Using different classification to calculate antibiotic susceptibility of bacteria isolated from Chang Gung Memorial Hospital

• Main Authors: Hsing-Yu Chang, 張幸宇
• Other Authors: Der-Ming Liou
• Format: Others
• Language: en_US
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/44149388000894992038

碩士 === 國立陽明大學 === 衛生資訊與決策研究所 === 92 === Surveillance of antimicrobial susceptibility of bacteria is an important role for the control of infectious disease. The antibiotic susceptible reports can be used for trends prediction, rules creation, direction for prescription, and comparison between different areas. Traditionally, source of antimicrobial susceptibility was usually depended on manual entry and calculation for each species and antimicrobial agents. The data might be incomplete, loss of integrity, heterogeneous, and complex. Currently, there are many studies which process the data into several groups for susceptibility calculation by different criteria. These improved methods of surveillance give another choice of susceptible frequencies of bacteria from all isolates processed by traditional methods. However, there are few flexible models suitable for different criteria. The organisms analyzed in our thesis were Acinetobacter baumanii (n = 3163), Pseudomonas Aeruginosa (n = 6018), Escherichia coli (n = 11111), Klebsiella pneumonia (n = 3831), and Staphylococcus aureus (n = 8341) from 1 Jan 2003 to 31 Dec 2003 at Chang Gang Memorial Hospital, Linkou branch, Taipei branch, and Chang Gung children’s hospital. To investigate the effects of different criteria on antibiotic susceptible rates, there were three criteria in the thesis. First criteria, we excluded duplicates of the same patient with a given species during specific time intervals of 7, 15 and 30 days. Second criteria, we divided isolates into two groups of adult and children, and calculated antibiotic susceptibility of bacteria based on inpatients, outpatients and ICU patients. Last criteria, we classified isolates according to the type of specimen. Five kinds of specimen (blood, urine, body fluid, respiratory tract, and others) were defined for analysis. In the first criteria, it was discovered that isolates from the first time per patient of a given species showed the highest susceptibility among all isolates and other subgroups of different time periods. First isolates from pediatric samples, the difference of susceptibility between OPD, IPD and ICU were smaller than adults. In the adults’ group, it was shown that the general rank order of susceptible rate was: OPD > IPD > ICU. In the aspect of different types of specimen, it was appeared an inverse relationship between susceptible rate and the frequencies of ESBL strains. Event-oriented reports showed that antimicrobial susceptibility of bacteria was varied from different wards, specimen types and ages. Through our models, the hidden information from traditional antibiotic susceptibility of total isolates would be revealed. It not only gave more information through event-oriented susceptible reports, but also shortened the time-consuming process on surveillance by automated classification and calculation. It resulted in more precise and detail of the antimicrobial susceptibility of bacteria for clinicians in the selection of empiric therapy and contributed to the prevention of antibiotic resistance.