National and State Healthcare-associated Infections Standardized Infection Ratio Report

Using Data Reported to the National Healthcare Safety Network

January – December 2010

Discussion

The HAI data summarized in this report demonstrate healthcare facilities reporting to NHSN during 2010, as a group, reported fewer CLABSIs (32%), CAUTIs (6%), and SSIs (8%) than predicted based on the case-mix of patients and locations that were monitored. Moreover, the CLABSI prevention success improved between reporting periods, as the SIR during 2010 was significantly decreased (SIR 0.684, 32% reduction in CLABSI) compared to 2009 (SIR 0.854, 15% reduction in CLABSI). This suggests that the facilities reporting during both years not only sustained the prevention success of 2009, but improved even more in 2010. Such improvement was more modest for SSIs: the overall SSI SIR decreased from 0.981 to 0.916 when including all reporting facilities, but the decrease lost statistical significance when limiting the comparison to only facilities reporting in both years. This may be a combination of a loss of power (only 904 facilities reported continuously in both years) and lack of substantial progress across all surgery types included. Regardless, there was sustained prevention success, with SIRs significantly lower than 1.0 in both 2009 and 2010. Interestingly, improved prevention success was observed among facilities reporting specifically on coronary artery bypass graft procedures (CABG), with a statistically significantly lower SIR in 2010 (0.820, 18% reduction) compared to that in 2009 (0.962, 4% reduction). While this demonstrates substantial and improved success in prevention of SSIs related to this procedure, there is opportunity for substantial SSI prevention in other procedures.

Another perspective on the potential improvements that could occur can be made by the evaluation of the facility-specific SIRs reported for each type of HAI, and in each of the major patient-location groups. In most cases, the calculated SIR for the highest 90th percentile of facilities reporting was >1.5, translating to over 50% more HAIs than would have been predicted based on the case-mix. If these worst performing facilities reduced their SIRs to about 1.0, great progress will be realized nationally.

Overall during 2010, 2,403 facilities reported CLABSI data to NHSN, an increase of about 800 facilities compared to the previous year. This increase may be the result of federal funding to support state-based HAI detection and prevention programs in the latter half of 2009. In addition, this increase also can be attributed to the advent of new CMS reporting requirements for hospitals participating in the CMS Hospital Inpatient Quality Reporting Program, which requires participating facilities to report to CMS through NHSN starting in 2011.⁵ Summary data reported through NHSN to CMS as part of this program will be posted quarterly beginning early 2012. Because these data may be a subset of all data reported to NHSN (e.g., some facilities report to NHSN but do not participate in the CMS Reporting Program), the summary statistics are expected to vary slightly.

Regarding CLABSI prevention success regionally, almost half of the states reported CLABSI SIRs in 2010 significantly <1.0, confirming that the national progress has not been limited to select geographic areas. Furthermore, most of these states reported accelerated prevention success in 2010 compared to 2009.

Roughly 400 more facilities reported SSI data nationally in 2010 compared to 2009, when 946 reported SSI data. This is a reversal of the trend reported in 2009.⁴ Much of this increase is due to state-specific mandates and facilities beginning to comply with CMS’s Hospital Inpatient Quality Reporting Program, which required facilities to report SSI data through NHSN starting in January 2012.¹¹ Although the number of facilities reporting increased, the proportion of total months of data that could have been submitted decreased slightly in 2010. This again may be due to facilities just entering into the system and reporting in the latter part of the calendar year.

A major consideration for interpretation of these data and for future reports is assessing the confidence in the validity of the data reported. First, specific validation efforts have been focused at the state level, and there is a need for more widespread validation of HAI data reported to NHSN. In this report, completion of validation studies of CLABSI data was reported from 16 states during 2009, and 21 in 2010; evaluations included data quality assessment of missing or implausible values and/or detection of outlier facilities (e.g., number of infections, rates, denominators) in all 21 states, and an audit of medical records in 16. Information on validation efforts was requested from all states, regardless of presence of a legislative mandate for the particular HAI type. Some states without mandatory reporting of a given HAI have performed validation on NHSN data that are voluntarily shared with them by facilities. Validation efforts by state health departments represent an important step toward a more complete understanding of the HAI data reported to NHSN. In previous SIR reports including state-specific data4, validation activities including a medical record audit (YESa in Table 1) by an external authority were anecdotally noted to be associated with higher SIRs. This phenomena is less apparent in this report.

