15; CI − 0.48, 0.79) | Low |
  Incidence rate of improved medication prescribing | N = 63,144 | H, IMP | Provider interventions not statistically significantly different from comparator group (IRR 1.02; CI 0.44, 2.36) | Low |
  Odds for increased contact with patients | N = 710 | H, IMP | Provider interventions not statistically significantly different from comparator groups (OR 6.40; CI 0.13, 322.40) | Low |
  Mean difference in contact with patients | N = 225 | IMP | Provider interventions not statistically significantly different from comparator groups (SMD 0.17; CI − 0.84, 1.19) | Moderate |
  Incidence rate of number of consultations (contact with patients) | 1 RCT [51] N = 444 | S | Provider intervention statistically significantly different from comparator group (IRR 1.78; CI 1.14, 2.78) favoring the intervention | Very low |
 Odds of general adherence to intervention | 6 RCTs [57, 61, 64, 79, 82, 83] N = 1375 | H, IMP | Provider interventions not statistically significantly different from comparator groups (OR 2.26; CI 0.50, 10.28) | Low |
  Mean difference in general adherence to intervention | N = 597 | H, DE, IMP | Provider interventions not statistically significantly different from comparator groups (SMD 0.23; CI − 1.42, 1.89) | Very low |
  Odds of referral offered to patient | N = 896 | IMP | Provider interventions not statistically significantly different from comparator groups (OR 1.11; CI 0.33, 3.70) | Moderate |
 Provider intervention vs practice redesign | ||||
  Odds of achieved provider adherence (main indication) | N = 867 | IMP | Provider interventions not statistically significantly different from comparator groups (OR 0.81; CI 0.30, 2.19) | Moderate |
  Mean difference in achieved provider adherence (main indication) | 1 RCT [52] N = 24 | S | Provider intervention not statistically significantly different from comparator group (SMD 0.07; CI − 0.73, 0.87) | Low |
  Odds of improved medication prescribing | N = 1738 | DE, IMP | Provider interventions not statistically significantly different from comparator groups (OR 0.96; CI 0.18, 5.08) | Low |
  Mean difference in contact with patients | 1 RCT [52] N = 24 | S | Provider intervention not statistically significantly different from comparator group (SMD 0.07; CI − 0.73, 0.87) | Low |
  Odds of general adherence to intervention | 1 RCT [53] N = 61 | Poor RoB, IP, S | Provider interventions not statistically significantly different from comparator groups (OR 0.30; CI 0.08, 1.14) | Very low |
 Provider intervention vs other interventions | ||||
  Odds of achieved provider adherence (main indication) | 1 RCT [59] N = 171 | S, IMP, PND | Provider intervention not statistically significantly different from comparator group (OR 0.85; CI 0.43, 1.69) | Very low |
  Odds of improved medication prescribing | 1 RCT [59] N = 171 | S, IMP, PND | Provider intervention not statistically significantly different from comparator group (OR 0.85; CI 0.43, 1.69) | Very low |
  Odds of general adherence to intervention | 1 RCT [59] N = 171 | S, IMP, PND | Provider intervention not statistically significantly different from comparator group (OR 0.45; CI 0.20, 1.01) | Very low |
Effects by intervention type | ||||
 Comparative effectiveness | ||||
  Guideline distribution plus implementation recommendations vs guideline distribution alone: odds of achieved provider adherence (main indication) | 1 RCT [57] N = 378 | S, IMP, IP | Provider interventions not statistically significantly different (OR 1.62; CI 0.64, 4.06) | Very low |
  Guideline distribution and education vs guideline distribution, education, and nurse disease management (system redesign): odds of achieved provider adherence (main indication) | 1 RCT [53] N = 61 | S, IMP, poor RoB, IP | Provider interventions not statistically significantly different (OR 0.30; CI 0.08, 1.14) | Very low |
  Academic detailing vs academic detailing plus continuous quality improvement: odds of achieved provider adherence (main indication) | 1 RCT [58] N = 389 | S, IMP | Provider interventions not statistically significantly different (OR 1.01; CI 0.48, 2.11) | Very low |
