Speeding up the predictive analytics process with automated machine learning


How a leading CRM provider leverages DataRobot’s predictive modeling technology to help optimize healthcare marketing

The Situation

In an increasingly dynamic and competitive healthcare environment, the need is clear: a healthcare organization must grow its business through patient, non-patient, and physician engagement. As healthcare systems balance both volume-to-value-based service delivery and marketing approaches, there is an even greater need to access and leverage all of an organization’s data, analytics, and digital communication assets. A significant piece of capitalizing on these assets requires predictive analytics. Consider, for example, the ability to predict high-risk individuals who would benefit from health screenings, to determine the propensity for certain cancers, or to aid physicians with optimal diagnoses and hospital readmission risk predictions.

Building and deploying predictive analytics can be costly and time-consuming — it is often fraught with inaccuracies. Regardless, the need for healthcare providers to turn clinical and diagnostic data into meaningful insights remains.

Evariant has recognized the transformational impact of accurate predictive analytics in healthcare. As a rapidly growing company, their resources limited their predictive modeling output to several specific models per month, not nearly enough to support their growing base of healthcare providers. Additionally, because the complexity of their healthcare data demanded a high level of hands-on data preparation, their existing solution was adequate, but not optimal. They needed high-quality predictive analytics that could be generated both automated and semi-automated – and with an extremely high degree of reliability and validity.

“It makes my job quicker and more efficient. I just get more done in a shorter period of time. The reason we use DataRobot is simple. The speed, the quality, and the ability to tweak on the fly as we are running the models. It’s really about allowing us as an organization to focus on the different applications of predictive analytics.”

The Solution

Predictive analytics requires predictive modeling within a context or package. Rather than the Evariant team spending weeks on model building, validation, and scoring, the DataRobot platform, hosted on Amazon Web Services (AWS), allowed for automation and semi-automation of the predictive analytic processes, speeding model building and extending the deployment lifecycle.

The Evariant/DataRobot collaboration resulted in thousands of validated predictive models, in an automated context. The Evariant team was able to choose the most statistically reliable, valid, and appropriate model results using a collaborative and sophisticated cross-validation framework. The results can be seen in the volume of models Evariant is creating and deploying — nearly 10x the previous pace.

Having a repeatable predictive model has afforded the Evariant team the ultimate luxury — time. In a nutshell, increased time allows for better utilization of resources, more one-to-one involvement with clients, and an increased ability to use scored data to optimize marketing and other efforts. The quality of cross-sell, up-sell, retention, acquisition, re-acquisition, and other strategies has significantly increased. In the end, these efficiencies lead to increased ROI, not only for Evariant’s clients, but for Evariant as well.

Evariant: Moving Healthcare Forward

Evariant, founded in 2008, has emerged as one of the fastest- growing SaaS companies in the healthcare provider market. The company delivers a suite of innovative CRM solutions that help healthcare systems identify and execute on the most important strategic growth initiatives, including patient engagement, physician alignment, and optimized and innovative marketing. As large and complex as the healthcare industry has become, the core focus remains on the patient. Evariant believes that technology and innovation can empower the connection between patients and healthcare providers, and can ultimately revolutionize the way healthcare service delivery and marketing are practiced.

“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.”