Begin main content

Systematic Review

Interventions to improve antibiotic prescribing practices for hospital inpatients.

Davey P, Brown E, Fenelon L, Finch R, Gould I, Hartman G, Holmes A, Ramsay C, Taylor E, Wilcox M, Wiffen P. Interventions to improve antibiotic prescribing practices for hospital inpatients. Cochrane Database of Systematic Reviews 2005 4:CD003543. [PubMed]

Sources of funding: British Society for Antimicrobial Chemotherapy UK, Hospital Infection Society UK. No perceived conflict of interest.

Main author affiliation: Davey P, Ninewells Hospital and Medical School, Dundee, Scotland.

Q: Which professional interventions are effective in changing antibiotic prescribing behaviour in hospital settings?

BACKGROUND

The prevalence of over prescription of antibiotics in hospital settings necessitates an examination of all potential interventions. The interventions could be organisational, financial, regulatory or professional. It is also common to use multifaceted interventions to assist with changing this specific prescribing behaviour. Different interventions are likely preferred in hospitals compared to community settings and the effectiveness of the interventions also probably differs.

INCLUSION

Individual studies from systematic review

SEARCH FOR EVIDENCE: 1966 to Nov 2003

INCLUDED STUDIES: 69

STUDIES RELATED TO PRESCRIBING: 66

STUDY DESIGN: RCT, CBA, ITS.

STUDY CHARACTERISTICS:

  • Target Population: Physician, Nurses, Pharmacists, Other.
  • Intervention(s): Organisational (provider) — other, Educational outreach visits, Formulary, Professional — other, Revision of professional roles — pharmacy, Reminders — general, Distribution of educational materials, Audit and feedback, Multifaceted.
  • Setting: Hospitals, Ownership — Not specified/unclear, Academic — Not specified/unclear.
  • Patients Reviewed: Hospitalized.
  • Drug Related Outcome Categories: Appropriate use, Appropriate use — route, Appropriate use — choice, Cost containment.
  • Other Outcome Categories: Other resource use, Other.

AUTHORS' ANALYSIS OF DATA:

Vote counting based on direction of effect.

REVIEW QUALITY: 7

RESULTS - OVERALL

Formulary changes (n=6), multifaceted interventions (n=26), audit and feedback (n=6) and other organisational changes (n=11) were found to be generally effective for improving appropriate care behaviours. A limited number of the studies were of high quality.

RESULTS - RELATED TO PRESCRIBING

Organisational (other) interventions (n=7), formulary changes (n=4), reminders (n=3) and audit and feedback (n=2) were found to be generally effective for improving other appropriate prescribing behaviour. Multifaceted interventions were effective for improving choice (n=3), dosage (n=3) and cost containment (n=8) yet had mixed effects for other appropriate use outcomes (n=12). Audit and feedback (n=3) and other organisational intervetions (n=2) were found to be generally effective for improving cost containment outcomes with a limited number of high quality studies. Pharmacy led care had mixed effects for improving cost containment (n=2). Educational outreach had no effect on improving other appropriate use outcomes (n=2). There was insufficient evidence to determine the effectiveness of formulary changes, reminders, pharmacy led care, distribution of educational materials and other professional interventions on other prescribing related outcomes.

CONCLUSIONS - OVERALL

Medium quality review. Formulary changes, audidt and feedback, other organisational and multifaceted interventions were found to be effective for improving appropriate care behaviours.

CONCLUSIONS - RELATED TO PRESCRIBING

Medium quality review. Multifaceted interventions were generally effective for improving cost containment, however, mixed effects were observed for improving other prescribing behaviours. Organisational (other) interventions were found to be effective for improving prescribing related behaviour. A broad definition of primary care was adopted.


Table of Results

Comparison Outcome N Analysis Results
Multifaceted vs control Appropriate care 26 Vote counting based on direction of effect Vote counting 19/26 studies (mixed designs) favoured intervention: generally effective. 5 studies with mixed results.
1/2 RCTs favoured intervention: mixed effects. 1 study with mixed results.
15/20 ITS studies favoured intervention; 3 studies observed a deterioration in slope; 2 studies with mixed results: generally effective.

Prescribing related outcomes:
Other appropriate use:
Vote counting 8/12 studies (mixed designs) favoured intervention: mixed effects. 2 studies with mixed results.
2/2 RCTs favoured intervention: generally effective.
5/9 ITS studies favoured intervention; 2 studies observed a deterioration in slope; 2 studies with mixed results: generally ineffective.

