Ahead of Print -Refining Historical Limits Method to Improve Disease Cluster Detection, New York City, New York, USA - Volume 21, Number 2—February 2015 - Emerging Infectious Disease journal - CDC

full-text ►

Ahead of Print -Refining Historical Limits Method to Improve Disease Cluster Detection, New York City, New York, USA - Volume 21, Number 2—February 2015 - Emerging Infectious Disease journal - CDC

Volume 21, Number 2—February 2015

Research

Refining Historical Limits Method to Improve Disease Cluster Detection, New York City, New York, USA

On This Page

• Methods
• Results
• Discussion
• Suggested Citation

Figures

• Figure 1
• Figure 2
• Figure 3
• Figure 4

Tables

• Table 1
• Table 2
• Table 3
• Table 4

Technical Appendicies

• Technical Appendix

Downloads

• RIS[TXT - 2 KB]

Alison Levin-Rector1 , Elisha L. Wilson2, Annie D. Fine, and Sharon K. Greene

Author affiliations: New York City Department of Health and Mental Hygiene, Queens, New York, USA

Suggested citation for this article

Abstract

Since the early 2000s, the Bureau of Communicable Disease of the New York City Department of Health and Mental Hygiene has analyzed reportable infectious disease data weekly by using the historical limits method to detect unusual clusters that could represent outbreaks. This method typically produced too many signals for each to be investigated with available resources while possibly failing to signal during true disease outbreaks. We made method refinements that improved the consistency of case inclusion criteria and accounted for data lags and trends and aberrations in historical data. During a 12-week period in 2013, we prospectively assessed these refinements using actual surveillance data. The refined method yielded 74 signals, a 45% decrease from what the original method would have produced. Fewer and less biased signals included a true citywide increase in legionellosis and a localized campylobacteriosis cluster subsequently linked to live-poultry markets. Future evaluations using simulated data could complement this descriptive assessment.

Detecting aberrant clusters of reportable infectious disease quickly and accurately enough for meaningful action is a central goal of public health institutions (1–3). Clinicians’ reports of suspected clusters of illness remain critical for surveillance (4), but the application of automated statistical techniques to detect possible outbreaks that might otherwise not be recognized has become more common (5). These techniques are particularly important in jurisdictions that serve large populations and receive a high volume of reports because manual review and investigation of all reports are not feasible.

Challenges such as lags in reporting and case classification and discontinuities in surveillance case definitions, reporting practices, and diagnostic methods are common across jurisdictions. These factors can impede the timely detection of disease clusters. Statistically and computationally simple methods, including historical limits (6), a log-linear regression model (7), and cumulative sums (8), each have strengths and weaknesses for prospective cluster detection, but none adequately address these common data challenges. As technology advances, statistically and computationally intensive methods have been developed (2,3,5,9–12), and although these methods might successfully correct for biases, many lack the ease of implementation and interpretation desired by health departments.

Figure 1. Following Stroup et al. (21), a schematic of the periods included in analyses using the historical limits method.

Since 1989, the US Centers for Disease Control and Prevention has applied the historical limits method (HLM) to disease counts and displayed the results in Figure 1 of the Notifiable Diseases and Mortality Tables in the Morbidity and Mortality Weekly Report (13). Because the method relies on a straightforward comparison of the number of reported cases in the current 4-week period with comparable historical data from the preceding 5 years, its major strengths include simplicity, interpretability, and implicit accounting for seasonal disease patterns. These strengths make it a potentially very useful aberration-detection method for health departments (12,14–18). The Bureau of Communicable Disease (BCD) of the New York City (NYC) Department of Health and Mental Hygiene (DOHMH) implemented the HLM in the early 2000s (HLMoriginal) as a weekly analysis for all reportable diseases for which at least 5 years of historical data were available.

In HLMoriginal, 4 major causes of bias existed: 1) inconsistent case inclusion criteria between current and historical data; 2) lack of adjustment in historical data for gradual trends; 3) lack of adjustment in historical data for disease clusters or aberrations; and 4) no consideration of reporting delays and lags in data accrual. Our objectives were to develop refinements to the HLM (HLMrefined) that preserved the simplicity of the method’s output and improved its validity and to characterize the performance of the refined method using actual reportable disease surveillance data. Although we describe the specific process for refining BCD’s aberration-detection method, the issues presented are common across jurisdictions, and the principles and results are likely to be generalizable.

