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

349
HealthStableModerate evidence · 1 source

A Random Forest trained on 22 routine clinical variables predicted unplanned 90-day psychiatric readmission in adults with MDD with moderate discrimination.

The DEEP READ prospective study across 13 Italian provinces followed 322 adults with major depressive disorder discharged from inpatient care and tested whether routinely collected clinical information could predict unplanned psychiatric readmission within 90 days. A Random Forest classifier trained on 22 predictors achieved a mean test AUC of 0.74 in internal cross-validation, with 50 patients (15.5%) readmitted.

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

Updated Sep 15, 2026 · TRV-2026-1096

68Index score
350
HealthStableModerate evidence · 1 source

In NSCLC immunotherapy selection, CB-only AI models outperformed PD-L1 and clinical scores in the independent test set, and both expert and nonexpert physicians improved their predictions when using the explainable AI decision support tool.

The I3LUNG study enrolled 2,396 patients with non-small cell lung cancer to develop AI models for immunotherapy selection, integrating clinical and blood data, CT, digital pathology and genomics into early and intermediate fusion models. CB-only models reached AUC up to 0.77 in the independent TEST set and outperformed PD-L1, ECOG PS, NLR, LDH and LIPI, and a usability study found physicians improved predictions with the explainable AI tool.

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

Updated Sep 15, 2026 · TRV-2026-1095

68Index score
351
HealthStableModerate evidence · 1 source

A 3D-ResNet18 multi-task model reduced error in estimating rib fracture age from chest CT to 7.94 days MAE, outperforming manual expert evaluation at 10.65 days, while also classifying healing stage and fracture type.

By September 2026, researchers reported a multi-task deep learning model based on 3D-ResNet18 designed to assist expert judgment of rib fracture age from chest CT. Trained on 1,848 fractures from multiple centers, the model was tested externally and reported to predict fracture age, healing stage, and fracture type concurrently.

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

Updated Sep 15, 2026 · TRV-2026-1094

68Index score
352
HealthStableModerate evidence · 1 source

AI systems including large language models can process unstructured clinical text and interpret complex laboratory findings to support clinical decision-making in laboratory medicine.

This narrative review from Chinese Medical Journal examines how AI in laboratory medicine has evolved from conventional machine learning on structured results to deep learning and large language models that handle unstructured clinical text. It compares four paradigms and reviews evidence across blood cell morphology, autoverification, infectious risk stratification, urinalysis interpretation, decision support, and report generation.

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

Updated Sep 15, 2026 · TRV-2026-1093

68Index score

AI problems · 771

349
HealthStableModerate evidence · 1 source

Machine learning's black-box nature has hindered clinical adoption of mortality prediction models for non-dialysis CKD despite their clinical importance.

Researchers developed and validated an interpretable machine learning model to predict 5-year all-cause mortality in non-dialysis chronic kidney disease using data from 1,858 patients in the KNOW-CKD prospective cohort, with 94 deaths observed. The CatBoost model achieved AUC 0.813 versus 0.747 for logistic regression, and a simplified version using age, eGFR, albumin, urine protein-to-creatinine ratio, and total calcium retained AUC 0.795 in an external cohort of 348 patients.

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

Updated Aug 8, 2026 · TRV-2026-0688

67Index score
350
HealthStableModerate evidence · 1 source

AI tools used in healthcare can make errors that result in patient harm, with responsibility difficult to assign because decisions are distributed across human and technological agents.

The peer-reviewed article examines responsibility gaps when AI tools in healthcare cause patient harm. It notes that traditional models struggle because decisions are spread across clinicians, developers, institutions and the AI systems themselves.

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

Updated Aug 8, 2026 · TRV-2026-0686

67Index score
351
HealthStableModerate evidence · 1 source

Users also reported risks from ChatGPT mental health conversations, including exposure to incorrect health advice, overly validating responses, and privacy concerns.

Researchers analyzed Reddit posts and comments about mental health conversations with ChatGPT to understand how large language models are being used for support outside clinical settings. By October 2025, they found users described ChatGPT as accessible and non-judgmental, providing emotional support, validation, and practical help like navigating difficult conversations and preparing for therapy.

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

Updated Aug 7, 2026 · TRV-2026-0681

67Index score
352
LifestyleStableModerate evidence · 1 source

Users of the Replika companion chatbot experienced AI-induced sexual harassment including unsolicited sexual advances and persistent boundary violations, leading to discomfort and privacy concerns.

A peer-reviewed study published October 16, 2025 examined user reviews of the Replika companion chatbot to investigate reports of inappropriate sexual behavior. From 35,105 negative Google Play Store reviews, researchers identified 800 cases for thematic analysis.

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

Updated Aug 7, 2026 · TRV-2026-0680

67Index score

Recomputed live from the record · Oct 12, 2026, 9:34 PM