Longitudinal Trends in Infection Rates in US Nursing Homes, 2006 - 2011
Background: Infections are a leading cause of morbidity and mortality in US nursing home (NH) residents with an estimated 1.6 – 3.8 million occurring annually. Minimum Data Set (MDS) assessments are performed on NH residents every quarter and provide data that can be used to track infections. A revised MDS (version 3.0) was implemented in 2010. This study aimed to estimate longitudinal trends in infection rates in US NH using MDS 2.0 data, describe infection item differences in the MDS 2.0 and 3.0, and evaluate impacts of MDS revisions on infection measurement.
Methods: MDS 2.0 (2006 – 2010) and 3.0 (2010 - 2011) annual and quarterly assessment data were used to estimate infection rates by quarter; MDS 2.0 data were used to estimate percent changes in infection rates between 2006 and 2010. Infection items on both MDS versions included multidrug-resistant organisms (MDRO), pneumonia, septicemia, tuberculosis (TB), urinary tract infection (UTI), viral hepatitis, and wound infection. Items were compared to assess differences.
Results: MDS data from 24 quarters and over 14,000 NH (n = 25,903,977 assessments) were used. With the exception of TB, infection rates in the MDS 2.0 changed with increases in viral hepatitis (69.7%), septicemia (25.2%), pneumonia (24.1%), MDRO (15.7%), and wound infections (4.6%); UTI rates decreased by 4.2% (all p-values <0.001). Changes in all infection items were noted between versions 2.0 and 3.0 and impacted all rates except for MDRO and TB (Figure). Items had two look-back periods in the MDS 3.0 (a 60-day disease identification period and a shorter diagnosis status period) but only one in the MDS 2.0 (the diagnosis status period). Also, compared with criteria used to identify UTI in the MDS 2.0, MDS 3.0 criteria included more specific signs or symptoms as well as evidence that the resident was receiving treatment.
Conclusion: Substantively important growth in infection rates in NH were identified for most items evaluated. Further research is needed to understand these increases as well as best practices for infection prevention. Additionally, MDS revisions should be accounted for when evaluating longitudinal trends in NH infections over this time period.
Figure. Longitudinal trends in quarterly infection rates among residents in US NH from 2006 through 2011. The gap represents implementation of the MDS 3.0.
M. Sorbero, None
M. Pogorzelska-Maziarz, None
C. C. Cohen, None
P. Stone, None