OBJECTIVES: in order to optimize prescriptions, we conducted a qualita
tive evaluation of antibiotic prescription in an intensive care unit.
METHODS: A prospective observational study was performed on 100 consec
utive prescriptions from 11/95 to 4/96. RESULTS: Among 14 documented c
ases, initial antibiotic therapy was in accordance with antimicrobial
susceptibility patterns in an but one case Among 86 empirical cases, 3
8 were secondarily documented, yielding 43 microorganisms. Of these 38
, 27 were susceptible to 2 or more empirical antibiotics, 3 to only 1
and 8 to none. Antibiotics were modified in 23/38 (60%) cases, resulti
ng in drug changes (n=21) or drug addition (n=2). In all cases, the ne
w prescription was consistent with the antibiogram. In the 48 cases wh
ere no microorganism was isolated, antibiotic change was guided by cli
nical course and occured in 6 (12.5%) cases. A switch to older, cheape
r or more narrow spectrum antibiotics was possible in 18 cases, but wa
s actually done in only 4 (22%). Dosage errors were observed in 5 case
s of initial therapy. Second line therapy contained 8(21%) dosage erro
rs. Most frequently, isolated organisms at admission were: Staphylococ
cus sp. (n=15), P. aeruginosa (n=11) and S. pneumoniae (n=10). New pat
hogens emerged in 16 patients (16%) receiving antibiotics. The most fr
equent was P. aeruginosa in 4 patients receiving ofloxacin + amoxicill
in +/- davulanic acid. CONCLUSION: These results are encouraging, howe
ver, the use of guidelines and periodic evaluation of antibiotic presc
ription practices might improve the efficiency of empirical antibiotic
prescriptions and reduce overall antibiotic costs. (C) 1998, Masson,
Paris.