Nosocomial Infection Surveillance, 1984
Teresa C. Horan, M.P.H. John W. White, Ph.D. William R. Jarvis, M.D. T. Grace Emori, R.N., M.S. David H. Culver, Ph.D. Van P. Munn, B.S. Clyde Thornsberry, Ph.D. David R. Olson, Ph.D. James M. Hughes, M.D. Hospital Infections Program Center for Infectious Diseases
Nosocomial infections cause substantial morbidity and mortality, prolong the hospital stay of affected patients, and increase direct patient-care costs (1-5). Since 1970, the National Nosocomial Infections Surveillance System (NNIS) has collected and analyzed data on the frequency of nosocomial infections in U.S. hospitals. This report provides descriptive data on nosocomial infections in a sample of U.S. hospitals in 1984. Materials and Methods
The methods used in this surveillance system and the characteristics of participating hospitals have been described in detail (4,6). In brief, hospitals participating in NNIS conduct active hospital-wide surveillance using uniform definitions of nosocomial infections. Although the definitions are specific for different sites of infection, onset must occur during hospitalization or shortly after discharge, and the infection may not be present or incubating at the time of the patient's admission. Each month data are recorded on standardized forms that are sent to CDC, where they are coded, edited, and entered into a computer before being analyzed. In 1984, 51 hospitals regularly (greater than or equal to 9 months) reported data to CDC. For each nosocomial infection detected, the following information was reported: site of infection; date of onset; hospital service on which the patient was placed; age and sex of the patient; pathogens isolated; occurrence of secondary bacteremia; antimicrobial susceptibility of bacterial pathogens; and, for those patients who died with a nosocomial infection, the relationship of the infection to death. In addition, the hospitals reported the number of patients discharged each month from six primary services: medicine, surgery, obstetrics, gynecology, pediatrics, and newborn. Results
The NNIS Sample. The hospitals participating in NNIS are not a probability sample of U.S. hospitals; however, those hospitals that regularly reported data in 1984 ranged in size from 80 to 1,200 beds, were located throughout the United States, and were owned by state and local governments, as well as by profit and nonprofit organizations. The geographic distribution of the 51 hospitals among the four regions of the country (Northeast, North Central, South, and West) was roughly the same as that for all 6,375 U.S. acute-care hospitals included in the American Hospital Association Annual Survey of Hospitals (7). Hospitals affiliated with medical schools, referred to as teaching hospitals, are still greatly overrepresented among the NNIS hospitals; 61% (31/51) of the NNIS hospitals are teaching hospitals, whereas only 17% of the hospitals across the country are affiliated with a medical school. Similarly, the 51 NNIS hospitals tend to be large, with a median size of 406 beds, compared with a median size of only 112 beds for the 6,375 U.S. acute-care hospitals (7).
Despite these limitations, previous analyses have shown that data collected in NNIS can be usefully interpreted by stratifying the 51 reporting hospitals into three groups: 1) 20 (39%) nonteaching hospitals, 2) 18 (35%) small teaching hospitals of 500 or fewer beds, and 3) 13 (26%) large teaching hospitals of more than 500 beds (4,6).
The overall infection rate (number of hospital-acquired infections per 1,000 patients discharged) was highest in the large teaching hospitals and lowest in the nonteaching hospitals (Table 1). In all three hospital categories, the infection rate was highest on the surgery service, followed by the medicine and gynecology services (Table 2). On each of the six primary services, the infection rate was highest at the large teaching hospitals and lowest at the nonteaching hospitals, with the exception of the gynecology service rate, which was highest at small teaching hospitals.
In all three hospital categories, the urinary tract was the site most frequently infected, followed by lower respiratory tract or surgical wound infections (Table 3). For each site of infection, the infection rates were highest in the large teaching hospitals and lowest in the nonteaching hospitals.
Infections of the urinary tract, of surgical wounds, and of the lower respiratory tract accounted for almost three-fourths of the infections in all three hospital categories (Table 4). Primary bacteremia and cutaneous infections accounted for a higher percentage of infections in the large teaching hospitals than in the other hospitals.
Combined Rates by Service and Site. In general, the site-specific infection rate on each service was highest in the large teaching hospitals and lowest in the nonteaching hospitals (Table 5). In each hospital category, urinary tract infections occurred predominantly on the medicine, surgery, and gynecology services. Surgical wound infections occurred primarily on the surgery, gynecology, and obstetrics services. Lower respiratory infections occurred predominantly on the surgery and medicine services. Primary bacteremia occurred most frequently on the medicine and surgery services at nonteaching and large teaching hospitals. At small teaching hospitals, primary bacteremia was most frequently seen on the medicine and pediatrics services, followed by the newborn and surgery services. Cutaneous infections occurred primarily on the newborn service in each hospital category.
Pathogens. Of the 26,965 infections reported, 64% were caused by single pathogens, and 20% were caused by multiple pathogens (Figure 1). No pathogen was identified in 6% of the infections, and no culture was obtained in 10%. Of the 84% of infections in which a pathogen was identified, 86% were caused by aerobic bacteria, 2% by anaerobic bacteria, and 8% by fungi. Viruses, protozoa, and parasites collectively accounted for 5% of the infections of known etiology.
Escherichia coli, Pseudomonas aeruginosa, enterococci, and Staphylococcus aureus were the most frequently reported pathogens (Table 6). E. coli was the pathogen most often reported on all services except pediatrics and newborn, where S. aureus was the most common. P. aeruginosa was the second most frequently identified pathogen on the medicine and surgery services, whereas enterococci were second on the gynecology and obstetrics services. Coagulase-negative staphylococci were the second most frequently identified pathogens on the pediatrics and newborn services.
E. coli was the pathogen most frequently associated with urinary tract infections, followed by enterococci and P. aeruginosa (Table 7). S. aureus was the pathogen most often associated with surgical wound infections, followed by enterococci and E. coli. P. aeruginosa was the pathogen most frequently associated with lower respiratory tract infections, followed by S. aureus and Klebsiella spp. Coagulase-negative staphylococci were the pathogens most commonly associated with primary bacteremia, followed by S. aureus and E. coli.
When the pathogens causing infections at the five major sites were examined by service, interesting differences were noted (Table 8). On all six services, E. coli was the pathogen most often isolated from the urinary tract. Enterococci were the second most frequently isolated pathogens from the urinary tract on the obstetrics, gynecology, and medicine services, whereas P. aeruginosa was the second most commonly isolated pathogen from the urinary tract on the surgery and pediatrics services, and Klebsiella spp. were second on the newborn service. S. aureus was the pathogen most often associated with surgical wound infections on all services except gynecology, where E. coli was isolated most frequently. The pathogen most frequently associated with lower respiratory infections on all services was P. aeruginosa, with S. aureus second on all but the gynecology and newborn services. Coagulase-negative staphylococci were most often associated with primary bacteremia on the pediatrics, newborn, and surgery services, whereas S. aureus was the pathogen most frequently associated with bacteremia on the medicine and obstetrics services. E. coli and Bacteroides spp. were isolated with the highest frequency in association with primary bacteremia on the gynecology service. S. aureus was most commonly associated with cutanesous infections and was followed by coagulase-negative staphylococci on all on all services except surgery and obstetrics. On surgery, S. aureus was first, followed by P. aeruginosa; on obstetrics, E. coli was isolated most frequently, followed by S. aureus.
Secondary Bacteremia. Secondary bacteremia was defined as a bloodstream infection with an organism that was also isolated from an infection at another site. Secondary bacteremia was reported most frequently by large teaching hospitals and least frequently by nonteaching hospitals (Table 9). Secondary bacteremia occurred most often on the pediatrics service in teaching hospitals, followed by the medicine, newborn, and surgery services, and it occurred least frequently on the obstetrics and gynecology services. In nonteaching hospitals, secondary bacteremia occurred most often on the medicine, obstetrics, and surgery services and least often on the newborn, gynecology, and pediatrics services. For all hospital categories, secondary bacteremia was associated less frequently with urinary tract, surgical wound, lower respiratory tract, and cutaneous infections than with infections, collectively, at "other" sites (Table 10). With respect to the four major sites, and excluding primary bacteremia, secondary bacteremia occurred most often following cutaneous infections. It occurred most frequently in all hospitals following infections with Acinetobacter spp., Bacteroides spp., S. aureus, Serratia spp., and coagulase-negative staphylococci (Table 11), but this varied greatly within each hospital category. For example, in nonteaching hospitals, S. aureus was the main pathogen that caused secondary bacteremia. In small teaching hospitals, the frequency of secondary bacteremia due to coagulase-negative staphylococci has nearly doubled since 1983 (6). In addition, Bacteroides spp., S. aureus, Group B Streptococcus, and Acinetobacter were frequently associated with secondary bacteremia in small teaching hospitals. In large teaching hospitals, no pathogen predominated as the causative agent of secondary bacteremia.
Antimicrobial Resistance. Resistance was defined as the number of resistant isolates divided by the number of organisms that were either sensitive or resistant, multiplied by 100. Methicillin-resistant S. aureus was most commonly reported from the large teaching hospitals (Table 12). In fact, for all the antimicrobials listed in Table 12, resistance was most often reported from the large teaching hospitals.
The percentages of E. coli, Klebsiella pneumoniae, Serratia marcescens, and P. aeruginosa organisms that were resistant to aminoglycosides and selected beta-lactam antibiotics varied according to the three hospital categories (Tables 13-16). Aminoglycoside resistance was most common in P. aeruginosa and S. marcescens, and cefotaxime or moxalactam resistance was most common in P. aeruginosa.
Mortality. Data from individual hospitals are included in the mortality analyses if the hospital assessed and reported the relationship of infection to death for more than 50% of the infections in patients who died while hospitalized. The 42 hospitals that met this criterion reported a total of 22,432 infections; among the 1,253 patients who died, there were 1,811 infections for which the relationship of the infection to death was recorded. Approximately 1% of all nosocomial infections caused death, and 3% contributed to death (Table 17). Infections were more often reported to cause or contribute to death in small teaching and in nonteaching hospitals than in large teaching hospitals. Among infected patients who died while hospitalized, 9% of the infections reportedly caused death, 38% contributed to it, and 37% were not related to death; in 15% of these infections, the relationship of the infection to death could not be determined (Table 18).
Discussion Nosocomial infections remain an important cause of morbidity and mortality in U.S. hospitals. Data from NNIS, the only national source of prospectively collected data on hospital-acquired infections, show that the overall rate of nosocomial infections during 1984 was 3.4 infections per 100 patients discharged. This is similar to the infection rate reported for the 3-year period 1980-1982 (4) and for 1983 (6). By comparison, the Study on the Efficacy of Nosocomial Infection Control (SENIC) found that a nosocomial infection develops in 5%-6% of hospitalized patients (8). SENIC was a retrospective study involving a representative sample of U.S. hospitals in 1975-1976. NNIS data suggest that the true incidence of nosocomial infections is underestimated. Factors contributing to the underestimation include variability of the intensity of surveillance and availability of laboratory support, especially in diagnostic virology. Since identification of nosocomial viral infections depends on both laboratory detection and surveillance intensity, hospitals without virology laboratory support will be unlikely to detect most of these infections.
Since 1980 (4), nosocomial infection rates have been consistently highest in large teaching hospitals and lowest in nonteaching hospitals for all services and sites of infection, suggesting that the three-level stratification effectively defines hospital categories in which patients have different levels of risk for acquiring nosocomial infections. This difference in risk undoubtedly reflects severity of underlying illness (patient mix) and the extent to which invasive diagnostic and therapeutic procedures are performed in these hospitals.
As in 1980-1982 (4) and in 1983 (6), the infection rates were highest on the surgery and medicine services, probably because of their high-risk patient populations. The lowest infection rates were on the pediatrics and newborn services. One explanation for this lower rate may be that in NNIS hospitals, there are fewer high-risk children and newborns than adults, particularly in the small hospitals. Furthermore, most of the infants included in the newborn service are in well-baby nurseries, where the infection risk is expected to be lower. Another factor that may help explain the lower rates of infection on the pediatrics and newborn services is that only a small proportion of NNIS hospitals have diagnostic virology laboratories; therefore, many viral infections probably go undetected. Since children more often acquire nosocomial viral infections than adults (9), and since in one study viruses accounted for approximately 14% of nosocomial infections in children (10), NNIS hospitals are probably underreporting viral infections. In addition, other factors, such as the short time that many pediatric patients are hospitalized and the frequent use of isolation precautions on the pediatrics and newborn services, may reduce the incidence of nosocomial infections on these services.
