Scolaris Content Display Scolaris Content Display

Study flow diagram.
Figures and Tables -
Figure 1

Study flow diagram.

Table 1. List of reviews, interventions, and trials that contributed to the overview with ROBIS assessment

Intervention

Review

RCTs contributed and sample size (n)

ROBIS assessment

Phase 2 domains

Phase 3

1. Study eligibility criteria

2. Identification and selection of studies

3. Data collection and study appraisal

4. Synthesis and findings

5. Risk of bias of review

POCT: C‐reactive protein

Aabenhus 2014

Andreeva 2014 (179)

Cals 2009 (431)

Cals 2010 (258)

Diederichsen 2000 (812)

Little 2013 (4264)

Melbye 1995 (239)

Low risk

Low risk

Low risk

Low risk

Low risk

Huang 2013

Cals 2009 (431)

Cals 2010 (258)

Cals 2011 (330)

Cals 2013* (379)

Diederichsen 2000 (812)

Gonzales 2011 (131)

Melbye 1995 (229)

Low risk

Low risk

Low risk

High risk

Low risk

POCT: rapid viral detection test

Doan 2014

Bonner 2003 (391)

Doan 2009 (200)

Poehling 2006 (300)

Low risk

Low risk

Low risk

Low risk

Low risk

POCT: procalcitonin

Schuetz 2012

Briel 2008 (458)

Burkhardt 2010 (550)

Christ‐Crain 2004 (243)

Christ‐Crain 2006 (302)

Kristoffersen 2009 (210)

Long 2009 (127)

Long 2011 (156)

Schuetz 2009 (1359)

Stolz 2007 (208)

Low risk

Low risk

Low risk

Low risk

Low risk

Educational materials for clinician or decision support, or both

Boonacker 2010

Christakis 2001 (NR)

Wilson 2002 (NR)

Low risk

Low risk

Low risk

High risk

Low risk

Interventions to support shared decision making

Coxeter 2015

Altiner 2007 (2164)

Briel 2006 (552)

Butler 2012 (479,502)

Cals 2009 (431)

Cals 2013* (379)

Francis 2009 (558)

Légaré 2011 (151)

Légaré 2012 (359)

Little 2013 (4264)

Welschen 2004 (1723)

Low risk

Low risk

Low risk

High risk

Low risk

Patient information leaflets to be used by clinician

de Bont 2015

Francis 2009 (558)

Macfarlane 1997 (1014)

Low risk

Low risk

Low risk

High risk

Low risk

Multifaceted interventions

Arnold 2005

Finkelstein 2001 (NR)

Flottorp 2002 (NR)

McIsaac 1998 (NR)

McIsaac 2002 (NR)

Mainous 2000 (NR)

Low risk

Low risk

Low risk

High risk

Low risk

*Cals 2013 is a follow‐up study of the trial reported in Cals 2009.

NR: not reported
POCT: point‐of‐care‐test
RCT: randomised controlled trial

Figures and Tables -
Table 1. List of reviews, interventions, and trials that contributed to the overview with ROBIS assessment
Table 2. Characteristics of excluded reviews

Reason for exclusion

Reviews excluded (n = 15)

Does not include RCTs including parallel‐group, cluster, or factorial RCTs

Petrozzino 2010

Does not include studies that include patients presenting to primary care with acute respiratory infection

Patel 2007

Does not include interventions aimed at health professional with the primary goal of reducing antibiotic prescribing

Andrews 2012; Arroll 2003; Petrozzino 2010; Spurling 2013

Does not investigate the effect of the intervention on antibiotic prescribing compared to usual care or control

Andrews 2012; Arroll 2003; Rausch 2009; Schuetz 2011; Spurling 2013

Duplication of included review

Schuetz 2013 (duplicate of included Cochrane review Schuetz 2012)

Data were not reported at an individual‐study basis.

Gross 2001; van der Velden 2012

No novel coverage in addition to included Cochrane Review

Cooke 2015; Engel 2012 (both fully overlap with Aabenhus 2014 in terms of included trials)

Rated as high risk in ROBIS quality assessment

Ranji 2008; van der Does 2016; Vodicka 2013

RCT: randomised controlled trial

Figures and Tables -
Table 2. Characteristics of excluded reviews
Table 3. Characteristics of included reviews

Review

Date assessed as up‐to‐date

Included RCTs (n)*

Population*

Intervention

Comparison intervention

Primary outcome*

Limitations

Aabenhus 2014

Around January 2014

6

Patients presenting with ARI in general practice

Point‐of‐care tests including C‐reactive protein, procalcitonin, and white blood cell count

Usual care

Number of patients given antibiotic prescription at index consultation and at 28 days' follow‐up

Small number of included studies

Arnold 2005

December 2002

5

Patients presenting with RTI in primary care

Professional interventions in the Cochrane Effective Practice and Organisation

of Care Group. To include: education materials for clinician, educational meetings, local consensus processes, educational outreach visits, local opinion leaders, patient‐mediated interventions, audit and feedback, reminders, marketing, mass media, financial interventions

Usual care or other intervention

Decision to prescribe an antibiotic or not

None reported.

