Results, lessons learned and recommendations from the REST randomised controlled trial of antibiotic strategies for children with acute otitis media with discharge

Abstract

BackgroundAcute otitis media (AOM) is a painful infection of the middle ear commonly seen in children. In some children, the ear drum spontaneously bursts, discharging visible pus (otorrhoea) into the outer ear (AOMd).ObjectiveTo investigate the clinical effectiveness and economic impact of immediate topical or delayed oral antibiotics, compared with immediate oral antibiotics for symptom duration in children presenting to primary care with AOMd.DesignPragmatic, three arm, individually randomised (stratified <2 vs. ≥2 years), non-inferiority, open trial with economic and qualitative evaluations, supported by a health-record-integrated electronic trial platform (TRANSFoRm) with internal pilot.Setting44 English GP practicesParticipantsChildren aged ≥12 months and <16 years whose parents (carers) were seeking medical care for unilateral otorrhoea (ear discharge) following recent (≤7 days) onset AOM.InterventionsInterventions: (i) immediate ciprofloxacin (0.3%) solution, four drops three times daily for seven days; or (ii) delayed ‘dose-by-age’ amoxicillin suspension three times daily (clarithromycin twice daily if penicillin allergic) for seven days, with structured delaying advice. All parents given standardised information regarding symptom management (paracetamol /ibuprofen/ fluids) and to complete the course.ComparatorImmediate ‘dose-by-age’ oral amoxicillin three times (clarithromycin twice) daily for seven days. Parents received standardised symptom management advice along with advice to complete the course.Main outcome measureTime from randomisation to the first day on which all symptoms (pain, fever, being unwell, sleep disturbance, otorrhoea, and episodes of distress/crying) were rated ‘no’ or ‘very slight’ problem (without need for analgesia).MethodsParticipants recruited from routine primary care appointments. Planned sample size 399 children. Follow up using parent completed, validated symptom diary. ResultsDelays in software deployment and configuration led to low recruitment and trial closure at the end of the internal pilot. Twenty-two children (median age 5 years; 62% boys) were randomised: five, seven and ten respectively to immediate oral, delayed oral and immediate topical antibiotics. All received prescriptions as randomised. Seven (33%) fully adhered to treatment as allocated.Symptom duration data were available for 17 (77%) children. The respective median (IQR) number of days until symptom resolution in the immediate oral, delayed oral, and immediate topical antibiotic arms were 6 (4, 9), 4 (3, 7), and 4 (3, 6). Comparative analyses not conducted due to low numbers. There were no serious adverse events, and six reports of new or worsening symptoms.Qualitative clinician interviews showed the trial question was important. When the platform functioned it was liked and worked as intended. However, staff reported malfunctioning software for long periods resulting in missed recruitment opportunities. Troubleshooting the software placed significant burdens on staff. LimitationsThe overriding weakness was the failure to recruit enough children. Conclusions and future workWe were unable to answer the main research question due to a failure to reach the required sample size. Our experience of running an electronic platform supported trial in primary care has highlighted challenges from which we have drawn recommendations for the NIHR and the research community. These should be considered before such a platform is used again.

More information

Identifiers

ORCID for Michael Moore: orcid.org/0000-0002-5127-4509

ORCID for Paul Little: orcid.org/0000-0003-3664-1873

Catalogue record

Export record