Predicting Healthcare Readmissions
Every life is infinitely important. To protect more lives, hospital clinicians need to predict and decide how to prevent the likelihood of a patient being readmitted. During a global pandemic, this prediction process is especially important because every bed matters more than ever. AI is key in helping to minimize readmissions — by better understanding each patient as an individual and by ensuring that there will be enough staff and beds for the upcoming day.
Implementing AI has typically been a long and arduous process, maintained by deeply experienced but siloed data scientists. That is where Rob O’Neill stepped in. Leading an effort to democratize AI, his team has created dashboards that non-data scientists can easily understand and act on. This results in better accuracy and adoption throughout several hospitals. Saving lives. Improving health. Reducing tremendous financial strain on both healthcare providers and payers.
Data-Based Predictions with Transparency
Rob wanted a solution that could proactively reduce 30-day readmissions rates by predicting in advance which patients were likely to be readmitted. He also wanted the transparency of understanding the top reasons why — on a patient-by-patient, individual hospital, or diagnosis basis.
Users can look for statistical patterns or anomalies, investigate distribution, review emergency department predictions by the hour, expected bed occupancy of inpatient wards, and much more. Providers can often recommend a wide array of options post-discharge, ranging from home healthcare services to temporary placement in a skilled nursing facility. These insights are based on a wealth of information: daytime, whether it is a major or minor patient, whether they arrive by ambulance, temperature, weather, holidays (especially bank holidays), and the proximity of a hospital near a university or college.
Snowflake is a scalable database that hosts the entire patient history and live data. Qlik migrates the data and feeds to DataRobot for generating AI-driven predictive models. This gets deployed with no coding, and the Qlik Analytical Command Centre is immediately and easily used by healthcare decision-makers for broader usage across the hospital system.
Being able to understand our current patient population’s risk for readmittance is crucial to effectively executing unscheduled demand forecasting to manage an influx of crisis-related patients. Combining data science and analytics through an integrated platform of DataRobot, Qlik, and Snowflake will help us more confidently understand our current needs, and predict what resources we will need to deliver complete care to our entire patient population.
DataRobot’s platform makes my work exciting, my job fun, and the results more accurate and timely – it’s almost like magic!
I think we need to take it upon ourselves in the industry to build the predictive models that understand what the needs and wants of our customers are, and go through the whole curation process, become their concierge.
At LendingTree, we recognize that data is at the core of our business strategy to deliver an exceptional, personalized customer experience. DataRobot transforms the economics of extracting value from this resource.
We know part of the science and the heavy lifting are intrinsic to the DataRobot technology. Prior to working with DataRobot, the modeling process was more hands-on. Now, the platform has optimized and automated many of the steps, while still leaving us in full control. Without DataRobot, we would need to add two full-time staffers to replace what DataRobot delivers.