Wild Fish Health Database: Glossary and Warnings About Data Use
Topics
This database has certain limitations that need to be understood
before drawing conclusions from its use.
Important topics related to these limitations are documented below
(along, also, with a glossary of technical terms used in the database).
These data were not collected as part of the National Wild Fish Health Survey.
They were obtained from other agencies, possibly using other methods and standards.
The historical compilation currently includes selected data from only Alaska, Colorado and Idaho.
Similar data exist from other states, but they have not been entered into the database.
The Fish and Wildlife Service does not take responsibility for the quality of the
fish testing in the historical data.
The data in this database record the presence or absence of pathogens
(potential agents of disease) or antibodies to pathogens
(evidence of previous exposure) in fish.
Pathogen or antibody presence in an individual fish does not necessarily indicate
that the fish is suffering from the associated disease, nor does it necessarily
indicate that the local fish population is experiencing effects of the disease.
Absence of the pathogen (or its antibodies) in the results of the testing reported
in this database indicates that the pathogen (or its antibodies) were absent
from the sampled fish that were tested; this does not necessarily mean that the
pathogen is absent from the population.
Clinical signs are indications of disease.
The National Wild Fish Health Survey does not focus on searching for clinical signs,
and the protocols do not require examination for clinical signs.
The database does provide the option for reporting clinical signs if the lab chooses to do so.
Quality Control of Fish Testing-
The fish testing conducted as part of the National Wild Fish Health Survey was done
in the regional Fish and Wildlife Service Fish Health Laboratories according to known
standard protocols, using methods with known performance.
These quality control measures do not apply to the historical data (see above).
Quality Control of Data Entry-
The fish testing data for the National Wild Fish Health Survey were transcribed to
electronic form by the respective Fish and Wildlife Service laboratories that did the testing.
The data entry for the historical data was done by a contractor.
All data in the central computer database are associated with laboratory case numbers.
The case numbers are assigned by the laboratories.
Each laboratory uses the case numbers for tracking data within its own record keeping system.
In the event of questions about particular results from the central database,
the data in the archives of the individual laboratories have ultimate precedence.
Quality Control of the Database and Access System-
The database is still being modified, both to add more data, and to provide new features.
During the course of modification, spot tests are carried out to verify proper functioning.
However, the database and access system has not yet been subjected to systematic testing to
compare against the records of the individual laboratories.
When such systematic testing is done, the details of the test, and the results, will be reported here.
The field collections for the National Wild Fish Health Survey were normally conducted by
Service partners, and were not under the direct control of the Service.
Cooperating partners were typically State, Tribal and other Federal agencies.
In some instances, Service Fish Health Center personnel either lead or assisted in the actual
sample collection and on-site processing.
The locations sampled for the National Wild Fish Health Survey were selected by the cooperators.
There was not a nation-wide design for selecting locations.
The database does record for each collection whether that collection was routine
or motivated by a particular purpose, such as investigating a fish kill, that might
have suggested unusual circumstances.
When whole fish samples are received for processing, the individual fish are sorted into cases
and assigned case history numbers.
Tissues samples are then taken.
Tissue samples may be pooled together from multiple fish before testing or,
the tissue sample from a single fish can be tested as a pool of 1.
The collection, field processing and preparation of tissue samples may be done by
Fish and Wildlife Service Fish Health Laboratories or by cooperators following Service protocols.
Case-
This refers to a batch of fish of all the same species, collected on the same date at a single location.
There could be many individual fish in a single case.
Tissue sample-
This refers to particular organs, or parts (or specific combinations of organs or parts)
that were dissected out and pooled for a particular pathogen test.
For brevity, the word "sample" may be used for "tissue sample".
There may be many different tissues samples prepared from a single case.
A single pathogen may be tested for in more than 1 tissue sample of a single case.
The current list of tissues types is available.
Pool-
This refers to a batch of tissue, from 1 or more individual fish in a case,
submitted as an aggregate for testing.
There could be many pools for a single tissue sample in a single case.
The database offers the option for recording, at the time of data entry,
which individual fish contributed to which particular pool.
Individual Fish Preparation-
This refers to 1 fish prepared for testing 1 tissue for 1 pathogen with
1 set of testing procedures.
This number is always precisely known, unlike the total number of individual fish involved.
It has special meaning when all of the test results are negative.
In this case the number of individual fish preparations relates to the upper
confidence limits for the infection rare in the individual fish (not pools)
of the population in question.
Once the collection is prepared for testing, each pool is tested with a preliminary testing procedure.
If the preliminary test result is positive, a corroborative procedure may also be used.
A single pathogen may be tested for in multiple tissues and with multiple procedures within a single case.
The procedures used for testing for a particular pathogen are normally the best available at the time,
but these may change over time as new procedures are developed and evaluated.
Preliminary procedure-
This is the first method that is applied to all pools to be tested.
