Skip to main content
Doctors and nurse predictive analysis

How Providers Can Use Predictive Tools to Inform End-of-Life Referrals

Thought Leadership

By Robin Stawasz

One of the major challenges for hospice organizations is to receive referrals from provider partners when patients are ready for end-of-life care. Predicting longevity runs counter to a doctor’s primary focus: to do everything they can to delay a patient’s death. Few have been clinically trained to estimate prognosis, and when they try, too often they vastly overestimate their patient’s life expectancy and miss the appropriate timing for a palliative care intervention.

Hospice and other palliative care providers are rightfully frustrated when providers refer a patient so late in the patient’s disease process that they are left with days, if not hours, to provide the medical, emotional, and social support that patients and their families so desperately need at the end of life.

Fortunately, there is a solution. Today hospice and palliative care organizations can work with providers using clinical algorithms to determine when a palliative treatment approach will be more effective than a curative treatment approach.

Technology Can Prognosticate Hospice Eligibility

For years, physicians have relied on clinical algorithms to indicate best practice for each disease process — algorithms that use specific data points as validated, predictive indicators of how the disease process will develop. For example, if a person with diabetes has a certain A1C level, then a specific treatment paradigm is called for. Cardiac intervention is determined by the classification of the patient’s level of heart disease, again determined by algorithm.

Dying is a disease process too, and specific data points can indicate a life expectancy prognosis for seriously ill patients. These mortality risk models have been validated and used within healthcare systems throughout the world and are, by far, the most accurate predictors of a patient’s prognosis. While physician estimates of prognosis are around 10% to 20% accurate, these risk scores are up to 90% accurate. 

Besides the algorithms that predict prognosis, there are other validated indicators of the many events that impact symptom burden and lead to coordination issues, poor clinical outcomes, and increased cost. These include risk scores for initial hospitalization, rehospitalization, emergency department utilization, caregiver breakdown, medication noncompliance, frailty, and psychosocial needs. With insight across a patient population, providers can proactively identify individuals within that population who would benefit from a palliative care intervention in a time frame that allows such intervention to have maximal benefits. 

A New Paradigm in Serious Illness Management

Eventually all providers will have access to a range of tools that provide high level insights to help them recognize the need for a palliative intervention earlier in the disease process. 

Currently, the Acclivity Connected Care Platform stands alone in the integration of these predictive algorithms and the application of proactive identification within the seriously ill population. Many providers may hesitate to trust such data analytics as a stand-in for their own diagnostic capabilities.  Acclivity addresses that concern with a Consultation Report that takes a provider’s patient and compares them across the national, multi-million patient population represented by the Acclivity Platform. The Platform finds clinical “twins” of the individual patient — patients with the same age, gender, and specific diagnostic conditions — and evaluates the outcomes for all those similar cases. This allows for an accurate prediction of clinical outcomes based on specific interventions and medications, as well as care costs and mortality. The report can further strengthen a provider’s confidence in the predicted clinical picture and prognosis for a patient.

It is time to move away from the assumption that a provider can and should accurately predict mortality on their own. The clinical tools needed to recognize the best time for a palliative care referral exist and they offer significant benefits to primary care providers, their practice, the palliative care provider, and, most importantly, to the patient and their family. 

For more information about Acclivity’s predictive analytics capabilities to inform earlier referrals, please contact us at

Jim Boborci