Rapid Respiratory Point-Of-Care Testing and Antibiotic Use in Primary Care
Key Takeaways
- Same-day antibiotic prescribing was the same in both groups, with no overall difference in the primary outcome.
- Symptom severity, later antibiotic use, and subsequent healthcare seeking were similar across groups during follow-up.
- Subgroup differences were reported for viral detection and chronic lung disease, while the overall trial result remained unchanged.
Investigators enrolled 552 patients and assigned 276 to RM-POCT and 276 to usual care. Eligibility required age 12 months or older, clinician-diagnosed acute respiratory tract infection within 21 days, and patient or clinician belief that antibiotics were or might be necessary. Mean age was 40.0 years, 63% were female, and 25% had chronic lung disease. The intervention used BioFire FilmArray Torch 1 with RP2.1 plus reagent pouches, detecting 19 viruses and 4 atypical bacteria in about 45 minutes. Clinicians received an information sheet but no training or communication-skills guidance, so the assay was tested within routine primary care workflow.
For the primary outcome, 124 of 276 participants in each group received same-day antibiotics, yielding an OR of 1.00, 95% CI 0.71 to 1.41, with P > .99. The prespecified safety outcome also showed no difference, with a mean symptom-severity difference on days 2 to 4 of 0.09, 95% CI -0.10 to 0.27, P = .36. In prespecified subgroup analyses, lower prescribing with RM-POCT was reported among participants in whom a virus was detected, with OR 0.35, 95% CI 0.20 to 0.63, and interaction P < .001. A subgroup signal was also reported among participants with chronic lung disease, with OR 0.55, 95% CI 0.28 to 1.09, and interaction P = .046. No clear effect appeared for age younger than 16 years or baseline disagreement about antibiotic necessity, and these subgroup patterns did not change the null overall result.
Antibiotic prescribing from days 2 to 28 was similar between groups at 38 of 270 versus 32 of 273, with OR 1.24 and 95% CI 0.75 to 2.06. Participant-reported antibiotic consumption from days 1 to 28 was also similar at 107 of 236 versus 99 of 233, with OR 1.11 and 95% CI 0.78 to 1.62. There was no evidence of increased subsequent healthcare seeking, and 2 intervention participants and 1 usual-care participant were hospitalized for respiratory tract infection within 28 days. The linked rapid respiratory microbiological point-of-care test produced a valid result in 273 intervention participants, and no microbe was detected in about half. Mycoplasma pneumoniae was the most common atypical bacterium, human rhinovirus or enterovirus was the most common virus, and downstream outcomes were similar across groups.
The study stopped as prespecified when symptom-severity data were available for 412 participants. Authors highlighted higher-than-expected attrition for symptom outcomes, recruitment from lower-prescribing and more affluent or less diverse practices, and fewer children than expected. They also noted that the trial evaluated this assay-and-workflow configuration without communication-skills training.
With viruses detected in fewer than half of participants, the authors suggested a negative result may have supported presumed bacterial illness because typical bacteria were not tested. They concluded this virus-and-atypical-bacteria panel did not reduce overall same-day prescribing or worsen outcomes and was unlikely to be clinically effective, let alone cost-effective, in primary care.