eLetters

1568 e-Letters

  • Regulated vs. Non-Regulated AI in UK Medicine: Commentary on Warrington and Holm (2024) and the Role of LLM Risk Assessment

    Comments related to the findings reported by Warrington and Holm regarding the use of Artificial Intelligence (AI) by UK General Medical Council (GMC) registered doctors (Warrington DJ, Holm S. BMJ Open 2024;14:e089090. doi:10.1136/bmjopen-2024-089090).
    A key observation regarding the study is its apparent lack of distinction between participants' use of regulated AI products (classified as medical devices) and non-regulated AI tools (such as general-purpose LLMs). The wide range of respondent specialties reported further highlights this potential issue; for instance, clinicians in radiology or pathology are more likely to encounter regulated, task-specific AI, whereas those in public health or psychiatry might be more likely to experiment with non-regulated, general-purpose models.
    During the review process, I used Gemini Advanced (specifically, the model designated by the user as 2.5 Pro Experimental) to assist with processing screenshots of table1, table 2 and fig 1 into spreadsheet processable data. The same Large Language Model (LLM) was also prompted to categorize the clinical risks associated with the AI uses listed in Figure 1 of the original paper. The author is of the opinion that the LLM’s risk categorisation (column 2 in table A below) adopted a patient-centric perspective.
    However, the author is of the opinion that a "composite" clinical risk assessment, which considers both the nature of the specific usage instance and the pot...

    Show More
  • Overdiagnosis: Misuse in Breast Cancer Screening

    In breast cancer screening, the term "overdiagnosis" is a misnomer. It would be more accurate to state that the natural history of screen-detected cancer has not been adequately verified.
    Overdiagnosis is typically defined as the diagnosis of a lesion as cancer that will not cause symptoms or result in death. This definition assumes that cancer detection is appropriate and occurs in individuals for whom the diagnosis would be clinically relevant. For instance, in elderly patients, cancer may not lead to symptoms or death within their life expectancy. However, the issue in breast cancer screening is not related to the duration of observation but rather to the diagnosis itself.
    Cancer is generally diagnosed when a mass is detected through imaging or endoscopy and its malignant nature is confirmed histopathologically. However, early-stage breast cancer is an unusual case. These lesions may not form a detectable mass and are diagnosed as cancer based solely on histopathological findings. There is no scientific or clinical verification that cancers identified in this manner are biologically cancerous. Consequently, most clinical studies on breast cancer screening are essentially uncontrolled case series, lacking rigorous controls and, as such, are not scientifically interpretable or statistically reliable.
    A systematic review by the U.S. Preventive Services Task Force (USPSTF) did not provide clear evidence that breast cancer screening reduces cancer...

    Show More
  • Spa Therapy Is Not a Pill: Reconsidering Methods in the Evaluation of Complex Interventions

    RJ Forestier, FB Erol Forestier, I Santos, A Muela Garcia, A Françon.
    Centre de Recherches Rhumatologiques et Thermales d’Aix-les-Bains, Aix Les Bains, France.

    This meta-analysis approaches spa therapy as if it were a pharmaceutical intervention, which we believe does not fully reflect the complex and multifaceted nature of such treatments.
    We have been conducting clinical trials and systematic reviews in this field for over 30 years. In our experience, spa therapy is a complex intervention traditionally based on the use of thermal mineral water, often combined with massages, baths, showers, mud applications, and supervised pool-based exercises -each of which may have therapeutic effects of its own.
    We were surprised by the conclusions of this meta-analysis regarding both the therapeutic effect and the risk of bias, as they differ markedly from our own findings and appear to stem from several questionable methodological choices.
    Bibliographic Incompleteness
    The limited scope of the literature search is particularly problematic. In 2020, we identified 122 comparative trials on balneotherapy, whereas this meta-analysis included only 42 randomized controlled trials. Our complementary search updated to 2025 identified 42 trials focused solely on knee osteoarthritis, and a total of 141 trials after removing duplicates related to multiple conditions. The highly selective inclusion criteria adopted in this meta-analysis substantially reduced t...