  Guideline distribution vs guideline distribution and motivational interviewing training: odds for achieved provider adherence (main indication) | 1 RCT [59] N = 171 | S, IMP, PND | Provider interventions not statistically significantly different (OR 0.85; CI 0.43, 1.69) | Very low |
  Education plus additional training sessions vs education alone: odds for achieved provider adherence (main indication) | 1 RCT [60] N = 55 | S, IMP, PND | Provider interventions not statistically significantly different (OR 1.17; CI 0.33, 4.19) | Very low |
  Education plus additional training sessions vs education alone: mean difference in achieved provider adherence (main indication) | 1 RCT [60] N = 55 | S, IMP, PND | Provider interventions not statistically significantly different (SMD 0.67; CI 0.06, 1.28) | Very low |
  Patient-specific treatment recommendations vs recommendations and care management: odds for achieved provider adherence (main indication) | 1 RCT [52] N = 417 | S, IMP | Provider interventions not statistically significantly different (OR 0.85; CI 0.58, 1.25) | Very low |
  Patient-specific treatment recommendations vs recommendations and care management: mean difference in achieved provider adherence (main indication) | 1 RCT [52] N = 417 | S, IMP | Provider interventions not statistically significantly different (SMD 0.07; CI − 0.73, 0.87). | Very low |
  Training plus tailored implementation vs training alone: odds for achieved provider adherence (main indication) | 1 RCT [51] N = 444 | S, IMP | Provider interventions not statistically significantly different (OR 1.07; CI 0.52, 2.19). | Very low |
  Training plus tailored implementation vs training alone: incidence rate for achieved provider adherence (main indication) | 1 RCT [51] N = 444 | S, IMP | Provider interventions statistically significantly different (IRR 1.78; CI 1.14, 2.78), favoring the intervention of training plus tailored implementation | Very low |
  Guideline distribution plus workshop and consultation vs guideline distribution alone: odds of achieved provider adherence (main indication) | 1 RCT [62] N = 147 | S, IMP | Provider interventions not statistically significantly different (OR 1.25; CI 0.40, 3.90) | Very low |
  Guideline distribution plus workshop and consultation vs guideline distribution alone: mean difference in achieved provider adherence (main indication) | 1 RCT [62] N = 147 | S, IMP | Provider interventions not statistically significantly different (SMD − 0.08; CI − 0.42, 0.26) | Very low |
  Education plus other components vs guidelines and education without tailoring to stages of change: mean difference in achieved provider adherence (main indication) | 1 RCT [63] N = 36 | S, IMP | Provider interventions statistically significantly different (SMD 0.89; CI 0.59, 1.18), favoring intervention with education plus other components tailored toward stages to change | Very low |
  Guideline distribution (passive) vs guideline distribution (active): odds of achieved provider adherence (main indication) | 1 RCT [61] N = 138 | S, IMP, IP | Provider interventions not statistically significantly different (OR 1.76; CI 0.64, 4.86) | Very low |
 Indirect comparison | ||||
  Meta-regression education only vs education plus for odds of achieved provider adherence (main indication) | 10 RCTs [51, 52, 58, 60, 62, 64, 78, 79, 82, 83] N = 2957 | I, IMP | No systematic effect detected (p = 0.574) | Very low |
  Meta-regression education only vs education plus for mean difference in achieved provider adherence (main indication) | 8 RCTs [52, 60, 62, 63, 84, 85,86,87, 92, 93] N = 712 | I, IMP | No systematic effect detected (p = 0.238) | Very low |
  Meta-regression unidimensional vs multidimensional for odds of achieved provider adherence (main indication) | 13 RCTs [52, 57, 58, 60,61,62,63, 78,79,80,81,82,83, 92] N = 2953 | I, IMP | No systematic effect detected (p = 0.707) | Very low |
  Meta-regression unidimensional vs multidimensional for mean difference in achieved provider adherence (main indication) | 9 RCTs [52, 60, 62, 63, 83,84,85,86,87, 92, 93] N = 1236 | I, IMP | No systematic effect detected (p = 0.055) | Very low |