Choice:
Vote counting 3/3 ITS studies favoured intervention with no observed a deterioration in slope: generally effective.

Dosage:
Vote counting 2/3 studies favoured intervention: generally effective. 1 study with mixed results.
1/1 ITS favoured intervention and no deterioration in slope: insufficient evidence.

Cost containment:
Vote counting 8/8 ITS studies favoured intervention; 1 study observed a deterioration in slope, generally effective.
Audit and feedback vs control Appropriate care 6 Vote counting based on direction of effect Vote counting 5/6 studies (mixed designs) favoured intervention generally effective.
1/1 RCT favoured intervention: insufficient evidence.
4/4 ITS studies favoured intervention, and 0/4 observed a deterioration in slope.

Prescriging related outcomes:
Cost containment:
Vote counting 2/3 studies (mixed designs) favoured intervention: generally effective.
1/1 RCT favoured intervention: insufficient evidence.
1/1 ITS favoured intervention with no deterioration in slope: insufficient evidence.

Other appropriate use:
Vote counting 2/2 studies (mixed designs) favoured intervention: generally effective.
1/1 RCT favoured intervention: insufficient evidence.
1/1 ITS favoured intervention with no deterioration in slope: insufficient evidence.

Dosage:
Vote counting 1/1 ITS favoured intervention with no deterioration in slope: insufficient evidence.
Formulary vs control Appropriate care 6 Vote counting based on direction of effect Vote counting 5/6 studies (mixed designs) favoured intervention: generally effective. 1 study with mixed results.
4/5 ITS studies favoured intervention; no studies observed a deterioration in slope; 1 study with mixed results: generally effective.

Prescribing related outcomes:
Other appropriate use:
Vote counting 3/4 studies (mixed designs) favoured the intervention: generally effective.
3/3 ITS studies favoured intervention; no studies observed a deterioration in slope: generally effective.

Choice:
Vote counting 1/1 ITS study favoured intervention with no observed a deterioration in slope: insufficient evidence.

Cost containment:
Vote counting 1/1 ITS study with mixed results: insufficient evidence.
Organisational — other vs control Appropriate care 11 Vote counting based on direction of effect Vote counting 9/11 studies (mixed designs) favoured intervention: generally effective. 1 study with mixed results.
3/3 RCTs favoured intervention: generally effective.
4/5 ITS studies favoured intervention, and one study observed a deterioration in slope: generally effective.

Prescribing related outcomes:
Cost containment:
Vote counting 2/2 studies favoured the intervention: generally effective.
1/1 ITS favoured the intervention with a deterioration in slope: insufficient evidence.

Other appropriate use:
Vote counting 9/10 studies (mixed designs) favoured the intervention: generally effective.
5/6 ITSs studies favoured intervention and no studies observed a deterioration in slope: generally effective.
2/2 RCTs favoured the intervention: generally effective.
Educational outreach visits vs control Appropriate use 2 Vote counting based on direction of effect Other appropriate use:
Vote counting 0/2 studies (mixed designs) favoured intervention: no effect.
0/1 RCT favoured intervention: insufficient evidence.
Distribution of educational materials vs control Appropriate use 1 Vote counting based on direction of effect Vote counting 0/1 ITS favoured intervention: insufficient evidence.
Professional — other vs control Appropriate use 1 Vote counting based on direction of effect Vote counting 1/1 ITS favoured intervention with evidence of slope deterioration over time: insufficient evidence.
Revision of professional roles — pharmacy vs control Appropriate use 2 Vote counting based on direction of effect Vote counting 1/2 studies (mixed designs) favoured intervention: mixed effects. 1 study with mixed results.
1/1 RCT favoured intervention: insufficient evidence.

Other appropriate use:
Vote counting 1/1 RCT favoured the intervention: insufficient evidence.

Cost containment:
Vote counting 1/2 studies favoured intervention: mixed effects.
0/1 RCT favoured the intervention: insufficient evidence.
Reminders vs control Appropriate use 5 Vote counting based on direction of effect Vote counting, 3/5 RCTs favoured intervention, mixed results.

Prescribing related outcomes:
Other appropriate use:
Vote counting, 3/3 RCTs favoured intervention, generally effective.

Route of administration:
Vote counting, 1/1 RCT favoured intervention, insufficient evidence.

Cost containment:
Vote counting 1/1 study favoured intervention, insufficient evidence.