Richard Heffernan

A Space–Time Permutation Scan Statistic for Disease Outbreak Detection

Background The ability to detect disease outbreaks early is important in order to minimize morbidity and mortality through timely implementation of disease prevention and control measures. Many national, state, and local health departments are launching disease surveillance systems with daily analyses of hospital emergency department visits, ambulance dispatch calls, or pharmacy sales for which population-at-risk information is unavailable or irrelevant. Methods and Findings We propose a prospective space–time permutation scan statistic for the early detection of disease outbreaks that uses only case numbers, with no need for population-at-risk data. It makes minimal assumptions about the time, geographical location, or size of the outbreak, and it adjusts for natural purely spatial and purely temporal variation. The new method was evaluated using daily analyses of hospital emergency department visits in New York City. Four of the five strongest signals were likely local precursors to citywide outbreaks due to rotavirus, norovirus, and influenza. The number of false signals was at most modest. Conclusion If such results hold up over longer study times and in other locations, the space–time permutation scan statistic will be an important tool for local and national health departments that are setting up early disease detection surveillance systems.

Monitoring the Impact of Influenza by Age: Emergency Department Fever and Respiratory Complaint Surveillance in New York City

Background The importance of understanding age when estimating the impact of influenza on hospitalizations and deaths has been well described, yet existing surveillance systems have not made adequate use of age-specific data. Monitoring influenza-related morbidity using electronic health data may provide timely and detailed insight into the age-specific course, impact and epidemiology of seasonal drift and reassortment epidemic viruses. The purpose of this study was to evaluate the use of emergency department (ED) chief complaint data for measuring influenza-attributable morbidity by age and by predominant circulating virus. Methods and Findings We analyzed electronically reported ED fever and respiratory chief complaint and viral surveillance data in New York City (NYC) during the 2001–2002 through 2005–2006 influenza seasons, and inferred dominant circulating viruses from national surveillance reports. We estimated influenza-attributable impact as observed visits in excess of a model-predicted baseline during influenza periods, and epidemic timing by threshold and cross correlation. We found excess fever and respiratory ED visits occurred predominantly among school-aged children (8.5 excess ED visits per 1,000 children aged 5–17 y) with little or no impact on adults during the early-2002 B/Victoria-lineage epidemic; increased fever and respiratory ED visits among children younger than 5 y during respiratory syncytial virus-predominant periods preceding epidemic influenza; and excess ED visits across all ages during the 2003–2004 (9.2 excess visits per 1,000 population) and 2004–2005 (5.2 excess visits per 1,000 population) A/H3N2 Fujian-lineage epidemics, with the relative impact shifted within and between seasons from younger to older ages. During each influenza epidemic period in the study, ED visits were increased among school-aged children, and each epidemic peaked among school-aged children before other impacted age groups. Conclusions Influenza-related morbidity in NYC was highly age- and strain-specific. The impact of reemerging B/Victoria-lineage influenza was focused primarily on school-aged children born since the virus was last widespread in the US, while epidemic A/Fujian-lineage influenza affected all age groups, consistent with a novel antigenic variant. The correspondence between predominant circulating viruses and excess ED visits, hospitalizations, and deaths shows that excess fever and respiratory ED visits provide a reliable surrogate measure of incident influenza-attributable morbidity. The highly age-specific impact of influenza by subtype and strain suggests that greater age detail be incorporated into ongoing surveillance. Influenza morbidity surveillance using electronic data currently available in many jurisdictions can provide timely and representative information about the age-specific epidemiology of circulating influenza viruses.

Changes in Invasive Pneumococcal Disease among HIV-Infected Adults Living in the Era of Childhood Pneumococcal Immunization

