Comparative analytics of infusion pump data across multiple hospital systems.

PURPOSE: A Web-based analytics system for conducting inhouse evaluations and cross-facility comparisons of alert data generated by smart infusion pumps is described. SUMMARY: The Infusion Pump Informatics (IPI) project, a collaborative effort led by research scientists at Purdue University, was launched in 2009 to provide advanced analytics and tools for workflow analyses to assist hospitals in determining the significance of smart-pump alerts and reducing nuisance alerts. The IPI system allows facility-specific analyses of alert patterns and trends, as well as cross-facility comparisons of alert data uploaded by more than 55 participating institutions using different types of smart pumps. Tools accessible through the IPI portal include (1) charts displaying aggregated or breakout data on the top drugs associated with alerts, numbers of alerts per device or care area, and override-to-alert ratios, (2) investigative reports that can be used to characterize and analyze pump-programming errors in a variety of ways (e.g., by drug, by infusion type, by time of day), and (3) "drill-down" workflow analytics enabling users to evaluate alert patterns­both internally and in relation to patterns at other hospitals­in a quick and efficient stepwise fashion. CONCLUSION: The formation of the IPI analytics system to support a community of hospitals has been successful in providing sophisticated tools for member facilities to review, investigate, and efficiently analyze smart-pump alert data, not only within a member facility but also across other member facilities, to further enhance smart pump drug library design.

Using blood glucose data as an indicator for epidemic disease outbreaks.

In the future, transfer of vital sensor data from patients to the public health care system is likely to become commonplace. Systems for automatic transfer of sensor data are now at the prototype stage. As electronic health record (EHR) systems adapt such functionality, widespread use may become an actuality in the foreseeable future.To prevent spreading of diseases, an early detection of infection is important. At the time an outbreak is diagnosed, many people may already be infected due to the incubation period. This study suggests an approach for detecting an epidemic outbreak at an early stage by monitoring blood glucose data collected from people with diabetes. Continuous analysis of blood glucose data may have the potential to prevent large outbreaks of infectious diseases, such as different strains of Influenza, Cholera, Plague, Ebola, Anthrax and SARS.When a person gets infected, the blood glucose value increases. If the blood glucose data from a large number of patients with diabetes are collected in a central database, it may be possible to detect an epidemic disease outbreak at an early stage. Advanced data analysis on the data may detect predominant numbers of incidences, indicating a possible outbreak. This gives the health authorities the possibilities to take actions to limit the outbreak and its consequences for all the inhabitants in an affected area.At the Norwegian Centre for Telemedicine, a mobile system for automatic transfer of blood glucose values has been constructed. By using wireless communication standards such as Bluetooth and GSM, the system transfers blood glucose data to an electronic health record system. Combined with a system accessing and querying data from EHR systems for patient surveillance we are extending our work into an Epidemic Disease Detection using blood Glucose (EDDG) system.