Using Data Science to Improve Patient Care and Satisfaction

In healthcare, one of the most important measures of success is patient satisfaction. Every hospital patient in the U.S. is asked to complete a Consumer Assessment of Healthcare Providers and Systems survey. A large health network asked us to analyze its data using advanced Artificial intelligence and computer science techniques.

The goal was to help our healthcare client fully understand patients’ needs so the network could improve its ratings and develop better, more customized care. The challenge was to deliver specific, actionable recommendations and advice to our client using a combination of patient feedback and clinical background data.

Our Approach

We investigated data from 60,000 patients who visited the organization’s health centers over the course of six months. Our team evaluated key patient-care issues such as communication, responsiveness and pain management, and then compiled a holistic set of analytics correlating the patients’ clinical and social backgrounds along with their satisfaction feedback.

In analyzing the data, we used decision-tree models and regression models, as well as intensive hypothesis testing to fully understand all the variables and how each interacts with the others.

Turning Data Into Business Intelligence

We delivered a detailed set of patient satisfaction analytics, along with a set of observations and very informed, specific recommendations for the health network. We made a number of suggestions regarding communications with patients, especially during the hospital discharge process. For instance, older patients need to be carefully briefed on their medications and new mothers need extra attention on discharge day. We advised the client on resource planning to ensure proper staffing for these special circumstances.