Info

FIGURE 13.2 Consultation mode of Iliad showing the patient findings that the physician entered (lower right panel), and the differential diagnosis (panel on left) that Iliad generated from those findings. (Courtesy of LE Widman, http://www.informaticamedica.org.br/informaticamedica/n0105/widman.htm.)

FIGURE 13.2 Consultation mode of Iliad showing the patient findings that the physician entered (lower right panel), and the differential diagnosis (panel on left) that Iliad generated from those findings. (Courtesy of LE Widman, http://www.informaticamedica.org.br/informaticamedica/n0105/widman.htm.)

figure 13.4 Log-in page of the web-based bovine syndromic surveillance system.

differential diagnosis for the patient's illness. It also uses value-of-information calculations to recommend additional tests and observations.

Shephard and colleagues developed BOSSS to meet a need to capture disease and syndrome observation data from field observations made by veterinarians, producers, and lay observers. These observations are important to population-based surveillance of cattle herds and to the business of cattle production, but are largely unrecorded and, therefore, unavailable for biosurveillance purposes. BOSSS encourages use by rewarding a user with diagnostic support in his efforts to determine the cause of illness or death in cattle.

Like Iliad, BOSSS contains information on the prevalence of diseases (for approximately 1000 diseases of cattle). It also contains the sensitivity and specificity of each finding for each disease. (We discuss sensitivity and specificity in Sect. 5.1.) The development team compiled this information from veterinary literature and the opinions of veterinary experts. Like Iliad, it

figure 13.5 A graphic user interface for entering findings of a sick animal.
figure 13.6 Differential diagnosis results ranked by posterior probabilities.

uses Bayes rules to determine the posterior probability of each disease given the findings (Figure 13.6).The program promotes the capture of extra information on each case through the use of an interrogation module that presents questions to the user (Figure 13.7). These questions are key differentiating signs for the most likely diseases identified by the system that have not already been recorded by the user. Unlike Iliad, BOSSS has an explicit biosurveillance mission. The "syn-dromic-surveillance" component of BOSSS is illustrated by the mapping of cases shown in Figure 13.8.

The developers of BOSSS, learning lessons from Iliad and other diagnostic expert systems about the importance of fitting

Guided Examination

0 0

Post a comment