Prescriptive Analytics in Healthcare - What Does It Mean?
My last two posts addressed a general overview of the healthcare analytics continuum, and the importance of foundational data integrity to healthcare analytics. For today's post I'd like to dig a little deeper into the difference between descriptive, predictive and prescriptive analytics, with a focus on prescriptive analytics, as I've found it has different meanings for different users.
It goes without saying that data analytics has become increasingly integral to the health care industry. From electronic health records to CMS reporting requirements, more health care data are being collected now than ever before. Over the previous two decades, health care data management systems' capabilities have expanded from simply recording information to analyzing pertinent data to provide evidence-based decision support.
Advanced analytics functionality can improve patient care by producing data-driven actionable insights. However, payers, providers, employers, brokers and other stakeholders are still figuring out how to utilize all of the information available to them. As in other industries, health care data analytics solutions can be categorized in three levels: descriptive, predictive and prescriptive. Prescriptive analytics builds upon the foundation of descriptive and predictive solutions.
Descriptive analytics: Recording what is
When healthcare analytics applications were first introduced, their objectives were to track and report plan performance and trends (cost, quality, utilization) in the past tense. What is my average claim cost compared to last year? How many diabetics in my population have had their HbA1c test? What are ER visits per 1000 compared to a benchmark?
Deerwalk's Executive Analytics ER Visit Utilization Dashboard
The digitization of medical records and the ability to collect many unique data types (eligibility, medical, Rx, lab, biometric, HRA, wellness, etc) has enabled healthcare stakeholders to quantify information, and make the resulting data more accessible. Authorized entities are now able to transmit data more easily and generate detailed reporting and analytics. Collecting of this available data in an organized systematic way is the first step in data analytics.
Utilizing descriptive analytics allows payers, providers and other key stakeholders to better understand the facts, including health history, costs and population statistics. When health plans can identify populations that are consuming more resources, they can begin developing health management programs to improve outcomes and reduce costs.
Predictive analytics: Forecasting what will be
Organizations that are running a comprehensive descriptive analytics program can begin to use that data over time to predict future outcomes. When complete, accurate data are available and when properly analyzed, payers can develop evidence-based projections. Predictive analytics is especially valuable for utilization management. Identifying correlations, trends and probabilities allows payers to better identify high risk members, evaluate overall risk and prepare for future needs.
Deerwalk's Demographic Risk Analysis Dashboard
Predicting future outcomes requires more than just parsing past events. Effective predictive analytics tools incorporate near real-time data to enable more dynamic decision-making. And to create really valuable analytics tool, organizations need a more robust platform than a standard data warehouse.
Forecasts are only as valuable as the data that they are based on, so uniform collection and strong data governance standards are critical. Analyzing this information requires integrating with a robust population health analytics tool, like Deerwalk's Plan Analytics. Implementing a predictive analytics program enables payers and employers to move from reflecting on what is to planning for what's next.
Prescriptive analytics: Influencing outcomes
The natural progression for organizations that have strong predictive capabilities is to take action on those insights. Prescriptive analytics expands upon the foundation built by descriptive and predictive analytics to provide actionable recommendations and to change predicted outcomes.
For example, if a payer was experiencing an increase in ER utilization, a prescriptive analytics tool would do more than note the issue (descriptive) or project future ER utilization (predictive). The system could also pinpoint avoidable ER visits, and calculate potential savings associated with those avoidable ER visits (prescriptive). When used consistently, the tool could evaluate the outcomes of its decision support recommendations and use that information to improve future suggestions.
Deerwalk's Avoidable Care Continuum Dashboard
To offer evidence-based insights, a prescriptive data analytics system needs to be fully integrated with all of the pertinent sources of data. A well-defined QA process is also necessary to scrub, filter and enrich the data to transform individual components into seamless data sets. SaaS tools such as Deerwalk's Data Factory offer a cost-effective way for health plans to ensure data quality and develop a robust analytics program.
The possible uses for prescriptive analytics are limited only by the availability and reliability of data, and the willingness to build prescriptive models, or partner with an analytics firm like Deerwalk that is already building these models for its clients. Ultimately, technology and organizations should have the same goals: to improve patient care, reduce costs and increase efficiency and quality, and investing in prescriptive analytics solutions will assist in these goals.
To learn more about Deerwalk's Population Health Analytics solutions, click below to contract us and schedule a demo.
What do you think? Subscribe to our blog, facebook, twitter or LinkedIn page to join the conversation and tell us what you think.