Implementation of predictive analytics in the global healthcare market reached $1.8 billion in 2017 and is estimated to reach $8.46 billion by 2025, according to Allied Market Research.
Buoyed by fast-developing advances in artificial intelligence and machine learning as applied to business intelligence applications, healthcare organizations are able to deploy predictive tools enterprise-wide that simultaneously improve patient care, reduce costs, and increase profitability.
From risk assessment to predicting imminent emergencies, applied artificial intelligence is helping to make significant inroads in trimming unnecessary costs, increasing efficiency, improving patient satisfaction, and managing healthcare staff in the healthcare sector.
As increasing numbers of healthcare executives are embracing more sophisticated big data analytics, they are moving from basic descriptive analytics to predictive insights that can inform and guide more precise decision-making across the board. They see digital technology as an effective means to support value-based care that provides better patient outcomes while reducing costs.
Predictive analytics can provide alerts to everyone from clinician and financial experts to administrative staff to inform them when events they are concerned with are likely to happen. As a result, they can make more informed, more timely, and usually better choices at times when small nuance can have big consequences. Whether a patient’s life depends on a short response time, or whether a proper candidate for clinical trials can be identified in time for treatment, predictive analytics benefit the patient and the organization.
Improving Care While Reducing Costs
Identifying, classifying, and managing high-risk patients “is central to improving quality and cost outcomes,” according to the Association of American Medical Colleges (AAMC).
Prediction and prevention work together to trim spending and improve health care. When predictive tools allow an organization to identify individuals at high risk of developing poor health outcomes and chronic conditions early in its progress, they help them avoid costly, long-term health problems and their complications. They can play a large role in reducing costs as well.
Developing risk scores—based on data from lab tests, biometrics, claims, and patient health data, among others—can provide information to identify the patients who might benefit from a selection of services and activities and allow healthcare organizations to provide more effective treatment solutions.
Predictive analytics can also warn providers when a patient’s risk factors indicate a high likelihood for readmission within a 30-day window. By identifying patients with a higher chance of being readmitted, providers gain the advantage of knowing when to focus resources on follow-up appointments and how to build discharge protocols that reduce the likelihood of patient returns to the hospital.
In a 2016 study from the University of Texas Southwestern, researchers found that certain occurrences during a hospital stay—including certain infections, instability of vital signs, and longer duration of stay—resulted in a significantly greater chance of patients being readmitted within 30 days.
The LACE Index factors in the length of stay, co-occurring diseases, and visits to the emergency room to identify patients at risk of readmission or death within 30 days of the date of discharge from a hospital. Based on these health factors, the index helps caregivers predict the course of health of the patient, make optimal decisions for their treatment, and reduce mortality rates.
Medicare’s Hospital Readmissions Reduction Program (HRRP) adds a financial incentive to implement these analytics sooner rather than later. Under HRRP, hospitals and healthcare organizations are subject to significant penalties for excess admissions. Predictive analytics can help organizations reduce the number of unexpected re-admissions and avoid these extra costs.
Predicting and responding to patient decline
Hospital-acquired conditions (HAC) increase hospital costs by $2 billion, or about $41 thousand per patient, according to a research brief by IBM Watson Health.
Once admitted to a hospital, patients often face a variety of possible threats to their health that can add extra risks and extra costs to treatment. Conditions such as sepsis, pressure ulcers, hard-to-treat infections, kidney injury, or poor reactions to treatment can impede patient improvement, or worse.
Data analytics and predictive analytics can help providers respond more quickly and precisely to downturns in patient health by monitoring vital metrics and alerting providers about signs of decline before the condition might be detected otherwise. Machine learning applications are especially well-suited to predicting trends and clinical events in hospitalized patients.
For example, a 2017 study reported that a predictive analytics tool using machine learning and electronic health record (EHR) data at the University of Pennsylvania helped to identify impending cases of severe sepsis or septic shock 12 hours before the onset of the condition.