Context Routine pneumococcal conjugate vaccination for infants began in 2000. Its use markedly decreased invasive pneumococcal disease among children, but did it influence rates of disease among HIV-infected adults? Contribution Between 1998 and 2003, invasive pneumococcal disease among adults infected with HIV living in 7 surveillance areas in the United States decreased from 1127 to 919 cases per 100000 adults with AIDS. Disease caused by serotypes in the vaccine decreased 62%, whereas disease caused by nonvaccine serotypes increased 44%. Implications Indirect evidence suggests that pediatric vaccine use is associated with a decreased incidence of pneumococcal disease among HIV-infected adults. The Editors Apneumococcal conjugate vaccine containing 7 serotypes was recommended for routine use in infants in the United States beginning in 2000 (1). Widespread use of the vaccine caused steep declines in invasive pneumococcal disease among young children (2-4) and was associated with decreased disease attributable to vaccine serotypes among adults, for whom the vaccine is not licensed (3). Effects on disease among unvaccinated persons, often called herd effects, are presumably due to reduced transmission from immunized children. Because 90 pneumococcal serotypes cause human disease, there were concerns that the introduction of a conjugate vaccine containing 7 serotypes would lead to increased disease caused by nonvaccine-type organisms, a phenomenon called serotype replacement. Early postintroduction surveillance showed limited serotype replacement disease in the target age group, with no consistent trend toward increasing disease caused by nonvaccine serotypes among adults (3). To our knowledge, the effects of pediatric use of pneumococcal conjugate vaccine on immunocompromised adults, including those infected with HIV, have not previously been investigated. Persons infected with HIV are particularly susceptible to invasive pneumococcal disease, with a 50- to 100-fold higher incidence than the general U.S. population (5, 6). After the introduction of highly active antiretroviral therapy in the mid-1990s, surveillance in 3 geographic areas of the United States showed a 50% reduction in invasive pneumococcal disease among persons with AIDS (7). However, the incidence of pneumococcal disease among persons with AIDS leveled off by mid-1997 and continued to be approximately 35-fold higher in persons with AIDS than in those without HIV infection or AIDS through 2000 (7). We investigated trends in invasive pneumococcal disease among HIV-infected adults to document changes since the use of pneumococcal conjugate vaccine became widespread in children. Methods Active, laboratory-based surveillance for cases of invasive pneumococcal disease, defined as isolation of Streptococcus pneumoniae from a normally sterile site, was conducted through Active Bacterial Core surveillance of the Emerging Infections Program network (8). We included cases diagnosed between 1 January 1998 and 31 December 2003 among surveillance-area residents who were 18 to 64 years of age. We limited the analyses to 7 surveillance sites, including California (San Francisco County), Connecticut (entire state), Georgia (8-county Atlanta metropolitan area), Maryland (City of Baltimore and 5 neighboring counties), Minnesota (7-county MinneapolisSt. Paul metropolitan area), Oregon (3-county Portland metropolitan area), and Tennessee (Davidson, Hamilton, Knox, Shelby, and Williamson Counties). Information was systematically collected on the HIV status of case-patients at these sites. In 2003, the resident adult population in these 7 areas was 10.8 million (4.5% of the U.S. population between 18 and 64 years of age) (9) and included 9.5% of the estimated number of adults living with AIDS in the United States (10). Surveillance officers routinely contacted all clinical laboratories in their areas to identify cases and conducted audits of laboratory records to ensure complete ascertainment. Recurrent episodes, defined as invasive pneumococcal disease occurring more than 7 days after a previous case in a surveillance-area resident, were included in this analysis. The race and ethnicity as well as HIV status or previous AIDS diagnosis of case-patients were extracted from medical records by using standardized case report forms. Surveillance in Georgia did not prospectively collect information on HIV infection or AIDS for case-patients until 2000; for case-patients in 1999, we retrospectively reviewed medical records to collect this information. Analyses for 1998 exclude Georgia. Pneumococcal isolates were sent to reference laboratories at the Minnesota Department of Health (for case-patients from Minnesota), the Centers for Disease Control and Prevention, or the University of Texas Health Science Center at San Antonio for susceptibility testing by broth microdilution using standard protocols and quality control procedures (3, 11). Nonsusceptible isolates were defined