Regardless of the success of validation efforts, inherent variability in case findings of HAIs will occur between facilities, explaining some of the differences in observed infection rates and facility-specific SIRs. Several efforts are in place to improve the accuracy and confidence in these HAI data. These include the availability of web-based NHSN surveillance training modules (http://www.cdc.gov/nhsn/training.html), including webinars, slide sets, and new, self-paced, interactive, online training courses with continuing education credits available upon successful completion of an assessment; the provision of NHSN training during CDC-hosted events and at professional meetings and conferences; continued improvements to the NHSN system including software changes such as business rules and cross-field edit checks to prevent data entry errors, system alerts to inform users of missing data, and the availability of data quality reports to inform users of aberrant data. In addition, CDC is exploring changes in methodology to minimize unreliable application of the standard definitions and data collection protocols. Finally, CDC is developing guidance and tools for efficient validation work to be implemented by states as resources become available.

The SIRs summarize complex data related to HAIs in a single set of indicators that use national data for a specified time period as a common referent. The indirect standardization technique used to calculate SIRs is the same as for standardized mortality ratios (SMRs), a commonly used method in epidemiology for comparing mortality between a group and a referent population.¹² This summary measure should not be used to derive any absolute ranking of facilities or regions, but rather as a tool to identify facilities or regions that may deserve targeted evaluations, which may include validation efforts or assessing potential prevention programs.

When interpreting data in this report, it is important to understand the extent to which SIRs are risk adjusted. For device-associated infections, the risk stratification is mostly by the location of the patient, often split into different strata further by status as a teaching facility, and several times further split by number of beds in the location.¹ Additional data, such as hospital-level case-mix indices, or patient-specific device use data, may result in improved risk adjustment and are being explored for incorporation into future evaluations. For SSIs, risk stratification includes procedure-and patient-specific factors.² Secondly, despite efforts through validation and training, infection prevention staff often interprets or implements surveillance methods differently. To minimize the variability in application of standardized methods, changes in NHSN methods are planned for 2013 and beyond. As these planned changes are finalized and implemented, their impact on our ability to report consistently over time will be an ongoing challenge.

Conclusion

This report presents a set of national summary statistics for CLABSIs, CAUTIs, and SSIs for 2010, including serial SIRs for CLABSI and SSI for 2009-2010. As a single summary measure of prevention success, there has been a large reduction (32%) in CLABSIs among reporting hospitals compared to predictions and more modest reductions for CAUTI and SSI. Prevention success improved in 2010 compared to the 2009 level of success for CLABSI. For SSI, improved prevention success over the two years was documented most significantly for coronary artery bypass graft operations (in 2010, 18% SSIs prevented), while stable reductions in SSIs were evident for two of the nine other operative procedures evaluated (knee arthroplasty, 11% reduction; colon surgery, 9% reduction). Overall, there appears to be great room for improvement across the variety of operative procedures. Serial comparisons of CLABSI at the state-level provide an improved means for monitoring the impact of interventions and indicate the successes of state-based and national HAI reduction efforts. Ongoing interactions with state health departments will be critical to determine ways to improve the reporting of HAIs and to act on these data to prevent HAIs. Although comparative data on HAIs (e.g., comparing the local facility to the referent group) are available to each participating facility at all times through the NHSN system, facility-specific SIRs have been used by an increasing number of state departments to present annual HAI summary.^13-20 CDC will continue to report SIRs at the national level as a measure of progress toward the HHS HAI Action Plan targets and to gauge the impact of federal support to the states for HAI prevention. However, first and foremost, these summary data add to a comprehensive body of data related to HAI occurrence for analysis and action at the local, state, and national levels.