Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-2082
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-10-21T06:55:12.257764Z
status: published
lens: g_space
sector: health
headline: Psychosocial Distress and Radiation Treatment Interruptions in Patients With Cancer Using Machine Learning and Regression: Retrospective Cohort Study
dek: Background Uninterrupted delivery of radiation therapy is critical to cancer treatment quality. Unplanned radiation treatment interruptions compromise tumor control and survival. Although previous studies have explored clinical and sociodemographic factors influencing treatment adherence, the role of psychosocial distress and related symptoms remains unclear. Objective This study aims to identify patient-reported psychosocial predictors of radiation treatment interruptions from patient-reported distress screenin…
gain_title: An XGBoost model and a negative binomial regression model gave closely aligned predictions of missed radiation days and identified the same key predictors, supporting use of distress screening to target interventions.
problem_title: (none)
trace_subject: (none)
gain_reading: An XGBoost model and a negative binomial regression model gave closely aligned predictions of missed radiation days and identified the same key predictors, supporting use of distress screening to target interventions.
gain_evidence: Both parametric and machine learning models provided stable and closely aligned predictions of missed radiation treatment days.
problem_reading: (none)
problem_evidence: (none)
quick_read: Researchers at a single institution analyzed 1953 patients who received external beam radiation therapy between August 2020 and May 2024. They linked electronic health record data, Distress Thermometer scores, and census tract socioeconomic indicators, then modeled missed treatment days with regression and a machine learning model.

Higher distress, learning or communication barriers, and genitourinary cancer were tied to more missed days, and the machine learning model ranked the same factors as top contributors. The associations were described as modest, and the machine learning model improved prediction only slightly, so its practical advantage remains unproven.
limitation: The associations between distress, education barriers and missed treatment days were only modest, and the machine learning model improved error by just 0.02 over regression, so the predictive gain is incremental.
tag: Evidence-backed gain
key_points: The cohort missed a mean of 1.19 treatment days (SD 2.18), with a range of 0 to 15. | Higher distress scores, learning or communication barriers, and genitourinary malignancy were significantly associated with more interruptions in the regression model. | XGBoost with a Poisson objective had a root mean squared error of 2.19 versus 2.21 for the regression model. | The authors conclude that psychosocial assessment should be built into radiation oncology workflows.
rundown: The study used a hospital-level electronic health record repository for demographics, clinical characteristics, and treatment details. Baseline distress was measured with the National Comprehensive Cancer Network Distress Thermometer and combined with census tract-level socioeconomic indicators.

Treatment interruption was measured as missed treatment days and modeled with negative binomial regression and an extreme gradient boosting model. Both approaches agreed on the leading predictors, which the authors say reinforces the robustness of the findings.

The authors present the work as support for future intervention strategies, motivated by earlier evidence that interruptions compromise tumor control and survival.
sources:
- peer_reviewed | JMIR Cancer | https://doi.org/10.2196/92899 | 2026-10-19
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

f3d3a92d7fc07a7ce11813eaef7428ee345190a18f1f1c66c636caa47795551e

previous

0000000000000000000000000000000000000000000000000000000000000000