Advertisement

Epidemiology and risk factors for Clostridium difficile-associated diarrhea in adult inpatients in a university hospital in China: Methodological issues

  • Saeid Safiri
    Correspondence
    Address correspondence to Saeid Safiri, MSc, PhD, Managerial Epidemiology Research Center, Department of Public Health, School of Nursing and Midwifery, Maragheh University of Medical Sciences, Maragheh, Iran. (S. Safiri).
    Affiliations
    Managerial Epidemiology Research Center, Department of Public Health, School of Nursing and Midwifery, Maragheh University of Medical Sciences, Maragheh, Iran
    Department of Epidemiology & Biostatistics, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran
    Search for articles by this author
  • Mark J.M. Sullman
    Affiliations
    Middle East Technical University, Northern Cyprus Campus, Güzelyurt/Morphou, Northern Cyprus
    Search for articles by this author
Published:March 16, 2018DOI:https://doi.org/10.1016/j.ajic.2018.01.022
      To the Editor:
      We read with interest the article recently published in the American Journal of Infection Control by Tang and colleagues.
      • Tang C.
      • Li Y.
      • Liu C.
      • Sun P.
      • Huang X.
      • Xia W.
      • et al.
      Epidemiology and risk factors for Clostridium difficile-associated diarrhea in adult inpatients in a university hospital in China.
      The authors investigated the risk factors for Clostridium difficile-associated diarrhea at a university hospital in eastern China.
      • Tang C.
      • Li Y.
      • Liu C.
      • Sun P.
      • Huang X.
      • Xia W.
      • et al.
      Epidemiology and risk factors for Clostridium difficile-associated diarrhea in adult inpatients in a university hospital in China.
      Although the study reported several interesting findings, several methodological issues must be considered.
      First, the article does not mention which method was used to select the predictors included in the multivariable analysis. As shown in Table 2,
      • Tang C.
      • Li Y.
      • Liu C.
      • Sun P.
      • Huang X.
      • Xia W.
      • et al.
      Epidemiology and risk factors for Clostridium difficile-associated diarrhea in adult inpatients in a university hospital in China.
      the authors included only length of hospital stay (≥6 days), comorbidity (e.g., diabetes), and treatment type (e.g., coloclysis and proton-pump inhibitor) in the multivariable analysis. This is puzzling, as their univariable analysis showed statistically significant (P < .05) associations for other predictors, such as the use of different types of antibiotics (e.g., cephalosporin and fluoroquinolones). Investigators normally conduct multivariable analyses using a stepwise method, selecting the variables to retain using statistical criteria, or they select predictors according to their clinical relevance.
      • Steyerberg E.W.
      Clinical prediction models: a practical approach to development, validation, and updating.
      However, the approach used by Tang and colleagues
      • Tang C.
      • Li Y.
      • Liu C.
      • Sun P.
      • Huang X.
      • Xia W.
      • et al.
      Epidemiology and risk factors for Clostridium difficile-associated diarrhea in adult inpatients in a university hospital in China.
      is not clear.
      Second, the differences in odds ratios between the univariable and multivariable models for some predictors (e.g. comorbidity with diabetes) were relatively small. Generally, it has been reported that the unadjusted exposure-outcome should be changed by a certain percentage (e.g., 10%) in the multivariable analysis. When this is not the case, it is likely to be caused by the degree of residual confounding.
      • Lee P.H.
      Is a cutoff of 10% appropriate for the change-in-estimate criterion of confounder identification?.

      Acknowledgment

      This work was not supported by any organization.

      References

        • Tang C.
        • Li Y.
        • Liu C.
        • Sun P.
        • Huang X.
        • Xia W.
        • et al.
        Epidemiology and risk factors for Clostridium difficile-associated diarrhea in adult inpatients in a university hospital in China.
        Am J Infect Control. 2018; 46: 285-290
        • Steyerberg E.W.
        Clinical prediction models: a practical approach to development, validation, and updating.
        Springer Science & Business Media, 2008
        • Lee P.H.
        Is a cutoff of 10% appropriate for the change-in-estimate criterion of confounder identification?.
        J Epidemiol. 2014; 24: 161-167

      Linked Article