Latest Health Informatics research

Top 10 this week

Week of 1 November 2026

1. 1

   Effectiveness of Wrist-Worn Wearable-Guided Cardiac Rehabilitation Delivered Through Fully Remote or Hybrid Programs in Patients With Cardiovascular Disease: Systematic Review and Meta-Analysis of Randomized Controlled Trials

   JMIR mHealth and uHealth · Q1 journal · SR/MA of RCTs · Evidence level 1 (High) · 27 October 2026

   Across 18 randomized trials involving 1,993 people with cardiovascular disease, wearable-guided remote or hybrid rehab improved walking distance by an average of 31 meters and daily activity by about 1,063 steps. Results varied widely, and the programs also included feedback and behavioral support, so the benefit cannot be attributed to wearables alone; consistency and long-term effects remain uncertain.

2. 2

   Effectiveness of Artificial Intelligence Systems in Healthcare: A Systematic Review of Clinical and Organizational Outcomes

   Healthcare (Basel, Switzerland) · Q2 journal · SR/MA of RCTs · Evidence level 1 (High) · 20 October 2026

   In this review of 54 randomized trials, 40 reported a positive direction of effect, though that does not mean every intervention improved patient outcomes. Diagnostic AI showed consistent accuracy gains, while conversational tools had positive signals for self-management and adherence; decision-support results depended on workflow integration. Larger pragmatic trials with standardized outcomes are still needed to show where AI works reliably.

3. 3

   Reporting Gaps and Challenges in Randomized Trials of Generative AI for Medical Use: Insights From a Systematic Review

   Journal of medical Internet research · Q1 journal · SR/MA of RCTs · Evidence level 1 (High) · 28 October 2026

   A review of 74 reports from 73 randomized trials found major gaps in reporting how generative AI systems received and processed inputs: only 4% described input data eligibility, and none reported input quality handling. Some checklist items may not fit generative AI, and the review used nonvalidated assessment questions, but missing technical details make trial findings harder to reproduce and assess.

4. 4

   Clinical Effectiveness, Safety, and Implementation of Artificial Intelligence-Assisted Insulin Management Systems in Type 1 and Type 2 Diabetes: A Systematic Review of Randomized Controlled Trials

   Diseases (Basel, Switzerland) · Q2 journal · SR/MA of RCTs · Evidence level 1 (High) · 13 October 2026

   Across seven randomized trials, AI-assisted insulin systems showed no increase in severe hypoglycemia, diabetic ketoacidosis, or serious adverse events; some improved time in range, adherence, or insulin dosing. Overall, results were comparable to conventional care, but small studies, short follow-up, and varied technologies leave their long-term benefits and best role in practice uncertain.

5. 5

   Evaluating the effectiveness of the eHealth-faciLitated SMILe-integrated care model on treatment burden, quality of life and graft-versus-host disease after allogeneic stem cell transplantation: results of two European randomized controlled implementation studies

   Annals of hematology · Q2 journal · RCT · Evidence level 1 (High) · 24 October 2026

   In two European randomized studies involving 152 stem cell transplant recipients, nurse-led integrated care with eHealth was linked to fewer cases of moderate-to-severe acute graft-versus-host disease: 34% versus 55% with usual care. Treatment burden was lower across five dimensions, but quality of life and chronic GVHD did not differ; the acute GVHD result needs further investigation.

6. 6

   Implementing Australia's psychotropic clinical practice guidelines in residential aged care: the EMBRACE helix counterbalanced randomised controlled trial comparing knowledge brokers with pharmacist education and passive distribution

   Implementation science : IS · Q1 journal · RCT · Evidence level 1 (High) · 29 July 2026

   Implementation trial of knowledge‑broker vs pharmacist education

7. 7

   Incorporating Objective Behavioral and Biological Measures in a Large-Scale Decentralized Mobile Health Trial for Depression: Randomized Controlled Trial

   JMIR mental health · Q1 journal · RCT · Evidence level 1 (High) · 9 October 2026

   In a fully remote trial of 1,156 adults with elevated depressive symptoms, 84% completed at least one follow-up measure; at least 75% submitted biological samples, and most completed measures yielded analyzable data. Older participants and those with college degrees completed more follow-up assessments, so this approach looks feasible but may not engage all groups equally.

8. 8

   Implementation strategies for digital relapse prevention in schizophrenia identified during a randomized controlled trial

   NPJ digital medicine · Q1 journal · RCT · Evidence level 1 (High) · 30 October 2026

   In a workshop midway through an ongoing multisite trial, 16 stakeholders pointed to changes in discharge planning, team collaboration, staff responsibilities and technical workflows to support routine use of a smartphone relapse-prevention app. These findings offer implementation guidance, not evidence that the app prevents relapse or that the recommended changes improve uptake; their value in routine care remains unknown.

9. 9

   AI-generated vs. human therapist feedback in internet-based CBT for social anxiety disorder: A six-arm randomized controlled trial with deception design

   Internet interventions · Q1 journal · RCT · Evidence level 1 (High) · 16 October 2026

   In this trial, psychologist-reviewed AI feedback and human therapist feedback were not statistically different in reducing social anxiety, while active treatment produced large improvements compared with waitlist. The study did not establish that AI feedback was non-inferior, so the results do not prove the approaches are interchangeable; participants initially rated treatment as more credible when they expected human feedback.

