A new national study examined US Medicare fee-for-service beneficiaries over age 65 who were admitted to an ICU between 2011 and 2023. The authors found that discharges to hospice after ICU admission increased substantially over time, including an increase in hospice discharges per 100,000 beneficiaries and a rise in the proportion of ICU patients discharged to hospice. Notably, these changes occurred while in-hospital mortality and 30-day mortality remained stable, suggesting the trend is not simply explained by sicker hospitalized patients.
What else changed over time
The study also adds important context for hospice and palliative care teams working with hospitals. Over the same period, documented DNR orders and claims for palliative care delivery increased, while median hospital length of stay remained relatively stable. The authors note that the COVID-19 years were associated with temporary increases in mortality and other measures, but without a parallel rise in discharge to hospice, followed by a return toward pre-COVID trend trajectories. At the hospital level, the analysis suggests a growing share of short-term decedents were transitioned to hospice prior to death but variation across hospitals persists.
Questions hospice clinicians and leaders can ask next
For hospice professionals, the most useful question may be what we should measure next. If hospice discharge is increasing, how do we evaluate whether the transition was timely, informed, and truly goal-concordant? What does a high-quality ICU-to-hospice transition look like in your community? How should we track outcomes after enrollment such as symptom burden, caregiver experience, and avoided unwanted interventions? Where might variation reflect thoughtful local practice, and where might it reflect inequities in access, communication, or referral pathways that we can actively improve?
A recent narrative review published in Cureus examined trends in U.S. academic publications on Medical Aid in Dying (MAiD). Rather than arguing for or against MAiD, the authors analyzed the scholarly literature itself. They reviewed recent peer-reviewed articles and categorized them by position (supportive, critical, or neutral), author affiliation, specialty, and publication trends over time. One notable finding was the distribution between supportive and critical perspectives, as well as an observed increase in publications expressing concerns about MAiD in more recent years.
The paper provides a snapshot of who is contributing to the academic debate and how that debate is evolving within medical and ethical literature.
For hospice clinical staff, directors, and operators, this raises important questions. How does the academic conversation influence bedside care, team discussions, and policy development? Are hospice teams prepared to navigate questions about MAiD from patients and families, regardless of personal or organizational stance? How should leaders ensure that staff education reflects a balanced understanding of the broader ethical and clinical discourse? And as scholarly perspectives shift over time, how might that shape public perception, referral patterns, or regulatory scrutiny?
Whether one agrees or disagrees with MAiD, understanding how it is being discussed in academic medicine is valuable. The literature does not just reflect the debate – it often shapes it.
Recent reporting has highlighted how major health systems such as Mass General Brigham and Mayo Clinic are using Generative AI (GenAI) for clinical reasoning. These applications extend beyond administrative efficiencies and into the generation of preliminary diagnoses and treatment plans.
For hospice and palliative care, this development is not simply a technological milestone. It raises a more practical question: How might AI reduce the clinical noise that overwhelms families and clinicians in home-based care?
The Limits of “Dr. Google”
Anyone who has served as a caregiver (or even grappled with their own health condition) recognizes the late-night search for answers. In the era of “Dr. Google,” symptom searches often increase anxiety rather than relieve it. Generic results cannot account for a patient’s diagnosis, medication profile, disease trajectory, or individualized care plan.
Information without context rarely provides reassurance.
As personal health records become increasingly integrated with AI systems, we may see the emergence of a clinical “co-pilot.” Rather than offering generalized internet results, these systems could generate responses grounded in a specific patient’s documented history. For example, responses may consider factors such as:
The patient’s diagnosis
Current medications
Recent symptom trends
Established care plan
That is a fundamentally different experience. This contextualization could significantly reduce uncertainty.
The 3:00 AM Problem in Home Hospice
One of the most stressful aspects of home hospice care is the appearance of new symptoms outside of normal office hours. A family observes increased agitation, breathing changes, or decreased intake. They call. They wait. Anxiety escalates.
A thoughtfully designed AI triage layer, tightly scoped and carefully governed, could reinforce existing plan-of-care instructions, clarify when symptoms fall within expected patterns, and guide families through non-emergent comfort measures already approved by the clinical team. The key is to properly scope and govern the AI triage layer. One could envision that some of its potential services could include:
Reinforce existing care plan instructions
Clarify when a symptom is expected vs. urgent
Provide step-by-step comfort measures already approved in the plan of care
Escalate appropriately when thresholds are met
This would not replace clinical judgment. It would provide structured support while human clinicians remain central to decision-making.
