Predictive Analytics: Forecasting the Future of Patient Health
Who knows what tomorrow may bring?
No one holds a crystal ball to perfectly foretell the future. Yet today—thanks to predictive analytics—we have many clues. This technology uses data, statistical algorithms, applied artificial intelligence, and machine learning techniques to predict the likelihood of future outcomes. Used in many industries, from financial services and insurance to retail and manufacturing, it is now rapidly gaining ground in healthcare. Providers and payers can continually monitor and analyze huge amounts of data from patients to understand their current state—and predict what might happen next along their healthcare journey.
“As virtual care becomes less of an art and more of a science, clinicians will rely increasingly on predictive analytics for decision support to help determine the best course of action for each patient,” says Jeff Haggard, vice president of customer development at Vivify Health, an international digital health innovator.
Transforming data into better patient outcomes
Through predictive analytics, providers are discovering new ways to leverage patient-generated health data—delivered through remote patient monitoring (RPM) technology—and use it to improve patient outcomes. With this data, providers can efficiently manage rising-risk and at-risk populations, triage patients who need early intervention, and reduce preventable hospitalizations, especially for those with complex chronic conditions.
“We give the caregiver teams the tools they need to look at data sets and identify individual patient patterns,” Haggard says, “For example, with Vivify’s managed kit and BYOD (bring-your-own-device) technology solutions, patients are monitored continually and remotely, providing daily biometric measurements back to their caregivers. As information is collected and transmitted, care teams look at the data and can intervene as needed to keep patients out of the hospital. They can also aggregate data from populations of patients, such as those with primary diagnoses of diabetes or congestive heart failure, to look at—and learn from—their cumulative patterns.”
Taking in the total picture, real time
“The key differentiation is that our data is not based on clinical episodes or claims data, but on the real-time, day-to-day picture of a patient’s health,” says Vivify founder and CEO Eric Rock. “It includes not only biometric readings and symptomatic responses but also subjective data that may reflect on social determinants of health or changing conditions in a patient’s home. Data is provided in real time, enabling care teams to intervene on the spot, continually adjusting clinical engagement protocols for improved health.”
Vivify’s leadership team has a combined 30 years of experience using predictive analytics. Rock’s first company, acquired by OpenTable.com, used predictive analytics inside restaurants to help solve the complexities of balancing resources, improving workflows, and accurately forecasting wait times. His second company, MEDHOST, provided predictive analytics and process management technology for hospital emergency departments. “With Vivify, we have the opportunity to take these innovations to larger and larger populations,” he says.
Testing the power of machine learning
In the future world of healthcare, we can look forward to even more precise predictive analytics through machine learning, the science of getting computers to act without being explicitly programmed. “Predictions will become even more real time and more dynamic,” says Rock. “Massive data sets will be transformed into actionable intelligence—finding personalized patterns and then translating them in a way that enables providers to accurately predict any impending health events for a specific patient. The more patients the system ‘sees’ and the more information it receives, the more the system will learn and the more precise its predictions will be.”
For example, if data analysis suggests a hospital admission in five days, the machine will learn whether or not this prediction is accurate. Predictive models will become more effective, “false positives” will decline, and hospitalizations will be predicted further in advance—with increasingly high accuracy.
Transitioning the role of payers and patients
Providers aren’t the only healthcare stakeholders to benefit from predictive analytics. Payers, too, can derive significant value. “As payers continue to expand their role in care delivery, the role of predictive analytics to enhance virtual care is paramount,” Haggard says. “Payers have incentives to keep patients healthier and out of high-cost settings of care delivery. RPM and predictive analytics are exceptional tools for managing rising costs.”
Patients, too, gain from improved health outcomes and a better healthcare experience. “Receiving more frequent care in their own time and in their own setting feels natural to patients,” Rock notes. “Often we hear them say, ‘This is what I imagined healthcare could be all along—delivered to me, on my own terms.’ As a result, they are more engaged, provide even more data, and are more likely to stay the course for health improvement.”
Thriving in the new world of care
Today, with leading technology such as Vivify’s, patients can be connected to their care, anytime and anywhere. “Compared with episodic inpatient visits, continually captured patient data—obtained through RPM and backed by predictive analytics—fills a huge void in care,” Haggard says. “It can reveal health deterioration in high-risk patients earlier and with higher accuracy, extend care manager reach, and forecast the future of patient health.”