Costs attributable to healthcare-acquired infection in hospitalized adults and a comparison of economic methods.

BACKGROUND: Hospitals will increasingly bear the costs for healthcare-acquired conditions such as infection. Our goals were to estimate the costs attributable to healthcare-acquired infection (HAI) and conduct a sensitivity analysis comparing analytic methods. METHODS: A random sample of high-risk adults hospitalized in the year 2000 was selected. Measurements included total and variable medical costs, length of stay (LOS), HAI site, APACHE III score, antimicrobial resistance, and mortality. Medical costs were measured from the hospital perspective. Analytic methods included ordinary least squares linear regression and median quantile regression, Winsorizing, propensity score case matching, attributable LOS multiplied by mean daily cost, semi-log transformation, and generalized linear modeling. Three-state proportional hazards modeling was also used for LOS estimation. Attributable mortality was estimated using logistic regression. RESULTS: Among 1253 patients, 159 (12.7%) developed HAI. Using different methods, attributable total costs ranged between $9310 to $21,013, variable costs were $1581 to $6824, LOS was 5.9 to 9.6 days, and attributable mortality was 6.1%. The semi-log transformation regression indicated that HAI doubles hospital cost. The totals for 159 patients were $1.48 to $3.34 million in medical cost and $5.27 million for premature death. Excess LOS totaled 844 to 1373 hospital days. CONCLUSIONS: Costs for HAI were considerable from hospital and societal perspectives. This suggests that HAI prevention expenditures would be balanced by savings in medical costs, lives saved and available hospital days that could be used by overcrowded hospitals to enhance available services. Our results obtained by applying different economic methods to a single detailed dataset may inform future cost analyses.

The impact of hospital practice on central venous catheter associated bloodstream infection rates at the patient and unit level: a multicenter study.

OBJECTIVE: Little is known about factors driving variation in bloodstream infection (BSI) rates between institutions. The objectives of this study are to (1) identify patient, process of care, and hospital factors that influence intensive care unit (ICU)-level BSI rates and (2) compare those factors to individual risk factors identified in a cohort analysis. DESIGN: In this multicenter prospective observational study, the authors measured the process of care for 2970 randomly sampled central venous catheter insertions over 13 months. SETTING: Medical, surgical, and medical/surgical ICUs of 37 domestic and 13 international hospitals. RESULTS: Significant correlates of unit-level BSI rates were percentage of female patients, patients on dialysis, ICU bed size, percentage of practitioners with low numbers of previous insertions, and percentage inserted by nurses. Patient-level analysis identified gender, age, posttransplant, postsurgery, and use of the line for parenteral nutrition. CONCLUSIONS: Factors that influence unit-to-unit variation may differ from factors identified in studies of individual patient risk.

Comparing bloodstream infection rates: the effect of indicator specifications in the evaluation of processes and indicators in infection control (EPIC) study.

OBJECTIVE: Bloodstream infection (BSI) rates are used as comparative clinical performance indicators; however, variations in definitions and data-collection approaches make it difficult to compare and interpret rates. To determine the extent to which variation in indicator specifications affected infection rates and hospital performance rankings, we compared absolute rates and relative rankings of hospitals across 5 BSI indicators. DESIGN: Multicenter observational study. BSI rate specifications varied by data source (clinical data, administrative data, or both), scope (hospital wide or intensive care unit specific), and inclusion/exclusion criteria. As appropriate, hospital-specific infection rates and rankings were calculated by processing data from each site according to 2-5 different specifications. SETTING: A total of 28 hospitals participating in the EPIC study. PARTICIPANTS: Hospitals submitted deidentified information about all patients with BSIs from January through September 1999. RESULTS: Median BSI rates for 2 indicators based on intensive care unit surveillance data ranged from 2.23 to 2.91 BSIs per 1000 central-line days. In contrast, median rates for indicators based on administrative data varied from 0.046 to 7.03 BSIs per 100 patients. Hospital-specific rates and rankings varied substantially as different specifications were applied; the rates of 8 of 10 hospitals were both greater than and less than the mean. Correlations of hospital rankings among indicator pairs were generally low (rs=0-0.45), except when both indicators were based on intensive care unit surveillance (rs = 0.83). CONCLUSIONS: Although BSI rates seem to be a logical indicator of clinical performance, the use of various indicator specifications can produce remarkably different judgments of absolute and relative performance for a given hospital. Recent national initiatives continue to mix methods for specifying BSI rates; this practice is likely to limit the usefulness of such information for comparing and improving performance.

