A retrospective analysis of clinical characteristics, hospitalization, and functional outcomes in residents with and without Clostridium difficile infection in US long-term care facilities

作者：Friedman Howard S^*; Navaratnam Prakash; Reardon Gregory; High Kevin P; Strauss Marcie E

来源：Current Medical Research and Opinion, 2014, 30(6): 1121-1130.

DOI：10.1185/03007995.2014.895311

• Clostridium difficile infection
• Infectious diarrhea Long-term care
• Outcomes
• Retrospective analysis
• Risk factors
• Survival analysis

Objective: Patients in long-term care (LTC) are at increased risk for acquiring Clostridium difficile infection (CDI). We compared the characteristics and outcomes of matched cohorts with and without CDI in the LTC setting.
Methods: Using a retrospective cohort design, demographic characteristics, Minimum Data Set (MDS 2.0) assessments (years 2007-2010), and pharmacy records of residents were analyzed. Residents were required to have a CDI diagnosis, >= 1 MDS 2.0 assessment <= 120 days pre- and post-index event, and receipt of metronidazole (MET) or vancomycin (VAN) within +/- 7 days of index date. Baseline characteristics were compared between cases and controls matched 1: 3 on age, gender, and index year. Cox regression (CR) analysis evaluated the relationship between CDI status, and post-index mortality and hospitalization.
Results: A total of 1145 CDI residents were matched with 3488 non-CDI residents. A second sample used propensity score methods. CDI vs. non-CDI residents had a higher baseline comorbidity burden (Charlson score: 3.0 +/- 1.9 vs. 2.2 +/- 1.8, respectively), and were more likely to have had a recent hospitalization (63% vs. 9%, respectively) and shorter mean pre-index continuous length of stay (cLOS) in the LTC (386.4 d +/- 536.3 d vs. 568.3 d +/- 567.4 d, respectively), all P<0.0001. CR analyses of both samples indicated that CDI was strongly associated with shorter times to hospitalization and mortality (hazard ratio (HR) = 1.3, P = 0.023 and 2.2, P<0.0001, respectively; propensity-matched group). Pre-index LTC cLOS also remained an important variable in the CR analysis and was the strongest predictor of post-index hospitalization and mortality (HR = 0.999 and 0.996, respectively, P<0.0001), indicating that residents with longer pre-index LTC cLOS had longer times to post-index hospitalization and mortality. Our reliance on the MDS records for case identification was our chief limitation; misclassification was mitigated by our requirement to include CDI treatment as part of our inclusion criteria.
Conclusions: Understanding factors that put LTC patients at risk for CDI can help guide better management and improvement of patient outcomes.

• 出版日期2014-6