Table 2

Features appeared to be associated with successful implementation

StrategyActive features/characteristicsInactive features/characteristics
Printed educational materials (PEM)
  • Tailoring

Purpose (eg, increase or decrease in, modification of behaviour)
  • Type of targeted behaviour

  • Clinical area

  • Format

*Based on very limited evidence and box plots presented only
▸ Mode,
  • Frequency,

  • Duration of delivery are not associated with improvement in outcomes

*Due to the lack of variability, not able to assess the importance of these characteristics to determine PEM effectiveness
Educational strategies
  • Mixed interactive and didactic formats

  • High attendance at educational meetings

  • Low complexity of the targeted behaviour

  • Tailoring

  • Relevance or identify needs with a facilitator

  • Interaction/active participation

  • Facilitate and (small) team based

  • Training support

  • Management support

  • Clear goals

  • Led by senior colleagues/superior

  • Intensity and frequency

  • Programmes directed at trainee physicians

  • Focus on serious outcomes

  • Didactic sessions/lectures alone

  • Seminar-based sessions

  • High complexity of the targeted behaviour

  • Minimal interaction/discussion

  • Passive strategies (eg, mailed educational materials)

  • Programmes directed at established physicians

Educational outreach visits
  • Most effective when the educators are known to and respected by the target group

No data reported
Audit and feedback (A&F)
  • Source—(p<0.001) supervisor/senior colleague

  • Format—(p=0.02) feedback provided both verbally and written

  • Measurable targets and action plan (p<0.001)

  • Timing—concurrent feedback, presented close to the time of decision-making

  • Active

  • Tailoring

  • Part of an overall strategy

  • Low/ non-existent baseline

  • Effect size was not influenced by the number of implementation strategies in addition to A&F.

    A&F alone vs A&F in a multifaceted intervention: not significant; Dichotomous: estimated absolute difference in adjusted RD=3.3%, p=0.27)

Practice facilitation
  • Tailoring to the context and needs of the practice (SMD=0.62, 95% CI 0.48 to 0.75; p=0.05)

  • Higher intensity of the intervention (average number of contacts by the average meeting time in hours; p=0.03)

  • Smaller number of practices per facilitator (p=0.004)

  • No tailoring (SMD=0.37, 95% CI 0.16 to 0.58)

  • Lower intensity of the intervention

  • Larger number of practices per facilitator

Financial strategies
  • Larger size of payment

  • Clear goal

  • Low complexity of task

  • Concurrent or intermittent payment

  • Sustainability of new behaviour—incentives may only buy temporary priority

  • Positive effect was greater for initially low performers (low baseline performance, more room for improvement) compared with already high performers

  • Involvement of stakeholders in target selection and incentive programme development

  • Context (national level gave more uniform results than fragmented programmes)

  • Design choices (process indicators gave higher improvement than outcome measures)

  • High awareness of the existence of an incentive programme

  • Incentives based on financial rewards only showed more positive effects

  • Size of payment—small rewards may not motivate doctors to change their behaviour or practices

  • High complexity of task

  • End of year payment (infrequent performance feedback)

  • Continuing adding additional funding or payment in the long term is not effective.

  • Low awareness of the existence of an incentive programme

  • Incentives based on a competitive approach (reward for high performers, as well as penalty for low performers)

Local opinion leaders
  • Multidisciplinary opinion leader teams

  • Single opinion leaders

  • A&F, audit and feedback; SMD, standardised mean difference.