The current list of preliminary procedures is available.
Corroborative procedure-
This is the second method that might be used to retest a pool when the
preliminary procedure had a positive test result.
The current list of corroborative methods is available.
Test Result-
This is the outcome of testing 1 pool for 1 pathogen with 1 procedure.
The pool naturally comes from a single tissue sample of a single case.
The procedure may be either a preliminary procedure or a corroborative procedure.
Test results are simply positive (+) or negative (-) for each pool.
The test results are given only on the detailed case information page and only
when individual tissue pools are tracked.
Positive result-
This refers to a result of a test indicating presence of that pathogen, or its antibody, in that pool.
Negative result-
This refers to a result of a test indicating absence of that pathogen, or its antibody, in that pool.
Tabular and map results from a query to the database report the test status for all of the
tissue samples that meet the search criteria for a site, state or hydrologic unit.
Search Criteria-
These include the beginning and ending dates, the database, the fish species
and the pathogens that were selected on the initial query form.
The dates refer to the actual date when the samples were collected in the field.
Samples must meet all of the search criteria to be displayed in any way.
The search criteria are reported and associated with a unique query number.
Query Number-
Each request for a search submitted to the database is assigned a unique query number,
which will appear at the top of all the results web pages (maps as well as tables)
from this search and on the Query Parameters page that archives the specifications of that query.
If the user chooses to download or print results from a search, the
query number provides a tracking mechanism to identify the search
parameters (specified fish species, pathogen, date window, etc)
associated with each table or map report that is generated, provided
the Query Parameters page is saved.
Test Status-
This refers to the outcome of testing all the pools of 1 tissue sample for 1 pathogen with
1 set of preliminary and corroborative procedures.
The tissue sample naturally comes from a single case,
but it may be organized into any number of pools for testing.
The test status may indicate "pathogen detected", "pathogen not detected" or "status inconclusive".
Each line of the "database search results" tables represents the test status of 1 tissue sample.
Each dot on the dot maps also represents the test status of 1 tissue sample,
but depending on the search criteria, many tissue samples may fall in the same location
and thus be displayed as a single dot.
The color of the most positive test status is always the one displayed.
The test status for all of the individual tissue samples that meet the search criteria,
can be seen by clicking on the dot.
Pathogen detected-
This indicates that 1 or more samples met all of the search criteria and that at least one
pool tested positive with a corroborative procedure.
This indicates presence of the pathogen (or its antibodies) in at least 1 individual fish,
at the time of the collection or earlier in the water body where that collection was obtained.
This status is indicated with the color red.
Inconclusive status-
This indicates that 1 or more samples met all of the search criteria and that at least one
pool tested positive with a preliminary procedure, and that at least 1 pool was
tested with a corroborative procedure,
but none of these pools tested positive with the corroborative procedure.
When the preliminary test is positive and the corroborative test is negative,
the preliminary result is likely spurious.
Any samples which tested positive with a preliminary procedure but had no corroborative testing
are suppressed from the database.
These results,however, may be seen in the case details if any other samples in that case qualify.
Inconclusive status is indicated with the color yellow.
Pathogen not detected-
This indicates that 1 or more samples met all of the search criteria and that all of the involved
pools tested negative with the preliminary procedure.
This indicates absence of the pathogen, or of previous exposure to the pathogen,
in the set of fish constituting the tested pools.
This does not, however, prove absence of the pathogen from this fish species in the
body of water where the collection originated, because the collection is a sample,
and other individual fish that were not included in the sample might
still have been positive for the pathogen.
This status is indicated with the color green.
The results generated from any given query of the Database, do not suggest nor imply any
occurrence of the subject pathogen(s) beyond that indicated in query results for the requested
fish/pathogen(s) combination(s).
It is formally possible to construe a defined spatial area, such as a state or a watershed, as a
statistical aggregate, and to interpret the set of test results from collections falling in that area
as a statistical sample, and on this basis to calculate the frequency of that pathogen
in the area as a whole. However, this statistical operation, interpreted in this way, makes
many assumptions that need to be considered by the user carefully.
The most serious assumption is that the locations where the tested collections were taken
were water bodies that are representative of the entire set of water bodies within the defined
area. Evaluation of the validity of such an assumption is a difficult scientific undertaking.
Sampling designs intended to assure representativeness of this sort require dense spatial
coverage of samples and control over the selection of sampling locations. Depending on the
query parameters and when the query is actually conducted there may be insufficient samples
included within the results to validate this assumption.
The database program will compute the fraction of pools testing positive with a
preliminary procedure in the set of collections in a defined area at specified
spatial scale, for the selected pathogen and fish species.
For each such report, the database program will also report confidence limits
on that fraction, treating that fraction as an estimate of the
fraction positive in hypothetical samples from all the water bodies
in the defined spatial area.
