A featured contribution from Leadership Perspectives, a curated forum for enterprise technology leaders, nominated by our subscribers and vetted by the CIOApplications Editorial Board.
Express Scripts
3 Ways Data Can Improve Health Outcomes
The healthcare industry also produces an incredible amount of data. Yet, in many cases, the information and insights derived from this data are inadequate. I believe big data is useless until it is made actionable. Using big data to generate actionable insights can help make better health more affordable and accessible. Here are three ways:
1. Addressing the Opioid Epidemic through Proactive Intervention
The fast-growing opioid epidemic has been called the worst drug crisis in American history. There is no silver bullet for addressing this epidemic. Solutions need to work across the care continuum of physician, pharmacy and patient, leveraging data to proactively identify areas of risk.
Real-time pharmacy data provides the opportunity to uncover insights that can be acted on — quickly and efficiently. Unlike medical claims data, which may appear weeks after the healthcare action and only a few times per year, pharmacy data come in real-time upon each prescription fill. Using this data, outliers can be identified – a doctor, pharmacy, or patient behaving in ways we would not expect and that could indicate opioid drug abuse.
|Using data, clinicians can be empowered to identify and address risks before it’s too late. This could include outreach from a doctor or pharmacist to the patient to discuss the risks associated with opioid use even before filling a prescription. It could also include academic detailing to empower physicians with more information as they make opioid prescribing decisions.
Advanced analytics and artificial intelligence provides a cutting edge capability to determine the health risk of individuals
2. Avoiding Downstream Costs and Negative Health Events
When patients don’t take their medication, it creates serious issues. Non-adherence can prolong a condition, leave symptoms untreated and, in some cases, lead to death. And that’s just the physical toll on patients. It’s also estimated to cost the healthcare industry more than $300 billion a year in more expensive medical treatments such as ER visits and hospital stays.
To address non-adherence, machine learning algorithms can uncover features about patients, physicians, diseases, and prescribed therapies to identify the individuals most likely to stop taking their medication.
For example, our machine learning algorithms uncovered that males with female physicians are less likely to be adherent to their treatment plan. Conversely, living with a partner, or having a higher income, will increase your likelihood for adherence.
It’s not enough to know who is going to be non-adherent. It also requires proactive intervention.
If a patient is likely to become non-adherent due to behavioral factors such as procrastination or forgetfulness (the majority of the instances), we could send the patient daily alerts, 90-day fills, or auto-renewals. If a patient is likely to have clinical questions or concerns about the medication, the patient would receive a pharmacist consultation. If high cost is the primary concern, the patient maybe a candidate for payment assistance programs, lower cost medication alternatives and lower cost pharmacy options such as home delivery.
3. Predicting Personalized Risk to Influence Better Health
Advanced analytics and artificial intelligence provides a cutting edge capability to determine the health risk of individuals. By layering medical, lab and non-traditional data onto already robust pharmacy data, we can identify and/or predict the potential for very personalized healthcare gaps. More importantly, we can recommend actions necessary to change patient behavior and close the identified gaps. For example, my team identified a patient with diabetes, depression, and chronic pain with open pharmacy and medical gap opportunities. A diabetes specialist pharmacist counseled the patient to work with her physician to measure her A1C, which is leading indicator of well-managed diabetes. We highlighted clinically-equivalent generic medication options that may be more affordable to the patient. And finally, we shared tips for better adherence to treatment plan, which prevents downstream health costs. By closing these very specific, very actionable gaps, we can move a patient with high risk index to low risk index. Advanced analytics and big data are not intended to replace clinical expertise or automate diagnosis. Rather, advanced insight allows prescribers and patients to optimize clinical care. This helps save patients money, lower plan costs and – most importantly – avoid future health risks to ensure better outcomes.
