494. Use of a Clinical Prediction Tool to Predict Clostridium difficile Infection
Session: Poster Abstract Session: Healthcare Epidemiology: Updates in C. difficile
Thursday, October 4, 2018
Room: S Poster Hall
  • Conference Poster.pdf (441.5 kB)
  • Background: Clostridium difficile is a pathogen that may be a component of normal microbiota. In 2011 there were an estimated 453,000 cases of CDI in the United States and 29,300 deaths. Diagnosis of CDI is of often accomplished through nucleic acid amplification testing (NAAT) for C. difficile toxin genes, which carries a risk of false positive results. In 1996, Katz et al. created a screen for CDI that was positive if the patient had significant diarrhea and either abdominal pain or prior antibiotic usage. Today, we believe this tool is worth revisiting with increased incidence of CDI and improved testing methods. Our aim is to determine the current usefulness of the Katz et al. 1996 clinical decision tool for CDI.

    Methods: We conducted a retrospective cross-sectional chart review at a Midwestern teaching hospital. All patients tested for CDI between June 1, 2016 and May 31, 2017 were initially eligible. Participants were excluded from data collection on the basis of missing information, a previous positive CDI test in the last 8 weeks or age <18 years. Charts were reviewed for age, sex, diarrhea, abdominal pain, antibiotic use, prior positive testing for CDI and length of hospitalization. Data was analyzed using SAS Software

    Results: Of the initial 432 charts analyzed, 202 (46.8%) had no documented amount of diarrhea and 16 more were missing other data points, leaving 214 of 432 (49.5%) charts that included all data to be used for analysis. Of these 18 of 214 (8.4%) were positive results. The Katz screen was positive in 85 of 214 (40.2%) cases. The sensitivity, specificity, positive predictive value and negative predictive value respectively were 61, 62, 13 and 95.

    Conclusion: Katz et al. found a sensitivity, specificity, positive predictive value and negative predictive value of 80, 45, 18 and 94 respectively. The differences between these values and our own may be due to changes in the testing methodology and prevalence of CDI compared to a 1992 study population. The negative predictive value remains a strength. If this screening tool had been applied to our population, there may have been 128 (59.8%) fewer tests, but 7 (38.9%) missed positive results.

    Marten Hawkins, MPH, Michael Baumgartner, BS Neurobiology, Danielle VanBeckum, BS, MS, Julia Buck, BS, BA and Crystal Holley, BS, Michigan State University College of Human Medicine, East Lansing, MI


    M. Hawkins, None

    M. Baumgartner, None

    D. VanBeckum, None

    J. Buck, None

    C. Holley, None

    Findings in the abstracts are embargoed until 12:01 a.m. PDT, Wednesday Oct. 3rd with the exception of research findings presented at the IDWeek press conferences.