Thursday, April 28, 2005

Use of Hospital Antibiograms as a Quality Indicator

Martha S. Boehme and Patricia A. Somsel.

Antimicrobial susceptibility testing (AST) is a highly complex test in clinical laboratories. Prior to NCCLS M39-A publication in 2002, no guidance existed for standardized preparation of cumulative AST data (antibiograms).

To measure effectiveness of efforts to promote AST practice guidelines and determine if guidelines are being implemented in Michigan hospital laboratories by analyzing antibiograms submitted to Michigan Department of Community Health (MDCH) Bureau of Laboratories.

Clinical microbiology laboratories in Michigan were asked, on a voluntary basis, to submit antibiograms from years 2000 through 2003 to the Michigan Department of Community Health. Representative antibiograms were obtained from a total of 40 (36%) out of 110 laboratories. Antibiograms were analyzed for unlikely resistance patterns, appropriateness of reported drugs and compliance with selected recommendations from M39-A. Errors found were categorized as major (reporting misleading or inappropriate organism/drug combinations, reporting impossible/ unlikely resistance patterns) or minor (misspelled organism names/antimicrobials, obvious math errors).

Compliance with NCCLS M-39 A is increasing in some areas, and most errors are decreasing. Major errors decreased, from 56% of antibiograms in year 2000, to 20% of antibiograms in year 2003. Minor errors also decreased, from 13% in 2000, to 9% in 2003. The percentage of laboratories reporting Streptococcus pneumoniae data increased, from 75 % in 2000 to 89% in 2003, with an increase in dual breakpoint reporting noted. Forty-four percent of antibiograms from 2000 and 61% from 2003 presented data for organisms with <10 isolates, indicating this recommendation may be problematic, especially for smaller laboratories.

Laboratories may have difficulty implementing some recommendations in NCCLS M39-A. The decrease in major errors indicates that laboratories are increasing compliance with recommendations to verify unusual patient results before release. Further adoption of guidelines will provide more reliable data to clinicians to guide antibiotic choice.

See more of Poster Session
See more of The 2005 Institute for Quality in Laboratory Medicine Conference