Boonacker 2010

February 2009

2

Children presenting with URTI in primary care

Computerised evidence‐based decision support, development of clinical practice guidelines, and educational materials for clinician

Usual care

Antibiotic prescribing and reduced proportion of prescribed antibiotic courses

Potential for publication bias in search. Cost‐effectiveness not evaluated. Insufficient information to judge risk of bias in some areas

Coxeter 2015

December 2014

9 (reported in 10 papers)

Patients presenting with ARI in primary care

Shared decision making

Usual care or other intervention

Prescription of antibiotics

Intraclass correlation was imputed for 2 studies.

de Bont 2015

April 2014

2

Patients presenting with ARI in general practice

Patient information leaflets (PIL) for use by clinicians in consultations

No PIL

Antibiotic prescribing

Studies were at high risk of bias as blinding to the intervention was not possible. Patients included and outcomes assessed were diverse. No heterogeneity analysis was conducted.

Doan 2014

Around July 2014

3

Children presenting with ARI in the emergency department

Rapid viral diagnosis by testing

No test or test result not made known to clinician

Antibiotic prescribing rate

None reported.

Huang 2013

June 2013

6 (reported in 7 papers)

Patients presenting with ARI in general practice

C‐reactive protein point‐of‐care test

No test or usual care, or both

Antibiotic prescribing at index consultation

Meta‐analysis is based on aggregate data rather than individual‐participant data, so difficult to explore sources of heterogeneity.

Schuetz 2012

2011

9

Adults presenting with ARI in primary care

Strategies to initiate or discontinue antibiotic therapy based on procalcitonin cut‐off ranges

No use of procalcitonin

Initiation of antibiotics

Variation in patient populations/settings and treatment failure was defined differently for each context.

*relevant to this overview
ARI: acute respiratory infection
RCT: randomised controlled trial
RTI: respiratory tract infection
URTI: upper respiratory tract infection

Figures and Tables -
Table 3. Characteristics of included reviews
Table 4. Quality of included studies: the proportion of studies within each review judged as at low risk of bias by review authors according to 'Risk of bias' domains

Review ID

Random sequence generation

Allocation concealment

Blinding of participants

Blinding of outcome assessment

Incomplete outcome data

Selective reporting

Other bias

Aabenhus 2014

(6 RCTs)

5 studies*

(83%1)

1 study

(17%)

None

(0%)

5 studies*2

(83%)

6 studies

(100%)

4 studies

(67%)

1 study

(17%)

Arnold 2005

(5 RCTs)

Not reported

1 study

(20%)

Not reported

3 studies

(60%)

2 studies

(20%)

Not reported

Not reported

Boonacker 2010

(2 RCTs)

2 studies

(100%)

1 study

(50%)

2 studies

(100%)

2 studies

(100%)

2 studies

(100%)

2 studies

(100%)

1 study

(34%)

Coxeter 2015

(10 RCTs2)

10 studies

(100%)

7 studies

(70%)

1 study

(10%)

7 studies

(70%)

9 studies

(90%)

10 studies (100%)

7 studies

(70%)

de Bont 2015

(2 RCTs)

1 study

(50%)

2 studies

(100%)

None

(0%)

1 study

(50%)

2 studies

(100%)

2 studies

(100%)

Not reported

Doan 2014

(3 RCTs)

3 studies

(100%)

1 study

(34%)

None

(0%)

None

(0%)

2 studies

(67%)

3 studies

(100%)

3 studies

(100%)

Huang 2013

(7 studies3)

1 study*

(14%)

1 study

(14%)

None

(0%)

None*

(0%)

5 studies

(71%)

7 studies

(100%)

None

(0%)

Schuetz 2012

(9 RCTs)

7 studies

(78%)

4 studies

(33%)

None

(0%)

4 studies

(33%)

9 studies

(100%)

9 studies (100%)

3 studies

(43%)

RCT: randomised controlled trial

*Results differ between reviews where RCTs had significant overlap. Aabenhus 2014 provided more detailed reporting of risk of bias than Huang 2013.
1Percentages report the proportion of RCTs within each review judged as at low risk of bias on the relevant item.
2Aabenhus 2014 reported blinding for primary outcome of antibiotic prescribing separately to blinding for other outcomes.
3Coxeter 2015 and Huang 2013 reported results for 10 and 7 publications of 9 and 6 original RCTs, respectively.

