I was surprised to see the discouraging New York Times headline: “What Works for Low Back Pain? Not Much, a New Study Says” [1]. This headline referred to a study by Cashin et al [2] that concluded that “The current evidence shows that only one in 10 non-surgical and non-interventional treatments for low back pain are efficacious, providing only small analgesic effects beyond placebo.” This conclusion surprised me because our research team had conducted randomized controlled trials evaluating various non-pharmacological treatments for chronic low back pain (CLBP) and found most of them more effective than routine care [3,4,5]. Our research, and that of many others, contributed evidence that led the American College of Physicians (ACP) to publish clinical guidelines recommending 13 evidence-based non-pharmacological treatments options for CLBP as first line treatments [6].
Several explanations for the apparently conflicting conclusions of the Cashin review [2] and the ACP guidelines [6] come to mind:
1) Review [2] was restricted to placebo-controlled trials. As the authors acknowledged, identifying credible placebos for non-pharmacological treatments and interpreting the results of such trials is challenging. what are credible placebos for yoga, massage, or mindfulness? Placebo-controlled trials work well for evaluating the efficacy of medications, but have serious limitations for studying more complex treatments.
2) Review [2] was restricted to pain...
I was surprised to see the discouraging New York Times headline: “What Works for Low Back Pain? Not Much, a New Study Says” [1]. This headline referred to a study by Cashin et al [2] that concluded that “The current evidence shows that only one in 10 non-surgical and non-interventional treatments for low back pain are efficacious, providing only small analgesic effects beyond placebo.” This conclusion surprised me because our research team had conducted randomized controlled trials evaluating various non-pharmacological treatments for chronic low back pain (CLBP) and found most of them more effective than routine care [3,4,5]. Our research, and that of many others, contributed evidence that led the American College of Physicians (ACP) to publish clinical guidelines recommending 13 evidence-based non-pharmacological treatments options for CLBP as first line treatments [6].
Several explanations for the apparently conflicting conclusions of the Cashin review [2] and the ACP guidelines [6] come to mind:
1) Review [2] was restricted to placebo-controlled trials. As the authors acknowledged, identifying credible placebos for non-pharmacological treatments and interpreting the results of such trials is challenging. what are credible placebos for yoga, massage, or mindfulness? Placebo-controlled trials work well for evaluating the efficacy of medications, but have serious limitations for studying more complex treatments.
2) Review [2] was restricted to pain outcomes. Our research team’s RCTs evaluating acupuncture, yoga, and Mindfulness-Based Stress Reduction found stronger effects on functional outcomes than on pain [3,4,5]. Thus, had the review [2] been able to study treatment effects on patient function, it might have found more encouraging results. Furthermore, improving function may lead to reduced pain.
3) The review [2] compared mean differences between intervention and control groups. Such a comparison could obscure differences between the intervention and placebo groups in the percentages of study participants who experienced clinically meaningful improvements. Given the high prevalence of chronic pain and the generally modest effects of all treatments for CLBP, even if each of several additional treatment options meaningfully benefitted only an additional 10% of persons with CLBP (compared with placebo) the quality of life of millions of Americans could be improved.
Beyond the technical issues mentioned above, the contrast between research questions underlying the review [2] and ACP [6] studies raises an important philosophical question: What type of evidence is most valuable for improving the care of patients with CLBP? Is it evidence of superiority of a treatment compared with a placebo? Or is it evidence that the introduction of a “new” treatment into usual care improves the lives of persons with CLBP? I would argue that the second question is more important given the evidence that most treatments recommended by the ACP guidelines have been found more effective than typical primary care for CLBP, are widely available, relatively safe, and inexpensive. Furthermore, without credible placebo controls for many non-Rx treatments, efficacy trials may never produce credible results.
CLBP is a complex condition and different treatments seem to work for different patients. Requiring evidence from placebo-controlled trials to justify use of ACP-recommended treatments [6] would greatly reduce patient access to treatments that appear superior to the often ineffective, costly, and sometimes harmful interventions that have already failed to benefit them.
References
1. https://www.nytimes.com/2025/03/18/well/lower-back-pain-treatments-nsaid...
2. Cashin AG, Furlong BM, Kamper SJ, De Carvalho D, Machado LA, Davidson SR, Bursey KK, Abdel Shaheed C, Hall AM. Analgesic effects of non-surgical and non-interventional treatments for low back pain: a systematic review and meta-analysis of placebo-controlled randomised trials. BMJ Evid Based Med. 2025 Mar 18:bmjebm-2024-112974. doi: 10.1136/bmjebm-2024-112974. Epub ahead of print. PMID: 40101974.
3. Cherkin DC, Sherman KJ, Avins AL, Erro JH, Ichikawa L, Barlow WE, Delaney K, Hawkes R, Hamilton L, Pressman A, Khalsa PS, Deyo RA. A randomized trial comparing acupuncture, simulated acupuncture, and usual care for chronic low back pain. Arch Intern Med. 2009 May 11;169(9):858-66. doi: 10.1001/archinternmed.2009.65. PMID: 19433697; PMCID: PMC2832641.