Ms. Levin-Rector is a public health analyst within the Center for Justice, Safety and Resilience at RTI International. Her primary research interests are developing or improving upon existing statistical methods for analyzing public health data.

Acknowledgments

We thank the members of the analytic team who work to detect disease clusters each week, including Ana Maria Fireteanu, Deborah Kapell, and Stanley Wang. We also thank Nimi Kadar who contributed substantially to the original SAS code for this method.

A.L.R., E.L.W., and S.K.G. were supported by the Public Health Emergency Preparedness Cooperative Agreement (grant 5U90TP221298-08) from the Centers for Disease Control and Prevention. A.D.F. was supported by New York City tax levy funds. The authors declare no conflict of interest.

References

1. Hutwagner L, Thompson W, Seeman GM, Treadwell T. The bioterrorism preparedness and response Early Aberration Reporting System (EARS). J Urban Health. 2003;80(Suppl 1):i89–96 .PubMed
2. Farrington P, Andrews N. Outbreak detection: application to infectious disease surveillance. In: Brookmeyer R, Stroup DF, editors. Monitoring the health of populations. New York: Oxford University Press; 2004. p. 203–31.
3. Choi BY, Kim H, Go UY, Jeong J-H, Lee JW. Comparison of various statistical methods for detecting disease outbreaks. Comput Stat.2010;25:603–17. DOI
4. Schuman SH. When the community is the “patient”: clusters of illness. environmental epidemiology for the busy clinician. London: Taylor & Francis;1997.
5. Unkel S, Farrington CP, Garthwaite PH. Statistical methods for the prospective detection of infectious disease outbreaks: a review. J R Stat Soc Ser A Stat Soc. 2012;175:49–82. DOI
6. Stroup DF, Williamson GD, Herndon JL, Karon JM. Detection of aberrations in the occurrence of notifiable diseases surveillance data. Stat Med.1989;8:323–9. DOIPubMed
7. Farrington CP, Andrews NJ, Beale D, Catchpole MA. A statistical algorithm for the early detection of outbreaks of infectious disease. J R Stat Soc Ser A Stat Soc. 1996;159:547–63. DOI
8. Hutwagner LC, Maloney EK, Bean NH, Slutsker L, Martin SM. Using laboratory-based surveillance data for prevention: an algorithm for detectingSalmonella outbreaks. Emerg Infect Dis. 1997;3:395–400 . DOIPubMed
9. Strat YL. Overview of temporal surveillance. In: Lawson AB, Kleinman K, editors. Spatial and syndromic surveillance for public health. Chichester (UK): John Wiley & Sons; 2005. p. 13–29.
10. Serfling RE. Methods for current statistical analysis of excess pneumonia-influenza deaths. Public Health Rep. 1963;78:494–506. DOIPubMed
11. Noufaily A, Enki DG, Farrington P, Garthwaite P, Andrews N, Charlett A. An improved algorithm for outbreak detection in multiple surveillance systems. Stat Med. 2013;32:1206–22. DOIPubMed
12. Wharton M, Price W, Hoesly F, Woolard D, White K, Greene C, Evaluation of a method for detecting outbreaks of diseases in six states. Am J Prev Med. 1993;9:45–9 .PubMed
13. Centers for Disease Control and Prevention. Proposed changes in format for presentation of notifiable disease report data. MMWR Morb Mortal Wkly Rep. 1989;38:805–9 .PubMed
14. Centers for Disease Control and Prevention. Notes from the field: Yersinia enterocolitica infections associated with pasteurized milk—southwestern Pennsylvania, March–August, 2011. MMWR Morb Mortal Wkly Rep. 2011;60:1428 .PubMed
15. Rigau-Pérez JG, Millard PS, Walker DR, Deseda CC, Casta-Velez A. A deviation bar chart for detecting dengue outbreaks in Puerto Rico. Am J Public Health. 1999;89:374–8. DOIPubMed
16. Pervaiz F, Pervaiz M, Abdur Rehman N, Saif U. FluBreaks: early epidemic detection from Google flu trends. J Med Internet Res. 2012;14:e125.DOIPubMed
17. Winscott M, Betancourt A, Ereth R. The use of historical limits method of outbreak surveillance to retrospectively detect a syphilis outbreak among American Indians in Arizona. Sex Transm Infect. 2011;87:A165. DOI
18. Hutwagner L, Browne T, Seeman GM, Fleischauer AT. Comparing aberration detection methods with simulated data. Emerg Infect Dis.2005;11:314–6. DOIPubMed
19. New York City Department of Health and Mental Hygiene. Communicable disease surveillance data [cited 2013 Nov 15].http://www.nyc.gov/html/doh/html/data/cd-epiquery.shtml
20. Nguyen TQ, Thorpe L, Makki HA, Mostashari F. Benefits and barriers to electronic laboratory results reporting for notifiable diseases: the New York City Department of Health and Mental Hygiene experience. Am J Public Health. 2007;97(Suppl 1):S142–5. DOIPubMed
21. Stroup DF, Wharton M, Kafadar K, Dean AG. Evaluation of a method for detecting aberrations in public health surveillance data. Am J Epidemiol.1993;137:373–80 .PubMed
22. United Hospital Fund. Neighborhoods. New York City community health atlas: sources, methods and definitions. New York: United Hospital Fund;2002. p. 2–3.
23. Centers for Disease Control and Prevention. Comparison of provisional with final notifiable disease case counts—National Notifiable Diseases Surveillance System, 2009. MMWR Morb Mortal Wkly Rep. 2013;62:747–51 .PubMed
24. Lotze T, Shmueli G, Yahav I. Simulating multivariate syndromic time series and outbreak signatures [cited 2014 Dec 3].http://papers.ssrn.com/sol3/papers.cfm?abstract_id=990020
25. Centers for Disease Control and Prevention. Simulation data sets for comparison of aberration detection methods. 2004 April 16, 2004 [cited 2013 Aug 30]. http://www.bt.cdc.gov/surveillance/ears/datasets.asp
26. Layton M. Increase in reported legionellosis cases. 2013 [cited 2013 Sep 18]. https://a816-health29ssl.nyc.gov/sites/NYCHAN/Lists/AlertUpdateAdvisoryDocuments/2013-07-03%20HAN_Legionella%20final2.pdf
27. Kulldorff M, Heffernan R, Hartman J, Assuncao R, Mostashari F. A space–time permutation scan statistic for disease outbreak detection. PLoS Med.2005;2:e59. DOIPubMed

