Martin Kulldorff

A spatial scan statistic

The scan statistic is commonly used to test if a one dimensional point process is purely random, or if any clusters can be detected. Here it is simultaneously extended in three directions:(i) a spatial scan statistic for the detection of clusters in a multi-dimensional point process is proposed, (ii) the area of the scanning window is allowed to vary, and (iii) the baseline process may be any inhomogeneous Poisson process or Bernoulli process with intensity pro-portional to some known function. The main interest is in detecting clusters not explained by the baseline process. These methods are illustrated on an epidemiological data set, but there are other potential areas of application as well.

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.

Prospective time periodic geographical disease surveillance using a scan statistic

Most disease registries are updated at least yearly. If a geographically localized health hazard suddenly occurs, we would like to have a surveillance system in place that can pick up a new geographical disease cluster as quickly as possible, irrespective of its location and size. At the same time, we want to minimize the number of false alarms. By using a space–time scan statistic, we propose and illustrate a system for regular time periodic disease surveillance to detect any currently ‘active’ geographical clusters of disease and which tests the statistical significance of such clusters adjusting for the multitude of possible geographical locations and sizes, time intervals and time periodic analyses. The method is illustrated on thyroid cancer among men in New Mexico 1973–1992.

Evaluating cluster alarms: a space-time scan statistic and brain cancer in Los Alamos, New Mexico.

OBJECTIVES This article presents a space-time scan statistic, useful for evaluating space-time cluster alarms, and illustrates the method on a recent brain cancer cluster alarms in Los Alamos, NM. METHODS The space-time scan statistic accounts for the preselection bias and multiple testing inherent in a cluster alarm. Confounders and time trends can be adjusted for. RESULTS The observed excess of brain cancer in Los Alamos was not statistically significant. CONCLUSIONS The space-time scan statistic is useful as a screening tool for evaluating which cluster alarms merit further investigation and which clusters are probably chance occurrences.

Dead Bird Clusters as an Early Warning System for West Nile Virus Activity

An early warning system for West Nile virus (WNV) outbreaks could provide a basis for targeted public education and surveillance activities as well as more timely larval and adult mosquito control. We adapted the spatial scan statistic for prospective detection of infectious disease outbreaks, applied the results to data on dead birds reported from New York City in 2000, and reviewed its utility in providing an early warning of WNV activity in 2001. Prospective geographic cluster analysis of dead bird reports may provide early warning of increasing viral activity in birds and mosquitoes, allowing jurisdictions to triage limited mosquito-collection and laboratory resources and more effectively prevent human disease caused by the virus. This adaptation of the scan statistic could also be useful in other infectious disease surveillance systems, including that for bioterrorism.

Power comparisons for disease clustering tests

Many different methods have been proposed to test for geographical disease clustering, and more generally, for spatial clustering of any type of observations while adjusting for an inhomogeneous background population generating the observations. Despite the many proposed test statistics, there has been few formal comparisons conducted. We present a collection of 1,220,000 simulated benchmark data sets generated under 51 different cluster models and the null hypothesis, to be used for power evaluations. We then use these data sets to compare the power of the spatial scan statistic, the maximized excess events test and the nonparametric M statistic. All have good power, the first having an advantage for localized hot-spot type clusters and the second for global clustering where randomly located cases generate other cases close by. By making the simulated data sets publicly available, new tests can easily be compared with previously evaluated tests by analyzing the same benchmark data.

Measles-Mumps-Rubella-Varicella Combination Vaccine and the Risk of Febrile Seizures

OBJECTIVE: In February 2008, we alerted the Advisory Committee on Immunization Practices to preliminary evidence of a twofold increased risk of febrile seizures after the combination measles-mumps-rubella-varicella (MMRV) vaccine when compared with separate measles-mumps-rubella (MMR) and varicella vaccines. Now with data on twice as many vaccine recipients, our goal was to reexamine seizure risk after MMRV vaccine. METHODS: Using 2000–2008 Vaccine Safety Datalink data, we assessed seizures and fever visits among children aged 12 to 23 months after MMRV and separate MMR + varicella vaccines. We compared seizure risk after MMRV vaccine to that after MMR + varicella vaccines by using Poisson regression as well as with supplementary regressions that incorporated chart-review results and self-controlled analyses. RESULTS: MMRV vaccine recipients (83 107) were compared with recipients of MMR + varicella vaccines (376 354). Seizure and fever significantly clustered 7 to 10 days after vaccination with all measles-containing vaccines but not after varicella vaccination alone. Seizure risk during days 7 to 10 was higher after MMRV than after MMR + varicella vaccination (relative risk: 1.98 [95% confidence interval: 1.43–2.73]). Supplementary analyses yielded similar results. The excess risk for febrile seizures 7 to 10 days after MMRV compared with separate MMR + varicella vaccination was 4.3 per 10 000 doses (95% confidence interval: 2.6–5.6). CONCLUSIONS: Among 12- to 23-month-olds who received their first dose of measles-containing vaccine, fever and seizure were elevated 7 to 10 days after vaccination. Vaccination with MMRV results in 1 additional febrile seizure for every 2300 doses given instead of separate MMR + varicella vaccines. Providers who recommend MMRV should communicate to parents that it increases the risk of fever and seizure over that already associated with measles-containing vaccines.

Real-time vaccine safety surveillance for the early detection of adverse events.