4. Sherman KJ, Cherkin DC, Wellman RD, Cook AJ, Hawkes RJ, Delaney K, Deyo RA. A randomized trial comparing yoga, stretching, and a self-care book for chronic low back pain. Arch Intern Med. 2011 Dec 12;171(22):2019-26. doi: 10.1001/archinternmed.2011.524. Epub 2011 Oct 24. PMID: 22025101; PMCID: PMC3279296.
5. Cherkin DC, Sherman KJ, Balderson BH, Cook AJ, Anderson ML, Hawkes RJ, Hansen KE, Turner JA. Effect of Mindfulness-Based Stress Reduction vs Cognitive Behavioral Therapy or Usual Care on Back Pain and Functional Limitations in Adults With Chronic Low Back Pain: A Randomized Clinical Trial. JAMA. 2016 Mar 22-29;315(12):1240-9. doi: 10.1001/jama.2016.2323. PMID: 27002445; PMCID: PMC4914381.
6. Qaseem A, Wilt TJ, McLean RM, Forciea MA; Clinical Guidelines Committee of the American College of Physicians; Denberg TD, Barry MJ, Boyd C, Chow RD, Fitterman N, Harris RP, Humphrey LL, Vijan S. Noninvasive Treatments for Acute, Subacute, and Chronic Low Back Pain: A Clinical Practice Guideline From the American College of Physicians. Ann Intern Med. 2017 Apr 4;166(7):514-530. doi: 10.7326/M16-2367. Epub 2017 Feb 14. PMID: 28192789.
The study “Analgesic effects of non-surgical and non-interventional treatments for low back pain: a systematic review and meta-analysis of placebo-controlled randomised trials” [1] which was published in this journal has generated a lot of comments. The authors performed a rigorous statistical analysis of several hundred trials.
While the publication has been widely covered, some media reports have implied that nothing can be done for low back pain. The publication specifically addresses "non-surgical and non-interventional treatments." Some media reports also overlooked the remark of the authors that “This study supports the efficacy of several non-surgical and non-interventional treatments for reducing pain intensity compared with placebo in low back pain.” The authors pointed out the limitations such as some treatments having only one trial with less than 100 participants in each group. They cautioned “There is a clear need for large, high-quality, placebo-controlled trials to reduce uncertainty in efficacy estimates for many non-surgical and non-interventional treatments.”
Given the widespread interest in the topic, many individuals will undoubtedly share news of the publication with their healthcare providers. It is our responsibility to keep the train on track and ensure that the conversation is focused, accurate, and compassionate.
Low back pain is a leading cause of years lived with disability. Chronic lo...
The study “Analgesic effects of non-surgical and non-interventional treatments for low back pain: a systematic review and meta-analysis of placebo-controlled randomised trials” [1] which was published in this journal has generated a lot of comments. The authors performed a rigorous statistical analysis of several hundred trials.
While the publication has been widely covered, some media reports have implied that nothing can be done for low back pain. The publication specifically addresses "non-surgical and non-interventional treatments." Some media reports also overlooked the remark of the authors that “This study supports the efficacy of several non-surgical and non-interventional treatments for reducing pain intensity compared with placebo in low back pain.” The authors pointed out the limitations such as some treatments having only one trial with less than 100 participants in each group. They cautioned “There is a clear need for large, high-quality, placebo-controlled trials to reduce uncertainty in efficacy estimates for many non-surgical and non-interventional treatments.”
Given the widespread interest in the topic, many individuals will undoubtedly share news of the publication with their healthcare providers. It is our responsibility to keep the train on track and ensure that the conversation is focused, accurate, and compassionate.
Low back pain is a leading cause of years lived with disability. Chronic low back pain is often multifactorial and commonly linked to a concurrent issue that remains a source of pain, even after another contributing factor has been resolved. Some patients who have had surgery for conditions like a herniated disc with sciatica may still suffer from persistent back pain, which we described as “Post-surgical spine syndrome” [2,3]. It is also known as Persistent Spinal Pain Syndrome (PSPS) [4].
While it may not have a definitive “cure,” a compassionate conversation with the patient and their family about ways to improve comfort and enhance quality of life can make a significant difference. Certain measures can help to alleviate symptoms and slow their progression.
Medicine is not solely focused on curing diseases; prevention is equally vital. Patients and individuals should receive guidance on the importance of preventive measures like stretching, exercise, physical activity, proper ergonomics, a balanced diet, and weight control.
References
1. Cashin AG, Furlong BM, Kamper SJ, et al. Analgesic effects of non-surgical and non-interventional treatments for low back pain: a systematic review and meta-analysis of placebo-controlled randomised trials
BMJ Evidence-Based Medicine Published Online First: 18 March 2025. doi: 10.1136/bmjebm-2024-112974
4. Christelis N, Simpson B, Russo M, et al. Persistent Spinal Pain Syndrome: A Proposal for Failed Back Surgery Syndrome and ICD-11. Pain Med. 2021;22(4):807-818. doi:10.1093/pm/pnab015
We read with great interest the article by MacCradden et al. titled “What makes a ‘good’ decision with artificial intelligence? A grounded theory study in paediatric care” published recently in BMJ Evidence-Based Medicine.1 The authors present a commendable effort to develop a framework for integrating machine learning (ML) models into clinical decision-making, particularly in the complex and high-stakes environment of paediatric intensive care. While the study offers valuable insights into the interplay between medical knowledge, contextual factors, and ML tools, we believe several underexplored aspects warrant further discussion to enhance the framework’s applicability and relevance to evidence-based medicine.
