Secular trends in bloodstream infection caused by antimicrobial-resistant bacteria in New Jersey hospitals, 1991 to 1995.

INTRODUCTION: Antimicrobial resistance among bacteria is an increasing public health problem. In 1991, New Jersey was the first state to establish statewide, hospital-based surveillance for antimicrobial-resistant bacteria. METHODS: Each month, all 96 nonfederal New Jersey hospital laboratories complete a form listing the species identity and drug susceptibility results for selected antimicrobial-resistant bacteria isolated from blood cultures from hospital inpatients. Penicillin-resistant Streptococcus pneumoniae and aminoglycoside-resistant gram-negative rods were studied from 1991 to 1995. Vancomycin-resistant enterococci and imipenem-resistant gram-negative rods were studied from 1992 through 1995. RESULTS: From 1992 to 1995, the vancomycin-resistant enterococci bloodstream infection prevalence rate increased from 11 to 29 per 100,000 hospital admissions (p < 0.001); the rate was higher at larger hospitals, urban and inner-city hospitals, and teaching hospitals. From 1991 to 1995, the penicillin-resistant S. pneumoniae bloodstream infection rate increased from 1.1 to 9.9 per 100,000 admissions (p < 0.001). In contrast, bloodstream infection rates did not change significantly for imipenem-resistant (12.5 during 1992 and 14.1 during 1995, p = 0.4) or aminoglycoside-resistant (8.0 during 1991 and 6.8 during 1995, p = 0.4) gram-negative rods. CONCLUSIONS: We found that vancomycin-resistant enterococci and penicillin-resistant S. pneumoniae, but neither of two groups of antimicrobial-resistant gram-negative rods, are increasing rapidly in prevalence in New Jersey. Continued monitoring and interventions to slow these increases are needed.

A statewide surveillance system for antimicrobial-resistant bacteria: New Jersey.

OBJECTIVES: To determine the validity of an active, hospital laboratory isolate-based surveillance system in estimating rates of infection and to evaluate the use of surveillance data in describing institutional risk factors for increased rates of infection. Methicillin-resistant Staphylococcus aureus (MRSA) was chosen as the prototype organism for these evaluations. DESIGN: Correlation Study: linear regression analysis and Student's t test were used to evaluate the correlation between number of MRSA isolates and number of MRSA infections in acute-care hospitals. Cross-Sectional Study: Student's t test, analysis of variance, and multiple linear regression analysis were used to evaluate the association between mean annual rate of MRSA blood isolates and institutional risk factors for increased rates of infection. SETTING: Acute-care hospitals, New Jersey. RESULTS: The number of MRSA blood isolates was significantly correlated with MRSA blood infections (R, 0.78; P < .01) and provided a good proxy measure for number of infections. Multivariate analysis demonstrated hospital location in the inner city (P = .02) and number of occupied beds (P < .01) to be independently associated with increased mean annual rates of MRSA blood isolates in acute-care hospitals. CONCLUSIONS: This surveillance system is a valid tool for the estimation of institutional rates of infection and for the determination of institutional risk factors for increased rates of infection. It is ideal for further population-based investigations of antimicrobial-resistant bacteria.

AIDS and changing patterns of mortality.

We compared changes in all-causes mortality rates, 1986 versus 1980, among members 25 to 44 years of age of demographically defined groups with high AIDS cumulative incidence to the changes among same-age, same-sex members of groups with low AIDS cumulative incidence. Among nonwhite men ages 25-44 residing in northeastern New Jersey (NJ) counties, AIDS cumulative incidence was 1,409 cases per 100,000; all-causes mortality was 413.8 deaths per 100,000 per year in 1980 and increased 74% to 726.6 deaths per 100,000 per year by 1986. In contrast, among white men ages 25-44 residing in other NJ counties, AIDS cumulative incidence was 75 cases per 100,000; all-causes mortality fell slightly from 192.6 deaths per 100,000 per year in 1980 to 189.2 deaths per 100,000 per year in 1986. Among nonwhite women ages 25-44 residing in northeastern NJ counties, AIDS cumulative incidence was 435 cases per 100,000; all-causes mortality was 162.07 deaths per 100,000 per year in 1980 and increased 70% to 276.3 deaths per 100,000 per year by 1986. Among white women ages 25-44 residing in other NJ counties, AIDS cumulative incidence was 9.1 cases per 100,000; all-causes mortality was 90.5 deaths per 100,000 per year in 1980 and fell slightly to 83.0 deaths per 100,000 per year in 1986. A substantial portion of the increased mortality of the groups with high AIDS cumulative incidence resulted from causes that have not been associated with HIV infection.

Cumulative AIDS incidence and altered pneumonia mortality.