A multicenter intervention to prevent catheter-associated bloodstream infections.

BACKGROUND: Education-based interventions can reduce the incidence of catheter-associated bloodstream infection. The generalizability of findings from single-center studies is limited. OBJECTIVE: To assess the effect of a multicenter intervention to prevent catheter-associated bloodstream infections. DESIGN: An observational study with a planned intervention. SETTING: Twelve intensive care units and 1 bone marrow transplantation unit at 6 academic medical centers. PATIENTS: Patients admitted during the study period. INTERVENTION: Updates of written policies, distribution of a 9-page self-study module with accompanying pretest and posttest, didactic lectures, and incorporation into practice of evidence-based guidelines regarding central venous catheter (CVC) insertion and care. MEASUREMENTS: Standard data collection tools and definitions were used to measure the process of care (ie, the proportion of nontunneled catheters inserted into the femoral vein and the condition of the CVC insertion site dressing for both tunneled and nontunneled catheters) and the incidence of catheter-associated bloodstream infection. RESULTS: Between the preintervention period and the postintervention period, the percentage of CVCs inserted into the femoral vein decreased from 12.9% to 9.4% (relative ratio, 0.73; 95% confidence interval [CI], 0.61-0.88); the total proportion of catheter insertion site dressings properly dated increased from 26.6% to 34.4% (relative ratio, 1.29; 95% CI, 1.17-1.42), and the overall rate of catheter-associated bloodstream infections decreased from 11.2 to 8.9 infections per 1,000 catheter-days (relative rate, 0.79; 95% CI, 0.67-0.93). The effect of the intervention varied among individual units. CONCLUSIONS: An education-based intervention that uses evidence-based practices can be successfully implemented in a diverse group of medical and surgical units and reduce catheter-associated bloodstream infection rates.

The use of economic modeling to determine the hospital costs associated with nosocomial infections.

Hospital-associated infection is well recognized as a patient safety concern requiring preventive interventions. However, hospitals are closely monitoring expenditures and need accurate estimates of potential cost savings from such prevention programs. We used a retrospective cohort design and economic modeling to determine the excess cost from the hospital perspective for hospital-associated infection in a random sample of adult medical patients. Study patients were classified as being not infected (n=139), having suspected infection (n=8), or having confirmed infection (n=17). Severity of illness and intensive unit care use were both independently associated with increased cost. After controlling for these confounding effects, we found an excess cost of $6767 for suspected infection and $15,275 for confirmed hospital-acquired infection. The economic model explained 56% of the total variability in cost among patients. Hospitals can use these data when evaluating potential cost savings from effective infection-control measures.

Impact of ring wearing on hand contamination and comparison of hand hygiene agents in a hospital.

We determined risk factors for hand contamination and compared the efficacy of 3 randomly allocated hand hygiene agents in a group of surgical intensive care unit nurses. We cultured samples of one of the subjects' hands before and samples of the other hand after hand hygiene was performed. Ring wearing was associated with 10-fold higher median skin organism counts; contamination with Staphylococcus aureus, gram-negative bacilli, or Candida species; and a stepwise increased risk of contamination with any transient organism as the number of rings worn increased (odds ratio [OR] for 1 ring worn, 2.6; OR for >1 ring worn, 4.6). Compared with use of plain soap and water, hand contamination with any transient organism was significantly less likely after use of an alcohol-based hand rub (OR, 0.3; 95% confidence interval [CI], 0.1-0.8) but not after use of a medicated hand wipe (OR, 0.9; 95% CI, 0.5-1.6). Ring wearing increased the frequency of hand contamination with potential nosocomial pathogens. Use of an alcohol-based hand rub resulted in significantly less frequent hand contamination.

Electronic interpretation of chest radiograph reports to detect central venous catheters.