as those with minimum inhibitory concentrations classified as intermediate or resistant for the antibiotic tested, according to the 2002 definitions of the National Committee for Clinical Laboratory Standards (12). Serotyping by the Quellung reaction was done at the Centers for Disease Control and Prevention or the Minnesota Department of Health (Minnesota cases only). The study personnel are listed in the Appendix. AIDS Surveillance Data For aggregated counties in each of the 7 surveillance areas, we obtained the estimated number of persons 18 to 64 years of age living with AIDS (as outlined in the 1993 Centers for Disease Control and Prevention case definition) (13), according to race and ethnicity and sex, on 30 June of each year. We obtained this number from the Centers for Disease Control and Prevention with permission from state AIDS surveillance coordinators. These estimates are derived from case report data by using a maximum likelihood method to account for delays in reporting new AIDS diagnoses and deaths among persons with AIDS (10), 14. Estimates of the number of persons living with HIV infection, not AIDS, were unavailable from 5 sites (California, Connecticut, Georgia, Maryland, and Oregon) that accounted for more than 80% of adults living with AIDS in the surveillance areas. Statistical Analysis For each surveillance area, we calculated the annual incidence rates of invasive pneumococcal disease among persons with AIDS as follows. We divided the number of cases of pneumococcal disease diagnosed during the calendar year among patients documented as having AIDS by the estimated number of persons 18 to 64 years of age living with AIDS. To calculate incidence among adults not infected with HIV, we used cases of pneumococcal disease in persons without documented HIV infection or AIDS in the numerator. For the denominator, we subtracted the number of adults living with AIDS from the total population of persons 18 to 64 years of age who lived in the surveillance areas; this number was obtained from the U.S. Census Bureau (9). As a proxy for the incidence rates among HIV-infected adults, we calculated a ratio of cases of pneumococcal disease among adults with HIV infection or AIDS to the estimated adult population living with AIDS, multiplied by 100000. Ratios for specific race and ethnicity categories were adjusted, assuming that the distribution of race and ethnicity for case-patients missing these data (3% of cases of pneumococcal disease) was the same as the sex-specific distribution of case-patients within each surveillance area for which race and ethnicity data were available. We adjusted data for selected serotypes (that is, conjugate vaccine or nonvaccine serotypes) or antibiotic-resistant pneumococci on the basis of the percentage of cases with isolates available for testing, assuming the distribution of cases missing serotype (9%) or antibiotic susceptibility data (7%) was the same as that of cases with isolates tested across all sites. We calculated 95% CIs for these ratios by using the standard error of the proportion of cases with isolates tested. To assess changes in the ratio of the number of cases of pneumococcal disease among HIV-infected adults to the number of adults living with AIDS before and after the introduction of the pneumococcal conjugate vaccine for children, we calculated percentage change and exact 95% CIs by comparing the average ratio during 1998 and 1999, called the baseline period, with that during 2002 or 2003. For differences or linear trends in proportions of cases or pneumococcal isolates, we calculated P values using chi-square tests; P values less than 0.05 indicated statistical significance. Statistical analyses were conducted with SAS, version 9.0 for Windows (SAS Institute, Inc., Cary, North Carolina), and EpiInfo, version 6.0 (Centers for Disease Control and Prevention, Atlanta, Georgia). Role of the Funding Source The funding source had no role in the design, analysis, or interpretation of the data or in the decision to submit the manuscript for publication. Results From 1998 through 2003, 8582 cases of invasive pneumococcal disease occurred in surveillance-area residents who were 18 to 64 years of age. Of these, 2013 cases occurred in persons with HIV infection or AIDS (Figure 1). When the latter group was excluded, the annual incidence rate in persons who were 18 to 64 years of age during the 1998 to 1999 baseline period averaged 13 cases per 100000 adults without AIDS. By 2003, this rate decreased to 9 cases per 100000, a decrease of 30% (95% CI, 25% to 35%; P< 0.001). Of the 2013 cases of pneumococcal disease among HIV-infected adults, 759 (38%) occurred among adults documented as having AIDS. Based on estimates of the number of adults living with AIDS in the surveillance areas, the incidence of invasive pneumococcal disease among persons with AIDS during the baseline period was 441 cases per 100000 adults. In 2002, the rate was 360 cases per 10000

