ACT: Closing the Loop

Data in healthcare is good. Data applied for something useful – like a clinical decision that informs individual care – is even better.

Despite our progress to date in acquiring data and our advances in artificial intelligence and machine learning for analysis, humans remain central to decisioning in healthcare.

Decisioning requires the automated delivery of necessary, specific information to the point of decision. Some of the necessary elements include access to requested data, standardization of data, analytics operations and insights from machine learning. These inputs should be delivered automatically and organized in an interface that is familiar to the decision maker.

In this context, data are only useful if they result in an action or a professional judgment. Most of the available clinical data about an individual is not needed for the next clinical decision, and thus should not be presented unless requested.

Post-decisioning status creates a set of Actions. Much of Action can be systematized and automated because these functions are largely the assemblage of information and the execution of predictable activities.

Post-decisioning automation includes actions like the delivery of prescriptions, standard work flows, requests for testing or procedures, reports and communication, analytics and output from machine learning. As with the automation of information flow into the decision process, a wealth of applications can be developed to support these downstream activities as well.

The result of post-decisioning actions are included in the next iteration of inputs, completing the automation learning loop.

The process of analyzing and presenting ever larger data sets at the point of clinical care requires advances in automation and not only machine learning but also systems-level learning. The learning loops and outcomes feedback loops support the systems-level learning.

We have reached an inflection point where generating data vastly exceeds our capability to use it in clinical settings or for self-care due to a lack of effective process automation.

The process for good human decisioning starts with asking “Why are certain data being presented? What is the step next required?” Then, after the decision, there are opportunities to develop innovative options for carrying out the downstream activities, the results of which become part of the data that feeds the next Assess function.

Bob Teague, MD, Chief Medical Officer of SocialCare

Bob brings a wealth of experience to his articles based on a career spanning clinical practice in major healthcare institutions as well as leadership roles in multiple entrepreneurial enterprises and Fortune 50 technology enterprises. These financially successful enterprises were transformative in their markets for respiratory home care services, diabetes chronic care management, and Medicare Advantage risk management through transitional care. Bob’s first blog series focuses on a central SocialCare paradigm, “Acquire + Assess + Act”.

0 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *