Current pressure to deliver affordable, convenient, consistent, and personalized healthcare requires organizations to make financial, operational, and clinical decisions using foresight and not hindsight. Forecasting what may happen and intervening to avoid negative events brings organizations better control over outcomes. For example, predicting that a patient is likely to be re-admitted allows the health system to avoid that occurrence by addressing the drivers such as confusion over medications (i.e., medication reconciliation), or by confirming follow up appointments with the patient and providing needed transportation, or by installing remote monitoring devices to ensure patients remain stable. Without these or other interventions, a patient may end up back in the hospital.
Predictive analytics use cases span clinical, financial, and administrative healthcare domains. From a clinical perspective, understanding that a negative clinical event like sepsis, the onset of a chronic condition, or admission/re-admission to the emergency room will occur means the clinical team can intervene early to avoid the event entirely or reduce the negative impact. Most of the commercialized predictive analytics use cases involve artificial intelligence (AI). These applications provide algorithms that have embedded AI technologies that are central and critical to the application’s function. If those AI technologies were removed, the application would fail to work.
Among the greatest challenges to the deployment of predictive analytics is the harvesting, cleansing, and normalizing of data to drive the analytics. Healthcare data is known for its heterogeneity, which forces tedious extraction and mapping of data from multiple electronic medical record systems as well as core administrative and financial systems. If an organization is going to adopt an AI-predictive analytics tool, then large amounts of its data will need to be available “on demand” during the early phase of training and validating the AI models.
The good news is that healthcare organizations have access to more electronic health data than ever today. Our research shows that they have been making steady investments in big data and analytics platforms. Forty percent of healthcare providers and 69% of payers report that they deployed big data technologies enterprise-wide, in departments or business units, or have pilot projects underway, according to IDC’s 2018 Industry IT & Communications Survey. Results are similar for business analytics with 45% of provider organizations and 68% of payers indicating that they have deployed that technology. Managing this increasing volume of data from multiple sources, including nontraditional healthcare data (e.g., weather patterns, social determinants of health) will require CIOs to reconsider their data infrastructure strategy.
Among the considerations is that legacy architecture will only get healthcare organizations so far. Key pain points for the IT organization include inflexible architecture and constrained networks. Healthcare organizations are evaluating cost-effective technologies that provide the foundation for networks across the distributed enterprise. Software-defined wide area networks (SD-WANs) enable enterprises to optimize the utilization of cloud/SaaS applications by providing a dynamic and secure WAN fabric for distributed enterprises. Flexibility, scalability, and simplification thus become important considerations for SD-WANs.
To learn more about the adoption of predictive analytics in healthcare, read the IDC Analyst Connection, “What are the Infrastructure Implications When Adopting Predictive Analytics,” sponsored by CenturyLink.
Cynthia Burghard is a Research Director with IDC Health Insights where she is responsible for the value-based healthcare practice.
This blog is provided for informational purposes only and may require additional research and substantiation by the end user. In addition, the information is provided “as is” without any warranty or condition of any kind, either express or implied. Use of this information is at the end user’s own risk. CenturyLink does not warrant that the information will meet the end user’s requirements or that the implementation or usage of this information will result in the desired outcome of the end user.