Women at a sexually transmitted disease clinic who reported same-sex contact: their HIV seroprevalence and risk behaviors.

OBJECTIVES This study compares characteristics, behaviors, and human immunodeficiency virus (HIV) infection in women who reported same-sex contact and women who had sex only with men. METHODS Participants were patients attending a New York City sexually transmitted disease clinic. Structured questionnaires were administered by interviewers. RESULTS Overall, 9% (135/1518) of women reported same-sex contact; among these, 93% also reported contact with men. Women reporting same-sex contact were more likely than exclusively heterosexual women to be HIV seropositive (17% vs 11%; odds ratio [OR] = 1.7, 95% confidence interval [CI] = 1.0, 2.6), to exchange sex for money/drugs (48% vs 12%, OR = 6.7, 95% CI = 4.6, 9.8), to inject drugs (31% vs 7%, OR = 6.3, 95% CI = 4.1, 9.5), and to use crack cocaine (37% vs 15%, OR = 3.3, 95% CI = 2.2, 4.8). HIV in women reporting same-sex contact was associated with history of syphilis (OR = 8.8), sex for crack (OR = 5.7), and injection drug use (OR = 4.5). CONCLUSIONS In this study, women who reported same-sex contact were predominantly bisexual. They had more HIV risk behaviors and were more often HIV seropositive than women who had sex only with men. Among these bisexual women, heterosexual contact and injection drug use were the most likely sources of HIV. There was no evidence of female-to-female transmission.

Evaluation of school absenteeism data for early outbreak detection, New York City

BackgroundSchool absenteeism data may have utility as an early indicator of disease outbreaks, however their value should be critically examined. This paper describes an evaluation of the utility of school absenteeism data for early outbreak detection in New York City (NYC).MethodsTo assess citywide temporal trends in absenteeism, we downloaded three years (2001–02, 2002–03, 2003–04) of daily school attendance data from the NYC Department of Education (DOE) website. We applied the CuSum method to identify aberrations in the adjusted daily percent absent. A spatial scan statistic was used to assess geographic clustering in absenteeism for the 2001–02 academic year.ResultsModerate increases in absenteeism were observed among children during peak influenza season. Spatial analysis detected 790 significant clusters of absenteeism among elementary school children (p < 0.01), two of which occurred during a previously reported outbreak.ConclusionMonitoring school absenteeism may be moderately useful for detecting large citywide epidemics, however, school-level data were noisy and we were unable to demonstrate any practical value in using cluster analysis to detect localized outbreaks. Based on these results, we will not implement prospective monitoring of school absenteeism data, but are evaluating the utility of more specific school-based data for outbreak detection.

An outbreak of Legionnaires disease associated with a decorative water wall fountain in a hospital.

OBJECTIVE To detect an outbreak-related source of Legionella, control the outbreak, and prevent additional Legionella infections from occurring. DESIGN AND SETTING Epidemiologic investigation of an acute outbreak of hospital-associated Legionnaires disease among outpatients and visitors to a Wisconsin hospital. PATIENTS Patients with laboratory-confirmed Legionnaires disease who resided in southeastern Wisconsin and had illness onsets during February and March 2010. METHODS Patients with Legionnaires disease were interviewed using a hypothesis-generating questionnaire. On-site investigation included sampling of water and other potential environmental sources for Legionella testing. Case-finding measures included extensive notification of individuals potentially exposed at the hospital and alerts to area healthcare and laboratory personnel. RESULTS Laboratory-confirmed Legionnaires disease was diagnosed in 8 patients, all of whom were present at the same hospital during the 10 days prior to their illness onsets. Six patients had known exposure to a water wall-type decorative fountain near the main hospital entrance. Although the decorative fountain underwent routine cleaning and maintenance, high counts of Legionella pneumophila serogroup 1 were isolated from cultures of a foam material found above the fountain trough. CONCLUSION This outbreak of Legionnaires disease was associated with exposure to a decorative fountain located in a hospital public area. Routine cleaning and maintenance of fountains does not eliminate the risk of bacterial contamination. Our findings highlight the need to evaluate the safety of water fountains installed in any area of a healthcare facility.

Diarrheal Illness Detected Through Syndromic Surveillance After a Massive Power Outage: New York City, August 2003

OBJECTIVES We investigated increases in diarrheal illness detected through syndromic surveillance after a power outage in New York City on August 14, 2003. METHODS The New York City Department of Health and Mental Hygiene uses emergency department, pharmacy, and absentee data to conduct syndromic surveillance for diarrhea. We conducted a case-control investigation among patients presenting during August 16 to 18, 2003, to emergency departments that participated in syndromic surveillance. We compared risk factors for diarrheal illness ascertained through structured telephone interviews for case patients presenting with diarrheal symptoms and control patients selected from a stratified random sample of nondiarrheal patients. RESULTS Increases in diarrhea were detected in all data streams. Of 758 patients selected for the investigation, 301 (40%) received the full interview. Among patients 13 years and older, consumption of meat (odds ratio [OR]=2.7, 95% confidence interval [CI]=1.2, 6.1) and seafood (OR=4.8; 95% CI=1.6, 14) between the power outage and symptom onset was associated with diarrheal illness. CONCLUSIONS Diarrhea may have resulted from consumption of meat or seafood that spoiled after the power outage. Syndromic surveillance enabled prompt detection and systematic investigation of citywide illness that would otherwise have gone undetected.