    Show More
  • Overdiagnosis: Misconceptions About Prostate and Breast Cancer Screening

    Overdiagnosis occurs when a lesion is identified as cancer despite the absence of symptoms or subsequent cancer-related death. (1) This is particularly common in elderly patients with shorter follow-up periods, assuming the cancer diagnosis is accurate. For prostate and breast cancers, the natural history of screen-detected cases remains unverified, and many of these cancers remain asymptomatic or non-lethal even with extended observation. Therefore, the issue lies with the diagnostic tests, not the duration of observation. While it is accurate to state that "colorectal cancer screening results in overdiagnosis in the elderly," describing prostate cancer as having an overdiagnosis issue is misleading. Instead, it should be clarified that "the natural history of screen-detected prostate cancer has not been verified" or that "there are limitations in the diagnostic tests."
    Cancer is typically diagnosed by detecting the formation of a mass through endoscopy or imaging, followed by histopathological confirmation that the cells are malignant. However, screen-detected prostate and breast cancers do not form masses and are diagnosed based solely on histopathology. Whether a lesion diagnosed in this way is biologically cancerous—that is, its natural history—has not been clinically or scientifically verified. (2-4)
    In cancer statistics, both prostate and breast cancers rank among the highest in terms of incidence. (5) However, most of these...

    Show More
  • Pre- and in-hospital pre-ROSC EEG during cardiopulmonary resucitation

    February 23, 2025, Eichinger et al. published a plan for pre-hospital bispectral index measurements (BIS) during CPR in out-of-hospital cardiac arrest patients and at hospitalization to
    a) determine feasibility of BIS during CPR,
    b) assess neurological outcomes -particularly return of consciousness before ROSC and
    c) grade cerebral performance category (CPC) at 1 month

    Highlighted strenghts were systematic data collection and focus on feasibility of BIS during and after CPR, whereas limitations included recruiting 45 patients, reliance on a particular EMS system and prioritization study feasibility over clinical outcomes (1).

    However, there are more weaknesses:
    Muscle signals (EMG) and artifacts in BIS due heart massage are expected to blur the EEG assessments – but easily eliminated by sedatives and paralytic agents. The major problem is 'noise' from EEG devices, electrical installations and human activities at the scene of arrest because the 'noise level' dictates whether electrocortical activities can be identified (2).
    Quality of Life (QoL) interviews of family members are suggested although one's own
    perception of life quality before CPR does not relate to qualities one month later.
    The reference list includes 18 articles of which 17 were published in 2017 or later and one neurosurgical from 1975.
    The number of patients is 45 whivch weakens the upper 95% confidence interval for th...

    Show More
  • Fair Informed Consent: is the tide turning?

    This article continues the evidence that the current Participant Information Sheet isn’t fit for its purpose. It no longer supports fair consent, rather, it has become a legal document focused on the needs of the research team, not the potential participant. Having spent much time as Research Ethics Advisor to the National Research Ethics Advisor and subsequently the Health Research Authority trying (unsuccessfully) to address this, I would support the authors' concluding sentence.
    But I believe the tide is turning (https://blogs-bmj-com.ezproxy.u-pec.fr/medical-ethics/posts-page/). The USA Food and Drug Administration draft guidance / requirements support and expect all patient-facing material to start with a comprehensible summary of information of importance to potential participants. Their requirements, if accepted, will carry great weight (https://www.fda.gov/media/176663/download) . On the Oxford A UK Research Ethics Committee, we were similarly concerned about this issue and together drew up and published proposals, which, perhaps unsurprisingly, overlap with the USA material (https://journals-sagepub-com.ezproxy.u-pec.fr/doi/10.1177/17470161231176932l ).
    These were shown to patient groups who responded very positively, but they now need to be tested in real practice, involving all with...