OBJECTIVE: To evaluate whether a natural language processing system, SymText, was comparable to human interpretation of chest radiograph reports for identifying the mention of a central venous catheter (CVC), and whether use of SymText could detect patients who had a CVC. DESIGN: To identify patients who had a CVC, we performed two surveys of hospitalized patients. Then, we obtained available reports from 104 patients who had a CVC during one of two cross-sectional surveys (ie, case-patients) and 104 randomly selected patients who did not have a CVC (ie, control-patients). SETTING: A 600-bed public teaching hospital. RESULTS: Chest radiograph reports were available from 124 of the 208 participants. Compared with human interpretation, SymText had a sensitivity of 95.8% and a specificity of 98.7%. The use of SymText to identify case- and control-patients resulted in a sensitivity of 43% and a specificity of 98%. Successful application of SymText varied significantly by venous insertion site (eg, a sensitivity of 78% for subclavian and a sensitivity of 3.7% for femoral). Twenty-six percent of the case-patients had a femoral CVC. CONCLUSIONS: Compared with human interpretation, SymText performed well in interpreting whether a report mentioned a CVC. In patient populations with less frequent CVC placement in femoral veins, the sensitivity for CVC detection likely would be higher. Applying a natural language processing system to chest radiograph reports may be a useful adjunct to other data sources to automate detection of patients who had a CVC.

Detection of postoperative surgical-site infections: comparison of health plan-based surveillance with hospital-based programs.

BACKGROUND: Review of health plan administrative data has been shown to be more sensitive than other methods for identifying postdischarge surgical-site infections (SSIs), but there has not been a direct comparison between this method and hospital-based surveillance for all infections, including those diagnosed before discharge. We compared these two methods for identifying SSIs following coronary artery bypass graft (CABG) procedures. METHODS: We studied 1,352 CABG procedures performed among members of one health plan from March 1993 through June 1997. Health plan administrative records were reviewed based on claims containing diagnoses or procedures suggestive of infection or outpatient dispensing of antibiotics appropriate for SSI. Hospital-based surveillance information was also reviewed. SSI rates were calculated based on the total events identified by either mechanism. RESULTS: Postdischarge information was reviewed for 328 (85%) of 388 procedures. SSIs were confirmed in 167 patients (13% overall risk of confirmed SSI; range, 3% to 14% in the 5 hospitals). The overall sensitivity of hospital-based surveillance was 49.7% (83 of 167), and that of health plan data was 71.8% (120 of 167). There was no significant difference among hospitals in the sensitivity of either surveillance mechanism. CONCLUSIONS: Surveillance based on health plan data identified more postoperative infections, including those occurring before discharge, than did hospital-based surveillance. Screening administrative data and pharmacy activity may be an important adjunct to SSI surveillance, allowing efficient comparison of hospital-specific rates. Interpretation of differences among hospitals' infection rates requires case mix adjustment and understanding of variations in hospitals' discharge diagnosis coding practices.

CDC central-line bloodstream infection prevention efforts produced net benefits of at least $640 Million during 1990-2008.

The prevention of central line-associated bloodstream infections in patients in hospital critical care units has been a target of efforts by the Centers for Disease Control and Prevention (CDC) since the 1960s. We developed a historical economic model to measure the net economic benefits of preventing these infections in Medicare and Medicaid patients in critical care units for the period 1990-2008-a time when reductions attributable to federal investment resulted primarily from CDC efforts-using the cost perspective of the federal government as a third-party payer. The estimated net economic benefits ranged from $640 million to $1.8 billion, with the corresponding net benefits per case averted ranging from $15,780 to $24,391. The per dollar rate of return on the CDC's investments ranged from $3.88 to $23.85. These findings suggest that investments in CDC programs targeting other health care-associated infections also have the potential to produce savings by lowering Medicare and Medicaid reimbursements.

Methicillin-resistant-Staphylococcus aureus hospitalizations, United States.

Methicillin-resistant Staphylococcus aureus (MRSA) is increasingly a cause of nosocomial and community-onset infection with unknown national scope and magnitude. We used the National Hospital Discharge Survey to calculate the number of US hospital discharges listing S. aureus-specific diagnoses, defined as those having at least 1 International Classification of Diseases (ICD)-9 code specific for S. aureus infection. The number of hospital discharges listing S. aureus-specific diagnoses was multiplied by the proportion of methicillin resistance for each corresponding infection site to determine the number of MRSA infections. From 1999 to 2000, an estimated 125,969 hospitalizations with a diagnosis of MRSA infection occurred annually, including 31,440 for septicemia, 29,823 for pneumonia, and 64,706 for other infections, accounting for 3.95 per 1,000 hospital discharges. The method used in our analysis may provide a simple way to assess trends of the magnitude of MRSA infection nationally.

Patient colonization and environmental contamination by vancomycin-resistant enterococci in a rehabilitation facility.