Firstly, while the study acknowledges the importance of contextual factors, it does not sufficiently address the dynamic and evolving nature of clinical contexts. The hypothetical case of Siri, a 4-month-old with trisomy 18, provides a useful starting point, but the framework’s generalisability to other clinical scenarios remains unclear. For instance, how might the framework adapt to decisions in outpatient settings, where longitudinal patient interactions and resource constraints differ significantly from the ICU? Additionally, the study’s reliance on a single hypothetical case may limit the transferability of its findings.2 Future iterations of the framework could benefit from incorporating a broader range of clinical scenarios, includ...
We read with great interest the article by MacCradden et al. titled “What makes a ‘good’ decision with artificial intelligence? A grounded theory study in paediatric care” published recently in BMJ Evidence-Based Medicine.1 The authors present a commendable effort to develop a framework for integrating machine learning (ML) models into clinical decision-making, particularly in the complex and high-stakes environment of paediatric intensive care. While the study offers valuable insights into the interplay between medical knowledge, contextual factors, and ML tools, we believe several underexplored aspects warrant further discussion to enhance the framework’s applicability and relevance to evidence-based medicine.
Firstly, while the study acknowledges the importance of contextual factors, it does not sufficiently address the dynamic and evolving nature of clinical contexts. The hypothetical case of Siri, a 4-month-old with trisomy 18, provides a useful starting point, but the framework’s generalisability to other clinical scenarios remains unclear. For instance, how might the framework adapt to decisions in outpatient settings, where longitudinal patient interactions and resource constraints differ significantly from the ICU? Additionally, the study’s reliance on a single hypothetical case may limit the transferability of its findings.2 Future iterations of the framework could benefit from incorporating a broader range of clinical scenarios, including those with varying levels of diagnostic uncertainty, resource availability, and patient demographics.
Secondly, the study highlights the importance of model interpretability and transparency, yet it does not fully explore the implications of algorithmic bias and data representativeness. While participants expressed concerns about whether patients “like Siri” were adequately represented in the training data, the study does not delve into how biases in historical data might perpetuate inequities in care. For example, how might the framework account for disparities in access to care or differences in clinical outcomes across socioeconomic or racial groups? Addressing these issues is critical to ensuring that AI tools do not exacerbate existing health inequities,3 particularly in paediatric populations where vulnerability is heightened.
Thirdly, the study’s reliance on a simulated ML model, while pragmatic, raises questions about the framework’s applicability to real-world clinical settings. The authors acknowledge this limitation but do not explore the potential discrepancies between simulated and actual model performance. In practice, clinicians may encounter models with varying levels of accuracy, robustness, and generalisability. How might the framework guide decision-making when model predictions are inconsistent or when models fail to perform as expected in specific patient populations? Incorporating real-world validation studies into the framework’s development could provide more nuanced insights into its practical utility.4
Finally, while the study emphasises the importance of shared decision-making, it does not fully address the ethical and practical challenges of integrating AI tools into these processes.5 For instance, how might clinicians communicate the limitations and uncertainties of AI predictions to families, particularly in emotionally charged situations? The framework could be strengthened by including specific guidance on how to navigate these conversations, ensuring that families are empowered to make informed decisions without undue reliance on algorithmic outputs.
In conclusion, MacCradden et al. have made a valuable contribution to the growing body of literature on AI in healthcare.1 However, to fully realise the potential of their framework, we urge the authors and the broader research community to address the limitations outlined above. By incorporating a wider range of clinical contexts, addressing algorithmic bias, validating the framework in real-world settings, and providing clearer guidance on ethical communication, the framework could become an even more robust tool for supporting evidence-based decision-making in paediatric care and beyond.
References
1. McCradden MD, Thai K, Assadi A, Tonekaboni S, Stedman I, Joshi S, Zhang M, Chevalier F, Goldenberg A. What makes a 'good' decision with artificial intelligence? A grounded theory study in paediatric care. BMJ Evid Based Med. 2025 Feb 12:bmjebm-2024-112919. doi: 10.1136/bmjebm-2024-112919.
2. Schloemer T, Schröder-Bäck P. Criteria for evaluating transferability of health interventions: a systematic review and thematic synthesis. Implement Sci. 2018 Jun 26;13(1):88. doi: 10.1186/s13012-018-0751-8.
3. d'Elia A, Gabbay M, Rodgers S, Kierans C, Jones E, Durrani I, Thomas A, Frith L. Artificial intelligence and health inequities in primary care: a systematic scoping review and framework. Fam Med Community Health. 2022 Nov;10(Suppl 1):e001670. doi: 10.1136/fmch-2022-001670.
