Delayed Antibiotic Prescribing for Respiratory Tract Infections: an Individual Patient Data Meta-Analysis

Talk Code: 
Beth Stuart
Beth Stuart, Taeko Becque, Hilda Hounkpatin, Guiqing Yao, Shihua Zhu, Dankmar Böhning, Jennifer Bostock, Heiner C C Bucher, Michael Moore, Paul Little
Author institutions: 
University of Southampton, University of Leicester, Basel Institute for Clinical Epidemiology and Biostatistics


Antibiotics are still frequently prescribed for respiratory tract infections (RTIs), most of which are self-limiting and for which symptomatic benefit from antibiotics is modest at best. Delayed prescribing can be a useful strategy to reduce antibiotic prescribing. Whilst in some situations delayed antibiotic prescribing is appropriate, for other patients it may be unsuitable. It is important to understand which subgroups of patients may require immediate antibiotics and which patients might benefit from a delayed or no prescribing strategy.


This study undertook a systematic review in Cochrane Central Register of Controlled Trials, Ovid MEDLINE, Ovid Embase, EBSCO CINAHL Plus and Web of Science to identify all randomised controlled trials (RCTs) and observational cohort studies of delayed prescribing. The primary outcome of interest was symptom severity scores. Secondary outcomes include duration of illness and reconsultation. Interaction terms for differential effects in subgroups were explored in the following pre-defined groups: shorter prior duration of illness, age, fever at baseline consultation, comorbid lung condition and severity of symptoms at baseline consultation. IPD meta-analysis was conducted using a one-stage approach, using generalised linear mixed modelling with a random effect for study. All models controlled for baseline severity of illness and diagnosis. Propensity scores were used as inverse probability weights to control for confounding in observational studies.


We obtained data from 4 observational studies and 9 RCTs, totalling 56,301 patients. For the primary outcome, there was no statistically significant difference in symptom scores between delayed and immediate antibiotics (mean difference 0.04;95% CI -0.05,0.13) nor between delayed and no antibiotics (MD 0.03;95% CI -0.13,-0.19). The symptom duration was slightly shorter in those given immediate antibiotics (RR 1.07;95% CI 1.00, 1.14). Reconsultation was significantly less likely in those given a delayed prescription compared to those given no prescription (OR 0.69;95% CI 0.56,0.84). Subgroup analyses showed a statistically significant interaction term for those under 16 years and those with lung disease. Compared to those aged 16-64, those aged under 16 were more likely to have improved symptom scores with delayed prescribing compared to none and slightly poorer symptom scores with delayed compared to immediate. Symptom scores in those with lung disease were slightly higher in those who received a delayed prescription compared to none. However, none of these differences were represented a clinically meaningful difference. There were no statistically significant interactions in the other subgroups.


Delayed prescribing appears to be a safe and effective strategy for most patients, with no clinically significant increase in symptom severity in any of the pre-defined subgroups. Encouraging delayed prescribing as a tool in consultations may reduce reconsultation and is unlikely to be associated with an increase in symptoms or illness duration.

Submitted by: 
Beth Stuart
Funding acknowledgement: 
This paper presents independent research funded by the National Institute for Health Research (NIHR) under its Research for Patient Benefit (RfPB) Programme (Grant Reference Number PB-PG-0416-20005). The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health & Social Care.