Background: Rural hospitals generally lack staffing with infectious disease specialists or pathologists. Without on-site pathologists, the range of microbiology services offered by clinical laboratories may be limited as well. Objective: To study the feasibility of using static-image telepathology to evaluate Gram stains of microbiologic preparations. Materials and Methods: In this retrospective feasibility study, three pathologists evaluated Gram stains of slides from 50 cases by two viewing modalities: static-image telepathology and conventional light microscopy. Digital video images of slides were captured at two magnifications (using 40x and 100x objective lenses) at 1024 x 768 x 24-bit color and transmitted over standard telephone lines at 14,400 kbps. Pathology reports and culture results served as 'truth diagnoses.' Categories of interpretations were correct, minor discrepancy, or major discrepancy with regard to the implications for patient care. Results: The diagnostic accuracy of video image readings and conventional light microscopy readings were nearly identical, with no statistically significant differences in the performances of specialty and nonspecialty pathologists (P > 0.05). The mean accuracies of readings of the video images and light microscopy images were 95.3% and 95.4%, respectively. Taking into account the time required by a referring pathologist to capture video digital images, telemicrobiology was somewhat less efficient than conventional light microscopy. Conclusions: Pathologists can accurately evaluate digital video images of preselected fields on Gram-stained slides. In clinical practice, however, a limiting factor may be the availability of local personnel qualified to select the microscopic fields for evaluation by telepathologists. The adequacy of the video images suggests that telepathology may also be used for remote supervision of quality assurance programs in microbiology laboratories, as well as for remote proficiency training of laboratory personnel.
ASJC Scopus subject areas