A Deluge of Data: Ambulatory Physiology Data Strategy

With all of the change occurring in both the economic models of healthcare delivery and technologies, a key challenge faced by healthcare delivery executives is how to adopt technologies that reflect the realities of today while preparing for success in the future. Such technology investments are significant, and are best made with a sense of the intersection between the three axis of the triple aim: improved care experience, improved health outcomes, and lowered cost. To support this, we offer a framework by which to assess needs and develop strategies.

Data acquisition alone should not be goal. Rather, an organization’s goal should be to provide information that supports clinical decision-making. To get here, the questions to answer are:

1. What information will the clinician need?
2. How can we provide that information?

A useful approach for answering these questions is to look through the lens of the classic clinical delivery of care framework:

1. Evaluate the patient – Validate health status and/or look for indications of health issues
2. Problem solve/diagnose – Investigate indicators of health issues to identify the specific, underlying problem
3. Treat – Attempt to address the underlying issue and help manage symptoms
4. Monitor – Provide ongoing assessment of impact of treatment and/or progression of the issue
5. Respond/refine treatment – Evolve the treatment approach based on patient needs, according to evidence-based guidelines (1^st line, 2^nd line, etc)

By overlaying the role of ambulatory physiological monitoring and the associated analytical requirements of a solution, we get a framework as represented here:

 

Clinical Process	The opportunity associated with ambulatory physiology monitoring	Analytics Requirements
1. Evaluate/asses health status	Provide early indication of issues, pre-symptomatic	•     Detect subtle physiological changes, with precision•     Extract the signal from the ambulatory noise•     Characterize physiology and account for the interplay between variables within a dynamic system•     Indicate issues in advance of symptoms
2. Problem solve/diagnose	Limited	N/A
3. Assign treatment regimen	Limited	N/A
4. Monitor	Indicate success of treatment choice in terms of addressing the underlying issue, managing symptoms and side effects	Indicate patient responsiveness to treatment
5. Respond/ refine treatment	Inform follow-on treatment decisions	Indicate patient responsiveness to treatment

 

By using this framework, we see there are several key requirements of the analytics that transform data into information:

1. It must detect subtle changes, with precision.   It’s important that the analytics indicate small changes in physiology and do so in a way that is reliably precise. Without this, clinicians will not have the confidence in the analytics, which will negatively impact adoption and return on investment.

2. It must remove the normal bio-signal variability associated with ambulatory living. A person’s physiology is changing throughout the day in an ambulatory environment (eg. sleeping vs. awake, walking up stairs, etc). Thus, an analytical solution must differentiate between these various states and only indicate clinically meaningful change. That is, change in physiology that is indicative of an underlying problem, not a normal response to increasing or decreasing system load.

3. It must characterize physiology and account for the interplay between variables within a dynamic system. Human cardiopulmonary physiology is a system defined by interrelated subsystems that work in concert to perfuse tissue with oxygenated blood and remove carbon dioxide. For example, upon walking up stairs, one’s muscles require more oxygen, which results in a faster heart rate, which, in turns, impacts respiration. This interplay enables compensatory action to keep things working when one of the subsystems (eg. cardiac stroke volume) is compromised. Thus, to truly see what’s going on with someone’s cardiopulmonary system, the analytical solution must account for multiple variables and capture a comprehensive picture of the interplay in order to yield the compensatory action.

4. It’s most effective when it indicates issues early, in advance of symptoms. For conditions such as heart failure or COPD, its’ often too late when symptoms are present; hospitalization often follows shortly thereafter. Thus, in order to leverage lower-cost outpatient or telephonic interventions, the analytic must indicate change in advance of symptoms.

5. It must indicate personal change in physiology. Monitoring, by definition, is a function of assessing patient progress relative to where they have started. Thus, an ambulatory monitoring solution must integrate a relatively sophisticated level of personalization to help the clinician understand if a patient is deteriorating or improving.

Image Source: Unsplash, Marcus Schorn