In 1984, infection rates on different services and at different sites of infection within the three hospital categories varied little from those reported for 1983 (6). Since 1980-1982 (4), the primary bacteremia and lower respiratory tract infection rates have increased. The overall lower respiratory tract infection rate surpassed the rate of surgical wound infections in 1984. Whether this is an artifact of reporting or a true shift in the rates is not known.
Specimens for microbiologic testing were obtained from 90% of the patients reported to have a nosocomial infection. Aerobic bacteria were the most commonly identified etiologic agents. Anaerobic bacteria, fungi, parasites, and viruses were seldom reported, reflecting in part the frequency with which these pathogens are looked for, as well as the diagnostic laboratory capabilities of the hospitals.
E. coli was the most frequently identified pathogen on the four adult services, reflecting the fact that this organism was the primary cause of urinary tract infections on these services. In contrast, S. aureus was the pathogen most often identified on the newborn and pediatrics services. Coagulase-negative staphylococci were the second most frequent cause of nosocomial infections on the newborn and pediatrics services and were an important cause of bacteremia on all services except gynecology and obstetrics. Recent studies suggest that the increasing use of long-line catheters may be contributing to the emergence of coagulase-negative staphylococci as an important cause of primary bacteremia (11,12).
Previous analyses of NNIS data have suggested that secondary bacteremia carries an increased risk of death (13). In all hospitals, the major sites of infection that were most likely to result in secondary bacteremia were cutaneous infections, followed by surgical wound and lower respiratory tract infections. Infections at sites other than the four major sites were, collectively, more frequently associated with secondary bacteremia. These include cardiovascular and intra-abdominal infections. An increase in cardiovascular surgery and in the use of long-line venous and arterial catheters may have accounted for the rise since 1983 in the percentage of infections associated with the cardiovascular system (6). Because of the increased risk of death associated with secondary bacteremia, these infections continue to be a high priority for prevention and control (13).
As in the past, the incidence of methicillin-resistant S. aureus (MRSA) infections was highest at large teaching hospitals (4,6,14), and between 1983 and 1984, these infections increased by more than 25% in each hospital category (6). Since 1983, the proportion of S. aureus organisms resistant to gentamicin and clindamycin increased at small teaching and at nonteaching hospitals, but at large teaching hospitals the proportion resistant to gentamicin decreased and that resistant to clindamycin remained about the same (6). The factors responsible for these resistance trends require further study. Recent work suggests that risk factors for MRSA may differ by type of hospital (15).
In 1984, compared with 1983, the proportion of K. pneumoniae organisms resistant to gentamicin and tobramycin decreased in large teaching hospitals and increased in nonteaching hospitals; however, resistance to amikacin increased in large teaching hospitals and decreased in both nonteaching and small teaching hospitals (6). Since 1982, the resistance of P. aeruginos to both cefotaxime and moxalactam has increased in the small teaching hospitals, but the trend has been variable in nonteaching hospitals (4,6). Over the same period cefotaxime resistance has continued to decrease, and moxalactam resistance has been rising in the large teaching hospitals (4,6). Since the proportion of isolates tested against cefotaxime and moxalactam was small, these data should be interpreted with caution.
When compared with NNIS mortality data for 1980-1982 (4) and 1983 (6), the overall percentage of infections reported to cause or contribute to death in 1984 has not changed significantly. Since 1980, the large teaching hospitals have reported a slightly lower percentage of infections each year as causing or contributing to a patient's death (4,6). The small teaching hospitals reported about the same frequency, and the nonteaching hospitals reported a slight increase each year (4,6). Mortality data should be interpreted with caution, since standard criteria for assessing the relationship of infection to death do not exist.
