Predictive Analytics and the Future of Healthcare

Intel provides a foundation for big data platforms and AI to advance health analytics.

Benefits of Predictive Analytics in Healthcare:

  • Predictive models can help keep patients healthier by anticipating the need for emergency care or the likelihood of a serious condition before it develops.

  • By identifying which patients are likely to be readmitted to a hospital, predictive analytics can help providers direct additional care where and when it’s needed.

  • Hospitals can schedule resources more effectively by predicting the length of patient stays.



Rising costs, an aging population, and the prevalence of chronic conditions are transforming the healthcare industry. By 2030, global healthcare spending is expected to reach an unprecedented USD 18.3 trillion.1 In response to these trends, payment models are already shifting from volume based to outcome or value based.

Predictive analytics is helping health organizations align with these new models while helping to enhance patient care and outcomes. From anticipating critical conditions, like septic shock and heart failure, to preventing readmissions, the latest advances in big data analytics and AI are fueling new predictive analytics solutions that help clinicians improve outcomes and reduce costs.

Harnessing Data for Predictive Health Analytics

Healthcare has become digitized, creating massive new data sets. These include electronic medical record (EMR) systems, health claims data, radiology images, and lab results. In the near future, genomics data will also grow significantly.

New data is also being generated by a growing number of medical devices at the edge, including patient wearables and monitors. Outside the clinical setting, patients are generating quasi-health data through the use of personal wearable devices, fitness trackers, and health applications.

By incorporating data from these sources, health providers can power new solutions in predictive analytics for medical diagnosis, predictive modeling for health risks, and even prescriptive analytics for precision medicine.

However, converting data into clinical results requires a foundation of hardware and software designed to extract value from disparate data sets. One survey found that more than half of health organizations do not have a comprehensive data governance plan in place.2 As a result, a significant portion of healthcare data remains untapped.

With a portfolio of technologies designed to efficiently move, store, and process data, power big data platforms, and run AI models, Intel and our partners are working with health organizations to put predictive analytics to work.

Predictive analytics powered by an Intel® Xeon® processor-based big data platform empowered a large hospital group to save $120 million in annual costs.

Benefits of Predictive Modeling in Healthcare

Predictive analytics has become a key piece of any health analytics strategy. Today, it’s a critical tool for measuring, aggregating, and making sense of behavioral, psychosocial, and biometric data that until recently was not available or exceedingly hard to capture.

At the individual level, predictive analytics can help health providers deliver the right care to the right patient at the right time. On a larger scale, it can enable health systems to identify and understand larger trends, leading to improved population health strategies.

In one example, researchers developed a model of how Ebola is spread using big data analytics and massive amounts of data, including information from social media and search engines. Individuals who have potentially been exposed to Ebola can enter their symptoms into a mobile app, which uses geocoordinates to check whether the person has been in proximity to someone from a community in which Ebola has been active.3

Not only can predictive analytics enhance care, but it can also dramatically reduce costs. For example, more accurate prediction models for patient length of stay and readmission rates enable hospitals to avoid penalties and reduce operational expenses. By tapping into electronic health records (EHRs) and predictive analytics, providers can flag patients who are likely to miss an appointment. Once identified, those patients could be reminded or otherwise supported in keeping their appointment.

The enormous potential of predictive analytics includes helping identify patients at risk for chronic conditions, developing evidence-based best practices, and proactively spotting potential obstacles to care plan adherence. Data can help clinicians remain one step ahead of events, delivering proactive care to patients before their health becomes critical.

Examples of Predictive Analytics in Healthcare

Today, health systems and providers are exploring different ways to use big data platforms and AI for predictive analytics. These solutions are helping health organizations transition from simply using data to learn what already happened to using that data to more reliably forecast what will happen.

Speeding Treatment of Critical Conditions

Working with Intel, Penn Medicine created a collaborative data science platform to help predict and prevent two of the most common and costly issues for hospitals: sepsis and heart failure.

The predictive model was able to identify about 85 percent of sepsis cases (up from 50 percent) as much as 30 hours before the onset of septic shock (as opposed to two hours using traditional methods).4 It was also able to identify between 20 and 30 percent of heart failure patients who had not been properly identified.4 These efforts empowered clinicians to deliver treatment sooner, speed time to recovery, and save resources for the hospital.

Predicting Length of Stay

Intel and Cloudera helped a large hospital group use predictive analytics to provide enhanced accuracy in predicting length of stay. The big data platform based on Intel® Xeon® processor clusters allowed the hospital group to ingest unrelated, unstructured, and semistructured data.

With the ability to plan and staff more efficiently, the hospital group saved USD 120 million in annual costs (about USD 12,000 per patient) and boosted facility utilization by 5 percent, allowing hospitals to potentially serve an extra 10,000 patients annually.5

Reducing Readmissions

In another effort, Intel and Cloudera used socioeconomic data, EHRs, and predictive analytics to help a hospital group identify patients with a high readmission risk at the time of diagnosis. Hospital staff could then provide additional medical care to reduce readmission rates.

The big data platform, powered by Intel® Xeon® processors, has enabled the hospital group to reduce 6,000 occurrences of patient readmission, avoid USD 4 million in potential Medicare penalties, and save about USD 72 million annually in medical service costs.6

Exploring the Potential of AI

Intel is passionate about using AI to help health systems and providers fight disease and personalize treatments. From our sponsorship of an AI-focused cancer-screening competition to the many Intel® technology-powered AI solutions in health and life sciences, Intel is helping health organizations find the right technologies to deploy predictive analytics.

Identifying Patients at Risk of Decline

Sharp HealthCare used technologies from Intel and Cloudera to successfully deploy a predictive clinical analytics model. The model used machine learning and data from the hospitals’ EMR system to identify patients at risk of needing an intervention from the rapid response team in the next hour.

The model was 80 percent accurate in predicting the likelihood of an event within an hour.7 This empowered rapid response teams to intervene proactively, improve the quality and cost of care, and enhance resource utilization.

Intel Supports Clinicians with Predictive Analytics

By providing a foundation of technology for AI and big data platforms, Intel and our ecosystem of partners are helping health providers tap into the vast amount of patient and health data that remains unused. The resulting solutions can help providers advance patient safety, enhance operational efficiency, and most importantly, improve patient outcomes.