The workshop brings together a group of scientists with complementary expertise in health informatics, population health, and computer science to identify the research challenges and opportunities in population health data measurement, representation, and predictive modeling.
Despite great advances in measurement, computing, and communication technologies, the health data measurement relies on legacy practices (e.g., phone surveys). In contrast, billions of persons worldwide have mobile devices, such as cell phones and music players, which contain measurement sensors and are increasingly network aware. If these devices could be appropriately employed for population health data measurement, they could revolutionize the acquisition and use of population health data. However, there are many research challenges that need to be addressed in order to make widespread use of actionable population health data.
The workshop is organized around a vision of actionable data for population health. It covers a broad range of questions such as: What data should be recorded to measure everyday health? How should this data be most helpfully collected? How should the sensor data be classified into actionable data? How should the diverse sources be judged for quality? How should this data be mined and correlated? How can population data be transformed into usable knowledge? How should this data be used to develop practical health systems? How can multiple knowledge sources be integrated for multiple users? How can existing data (medical records and clinical trials) be leveraged using model-based inference to support customized decision making and refine predictive models? What is the impact of this new data on health quality and cost?
Workshop participants include experts in the areas of health informatics, knowledge representation and inference, machine learning and data mining. Thw workshop aims to increase the awareness of research challenges and opportunities in health informatics in general, and population data measurement, representation, and predictive modeling in particular, among researchers in data mining, knowledge representation and inference, machine learning, text analysis, human-computer interaction, social networks and social media, semantic web, decision theory. It also aims to make researchers in health informatics, public health, and related areas better aware of the state of the art informatics approaches that could be leveraged to develop the next generation health informatics infrastructure. The workshop results, including new research challenges and opportunities in discovery informatics, will be broadly disseminated through the workshop report, publications by workshop participants, and outreach efforts through follow-up activities that engage the research community.
At the core of the healthcare crisis is a fundamental lack of actionable data, needed to stratify individuals within populations, to predict which persons have which outcomes. A new health system with better health management will require better health measurement, to improve cost and quality. It is now possible to use new technologies to provide rich datasets necessary for adequate health measurement, to enable new information systems for new health systems. This report is a summary of a workshop on Measuring Data for Population Health, sponsored by the NSF SmartHealth program with assistance from the NIH mHealth initiative, held on January 12-13, 2012 in Washington DC. There were 42 attendees, including invited researchers from academia, government and industry, plus program officers from NSF and NIH. The goal was to recommend new initiatives for future programs in health informatics. The workshop had background talks by leaders in health systems and information systems, followed by breakout discussions on future challenges and opportunities in measuring and managing population health. This report describes the observations on what problems of health systems should be addressed and what solutions of information systems should be developed. The recommendations cover how new information technologies can enable new health systems, given funding support from federal initiatives to build working testbeds with real patients. The workshop and its report identify research challenges that utilize new computing and information technologies to enable better measurement and management for practical healthcare. The measurement technologies focus on deeper monitoring of broader populations. The management technologies focus on utilizing new personal health records to provide personalized treatment guidelines, specialized for each population cohort. This would enable predictive modeling for health systems to support viable healthcare at acceptable cost and quality. A workshop website contains background and discussion notes: https://wiki.engr.illinois.edu/display/hiworkshop/NSF+Workshop+Population+Health
