Antibiotic Prescribing for Acute Respiratory Tract Infections 12 Months After Communication and CRP Training: A Randomized Trial

Paul Little; Beth Stuart; Nick Francis; Elaine Douglas; Sarah Tonkin-Crine; Sibyl Anthierens; Jochen W.L. Cals; Hasse Melbye; Miriam Santer; Michael Moore; Samuel Coenen; Chris C. Butler; Kerenza Hood; Mark Kelson; Maciek Godycki-Cwirko; Artur Mierzecki; Antoni Torres; Carl Llor; Melanie Davies; Mark Mullee; Gilly O’Reilly; Alike van der Velden; Adam W.A. Geraghty; Herman Goossens; Theo Verheij; Lucy Yardley

doi:10.1370/afm.2356

Article Figures & Data

Figures

Tables
Additional Files

Download figure
Open in new tab
Download powerpoint
Figure 1
CONSORT flow diagram.
CONSORT =Consolidated Standards of Reporting Trials; CRP = C-reactive protein.

View popup

Table 1

Patient Characteristics

Combination	Baseline (n = 6,771)	3 Months of Follow-up				12 Months of Follow-up

		CRP Training		Communication Training		CRP Training		Communication Training

		Control (n = 2,040)	Intervention (n = 2,224)	Control (n = 1,932)	Intervention (n = 2,332)	Control (n = 2,463)	Intervention (n = 2,699)	Control (n = 2,464)	Intervention (n = 2,698)
Sex, female, No. (%)	4,218 (62)	1,311 (64)	1,423 (64)	1,223 (63)	1,511 (65)	1,470 (60)	1,620 (60)	1,466 (60)	1,624 (60)
Age, mean (SD), y	49.6 (18.6)	50.9 (17.3)	51.0 (17.5)	50.8 (17.6)	51.1 (17.2)	51.7 (18.8)	51.2 (18.7)	51.1 (18.6)	51.7 (18.8)
Never-smoker, No. (%)	N/A	1,067 (52)	1,147 (52)	1,041 (54)	1,173 (50)	N/A	N/A	N/A	N/A
Illness duration before index consultation ≥5 days, No. (%)	3,542 (53)^a	1,038 (51)^a	1,128 (51)^a	994 (52)^a	1,172 (51)^a	1,026 (42)	1,198 (44)	1,032 (42)	1,192 (44)
Respiratory rate ≥25 breaths/minute, No. (%)	N/A	101 (5)^a	123 (6)^a	145 (8)^a	79 (3)^a	65 (3)^a	133 (5)	96 (4)	102 (4)
Temperature ≥38°C, No. (%)	N/A	228 (11)^a	280 (13)^a	202 (11)^a	306 (13)^a	291 (12)	380 (14)	285 (12)	386 (14)

CRP = C-reactive protein; N/A = not available.
↵a Data missing for some patients.

View popup

Table 2

Factorial Analysis: Efficacy of Interventions in Reducing Antibiotic Prescribing at 12 Months

Measure	No CRP Training (n = 2,360)	CRP Training (n = 2,462)	No Communication Training (n = 2,246)	Communication Training (n = 2,576)
Patients prescribed antibiotics, No. (%)	1,078 (45.7)	1,097 (44.6)	1,069 (47.6)	1,106 (42.9)
Basic risk ratio (95% CI) [P value]^a	ref	0.91 (0.77-1.05) [.22]	ref	0.90 (0.77-1.04) [.17]
Fully adjusted risk ratio (95% CI) [P value]^b	ref	0.87 (0.68-1.06) [.18]	ref	0.81 (0.64-1.00) [.049]

CRP = C-reactive protein; ref = reference group.
↵a The basic model adjusted for baseline prescribing and clustering by physician and practice.
↵b The fully adjusted model controlled for diagnosis (lower respiratory tract infection, upper respiratory tract infection, pneumonia), sex, age, presence of cough, phlegm, shortness of breath, blocked/runny nose, chest pain, fever, muscle ache, headache, disturbed sleep, feeling generally unwell, interference with social activities, earache, sore throat, facial/sinus pain, crackles, wheeze, pulse >100 beats per minute, temperature >37.8°C, respiratory rate, physician’s rating of severity, low blood pressure, duration of cough, and duration of illness before consultation.