4. Weinstein EJ, Ritchey ME, Lo Re V 3rd. Core concepts in pharmacoepidemiology: Validation of health outcomes of interest within real-world healthcare databases. Pharmacoepidemiol Drug Saf. 2023 Jan;32(1):1-8. doi: 10.1002/pds.5537.
5. Gerke S, Minssen T, Cohen G. Ethical and legal challenges of artificial intelligence-driven healthcare. Artificial Intelligence in Healthcare. 2020:295–336. doi: 10.1016/B978-0-12-818438-7.00012-5.
It is misleading (as stated in “what this study adds”) to described this trial as placebo controlled. Although dummy capsules were used to blind participants to which combination of curcumin or omeprazole they were receiving, no group received placebo only. One interpretation of the findings therefore remains that they are due to a placebo effect. It is unfortunate that this misrepresentation of the study design has already been picked up by a UK National newspaper (Guardian 12 September)
I want to express my concern regarding “Curcumin and proton pump inhibitors for functional dyspepsia: a randomised, double blind controlled trial” by Kongkam et al(1). It was published against the journal’s editorial policy and has serious issues with reporting and interpretation of results.
The article shouldn’t have been published in the first place. It lacks prospective registration, which directly contradicts the BMJ Evidence-based medicine editorial policy stating that a prospective registration is mandatory for any clinical trials(2). The Thai Clinical Trials Registry(3) registration TCTR20221208003 is retrospective which is clearly stated in the registry. The registration was submitted on 07 December 2022, just before a preprint was posted on medRxiv on 09 December 2022, while the study was completed on 30 April 2020.
On top of that, there are serious issues with the reporting and interpretation of results.
According to the authors an equivalence design was used with the equivalence margin of 2 points in the SODA score. Nine comparisons of SODA scores in the curcumin plus omeprazole (C+O), curcumin only (C), and omeprazole only (O) groups were reported. For three of those confidence intervals include equivalence margin. The only available interpretation here is that the trial failed to demonstrate equivalence. To demonstrate equivalence the confidence intervals should be between the two equivalence margins rath...
I want to express my concern regarding “Curcumin and proton pump inhibitors for functional dyspepsia: a randomised, double blind controlled trial” by Kongkam et al(1). It was published against the journal’s editorial policy and has serious issues with reporting and interpretation of results.
The article shouldn’t have been published in the first place. It lacks prospective registration, which directly contradicts the BMJ Evidence-based medicine editorial policy stating that a prospective registration is mandatory for any clinical trials(2). The Thai Clinical Trials Registry(3) registration TCTR20221208003 is retrospective which is clearly stated in the registry. The registration was submitted on 07 December 2022, just before a preprint was posted on medRxiv on 09 December 2022, while the study was completed on 30 April 2020.
On top of that, there are serious issues with the reporting and interpretation of results.
According to the authors an equivalence design was used with the equivalence margin of 2 points in the SODA score. Nine comparisons of SODA scores in the curcumin plus omeprazole (C+O), curcumin only (C), and omeprazole only (O) groups were reported. For three of those confidence intervals include equivalence margin. The only available interpretation here is that the trial failed to demonstrate equivalence. To demonstrate equivalence the confidence intervals should be between the two equivalence margins rather than include them. The fact that “no significant differences were observed among the three groups” is also fully irrelevant, it does not demonstrate equivalence, as the latter cannot be claimed on the basis of nonsignificant tests(4).
Another striking deficiency in reporting is the unexplained loss to follow-up difference between the study arms. While 17 participants were lost to follow-up in the curcumin plus omeprazole and 17 in the curcumin only groups, only 1 was lost to follow-up in the omeprazole only arm. At the same time, the numbers of subjects who withdrew consent were also noticeably different – 2, 2 and 18 respectively. This difference is unlikely to have arisen by chance alone. There are two possible explanations: either the numbers of subjects who withdrew consent and those lost to follow-up were mistakenly swapped, or the loss to follow-up was systematically different in the omeprazole only arm compared to the two arms with curcumin. The former would question the peer review, the latter would question the blinding. Given the taste and smell of curcumin the blinding should have been questioned even if there were no differences in attrition.
It’s hard to believe that CONSORT requirements, explicitly stating that the interpretation must be consistent with the results and the primary and secondary outcome measures need to be completely defined and pre-specified, were sufficiently taken into consideration during the review process. I hope appropriate actions will be taken by the journal to correct for that.
(1) Kongkam P, Khongkha W, Lopimpisuth C, et alCurcumin and proton pump inhibitors for functional dyspepsia: a randomised, double blind controlled trialBMJ Evidence-Based Medicine 2023;28:399-406.
(4) Piaggio G, Elbourne DR, Pocock SJ, Evans SJW, Altman DG, CONSORT Group FT. Reporting of Noninferiority and Equivalence Randomized Trials: Extension of the CONSORT 2010 Statement. JAMA. 2012;308(24):2594–2604. doi:10.1001/jama.2012.87802
Health & healthcare is a human rights issue. Persons with intellectual disabilities use disproportionately more health care resources than the population without intellectual disabilities. In spite of this, they experience poorer health outcomes and they and their carers are significantly less satisfied with the quality of care provided to them by a variety of healthcare personnel. This can include doctors, pharmacists, nurses and other personnel.