The larger the number of pools that form the basis for the estimate,
the narrower will be the calculated confidence limits.
But it must be remembered that the confidence limits calculation assumes
representativeness of the sample, which of course cannot be assured.
For this reason, the calculated confidence limits should realistically be treated as a minimal
measure of the uncertainty.
The actual uncertainty is greater.
All the pools that are tested for a particular pathogen are tested with the preliminary procedure,
so the fraction of pools with positive test results has direct meaning for the frequency of positives,
once the number of fish in the pools, and the false positive rate of the method, are taken into account.
The database program automatically calculates confidence limits on the
fraction of pools that test positive with the preliminary procedure.
Some, but not necessarily all, of the pools that tested positive with the preliminary procedure
may be subsequently submitted for testing with a corroborative procedure.
Tissue samples with any positive test results for the preliminary procedure, but no testing with a
corroborative procedure are excluded from the database.
These results,however, may be seen in the case details if any other samples in that case qualify.
Because of this protocol, the fraction of pools tested with a corroborative procedure that are actually
positive does not, in itself, bear on the frequency of positives.
Rather, the fraction of pools that tested positive with the corroborative procedure is an indication
of the false positive rate of the preliminary procedure.
The sample handling protocol used by the laboratories does not generally allow
a direct tally of the number of fish testing positive, because testing is done on pools of tissue
which often were composited from more than 1 fish.
The number of fish that contributed to a pool is known, but a positive result for testing
of a given pool only indicates that at least 1 of those fish was positive.
If there are more than 1 individual fish contributing to a single pool, the test
results for that pool represent the set of fish, and do not distinguish between individuals.
Thus, if a particular pool tests positive for a pathogen, where more than 1 fish contributed to the
pool, the result does not indicate how many more fish than 1 were positive.
Since the number of fish contributing to each pool is always reported in the
Wild Fish Health Survey data, it is possible to calculate exactly how many total fish
were tested in arriving at a test result for a single pathogen tested in a single tissue
with a single set of procedures.
However, the sampling handling protocol does not always track individual fish through the
pooling process, so if a query concerns multiple pathogens or involves multiple tissues
or multiple sets of procedures, the total number of fish tested in arriving at that result is ambiguous.
When individual fish are not tracked through the pooling processes, it is not possible
to determine from the database whether pools for different tissues and procedures or pools submitted
to tests for different pathogens were drawn from the same fish or from different fish,
so the number of fish involved could be as small as the total for 1 tissue and 1 pathogen
and 1 set of procedures, or as large as the sum of the fish numbers for all the tissues, pathogens
and sets of procedures.
Information would be maximized if all the fish from pools that
test positive with a preliminary procedure were tested individually with a corroborative procedure.
This maximum is now achieved in some cases in the database, where pools
consist of 1 fish each and all the pools that test positive with a preliminary procedure
are tested with a corroborative procedure.
But this is not the general pattern, and it does
not capitalize on the efficiency of testing pools of more than 1 fish with a preliminary
procedure while testing with a corroborative procedure 1 fish at a time (which requires
tracking which individual fish contributed to which pool that was tested with a preliminary procedure).
Constraints on the minimum volume of
tissue required for a test might preclude 1 fish pools in some
instances; and the desire to limit the number of pools tested, for
time or cost reasons, while still testing a large number of fish,
might argue for multiple fish per pool even when testing with a corroborative procedure.
At the time of data entry, the laboratories provide a latitude and
longitude for each collection, and also assign the collection to a state.
There are many instances where state boundaries run through water bodies.
The location of the official boundary is known, and so state assignment could be made
on the basis just of the latitude and longitude information.
However, the state assignments provided by the laboratories take precedence within the database,
for boundary waters, even if this disagrees with the latitude and longitude coordinates.
HUC refers to a "hydrologic unit code" number.
This is a numbering system, adopted by the USGS, that assigns unique numbers to land
areas defined by watershed boundaries.
The system is hierarchical, with smaller units (designated by a larger number of digits in the
code) nested inside larger units (for which the code number is common
with the first few digits of the smaller units it encompasses).
The database present HUC information according to region (2 digit code),
subregion (4 digit code), accounting unit (6 digit code), and cataloging unit (8 digit code).
Sample dots (or circles), which are close to HUC boundaries, could be incorrectly assigned
and inaccurately placed within a HUC.
Therefore, if the exact location is important to the user it should be verified independently
by accessing the applicable case sheet.
The HUC maps used in the database were drawn with a normal resolution of 100 m. However, in rare
instances (e.g., where the HUC boundary is in an area of flat topography and anastomosing
streams, or where stream features are distant from the boundary), the resolution may be as
much as a few km and thus the sample dot could be erroneously, albeit unavoidably, placed
in the wrong HUC.
29 NOV 2000, Last Updated 2 JULY 2001
dlg@rapid.msu.montana.edu