Consistency in a High-Burnout Environment
A study published in Annals of Internal Medicine found that AI-generated recommendations were rated “optimal” in 77% of cases compared to 67% for physicians. One explanation is that AI does not experience fatigue or cognitive overload.
In hospice, symptom management frequently follows defined titration schedules and established care pathways. However, staffing strain and emotional intensity can contribute to variability in protocol adherence.
When used appropriately, GenAI could monitor key data points, reinforce established pathways, and flag inconsistencies. In doing so, it may enhance consistency without diminishing professional autonomy.
Guardrails are Essential
However, generative AI is not without risk. Generative AI systems can produce inaccurate or overconfident outputs and their reliability depends entirely on the quality of the underlying data. Embedded guardrails are required to ensure safe use of these systems.
Any hospice application would require:
Clearly defined scope limitations
Strong privacy and data protections
Transparent documentation practices
Defined escalation protocols
Unambiguous human oversight
The goal of introducing generative AI must be augmentation rather than automation.
Protecting What is Distinctly Human
Hospice is not algorithmic at its core. It is relational, emotional, and spiritual.
If technology can responsibly assume more of the metric-heavy and protocol-driven aspects of care, clinicians may gain cognitive space to focus on what only humans can provide: presence, listening, mediation, and meaning-making at the end of life.
In that sense, the true opportunity of GenAI in hospice may not be efficiency alone but the preservation of human attentiveness.
The Strategic Question for Hospice Leaders
The role of Generative AI in hospice care remains uncertain. What is clear, however, is that healthcare innovation is accelerating, and hospice leaders cannot afford to ignore emerging developments.
Whether GenAI ultimately finds a meaningful place in hospice workflows – or proves ill-suited to its relational model – will depend on thoughtful evaluation, ethical guardrails, and mission-driven leadership.
For now, the more important task may be asking better questions rather than rushing toward definitive answers.
Many hospice administrators are familiar with the hospice PEPPER report. A smaller number of hospice leaders, however, are familiar with its counterpart: the Comparative Billing Report (CBR) and eCBR, the electronic version of the Comparative Billing Report. While the PEPPER provides a general overview of a hospice agency’s billing data, the CBR is a specific tool used by Medicare to identify individual agencies whose billing patterns differ significantly from their peers. To manage hospice compliance effectively, it is important to understand the purpose and the mechanics of the CBR.
What is a CBR and Who Generates It?
A CBR is a formal educational resource produced by CMS (the Centers for Medicare & Medicaid Services). It provides data-driven insights into a hospice agency’s billing patterns compared to state and national averages.
The report is a collaborative effort:
CMS: defines the metrics (like length of stay) used to monitor for billing errors.
National Contractors: CMS hires private companies to do the data analysis. CMS may hire different companies to handle the data work versus managing distribution of the reports to providers.
MACs (Medicare Administrative Contractors): An agency’s local contractor (like Palmetto GBA, CGS, or NGS) may also provide their own version, called an eCBR, through their provider portal.
How the CBR Highlights Outliers
The CBR is a proactive tool designed to encourage providers to review their own data before a formal audit occurs. It identifies “outliers” – agencies that fall outside of normal billing ranges – by following a simple comparison process:
“Finding your Neighbors”: Instead of comparing a small local hospice to a massive national chain, the CBR groups each hospice agency with similar agencies. An agency’s data is compared against other hospices in their specific state. This ensures the comparison is fair and based on the local market.
The 1-to-100 Ranking: For each metric that is measured, the report provides the agency’s “percentile” score, ranking the agency on a scale from 1 to 100.
Imagine 100 hospices are standing in a line, ordered from the lowest billing to the highest.
If an agency’s CBR report indicates that the agency is in the 90th percentile, it means that the agency is billing more than 90 of the other hospices.
This is why the 90th percentile is the “red flag” area; it tells Medicare that the agency is at the very edge of the pack.
Clear Results There is no need to be a trained statistician to read or interpret the CBR report. The report uses clear labels:
Significantly Higher: This is the primary “outlier” signal. It indicates that the agency’s billing for a specific metric is in the top 10% (90%th percentile) of all providers. While not proof of wrongdoing, it is a high-priority “red flag” that warrants an immediate internal review of patient charts.
