In hospice care, there’s little margin for missed signals. When a patient’s condition changes, timely intervention can mean the difference between comfort and crisis.
That challenge is compounded by increasing regulatory complexity. With the introduction of Hospice Outcomes and Patient Evaluation (HOPE) requirements, clinicians must consistently assess and document eight key symptoms, including pain, shortness of breath, anxiety, and gastrointestinal issues. When symptoms escalate, follow-up visits are required, adding pressure to already limited clinical resources.
The challenge is clear: How do you know which patients are most likely to need immediate attention?
A new Axxess white paper examines how predictive analytics can help clinicians identify patients at risk for symptom exacerbation before it occurs.
“Identifying when symptoms are worsening and require fast action is critical,” explained Axxess Senior Data Scientist Phil Gigliotti, PhD. “The goal is to help clinicians focus their attention where it matters most.”
Moving From Reactive to Predictive Care
Traditionally, hospice providers rely on broad monitoring strategies to identify symptom changes. While effective in some cases, this approach can lead to inefficiencies, especially when patient symptom burden increases gradually in any given week.
To address this, Axxess developed a predictive model powered by Axxess intelligence designed to identify patients at high risk for severe symptom exacerbation in the next seven days.
Using hundreds of patient-specific data points, the model analyzes historical patterns to generate a forward-looking risk score. These predictions are updated nightly, giving care teams a continuously refreshed view of where intervention may be needed next.
“We’re not just looking at what’s happening right now,” Gigliotti said. “We’re leveraging patterns across time to anticipate what’s likely to happen next. That’s what enables earlier, more proactive care.”
What the Data Reveals
To evaluate performance, the model was tested across rolling seven-day windows throughout May 2026.
The results highlight just how difficult symptom exacerbation is to predict. These events occur in a small group of patients each week, yet the model is able to concentrate risk into an even smaller, more actionable subset.
That means clinicians aren’t spreading attention across the entire population; they’re focusing on a narrow group where the likelihood of a flare is meaningfully higher.
“When we concentrate risk into a select group, the probability shifts in a meaningful way,” Gigliotti explained. “Instead of guessing across the entire patient population, clinicians can focus on a subset where the likelihood of an event is significantly higher.”
That improvement is reflected in the model’s 11x lift, increasing the likelihood of identifying a true case of symptom exacerbation more than elevenfold compared to random selection.
Why Performance Transparency Matters
Predictive accuracy is only part of the equation. Equally important is helping clinicians understand how much confidence to place in those predictions.
Axxess intelligence addresses this with built-in transparency, surfacing key performance metrics (including recall, precision, and lift) directly within the dashboard.
“Trust is critical when you’re introducing predictive analytics into clinical workflows,” Gigliotti noted. “By showing exactly how the model is performing, we’re giving clinicians the context they need to make informed decisions.”
This transparency enables providers to:
- Quantify uncertainty in real time
- Better interpret risk scores
- Integrate predictions into clinical judgment with greater confidence
Rather than relying on “black box” outputs, clinicians gain clear, measurable insight into how the model is performing and where it adds value.
What This Means for Hospice Providers
Even small gains in early identification can have a significant impact, especially when dealing with high-stakes events.
By narrowing the focus to less than 4% of patients, the model enables:
- More proactive symptom management
- Better prioritization of follow-up visits
- Improved allocation of clinician time and resources
- Stronger alignment with HOPE compliance requirements
“The real value isn’t just in predicting events,” Gigliotti said. “It’s in enabling earlier, more targeted intervention that improves both patient experience and operational efficiency.”
A Smarter Approach to Symptom Management
As hospice providers face increasing demands, the ability to anticipate patient needs, rather than react to them, is essential.
Predicting symptom exacerbation helps hospice teams move beyond reactive care, prioritizing high-risk patients, staying ahead of HOPE requirements, and intervening before crises occur.
With Axxess intelligence, hospice providers can shift from broad, reactive workflows to precise, data-driven strategies, improving outcomes while optimizing operations.
