Percentage of Patients who Died with Cancer Receiving Systemic Cancer-Directed Therapy in the Last 14 / 30 Days of Life (lower score – better) (Registry version)

Importance

Strengths:

A clear logic model is provided, depicting the relationships between inputs (e.g., specialized staff, clinician training, and guidelines), activities (e.g., prognostic disclosure, early palliative integration, shared decision-making), and desired outcomes (e.g., fewer emergency department [ED] visits, hospitalizations, and intensive care unit [ICU] visits). This model demonstrates how the measure’s implementation will lead to the anticipated outcomes.
The problem this measure addresses presents a significant burden for patients, as cancer is the second leading cause of death overall and the leading cause of death among people younger than 85 years old in the United States. There are projected to be approximately 2.1 million new cancer diagnoses and over half a million cancer deaths in 2026.
The measure is supported by a comprehensive literature review, including systematic reviews with high evidence quality and clinical practice guidelines with evidence grading of strong/high and high quality empirical studies, including National Comprehensive Cancer Network (NCCN) guidelines and a retrospective cohort study, demonstrating a clear net benefit in terms of improved outcomes and reduced cost/resource use among people dying with cancer.
At the individual clinician performance level (158 entities), data from the US Oncology Network (USON)/McKesson databases from January 2023-December 2024 show a performance gap for the Percentage of Patients who Died with Cancer Receiving Systematic Cancer Directed Therapy in the Last 14 Days of Life and the Last 30 days of Life measure, with decile ranges from 0% to 40% with a median of 7.5% and 0% to 60% with a median of 24.4%, respectively, indicating variation in measure performance.
Description of patient input supports the conclusion that the measured intermediate outcome is meaningful with at least moderate certainty. The American Society of Clinical Oncology's (ASCO) end-of-life measures were originally developed using patient-centered methodology (e.g., focus groups consisting of patients with incurable cancer and family members of deceased patients) to capture outcomes meaningful to those with advanced illness. Patient and caregiver input has continued to be obtained through expert panel participation and public comment.

Limitations:

At the clinician practice level (10 entities), data from the USON/McKesson databases from January 2023-December 2024 show limited variation in performance for the Percentage of Patients who Died with Cancer Receiving Systematic Cancer Directed Therapy in the Last 14 Days of Life and the Last 30 days of Life measure, with decile ranges from 6% to 12% and 20% to 32%, respectively, suggesting a small performance gap. The measure submission could be strengthened by expanding performance gap testing to a larger population or exploring additional literature sources that demonstrate a gap in care at the clinician: practice level.
The use of 10 group practices only does raise a question about how widely this measure is used in the MIPS program.
Although the developer indicated they had gathered patient input through a family caregiver and a patient representative across two engagements (technical expert panel and public comment), direct engagement with and feedback from patients themselves is limited.

Rationale:

This maintenance measure meets all criteria for ‘Met’ for importance due to the significance of the problem it addresses, its robust evidence base, a documented performance gap at the clinician: individual level, and well-articulated logic model, making it essential for encouraging timely enrollment in palliative care among people dying from cancer.

Closing Care Gaps

Closing Care Gap Rating

Closing Care Gaps

Strengths:

The developer provided evidence of gaps in care related to the measure focus for subgroups, including systematic study, meta-analysis, retrospective cohort study, and literature review, but they did not clearly articulate how the measure will close gaps in care

Limitations:

Although the developer refers and reports results from studies that have analyzed gaps in care, they did not provide any description of their own measure testing for evaluating gaps in care for subgroups.
The developer did not provide recommended actions entities can take to close care gaps.

Rationale:

While the developer attempted to assess gaps in care across subgroups, statistical methods and results for their analyses were not reported.

Feasibility Assessment

Feasibility Assessment Rating

Feasibility Assessment

Strengths:

The developer described how all required data elements can be collected without risk to patient confidentiality, including strict accordance with Health Insurance Portability and Accountability Act (HIPAA) Privacy and Security Rules, de-identified Electronic Health Record (EHR) data, and a minimum threshold of five patients for performance reporting.
Licensing, or other requirements to use any aspect of the measure (e.g., value/code set, risk model, programming code, algorithm) are clearly described and justified. Non-profit entities are not charged a fee.

Limitations:

The fee for entities that are not non-profits is not stated. There is an outstanding question if the fee structure places the measure out of reach for rural and safety net clinicians and group practices.
The developer states that, the measure relies entirely on clinical data extracted from EHRs with no direct patient interaction or survey-based data collection. It seems that this is the case for the measure implemented within the McKesson Oncology Network Registry. It is unclear if these findings are the same with entities reporting the measure that are not reporting in the McKesson Oncology Network Registry. The fact that the measure can be reported using EHR data raised the question why the measure is not specified as a digital measure (eCQM, Fast Health Interoperability Resources (FHIR) digital quality measure), which may reduce reporting burden.
The measure specification is embedded in the measure submission, which makes review challenging. In addition, there is not a corresponding file with codes used in the measure specification, for ease of review. The developer may consider including the complete specification as an attachment, to ease review by endorsement and maintenance committee members.