Figures and Tables -
Table 4. Quality of included studies: the proportion of studies within each review judged as at low risk of bias by review authors according to 'Risk of bias' domains
Table 5. Overview of results

Interventions to reduce antibiotic prescribing in the management of acute respiratory infections for patients presenting in primary care

Outcome

Intervention and comparison intervention

Contributing reviews

Relative effect (95% CI) of an antibiotic being prescribed1

Number of participants (RCTs)

Quality of the evidence (GRADE)*

Comments

Change in antibiotic prescriptions for ARI (at consultation)

CRP point‐of‐care test / usual care

Aabenhus 2014

RR 0.78 (0.66 to 0.92)

3284 (6)

Moderate6

CRP point‐of‐care testing probably reduces antibiotic prescribing in general practice compared to usual care. However, CRP testing may have little or no effect on prescribing in emergency departments.

Huang 2013

General practice setting (individual trials reported):

RR 0.57 (0.44 to 0.74)

RR 0.58 (0.45 to 0.74)

Emergency department setting:

RR 1.23 (0.76 to 1.99)

330 (1)

379 (1)

131 (1)

Moderate3*

Low3,6*

Rapid viral diagnosis / usual care

Doan 2014

RR 0.86 (0.61 to 1.22)

891 (3)

Low3,6*

Rapid viral diagnosis may have little or no effect on antibiotic prescribing compared to usual care.

Procalcitonin‐guided management / usual care

Schuetz 2012

General practice setting:

adjusted OR 0.10 (0.07 to 0.14)

Emergency department setting:

adjusted OR 0.34 (0.28 to 0.43)

1008 (2)

2605 (7)

Moderate3

Moderate3

Procalcitonin‐guided management probably reduces antibiotic prescribing in general practice and the emergency department compared to usual care.

Clinician education and decision support / usual care

Boonacker 2010

Difference in behaviour change ‐12% (0.095)

OR 0.60 (0.43 to 0.83)

Not reported (1)

Not reported (1)

Very low3,5,6*

We are uncertain about whether clinician education and decision support reduces antibiotic prescribing compared to usual care.

Patient information leaflets / usual care

de Bont 2015

RR 0.47 (0.36 to 0.64)

RR 1.15 (0.89 to 1.48)

558 (1)

1014 (1)

Very low3,4,6*

We are uncertain as to whether patient information leaflets reduce antibiotic prescribing compared to usual care.

Shared decision making / usual care

Coxeter 2015

No pooled analysis of all trials:

OR 0.44 (0.26 to 0.75)

RR 0.64 (0.49 to 0.84)

adjusted risk difference ‐18.44 (‐27.24 to ‐9.65)

3274 (3)

4623 (2)

481,807 (4)

Moderate3

Shared decision making probably reduces antibiotic prescribing compared to usual care.

Multifaceted interventions

Printed educational materials for clinicians and reminders / usual care

Arnold 2005

Individual trials reported:

OR 0.44 (0.21 to 0.92)

OR 0.57 (0.27 to 1.17)

Not reported (1)

Not reported (1)

Very low3,4,6*

We are uncertain as to whether printed educational materials reduce antibiotic prescribing compared to usual care.

Audit and feedback (with patient education materials) / usual care

Arnold 2005

Difference in behaviour change: ‐7.3% (audit and feedback alone)

Difference in behaviour change: ‐7.2% (audit and feedback with patient materials)

Not reported (1)

Low3,6*

Audit and feedback may reduce antibiotic prescribing compared to usual care.

Educational materials and educational meetings for clinician and patient education materials / usual care

Arnold 2005

Difference in behaviour change:

16% (8% to 23%) (age 3 to 36 months)

Difference in behaviour change:

12% (2% to 21%) (age 36 to 72 months)

Not reported (1)

Low3,6*

Educational materials and meetings may reduce antibiotic prescribing compared to usual care.

Educational materials, computerised decision support, professional development, financial incentive, and patient education materials / usual care

Arnold 2005

Difference in behaviour change: ‐3.0% (P = 0.03)

Not reported (1)

Low3,6*

These interventions may slightly reduce antibiotic prescribing compared to usual care.

Change in antibiotic prescriptions for ARI (within 28 days of consultation)

CRP point‐of‐care test / usual care

Aabenhus 2014

RR 0.80 (0.67 to 0.96)

3284 (6)

Moderate6

CRP point‐of‐care testing probably reduces antibiotic prescribing for up to 28 days following consultation compared to usual care.

