As CMS announces readmission penalties for nearly 2,600 hospitals, some hospitals and health systems are finding precision medicine and big data efforts to be a new strategy for reducing 30-day readmissions
On October 1, 2019, the Centers for Medicare and Medicaid Services (CMS) announced payment cuts affecting nearly 2,600 hospitals, reported Kaiser Health News. The cuts follow up on CMS’ nearly decade-long pressure and threat of financial penalty for unacceptable 30-day unplanned hospital readmission rates.
Early signs of a new strategy for reducing unnecessary hospital readmissions point to precision medicine. Here are some examples that the Precision Medicine Institute has found:
Leaders at Intermountain Medical Center Heart Institute in Salt Lake City recently studied personalized “risk-score guided care,” which uses a precision medicine approach. Rolled out gradually to eight of Intermountain’s largest hospitals, risk-score guided care relies on patients’ personalized daily clinical information to create a score that determines where patients are directed for treatment for heart failure.
The results of Intermountain’s study showed thatrisk-score guided care reportedly lowered 30-day readmission rates by 25%, and mortality rates came down by nearly half. It should also be noted that three of the Intermountain hospitals that were part of the study are among the 20 hospitals in the country with the lowest readmission rates, according to an Intermountain news release.
Members of the Intermountain Heart Institute team involved in the risk-score guided care study include Colleen Roberts, RN, MS; Kismet Rasmusson, DNP, FNP; Jason Buckway, RN, MBA; Rami Alharethi, MD; Jalisa Cruz; R. Scott Evans, PhD, MS; James F. Lloyd; Tami Bair, RN; Abdallah Kfoury, MD; and Donald Lappe, MD.
Indeed, Intermountain has an overall high star rating, according to the CMS Hospital Compare tool, thanks in large part to its outcomes in patients with heart failure.
Condition-Specific Precision Medicine Gaining Momentum
Heart failure ranks among the top four condition-specific readmission measures as part of the federally-mandated Hospital Readmissions Reduction Program (HRRP), according to CMS. The others are acute myocardial infarction (AMI), chronic obstructive pulmonary disease (COPD), and pneumonia.
Therefore, it makes sense that researching the genetic origins of cardiac conditions and fostering advancements in cardiac care appear on the roadmap to precision medicine success.
As PMI reported recently, researchers at Temple University’s Lewis Katz School of Medicine are studying the unique genetic variants related to heart muscle disease and heart failure—evidence of even more dedication.
Additionally, through the development of Precision Medicine Centers of Excellence (PMCOE), the Johns Hopkins Precision Medicine Initiative includes 16 COEs, each focusing on a specific disease. Hopkins is now working to develop 50 Precision Medicine Centers in the next five years.
Johns Hopkins Medicine reports progress in the areas of prostate and other cancers, as well as scleroderma, arrhythmia, myositis, and neurofibromatosis, to name a few. The Johns Hopkins Health System also understands the importance of precision medicine in reducing readmissions.
Is Big Data Getting Better at Predicting Hospital Readmissions?
At Johns Hopkins, providers are taking advantage of “sophisticated computations of similar patients’ histories” using data in the Hopkins PMCOE for Prostate Cancer. The Centers of Excellence are allowing the health system to harness “the power of big data to help pinpoint the treatment—or watchful waiting protocol—that is most likely to be beneficial for each individual,” notes Patricia Brown, President of Johns Hopkins HealthCare.
When PMI advisory board member Pranil K. Chandra, DO, Chief Medical Officer of Genomic and Clinical Pathology at PathGroup in Brentwood, Tenn., thinks about the inroads made by Intermountain and Johns Hopkins, he says he can’t help but wonder how using big data to focus “on infectious disease organisms, particularly the microbiome, can help more precisely identify patients with infections so they can receive individualized treatment early and not go on to develop advanced disease.”
When it comes to assessing and reducing readmission rates, a common denominator appears to be data. A study published in March 2019 in the Journal of the American Medical Association (JAMA) demonstrated that machine learning worked better than standard predictive methods in predicting hospital readmissions. In July, researchers reported in Nature that a deep learning approach was able to predict 90% of future acute kidney injuries that required subsequent dialysis.
PMI advisory board member Michael Clare-Salzler, MD, Chair of Pathology at the University of Florida School of Medicine, notes that studies like these underscore the value of mining the electronic medical record (EMR) to develop risk scores for readmission.
Michael Clare-Salzler, MD (above), Chair of Pathology at the University of Florida School of Medicine, says utilization of the EMR has yet to be fully realized in most health systems, adding that it would behoove hospitals and health systems to mine EMR data for new initiatives, especially with the addition of genomic, metabolomic, and other advanced testing. (Photo copyright: University of Florida Health)
Dr. Clare-Salzler offers these “keys to success:”
- Strong informatics support including machine learning;
- Experts to guide the process;
- “Clean” and organized data contained within the EMR; and
- Clinical studies to validate the approach.
“In addition to medical history data, information from the laboratory and other ‘hard’ data sources will likely make the process stronger, he explains. “Finally, as we move forward, the addition of genomic, metabolomic, and other advanced testing will likely increase not only the ability to prevent readmissions, but also to potentially prevent or delay disease onset and manifestations.”
As findings like these are further validated, hospitals and health systems with datasets that include thorough genetic and demographic information will be in an enviable position when it comes to reducing readmission rates and improving outcomes.
In the meantime, Clare-Salzler says he likes Intermountain’s relatively simple approach, and he suggests others do the same.
“They employed data from a system already paid for and in place. They’re using the EMR for something other than a filing cabinet.” If the paradigms developed by Intermountain translate to other institutions, he adds, “there is a great opportunity to make this a national initiative.” Clare-Salzler says that the next logical step would be for federal agencies to fund studies similar to Intermountain’s to see if a case could be made for such a rollout.