View popup

Table 3

Individual Group Analysis: Efficacy of Interventions in Reducing Antibiotic Prescribing at 12 Months

Measure	Usual Care (n = 1,194)	CRP Training (n = 1,052)	Communication Training (n = 1,166)	Combination (n = 1,410)
Patients prescribed antibiotics, No. (%)	613 (51.3)	456 (43.4)	465 (39.9)	641 (45.5)
Basic risk ratio (95% CI) [P value]^a	ref	0.83 (0.66-1.02) [.08]	0.83 (0.66-1.01) [.07]	0.83 (0.66-1.00) [.053]
Fully adjusted risk ratio (95% CI) [P value]^b	ref	0.75 (0.51-1.00) [.052]	0.70 (0.49-0.94) [.02]	0.70 (0.49-0.93) [.01]

CRP = C-reactive protein; ref = reference group.
↵a The basic model adjusted for baseline prescribing and clustering by physician and practice.
↵b The fully adjusted model controlled for the variables listed under Table 1.

View popup

Table 4

Assessment of Attrition Bias: Completing Practices and All Practices With Imputation

Measure	Usual Care	CRP Training	Communication Training	Combination
Practices completing 12 months of follow-up
Patients prescribed antibiotics, %	51.0	42.9	39.8	47.1
Difference, % (95% CI) [P value]^a	ref	−7 (−15 to 2) [.11]	−7 (−15 to 1) [.09]	−4 (−12 to 4) [.30]
All practices, with imputation for missing data
Patients prescribed antibiotics, %	49.7	42.9	41.9	47.0
Difference, % (95% CI) [P value]^a	ref	−6 (−14 to 3) [.22]	−7 (−15 to 1) [.10]	−4 (−12 to 4) [.34]

CRP = C–reactive protein; ref = reference group.
Note: A practice-level analysis.
↵a Reduction in proportion compared with usual care controlling for baseline prescribing and practice-level averages of patient characteristics: age, type of infection (lower respiratory tract infection vs upper respiratory tract infection), presence of symptoms (crackles, wheeze, pulse higher than 100 beats per minute, temperature higher than 37.8°C), respiratory rate, physician’s rating of severity, and duration of cough.

Additional Files

Figures
Tables

Supplemental Appendixes

Supplemental Appendixes

Files in this Data Supplement:

Supplemental data: Appendixes - PDF file
The Article in Brief

Antibiotic Prescribing for Acute Respiratory Tract Infections 12 Months After Communication and CRP Training: A Randomized Trial

Paul Little , and colleagues

Background Training clinicians in communication skills to explore patients' concerns has been shown reduce antibiotic prescribing for lower respiratory tract infections. In addition, point-of-care testing for C-reactive protein can have diagnostic utility in cases of respiratory tract infection. A study of internet-based physician training for communication skills, including use of an interactive patient booklet, and testing for C-reactive protein previously found that, at 3 months, both interventions reduced antibiotic prescribing. This report examines the effect of the interventions after 12 months.

What This Study Found Although training physicians in communication skills and in testing for C-reactive protein has short-term effectiveness in reducing antibiotic prescriptions, this analysis finds that only communication training remains effective over time. In this six-country study, 246 general practices received internet-based training in patient-centered communication, including use of an interactive patient booklet, and point-of-care testing for C-reactive protein. Participants were cluster randomized to one of four groups: usual care (n=61), training for C-reactive protein point-of-care testing (n=62), training in communication skills and use of patient booklet (n=61), or combined interventions (n=62). At 12 months, antibiotic prescribing was reduced in usual care (from 58 percent at three months to 51 percent at 12 months), but increased in the C-reactive protein testing group (from 35 percent to 43 percent). When compared to three-month data, reductions in prescribing for communication training were maintained at 40 percent (465/1,166). Despite being freely provided, C-reactive protein testing was rarely used, and patient booklets were used only sparingly.

Implications

The authors suggest that, in routine primary care practice, training clinicians to use CRP testing is likely to yield short-term benefits. Training in enhanced communication skills provides the most useful long-lasting effects.