Internet-based telemedicine: An experimental study to provide decision support using real-time assessment of video quality

Telemedicine and related healthcare technologies aim to provide efficient health services to improve the well being of patients and bring medical expertise at a lower cost to individuals who are geographically separated by distance. The goal of this dissertation was three-fold: (1) to provide a classification framework for telemedicine/telehealth programs, (2) to develop new quality models/measures that are suitable for telemedicine applications over the Internet, and (3) to apply these models in a prototype that can provide real-time quality information to the users at the point of service.;Telemedicine taxonomy was developed as a framework to guide program planners and users of telemedicine during the planning, implementation, and evaluation of their programs. Telemedicine has various dimensions that require expertise in different domains. Relationships that occur between different dimensions in the existing pool of telemedicine programs were identified and reported.;Telemedicine has been quite slow in utilizing Internet technologies because of quality concerns. This study investigated the relationship between objective and subjective quality measures, as well as clinical decision making capability that is affected by the impairments that occur during the transmission of video over the Internet. Existing quality measures were tested on degraded videos generated over an emulated Internet testbed using NISTNet. The original telemedicine video sequence selected for the experiments was a telediagnosis case in ophthalmology (this is a video recording of a general eye examination conducted by a physician from a remote location) obtained from the Regenstrief Institute for Health Care. A new clinical decision making capability scale was also introduced and linked to the traditional objective and subjective values. Findings regarding the relationship between different telemedicine video quality measurements were reported. For example, experts were sometimes able to make clinical decisions with low objective quality scores because the impairment in the video occurred at non-critical frames. The results of the quality measurements were stored in a quality database. This database was utilized in a prototype decision support tool that predicts quality and decision making capability based on the previously collected data. Such tools will be very useful for practitioners during telemedicine sessions.