Photo: Dr. Nirav S. Shah
Wearable sensors that collect continuous physiologic data streams are becoming less expensive, more accurate and increasingly connected to cloud-based machine learning algorithms to identify perturbation – a deviation from a normal state caused by an outside influence.
To date, there has been limited work looking at how these platforms can remotely identify perturbation, allowing care providers ample time to intervene and prevent decompensation and readmission.
Dr. Nirav S. Shah, medical director, quality innovation, at NorthShore University HealthSystem, and colleagues evaluated a continuous remote patient monitoring platform using noninvasive wearable sensors that collect near real-time ambulatory vitals, a mobile device that collects patient-reported data, and machine-learning algorithms to identify heart failure patients at risk of physiological perturbation.
The continuous RPM platform is tied to structured operational workflows that involve home health nurses and the clinical team in a cascading alert system. This study has provided insight into scaling implementation of a continuous RPM platform to improve patient care across use cases and throughout NorthShore University HealthSystem.
Shah will be presenting on the subject of continuous RPM at the upcoming HIMSS22 Global Conference & Exhibition in a session entitled “Building a Continuous Remote Patient Monitoring Program.”
Shah, a practicing physician in the division of infectious diseases at NorthShore and a clinical assistant professor at the University of Chicago, sat down with Healthcare IT News to offer a sneak peek at his session.
Q. How do hospital staff members recognize the opportunities to monitor patients at home?
A. Our remote monitoring project is a clinical trial to understand the feasibility and preliminary efficacy of continuous remote monitoring in heart failure and colorectal surgery populations.
As a result, our research team screens patients to see if they meet predefined inclusion criteria such as being high-risk for readmission based on a custom 30d readmission prediction model called CAPE that we built in house at NorthShore and that scores every patient on presentation, qualifying for home health and able to perform the daily tasks of the remote monitoring program.
Within these general inclusion criteria, the question is, how do you identify the right patient? Some patients, despite meeting these criteria have difficulty following the protocol or are overwhelmed with their condition to worry about a clinical trial on remote monitoring.
One of the characteristics we have identified associated with withdrawal from the study is lack of social support. Family members and caregivers can provide a strong motivating force to help patients continue with the remote monitoring, and this criteria is one of our informal criteria to ensure that the patient will be able to comply.
In terms of how we settled on our first two use cases for remote monitoring, we were interested in finding a scalable remote monitoring solution that would allow us to identify physiologic perturbation in very different patient populations.
So we settled on a common initial remote monitoring use case of chronically ill heart failure patients, but we chose another high-risk patient population that is less commonly studied, which is patients immediately after colectomy who undergo ileostomy creation.
We landed on the colorectal patient population because this population is at high risk of readmission, there are only a few key complications that result in readmission, and physiologically they are similar to heart failure in that it has to do with volume status, specifically dehydration while heart failure is volume overload.
Our hope was that the similarity in volume status would help us with transfer learning from the heart failure to the colorectal use case.
Q. You suggest a human-centered approach that engages and empowers frontline workers is crucial in implementing advanced tech for continuous remote patient monitoring. Please elaborate a bit.
A. One of the key frameworks that we anchored on for this work is: Success in this work is not just reliant on the technology and process, but also on the people. All three of these components are critical, and the people component is the one that can be neglected, especially for advanced tech implementation.
The people component is one that comprises engagement, empowerment and education. What we found in our initial heart failure soft launch is we were not able to engage our heart failure nursing team, because there were staffing issues, and we weren’t sure who was going to participate in this program. As we developed the program, we relied on the heart failure physicians to map out the nurses’ workflows.
When we went live for our soft launch, while we educated our newly named heart failure nurses, they were not part of the design process. And we found that our protocol seemed to fail with this team member in our cascading alerting system. Despite educating the nursing team how to use the remote monitoring technology and telling them what they were required to do for each escalation to them, there was minimal engagement.
After the soft launch, we recognized the heart failure nurses were critical in the management of our remotely monitored patients, so we paused for a month to build the nurses’ process map from scratch utilizing their input on how to integrate this within their existing workflow, and with their clinical insights.
We also combined this with intense education and embedded our team in their weekly meetings to make sure our study was top of mind. After relaunching for our calibration period, we found immediate success and great engagement and empowerment. The nursing team intervened on numerous patients, and we are fairly convinced prevented one to two readmissions because of active and early management.
Q. What is one of the key learnings from the configuration period of your heart failure continuous remote patient monitoring study?
A. In the configuration period, we had no readmissions. During this part of the study, our home health nurse was contacting patients for every physiological alert, every new or worsening symptom and also if a patient did not have an alert after a set number of days.
With this very high-touch system we intervened on numerous patients by changing their diuretic dosing and provided counseling on diet and adherence. As a result, we are fairly convinced we prevented one to two readmissions from heart failure exacerbation.
While we were able to potentially prevent these readmissions, this degree of touch was only sustainable by the home health and clinical teams because of the low caseload. From this we learned that the alerting system can help us identify minor perturbations that can be intervened on with high engagement and active management.
Our goal after this period was to understand how best to leverage our alerting system while making the process more scalable and efficient.
Shah, along with his colleague Wei Ning Chi, a research associate at NorthShore University HealthSystem, will present “Building a Continuous Remote Patient Monitoring Program” at HIMSS22. It’s scheduled for Thursday, March 17, from 11:30 a.m. to 12:30 p.m. at the Orange County Convention Center in room W230A.
Twitter: @SiwickiHealthIT
Email the writer: [email protected]
Healthcare IT News is a HIMSS Media publication.
Source: Read Full Article