Value-based care in oncology continues to gain traction as a model that rewards patient outcomes and cost efficiency over volume. One of the toughest challenges within this framework is managing inpatient admissions, since oncology patients often need extended hospital stays. New tools, from machine learning to CMS payment models, are helping practices meet that challenge.
Value-based care is a core quality strategy in healthcare reform that prioritizes better care over more care. The model rewards providers for improving patient health, preventing complications, and promoting healthier living. Telehealth and remote patient monitoring support this approach by extending care beyond the clinic.
This article reviews a machine learning study on reducing hospital length of stay and connects it to the current state of value-based care in oncology under CMS.
The State of Value-Based Care in Oncology
Value-based care in oncology has a dedicated federal test model. The CMS Enhancing Oncology Model (EOM) began July 1, 2023, and holds participating oncology practices accountable for quality and total cost of care during six-month episodes tied to chemotherapy.
CMS has since expanded the program. The model was extended through June 30, 2030, and a second cohort of participants joined on July 1, 2025. Monthly Enhanced Oncology Services payments rose from $70 to $110 per patient per month to help practices fund care coordination, patient navigation, and around-the-clock clinician access. In August 2025, CMS released first-reconciliation results showing cost savings for more than three quarters of participants. These moves signal that value-based care in oncology is maturing from pilot to practice.
Research in Value-Based Care in Oncology
Against that backdrop, predictive tools that lower cost without harming quality carry real weight. A study published in the Journal of Clinical Oncology explored how combination machine learning (ML) models can predict and potentially reduce hospital length of stay (LOS) for oncology patients.
Srisairam Achuthan and colleagues focused on five major cancer types: lung cancer, multiple myeloma, lymphoma, small intestinal and colorectal cancer, and high-risk breast cancer. These types account for roughly 50% of inpatient admissions and hospital days. The goal was to classify and predict LOS in order to reduce it, which lowers hospital-acquired conditions and associated costs and supports value-based care in oncology.
Methodology
The researchers used two machine learning models:
- Model 1: A multi-classification model (eXtreme Gradient Boosted Trees Classifier) sorting LOS into four categories:
- Class 1: 1 to 3 days
- Class 2: 4 to 8 days
- Class 3: 9 to 15 days
- Class 4: More than 15 days
- Model 2: A regression model (eXtreme Gradient Boosted Trees Regressor with Early Stopping) predicting LOS for Classes 1 to 3. Class 4 was excluded because of its wide range (16 to 90+ days).
The training data included 4,280 randomly selected inpatient admissions for Model 1 and 19,636 for Model 2, both using 102 features.
Results
Tested on new claims data, the models produced findings that support value-based care in oncology:
- Model 1: Predicted 296 of 3,945 inpatient admissions as Class 4, with 88 true positives. A 10% reduction in LOS could lower overall LOS by 1% to 3% for most practices.
- Model 2: Predicted LOS for 3,649 admissions in Classes 1 to 3. The gap between actual and predicted LOS ran 18% to 24%. Setting the target to the predicted LOS plus an upper bound of at least 10% achieved LOS reductions of 6% to 16% across practices.
Combined, the two models suggested a potential length-of-stay reduction of roughly 6% to 19%, a meaningful result for value-based care in oncology.
Why Length of Stay Matters for Value-Based Care in Oncology
Reducing LOS lowers hospital costs and cuts the risk of complications such as hospital-acquired infections. That aligns with the core principles of value-based oncology care: high-quality care and efficient use of resources. Shorter stays also tend to improve patient satisfaction and recovery, which supports the outcomes that value-based models reward.
Cost Reduction and Quality Improvement
Predicting and reducing LOS lets providers improve outcomes while lowering cost, the central aim of every value-based arrangement. In a model like EOM, where practices carry financial accountability for total episode spending, tools that trim avoidable inpatient days directly affect performance against the benchmark.
Where Remote Monitoring Fits
Remote patient monitoring extends value-based care in oncology beyond the hospital. By tracking symptoms and vital signs at home, care teams can catch problems earlier and prevent avoidable admissions and re-hospitalizations. Combined with predictive models that manage length of stay, remote monitoring helps practices deliver the coordinated, patient-centered care these models require.
Frequently Asked Questions
1) What is value-based care in oncology?
Value-based care in oncology is a payment and delivery approach that rewards oncology practices for improving patient outcomes and controlling total cost of care, rather than for the volume of services delivered. The CMS Enhancing Oncology Model is the leading federal example.
2) How does machine learning support value-based oncology care?
Machine learning can predict hospital length of stay and flag high-risk patients, helping practices reduce avoidable inpatient days and complications. One study estimated potential length-of-stay reductions of 6% to 19% using combined ML models.
3) Is the Enhancing Oncology Model still active?
Yes. CMS extended the model through June 30, 2030, added a second cohort in July 2025, and raised monthly enhanced-services payments to $110 per patient per month.
Understanding Value-Based Care in Oncology
Machine learning models that predict and reduce hospital length of stay advance value-based care in oncology, especially where patient management is complex and costly. Paired with CMS models like EOM and with remote patient monitoring, these tools help practices improve outcomes while keeping care affordable. Embracing them is central to the continued evolution of value-based care in oncology and to more sustainable cancer care.
Tenovi works exclusively with companies that offer remote patient monitoring to healthcare providers, including RPM software companies, chronic care management companies, and telehealth platforms. Our cellular-connected, no sync, no app devices help partners extend value-based cancer care into the home. Book a free demo and consultation with Tenovi today.
