Changing the way we manage patients with chronic diseases is one of the major challenges to improving outcomes while controlling costs, and technology used in the home can have a dramatic impact on the results.
These patients already account for 70 percent of the total amount spent on healthcare in the U.S. Thanks to the growing effectiveness of medical treatments, the prevalence of chronic disease is increasing because patients are living longer. As their average age goes up, so does the cost of treating them.
This is especially true for patients who have chronic diseases that get worse “in secret” – diseases such as congestive heart failure (CHF) and chronic obstructive pulmonary disease (COPD), where deterioration takes place out of sight of clinicians. With these “silent” diseases, early signs of problems are too often missed, along with the opportunity to intervene before a trip to the hospital is required.
On the upside, CHF and COPD are considered “ambulatory care sensitive,” meaning that better outpatient care can improve health and reduce hospitalizations. That recognition was the impetus behind the creation of the Cedars-Sinai Cardiac Optimization Program, an intensive hands-on program that supports frequent contact with CHF patients to track and identify their problems before they become acute.
The Cedars program recently enhanced its effectiveness, and its ability to scale to cover large numbers of patients, through an experiment in reimagining the monitoring of patients in their own homes.
The practice of monitoring chronic disease patients in their homes, particularly those with CHF, has been around for more than 20 years. The first programs showed limited value in reducing avoidable hospitalizations. They were typically labor-intensive, expensive and unable to scale beyond a small group of patients.
In addition, most of the early efforts at home monitoring were overly simplistic, lacking a realistic model of the circumstances faced by a chronic disease patient at home. For example, almost all the early home monitoring programs for CHF patients ignored the fact that most of the patients had other chronic diseases, such as diabetes or COPD. Focusing only on one disease in a patient with multiple co-morbidities inevitably leads to missed opportunities for avoiding hospitalization.
The early programs also took a “one diagnosis fits all” approach to selecting patients for home monitoring, although not all CHF patients are well-suited for home monitoring. Even those programs that tried, correctly, to predict which patients would be heavy users of hospitalization failed to assess whether patients would benefit from home monitoring.
Finally, the early home monitoring programs tracked only one parameter, such as body weight, and only one-day changes in that parameter. This approach results in an excess of false alerts of deterioration, which wastes staff time and adds significant expense.
For example, if a patient experiences a two- or three-pound weight gain from one day to the next, it might indicate sudden fluid retention – a clear sign of deterioration in a CHF patient – but the reality is that there are many reasons why weight can change. Early programs didn’t track potentially corresponding parameters that might have explained such short-term weight gains. Even if they had, weight gain occurs late in the deterioration process, reducing the likelihood of avoiding hospitalization.
As we now know, there are better ways to predict deterioration.
For one example, Cedars-Sinai provides a “high touch” care program to improve the care of patients with advanced CHF. The high-touch approach helps anticipate problems so they can be treated pre-emptively. The next step is to increase the level of communication between patients and their care team by tracking physiologic measurements such as heart rate, blood pressure, oxygen saturation, thoracic fluid, activity and other indicators to detect clinical decline even earlier.
The Cedars-Sinai team provides the human interaction and clinical support; Sentrian provides the technology and analytics to detect the patterns in the data that are predictive of worsening CHF, days before it would otherwise be apparent. Together, both Cedars and Sentrian are investigating the potential application of these emerging technologies with the goal of personalizing care and improving heart failure management at home rather than in the hospital.
A realistic model for effective home monitoring, such as the one deployed in the Cedars-Sinai program, is designed to correct for the following complexities:
* An accurate assessment of health status requires tracking multiple parameters, including more than one physiologic measure, patient-reported data, observations from loved ones and caregivers in the home and even environmental data.
* Analyzing the data to detect patterns in multiple data streams will permit more accurate and earlier prediction of patients that are on a path to hospitalization.
* Because deterioration occurs over a period of days or even weeks, we need to look at longer term trends, not just one day changes, if we are going to be able to provide enough advance warning to enable a modest intervention to reverse the process.
* Achieving better specificity and sensitivity, to provide the best clinical and economic value for patient and provider, requires that the system adapt to become more personalized to the individual patient.
* Not all patients with CHF are the same, so it is likely that different patients will exhibit different predictive patterns.
Patient participation is a key to the success of a home monitoring program. However, the more we require of patients in taking and reporting their measurements, the less likely they are to participate.
Early home monitoring efforts made numerous demands of patients, such as having to manually record their blood pressure and phone results in. Today, new wearable technologies can automatically record and transmit data to the cloud to reduce this burden. By making home monitoring systems easier for patients to use, machine learning not only can produce more data but more reliable data, and a greater likelihood of value.
Advances in machine learning also are helping to personalize home monitoring systems to individual patients. The technology helps determine the unique patterns within individual patients that predict deterioration. This is critical because not all patients with CHF, or any other complex condition, are the same, so different patients will exhibit different predictive patterns.
Machine learning essentially is artificial intelligence that compares predictions with actual outcomes and feeds the findings back into the analytic system, and its use helps improve and fine tune individual predictive algorithms over time.
Taken together, machine learning and predictive analytics are driving a quantum leap in remote monitoring that is revolutionizing the practice of monitoring patients at home and preventing needless hospitalization.
Raj Khandwalla, MD, is director of cardiovascular education at Cedars-Sinai Medical Foundation.
Martin Kohn, MD, is chief medical scientist of Sentria.
Register or login for access to this item and much more
All Health Data Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access