The right to health contains freedoms. These freedoms include the right to be free from non-consensual medical treatment, such as medical experiments and research or forced sterilization, and to be free from torture and other cruel, inhuman or degrading treatment or punishment.
The right to health also contains entitlements. These entitlements include among others: The right to a system of health protection providing equality of opportunity for everyone to enjoy the highest attainable level of health; The right to prevention, treatment and control of diseases; Equal and timely access to basic health services; The provision of health-related education and information; Equal and timely access to basic health services etc.
Human rights are interdependent, indivisible and interrelated. This means that violating the right to health may often impair the enjoyment of other human rights, such as the rights to education or work, and vice versa. Persons with disabilities face various challenges to the enjoy...
Health & healthcare is a human rights issue. Persons with intellectual disabilities use disproportionately more health care resources than the population without intellectual disabilities. In spite of this, they experience poorer health outcomes and they and their carers are significantly less satisfied with the quality of care provided to them by a variety of healthcare personnel. This can include doctors, pharmacists, nurses and other personnel.
The right to health contains freedoms. These freedoms include the right to be free from non-consensual medical treatment, such as medical experiments and research or forced sterilization, and to be free from torture and other cruel, inhuman or degrading treatment or punishment.
The right to health also contains entitlements. These entitlements include among others: The right to a system of health protection providing equality of opportunity for everyone to enjoy the highest attainable level of health; The right to prevention, treatment and control of diseases; Equal and timely access to basic health services; The provision of health-related education and information; Equal and timely access to basic health services etc.
Human rights are interdependent, indivisible and interrelated. This means that violating the right to health may often impair the enjoyment of other human rights, such as the rights to education or work, and vice versa. Persons with disabilities face various challenges to the enjoyment of their right to health. Non-discrimination and equality are fundamental human rights principles and critical components of the right to health in this population group.
Healthcare professionals such as doctors, pharmacists and nurses and others should be mindful of the clinical effects associated with intellectual disabilities and how the individual patient's health and access to equitable healthcare is challenged and/or adversely affected. They must also take the time to understand the overall impact the condition has on the patient's life and on their carers and support network. Person centred care involves asking the right questions and listening carefully to patients with intellectual disability and their carers describing their unique challenges and experiences e.g. how to administer often complex medication regimens safely.
The right to health is a fundamental part of our human rights and of our understanding of a life with dignity. Regardless of our age, gender, socio-economic or ethnic background, ability/disability level our health is our most basic and essential asset. Improving health care providers' ( doctors, pharmacists, nurses and others) knowledge of intellectual disabilities and ability to provide culturally competent person centred care can limit the health disparities experienced by this disadvantaged population i.e. help ‘close the gap’.
Culturally competent doctors, pharmacists, nurses are important in eliminating health disparities experienced by people of all populations, including those with intellectual disabilities. Specialist pharmacists must be included in any multidisciplinary clinic as medication use is the main therapeutic intervention in the population with intellectual disabilities. Multidisciplinary teams with knowledge, expertise and experience provide ‘reasonable accommodations’ to the person with intellectual disability and equitable access to healthcare. The population with intellectual disabilities have a right to equitable access to healthcare leading to equal outcomes with the general population.
To: Juan VA Franco, MD
Editor-in-Chief
British Medical Journal, Evidence-Based Medicine
Dear Editor:
The study on cutaneous melanoma overdiagnosis attempts to tackle an important issue. However, we wish to address several methodological concerns that may warrant a critical evaluation of its conclusions.
First, one key study assumption is that the overdiagnosis of melanoma is due to over-screening by clinicians, including dermatologists. However, whether the patients were actually screened by clinicians is unknowable with the current study design. Thus, the lack of direct evidence to support this key assumption limits the study's capacity to attribute melanoma diagnoses to the prevalence of screening.
Second, the study's ecological methodology does not sufficiently account for variables that could affect melanoma diagnosis and mortality rate over time, such as advancements in diagnostic technologies, treatments, public awareness, and healthcare access, all factors that dermatologists have worked to improve over time. These factors could independently influence trends in melanoma incidence and mortality. This limitation is critical as it underlines the difficulty in drawing definitive conclusions from the ecological data presented.
Thirdly, the choice to manually input annual data from the SEER program into the DevCan software, deviating from the standard 3-year data aggregatio...
To: Juan VA Franco, MD
Editor-in-Chief
British Medical Journal, Evidence-Based Medicine
Dear Editor:
The study on cutaneous melanoma overdiagnosis attempts to tackle an important issue. However, we wish to address several methodological concerns that may warrant a critical evaluation of its conclusions.
First, one key study assumption is that the overdiagnosis of melanoma is due to over-screening by clinicians, including dermatologists. However, whether the patients were actually screened by clinicians is unknowable with the current study design. Thus, the lack of direct evidence to support this key assumption limits the study's capacity to attribute melanoma diagnoses to the prevalence of screening.