Background:Rare but serious adverse events associated with vaccines or drugs are often nearly impossible to detect in prelicensure studies and require monitoring after introduction of the agent in large populations. Sequential testing procedures are needed to detect vaccine or drug safety problems as soon as possible after introduction. Objective:To develop and evaluate a new real-time surveillance system that uses dynamic data files and sequential analysis for early detection of adverse events after the introduction of new vaccines. Research Design:The Centers for Disease Control and Prevention (CDC)-sponsored Vaccine Safety Datalink Project developed a real-time surveillance system and initiated its use in an ongoing study of a new meningococcal vaccine for adolescents. Dynamic data files from 8 health plans were updated and aggregated for analysis every week. The analysis used maximized sequential probability ratio testing (maxSPRT), a new signal detection method that supports continuous or time-period analysis of data as they are collected. Results:Using the new real-time surveillance system, ongoing analyses of meningococcal conjugate vaccine (MCV) safety are being conducted on a weekly basis. Two forms of maxSPRT were implemented: an analysis using concurrent matched controls, and an analysis based on expected counts of the outcomes of interest, which were estimated based on historical data. The analysis highlights both theoretical and operational issues, including how to (1) choose appropriate outcomes and stopping rules, (2) select control groups, and (3) accommodate variation in exposed:unexposed ratios between time periods and study sites. Conclusions:Real-time surveillance combining dynamic data files, aggregation of data, and sequential analysis methods offers a useful and highly adaptable approach to early detection of adverse events after the introduction of new vaccines.

Meat, Meat Cooking Methods and Preservation, and Risk for Colorectal Adenoma

Cooking meat at high temperatures produces heterocyclic amines (HCAs) and polycyclic aromatic hydrocarbons (PAHs). Processed meats contain N-nitroso compounds. Meat intake may increase cancer risk as HCAs, PAHs, and N-nitroso compounds are carcinogenic in animal models. We investigated meat, processed meat, HCAs, and the PAH benzo(a)pyrene and the risk of colorectal adenoma in 3,696 left-sided (descending and sigmoid colon and rectum) adenoma cases and 34,817 endoscopy-negative controls. Dietary intake was assessed using a 137-item food frequency questionnaire, with additional questions on meats and meat cooking practices. The questionnaire was linked to a previously developed database to determine exposure to HCAs and PAHs. Intake of red meat, with known doneness/cooking methods, was associated with an increased risk of adenoma in the descending and sigmoid colon [odds ratio (OR), 1.26; 95% confidence interval (95% CI), 1.05-1.50 comparing extreme quintiles of intake] but not rectal adenoma. Well-done red meat was associated with increased risk of colorectal adenoma (OR, 1.21; 95% CI, 1.06-1.37). Increased risks for adenoma of the descending colon and sigmoid colon were observed for the two HCAs: 2-amino-3,8-dimethylimidazo[4,5]quinoxaline and 2-amino-1-methyl-6-phenylimidazo[4,5]pyridine (OR, 1.18; 95% CI, 1.01-1.38 and OR, 1.17, 95% CI, 1.01-1.35, respectively) as well as benzo(a)pyrene (OR, 1.18; 95% CI, 1.02-1.35). Greater intake of bacon and sausage was associated with increased colorectal adenoma risk (OR, 1.14; 95% CI, 1.00-1.30); however, total intake of processed meat was not (OR, 1.04; 95% CI, 0.90-1.19). Our study of screening-detected colorectal adenomas shows that red meat and meat cooked at high temperatures are associated with an increased risk of colorectal adenoma.

Intussusception Risk after Rotavirus Vaccination in U.S. Infants

BACKGROUND International postlicensure studies have identified an increased risk of intussusception after vaccination with the second-generation rotavirus vaccines RotaTeq (RV5, a pentavalent vaccine) and Rotarix (RV1, a monovalent vaccine). We studied this association among infants in the United States. METHODS The study included data from infants 5.0 to 36.9 weeks of age who were enrolled in three U.S. health plans that participate in the Mini-Sentinel program sponsored by the Food and Drug Administration. Potential cases of intussusception and vaccine exposures from 2004 through mid-2011 were identified through procedural and diagnostic codes. Medical records were reviewed to confirm the occurrence of intussusception and the status with respect to rotavirus vaccination. The primary analysis used a self-controlled risk-interval design that included only vaccinated children. The secondary analysis used a cohort design that included exposed and unexposed person-time. RESULTS The analyses included 507,874 first doses and 1,277,556 total doses of RV5 and 53,638 first doses and 103,098 total doses of RV1. The statistical power for the analysis of RV1 was lower than that for the analysis of RV5. The number of excess cases of intussusception per 100,000 recipients of the first dose of RV5 was significantly elevated, both in the primary analysis (attributable risk, 1.1 [95% confidence interval, 0.3 to 2.7] for the 7-day risk window and 1.5 [95% CI, 0.2 to 3.2] for the 21-day risk window) and in the secondary analysis (attributable risk, 1.2 [95% CI, 0.2 to 3.2] for the 21-day risk window). No significant increase in risk was seen after dose 2 or 3. The results with respect to the primary analysis of RV1 were not significant, but the secondary analysis showed a significant risk after dose 2. CONCLUSIONS RV5 was associated with approximately 1.5 (95% CI, 0.2 to 3.2) excess cases of intussusception per 100,000 recipients of the first dose. The secondary analysis of RV1 suggested a potential risk, although the study of RV1 was underpowered. These risks must be considered in light of the demonstrated benefits of rotavirus vaccination. (Funded by the Food and Drug Administration.).