This post is provided by one of our Medicaid Star Search panelists. Be sure to check out Forecast Health on the April 3rd show. Sign up here if you haven’t already.
If asking the right questions is everything, why are so many Medicaid plans asking the wrong ones?
Today, plans and providers dedicate time and resources to identifying high risk patients — those who are more likely to experience a readmission, post-operative complication, higher costs, or other adverse event. To do this, they rely on a variety of popular analytics platforms that are only marginally better than clinical judgment (or making educated guesses, frankly) at predicting a patient/member’s risk.
For organizations charged with reducing and managing risk, not just measuring it, merely answering questions about a person’s future risk is wasteful. We believe answering the question about future risk is wasteful even if the prediction is perfect.
Why? Many predicted high-risk people, and rising risk people, are not impactable. Or if they are, the plan or provider doesn’t know which type of intervention is likely to have the biggest impact. Or, perhaps the provider or plan does not offer services that would make a difference in preventing the adverse event.
A more useful way to look at the data is to focus on the subset of high-risk patients who are impactable. By identifying this subgroup, Medicaid plans and providers can focus resources in ways that will make a difference.
So exactly what could this mean to a Medicaid plan? Consider one procedure with among the highest costs and risks for a Medicaid population: joint replacements. An organization that oversees approximately 3,000 joint replacements annually asked us to evaluate their decisions regarding post-acute care and managing risk. Using a combination of machine-learning predictive analytics and evaluation analytics with social determinants of health (SDH) data to predict risk and impactability, they were able to:
- More accurately identify joint replacement patients at high-risk for unplanned readmission in the 90-day post-discharge period
- Identify the 14% of patients who were discharged from the hospital to a skilled nursing facility, but could have safely been discharged home.
- Identify the 7% of patients who were discharged home, but would have had better outcomes and lower costs if they had been initially discharged to a SNF (the costs of the more expensive SNF were offset by the reduced emergency readmissions).
- Identify patients’ individual risk drivers—prior to admission—in order to better inform the care team of potential issues with affordability, caregiver support, medication non-adherence, transportation, material deprivation, and health behaviors. This information provides opportunities for care managers to spend more time addressing the needs of the patients versus trying to discover exactly what those needs might be via intake questioning and exhaustive chart review.
Together, these analytics can save on average over $800 per joint replacement patient, or $2.5 million annually based on 3,000 joint replacement patients.
The bottom line is this: Asking the right question isn’t just an academic exercise. By moving from just predicting risk to identifying who is impactable, and how, an organization can measurably impact costs and patient outcomes. For more information, please contact Sandy Shroyer at firstname.lastname@example.org