Second, the study's ecological methodology does not sufficiently account for variables that could affect melanoma diagnosis and mortality rate over time, such as advancements in diagnostic technologies, treatments, public awareness, and healthcare access, all factors that dermatologists have worked to improve over time. These factors could independently influence trends in melanoma incidence and mortality. This limitation is critical as it underlines the difficulty in drawing definitive conclusions from the ecological data presented.
Thirdly, the choice to manually input annual data from the SEER program into the DevCan software, deviating from the standard 3-year data aggregation, raises concerns about the accuracy and reliability of the findings. Such a departure from standardized methods without a robust justification may introduce bias or artifacts into the analysis. Furthermore, this approach complicates the interpretation of trends over time, as it may attribute undue significance to annual fluctuations that could be mere artifacts of random variation rather than indicative of meaningful trends.
In light of these concerns, we urge a cautious interpretation of the study's conclusions regarding the overdiagnosis of cutaneous melanoma. It is essential for future research in this area to address these methodological issues and be inclusive of all racial and ethnic groups, ensuring a more robust and reliable basis for understanding rate of melanoma diagnosis.
Sincerely,
Anna McNay M.D.
President
Melissa Shive M.D., M.P.H.
President Elect
On behalf of the California Society of Dermatology and Dermatologic Surgery (CalDerm)
we are writing in response to the article titled " How methodological pitfalls have created widespread misunderstanding about long COVID”(1).
We agree with the authors, that the existing epidemiological research on long COVID has suffered from overly broad case definitions and a striking absence of control groups, which may have led in an overestimation of risk.
It is important to acknowledge that Long- and Post-COVID syndrome are heterogeneous conditions, likely comprising different pathomechanistic groups such as autoimmunity, mitochondrial dysfunction, and virus persistence (2). This complexity, coupled with the lack of routine biomarkers, makes it difficult to accurately define and study this condition. Høeg et al et al. therefore raise some relevant points regarding the challenges faced in studying Long- and Post-COVID syndrome, particularly the need for properly matched control groups and internationally-established diagnostic criteria. Regarding the latter, the authors of the article themselves fail to use definitions accurately, particularly in distinguishing between the now consented WHO definitions of Long-COVID and Post-COVID (WHO/2019-nCoV/Post_COVID-19_condition/Clinical_case_definition/2021.1).
It is true, that initial studies depicted a high prevalence of Post-COVID syndrome (PCS). However, more recent population-based studies present a different perspective. In assessing the clinical picture, the primary focus isn...
we are writing in response to the article titled " How methodological pitfalls have created widespread misunderstanding about long COVID”(1).
We agree with the authors, that the existing epidemiological research on long COVID has suffered from overly broad case definitions and a striking absence of control groups, which may have led in an overestimation of risk.
It is important to acknowledge that Long- and Post-COVID syndrome are heterogeneous conditions, likely comprising different pathomechanistic groups such as autoimmunity, mitochondrial dysfunction, and virus persistence (2). This complexity, coupled with the lack of routine biomarkers, makes it difficult to accurately define and study this condition. Høeg et al et al. therefore raise some relevant points regarding the challenges faced in studying Long- and Post-COVID syndrome, particularly the need for properly matched control groups and internationally-established diagnostic criteria. Regarding the latter, the authors of the article themselves fail to use definitions accurately, particularly in distinguishing between the now consented WHO definitions of Long-COVID and Post-COVID (WHO/2019-nCoV/Post_COVID-19_condition/Clinical_case_definition/2021.1).
It is true, that initial studies depicted a high prevalence of Post-COVID syndrome (PCS). However, more recent population-based studies present a different perspective. In assessing the clinical picture, the primary focus isn't merely on the presence of symptoms but on the symptom-related reduction in quality of life. Our own study reveals that only 20% of symptomatic patients experience a significant reduction in their quality of life, indicating suffering from Post-COVID syndrome. This suggests a prevalence of Post-COVID syndrome at 5-6% for the entire population (3). The risk of developing Post-COVID syndrome after infection with currently circulating variants is notably lower than in patients from the initial waves of the infection (3). This aspect should be emphasized as part of a critical analysis of the issue, the authors can be expected to go into this in more depth and thus contribute to a more objective discussion. A systematic meta-analysis that considers different data collection times after infection and the method of symptom recording (self-reported versus documented through examinations) would have been more beneficial. This approach would provide a more comprehensive view rather than highlighting individual studies, thus avoiding polarized statements.
Furthermore, Høeg et al. suggest the inclusion of infection-negative control groups in future research. While this may be ideal, it is not realistic in the current state of the pandemic, which is now in its third year and has reached an endemic stage (Rahman et al., 2021). Additionally, the absence of antibodies does not exclude the presence of infection, as demonstrated in the Conan study, where half of the patients with PCR-confirmed infection had no detectable antibodies six weeks later (4).
Despite methodological limitations pointed out by Høeg et al. that apply to Post-COVID studies, it is crucial to recognize that Post-COVID syndrome is a relevant and frequent condition that requires further investigation in order to optimized diagnostic criteria and identify pathomechanism related biomarkers that may help to assign patients to appropriate randomized treatment studies according to the underlying mechanism. The findings regarding immune dysregulation described by A. Iwaski's working group are indeed promising and offer encouragement to many individuals affected by this condition (5). It is essential to continue studying and addressing post-COVID conditions to provide appropriate care and support for those affected.