This nationwide nosocomial surveillance system is expanding in four directions (6). First, microcomputer software called the Interactive Data Entry and Analysis System (IDEAS) has been developed to support nosocomial infection surveillance activities of NNIS hospitals. Beginning in October 1984, IDEAS was pilot tested in three hospitals and is now being used in 22 additional hospitals. This information management system not only helps to improve the quality and timeliness of nosocomial infection data collected in NNIS, but it also assists infection control practitioners in conducting more effective and efficient surveillance in their institutions. Second, additional hospitals are being added to the surveillance system so that the data obtained will be from a more representative sample of all acute-care hospitals in the United States. Since recruitment began in March 1985, 10 hospitals have been added and additional hospitals are being considered for enrollment. Third, in July 1985, the feasibility of collecting data on antimicrobial usage in NNIS hospitals was assessed. Hospitals with computerized pharmacy records wishing to participate in the study will report on the use of antimicrobial agents so that for selected nosocomial bacterial pathogens the relationship between usage and resistance can be evaluated. Fourth, strategies are being developed for determining a more sensitive indicator of patients' risk based on characteristics of both the patient and the hospital (such as the three size categories used in this report). When various levels of nosocomial infection risks can be calculated, infection rates among hospitals can be compared and the distribution of risks can be standardized; in addition, hospital-specific infection rates and secular trends can be evaluated more effectively.
Jr. Extra charges and prolongation of stay attributable to nosocomial infections: a prospective interhospital comparison. Am J Med 1981;70:51-8.
2. Haley RW, Schaberg DR, Von Allmen SD, McGowan JE. Estimating the extra charges and prolongation of hospitalization due to nosocomial infections: a comparison of methods. J Infect Dis 1980;141:248-57.
3. McGowan JE. Cost and benefit: a critical issue for hospital infection control. Am J Infect Control 1982;10:100-8.
4. CDC. Nosocomial infection surveillance, 1980-1982. In: Surveillance Summaries (published four times a year). 1983;32(No.4SS):1SS-16SS.
5. Hughes JM, Jarvis WR. Epidemiology of nosocomial infections. In: Lennette EH, Balows A, Hausler WJ Jr, Shadomy HJ, eds. Manual of clinical microbiology. Washington, D.C.: American Society for Microbiology 1985:99-104.
6. CDC. Nosocomial infection surveillance, 1983. In: Surveillance Summaries (published four times a year). 1984;33(no. 2SS):9SS-21SS.
7. American Hospital Association. American Hospital
Association guide to the health care field. Chicago: American Hospital Association, 1983.
8.Haley RW, Hooton TM, Culver DH, et al. Nosocomial infections in U.S. hospitals, 1975-1976: estimated frequency by selected characteristics of patients. Am J Med 1981;70:947-59.
9.Valenti WM, Hall CB, Douglas RG Jr, Menegus MA, Pincus PH. Nosocomial viral infections: 1. Epidemiology and significance. Infect Control 1980;1:33-7. 10. Welliver RC, McLaughlin S. Unique epidemiology of
nosocomial infection in a children's hospital. Am J Dis Child 1984;138:131-5. 11. Munson DP, Thompson TR, Johnson DE, Rhame FS, Van
Drunen N, Ferrieri P. Coagulase-negative staphylococcal septicemia: experience in a newborn intensive care unit. J Pediatr 1982;101:602-5. 12. Christensen GD, Bisno AL, Parisi JT, McLaughlin B,
Heaster MG, Luther RW. Nosocomial septicemia due to multiply antibiotic-resistant Staphylococcus epidermidis. Ann Intern Med 1982;96:1-10. 13. Hughes JM, Munn VP, Jarvis WR, Culver DH, Haley RW.
Mortality associated with nosocomial infections in the United States, 1975-1981. Abstract 701. Presented at the 22nd Interscience Conference on Antimicrobial Agents and Chemotherapy, Miami Beach, Florida, 1982 October 4-6. 14. Haley RW, Hightower AW, Khabbaz RF, et al. The
emergence of methicillin-resistant Staphylococcus aureus infections in United States hospitals. Ann Intern Med 1982; 97:297-308. 15. Jarvis WR, Thornsberry C, Boyce J, Hughes JM.
Methicillin-resistant Staphylococcus aureus at children's hospitals in the United States. Pediatr Infect Dis 1985; 4:651-5.
Disclaimer All MMWR HTML documents published before January 1993 are electronic conversions from ASCII text into HTML. This conversion may have resulted in character translation or format errors in the HTML version. Users should not rely on this HTML document, but are referred to the original MMWR paper copy for the official text, figures, and tables. An original paper copy of this issue can be obtained from the Superintendent of Documents, U.S. Government Printing Office (GPO), Washington, DC 20402-9371; telephone: (202) 512-1800. Contact GPO for current prices.**Questions or messages regarding errors in formatting should be addressed to firstname.lastname@example.org.
Page converted: 08/05/98
This page last reviewed 5/2/01