Keith Loria is a contributing writer to Medical Economics.
Operational efficiencies, patient-centered care, and outcomes will get a boost.
Complex problems that affect the world of pharmacy, such as the opioid crisis and avoidable hospital readmissions, require multidimensional answers. Artificial intelligence (AI) might be one solution. There are currently firms working to use AI, some through the use of machine learning, in the hope that it will contribute to solving these problems or at least minimize their impact. Taken together, AI and machine learning may have a massive multifaceted impact on pharmacy operational efficiencies, patient-centered care, and outcomes.
In today’s world, AI complements human interaction with patients. Adam Beacham, director of business intelligence at PDX-NHIN, a pharmacy software company, believes that 2019 and 2020 will be pivotal years for AI and predictive models to improve patient health from prescribers to pharmacies.
Sadiqa Mahmood, DDS, MPH, senior vice president of medical affairs for life sciences at Health Catalyst, says pharmacists can be empowered by AI to shift from prescription-filling roles to patient engagement and management of disease. Health Catalyst is a data/analytics vendor that has been working to predict and prevent readmissions through a combination of predictive analytics, machine learning, and new intervention strategies.
“Tools like AI and ML are enabling decision-making processes at the point of care; [and] quicker identification of patients who become high risk due to changes in their diagnoses, condition, or care plan,” she says. “Just as important, AI is optimizing pharmacy operations for inventory and supply chain management to enhance pharmacy productivity, and increase patient satisfaction and outcomes.”
Beachum adds, “AI and technology in the healthcare vertical continue to expand their use cases. As more organizations begin to leverage the available data, continuing to predict outcomes and patient adherence will expand its role.”
AI and machine learning are primary drivers within health systems as a way to lower the risk of readmissions among patients. Samir Manjure, CEO of KenSci, an AI-powered risk prediction platform, says the adoption of electronic health records and the availability of patient data sets have made it possible to predict more accurately which patients are at the highest risk of readmission. This not only offers the ability to intervene early but also mitigate the risk of infectious diseases and other health complications for the patient.
“With the Medicare Payment Advisory Committee (MedPAC) stating that 76% of hospital readmissions are potentially avoidable, AI and machine learning play a critical role in enabling hospitals to prevent cost leaks owing to potential cases of readmission,” he says.
A challenge to implementing AI is that, in its current state, administrative, clinical and financial systems in healthcare are not integrated, and, in some cases, are handled manually. To deploy AI successfully, a single technology platform that integrates and harmonizes data from disparate sources is critical.
“Institutions with such platforms are able to deploy AI successfully and offer hospital, health systems, and pharmacy leaders a more ironclad defense against preventable adverse drug events and avoidable care delays, and effective operational management,” Mahmood says. “Health Catalyst’s Data Operating System is a cloud-based digital platform, which enables health systems to integrate and analyze data from virtually any software system or other data source.”
DOS contains large and comprehensive data assets of its kind with more than 100 million patient records, encompassing trillions of facts sourced from more than 300 distinct siloed sources.
“Because we integrate data from so many highly disparate sources, we’re able to synthesize a single-source-of-truth data feed, and generate high-quality training sets that machine-learning algorithms can use to continuously improve their performance and accuracy,” Mahmood says. “We give the health systems the right validated information needed to support any AI-based pharmacy initiative.”
PDX offers Explore Dx, software that allows customers to analyze, trend, and drill into specific therapeutic categories, prescriber and patient demographics, and many other areas of interest. With more than 18,000 dimensions and measures available, Beacham explains that Explore Dx helps customers drive patient care today and has the ability to expand into AI to enhance patient care in the future.
“Many organizations struggle to gain access to their own data or provide the resources to become knowledgeable about their data. Explore Dx solves both of those problems,” Beacham says. “By providing a self-service data visualization platform and a knowledgeable team of business and data analysts, Explore Dx closes that gap.”
KenSci’s AI-powered risk prediction platform delivers a readmission solution built using machine learning models. By aggregating patient data from across medical sources, KenSci’s platform is able to predict which patients are at the highest risk of hospital readmissions, providing insight that enables caregivers and health professionals to act early.
“With KenSci, health systems are able to provide better quality care at lower costs,” Manjure says. “Predictive models for identifying patients at risk of readmissions can result in a positive rate of return for organizations by enhancing the patient experience.” Such models improve patient outcomes, including decreased risk of hospital-acquired conditions, complications, and mortality, mitigating costs related to returns to the emergency department, and hospital reimbursement.
Trending: Discount Cards: A Rebuttal
Variables used in the prediction are an aggregate of labs, comorbidity history, demographics, and inpatient history from an electronic medical record during the hospital stay. An interactive dashboard provides patient risk scores.
Preparing for What’s To Come
As business intelligence platforms expand in the market and AI grows in popularity, the need for knowledge of the industry and source systems becomes more important. The technology continues to evolve and gather data at a rapid pace, but needs to be structured and applied for real-world scenarios. Without properly using this knowledge, it will be a slow process to improve patient healthcare and adherence.
Patient-centered care can’t achieve its goal of improving outcomes if the mechanisms needed to support that effort aren’t supported with the right technology. For pharmacists in particular, AI and ML is an excellent way to achieve improved medication and care-pathway adherence, and reduce the avoidable adverse drug events at scale.
“AI and machine learning is the right technology to overcome many of the challenges pharmacists face today-and it’s the right moment in time for pharmacy leaders to address those challenges, because doing so will enable them to meet the future expectations that all payers, providers, and regulatory entities are so clearly signaling,” Mahmood says. “AI also provides pharmacy an opportunity for more collaboration across many different entities serving the same patient, including clinicians and payers.”
Read More: CVS Ramps Up Script Delivery
With the huge investments that health systems have made in creating electronic medical records, the use of AI in healthcare will usher in the era of systems of insights, where technology uses insights to turn data into effective action.
“This is critical to health systems that are focused on improving their quality of care while lowering costs,” Manjure says. “Moreover, with the right insights, healthcare organizations can better align to quadruple aim and drive the ability to forecast and take action based on predictive analytics.”
Data and predictive analytics help any industry understand decisions that have been made, the impact of those decisions, and improve the opportunity to make future decisions with more accuracy.
“As more data is captured and created, it creates better relationships between pharmacists and patients, prescribers and patients, and pharmacists and prescribers,” Beacham says. “As these relationships strengthen through data patient outcomes, medication adherence should improve rapidly over a short period of time.”