Adverse events

CRP point‐of‐care test / usual care

Aabenhus 2014

RR 2.45 (0.65 to 9.19)

4264 (1)

Moderate6

CRP point‐of‐care testing probably results in little or no difference in adverse events compared to usual care.

Symptom duration or severity

CRP point‐of‐care test / usual care

Aabenhus 2014

At 7 days: RR 1.03 (0.93 to 1.14)

At 28 days: RR 0.94 (0.69 to 1.28)

1309 (3)

849 (3)

Moderate6

CRP point‐of‐care testing probably results in little or no difference in symptom duration or severity compared to usual care.

Health‐related quality of life

Procalcitonin‐guided management / usual care

Schuetz 2012

Adjusted difference in days 0.05, ‐0.46 to 0.56, P = 0.854

1008 (2)

Moderate3

Procalcitonin‐guided management probably results in little or no difference in health‐related quality of life compared to usual care.

Patient satisfaction

CRP point‐of‐care test / usual care

Aabenhus 2014

RR 0.79 (0.57 to 1.08)

689 (2)

Moderate6

CRP point‐of‐care testing probably results in little or no difference in patient satisfaction compared to usual care.

Shared decision making / usual care

Coxeter 2015

RR 0.86 (0.57 to 1.30)

1110 (2)

Low3,6

Shared decision making may result in little or no difference in patient satisfaction compared to usual care.

Management failure – reconsultation and treatment failure

CRP point‐of‐care test / usual care

Aabenhus 2014

RR 1.08 (0.93 to 1.27)

5132 (4)

Moderate6

CRP point‐of‐care testing probably results in little or no difference in reconsultation compared with usual care at 28 days' follow‐up.

Shared decision making / usual care

Coxeter 2015

RR 0.87 (0.74 to 1.03)

1860 (4)

Moderate3

Shared decision making probably results in little or no difference in reconsultation compared to usual care.

Patient information leaflets / usual care

de Bont 2015

Individual trials reported:

RR 0.80 (0.52 to 1.21)

RR 0.70 (0.53 to 0.91)

558 (1)

1014 (1)

Very low3,4,6*

We are uncertain as to whether patient intervention leaflets result in a difference in reconsultation compared to usual care.

Rapid viral diagnosis / usual care

Doan 2014

RR 0.86 (0.59 to 1.25)

200 (1)

Low3,6*

Rapid viral diagnostics may result in little or no difference in reconsultation relative to usual care.

Procalcitonin‐guided management / usual care

Schuetz 2012

Treatment failure in general practice2:

adjusted OR 0.95 (0.73 to 1.24)

Treatment failure in emergency department2:

adjusted OR 0.76 (0.61 to 0.95)

1008 (2)

2605 (7)

Moderate3

Procalcitonin‐guided management probably results in little or no difference in treatment failure in general practice compared to normal care.

Procalcitonin‐guided management probably reduces treatment failure in the emergency department compared to usual care.

GRADE quality of evidence and definitions

High quality: Further research is very unlikely to change our confidence in the estimate of effect.

Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.

Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.

Very low quality: Any estimate of effect is very uncertain.

*GRADE criteria were applied retrospectively to outcomes when GRADE was not used by the original review authors.

ARI: acute respiratory infection; CI: confidence interval; CRP: C‐reactive protein; OR: odds ratio; RCT: randomised controlled trial; RR: risk ratio

1Effect estimates are shown as reported in the original reviews. Multiple effect estimates are reported for some outcomes when reviews did not pool data from trials but reported individual trials separately.
2Schuetz 2012 defined treatment failure in primary care as death, hospitalisation, ARI‐specific complications (e.g. empyema for lower ARIs, meningitis for upper ARIs), recurrent or worsening infection, and still having ARI‐associated discomfort at 30 days. Treatment failure in the emergency setting was defined as death, intensive care unit admission, rehospitalisation after index hospital discharge, ARI‐associated complications (e.g. empyema or acute respiratory distress syndrome for lower ARIs), and recurrent or worsening infection within 30 days of follow‐up.
3Quality of evidence was downgraded one level because of risk of bias: inadequate methods of sequence generation, lack of allocation concealment, and/or lack of blinding of participants.
4Quality of evidence was downgraded one level because of inconsistency: heterogeneity in results likely due to differences in the interventions trialled across studies.
5Quality of evidence was downgraded one level because of risk of publication bias: review only reported effective interventions.
6Quality of evidence was downgraded one level due to imprecision because trials included relatively few patients or when the confidence interval showed substantial variation in the effect of the intervention.

Figures and Tables -
Table 5. Overview of results