MOUNT HOLLY, NJ / ACCESSWIRE / March 9, 2023 / Clare Medical of New Jersey ("Clare") (www.getclare.com), a provider of in-home comprehensive medical care to seniors throughout New Jersey, announced compelling data, from its recently completed large scale trial, evaluating the Clare Artificial Intelligence (AI)-based diagnostic (i.e., CAID™) platform's utility in preventing ER visits and hospitalizations. Based on these results Clare will initiate a full-scale launch of the CAID platform and deploy it as a diagnostic tool for its entire patient panel.
In the trial, involving approximately 1000 elderly patients having multiple chronic conditions and significant disease burdens, it was found that use of CAID™ resulted in a drastic reduction in the number of patients who experienced an ER visit or hospitalization, after being identified by the platform as having a high probability of requiring one within a 30-day period. CAID™'s model identified 317 patients (31.7%) of the panel as being at high risk of an ER or hospital visit and of those patients, only 66 of them ended up in the ER or hospital, reflecting a 79.2% reduction . The platform did not identify the remaining 683 patients as high risk, no recommendations were made, and the usual course of treatment was deemed appropriate by the model. In addition to those patients identified as high risk who ended up in the ER/hospital, another 23 patients (2.3%) had an ER visit or hospitalization after having been identified as low-moderate risk. These most recent results are very consistent with highly successful results from earlier trials conducted by Clare announced by Clare in November 2021 and June 2022 and with previously announced findings confirming validation of the platform's ability to accurately forecast which patients were at an increased risk of a hospital admission or ER visit within an error range of only 3%.
There has been an incremental build to Clare's AI development strategy as the company has been laser focused on the all-important goal of reducing potentially avoidable ER visits or hospitalizations. The platform has been enhanced to provide continuous and instantaneous uploads and updates of data in real-time and to delineate the variables that drive the model's prediction and thereby identify, for the provider, the primary reason(s) that a given probability score was generated. The AI model, developed by Elie Donath, M.D, MPH, MBA, Clare's Director of Data Analytics, identifies patients who are at high risk of requiring an ER visit or hospitalization by evaluating a variety of data points contained in their medical records associated with their most recent set of clinical encounters and encompasses variables like vital signs, lab data and provider notes. Based on this data, the model forecasts which diseases or conditions need to be addressed to attempt to avoid the patient having an ER visit or hospitalization. When an alarm is triggered by the model identifying a patient at risk, it guides the providers (I.e., medication adjustment, diet/medication compliance, symptom awareness, etc.) as to which disease condition to focus their efforts on and provides a set of recommendations targeted to that specific disease. In the recently completed trial, the most common conditions identified included fall/fractures (7.7%) cardiovascular events (4.5%), UTI's (4.3%) and COPD exacerbations (3.6%).
Although some patients (approximately 10% of the trial population), were lost to follow-up and several of those patients may have needed or experienced an ER visit or hospitalization the results are sufficiently compelling that even including all those patients as having had an ER or hospital visit would not significantly impact the trial's findings.
Commenting on CAID™'s enhanced capabilities, Dr. Donath said, "This latest trial goes a long way towards proving the capabilities of this exciting new model. We consider our platform to be unique in several ways compared to other prediction models currently in deployment. Firstly, it incorporates a wider variety of variables including both conventional tabular data in the form of vital signs and labs as well as unstructured text data using novel transformer models (which are the backbone to large language models like ChatGPT). Furthermore, in stark contrast to other AI-based diagnostic strategies, which do not provide the tools to understand why a given recommendation is made, our approach both identifies a patient as being at high-risk and provides direction as to what can be done to prevent an unfortunate outcome. It does so in a quite simple manner as not to overburden already-overworked providers trying to process an endless stream of data points."
Commenting on the potential value of the platform, Ron Lipstein, CEO of Clare Medical, said, "The results of our large-scale trial underscore the value of CAID. We have not seen or are aware of comparable published results Incorporating a diagnostic tool which can consistently reduce hospitalizations and ER visits to the magnitude our platform has shown. The capabilities of CAID have significant value particularly for those companies managing patients in a performance based 'at risk' model and for payers looking to reduce the healthcare spend for its members. According to most estimates, there are 3-4 million avoidable ER visits or hospitalizations annually in the U.S. and they cost the system approximately $30-50 billion per year. We look forward to exploring ways our highly effective predictive model and diagnostic tool can generate value to Clare, our affiliates and other healthcare organizations."
About Clare Medical of New Jersey
Clare Medical provides comprehensive in-home medical care to seniors and individuals with multiple chronic conditions throughout New Jersey. Clare partners with over 60 organizations throughout the state including hospitals, homecare companies, nursing homes and rehabilitation centers.
FOR MORE INFORMATION CONTACT:
Ron Lipstein
[email protected]
SOURCE: Clare Medical