A proposed framework focusing on development and application of electronic healthcare predictive analytic (e-HPA) applications has been published this month in eGEMs (Generating Evidence & Methods to improve patient outcomes).
The framework builds on "earlier frameworks of model development and utilization" to show opportunities for e-HPA, ways to address challenges and ideas for motivating stakeholders to adopt and refine the framework.
It came together through a grant for Parkland Center for Clinical Innovation from the Gordon and Betty Moore Foundation. Staff from both organizations chose the framework participants, who included seven professionals from academia, two from private foundations, three from healthcare delivery systems, three model developers and two from governmental agencies.
Five areas provide the structure for the framework, according to the authors, which include:
- Data barriers and model development: This section addresses the need to make data sources available to the scientific community, as well as establish protocols to address data breaches and de-identification of information. In addition, "data from different sources must be harmonized, cleaned, and reliably linked before e-HPA models can produce reliable results," the authors write.
- Transparency and model evaluation: Standards must be created with regard to validation and transparency of predictive analytics. They should include developing best practices, creating clinical coherence and incorporating diverse data sources, among other standards. Transparency should increase trust for both patients and providers, the authors add.
- Ethics: "Users of healthcare predictive analytics should develop a risk-benefit analysis approach ... at the individual, organizational and societal levels to determine the adoption of these models," according to the framework.
- Regulation and certification: While the authors note that innovation when it comes to e-HPA should not be stifled, "appropriate regulation and certification framework" is crucial for organizations applying such technology.
- Education and training: Medical institutions and other training facilities should include e-HPA in curricula and training.
"It is the task of healthcare leaders, e-HPA practitioners and other stakeholders to ensure an infrastructure that ultimately promotes effective use of predictive analytics to improve patient outcomes, satisfaction and the value of healthcare resources," the authors conclude.
Predictive analytics in healthcare, however, are still in their infancy. Many organizations still lack a clear analytics strategy, according to a 2015 Deloitte survey. While spending on analytics is expected to grow overall in the industry, only five of the responding organizations said they expect analytics spending to grow significantly in the next three years, the survey found.
In addition, while data has the potential to improve healthcare, the industry must make better use of the information being generated, according to a paper from the National Quality Forum.
To learn more:
- here's the proposed framework (.pdf)