Sincerely,
Mathias W. Pletz, MD and Andreas Stallmach, MD
Bibliography
1. Høeg TB, Ladhani S, Prasad V. How methodological pitfalls have created widespread misunderstanding about long COVID. BMJ Evid Based Med. 2023 Sep 25;
2. Davis HE, McCorkell L, Vogel JM, Topol EJ. Long COVID: major findings, mechanisms and recommendations. Nat Rev Microbiol. 2023 Mar;21(3):133–46.
3. Giszas B, Trommer S, Schüßler N, Rodewald A, Besteher B, Bleidorn J, et al. Post-COVID-19 condition is not only a question of persistent symptoms: structured screening including health-related quality of life reveals two separate clusters of post-COVID. Infection. 2023 Apr;51(2):365–77.
4. Weis S, Scherag A, Baier M, Kiehntopf M, Kamradt T, Kolanos S, et al. Antibody response using six different serological assays in a completely PCR-tested community after a coronavirus disease 2019 outbreak-the CoNAN study. Clin Microbiol Infect. 2021 Mar;27(3):470.e1-470.e9.
5. Klein J, Wood J, Jaycox J, Dhodapkar RM, Lu P, Gehlhausen JR, et al. Distinguishing features of Long COVID identified through immune profiling. Nature. 2023 Sep 25;
Dr Juan Franco
Editor-In-Chief
BMJ Evidence Based Medicine
BMA House
Tavistock Square
London WC1H 9JP
UNITED KINGDOM
31 October 2023
Dear Editor-In-Chief,
We read with interest the recent article by Høeg and colleagues that describes how methodological limitations in long COVID research distort risk and overestimate prevalence.[1]
The authors propose criteria to improve epidemiological research of long COVID. We write in support of these criteria, and to suggest two additions. We recently compared outcomes three months after PCR-confirmed COVID-19 infection with PCR-confirmed influenza infection, and found no difference between these illnesses.[2] Our comparative observational study had limitations (which we acknowledged) but was noteworthy because it was conducted in an Australian population that was primarily exposed to the Omicron variant after achieving high vaccination rates (>90%).
As a result, our two proposed additions to Høeg et al’s criteria relate to the exposed population which, as they suggest, should have diagnostic evidence of infection.
The first addition is to document the COVID variant to which this population was exposed. Recent data from Sweden shows a progressive (and substantial) decrease in the risk of long COVID from the wild type to the Omicron variant.[3] In addition, the type and frequency of symptoms has changed as the virus evolves.[4] This inclusion would improv...
Dr Juan Franco
Editor-In-Chief
BMJ Evidence Based Medicine
BMA House
Tavistock Square
London WC1H 9JP
UNITED KINGDOM
31 October 2023
Dear Editor-In-Chief,
We read with interest the recent article by Høeg and colleagues that describes how methodological limitations in long COVID research distort risk and overestimate prevalence.[1]
The authors propose criteria to improve epidemiological research of long COVID. We write in support of these criteria, and to suggest two additions. We recently compared outcomes three months after PCR-confirmed COVID-19 infection with PCR-confirmed influenza infection, and found no difference between these illnesses.[2] Our comparative observational study had limitations (which we acknowledged) but was noteworthy because it was conducted in an Australian population that was primarily exposed to the Omicron variant after achieving high vaccination rates (>90%).
As a result, our two proposed additions to Høeg et al’s criteria relate to the exposed population which, as they suggest, should have diagnostic evidence of infection.
The first addition is to document the COVID variant to which this population was exposed. Recent data from Sweden shows a progressive (and substantial) decrease in the risk of long COVID from the wild type to the Omicron variant.[3] In addition, the type and frequency of symptoms has changed as the virus evolves.[4] This inclusion would improve our understanding of post-viral impacts by variant, and add important context to Høeg et al’s sensible suggestion of a symptom-based approach to support patients.
The second addition involves documenting the population’s vaccination status, including (if possible) the time since last dose. A systematic review found that COVID-19 vaccination could protect against long COVID, but also noted that study quality was generally low for the reasons argued by Høeg et al.[5] While ongoing reinfections and decreased diagnostic testing may cloud the benefits of vaccination, vaccination status offers important insights into the symptoms and impacts of each COVID variant.
Finally, we concur with the article’s observations about the ongoing impact of regular negative reports about long COVID. We have heard people say they are “more afraid of getting long COVID than they are of getting COVID”. Høeg and colleagues have provided a framework to challenge the many inflated claims that may contribute to this fear and anxiety. While we believe and support those who experience post-viral effects, we must remember the most likely outcome after COVID-19 infection is a full recovery.
Yours sincerely,
Matthew Brown (Program Manager, Queensland Long COVID Response)
John Gerrard (Chief Health Officer, Queensland)
Ross Andrews (Senior Consultant Epidemiologist)
References
1 Hoeg TB, Ladhani S, Prasad V. How methodological pitfalls have created widespread misunderstanding about long COVID. BMJ Evid Based Med 2023.