Another study by the American Medical Informatics Association revealed that by combining predictive analytics and clinical decision support tools, sepsis mortality was reduced by 53% and the 30-day readmission rate was reduced from 19% to 13%.
Improving Operational Efficiency
Anticipating patient flow patterns
Predictive analytics can also modernize methods for scheduling patients by setting up more flexible systems that respond more accurately to demand. Using analytics to predict usage and patient visitation patterns can help optimize staffing planning and schedules while reducing patient wait times, avoiding overflow in waiting rooms, and improving patient satisfaction.
Emergency facilities and urgent care centers must accommodate fluid staffing levels to account for fluctuations in patient populations. Inpatient services need to keep beds available for new patients, and outpatient clinics and physician offices are invested in minimizing wait times for patients.
Analyzing typical utilization rates can help care centers recognize trends such as spikes or lulls in capacity. Predictive tools allow them to aggregate a variety of doctors and treatments and smooth out the variances for optimal use of time and resources.
At Wake Forest Baptist Health in North Carolina, resources were “stressed to the max during peak hours,” according to Karen Craver, Clinical Practice Administrator. At other times, staff members had few patients to care for and little work to do. Using analytics, the center was able to “flatten out the bell curve” and experience a more even distribution that evened out the workflow and improved patient satisfaction.
Reducing costly no-shows
Most healthcare providers experience significant financial impact from appointment no-shows. With the help of predictive analytics, however, they can identify patients who are highly likely to miss upcoming appointments without notice, and in turn offer open slots to other patients and minimize the organization’s potential revenue losses.
Electronic health data can help identify patients who are most likely to no-show, according to a study from Duke University. Predictive models using clinical data captured 4800 more patient no-shows per year than previous attempts to forecast patient patterns had been able to identify.
Providers might be able to use the data to send appointment reminders to patients to reduce the risk of no-show or offer other dates and times that might fit their schedule better.
A pilot program at Boston-based Brigham and Women’s Hospital was able to reduce no-show rates by 30% for colonoscopies by texting digital prep guides, appointment reminders, and relevant links to patients with upcoming appointments. Patients reported feeling more prepared for the procedure.
That’s a significant ROI,” says Adam Landman, vice president and CIO of Brigham. “We have some very encouraging results, and we’re starting to expand that tool to other procedural areas and use cases. This is an example of where it’s meeting a real need, and there are some palpable ROI that resonates with our CFOs that lets us then build this platform out.”
Streamlining the supply chain
Not surprisingly, one of the most significant arenas for healthcare organizations to reduce costs and improve efficiency in their supply chain, and predictive tools can help. Predictive tools are in high demand among hospital executives looking to reduce variation and gain more actionable insights into ordering patterns and supply utilization.
While only 17% of hospitals currently use automated or data-driven solutions to manage their supply chains, predictive tools can quickly reduce variation and provide more nuanced actionable insights into ordering patterns of supply utilization.
Using analytics tools to monitor the supply chain and make proactive, data-driven spending decisions could save hospitals as much as $10 million per year, according to a survey by Navigant. Both descriptive and predictive analytics can help in the process of negotiating pricing, reducing variations in supplies, and optimizing the ordering process.
Organizations can benefit from intruder alerts and other changes in the data systems with analytics tools that monitor data access, sharing, and utilization. For example, predictive tools and machine learning algorithms can evaluate real-time risks for a variety of transactions and respond based on the results of the evaluation. Once a risk assessment is made in real time, the system can either grant access, require further authentication, or even block access based on the data.
In the future, it’s likely we’ll see a convergence of applications and solutions within a seamless platform that delivers unprecedented effectiveness of care, highly responsive and customized treatments, and exceptionally efficient systems that reduce cost and optimize profitability. Tools that combine hospital-derived data, patient-reported outcomes, clinical decisions, and analytics present exciting opportunities to streamline organizations within the healthcare sector with benefits that are felt by patients and providers alike.