Method and system for bio-surveillance detection and alerting

• US 7,249,006 B2
• Filed: 03/23/2001
• Issued: 07/24/2007
• Est. Priority Date: 03/23/2000
• Status: Expired due to Term

• Current Assignee: Johns Hopkins University
• Original Assignee: Johns Hopkins University
• Inventors: Newhall, Bruce K., Lombardo, Joseph S., Loschen, Wayne A., Chotani, Rashid A., Wojcik, Richard A., Burkom, Howard S., Pineda, Fernando J.
• Primary Examiner(s): Moran; Marjorie A.
• Application Number:
  US10/130,404
• Publication Number:
  US 20030009239A1
• Time in Patent Office: 2,314 Days
• Field of Search: 702/19
• US Class Current: 703/2
• CPC Class Codes: G16H 10/60 for patient-specific data, e.g. for electronic patient records
• CPC Class Codes: G16H 50/80 for detecting, monitoring or modelling epidemics or pandemics, e.g. flu
• CPC Class Codes: Y02A 90/10 Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Abstract

Background noise from relevant data sets, including for example over-the-counter sales data, absenteeism data, etc., is subtracted using a background estimation algorithm that outputs residual data. The effects of hypothetical anomalous events, such as a bio-terrorist attack, on the relevant data sets are modeled to create replica data. The replica data may be based on input from epidemiologists and various scenario templates including information on disease manifestation and other intelligence. The residual data and the replica data are then matched using a detector. Types of detectors include for example adaptive matched-filter detectors, change detectors and Bayesian Inference Networks. An alarm is triggered if a real anomalous event similar to a hypothetical anomalous event is detected. A Geographical Information System (GIS) may be used to display data from individual zip codes.

Claims

1. A method for bio-surveillance detection and alerting, comprising the steps of:

performing a center surround algorithm to create residual data related to population health, comprising

computing background noise by averaging background data sets related to population health from different regions neighboring a center region, and

subtracting said background noise from relevant data sets related to population health from the center region to create said residual data;

modeling the effects of a hypothetical anomalous health-impacting event on said relevant data sets to create replica data of different types that simulates the effects of the anomalous health-impacting event on the relevant data sets;

matching said residual data with said replica data using a detector to detect a real anomalous health-impacting event similar to said hypothetical health-impacting anomalous event; and

triggering an alert if a real anomalous health-impacting event similar to said hypothetical anomalous health-impacting event is detected.

2. The method for bio-surveillance detection and alerting as recited in claim 1, wherein said background estimation algorithm includes a Kalman filter.

3. The method for bio-surveillance detection and alerting as recited in claim 1, wherein said detector is an adaptive matched-filter detector.

4. The method for bio-surveillance detection and alerting as recited in claim 1, wherein said detector is a Neyman-Pearson detector.

5. The method for bio-surveillance detection and alerting as recited in claim 1, wherein said detector is a change detector.

6. The method for bio-surveillance detection and alerting as recited in claim 1, wherein said detector includes a Bayesian Inference Network.

7. The method for bio-surveillance detection and alerting as recited in claim 1, wherein said relevant data sets comprise infonnation on specific human behaviors exhibited during the onset of a disease.

8. The method for bio-surveillance detection and alerting as recited in claim 1, wherein said relevant data sets comprise over-the-counter drug sales data.

9. The method for bio-surveillance detection and alerting as recited in claim 1, wherein said relevant data sets comprise absenteeism data.

10. The method for bio-surveillance detection and alerting as recited in claim 1, wherein said relevant data sets comprise emergency room admissions data.

11. The method for bio-surveillance detection and alerting as recited in claim 1, wherein said relevant data sets comprise insurance claims billing data.

12. The method for bio-surveillance detection and alerting as recited in claim 1, wherein said relevant data sets comprise animal health data.

13. The method for bio-surveillance detection and alerting as recited in claim 1, wherein said relevant data sets comprise one or more of over-the-counter drug sales data, absenteeism data, emergency room admissions data, insurance claims billing data, and animal health data.

14. The method for bio-surveillance detection and alerting as recited in claim 1, wherein said relevant data sets comprise data from at least two different data sources.

15. The method for bio-surveillance detection and alerting as recited in claim 1, wherein said relevant data sets comprise data from at least five different simultaneous data sources.

16. The method for bio-surveillance detection and alerting as recited in claim 1, wherein said step of matching said residual data with said replica data uses more than one detector.

17. A method for bio-surveillance detection and alerting as recited in claim 1, further comprising the step of inputting information from a system user concerning a region of interest and a disease of interest.

18. A method for bio-surveillance detection and alerting as recited in claim 1, further comprising the step of sending an electronic message to disease control personnel in a jurisdiction where an alert has been triggered.

19. The method of claim 1, wherein:

said relevant data sets include data sets of different types;

said background data sets include data sets of different types corresponding to said different types of relevant data sets;

said subtracting step produces residual data of different types corresponding to said different types of relevant data sets;

said modeling step produces replica data of different types corresponding to said different types of relevant data sets; and

said matching step includes matching said residual date of different types with said replica data of different types.

20. The method of claim 1, wherein the replica data takes the form of an artificial signal having a magnitude proportional to a number of individuals in said population adversely affected by the health-impacting event, as indicated by said relevant data sets.

21. The method of claim 1, further comprising displaying via an internet web site said relevant data sets and said alert indicating said real anomalous health-impacting event.

22. A bio-surveillance detection and alerting system, comprising:

means for performing a center surround algorithm to create residual data related to population health, comprising

means for computing background noise by averaging background data sets related to population health from different regions neighboring a center region, and

means for subtracting said background noise from relevant data sets related to population health from the center region to create said residual data;

means for modeling the effects of a hypothetical anomalous health-impacting event on said relevant data sets to create replica data that simulates the effect of the anomalous health-impacting event on the relevant data sets; and

means for matching said residual data with said replica data to detect a real anomalous health-impacting event similar to said hypothetical health-impacting anomalous event; and

means for alerting a system user about said real anomalous health-impacting event.