Data Opportunities in Health Care

Data-driven patient engagement  |  October 5th, 2017

No one group in the health care industry can create a complete patient picture. Rather, insurance companies, drug companies, doctor offices, etc., each get snapshots of the patient at fixed points in time. Having one place in which to compile all patient data would be tremendously useful, but the lack of it shouldn’t stop you from using the data you do have.

While the specifics may look a bit different from company to company, this post contains high-level observations on how your health care organization can start with what you know about patients to build better, stronger patient relationships. For further thoughts on understanding customer engagement in product companies, see our Data-Driven User Engagement post.

Becoming patient-centric through data

It’s tempting to believe that shipping your data off to a firm for model-building, or uploading it to a product, will expose amazing insights. To truly be successful, though, you need to be intentional about the insights you’re after, and start by identifying your business objectives. With those in hand, think about the data you actually need to achieve those objectives.

How can you get from the data you have, to the data you need? The following observations will give you a place to start so you can use your data to drive patient engagement.

Establish patient engagement metrics

Before trying to “boil the ocean” and unlock all aspects of the patient experience, determine which specific results you most want to improve. Questions may include:

  • Who is the “patient” or consumer of your health care offering?
  • What is “success,” and how will you measure it?
  • How does your role in the continuum of care position you to achieve those results?

Once you have addressed questions like these, you need understand if the data supports your definitions and metrics in order to to quantify outcomes. Performing Exploratory Data Analysis (EDA), can help you evaluate the feasibility of your targets and develop an appreciation for the data you have as well as requirements for what you may need. EDA brings together samples of your patient related data and uses visual and quantitative methods to understand and summarize a dataset, without making any assumptions about its contents.

Understand patient segments

If you’re having trouble answering that first question—who is the “patient” or consumer of your health care offering—you can use your data to perform segmentation, and define different groups of patients with different sets of needs. This will allow you to create a more personalized experience for each segment of your patients. Here are some specific examples of things to try:

  • apply historical models and insights to new patients in order to quickly put them into segments to better manage their experience
  • test and iterate on programs targeted to specific segments
  • create opportunities for personalized health care, based on unique patient and segment characteristics
  • understand patient migration between segments, including leading indicators and remediation strategies
  • map channel and engagement behaviors since some demographics may “customize” how they manage their care beyond what is assumed

Engage the larger ecosystem

Each participant in the health care ecosystem is trying to optimize their services and outcomes around what they actually know; then they purchase, infer, or predict information to get at what they do not. Over the last few years, many organizations have been carefully developing data-sharing partnerships for very specific uses while keeping their competitive advantages close.

There are distinct opportunities for anyone aggregating health-related data. EDA will not only reveal any gaps in your own data, but will also help you understand the value of the data you create through your business operations. Think about how it can be used beyond its primary transaction or interaction purposes, and consider who else might find it valuable.

Next, explore who has puzzle pieces of data that would enrich your own data for better patient engagement. How might you share that data and what could you get in return?

Detect and predict patterns to drive adherence

In order to achieve the desired outcomes, a patient must adhere to their prescribed treatment regimen. Pattern detection can be used to understand the patient’s behaviors in following the desired protocols. Once you understand patterns, you can assess the challenges a patient may be encountering and potentially predict the path of their care through your offering.

Is the patient at risk of going off therapy? Is he showing signs of dealing with complications or comorbidities? Is she risking a lapse in coverage? Or are you simply seeing seasonality, such as the patient taking a break over the holidays?

Applying techniques such as machine learning to your patient related data can help you detect these patterns and predict potential outcomes—including who or what may be influencing the patient, the degree of the influence, and what types of intervention may lead to better adherence.

What next?

To truly make the most of your patient data, you need to first establish your business goals. Once those are in place, and agreed upon among your various stakeholders, look at the data you have available. Map that data to those business goals and identify any gaps that might exist. From there, you can prioritize, knowing that you are being deliberate in your choices.