Epidemiology in Action: Using Active Surveillance Systems to Profile Marine Animal Diseases
IAAAM Archive
Stephanie Wong; William Van Bonn; Cynthia Smith; Eric Jensen; Carrie Lomax; Sam Ridgway
United States Navy Marine Mammal Program
San Diego, CA, USA


The purpose of a surveillance system is not simply to collect data, but to turn targeted information into action. When used appropriately, standardized surveillance systems can be an invaluable means of monitoring the health of marine animal populations. While clinicians are well versed in case-based medicine, there is also a need to continually hunt for underlying factors that may represent health risks to marine animal populations. As part of its vigilant preventive medicine program, the U.S. Navy Marine Mammal Program (NMMP) uses ongoing and active surveillance to monitor the long-term health of Tursiops truncatus (bottlenose dolphin) and Zalophus californianus (California sea lion) populations. Below, we provide eight basic steps used by the NMMP for population disease surveillance; these steps can be implemented in any marine animal facility.

Why surveillance?

Subtle but potentially high risk factors for disease in marine animal populations include changes in climate, husbandry, animals, and microbes (e.g., antimicrobial-resistant bacteria). Such risk factors for disease can manifest over long periods of time (months to years) and may be difficult to detect by monitoring individual animals. For example, infection with an undetected virus may have caused acute Staphylococcus aureus sepsis in one animal and disseminated Aspergillus fumigatus infection in another. While treatment for Staphylococcus and Aspergillus infection may have affected the outcome in two individual cases, recognition and treatment of the underlying viral infection could lead to significantly decreased morbidity throughout the population. Therefore, marrying case-based medicine with population-based surveillance can be a powerful step toward decreasing economic and morbidity-related costs in marine animal populations.

Eight steps to implementing an active surveillance system

 Step 1. Identify questions. Identify questions of clinical or economic importance that a surveillance system can address.

 Step 2. Define the population of interest. As priorities and resources dictate, define the surveillance population.

 Step 3. Select measures of interest. Measures of interest that may be included in active surveillance systems are animal demographics (e.g., species, age, gender, origin, location); animal health (e.g., blood values, clinical observations, medical treatments); isolation of microbes (e.g., species, source, antimicrobial susceptibility); husbandry practices (e.g., diet, animal movement); and environmental factors (e.g., water temperature, turbidity, and salinity).

 Step 4. Define measures of interest. To increase the reliability and validity of a surveillance system, specific measures may need to be defined. For example, if a surveillance system is intended to monitor sick versus healthy animals, 'sick' and 'healthy' will need to be defined.

 Step 5. Develop a data repository. All surveillance systems require a repository to enter and store information. One user-friendly tool for entering and storing data is the 'Make View' function of the Centers for Disease Control and Prevention's statistical software, EpiInfo2000.1 The Make View function allows one to develop electronic questionnaires for user-friendly data entry. This software is free and can be downloaded from the Internet.

 Step 6. Standardize data collection and data entry methodologies. The reliability of data analysis and the validity of results greatly depend upon how well data collection and entry methodologies are standardized. Whenever possible, create drop-field categories in the electronic questionnaire (e.g., under 'Diagnosis,' force the user to choose a category--such as enteritis--instead of having an open-text field).

 Step 7. Generate standardized algorithms and analyze data. There are several statistical software packages available for data analysis. The U.S. Navy Marine Mammal Program uses three software programs. All of these programs are compatible with Microsoft Excel.

 EpiInfo2000.1 This software is free, can be downloaded from the Internet, and is relatively user-friendly. It is ideal for simple, descriptive and univariate analyses.

 SAS, Release 8.2.2 SAS is ideal for complex analyses involving large databases (records of thousands to tens of thousands). While it is an incredibly powerful statistical tool, there is a steep learning curve for users and costs range from $5,000 to $8,000.

 Miner3D Software, Release Relatively new to the market, Miner3D uses creative and effective three-dimensional means of communicating data. Impressive color, spatial, text, and sound-based graphs are translated directly from Excel spreadsheets. The cost for one user varies from $150 to $600.

 Step 8. Develop and disseminate routine reports that feed data-driven decisions. Routine data reporting helps clinical and administrative teams understand baseline population data, monitor trends over time, and recognize abnormal events. Interpretation of surveillance data by the clinical team will guide data-driven decision making to reduce morbidity, including potential changes in the population's diet, the environment, diagnostics, and clinical care.

An Example of Active Surveillance at the U.S. Navy Marine Mammal Program

Step 1. Identify Questions

What is the leading cause of morbidity in the NMMP Tursiops population?

Step 2. Define the Population of Interest

All Tursiops at the NMMP from which opportunistic samples and data are available.

Step 3. Select Measures of Interest

Animal demographics, CBC and chemistry values, microbial isolation data, medical treatment profiles, and observations from the clinical veterinarians.

Step 4. Define Measures of Interest

'Morbidity' = a condition, in a Tursiops housed at NMMP, that involves clinical signs observed by a veterinarian and at least one CBC or chemistry parameter outside the normal range (outside of 'normal'= greater than 1.5 standard deviations from the population mean).

Step 5. Develop a Data Repository

Electronic questionnaires, including fields for all the parameters outlined above, were created using EpiInfo 2000 software.

Step 6. Standardize Data Collection and Data Entry Methodologies

The clinical team developed standardized methodologies for all blood and tissue sample collection, sample storage and submission, and data entry.

Step 7. Generate Standardized Algorithms and Analyze Data

Data entered into EpiInfo2000 are routinely exported into SAS, Release 8.2 for data analysis. On a monthly basis, standardized algorithms are used to analyze all parameters in each updated morbidity dataset. Appropriate monthly tables are created in SAS, exported into Excel, and graphed using Miner3D software.

Step 8. Develop and Disseminate Routine Reports for Data-Driven Decisions

Monthly reports, including the number, gender, age, and location of animals affected; the frequency of clinical signs and diagnoses; the frequency and duration of medical treatments; microbial profiles from Tursiops samples; and trends in veterinary observations, are distributed to the clinical team for review and interpretation (pending).

Resources for Statistical / Surveillance Software

1.  EpiInfo2000. Centers for Disease Control and Prevention: http://www.cdc.gov/epiinfo/index.htm

2.  SAS Release 8.2, SAS, Inc. http://www.sas.com

3.  Miner3D Excel. http://miner3d.com


This work is graciously supported by the Biosciences Division, SSC San Diego, Space and Naval Warfare Systems Center, under the Department of Defense. The authors thank all the NMMP veterinary technicians who diligently collect routine samples for the health of our animals and the many students who have helped turn thousands of paper records into a thorough and powerful electronic database.

Speaker Information
(click the speaker's name to view other papers and abstracts submitted by this speaker)

Stephanie K. Wong

MAIN : Immunology, Epidemiology : Epidemiology In Action
Powered By VIN