Rationale:

The maintenance measure is rated as ‘Not Met, but Addressable’ for feasibility due to lack of clarity in the following areas: 1) Cost for use by clinicians and group practices that are NOT non-profits and if this is manageable for rural and safety net systems; and 2) Question regarding the electronic implementation by entities not reporting to the McKesson registry.

Scientific Acceptability

Scientific Acceptability Reliability Rating

Scientific Acceptability Reliability

Strengths:

The developer performed the required reliability testing for this maintenance measure, namely, they conducted accountable entity-level (“measure score”) reliability testing at the level(s) for which the measure is specified. Data sources used for reliability analysis are adequately described and include EHR data sourced from the US Oncology Network (USON)/McKesson databases consisting of a national representation of 13,812 patients across 10 practices and 8,939 patients across 147 clinicians during the period of 2023 and 2024.
The developer conducted signal-to-noise reliability testing at the practice level, both for patients in the last 14 days of life and the last 30 days of life. For patients in both the last 14 days of life and the last 30 days of life, more than 70% of accountable entities meet the expected threshold of 0.6 and 100% of entities meet the expected threshold of 0.4 for signal-to-noise testing.

Limitations:

The developer conducted signal-to-noise reliability testing at the at the clinician level, both for patients in the last 14 days of life and the last 30 days of life. For patients in both the last 14 days of life and the last 30 days of life, fewer than 30% of accountable entities meet the expected threshold of 0.6 and at least 30% of entities are below the expected threshold of 0.4 for signal-to-noise testing. The developer mentions that this "warrants caution" at the clinician level but offers no other justification for the low reliability.
Data used for reliability testing were sourced from a two-year period of performance. If this measure is intended to be calculated over shorter periods of performance, the accountable-entity level reliability will be lower. For a one-year period of performance, at least 70% of the clinicians are still likely to have a reliability greater than 0.6 for patients in the last 30 days of life. However, for patients in the last 14 days of life, it is likely that fewer than 30% of the clinicians will have a reliability greater than 0.6.

Rationale:

This maintenance measure is rated as ‘Not Met’ for reliability because the reliability testing results significantly fall below the established thresholds at the clinician level indicating major issues with the consistency and accuracy of the results across different settings and populations. The developer may consider recommending against the use of this measure at the clinician level. By addressing this issue, there is potential to enhance the reliability, provided the period of performance is intended to be at least two years.

Scientific Acceptability Validity Rating

Scientific Acceptability Validity

Strengths:

The developer performed the required validity testing for this maintenance measure, namely, they conducted accountable entity-level (“measure score”) validity testing at the levels for which the measure is specified (physician and practice). Data sources used for validity analysis were adequately described and include electronic health record (EHR) data from January 2023 through December 2024 for 158 physicians and 10 practices affiliated with USON.
The developer conducted empirical validity testing at both accountable entity levels using Pearson correlation to compare the measure score with the percentage of patients who died with cancer enrolled in hospice for at least 3 days immediately before death (CBE ID 0216). The developer hypothesized negative relationships between the therapy measure and the hospice measure, reasoning that these measures reflect opposing clinical priorites near end of life. Pearson correlation coefficients confirmed the expected direction and significance of associations of both strata of the therapy measure (14 days and 30 days) and the hospice measure at both the physician level (therapy at 14 days: r = -0.2478, p = 0.0017; therapy at 30 days: r = -0.1950, p = 0.0141; n=158) and practice level (therapy at 14 days: r = -0.6446, p = 0.0442; therapy at 30 days: -0.8045, p = 0.0050; n=10). While the developer did not state this, it my be reasonable to expect the stronger negative relationship between the therapy in the last 14 days of life and the hospice measure vs. between the therapy in the last 30 days of life and the hospice measure, as was observed at the physician level, due to the tighter time frame within which to potentially make a shift in clinical direction from agressive treatment to hospice. The developer also reported the expected strongly positive correlation between 14-day and 30-day therapy at the physician and practice levels (respectively: r=0.6669, p<0.0001, n=158; r=0.7998, p=0.0055, n=10).
A well-grounded, thorough logic model support an inference of validity for the measure.

