NOTICES
Health Care Associated Infection Benchmarking Areas for Hospitals under the Medical Care Availability and Reduction of Error (MCARE) Act; Final Notice
[42 Pa.B. 2543]
[Saturday, May 12, 2012]Section 408(9) of the Medical Care Availability and Reduction of Error (MCARE) Act (MCARE Act) (40 P. S. § 1303.408(9)) requires that the Department of Health (Department) publish a notice in the Pennsylvania Bulletin of the specific benchmarks the Department will use under section 408(8) to measure the progress health care facilities are making to reduce the incidence of healthcare associated infections (HAI). Before publishing a final notice on the benchmarks, the Department is to seek public comment and respond to any comments received during the public comment period.
The public notice was published at 42 Pa.B. 273 (January 14, 2012). The Department proposed to use catheter-associated urinary tract infections (CAUTI), central line-associated bloodstream infections and selected surgical site infections (SSI) as benchmarking conditions. For these HAI infection types, the Department proposed to define hospitals that are not meeting reduction targets as those that, for each of the two most recent years of analyzed data, had both standardized infection ratios (SIR) and infection rates that were above the 90th percentile when compared to other Commonwealth hospitals.
During the 30-day comment period, the Department received a total of four submissions. These comments were generally supportive of the approach proposed by the Department but raised concerns about the inherent difficulty of assessing performance across the range of hospitals found in this Commonwealth. Specific comments are summarized as follows along with the Department's response. Some respondents raised multiple concerns. Therefore the total number of items exceeds the number of submissions. The number of respondents who raised each concern is included.
1. Two respondents indicated the data used for analysis should be as accurate as possible. They felt that until the data used for analysis had been validated through auditing, it is possible that conclusions based on the current data set will be inaccurate and potentially biased.
The Department concurs that audits of data quality are important to assure the accuracy of reporting by Commonwealth hospitals and the validity of analyses of the data performed by the Department. The Commonwealth uses a number of auditing methods to accomplish these goals. All data submitted to the National Healthcare Safety Network (NHSN) by Commonwealth hospitals are assessed for potential errors and hospitals receive a quarterly data integrity validation (DIV) report and are asked to investigate and (if necessary) correct any errors. Since the DIV report has been in operation as a feedback mechanism, the number of identified errors has declined substantially. In addition, Department representatives routinely visit hospitals to determine whether HAI reporting requirements and data collection methods are being followed. Finally, the Department has conducted onsite chart audits of both 2009 and 2010 HAI data in a selected sample of hospitals. Findings of these audits were provided to the audited hospitals and seminars were conducted for all hospitals to review the overall results to identify common issues and improve reporting accuracy. The Department agrees that more extensive onsite chart audits would be beneficial. However, conducting onsite audits of all Commonwealth hospitals before the data are analyzed would result in substantial delays in producing any reports. In addition, the findings of the chart audits performed to date do not suggest that error rates are high enough (either over- or under-reporting) to invalidate the data analyses published to date by the Department. Therefore, the Department does not believe the published benchmarking methods require modification based on this comment.
2. Two respondents noted there are important differences between different types and groups of hospitals that cannot be easily accounted for through the use of an overall SIR. These respondents felt it was not appropriate to aggregate all Commonwealth hospitals to generate SIRs for the proposed ranking as described in the Notice.
The Department concurs that there are differences between hospitals (individually and by group). Each hospital differs by location, by size, by population served and by the level and type of care. The Department accounts for these differences through indirect adjustment of the predicted number of infections (which in turn determines the SIR) based on the ward types present in each hospital, by hospital size, medical school affiliation and hospital geographic location. These are factors commonly recognized as possibly confounding HAI rates. For SSIs, direct adjustment using patient-level data is performed. However, there are no adjustment methods that can perfectly account for differences between hospitals. Sub-stratification of hospitals for the purposes of generating SIRs based on hospital type or other factors would be problematic because the number of facilities in some categories would be small and would make generation of meaningful SIRs nearly impossible. The Department believes the use of an overall SIR for benchmarking purposes is a reasonable approach and therefore does not believe the proposed analytic method requires modification based on this comment.
3. Two respondents suggested that additional strategies are needed to standardize data from hospitals that provide care to children, as the risk factors for infection in children are different from the risk factors in adults. The respondents felt the adjustments used by the Department to generate SIRs are insufficient to account for these risk factors.
The Department concurs that risk factors for infection in children may differ from those in adults. This is why the Department calculates Statewide ward-specific rates (including rates for pediatric ward types) and incorporates these rates into its risk adjustment calculations by adjusting for the ward types present in each hospital. This Commonwealth has a small number of exclusively pediatric hospitals, but the number of general hospitals that have pediatric wards (including neonatal intensive care units) is much larger. Therefore it would be inappropriate to separately display or separately calculate SIRs for pediatric-only facilities with the knowledge that a large proportion of pediatric care occurs in other hospital types. The Department welcomes feedback that might improve current risk adjustment methods. But this feedback would not produce any alteration in the proposed approach to identifying hospitals not meeting benchmark reduction targets. It would only alter which hospitals would be identified by the proposed method. Therefore no changes need to be made in the methods based on this comment.