Figures

• Figure 1. Following Stroup et al. (21), a schematic of the periods included in analyses using the historical limits method.
• Figure 2. Confirmatory proportion of pending cases for diseases with any pending cases, New York City, New York, USA, July–December 2012. The confirmatory proportion was defined as the proportion of initially...
• Figure 3. Unadjusted and adjusted weekly citywide counts of campylobacteriosis cases to illustrate adjustment for a linear trend in historical data, New York City, New York, USA, November 2006–October 2011.
• Figure 4. Unadjusted and adjusted 4-week moving sum of citywide dengue fever cases to illustrate adjustment for outliers in historical data, New York City, New York, USA, November 2006–October 2011.

Tables

• Table 1. Case statuses in current and baseline periods included in HLMoriginal and HLMrefined, New York City, New York, USA
• Table 2. Diseases included in analyses using HLMrefined and details of customizations, New York City, New York, USA, May 20–August 5, 2013
• Table 3. Geographic resolution of signals produced by HLMoriginal and HLMrefined in 12 weekly analyses, New York City, New York, USA, May 20–August 5, 2013
• Table 4. Explanation of signals produced by HLMoriginal and HLMrefined in the 12 weekly analyses, New York City, New York, USA, May 20–August 5, 2013.

Technical Appendix

• Technical Appendix. Supplemental information, SAS code, and sample output for the refined historical limits method, New York City, New York, USA.  543 KB

Suggested citation for this article: Levin-Rector A, Wilson EL, Fine AD, Greene SK. Refining historical limits method to improve disease cluster detection, New York City, New York, USA. Emerg Infect Dis [Internet]. 2015 Feb [date cited]. http://dx.doi.org/10.3201/eid2102.140098

DOI: 10.3201/eid2102.140098

1Current affiliation: RTI International, Research Triangle Park, North Carolina, USA

2Current affiliation: Colorado Department of Public Health and Environment, Denver, Colorado, USA