Table 1.
Search terms by database.
Fig 1.
Note: PRISMA flow chart based on: Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(7): e1000097.
Table 2.
Characteristics of included studies: Country, setting, clinical condition and objective.
Table 3.
Characteristics of included studies: Study design, intervention, time periods, sample size and outcome measures.
Fig 2.
Decrease in the number of prescriptions for antibiotics in percentage, by studies assessing this outcome parameter.
Note: Chopra et al. (2014) and Zahabiyoun et al. (2015) do not report that outcome measure. Instead of reporting figures, odds ratios are reported in Seager et al. (2006). (a) Reduction of 1.0 antibiotic items per 100 NHS treatment claims in the intervention group compared to 0.4 antibiotic items in the control group. (b) Reduction from 2,951 prescriptions for antibiotics before the audit to 1,665 prescriptions after the audit. (c) Reduction from 2,316 prescriptions for antibiotics before the audit to 1,330 prescriptions after the audit. (d) The number of prescriptions was not stated in the paper, but the authors report an overall reduction of ~50%. (e) Reduction of the total number of prescriptions for antimicrobial agents from 253 prescriptions among 300 patients at baseline to 82 prescriptions among 300 patients at one month after the intervention. Three months after the intervention these prescriptions were at 63 among 300 patients and at 102 prescriptions among 300 patients six months after the intervention. (f) Reduction of preoperative prescriptions for antibiotics from 15 prescriptions among 80 patients before the audit to one prescription among 52 patients after the audit. A postoperative reduction of prescriptions for antibiotics was not intended.
Table 4.
Characteristics of the studies and their guidelines.
Fig 3.
Accuracy of the prescription before (pre) and after (post) intervention, by studies assessing this outcome parameter (measured as a percentage of adherence to guidelines).
Note: Palmer et al. (2001), Steed and Gibson (1997), Raunair et al. (2012) and Thomas and Hill (1997) did not report this outcome measure. Seager et al. (2006) provided information as odds ratios. (a) p<0.001. (b) p = 0.01. (c) p-values were not reported within this publication.
Fig 4.
Risk of bias within included studies.
Note: Low risk of bias is indicated by green colour, moderate risk of bias by yellow colour and high risk of bias by red colour. The question mark indicates an unknown risk.
Fig 5.
Risk of bias across included studies.
Note: * Allocation concealment and sequence generation apply only to RCTs (Seager et al. and Elouafkaoui et al.) and are not applicable to pre-post studies. RCT = randomised controlled trial.