  Meta-regression unidimensional vs multidimensional for odds of improved medical prescribing | 12 RCTs [51, 52, 57,58,59,60,61,62, 64, 78, 81, 82] N = 2678 | I, IMP | No systematic effect detected (p = 0.317) | Very low |
  Meta-regression unidimensional vs multidimensional for odds of referral offered to patients | N = 896 | I, IMP | No systematic effect detected (p = 0.195) | Very low |
  Meta-regression intervention intensity for odds of achieved provider adherence (main indication) | 13 RCTs [51, 52, 57, 58, 60,61,62, 64, 78,79,80,81,82,83] N = 3158 | I, IMP | No systematic effect detected (p = 0.973) | Very low |
  Meta-regression intervention intensity for mean difference in achieved provider adherence (main indication) | 9 RCTs [52, 60, 62, 63, 83,84,85,86,87, 92, 93] N = 1236 | I, IMP | The analysis suggested that the intensity of the intervention is associated with the effect size (p = 0.033) | Very low |
  Meta-regression intervention intensity for odds of improved medical prescribing | 12 RCTs [51, 52, 57,58,59,60,61,62, 64, 78, 81, 82] N = 2678 | I, IMP | No systematic effect detected (p = 0.414) | Very low |
  Meta-regression intervention intensity for odds of general adherence to intervention | 8 RCTs [53, 57, 59, 61, 64, 79, 82, 83] N = 2411 | I, IMP | No systematic effect detected (p = 0.542) | Very low |
 Subgroup analyses by intervention type | ||||
  Guideline distribution only: odds of achieved provider adherence (main indication) | N = 683 | IMP | Provider interventions not statistically significantly different from comparator groups (OR 1.28; CI 0.75, 2.19) | Low |
  Guideline distribution only: mean difference for achieved provider adherence (main indication) | 1 RCT [86] N = 281 | S, IMP, PND | Provider intervention statistically significantly different from comparator group (SMD − 0.44; CI − 0.68, − 0.20), favoring the comparator | Very low |
  Guideline distribution only: odds of improved medication prescribing | N = 854 | H, IMP | Provider interventions not statistically significantly different from comparator groups (OR 1.52; CI 0.60, 3.86) | Low |
  Guideline distribution only: odds of increased provider contact with patients | 1 RCT [61] N = 130 | S, IMP, IP | Provider intervention statistically significantly different from comparator group (OR 2.71; CI 1.24, 5.94) | Very low |
  Guideline distribution only: odds of general adherence to intervention | N = 679 | H, IMP | Provider interventions not statistically significantly different from comparator groups (OR 0.95; CI 0.17, 5.17) | Very low |
  Education only: odds of achieved provider adherence (main indication) | N = 338 | H, IMP | Provider interventions not statistically significantly different from comparator groups (OR 3.04; CI 0.01, 756.17) | Low |
  Education only: mean difference in achieved provider adherence (main indication) | 3 RCTs [86] N = 414 | IMP | Provider interventions not statistically significantly different from comparator groups (SMD 0.15; CI − 0.48, 0.79) | Moderate |
  Education only: odds of improved medication prescribing | 1 RCT [82] N = 48 | S, IMP | Provider intervention not statistically significantly different from comparator group (OR 2.78; CI 0.80, 9.59) | Very low |
  Education only: odds of increased provider contact with patients | 1 RCT [82]) N = 48 | S, IMP | Provider intervention statistically significantly different from comparator group (OR 6.42; CI 1.78, 23.18) | Very low |
  Education only: odds of general adherence to intervention | N = 399 | H, IMP | Provider interventions not statistically significantly different from comparator groups (OR 2.03; CI 0.06, 73.30) | Very low |
  Education plus other components: odds for achieved provider adherence (main indication) | 7 RCTs [51, 52, 58, 60, 62, 64, 78] N = 2090 | IMP | Provider interventions not statistically significantly different from comparator groups (OR 1.17; CI 0.62, 2.18) | Moderate |
  Education plus other components: mean difference in achieved provider adherence (main indication) | N = 938 | H, IMP | Provider interventions not statistically significantly different from comparator groups (SMD 0.37; CI − 0.16, 0.90) | Low |