10. 10

    Differential Effect of Endovascular Treatment in Medium or Distal Vessel Occlusion: A Distal Secondary Analysis

    Stroke · Q1 journal · RCT · Evidence level 5 (Expert Opinion) · 30 October 2026

    External model validation in randomized trial data informs treatment decisions

Paper of the day, last two weeks

• Digital and artificial intelligence-assisted medication management for older adults with chronic diseases: a systematic review and meta-analysis of randomized controlled trials

  Frontiers in public health · Q1 journal · SR/MA of RCTs · Evidence level 1 (High) · 22 October 2026

  A review of 19 randomized trials found no established benefit for hospitalization or mortality from digital or AI-assisted medication management in older adults with chronic diseases. Only six trials could be pooled, and adherence results varied widely, preventing a reliable combined estimate. Few comparable studies and limited high-quality evidence leave the value of routine use in geriatric care uncertain.

• Artificial intelligence in cardiovascular medicine: systematic review of randomized controlled trials

  European heart journal. Digital health · Q1 journal · SR/MA of RCTs · Evidence level 1 (High) · 12 October 2026

  A review of 25 randomized trials found that 17 reported AI strategies were non-inferior or superior to standard care, but 84% had high risk of bias or some concerns. Only 36% explicitly followed AI-specific reporting guidelines, and most studies were small in scope; robust, multicentre trials measuring patient-relevant outcomes are still needed before wider adoption.

• A local outpatient practice-level prediction model for short-term psychiatric emergency presentation

  JAMIA open · Q1 journal · Prospective cohort · Evidence level 2 (Moderate) · 2 November 2026

  In one Montana psychiatry practice, a model using health information exchange data distinguished snapshots preceding a psychiatric or substance-related emergency visit within 30 days, with an area under the curve of 0.85. It outperformed general risk scores, but only 68 patients had visits and estimates were imprecise; prospective testing must show whether using the model improves care.

• An integrated CT/MRI diagnostic system for pancreatic cystic lesions that supports pathological classification and risk stratification

  Cancer biology & medicine · Q1 journal · Prospective cohort · Evidence level 2 (Moderate) · 29 October 2026

  In a 1,102-patient retrospective study, the imaging-based system classified five pancreatic cyst types with 93% discrimination and was associated with fewer unnecessary surgeries than current practice: 11.5% versus 23.1%. In a prospective group of 76 patients, classification accuracy was 86.2%; the results apply to surgically evaluated cysts, and broader validation is still needed.

• Artificial intelligence-enabled interventions for smoking cessation: a systematic review and meta-analysis of randomised controlled trials

  International journal of medical informatics · Q1 journal · SR/MA of RCTs · Evidence level 1 (High) · 29 October 2026

  Across 14 trials, pooled results did not establish that AI-enabled support improves quitting at six months or longer; the estimate was imprecise and varied across studies. A favorable signal appeared in a smaller analysis of biochemically confirmed quitting, but it excluded the largest null trials. Current evidence also does not show that AI adds benefit over well-designed non-AI digital support.

• Subgroup deployment-readiness audit of calibration and clinical utility in perioperative infection prediction

  Journal of the American Medical Informatics Association : JAMIA · Q1 journal · Prospective cohort · Evidence level 2 (Moderate) · 30 October 2026

  In 10,719 procedures, the model discriminated well overall, but calibration and decision-making value differed by physiological reserve. Recalibration brought overall calibration close to ideal, yet predictions still overestimated infection risk in a later time period. Infection coding was largely absent before 2018 and differed across reserve groups, so prospective validation with comparable outcome tracking is needed before clinical use.

• Point of care randomization (POCR) engine: demonstration of a transportable clinical informatics technology that improves health systems' ability to conduct embedded clinical trials

  JAMIA open · Q1 journal · RCT · Evidence level 5 (Expert Opinion) · 28 October 2026

  EHR-embedded randomization deployed with measurable practice impact

• External validation and clinical implementation of clinical prediction models: a systematic review

  Journal of the American Medical Informatics Association : JAMIA · Q1 journal · SR of cohorts · Evidence level 2 (Moderate) · 30 October 2026

  Among 1,072 machine-learning prediction models, only about 1 in 10 underwent external validation and 1 in 20 reached clinical implementation; 39% of implemented models had no prior external validation. Code availability and better reporting were linked to validation, but not consistently to implementation, suggesting real-world deployment follows a distinct path with operational demands still to address.

• Association of ambient artificial intelligence scribe availability with patient instruction documentation and readability in ambulatory care

  Journal of the American Medical Informatics Association : JAMIA · Q1 journal · Retrospective cohort · Evidence level 3 (Low) · 22 October 2026

  Evaluates AI scribe effects on documentation workflow and patient instructions

• Precision targeting using machine learning to optimize medication therapy management delivery in value-based care: A quasi-experimental study using real-world evidence

  Journal of managed care & specialty pharmacy · Q1 journal · Non-randomized trial · Evidence level 2 (Moderate) · 1 November 2026

  In a matched analysis of nearly 184,000 patient pairs, machine-learning targeting was associated with an 8.6 percentage-point increase in acceptance of pharmacist consultations compared with traditional eligibility-based assignment. Because the study compared different years rather than randomly assigning patients, it cannot establish that the targeting model caused the improvement; whether results hold across other pharmacy programs remains unknown.

• Q-AortaECG: A Quantum-Inspired Multi-Agent 12-Lead ECG System for Early Triage of Acute Aortic Syndrome

  IEEE journal of biomedical and health informatics · Q1 journal · Prospective cohort · Evidence level 2 (Moderate) · 23 October 2026

  In a 700-patient locked test, Q-AortaECG identified 95.5% of confirmed acute aortic syndromes, but specificity was 72.1%, so false alarms remain a concern. In a reader study, assistance raised clinician sensitivity from 69.6% to 87.0%; silent-run alerts could have shortened time to the correct diagnostic pathway, but CT angiography and expert imaging remain the diagnostic standard.