Higher: The agency is billing more than the average hospice. This is a signal to monitor the trend to ensure it does not move toward the 90th percentile.
Does Not Exceed: The agency’s billing is aligned with or lower than the average of its peers.
Not Applicable (N/A): The agency did not have enough claims in that category during the reporting period to create a statistically valid comparison.
Real-World Metric Examples
Each CBR focuses on a narrow “Target Area” of vulnerability. Two common examples for hospices include:
Non-Cancer Length of Stay (NCLOS) > 210 Days: Comparing the agency’s percentage of long-stay non-cancer patients against jurisdiction benchmarks.
GIP Average Length of Stay (ALOS): Benchmarking the agency’s inpatient stays (Q5004–Q5009) against state averages to identify potential overutilization.
How CBR Differs from PEPPER
It is helpful to view these two reports as different tools for the same goal:
Feature
PEPPER
CBR / eCBR
Primary Intent
Provides a broad risk profile across many areas at once.
Focuses on a specific billing trend with educational detail.
Result
A list of scores for 12+ different categories.
A specific label like “Significantly Higher” for one area.
Comparison
National, MAC, and State averages.
Usually State, Region, and Jurisdiction National averages.
Converting Comparative Data into Operational Action
The value of these reports lies in the establishment of a functional loop that transforms comparative data into focused action. Hospice leadership should treat PEPPER and CBR/eCBR results as signals tied to specific topics and timeframes. By comparing results against national or state benchmarks, leadership can determine whether a specific billing pattern requires continued monitoring or a formal investigation.
Determining the Response: Monitor vs. Investigate
A hospice agency that appears as a statistical outlier should view the data as a prompt for a focused internal review rather than evidence of an error. The Hospice PEPPER User’s Guide clarifies that these reports do not identify improper payments directly; instead, they serve as guides for auditing and monitoring billing changes over time. When a hospice agency looks materially different from its peers – particularly when that difference persists across multiple quarters – leadership should initiate a targeted investigation.
Maintaining a Narrow Scope
If an investigation is necessary, the review should remain strictly tied to the specific topic that triggered the signal. Effective hospice leadership avoids broad “audit everything” exercises in favor of targeted analysis:
Length of Stay (LOS) > 210 Days: The agency should review the long-stay patient population and the documentation patterns driving extended care.
GIP Average Length of Stay (ALOS): Leadership should focus on General Inpatient stays, discharge transitions, and the specific documentation supporting those inpatient days.
Closing the Loop
To finalize the process, the hospice agency must document the scope of the review, the specific findings, and any subsequent process changes. Re-checking the trend in the following cycle ensures these reports become a permanent part of the agency’s operating rhythm. CMS and MACs intend for these tools to function as educational resources that support a self-audit culture within the hospice organization.
Summary: Key Takeaways for Hospice Leadership
Effective compliance management requires utilizing both the PEPPER and the CBR as complementary tools for agency oversight. While the PEPPER serves as a broad indicator across many categories, the CBR functions as a targeted tool for identifying specific billing trends that may require immediate attention.
Hospice leadership should prioritize internal reviews for any metric labeled “Significantly Higher” or landing in the 90th percentile, as these results indicate the agency is a statistical outlier compared to its peers. By conducting narrow, focused chart audits based on these specific signals and documenting the findings within the QAPI process, a hospice agency can demonstrate a proactive approach to billing accuracy. Ultimately, treating the CBR as an educational resource rather than a threat allows leadership to refine workflows and improve documentation standards before external audits are initiated.
The hospice team often meets families at one of the most vulnerable moments of their lives. A loved one is nearing the end, emotions are raw, and time feels both urgent and suspended. What many families don’t realize – until they are in it – is how much harder this moment becomes when conversations about death were never had in advance.
Avoiding discussions about death is deeply human. It can feel frightening, pessimistic, or even disrespectful to talk about dying while someone is still living. Yet, the members of the hospice team witness daily the cost of that silence.
What Hospice Reveals About Unspoken Wishes
By the time hospice is involved, families are often facing rapid decisions about comfort, care, and meaning. When wishes haven’t been discussed, loved ones are left asking painful questions in real time:
“What would they want?” “Are we doing the right thing?” “Did we miss something important?”
