Antibiotic policies in acute English NHS trusts: implementation of 'Start Smart-Then Focus' and relationship with Clostridium difficile infection rates.

OBJECTIVES The objective of this study was to establish how antibiotic prescribing policies at National Health Service (NHS) hospitals match the England Department of Health 'Start Smart-Then Focus' recommendations and relate to Clostridium difficile infection (CDI) rates. METHODS Antibiotic pharmacists were surveyed regarding recommendations for empirical treatment of common syndromes ('Start Smart') and antimicrobial prescription reviews ('Focus') at their hospital trusts. If no response was provided, policy data were sought from trust websites and the MicroGuide app (Horizon Strategic Partners, UK). Empirical treatment recommendations were categorized as broad spectrum (a β-lactam penicillin/β-lactamase inhibitor, cephalosporin, quinolone or carbapenem) or narrow spectrum. CDI rates were gathered from the national mandatory surveillance system. RESULTS Data were obtained for 105/145 English acute hospital trusts (72%). β-Lactam/β-lactamase inhibitor combinations were recommended extensively. Only for severe community-acquired pneumonia and pyelonephritis were narrow-spectrum agents recommended first line at a substantial number of trusts [42/105 (40%) and 50/105 (48%), respectively]. Policies commonly recommended dual therapy with aminoglycosides and β-lactams for abdominal sepsis [40/93 trusts (43%)] and undifferentiated severe sepsis [54/94 trusts (57%)]. Most policies recommended treating for ≥ 7 days for most indications. Nearly all policies [100/105 trusts (95%)] recommended antimicrobial prescription reviews, but only 46/96 respondents (48%) reported monitoring compliance. Independent predictors of higher CDI rates were recommending a broad-spectrum regimen for community-acquired pneumonia (P=0.06) and, counterintuitively, a recommended treatment duration of <48 h for nosocomial pneumonia (P=0.01). CONCLUSIONS Hospital antibiotic policies in the NHS 'Start Smart' by recommending broad-spectrum antibiotics for empirical therapy, but this may have the unintended potential to increase the use of broad-spectrum antibiotics and risk of CDI unless better mechanisms are in place to improve 'Focus'.