Top 5 Takeaways from the 2022 Healthcare Analytics Summit
10/28/2022 by Chris St. Jeor
As someone who is obsessed with healthcare analytics, I was thrilled to attend the Healthcare Analytics Summit (HAS) last month. The conference attracted people of diverse backgrounds who boast a wealth of knowledge and experience. However, because not every clinician, executive, or analyst within an accountable care organization ecosystem could attend, I’ve summarized the top 5 takeaways you should know about from the summit’s panel discussions, breakout sessions, and showcases.
With healthcare’s migration towards quality-of-care reimbursement models over fee-for-service models, healthcare is laser-focused on creating robust population health management strategies. Understanding the different segments in your patient population is critical to identifying emerging risks, caring for your high-risk populations, and avoiding unnecessary high-cost medical events.
Building a quality population health strategy requires quality data. The most successful organizations align their electronic medical records (EMR) data with their claims data. This robust view allows them to align their patients’ comorbidities with their other social determinants of health factors, which in turn, saves organizations millions of dollars by optimizing patient care with smarter insights and interventions.
Population health is key to driving improved clinical, financial, and overall health outcomes.
With the rise of machine learning and artificial intelligence in healthcare, we need to be cautious of an inherent pitfall: bias against minority populations. Research has shown that even in our most state-of-the-art algorithms, these predictive models perform worse when predicting outcomes of minority populations.
Dr. Marzyeh Ghassemi discussed that while machine learning models have proven the ability to perform at, or above, human doctors across the range of clinical tasks of the human lifespan, these same models often unintentionally identify proxies for other social determinants of health identifiers, and can use these variables to widen the equity gap in the care received by minority patients. When audited, these models have been found to have differences in model accuracy across race, gender, and insurance type.
Machine learning and artificial intelligence models must be audited before pushing them to production. Predictive performance should be stratified across the patient populations to ensure that the model is performing consistently across all segments of the population. Healthcare providers should also review the information from these models with a critical eye. Providers should not just accept the output at face value, but should critically think about whether the recommendations and predictions make clinical and ethical sense.
Annual wellness visits (AWV) have compounding benefits for your patient population’s health outcomes. One accountable care organization (ACO) outlined the benefits from an intervention campaign where registered nurses reached out to patients with a gap in their AWV. Following the campaign, the ACO found that patients who received an AWV had the following benefits:
Increasing participation in AWV is one of the best ways to improve your members participation in preventative and management care and decrease unnecessary or avoidable spending.
Emergency Department (ED) visits are expensive, and many of the visits are avoidable. Successful population health solutions track ED utilization by each patient. Members with high utilization should be assigned social care workers to make sure the patient is getting the care where and when they need it. Gaps in social determinants of health, such as transportation needs, should be filled to make sure that members can get to their scheduled appointments and participate in their preventative and maintenance care appointments.
Data literacy is one of the biggest roadblocks that ACOs and other quality care organizations face when trying to use analytics to drive better patient care outcomes. Clinicians, management teams, and executive teams should be trained on how to interpret the results of the content the analytics team provides. As the analyst builds the management team’s confidence in the analytics and their ability to interpret the results, the management team can make data-driven clinical and business decisions.
Healthcare is quickly moving its focus to quality-of-care versus quantity. Zencos has years of experience helping organizations adopt many of the best practices highlighted above. Whether working with commercial contracts, Medicare, Medicaid, or some combination of the three, understanding your population health and gaps in social determinants of health is the first step to thriving in the quality-based ecosystem. Contact us to see how we can help, or at the very least, attend the Healthcare Analytics Summit next year. I’ll see you there!