7 hospital managers explain how they use predictive analytics
AI-powered predictive analytics tools can improve hospital operations by optimizing capacity demands, alerting care teams of patients who may be at risk of adverse events, and anticipating supply shortages , among other uses.
Here, seven hospital executives share the most important way their healthcare system uses technology.
Editor’s Note: Answers have been edited slightly for clarity and style.
Atefeh Riazi. CIO at Memorial Sloan Kettering Cancer Center (New York): MSK uses predictive analytics in many areas. One is to identify patients who are at high risk for certain clinical events, including malignant airway obstruction, pelvic fracture due to bone metastasis, or cancer-associated thrombosis, in order to suggest prophylactic interventions. MSK is also applying it in digital pathology to identify and quantify various cancer biomarkers for targeted therapy. Finally, in the field of radiology, we integrate AI models into clinical workflows, such as lymphoscintigraphy examinations or the diagnosis of neuroendocrine tumors.
Allen Hsiao, MD. Director of Medical Information at Yale New Haven (Connecticut) Healthcare: Predictive analytics is one of the hottest promises of the digitalization of healthcare data. At Yale New Haven Health and Yale School of Medicine, we have implemented numerous algorithms to predict readmissions, emergency department volumes, sepsis, acute kidney injury, congestive heart failure, deterioration and one-year mortality, as well as to model COVID-19 patient volumes. , length of stay and state of preparation for discharge. Some are local, others are open source or commercially available and suitable for our population.
We are also studying the impact of predictive models to determine whether they actually work to change outcomes.
Zafar Chaudry, MD. Seattle Children’s Senior Vice President and CIO: Our data manager led our organization to HIMSS Adoption Model for Analytics Maturity Stage 7. In the area of ââpredictive analytics, our team of data scientists developed models to predict enumeration and the capacity to support it, such as tools for removing surgical cancellations. , shifting from diversions to directed placement for diverted children results in decreased visual acuity, improves operating room capacity, and eliminates clinical processes such as surgical regrouping.
Another model is used to predict whether the demand in the emergency department will be busier than expected over the next two hours; this allows the urgency to recruit staff in anticipation of this unexpected demand. Clinically, we have developed our own models for predicting sepsis and deterioration, and models for the sequential pediatric assessment of organ failure.
Mike Seim, MD. Head of Quality at WellSpan Health (York, PA): At WellSpan Health, we know that the application of predictive analytics, if done right, can be an essential tool in identifying patients at high risk for rapid decompensation of their condition and who may require transfer to the next level. superior care. To improve safety, our clinical teams engaged with technical and IT experts to review this data and recommendations, resulting in the development of evidence-based treatment / care pathways and integrated interventions into the EHR for adoption and distribution in our hospitals.
We have had great success in implementing this strategy to identify patients with sepsis, which has resulted in a significant decrease in mortality. We also implemented a similar strategy to measure SF ratios in our COVID-19 patients to identify those patients most at risk for developing acute respiratory distress syndrome. When patients are identified as having a decreasing SF ratio, we are able to begin predisposing patients earlier in the course of their disease. We have seen improved outcomes and decreased emergency transfers to the intensive care unit directly related to this type of predictive analytics and that is why we are looking at it at WellSpan.
Sunil Dadlani. CIO at Atlantic Health System (Morristown, NJ): Predictive analytics is not only an integral part of Atlantic Health System’s IT strategy, it is essential to how healthcare is delivered today and into the future. At Atlantic Health System, we use state-of-the-art AI and machine learning-based predictive and prescriptive analytics across the organization to help us better tailor care across our specialties and service lines. .
This helps us at all levels, from individual patients – predicting which patients are likely to develop adverse events such as congestive heart failure or being readmitted, or scheduling no-shows – to broader considerations at the same time. population health level, such as predicting which counties and postal codes will likely see an increase in COVID-19 patient volumes or a risk rating for chronic disease.
These analyzes can also help us predict supply chain and logistics bottlenecks, so that we can prepare and create contingency plans, ensuring that the quality of our care is never interrupted.
Deb Muro. CIO of El Camino Health (Mountain View, CA): As a healthcare industry, we have focused for years on building EHR platforms to support interoperability. The result of automating large amounts of patient information has created vast and sometimes unwieldy âdata lakesâ filled with challenges to deliver value or generate actionable results. It is important now, more than ever, to leverage the huge investments in data automation to harness the power of information to predict and predict aspects of the patient experience and care.
Our first foray into predictive analytics involved the use of algorithms capable of ingesting real-time patient data to predict expected changes in patient condition and condition. If a patient’s condition is expected to worsen within a defined period of time, a rapid response team is notified to intervene appropriately, resulting in a decrease in adverse events for the patient. We have only scratched the surface with these capabilities as the use cases are endless to use the data to predict what will happen in an hour, four hours or 24 hours regarding patient events, volumes, personnel and supply chain needs. Now is the time to disrupt the healthcare experience in a positive and impactful way.
Scott LaRosa. Executive Director of Business Analytics and IT at Southcoast Health (New Bedford, Massachusetts): Predictive analytics is ultimately the reward of clean workflows and accurate / aligned data, as well as an advanced, data-literate culture ready to understand and respond to such information. While predictive analytics may seem like the Ferrari of data science, we choose to start with the fundamentals of data governance, data management, and overall data usage education.
We’ve seen predictive analytics tools fail because of a standardized approach to a regionalised problem, like readmissions. The logic must match the need. The way this happens is to include the top executives of the C suite to take over the logic behind the build before launching the tool itself for end users to make decisions.
Where we’ve seen value, once we have that level of support, is when we mix predictive analytics with real-time reporting and daily workflows, for example by mixing a predictive metric of no-show on an existing outpatient calendar that shows today’s appointments. By doing this we have the ability to target patients who should have a more aggressive call / contact approach to ensure that they remember their visit that day, and that the provider’s busy schedule remains. on the right track.
Culture, management support, workflow, and education are all key aspects that should come before implementing a predictive tool in any organization.