Optimizing Antibiotic Prescribing: Are Physicians Following National Trends in MRSA Infections Rather than Local Data when Treating MRSA Wound Infections?

Schweizer ML, Perencevich EN, Eber MR, Shardell MD, Laxminarayan R. Optimizing Antibiotic Prescribing: Are Physicians Following National Trends in MRSA Infections Rather than Local Data when Treating MRSA Wound Infections? Poster session presented at: Interscience Conference on Antimicrobial Agents and Chemotherapy; 2010 Sep 13; Boston, MA.

Background: Physicians often prescribe antibiotics for outpatient (OP) wound infections before culture results are known. Local or national MRSA rates may be considered when prescribing anti-MRSA therapy. If physicians prescribe in response to national MRSA trends, not local, prescribing may improve by making local data accessible. Objective: to compare OP trends in antibiotic prescribing and MRSA wound infections across zip code (ZC), state, and national levels. Methods: Monthly MRSA+ wound culture counts were obtained from The Surveillance Network, a database of antibiotic susceptibilities from clinical laboratories across 133 ZCs from 1999-2007. Monthly OP retail sales of linezolid (LZ), clindamycin (CL) and trimethoprim/sulfamethoxazole (TMP) from 1999-2007 were obtained from the IMS Xponent database. Rates were created using Census 2000 populations. The proportion of variance in prescribing that could be explained by MRSA rates was assessed by the coefficient of determination (R2), using population weighted linear regression. Results: 107,215 MRSA+ wound cultures and 106,641,604 antibiotic prescriptions were assessed. The R2 was low ( < 0.1) when ZC antibiotic prescription rates were compared to MRSA rates at all levels. State level prescriptions of TMP were not well described by state or national rates of MRSA (R2 < 0.01). State level prescriptions of CL and LZ were correlated with state MRSA rates (CL R2 = 0.12, LZ R2 = 0.27; p-values < 0.01). The variance in state level prescribing of CL and LZ was best explained by national MRSA rates (CL R2 = 0.33, LZ R2 = 0.59; p-values < 0.01). Conclusion: Physicians may be relying on national, not local, MRSA data when prescribing CL and LZ for wound infections. Sharing local MRSA rates or including local data in clinical guidelines may improve antibiotic prescribing.