Limitations:

Given the small sample of facilities and providers, there may be gaps in the overall representativeness of accountable entities. As noted by the developer, the small sample size of the practice-level correlation analysis may result in less precise estimates than at the physician-level.
While the direction of the relationship between the hospice measure and both therapy measures aligns with their hypothesis at both levels of analysis, the developer could provide additional interpretation for the stronger relationships at the practice-level compared with the physician-level. Another limitation of the testing presented is that the developer did not report testing relationships of the therapy measure with one or more independent measures, which could help establish discriminant validity or provide more robust evidence of a shared quality mechanism.
The developer did not conduct risk adjustment or stratification, but provided the rationale that quality palliative care is a universal standard that should not vary based on patient complexity and that adjustment could normalize departures from palliative care best practices. The developer did not provide supporting literature, a conceptual model, or empirical analysis demonstrating that differences in patient characteristics do not affect measure results or inhibit fair comparisons.

Rationale:

This maintenance measure is rated as ‘Not Met But Addressable’ for validity; measure score testing results partially support an inference of validity, suggesting that the measure somewhat accurately reflects performance on quality and can distinguish good from poor performance to a limited extent.
The developer did not conduct risk adjustment or stratification and provided a rationale for the decision, but did not support the rationale with supporting literature or empirical analysis.

Use and Usability

Use and Usability Rating

Use and Usability

Strengths:

The version of the measure without recent material specifications is currently used in the MIPS reporting program, and developers are in consultation with CMS regarding inclusion in programs such as Prospective Payment System (PPS)-Exempt Cancer Hospital Quality Reporting (PCHQR) Program and Inpatient Quality Reporting (IQR) Program in addition to MIPS. This updated version of the measure will be used in MIPS beginning Payment Year (PY) 2027. Attributes of a suitable program for the measure are described, and these include adult patients (aged 18 and older) with a confirmed diagnosis of cancer, accountability at the group practice level.
The developer provided a summary of how accountable entities can use the measure results to improve performance. Specifically, entities can provide guidance on disease trajectory, reassessing prognostic awareness, and discuss the continuation of treatment; stop all treatments not contributing to comfort, initiating intensive symptom management, and placing hospice referrals; early referrals; and Advanced Care Plan (ACP) documentation. The developer include information as to why the mechanisms to improve performance can be difficult and offered strategies to overcome difficulties within this section of the submission.

Limitations:

In 6.2.2 Feedback on Measure Performance, the developer indicates that ASCO has not received any feedback on measure performance or implementation. It is unclear if a there is a systematic feedback approach that is used to better understand if challenges exist within entities implementing the measure or with their actionability toward improving performance on the measure. To strengthen the submission, the developer should clarify whether or not they have a systematic feedback approach.

Rationale:

This maintenance measure is rated ‘Met' for use and usabiltiy because it is actively used in one accountability application. The developer indicated they have not received any feedback on this measure. The measure demonstrates modest improvements in performance results and the developer noted performance trends are expected to continue to improve, affirming its ongoing usability.

Committee Independent Review

0210 Review

Importance

Importance Rating

Importance

I agree with the staff assessment

Closing Care Gaps

Closing Care Gaps Rating

Closing Care Gaps

I agree with the staff assessment that while the developer attempted to assess gaps in care across subgroups, statistical methods and results for their analyses were not reported.

Feasibility Assessment

Feasibility Assessment Rating

Feasibility Assessment

The staff assessment brought up some good questions: 1) Cost for use by clinicians and group practices that are NOT non-profits and if this is manageable for rural and safety net systems; and 2) Question regarding the electronic implementation by entities not reporting to the McKesson registry.

Scientific Acceptability

Scientific Acceptability Reliability Rating

Scientific Acceptability Reliability

As a Patient Partner, I am not a subject matter expert in this area and will rely on the staff assessment.

Scientific Acceptability Validity Rating

Scientific Acceptability Validity

As a Patient Partner, I am not a subject matter expert in this area and will rely on the staff assessment.

Use and Usability

Use and Usability Rating

Use and Usability

I agree with the staff assessment that this measure is actively used in one accountability application.

Summary

As a patient partner, I was hoping the developer provided additional data on why patients chose aggressive care at the end of their lives. Could it be the age of the patient (younger versus older), geographic location, beliefs, religion, hoping for a miracle? Should these decisions be excluded from the measure?

(No subject)

Importance

Importance Rating

Importance

This measure brings attention to systemic cancer treatments that may be non-beneficial for a patient so close to end of life. I rank it high in importance.

Closing Care Gaps

Closing Care Gaps Rating

Closing Care Gaps

There is a gap in quality of care when patients are receiving systemic therapy in their final days of life. This suggests a need for more discussions about the expected benefits (and risks, alternatives) or treatment in a dying patient.

Feasibility Assessment

Feasibility Assessment Rating

Scientific Acceptability

Scientific Acceptability Reliability Rating

Scientific Acceptability Validity Rating

Use and Usability

Use and Usability Rating