from perfect. Although blood culture assays are highly specific, they are insensitive (even under ideal conditions, only about 20 percent of pneumococcal disease can be accurately detected, for example); methods that look at lung aspirates, while producing more sensitive results, are difficult to perform even in the research setting; and because of the problem of frequent carriage, sputum sampling gives relatively nonspecific (and insensitive) information.
The availability of accurate diagnostics would make it possible to define the public health impact of pneumonia, characterize its epidemiology, and would facilitate studies of viral-bacterial interactions in ARI. Diagnostic capability also would give scientists a tool for estimating the potential impacts of conjugate vaccines, a critical factor in convincing developing-country health ministers to include ARI vaccines in their immunization programs. Finally, accurate diagnosis of pneumonia would have a direct, positive impact on the conduct of ARI vaccine studies (Box 3).
BOX 3 Impact of Accurate Diagnosis of Bacterial Pneumonia on Vaccine Trials
SOURCE: Broome, 1994.
A number of laboratories are exploring PCR (polymerase chain reaction) techniques to improve diagnostic accuracy. While some of their results look promising, PCR may not increase sensitivity beyond that achieved with blood culture. Diagnostic methods based on antigen detection were thought to show promise. Some tests using latex agglutination reagents—for example, to identify Hib infection—successfully flag 90 percent or more of bacteremic patients, based on antigens found in their serum or urine. Unfortunately, these sorts of studies have often failed to include a control group of healthy carriers. When that is done, antigen-detection assays frequently are found to misclassify as positive a significant proportion of individuals who are carriers. (Similarly misleading