4. One respondent indicated that the approach to identifying hospitals not making progress towards the reduction of HAIs should be amended, as there may be organizations identified as not making progress that have actually decreased their rates of infection, but the decreases are not as large as those that have occurred in other organizations. In addition, it may be difficult to identify reductions for organizations that have successfully reduced utilization of devices such as urinary catheters and central lines as an HAI prevention method.
The Department concurs that a hospital's annual SIR may paradoxically increase if the rates of HAI reductions occurring in most other hospitals are greater than the ones that have occurred at the hospital in question. This is because the predicted number of infections at each hospital is based on overall Statewide rates. This is the reason the Department has chosen to assess both the SIR and the actual HAI rate for each hospital, and to only cite hospitals that have both SIRs and rates that are above the 90th percentile. This approach identifies hospitals not performing as well as its peers, and does not penalize hospitals with very low HAI rates even though the rate may slightly increase from one year to the next. The Department considers its approach to be a broad one that fulfills the obligation to identify less well performing hospitals without citing hospitals that are sufficiently addressing HAI prevention.
5. One respondent suggested that organizations may be repeatedly identified as not meeting benchmark reduction targets.
The respondent is correct that the same hospital may be repeatedly identified as not meeting benchmark reduction targets. If certain hospitals are not making sufficient progress in comparison to other hospitals or have sustained high SIRs and rates, the Department's methods will continue to identify these institutions. Hopefully, this will allow the hospital to identify approaches that will reduce their HAI incidence in future years. The Department does not believe that citing a facility for one 2-year period should remove it from being cited again if the rates or SIRs remain high.
6. One respondent noted that the methodology appears to be biased against larger organizations that care for more complex patients.
The Department has performed and published analyses of HAI data for the last half of 2008, for 2009 and for 2010. The findings indicate that larger hospitals tend to have SIRs that are significantly higher or lower than predicted when compared to smaller hospitals. This is because the larger number of HAIs, device days and patient days allows more reliable estimates of actual rates to be calculated. This can be seen in the much narrower confidence intervals for larger hospitals than smaller hospitals. However, the proposed methodology for identifying hospitals with high SIRs or rates considers only the actual SIR and rate, not their significance. This approach would therefore not penalize larger hospitals, and no changes are needed based on this comment.
7. One respondent noted that the proposed approach does not use the most currently available information. Hospitals that are identified as not meeting benchmark targets may have actually made progress since the time period used for the calculations. The proposed methodology should be used as a screening tool only and a correction plan should only be required if there is no evidence of subsequent improvement with more recent data.
The Department produces an annual report on HAIs. This report is targeted for publication approximately 6 months after the close of the reporting period because data submission, DIV and data cleaning follow a 4-month cycle. In addition, implant-associated surgical procedures must be followed to see if an HAI develops for a full 12 months post-procedure. Therefore, SSI data has a 1-year lag period for reporting. The Department does not believe it is appropriate to use more recent (which would be partial year, provisional data) data to determine whether a hospital should be cited for the previous 2 years of high SIRs and rates. The goal of notifying the hospitals of this finding is to alert them to the problem and assist them in identifying and correcting the conditions that result in HAIs. If the hospital has been following the Department's published reports, it may already have noted its performance relative to other facilities and taken appropriate steps. In this situation, no further actions or interventions would be needed unless the interventions are unsuccessful. Since this comment does not suggest any changes in the proposed benchmark methods, the Department does not believe the proposed approach needs modification based on the comment.
Therefore, the final approach to benchmarking hospitals and identifying those not making sufficient progress in reducing HAIs is as follows:
The Department, under section 408(8) and (9) of the MCARE Act, publishes this notice regarding the methods to be used to measure progress of hospitals in reducing the occurrence of HAIs and identify hospitals not meeting benchmark reduction targets.
A. Purpose and Statutory Authority
Section 408(8) of the MCARE Act requires that the Department develop, in consultation with the Patient Safety Authority and the Pennsylvania Health Care Cost Containment Council, ''reasonable benchmarks to measure the progress [hospitals] make toward reducing health care-associated infections.'' The section further provides, ''Beginning in 2010, all health care facilities shall be measured against these benchmarks.''
Section 408(9) of the MCARE Act requires that the Department publish a notice in the Pennsylvania Bulletin of the specific benchmarks the Department will use under section 408(8) of the MCARE Act to measure health care facilities. Section 408(9) of the MCARE Act requires that prior to publishing the final notice, the Department is to seek public comment for at least 30 days on the benchmarks and respond to the comments received during the public comment period. The Department published a notice at 42 Pa.B. 273 announcing these benchmarks and requesting public comment. The Department's responses to these comments are included in the present notice.
