Portal de Periódicos da CAPES

The effect of implementing the British Thoracic Society community-acquired pneumonia guidelines on antibiotic prescribing and costs in a UK teaching hospital

2006; Elsevier BV; Volume: 12; Issue: 5 Linguagem: Inglês

10.1111/j.1469-0691.2006.01387.x

1469-0691

Gavin Barlow, Dilip Nathwani, Peter Davey,

Antibiotic Use and Resistance

The recent article in CMI by Ooesterheert et al. [1Oosterheert JJ Bonten MJM Schneider MME Hoepelman IM Predicted effects on antibiotic use following introduction of British or North American guidelines for community‐acquired pneumonia in The Netherlands.Clin Microbiol Infect. 2005; 11: 992-998Crossref PubMed Scopus (11) Google Scholar] predicted a 23% increase in antibiotic use following hypothetical implementation of the British Thoracic Society (BTS) community-acquired pneumonia (CAP) guidelines at a teaching hospital in The Netherlands. We have recently evaluated the implementation of a management pathway to improve the time to administration and the appropriateness of empirical antibiotics for CAP at a UK teaching hospital. The intervention was based on the BTS guidelines and emphasised the link between severity assessment and appropriate antibiotic prescribing in CAP [2British Thoracic Society BTS Guidelines for the Management of Community Acquired Pneumonia in Adults.Thorax. 2001; 56: 1-64PubMed Google Scholar]. For patients with none of the BTS-recommended adverse prognostic indicators, home therapy with oral amoxycillin was recommended (‘non-severe, home therapy’ group). For patients with one adverse prognostic indicator, inpatient care with oral/intravenous (IV) amoxycillin plus oral erythromycin or IV clarithromycin was recommended (‘non-severe, inpatient therapy’ group). For patients with two or more adverse prognostic indicators, IV co-amoxyclav plus IV clarithromycin was recommended (‘severe’ group). Upgrading of management was allowed if clinical judgement suggested that this was required. The intervention was implemented in October 2002 using reminders, educational sessions, audit and feedback, and was evaluated using a controlled before–after design. At baseline (November 2001 to April 2002), the proportion of patients in the non-severe, inpatient therapy group who received amoxycillin plus a macrolide was 49.5% (51/103) at the intervention site and 69% (22/32) at the control site (difference = +19.5%; χ2 3.64, p 0.06), compared with 21% (20/95) at the intervention site and 54% (20/37) at the control site in the 6-month period (November 2002 to April 2003) following implementation (difference = –33%; χ2 13.73, p 0.0002%; difference in absolute change from baseline = –13.5%). In contrast, the proportion of patients receiving either co-amoxyclav/cefuroxime / ceftriaxone plus a macrolide, or levofloxacin monotherapy, increased from 23.5% (24/103) to 62% (59/95) at the intervention site, and from 22% (7/32) to 32.5% (12/37) at the control site (difference post-implementation = +29.5%; χ2 9.43, p 0.002; difference in absolute change from baseline = +28%). In the severe group, the prescribing of amoxycillin plus a macrolide decreased from 24% (17/71) to 6.5% (7/109) at the intervention site, but increased from 14.5% (4/28) to 41.5% (5/12) at the control site (difference post-implementation = –35%, p 0.004 (Fisher's exact test); difference in absolute change from baseline = +44.5%). In contrast, prescribing of either co-amoxyclav/cefuroxime/ceftriaxone plus a macrolide, or levofloxacin monotherapy, increased from 49.5% (35/71) to 83.5% (91/109) at the intervention site, and from 35.5% (10/28) to 41.5% (5/12) at the control site (difference post-implementation = +42%; p 0.006 (Fisher's exact test); difference in absolute change from baseline = +28%). Overall, broad-spectrum antibiotic prescribing increased from 42.5% to 77.5% at the intervention site, but decreased from 45% to 41.5% at the control site (difference post-implementation = +36%, χ2 26.21, p < 0.0001; unadjusted difference in absolute change from baseline = +38.5%; relative percentage change post-implementation = +87%). There were no significant differences or trends towards improvement in length of hospital stay or 30-day mortality rates. Before implementation, the mean cost of the first dose of antibiotic(s) prescribed was £6.34 (Euro 8.9) at the intervention site and £4.83 (Euro 6.8) at the control site (unadjusted difference +£1.51 (Euro 2.1), 95% CI £0–3.03 (Euro 0–4.2), p 0.05; adjusted difference (for severity) +£1.62 (Euro 2.3), 95% CI £0.02–3.21 (Euro 0.3–4.5), p 0.05). Post-implementation, costs increased at both sites to £9.01 (Euro 12.6) at the intervention site and £6.42 (Euro 9.0) at the control site (unadjusted difference +£2.59 (Euro 3.6), 95% CI £1.02–4.16 (Euro 1.4–5.8), p 0.0001; adjusted difference +£1.56 (Euro 2.2), 95% CI £0.01–3.12 (Euro 0.1–4.4), p 0.05; unadjusted difference in absolute change from baseline = +£1.08 (Euro 1.5). These results confirm the prediction of Ooesterheert et al. [1Oosterheert JJ Bonten MJM Schneider MME Hoepelman IM Predicted effects on antibiotic use following introduction of British or North American guidelines for community‐acquired pneumonia in The Netherlands.Clin Microbiol Infect. 2005; 11: 992-998Crossref PubMed Scopus (11) Google Scholar] that implementation of the BTS CAP guidelines would increase broad-spectrum antibiotic consumption. Although our intervention reduced under-treatment of patients with severe CAP significantly, physicians appear to have taken a ‘just in case’ approach for less severe patients, which resulted in a significant increase in broad-spectrum prescribing for non-severe CAP at the intervention site, probably against a background of temporal trends at both sites. This was despite the emphasis placed on the link between severity assessment and appropriate antibiotic prescribing (both under- and over-treatment) in the intervention. Although the costs of the first dose of antibiotic(s) did not increase significantly compared with the control site, it is possible that the increase in broad-spectrum prescribing may have impacted on patient outcomes (e.g., by causing Clostridium difficile-associated diarrhoea or antibiotic-resistant infections), thereby increasing costs indirectly. An increase in adherence to the BTS CAP guidelines has been suggested previously as the cause of an increase in cephalosporin prescribing in a UK teaching hospital [3Monnet DL MacKenzie FM López‐Lozano JM et al.Antimicrobial drug use and methicillin‐resistant Staphylococcus aureus, Aberdeen, 1996–2000.Emerg Infect Dis. 2004; 10: 1432-1441Crossref PubMed Scopus (170) Google Scholar], which resulted in an increase in the prevalence of MRSA. In future, trials assessing clinical- and cost-effectiveness of antibiotic guidelines should measure and take account of such adverseevents.