Every year, 1.6 million Americans are diagnosed with cancer. Despite tremendous advances in treatment, the sad truth is that for many patients, cancer is still a terminal disease. Sadder still is that many patients have little to no control over their final days, dying unexpectedly without a chance to fully come to grips with their mortality. 

Palliative care services are designed to give patients support and agency over the end of their life. They may involve mental health counseling or pain management, but they always focus on aligning treatment with the goals and desires of the patient.

Unfortunately, relatively few patients have access to this resource that can dramatically improve their quality of life. One study found that 68% of cancer patients are never referred for palliative care before they die. These patients never get the opportunity to dictate the terms of their death or take control of the process.

Death is always going to be a delicate subject for care teams to discuss with their patients. The last thing most oncologists want to do is to prematurely tell a patient that death is imminent. But as difficult and emotional as these end-of-life conversations can be, they are necessary for patients to live out their final days in comfort and in control. 

A new study gives reason for hope, demonstrating that augmented intelligence can help oncologists know when it’s the right time to initiate these difficult end-of-life conversations. 

Augmented Intelligence Predicts 30-day Mortality in Patients with Cancer

The study, published in Future Oncology, aimed to determine whether an augmented intelligence product, in this case the Jvion CORE™, could accurately predict short-term mortality risk among cancer patients. Better mortality predictions can help oncologists identify those who would benefit from palliative care services sooner, improving their quality of life. 

The results were striking: patients that the CORE predicted to be the highest risk for 30-day mortality were indeed 7.4x more likely to die within 30 days than the predicted low-risk patients.

To be clear, not all patients who were scored as the highest risk for 30-day mortality ultimately passed away in that time frame. In fact, only 4.9% of the highest risk patients did so, which also means that not every high-risk patient necessarily needed palliative care. This is where the distinction between artificial intelligence (AI) and augmented intelligence becomes important. 

Augmented intelligence products like the Jvion CORE apply AI to support, rather than replace, human judgement. They nudge clinicians to look at patients more closely, and to consider possible risk factors that they may not have considered before. The care team is ultimately still responsible for referring the patient for palliative care or taking any other intervention. 

However, the key difference between the CORE and other augmented intelligence tools, such as predictive analytics, is that the CORE provides insight into the modifiable factors driving the high-risk trajectory, and connects those drivers to clinical guidance to change the outcome. Critically, the CORE views patients outside of a narrow clinical lens; surfacing non-clinical risk factors that would otherwise be invisible to the care team: social determinants of health (SDOH). 

Seeing the Unseen: Social Determinants of Health

SDOH factors can be powerful predictors of health outcomes, including mortality. In fact, it’s estimated that up to 80% of health outcomes are tied to SDOH. When it comes to mortality risk, some of the most important, hidden contributors are a patient’s access to transportation, both to their appointments and to pick up medication, and whether a patient has social support at home. 

For example, socially isolated patients are sadly more prone to give up their will to live and stop taking their medication. However, like poor transportation access, this is an addressable risk factor. Once clinicians are empowered with the knowledge of why their patient is vulnerable, they can intervene to reduce their risk — whether that means enrolling the patient in a transportation assistance program or assigning them a social worker for emotional support. 

The CORE’s ability to surface clinical and non-clinical risk factors for mortality, as well as the recommended interventions that can address these risk factors, can be a game-changer for oncology. In some cases, the patients that are predicted to be at high risk are already known and accounted for by the care team. In other cases, it may come as a surprise, simply because providers don’t have the full picture of the patient’s risk. 

“No matter how great you are, the best doctor in the world, you don’t know every second of what’s going on with the patient,” said Dr. Sibel Blau, MD, in a recent interview. Dr. Blau has been using the CORE for several years now at Northwest Medical Specialties (NWMS), the oncology practice she founded in the Tacoma, WA area. She was also a co-author of the Future Oncology study. “This tool allows us to get those sick people into the clinic earlier,” she added.

Applying Augmented Intelligence for Oncology in Practice

Beyond the study’s findings that the CORE can accurately predict mortality risk, real-world experience has shown that the CORE can increase the timely integration of palliative and hospice care. Within two years of implementing the CORE, NWMS nearly doubled their palliative care consults, and increased their hospice referrals twelvefold. 

As a result of the CORE’s AI-driven interventions, patients were better supported at the end of their life. NWMS increased depression screenings by 171%, depression diagnoses by 22%, and case management evaluations by 84%. Pain management agents were better targeted as well, resulting in a 33% reduction in moderate or severe pain reports from patients. Patient deterioration was reduced as well, as indicated by a 30% drop in loss of function. 

But perhaps the most important outcome is that terminal patients gained more agency over the end of their lives. So often, a patient’s risk evolves faster than the oncologist’s understanding of their risk. By nudging clinicians that it may be time to initiate difficult end-of-life conversations, the CORE helped give patients a level of control over their death that they rarely have.