Truace
Subscribe
The Index recomputed live from the record

What the evidence says.What the public feels.

The record holds 935 sourced gains and 771 sourced problems, averaging 68 and 66 on the index score. Readers have logged 22 public signals on the Pulse, which is kept apart and never counted as evidence.

935 gains
771 problems
Every sourced claim in the Index, one square each, shaded by the strength of its evidence. High Moderate Emerging

Ranks distinct AI gain and problem claims from the published record. Scores reward impact, independent source strength, scale, confidence, and recency.

1,706 results
Show filters and sorting

Download every matching row, not just this page:Export CSVExport JSON

AI gains · 935

309
HealthNewly addedModerate evidence · 1 source

In simulated RCT time-to-event data, random survival forest retained predictive performance better when treatment-covariate interactions were present than when they were absent.

By 2026-10-01, authors reported a neutral comparison of Cox proportional hazards regression and random survival forest for predicting patient-specific survival probabilities, using varied simulation scenarios based on two RCT reference datasets and TRIPOD-aligned metrics. They observed that C-index-only conclusions did not generalize, that overall performance measures appeared more reasonable, and that splitting rule choice mattered for RSF.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
98

Updated Oct 1, 2026 · TRV-2026-1232

68Index score
310
HealthNewly addedModerate evidence · 1 source

Researchers used machine learning and LASSO regression on single-cell data to identify seven PANoptosis key genes and build a validated IDD risk diagnosis model, with experimental knockdown of ACSBG1 reducing degeneration in a rat model.

By October 2026, researchers reported using single-cell sequencing and machine learning to pinpoint seven PANoptosis-related genes associated with intervertebral disc degeneration and to build a risk diagnosis model. They validated expression patterns in patient nucleus pulposus samples and tested ACSBG1 knockdown in rats.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
98

Updated Oct 1, 2026 · TRV-2026-1230

68Index score
311
HealthNewly addedModerate evidence · 1 source

Researchers built and externally validated a machine learning model that identifies DCIS at high risk of progressing to invasive disease, enabling more precise risk stratification and clinical management.

Researchers analyzed coexisting ductal carcinoma in situ and invasive ductal carcinoma using spatial transcriptomics, single-cell RNA sequencing, and single-cell DNA sequencing to map how malignant cells and their microenvironment change at invasion.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
98

Updated Sep 30, 2026 · TRV-2026-1224

68Index score
312
HealthNewly addedModerate evidence · 1 source

A Random Forest model combining two tryptophan metabolites (kynurenine and neopterin) with five routine clinical tests discriminated MASLD from healthy controls with high diagnostic performance.

In a 179-person study of 80 healthy controls and 99 MASLD patients published September 2026, investigators used non-targeted and targeted metabolomics to identify dysregulated tryptophan metabolites and trained seven machine learning classifiers for MASLD discrimination, selecting a Random Forest model that combines kynurenine, neopterin and five routine labs.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
98

Updated Sep 30, 2026 · TRV-2026-1223

68Index score

AI problems · 771

309
HealthStableModerate evidence · 1 source

Multi-omics biomarker approaches face persistent challenges in data heterogeneity, reproducibility, and clinical validation across diverse patient populations.

By November 2025, a review in Molecular Biomedicine synthesized multi-omics strategies integrating genomics, transcriptomics, proteomics and metabolomics, with emphasis on machine learning and deep learning for horizontal and vertical integration, including single-cell and spatial technologies.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
89

Updated Aug 18, 2026 · TRV-2026-0831

67Index score
310
HealthStableModerate evidence · 1 source

Oncology AI systems can reproduce or amplify existing disparities across patient populations, and efforts to enforce fairness definitions often conflict with overall predictive performance.

As of the August 2026 commentary, AI was increasingly integrated into oncology for detection, risk stratification, treatment planning, and documentation. The authors reviewed evidence that these systems can reproduce or amplify disparities and examined technical sources of bias and competing statistical definitions of fairness.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
89

Updated Aug 18, 2026 · TRV-2026-0829

67Index score
311
HealthStableModerate evidence · 1 source

When used for prostate MRI reporting, LLMs can hallucinate measurements, flip negations, misstate laterality, and overstate cancer likelihood, creating patient-safety and accountability risks especially if reports are copied outside clinical governance.

Published August 17 2026 in Abdominal Radiology, this Perspective examines large language models applied to prostate MRI reporting, a task where laterality, sector, size, PI-RADS, and staging language directly affect biopsy and treatment decisions and where patients often see reports via portals before clinician discussion.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
89

Updated Aug 18, 2026 · TRV-2026-0828

67Index score
312
HealthStableModerate evidence · 1 source

Current AI systems for psychological assessment and treatment have not shown ability to produce meaningful and sustained clinical change, falling short due to memory limits, sycophancy, and focus on short-term helpfulness.

On 2026-08-17, a peer-reviewed framework paper argued that while large language models could augment psychological assessment and treatment, current technologies have not demonstrated sustained clinical benefit. The authors attribute this to poor integration of clinical science and to a duration mismatch between brief AI chats and months-long evidence-based treatments.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
89

Updated Aug 18, 2026 · TRV-2026-0821

67Index score

Recomputed live from the record · Oct 12, 2026, 3:43 PM