    Show More
  • Emergency care for young people after self-harm: a realist review protocol

    Letter:
    Dear Editor,
    We are writing in response to the article published by Romeu et al., 2025, "Emergency Care for Young People After Self-Harm: A Realist Review Protocol"[1]. The article provides a detailed examination of the factors contributing to the rising number of young people seeking emergency care after self-harm, such as increased mental health issues and limited access to preventive care, effectively highlighting the urgency of this public health issue through detailed literature review. Given increasing concerns about youth mental health and the strain on emergency services, this research is highly relevant and well-timed. The study follows a structured realist review methodology, incorporating Pawson’s five iterative stages, which include clarifying the scope, searching for evidence, appraising the quality of evidence, synthesizing findings, and drawing conclusions. The use of methodical searches, stakeholder engagement, and program theory development adds depth and credibility. Nevertheless, some methodological concerns and excluded factors limit the study’s thoroughness.
    Pilot searches have confirmed the originality and feasibility of this review by demonstrating unique findings not previously covered in existing literature and by successfully identifying relevant data sources and methodologies applicable to the study's scope. However, the limited geographic scope restricts its generalizability to other healthcare systems. W...

    Show More
  • Interesting read

    Having worked in clinical research, seeing that one of the barriers is: “a ‘lack of incentives to take part’ can have on people’s motivation to offer their time and resources to research, especially when doing so is perceived as having the potential to be detrimental to their physical and mental well-being.” was eye-opening. I think as clinicians, we tend to look at clinical research through the lens of ‘the reward/incentive for taking part is the opportunity for an improved health outcome that is new and innovative and not accessible to everyone yet’ and obviously with it being a trial there’s always a chance it won’t work or that there will be side effects, but we generally don’t have the perception that the trial could be detrimental to a participants health (perhaps with the exception of phase 1 CTIMP studies and last-line experimental therapies). So, to see that nearly 50% of the phase 2 group were in agreement that this is a barrier is certainly something to consider.

  • The financial burden of mental health disability in the UK

    Mental health disabilities have led to a significant financial burden on the UK economy, affecting not only individuals and their families but also the wider society and economic framework. As lecturers in mental health and finance, it is important to analyze these economic implications so we can understand the multi-dimensional challenges created by mental health disabilities. These challenges include:

    According to the Mental Health Foundation (2022), mental health disabilities cost the UK economy at least £117.9 billion annually, equivalent to approximately 5% of the nation's GDP. This figure comprises of various factors, including healthcare expenditures, loss of productivity, and informal care costs. Particularly, nearly three-quarters (72%) of this cost is attributed to lost productivity and the efforts of unpaid informal carers.

    Families with children suffering from mental health disabilities face significant financial challenges. Solmi, Melnychuk & Morris (2018) revealed that such families require an additional £49.31 per week to maintain a standard of living comparable to families without disabled children. This financial strain is worsened in low socioeconomic families, who may need between £59.28 and £81.26 more per week, depending on the severity of the disability.

    The UK government allocates substantial funding to disability benefits, with the annual health and disability benefits bill standing at £65 billion and projections indic...

    Show More
  • RE: Chronic kidney disease progression in patients with previous type 2 diabetes and/or hypertension: a population-based cohort study from primary care in Spain

    Cunillera-Puértolas et al. conducted a prospective study to evaluate the association of hypertension and type 2 diabetes (T2D) with chronic kidney disease (CKD) progression (1). CKD progression were predominant in patients with low estimated glomerular filtration rate (eGFR) and albuminuria, even at mild-moderate levels. In addition, patients with hypertension and T2D presented the greatest risk of CKD progression, and the risk of CKD progression was higher in patients with hypertension than those with T2D. I present a comment.

    There is a need of speculating the difference between hypertension and T2D for predicting the risk of CKD progression. Although the severity of hypertension and T2D may contribute to the risk of CKD progression, medications for hypertension and T2D would also relate to the risk of CKD progression. James et al. conducted a 4-year prospective study, and the lower eGFRs and higher urine albumin-creatinine ratio (ACR) were significantly associated with higher risk of acute kidney injury (AKI), regardless of having T2D or hypertension (2). Kidney damage with lower eGFRs and higher ACR would contribute to the CKD progression, and I speculate that vascular damage in the kidney with T2D and hypertension would be definitely existed. Endothelial dysfunctions via T2D and hypertension may become one of the mechanisms, which contribute to the risk of CKD progression, and obesity is a key factor for the risk assessment (3).

    Additional epidemiologica...

    Show More

Pages