Table 3 Studies relating to outreach interventions (ordered by study design and risk of bias)
Study ID | Design | Setting | Length of follow-up | Intervention groups | Audience | Goal(s) of dissemination | Risk of bias assessment |
|---|---|---|---|---|---|---|---|
Skoglund 2013 [33] | Cluster randomised controlled trial (evidence-based drug information vs evidence-based drug information + motivational interviewing) | Primary care health centres in Sweden | 6 months follow-up post-intervention | Evidence-based drug information + motivational interviewing (408 GPs) vs evidence-based drug information alone (control) (583 GPs) | Practitioners: primary care GPs | Reach, motivation | Low risk of bias |
Ludden 2018 [37] | Cluster randomised controlled trial (facilitator-led dissemination vs traditional dissemination vs no intervention) | Primary care practices in North Carolina, USA | 3Â months follow-up | Facilitator-led approach to dissemination (10 clusters) vs traditional dissemination (10 clusters) vs control (no formal dissemination) (10 clusters) | Primary care providers | Reach, ability to implement change | Low risk of bias |
Acolet 2011 [35] | Cluster randomised controlled trial (control arm vs active dissemination) | 180 neonatal units in the UK | 6 months follow-up | Active dissemination: dissemination—as below + regional champions, workshops, benchmarking and email and telephone support (87 units) vs control arm: report, recommendations and slide set sent to units (93 units) | Staff at neonatal units | Reach, ability to implement change | Some concerns due to missing outcome data and discrepancies in the number of units reported as randomised to intervention/control in protocol vs results paper |
Bernal-Delgado 2002 [36] | Pragmatic, simple blind trial, with random assignation to one of three groups | Primary care practices in Spain | 6Â months post-intervention | Experimental group (participating in an evidence-based educational outreach visit) (48 GPs) vs placebo group (participating in a conventional educational session), and control group (with no intervention) (56 GPs) | GPs in Spain | Reach | Some concerns due to a lack of information about missing data and deviations from intended interventions |
Stafford 2010 [38] | Observational study using survey data and pharmacy data | Primary care practices in the USA | 9-months pre-intervention to 9Â months post-intervention | Comparing intensity (number of physicians contacted by investigator-educators per 100,000 population members 50Â years or older) of dissemination efforts (academic detailing) with outcomes from geographic areas | Community physicians in the USA | Reach, motivation, ability to implement change | Some concerns around potential for confoundinga |
Bartholomew, 2009 [39] | Process evaluation, including pre and post-intervention questionnaires | Primary care practices in the USA | 12Â months | Academic detailing | Internal medicine and family practitioners | Reach, motivation, ability to implement change | Some concerns around selection bias, lack of information on response rate and statistical methodsa |
Majumdar 2015 [34] | Observational study looking at the prescription of ramipril in two countries, one with no promotion (USA) and one with promotion (Canada) | Pharmacies in Canada and the USA | 4Â years of follow-up | Promotional activity undertaken by pharmaceutical companies (Canada) (4800 pharmacies) vs no promotional activity undertaken by pharmaceutical companies (USA) (51,355 pharmacies) | Cardiologists | Reach, ? | Serious risk of bias due to the potential for confoundingb |