OBJECTIVES: To determine the frequency of environmental contamination in patient and common-use rooms and patient colonization by vancomycin-resistant enterococci (VRE). DESIGN: Cross-sectional study. SETTING: A 146-bed rehabilitation facility. PARTICIPANTS: Rectal cultures were collected from 74 (80%) of 93 patients. Environmental cultures were obtained from surfaces in 15 patient rooms (5 floors) and common-use areas on 8 floors. INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: Gastrointestinal colonization of patients and environmental contamination of surfaces by VRE. RESULTS: VRE was detected from 13 (18%) of 74 patients and 32 (10%) of 319 surfaces. The frequency of positive environmental cultures varied by location; cultures were more likely to be positive in patient rooms (15%), followed by common areas on patient floors (9%) and common areas separate from patient floors (1.3%). Surfaces were more likely to be positive in rooms with a VRE-colonized patient (24%), compared with rooms in which patient colonization status was unknown (13%, P=.13) or the patient was not colonized (0%, P=.002). Surfaces were more likely to be contaminated in a room that housed an incontinent compared with continent patients (22% vs 7%, P=.01). CONCLUSIONS: Although environmental contamination by VRE was common in patient rooms, contamination of common-use areas separate from patient floors was infrequent. Despite use of common-use areas by colonized patients, isolation practices at this facility appear to have minimized environmental surface contamination in these areas.

Identification of serious drug-drug interactions: results of the partnership to prevent drug-drug interactions.

OBJECTIVE: To develop a list of clinically important drug-drug interactions (DDIs) likely to be encountered in community and ambulatory pharmacy settings and detected by a computerized pharmacy system. DESIGN: Cross-sectional, one-time evaluation. SETTING: United States in fall 2001. PARTICIPANTS: An expert panel comprising two physicians, two clinical pharmacists, and an expert on DDIs. INTERVENTIONS: Systematic review of drug interaction compendia and published literature, ratings (on a 1 to 10 scale) of various clinical aspects of DDIs (e.g., clinical importance, quality and quantity of evidence, causal relationship, risk of morbidity and mortality), and a modified Delphi consensus-building process. MAIN OUTCOME MEASURE: Panelists' opinions about clinical importance of DDIs. RESULTS: The expert panel considered 56 DDIs. Of these, 28 had a mean clinical importance score of 8.0 or more. The ratings for clinical importance ranged from 3.2 to 9.6, with a mean +/- SD of 7.5 +/- 1.5 across the combinations examined. The mean score for the quality of literature suggesting the interaction exists ranged from 1.0 to 9.6, with a mean +/- SD of 5.8 +/- 2.5. In terms of substantiation of the interactions evaluated, the mean +/- SD rating was 6.3 +/- 2.2, with a range from 1.4 to 9.2. Through the modified Delphi process, the panel determined that 25 interactions were clinically important. CONCLUSION: Using an expert panel and a standard evaluation tool, 25 clinically important drug interactions that are likely to occur in the community and ambulatory pharmacy settings were identified. Pharmacists should take steps to prevent patients from receiving these interacting medications, and computer software vendors should focus interaction alerts on these and similarly important DDIs.

Using automated health plan data to assess infection risk from coronary artery bypass surgery.

We determined if infection indicators were sufficiently consistent across health plans to allow comparison of hospitals' risks of infection after coronary artery bypass surgery. Three managed care organizations accounted for 90% of managed care in eastern Massachusetts, from October 1996 through March 1999. We searched their automated inpatient and outpatient claims and outpatient pharmacy dispensing files for indicator codes suggestive of postoperative surgical site infection. We reviewed full text medical records of patients with indicator codes to confirm infection status. We compared the hospital-specific proportions of cases with an indicator code, adjusting for health plan, age, sex, and chronic disease score. A total of 536 (27%) of 1,953 patients had infection indicators. Infection was confirmed in 79 (53%) of 149 reviewed records with adequate documentation. The proportion of patients with an indicator of infection varied significantly (p < 0.001) between hospitals (19% to 36%) and health plans (22% to 33%). The difference between hospitals persisted after adjustment for health plan and patients' age and sex. Similar relationships were observed when postoperative antibiotic information was ignored. Automated claims and pharmacy data from different health plans can be used together to allow inexpensive, routine monitoring of indicators of postoperative infection, with the goal of identifying institutions that can be further evaluated to determine if risks for infection can be reduced.