Tweet
Surveillance- How many H1N1 cases are there????
I think that it is time to revisit the surveillance pyramid concept.
In any system of case detection and reporting in a free-living population, actual reported case are ALWAYS an underestimate of true cases. The ability to extrapolate the ?true? number from the reported number is difficult and has variable utility.
The ratio of reported cases to the true number of cases depends upon for example: how rare the occurrence of illness is; how unusual the symptoms are; how aware people are of the illness; how severe the symptoms are; access to medical care; quality of care; how cases are defined; consistency of diagnostic procedures among caregivers; access to laboratory diagnostic support; diagnostic test characteristics (such as sensitivity, specificity, etc); reporting requirements and procedures; staffing levels in laboratories and care facilities; and finally, but not exhaustively, surveillance systems.
As you can imagine, even within a small region of one country, these variables may differ significantly and vary over time and space.
As an illustration of how surveillance methods can alter the number of reported case of illness, I will refer the interested reader to 2 bar graphs within the ?Connecticut Epidemiologist? newsletter from Sept. 2006.
These folks have been monitoring tickborne infections for years using multiple methods of surveillance. This short newsletter nicely illustrates 2 points about surveillance: the more sources you use for reporting cases, the greater the potential for detecting cases, and actively seeking cases yields higher numbers than passively waiting for cases to be reported.
The first bar graph on page 13 shows the enhanced number of cases of Lyme disease detected when laboratory based surveillance methods are added to physician reports. For the years both methods are utilized (1994 ? 2002), the number of reported cases almost double.
In the second bar graph on page 15, the effect of active case finding for another tick-borne disease is shown: during the active surveillance years (1997 ? 1999), the number of cases doubles to almost triples over the baseline number of passively-reported cases.
As you can see, reported cases are only a rough approximation of where the illness is occurring given the ability, resources, will and infrastructure to detect it.
Nothing more.
In any system of case detection and reporting in a free-living population, actual reported case are ALWAYS an underestimate of true cases. The ability to extrapolate the ?true? number from the reported number is difficult and has variable utility.
The ratio of reported cases to the true number of cases depends upon for example: how rare the occurrence of illness is; how unusual the symptoms are; how aware people are of the illness; how severe the symptoms are; access to medical care; quality of care; how cases are defined; consistency of diagnostic procedures among caregivers; access to laboratory diagnostic support; diagnostic test characteristics (such as sensitivity, specificity, etc); reporting requirements and procedures; staffing levels in laboratories and care facilities; and finally, but not exhaustively, surveillance systems.
As you can imagine, even within a small region of one country, these variables may differ significantly and vary over time and space.
As an illustration of how surveillance methods can alter the number of reported case of illness, I will refer the interested reader to 2 bar graphs within the ?Connecticut Epidemiologist? newsletter from Sept. 2006.
These folks have been monitoring tickborne infections for years using multiple methods of surveillance. This short newsletter nicely illustrates 2 points about surveillance: the more sources you use for reporting cases, the greater the potential for detecting cases, and actively seeking cases yields higher numbers than passively waiting for cases to be reported.
The first bar graph on page 13 shows the enhanced number of cases of Lyme disease detected when laboratory based surveillance methods are added to physician reports. For the years both methods are utilized (1994 ? 2002), the number of reported cases almost double.
In the second bar graph on page 15, the effect of active case finding for another tick-borne disease is shown: during the active surveillance years (1997 ? 1999), the number of cases doubles to almost triples over the baseline number of passively-reported cases.
As you can see, reported cases are only a rough approximation of where the illness is occurring given the ability, resources, will and infrastructure to detect it.
Nothing more.