Enhanced drop-in syndromic surveillance in New York City following September 11, 2001.

After the 2001 World Trade Center disaster, the New York City Department of Health was under heightened alert for bioterrorist attacks in the city. An emergency department (ED) syndromic surveillance system was implemented with the assistance of the Centers for Disease Control and Prevention to ensure early recognition of an increase or clustering of disease syndromes that might represent a disease outbreak, whether natural or intentional. The surveillance system was based on data collected 7 days a week at area EDs. Data collected were translated into syndromes, entered into an electronic database, and analyzed for aberrations in space and time within 24 hours. From September 14–27, personnel were stationed at 15 EDs on a 24-hour basis (first staffing period); from September 29–October 12, due to resource limitations, personnel were stationed at 12 EDs on an 18-hour basis (second staffing period). A standardized form was used to obtain demographic information and classify each patient visit into 12 syndrome categories. Seven of these represented early manifestations of bioterrorist agents. Data transfer and analysis for time and space clustering (alarms) by syndrome and age occurred daily. Retrospective analyses examined syndrome trends, differences in reporting between staffing periods, and the staff’s experience during the project. A total of 67,536 reports were received. The system captured 83.9% of patient visits during the first staffing period, and 60.8% during the second staffing period (P<01). Five syndromes each accounted for more than 1% of visits: trauma, asthma, gastrointestinal illness, upper/lower respiratory infection with fever, and anxiety. Citywide temporal alarms occurred eight times for three of the major bioterrorism-related syndromes. Spatial clustering alarms occurred 16 time by hospital location and 9 times by ZIP code for the same three syndromes. No outbreaks were detected. On-site staffing to facilitate data collection and entry, supported by daily analysis of ED visits, is a feasible short-term approach to syndromic surveillance during high-profile events. The resources required to operate such a system, however, cannot be sustained for the long term. This system was changed to an electronic-based ED syndromic system using triage log data that remains in operation.

Failure to control an outbreak of multidrug-resistant Streptococcus pneumoniae in a long-term-care facility: Emergence and ongoing transmission of a fluoroquinolone-resistant strain

OBJECTIVES To characterize risk factors associated with pneumococcal disease and asymptomatic colonization during an outbreak of multidrug-resistant Streptococcus pneumoniae (MDRSP) among AIDS patients in a long-term-care facility (LTCF), evaluate the efficacy of antimicrobial prophylaxis in eliminating MDRSP colonization, and describe the emergence of fluoroquinolone resistance in the MDRSP outbreak strain. DESIGN Epidemiologic investigation based on chart review and characterization of SP strains by antimicrobial susceptibility testing and PFGE and prospective MDRSP surveillance. SETTING An 80-bed AIDS-care unit in an LTCF PARTICIPANTS: Staff and residents on the unit. RESULTS From April 1995 through January 1996, 7 cases of MDRSP occurred. A nasopharyngeal (NP) swab survey of all residents (n=65) and staff (n=70) detected asymptomatic colonization among 6 residents (9%), but no staff. Isolates were sensitive only to rifampin, ofloxacin, and vancomycin. A 7-day course of rifampin and ofloxacin was given to eliminate colonization among residents: NP swab surveys at 1, 4, and 10 weeks after prophylaxis identified 1 or more colonized residents at each follow-up with isolates showing resistance to one or both treatment drugs. Between 1996 and 1999, an additional 6 patients were diagnosed with fluoroquinolone-resistant (FQ-R) MDRSP infection, with PFGE results demonstrating that the outbreak strain had persisted 3 years after the initial outbreak was recognized. CONCLUSIONS Chemoprophylaxis likely contributed to the development of a FQ-R outbreak strain that continued to be transmitted in the facility through 1999. Long-term control of future MDRSP outbreaks should rely primarily on vaccination and strict infection control measures.