  Education plus other components: odds of improved medical prescribing | 7 RCTs [51, 52, 58, 60, 62, 64, 78] N = 1710 | H | Provider interventions not statistically significantly different from comparator groups (OR 1.21; CI 0.85, 1.71) | Low |
  Education plus other components: odds of increased provider contact with patients | 1 RCT [64] N = 483 | S, IMP | Provider interventions statistically significantly different from comparator group (OR 101.34; CI 6.17, 1664.08) | Very low |
  Education plus other components: odds of general adherence to intervention | 1 RCT [64] N = 482 | S | Provider interventions statistically significantly different from comparator group (OR 2.56; CI 1.65, 3.97) | Very low |
Effects by provider type | ||||
  Meta-regression single provider vs team for odds of achieved provider adherence (main indication) | 13 RCTs [51, 52, 57, 58, 60,61,62, 64, 78,79,80,81,82,83] N = 3158 | I, IMP | The analysis suggested that the type of provider is associated with the effect size (p = 0.034); however, the analysis is based on only 1 team intervention | Very low |
 Subgroup analysis by provider type | ||||
  Single provider interventions: odds for achieved provider adherence (main indication) | 12 RCTs [51, 52, 57, 58, 60,61,62, 78,79,80,81,82,83] N = 1334 | H, IMP | Provider interventions not statistically significantly different from comparator groups (OR 1.42; CI 0.74, 2.73) | Low |
  Team provider interventions: odds of achieved provider adherence (main indication) | 1 RCT [64] N = 482 | S, IMP | Provider intervention statistically significantly different from comparator group (OR 101.34, CI 6.17, 1664.08), favoring the intervention | Very low |
Effect by setting | ||||
  Meta-regression primary care vs specialty care setting for mean difference in achieved adherence (main indication) | 9 RCTs [52, 60, 62, 63, 83,84,85,86,87, 92, 93] N = 1236 | I, IMP | No systematic effect detected (p = 0.385); however, the analysis is based on only 2 specialty care interventions | Very low |
Patient outcomes | ||||
 Provider intervention vs UCP | ||||
  Mean difference in depression rating scale scores | 9 RCTs [51, 52, 58, 61, 62, 79, 81, 84, 92] N = 2196 | DE | Provider interventions not statistically significantly different from comparator groups (SMD − 0.06; CI − 0.14, 0.01) | Moderate |
  Odds of depression treatment response | 6 RCTs [52, 57, 60, 61, 64, 81] N = 1312 | DE | Provider interventions statistically significantly different from comparator groups (OR 1.12; CI 1.04, 1.21) favoring the intervention | Moderate |
  Odds of depression recovery | 6 RCTs [52, 57, 60, 61, 79, 81] N = 1274 | DE | Provider interventions not statistically significantly different from comparator groups (OR 1.02; CI 0.91, 1.15) | Moderate |
  Odds of depression treatment adherence | N = 281 | IMP | Provider interventions not statistically significantly different from comparator groups (OR 1.52; CI 0.70, 3.31) | Moderate |
 Provider intervention vs system redesign | ||||
  Mean difference in depression rating scale scores | N = 861 | IMP | Provider interventions not statistically significantly different from comparator groups (SMD 0.09; CI − 0.48, 0.67) | Moderate |
  Odds of depression treatment response | N = 478 | IMP | Provider interventions not statistically significantly different from comparator groups (OR 0.53; CI 0.01, 40.38) | Moderate |
  Odds of depression recovery | N = 478 | IMP | Provider interventions not statistically significantly different from comparator groups (OR 0.41; CI 0.01, 17.89) | Moderate |
  Odds of depression treatment adherence | 1 RCT [53] N = 61 | S | Provider interventions not statistically significantly different from comparator groups (OR 0.16; CI 0.02, 1.39) | Very low |
 Provider intervention vs other interventions | ||||
  Odds of depression treatment adherence | 1 RCT [59] N = 171 | S, IMP | Provider intervention not statistically significantly different from motivational interviewing (OR 0.79; CI 0.30, 2.08) | Very low |
  Mean difference in treatment adherence | 1 RCT [59] N = 171 | S | Provider intervention not statistically significantly different from motivational interviewing (SMD − 0.43; CI − 0.76, − 0.11) | Very low |