Background
Inappropriate antibiotic use is a known driver of antimicrobial resistance. Primary care antibiotic prescribing accounts for approximately 80% of antibiotic consumption in England, with respiratory tract infections (RTIs) being the most common indication. RTIs are largely viral and self-limiting, and antibiotics are often inappropriate. The 2015/16 NHS England Quality Premium (QP) financially incentivised reductions in primary care antibiotic prescribing. This may have led to unintended consequences such as a reduction in appropriate antibiotic treatment with some patients developing more severe infections.
Aim
To assess the reduction in antibiotic prescribing following introduction of the 2015/16 QP, and the occurrence of unintended consequences in patients presenting to English general practices with RTIs, as measured by re-consultations, severe infections (in primary and secondary care) and death.
Methods
A systematic literature review and meta-analysis were undertaken pooling evidence on the risk of RTI complications where there was lack of exposure to timely antibiotic treatment. This contributed to a modified Delphi method, defining RTI infection pathways. Subsequent investigations of the potential impact of the QP used linkage of routinely collected national healthcare datasets. Interrupted time series analysis (ITSA) and hierarchical multivariable analysis were the statistical methods utilised, comparing antibiotic prescribing for RTIs in general practices across England, and unintended consequence (measured by re-consultations, severe infections and death, within 30-days of an initial RTI) pre- and post-QP.
Results
The systematic review found that RTI complications were rare. The pooled odds ratio favoured the use of antibiotics in preventing RTI complications. There was a high-level of heterogeneity between studies, high risk of bias (particularly indication bias) and studies were often not powered or designed to assess complications. ITSA demonstrated that antibiotic prescribing for RTIs decreased over the six-year study period, with a significantly decrease coinciding with the introduction of the QP (decline was particularly evident in children, <16y). The ITSA assessing potential impact on unintended consequences did not find evidence of greater risk of re-consultation, or in complications reported in general practices, hospital admissions, or mortality. However, increases in complications (e.g. pneumonia in primary care and bloodstream infections in secondary care [p>0.05]) were reported, particularly for elderly patients (≥65 years) and patients who had been prescribed antibiotics; increased mortality was also noted, although this was not sustained. Complications were shown to have been on a gradual rise prior to the QP, hence findings from the multivariable analysis may reflect this increase. Findings from this analysis also showed a significant reduction in antibiotic prescribing post-QP, and greater odds of complications in patients who had been prescribed antibiotics compared to those who had not.
Conclusions
The 2015/16 QP has been safely implemented with no significant unintended consequences. Future reductions in antibiotic prescribing should be tailored based on infection indication and patient risk factors (e.g. by age/elderly). Future surveillance would benefit from improvements in national primary care data acquisition, linkage and surveillance of unintended consequences.