Dr. Mary Charlson, chief of the Division of Clinical Epidemiology & Evaluative Sciences Research, worked over 40 years to identify patients at highest risk and intervene to improve outcomes. One critical question for all populations is how to reliably identify those most at risk for unplanned hospitalizations and using this information, can early interventions reduce that risk—and the healthcare costs associated with it? A study Dr. Charlson and her colleagues published June 30 in PLOS One is the latest entry in that work.
Dr. Mary Charlson (Provided)
The paper shows how the Charlson Comorbidity Health Analytics (CCHA) works cross-sectionally over six years, longitudinally over five years and with new beneficiaries over the following five years. It used de-identified insurance claims data from 27,190 patients from 2016 to 2021 to determine if the CCHA could accurately identify patients with the highest risk of hospitalization and repeated admission in the targeted timeframe. CCHA scored each patient by adding up weighted measures of 38 chronic conditions in adults and children. Higher scores indicated higher likelihood of future hospitalizations and healthcare expenses. The relationship between CCHA scores and health outcomes remained consistent across the six-year period and age groups, according to the study. Dr. Charlson is hopeful that this model could prevent health deterioration and become a framework for population health management.
The CCHA study is a result of cross campus collaboration. Dr. Charlson worked closely with Dr. James Hollenberg, associate professor of medicine, also at Weill Cornell Medicine, and Dr. Martin Wells, the Charles A. Alexander Professor of Statistical Sciences at Cornell University, to analyze patient data. Dr. Hollenberg did a detailed evaluation of the Chronic Conditions Warehouse (CCW30) measure so that it could be compared with the CCHA. Dr. Wells analyzed the CCHA as a predictor of hospital admissions and high healthcare costs, using statistical models that contrast patients with and without costs and assess whether healthcare costs were determined by hospital admissions.
Enterprise Innovation took a deep dive into the development of the CCHA comorbidity index and its application in a conversation with Dr. Charlson.
Please provide us with an overview of your CCHA comorbidity index. What motivated you to develop this index?
The original comorbidity index was written in 1987. It was designed to predict mortality from chronic diseases so that clinical trials could incorporate these patients, by controlling for the additional risk of mortality, rather than excluding them. That original comorbidity study is now the most cited paper in the literature with over 53,000 citations.
The new Charlson Comorbidity Health Analytics (CCHA) Index is focused instead on predicting hospitalization, repeated hospitalization and high cost. The issue today in healthcare, especially in the United States, is how we intervene in patient populations to identify those at highest risk and take steps to prevent unnecessary hospitalizations. If you don't know who's at high risk, you won't focus on the right group or have an impact.
Most of the studies aiming to improve outcomes for chronic illnesses have included only two chronic diseases, such as hypertension and diabetes, but the problem is that they don’t consistently identify high-risk patients. When a patient has multiple chronic diseases, which is what comorbidity is, it needs to be handled differently. There has to be a more holistic approach to these patients.
Could you give us some examples of the comorbidity index adoption? What settings do you think bring the most value to providers and patients?
The CCHA comorbidity index, in terms of day-to-day use, has been implemented in several healthcare settings. Geisinger is an example of an institution that has rolled this out with a large population of patients, increasing appointment duration from 15 minutes to 40 minutes for patients with higher comorbidity indices. Columbia Primary Care has also adopted it.
The index was developed in a series of studies, some of which were published in 2014: one on a Medicaid population, another on a large population from SEIU1199 healthcare workers and their families. It was further developed over the next few years. This final index incorporates both adults and children, which the original index didn't.
My hope is that as the CCHA paper is published, more settings will adopt the comorbidity index because it gives an approach to managing outpatient care in a different way.
The way we care for primary care patients with high comorbidity today is not good for the patients themselves, their family or their community. They usually see a primary care doctor for 15 minutes and are asked some general health questions. Then, they are referred to multiple specialists, each of whom may start or adjust a medicine, often all conflicting. And pretty soon some patients end up in urgent care or the ER in a hospital. Nobody along that chain knows what happened. In order to intervene, we have to identify those who are at high risk among this group of more complicated patients and act accordingly.
How have you worked with EI/CTL to protect and disseminate this innovation?
For the new CCHA comorbidity index, we have the IP. We are licensing it to Geisinger and Columbia Primary Care. William Pegg, director of intellectual property at CTL in Ithaca, has done a brilliant job handling the prosecution of the patent application in the U.S. Patent and Trademark Office. Brian Kelly, director of business development and licensing of CTL at Weill Cornell, helps manage the day-to-day administration. Enterprise Innovation plans to highlight this work in its commercial outreach.
What are some other tools or methods that you’re currently working on? How does the comorbidity index fit into the bigger picture or inform your current research projects?
The endpoint is to use the CCHA comorbidity index as a platform for interventions to make a difference. We’re trying to prevent unplanned hospitalizations and emergency department visits with this risk marker. We’re working on a project that's designed to stabilize patients over time so that they don’t bounce in and out of the hospital. That's in the final stages now.
To learn more about Dr. Charlson's study, read the Cornell Chronicle article here.