1375. Pre-admission Predictors of Clostridium difficile Infection (CDI) in Residents Receiving Long-term Care at Veterans Affairs Facilities
Session: Poster Abstract Session: Clostridium difficile
Saturday, October 5, 2013
Room: The Moscone Center: Poster Hall C
Background:  C. difficile carriage is predicted by prior antimicrobial use and recent hospitalization.  We examined whether these factors were associated with C. difficile infection (CDI) on admission and CDI during follow-up in VA Long Term Care Facility (LTCF) residents.    

Methods:  A nationwide retrospective cohort of admissions to 127 VA LTCFs with average daily census of >20 residents from Jan 2006-Dec 2011 was identified.  CDI was operationalized as a positive clinical test for C. difficile toxin. Independent variables including age as well as recent history (within 60 days prior to admission) of CDI, antibiotic use and VA inpatient acute care exposure were used to predict outcomes of CDI on admission (toxin positive < 3 days from admission) or during follow-up (toxin positive > 3 days from admission). Data were analyzed using logistic regression in SAS. The sensitivity and PPV of these predictors for CDI on admission or during follow-up were also determined.

Results:  The cohort consisted of 272,623 admissions with 17,761,299 resident days over six years. Median age in years [IQR] was 71 [61, 82]. 2.8% had history of recent CDI, and 53% were recently exposed to antibiotics and 65% had received care on a VA acute care location. Odds ratios and prediction model c-statistics for CDI are presented (see Table). A history of CDI within 60 days had a sensitivity and PPV for CDI on admission of 0.224 and 0.024 and for CDI during follow-up of 0.209 and 0.189.



CDI on Admission


CDI During Follow-up

Predictor Variables


95% CI




95% CI


Recent CDI








Recent Antibiotics








Recent VA Acute Care

















816 CDI in 272,623 admissions

C-statistic 0.6870


6,651 CDI in 271,807 admissions

C-statistic 0.7000

Conclusion:  Factors previously demonstrated to be associated with C. difficile carriage on admission are also predictive of CDI on admission and CDI during follow-up in LTCF residents. These predictors may be useful for implementation of control strategies in LTCFs that depend on early or even pre-emptive identification of subsets of high risk patients.

Jeanmarie Mayer, MD1,2, Makoto Jones, MD, MS3, Susan M. Pacheco, MD4, Dale N. Gerding, MD5, Curtis J. Donskey, MD6, Molly Leecaster, PhD7, Matthew Maw, MS1, Kevin Nechodom, BS1 and Matthew Samore, MD1, (1)University of Utah School of Medicine, Salt Lake City, UT, (2)IDEAS Center, VA Salt Lake City Health Care System, Salt Lake City, UT, (3)Internal Medicine, University of Utah School of Medicine Division of Epidemiology, Salt Lake City, UT, (4)Infectious Diseases, Northwestern University, Chicago, IL, (5)Edward Hines Jr. VA Hospital, Hines, IL, (6)Case Western Reserve University, Cleveland, OH, (7)Salt Lake City VA Health Care System, Salt Lake City, UT


J. Mayer, None

M. Jones, None

S. M. Pacheco, CDC: Grant Investigator, Research grant

D. N. Gerding, None

C. J. Donskey, None

M. Leecaster, None

M. Maw, None

K. Nechodom, None

M. Samore, None

Findings in the abstracts are embargoed until 12:01 a.m. PST, Oct. 2nd with the exception of research findings presented at the IDWeek press conferences.