We grouped New Jersey residents according to age, sex, race, and residence-specific cumulative incidence of AIDS since the onset of the AIDS epidemic: less than 15, 15-99, 100-499, and greater than or equal to 500 cases per 100,000 people, respectively. We determined mortality from bacterial and viral pneumonias (International Classification of Diseases [ICD] 480.0-486.9) from underlying cause of death files. Between 1980 and 1986, pneumonia mortality increased from 15.1 deaths per 100,000 per year (95% confidence interval [CI] 10.4, 19.7) to 25.0 deaths per 100,000 per year (95% CI 19.2, 30.8), an increase of 10.0 deaths per 100,000 per year (95% CI 2.6, 17.3), among those 25-44 years of age in the highest cumulative incidence group for AIDS. Increases in other population subgroups were approximately proportional to each subgroup's AIDS cumulative incidence. In particular, pneumonia mortality did not increase among those 25-44 years of age in groups with low cumulative incidence of AIDS. Deficiency of cell-mediated immunity, a diagnosis commonly applied in AIDS cases, was listed as a secondary diagnosis in 14% of the pneumonia deaths of persons 25-44 years of age in 1986 and in none of those in 1980.

Fatal and nonfatal immersion injuries among New Jersey residents.

We identified immersion injuries of New Jersey residents from mortality and hospital discharge data. The incidence rate was 2.3 immersion injuries (1.3 fatal and 1.0 nonfatal) per 100,000 population per year. Incidence rates were elevated among young children, men, blacks, and residents of counties in the southern part of the state. Case fatality ratios increased with age and were higher for men than for women from 10 to 50 years of age.

Cumulative AIDS incidence and altered mortality from mycobacterial disease: New Jersey.

Changes in mycobacterial disease mortality between 1980 and 1986 were examined among New Jersey residents aged 25 to 44 using single cause of death data. The demographic group with the highest cumulative incidence of acquired immune deficiency syndrome (AIDS) (non-white residents of the four urban counties adjacent to New York City) sustained an increase of 10.1 deaths/100,000 men/yr and 3.1 deaths/100,000 women/yr. Groups with lower cumulative incidence of AIDS sustained smaller increases in mycobacterial disease mortality. The group with the lowest cumulative incidence of AIDS (white residents outside the four urban counties adjacent to New York City) sustained the smallest increase in tuberculosis (TB) mortality. Using single cause of death data, it was not possible to identify a relationship between increased extrapulmonary TB deaths and AIDS cumulative incidence, but such a relationship was identifiable from multiple cause of death data. Of 30 mycobacterial disease deaths of all ages with cellular immune deficiency as a contributory diagnosis on the death certificate, 21 (70%) were known to the state's AIDS registry as AIDS cases and four more (13%) were known to the registry as having human immunodeficiency virus (HIV) disease not meeting the full clinical criteria for AIDS. Young populations with a high cumulative incidence of AIDS have experienced substantially increased mortality from mycobacterial diseases. The association of mycobacterial disease mortality with HIV disease may be underestimated from AIDS registry data and from searches of single cause of death data for mycobacterial disease deaths.

Cumulative AIDS incidence and altered mortality from bacterial infections.

To determine whether populations with high cumulative incidence of acquired immunodeficiency syndrome (AIDS) experienced increased deaths from sepsis, central nervous system abscess, or endocarditis, New Jersey AIDS patients were grouped according to their age, sex, race, and residence-specific cumulative incidence of AIDS since the onset of the AIDS epidemic. Between 1980 and 1986, among 25-44 year olds in the highest cumulative incidence group for AIDS, sepsis mortality increased from 3.3 to 15.2 deaths/100,000/year, an increase of 11.9 deaths/100,000/year (95% confidence interval (6.9, 17.0) deaths/100,000/year); mortality from central nervous system abscesses increased from zero to 1.7 (0.1, 3.2) deaths/100,000/year; and mortality from endocarditis increased from 0.8 deaths/100,000/year to 2.4 deaths/100,000/year, an increase of 1.6 (-0.5, 3.7) deaths/100,000/year. Age-matched New Jersey patient populations with low cumulative incidence of AIDS did not sustain a similar increase. The HIV disease-associated increase in sepsis mortality among young populations represents a new component of the substantial increase in U.S. sepsis mortality that occurred over the last two decades, but was previously limited to older populations.

Microcomputer method to extract subfiles from statewide multiple cause of death files.

A simple microcomputer system has been developed using off-the-shelf components. The system permits local access, in an acceptable time frame, to several years of New Jersey multiple cause of death data (approximately 70,000 records per calendar year). This report describes the system, including hardware, software, data storage space requirements, speed of data access, and the effect of abbreviating the list of diagnoses. Prospects for extension of such a system to larger data bases are discussed.