Healthcare-Associated Infections Community Interface

Healthcare-associated infections (HAI) are caused by a wide variety of common and unusual bacteria, fungi, viruses, and toxins encountered during the course of receiving medical care. These infectious agents can come from the patient themselves or from non-patient sources such as healthcare personnel, visitors, patient care equipment, medical devices, or the healthcare environment. HAI account for an estimated 1.7 million infections and 99,000 associated deaths each year._[1]

The direct medical cost associated with these infections was estimated, in 2007 dollars, to be $25,903 per HAI._[2]

The national EIP has distinct projects to conduct active population-based surveillance for HAI as well as evaluations of HAI data. NMEIP is currently contributing to two main Healthcare-Associated Infections-Community Interface (HAIC) projects.

Special Projects

HAI and Antimicrobial Use Prevalence Survey

Background

While HAI surveillance is often conducted for specific infections at both the state and national levels, there is limited data on overall prevalence of HAI and antimicrobial use in the US.

The objectives of the HAI and Antimicrobial Use Prevalence Survey are to:

1. Estimate HAI prevalence among inpatients in acute healthcare facilities. (i.e., hospitals)
2. Determine the distribution of HAI by pathogen and major infection site.
3. Estimate prevalence and describe the rationale for antimicrobial use in acute healthcare facilities.

Phase 2

New Mexico participated in phase 2 of this survey in 2010. During this phase, one acute healthcare facility was selected to assess HAI and antimicrobial use among a random sample of bed/patient combinations. Medical records were reviewed by EIP staff to ascertain antimicrobial use and determine if HAI were present.

To protect anonymity of facilities participating in phase 2 of this survey, data from all ten EIP sites have been aggregated by CDC. Healthcare-associated infections were found in about 6% of patients and were disproportionately found in patients in intensive care unit locations. Pneumonia and surgical site infections comprised about half of all HAI found. Antimicrobials were administered to about half of the surveyed patients, with treatment of active infection the most commonly cited reason for their use.

Phase 3

For the third phase of the project, each EIP state randomly selected and recruited facilities to participate in a single day survey during May – August 2011. Twenty facilities in New Mexico participated in this survey. Eligible patients for the survey included acute care inpatients of any age in randomly selected beds on the morning of the survey. Facilities randomly sampled up to 100 patient records, depending on the size of the facility. For those patients determined to be on antimicrobials on the survey date or day prior to the survey date, detailed medical record reviews were conducted to determine the rationale for antimicrobial use and if there was any evidence of HAI based on National Healthcare Safety Network (NHSN) definitions.

Phase 3 – New Mexico Results

In New Mexico, a total of 892 patient records were surveyed across 20 acute healthcare facilities. (See Table 1)

Table 1: Phase 3 Demographic Characteristics & Medical Device Use Among 892 Patient Records Surveyed

Among the patient records surveyed, 433 patients (48.5%) received a total of 768 antimicrobials on the survey day or day prior to the survey day. (See Table 2)

Table 2: Phase 3 Antimicrobial Use, Rationale for Antimicrobial Use, & Therapeutic Sites for Active Infection

Among the patients surveyed, the New Mexico EIP HAI staff determined that 25 HAIs were present in 22 patients (2.5%, 95% CI 1.6-3.7). The most common HAIs were urinary tract infection (28.0%) and pneumonia (24.0%). Among the HAIs attributable to a facility location, 43.5% were attributable to critical care locations and 56.5% were attributable to ward locations. CDC is in the process of performing aggregated data analysis across acute healthcare facilities in all 10 EIP sites to generate prevalence estimates of HAI and antimicrobial use.

The information from the Phase 3 HAI and Antimicrobial Use Prevalence Survey provides a snapshot of the prevalence of all HAIs across participating facilities in New Mexico, which can supplement the smaller subset of HAIs being tracked at the state and national levels through NHSN. In addition, this survey methodology can be used to follow trends in HAI and antimicrobial use patterns and data can be used to inform policy decisions and highlight targets for prevention efforts.

National Healthcare Safety Network (NHSN) Denominator Simplification Project

Collection of denominator data is an essential part of HAI surveillance in order to calculate infection rates. However, denominator data collection is often time intensive and can consume human resources that could be utilized for other infection prevention efforts.

The objectives of the NHSN Denominator Simplification Project are to:

1. Evaluate use of a simple, less labor intensive method for estimating device-days (i.e. use of central lines) denominator data.
2. Determine if denominator sampling methods can be successfully implemented.
3. Determine if denominator sampling methods can generate valid estimates of device-days.

The expected benefits of this project are to validate sampling methodology which could then contribute to reduction in HAI surveillance data collection burden, increased HAI surveillance participation, and improve denominator data reliability and accuracy.

This project has two phases:

1. Retrospective data collection.
2. Prospective data collection.

The retrospective phase involved facility infection preventionists providing 6-12 consecutive months of monthly device-days (i.e., central line) denominator worksheets. These data are being used by CDC to retrospectively assess the feasibility of using sampling methods to obtain estimates of central line-days in both Intensive Care Unit (ICU) and non-ICU settings. The data collection for the retrospective phase was completed in the fall of 2010. Five facilities in New Mexico participated in this phase.

The prospective phase involved facility staff continuing to collect denominator data using their usual methods during January-December 2011 while also conducting independent sampling (one day per week) of denominator data from the same inpatient units. In New Mexico, eight facilities collected data in 11 units for the prospective phase. This phase will allow for assessment of the feasibility of implementing the use of sampling methods to collect patient-days and central line-days denominator data. In addition, this phase will provide data to compare estimates of central line-days derived from data collected using sampling methodology against data collected using current denominator data collection practices. This comparison will determine if estimates of central line-days and central line-associated bloodstream infection (CLABSI) rates generated are suitable for the purposes of conducting HAI surveillance and submitting data to NHSN. Data from New Mexico have been combined with data from other states for analysis by CDC, currently in progress.