Without guidance, families are forced to guess – while grieving. This can lead to tension, doubt, and lingering regret that lasts far beyond the loss itself. Hospice teams work tirelessly to support families through these moments, but even the best care cannot replace clarity that could have come from earlier conversations.
Grief is inevitable. Chaos does not have to be.
The Chaos Left Behind When Death Isn’t Discussed
When someone dies without having shared their wishes, the aftermath often includes more than sadness. Families may struggle with practical uncertainty and emotional strain at the same time. Important information may be scattered or missing. Loved ones may disagree about care decisions or arrangements. Meaningful stories, values, and memories may never be voiced or preserved.
The hospice team sees how this uncertainty compounds pain. Families are not only saying goodbye. They are also navigating confusion, paperwork, and decisions they never felt prepared to make. Many later say the same thing: “I wish we had talked about this sooner.”
Why These Conversations Matter in Hospice Care
Hospice is not just about managing symptoms at the end of life. It is about honoring a person’s values, comfort, dignity, and legacy. When families arrive with clarity about wishes, hospice care can be more aligned, more peaceful, and more meaningful.
Talking about death earlier allows hospice to become a continuation of a thoughtful journey rather than a crisis response. It gives families permission to focus on presence, connection, and love rather than logistics and uncertainty.
Tools That Help People Start the Conversation
For many people, the hardest part is knowing how to begin. Conversations about death don’t need to be clinical or overwhelming. They can start with values, stories, and simple questions about what matters most.
There are tools designed specifically to make these conversations more approachable and human:
The Conversation Project offers gentle guides that help families talk about wishes in a non-medical, values-based way. https://theconversationproject.org
Death Over Dinner reframes the discussion by encouraging people to talk about death in familiar, communal settings using curated prompts. https://deathoverdinner.org
PREPARE for Your Care uses videos and step-by-step guidance to help people reflect on their values and clearly communicate healthcare wishes. https://prepareforyourcare.org
These tools don’t force decisions; they create space for understanding.
Tools That Help Bring Affairs in Order
Once conversations begin, organization becomes an act of compassion. When information is documented and accessible, families are spared unnecessary stress during already emotional times.
Several resources exist to help individuals gather and record important details:
Five Wishes blends medical, personal, emotional, and spiritual preferences into one guided document. https://fivewishes.org
CaringInfo, from the National Hospice and Palliative Care Organization, provides free advance directive forms and hospice education. https://www.caringinfo.org
Everplans helps people organize important documents, instructions, and information for loved ones in one place. https://www.everplans.com
These tools help transform good intentions into clarity families can rely on.
From Avoidance to Care
Choosing not to talk about death does not protect loved ones. Instead, it often leaves them unprepared. The hospice team typically sees how earlier conversations can ease fear, reduce conflict, and allow families to focus on what truly matters in the final chapter of life.
Talking about death is not about giving up hope. It is about giving a gift: guidance, reassurance, and peace of mind for those we love most. When death is acknowledged with honesty and compassion, the end of life can be met with greater calm, dignity, and connection.
Used to monitor number of Google Analytics server requests when using Google Tag Manager
1 minute
__utmv
Contains custom information set by the web developer via the _setCustomVar method in Google Analytics. This cookie is updated every time new data is sent to the Google Analytics server.
2 years after last activity
__utmx
Used to determine whether a user is included in an A / B or Multivariate test.
18 months
_ga
ID used to identify users
2 years
_gali
Used by Google Analytics to determine which links on a page are being clicked
30 seconds
_ga_
ID used to identify users
2 years
_gid
ID used to identify users for 24 hours after last activity
24 hours
__utma
ID used to identify users and sessions
2 years after last activity
__utmt
Used to monitor number of Google Analytics server requests
10 minutes
__utmb
Used to distinguish new sessions and visits. This cookie is set when the GA.js javascript library is loaded and there is no existing __utmb cookie. The cookie is updated every time data is sent to the Google Analytics server.
30 minutes after last activity
__utmc
Used only with old Urchin versions of Google Analytics and not with GA.js. Was used to distinguish between new sessions and visits at the end of a session.
End of session (browser)
__utmz
Contains information about the traffic source or campaign that directed user to the website. The cookie is set when the GA.js javascript is loaded and updated when data is sent to the Google Anaytics server
6 months after last activity
_gac_
Contains information related to marketing campaigns of the user. These are shared with Google AdWords / Google Ads when the Google Ads and Google Analytics accounts are linked together.
