Making organism identification cost-effective

Medical Laboratory Observer, March, 1986 by Michael T. Kelly

Making organism identification cost-effective

Budget constraints have deepened microbiologists' interest in cost-effective approaches to organism identification. Too often, however, the test that has the lowest supply, labor, and equipment costs wins out automatically.

Two other factors must also be considered in evaluating the cost-effectiveness of alternative methods: accuracy and impact on patient care. An error rate below 10 per cent and results in less than five hours are desirable features of an organism identification system, along with low cost.

At a minimum, the system should accurately identify more than 90 per cent of organisms likely to be encountered in clinical laboratories. Many systems today have accuracy in the range of 95 per cent, which is as good as can be achieved even with reference methods performed repeatedly.

As for turnaround time, rapid organism ID can shorten hospital stays, head off unnecessary or inappropriate antibiotic therapy, and prevent unnecessary repeat cultures. Some observers have questioned the ability of rapid methods to reduce length of stay. Rapid testing, however, can shorten hospital stays if results are conveyed without delay--by rapid reporting--and if clinicians make timely use of the data to institute appropriate therapy.

Five hours is an arbitrary but useful standard for rapid organism ID through newer techniques, as opposed to the 24 to 48 hours required by conventional methods. Individual organisms might sometimes even be identified in minutes, but no system enables microbiologists to identify a broad range of organisms such as Enterobacteriaceae in so short a time.

Rapid identification methods directly applicable to clinical specimens include fluorescent antibody tests (for Bacteroides, pertussis, Legionella); enzyme immunoassays (N. gonorrhoeae, rotavirus); and particle agglutination tests (group A strep, H. influenzae, N. meningitidis, S. pneumoniae, etc.). Reagents of this type are available for only a limited number of organisms.

Rapid identification methods requiring isolated colonies include spot biochemical tests, particle agglutination tests, and more elaborate commercial systems. It is most cost-effective to use colony morphology and the spot tests first. The latter apply only to a limited number of organisms, but they are a substantial portion of the organisms encountered in clinical labs. For every E. coli that can be identified by a spot test, the lab can avoid spending $2 to $3 on identification by a commercial system. Spot tests are 98 to 99 per cent accurate for P. aeruginosa, E. coli, and P. mirabilis.

As we noted, rapid organism identification without rapid reporting is not likely to improve patient care. While computerized reporting is the best approach, the results must be available at times when clinicians will see them. Traditionally, clinicians make rounds early in the morning and late in the afternoon. They do not see results issued at 9 a.m. until 5 or 6 p.m. Labs must get reports to the charts first thing in the morning whenever possible. Labs may also have to train clinicians to check the charts for later reports.

Laboratory staffing patterns may also limit the usefulness of rapid tests. Laboratories with eight-hour coverage, for example, usually set up identifications at 2 or 3 in the afternoon; technologists go home and read results the next morning. Thus the advantage of having results available in five hours or less is lost. If eight-hour staffing cannot be extended, it may be better to use a traditional overnight method--or use one of the few rapid instruments that reports results automatically into a computer system.

Hospital size also bears on selection of an organism identification system. In an institution with less than 500 beds, it will probably be most cost-effective to perform spot tests for the most commonly occurring organisms and identify the rest with manual commercial systems. In fact, spot tests tend to reduce costs in a hospital setting of virtually any size. When there are 500 or more beds, the workload may be great enough to also justify replica plating methods or an automated system.

Adequate laboratory staffing also influences the selection of identification systems. If the workload is less than or equal to 35 cultures per technologist per day, an inexpensive manual system may be the most cost-effective. With heavier workloads per technologist, labor becomes a limiting factor, and labor-saving automated systems may become more cost-effective.

In Figure I, the factors of cost, accuracy, and speed are used to evaluate different methods for identification of Enterobacteriaceae, on which more than 90 per cent of commercial systems are used. All of the systems listed, except one, have greater than 90 per cent identification accuracy and therefore are acceptable. Spot tests are recommended because they cost less than $1 and are highly accurate and rapid. Classic biochemical test media--the reference method--require overnight incubation and are labor-intensive. The cost of materials, although reasonable at $1 to $2, is higher than in spot tests.