Clostridium difficile Infection (CDI) Project

The second main HAIC project that NMEIP participates in aims to ascertain the scope of CDI through active, population-based surveillance. This project is being conducted in Bernalillo County with the specific aim to determine the incidence of CDI, the proportion of infections that are community- and healthcare-associated, and which Clostridium difficile strains are causing disease. For more information on this project, see our Clostridium difficile page.

New Mexico Phase 1 & 2 Central Line-Associated Bloodstream Infection (CLABSI) Data Validation Project

Background and Objectives

Data validation is a key element to assure quality, accuracy, and reliability of public reports. The objective of the Phase 1 and 2 CLABSI data validation project was to determine if adult Intensive Care Unit (ICU) CLABSI events in New Mexico were properly ascertained, defined, and reported to the National Healthcare Safety Network (NHSN), a secure web-based electronic repository for HAI data.

Phase 1 Methods and Findings

During the Phase 1 CLABSI data validation project, NMDOH conducted retrospective medical record reviews on individuals determined to have had both a positive blood culture result and an ICU stay during November 2009 – March 2010 at six New Mexico hospitals. Medical record reviews were conducted independently by two reviewers using a standardized data collection instrument. Reviewers were blinded both to each other’s findings and to NHSN data. ICU CLABSI event determinations were compared between NMDOH reviewers and with NHSN data. Discordant HAI case determinations were reviewed by NMDOH medical epidemiologists in collaboration with hospital Infection Preventionists (IP) to reach final determinations. All definitions used for determining the presence of an HAI were based on the CDC NHSN Patient Safety Component Manual Identifying HAIs.

Initially, 118 individuals were identified for medical record review based on matching the ICU and laboratory line-lists provided by the six hospitals. A search of data submitted to NHSN revealed five additional ICU CLABSI cases that had not been identified for medical record review due to lack of inclusion on either the ICU or laboratory line-lists.

A final total of 123 medical records were reviewed. At the conclusion of the review process, final determinations revealed that four additional ICU CLABSI events should have been entered into NHSN. These four cases were not originally entered into NHSN due to a difference in interpretation of NHSN definitions (two cases) and due to hospital systems issues that resulted in facility IP not having been alerted to the positive blood cultures (two cases). Events have been updated in NHSN by pertinent facilities to reflect these additional cases. Comparison of original NHSN status against the final case determinations for all 123 individuals resulted in a sensitivity of 66.7%, specificity of 100%, positive predictive value of 100%, and negative predictive value of 96.5% for ICU CLABSI surveillance across the six participating hospitals.

Phase 2 Methods and Findings

Blinded retrospective medical record reviews were conducted with twice as many hospitals using similar methodologies as Phase 1. Retrospective medical record reviews were conducted for individuals with positive blood cultures who were being cared for in adult ICUs of 12 New Mexico hospitals during May 2011-July 2011.

Matching of laboratory and ICU line-lists provided by participating hospitals identified 109 individuals eligible for medical record review. Search of the NHSN data revealed six additional ICU CLABSI events that had been entered by the 12 hospitals for events occurring during May 2011-July 2011. These six cases had not been identified for review by NMDOH due to lack of inclusion on the laboratory or ICU admission line-lists.

A final total of 115 potential cases had medical record reviews. The six cases originally in NHSN were determined to be true CLABSI events. Four additional CLABSI events that should have been entered into NHSN were identified. These four cases were not originally entered due to a difference in interpretation of NHSN definitions (one case) and due to hospital systems issues that resulted in facility IP not being alerted to the positive blood cultures (two cases). For one case, the reason for not completing the case review was unclear. Events have been updated in NHSN by pertinent facilities to reflect these additional cases. These findings resulted in a sensitivity of 60.0%, specificity of 100.0%, positive predictive value of 100.0%, and negative predictive value of 96.3% for CLABSI surveillance across the 12 facilities.

Limitations

One limitation of this project was the small number of CLABSI events identified due to the relatively short time period of evaluation (i.e. five months in Phase 1 and three months in Phase 2). Additionally, this project was conducted voluntarily with hospitals, and these facilities might not be representative of the entire state with regard to experience with CLABSI prevention, surveillance, and reporting.

Conclusions

This project validated adult ICU CLABSI data submitted to NHSN. The overall sensitivity, specificity, and positive and negative predictive values for adult ICU CLABSI surveillance were similar in Phase 1 and Phase 2 despite differences in the number of participating hospitals and timeframes validated. The reasons for discordant CLABSI event determinations between NMDOH reviewers and hospital IP staff were similar in both phases; hospital systems issues and differences in interpretation or application of NHSN definitions were the most common sources of discordance.

CLABSI surveillance is complex, and this validation project provided an opportunity to collaboratively assess and provide education on NHSN HAI definitions, surveillance processes, and use of NHSN, while improving NHSN data accuracy. In addition, the project provided the opportunity to build and strengthen relationships between New Mexico EIP HAIC and hospital personnel, including IP. There is need for ongoing quality improvement processes and validation to ensure the quality, accuracy, and reliability of NHSN data, which is used to inform prevention efforts. The lessons learned from this project will be used to inform future validation and education efforts.

References

1. National Action Plan to Prevent Healthcare-Associated Infections: Road Map to Elimination
   Tier 1 Action Plan (revised). Accessed on May 14, 2012.
2. Direct Medical Costs of Healthcare-Associated Infections in US Hospitals and the Benefits of Prevention
   R. Douglas Scott II, CDC, March 2009. Accessed on May 14, 2012.