B. Background
Since the HAI reporting provisions of Chapter 4 of the MCARE Act (40 P. S. §§ 1303.401—1303.411) took effect in February 2008, the Department has used the following list of HAIs for benchmarking purposes:
• Central Line Associated Blood Stream Infection (CLABSI)
• CAUTI
• SSI for:
o Coronary artery bypass graft with both chest and donor site incisions
o Coronary artery bypass graft with chest incision only
o Cardiac surgery
o Hip arthroplasty
o Knee arthroplasty
o Abdominal hysterectomy (HYST)
Data on these benchmark HAIs have been published annually. Currently, data from 2009 is considered to be the baseline year for trend analysis. As published at 41 Pa.B. 6454 (December 3, 2011), beginning January 1, 2012, the Department will also collect data regarding SSIs for colon surgeries for benchmarking purposes in the future.
Infections associated with surgeries that involve an implant may not develop or be detectable for some time following the surgical procedure. Accordingly, the United States Department of Health and Human Services, Centers for Disease Control and Prevention, NHSN, requires a full year of patient follow-up for complete identification and reporting of infections associated with procedures that involve an implant. Among the six surgical procedure types selected for benchmark SSI consideration, all but abdominal HYSTs may involve an implant. Consequently, annual data on SSIs are published in the year after data on CLABSIs and CAUTIs is published, that is, HAI data published in 2011 includes the CLABSI and CAUTI data for procedures completed in 2010 and the SSI data for procedures completed in 2009. At this time, the 2010 data for SSIs that is required for comparison to the 2009 baseline year is not available yet. However, the 2010 data needed to measure health care facilities against benchmarks are available for CLABSIs and CAUTIs.
C. Benchmark Methodology
The Department uses two metrics for calculating rates for CAUTIs and CLABSIs. The first metric is the incidence rate of infection. For CAUTIs, this is the number of infections per 1,000 urinary catheter days. For CLABSIs, this is the number of infections per 1,000 central line days. The second metric used by the Department is the SIR. The SIR consists of the number of infections observed (reported) by the health care facility divided by the number of infections predicted to be reported by the health care facility. The predicted number is a risk-adjusted calculation made by the Department based on Statewide rates of HAIs. The methodology for risk adjustment and calculation of the predicted number of infections can be found in the published annual HAI reports prepared by the Department and posted on its web site, http://www.health.state.pa.us. SIRs are produced for each hospital.
For the initial measurement of hospital progress in reducing HAIs, the Department proposes to rank all hospitals separately by their incidence rates of infection and SIRs for CAUTIs and CLABSIs, in two separate consecutive years, currently 2009 and 2010. Accordingly, hospitals will be ranked for:
CAUTI Benchmark Targets:
1) Incidence rate of infection for CAUTIs in year one
2) SIRs for CAUTIs in year one
3) Incidence rate of infection for CAUTIs in year two
4) SIRs for CAUTIs in year two
CLABSI Benchmark Targets:
1) Incidence rates of infection for CLABSIs in year one
2) SIRs for CLABSIs in year one
3) Incidence rates of infection for CLABSIs in year two
4) SIRs for CLABSIs in year two
The Department will then identify the hospitals that fall above the 90th percentile for all Commonwealth hospitals in each of the previously listed rankings. As an example, the 90th percentile for the 2010 CAUTI rate was 3.81 per 1,000 catheter days and the 90th percentile for the 2010 CAUTI SIR was 2.37. Any hospital identified as having both an incidence rate of CAUTIs and an SIR for CAUTIs above the 90th percentile for all Commonwealth hospitals in 2 consecutive years shall be considered not making progress towards the reduction of CAUTI HAI rates. Similarly, any hospital identified as having both an incidence rate of CLABSIs and an SIR for CLABSIs above the 90th percentile for all Commonwealth hospitals in 2 consecutive years shall be considered not making progress towards the reduction of CLABSI HAI rates.
When the 2010 data are available for SSIs, similar procedures will be used to identify the hospitals not making progress towards meeting the SSI benchmark targets for each individual procedure. In subsequent years, the Department will conduct a similar analysis for CAUTIs, CLABSIs and SSIs using the most currently available annual data for 2 consecutive years.
D. Affected Persons
All hospitals are currently required to comply with the HAI reporting requirements of the MCARE Act and will be measured for progress in meeting benchmark targets set forth. Section 103 of the MCARE Act (40 P. S. § 1303.103) defines a ''hospital'' as ''An entity licensed as a hospital under the act of June 13, 1967 (P. L. 31, No. 21), known as the Public Welfare Code, or the act of July 19, 1979 (P. L. 130, No. 48), known as the Health Care Facilities Act.''
For additional information, or for persons with a disability who require an alternative format of this notice (for example, large print, audiotape, Braille), contact the Office of Healthcare Associated Infection Prevention, 555 Walnut Street, 8th Floor Forum Place, Harrisburg, PA 17101, or for speech and/or hearing impaired persons V/TT (717) 783-6514, or the Pennsylvania AT&T Relay Service at (800) 654-5984.
ELI N. AVILA, MD, JD, MH, FCLM,
Secretary
[Pa.B. Doc. No. 12-856. Filed for public inspection May 11, 2012, 9:00 a.m.]
No part of the information on this site may be reproduced for profit or sold for profit.This material has been drawn directly from the official Pennsylvania Bulletin full text database. Due to the limitations of HTML or differences in display capabilities of different browsers, this version may differ slightly from the official printed version.