2 Brown M, Gerrard J, McKinlay L, et al. Ongoing symptoms and functional impairment 12 weeks after testing positive for SARS-CoV-2 or influenza in Australia: an observational cohort study. BMJ Public Health 2023; 1(1).
3 Hedberg P, Naucler P. Post COVID-19 condition after SARS-CoV-2 infections during the omicron surge compared with the delta, alpha, and wild-type periods in Stockholm, Sweden. J Infect Dis 2023.
4 Looi MK. How are covid-19 symptoms changing? BMJ 2023; 380: 3.
5 Byambasuren O, Stehlik P, Clark J, et al. Effect of covid-19 vaccination on long covid: systematic review. BMJ Medicine 2023; 2(1).
We were surprised that BMJ Evidence Based Medicine chose to publish the flawed article by Høeg and co-authors on methodological limitations of research on long COVID (1). This piece appears to be a ‘Trojan Horse’ article where a scientifically dubious proposition escapes proper scrutiny because it is cloaked in otherwise plausible research commentary.
As the authors state, we need well designed studies to provide a valid measure of the long-term effects of acute COVID-19 infection (Long COVID). Such studies require robust case definitions, adequate duration of follow-up, and suitable comparison groups.
But in a section titled “The most well-designed studies provide reassuring estimates”, the authors include just two studies to support that sweeping statement. This highly selective ‘mini meta-analysis’ subverts the very purpose of evidence-based medicine. The main message of the Høeg paper appears to be that there is a negligible risk of long COVID, based on the selection of papers they have cited. That message does not fit with the actual body of scientific evidence (2). There is now overwhelming research that SARS-CoV-2 infection carries a significant risk of long-term effects over and above the generic effects of post-ICU syndrome and pneumonia (3).
The evidence of long-term effects comes from multiple sources, including epidemiological studies and basic science research looking at the severe and lasting pathological changes that occur in some pati...
We were surprised that BMJ Evidence Based Medicine chose to publish the flawed article by Høeg and co-authors on methodological limitations of research on long COVID (1). This piece appears to be a ‘Trojan Horse’ article where a scientifically dubious proposition escapes proper scrutiny because it is cloaked in otherwise plausible research commentary.
As the authors state, we need well designed studies to provide a valid measure of the long-term effects of acute COVID-19 infection (Long COVID). Such studies require robust case definitions, adequate duration of follow-up, and suitable comparison groups.
But in a section titled “The most well-designed studies provide reassuring estimates”, the authors include just two studies to support that sweeping statement. This highly selective ‘mini meta-analysis’ subverts the very purpose of evidence-based medicine. The main message of the Høeg paper appears to be that there is a negligible risk of long COVID, based on the selection of papers they have cited. That message does not fit with the actual body of scientific evidence (2). There is now overwhelming research that SARS-CoV-2 infection carries a significant risk of long-term effects over and above the generic effects of post-ICU syndrome and pneumonia (3).
The evidence of long-term effects comes from multiple sources, including epidemiological studies and basic science research looking at the severe and lasting pathological changes that occur in some patients following SARS-CoV-2 infection. This research shows that COVID-19 is a multisystem disease that can cause microclots, changes to the immune system, viral persistence in tissues, and other effects even in mild cases (4). These effects provide a basis for well-described sequelae of COVID-19 such as impaired brain function, extreme fatigue, and stroke.
Furthermore, tissue-level effects can be clinically silent but they include known risk factors for heart disease and other conditions, raising concerns for future population health. Viral persistence suggests additional potential for long-term effects that may take years or decades to emerge, as we have seen with other infections (5).
Even the research selectively cited in the Hoeg article is not as reassuring as the authors imply. Antonelli et al. reported an overall lower risk of Long COVID for Omicron compared with Delta infection (6). But they noted that Omicron variants have caused far higher case numbers and the conclusion of their article is that "future numbers with long COVID will inevitably rise".
Full declaration of potential conflicts of interest is important in evidence-based medicine. Yet at least two of the authors here have made no secret of their unorthodox scientific views on the origins, transmission, severity and prevention of COVID-19 and appear to be associated with particular ideological positions and organisations (7, 8).
Rather than the robust science we might reasonably expect from the highly regarded BMJ Evidence-Based Medicine, the Høeg paper serves to illustrate the very biases and errors which evidence-based medicine was established to challenge. It appears to be a ‘Trojan Horse’ from a partisan group. Its publication significantly damages the reputation of the journal as a platform for rigorous, principled, and balanced scientific debate.
References
1. Hoeg TB, Ladhani S, Prasad V. How methodological pitfalls have created widespread misunderstanding about long COVID. BMJ Evid Based Med 2023 doi: 10.1136/bmjebm-2023-112338 [published Online First: 2023/09/26]
2. Altmann DM, Whettlock EM, Liu S, et al. The immunology of long COVID. Nature reviews Immunology 2023;23(10):618-34. doi: 10.1038/s41577-023-00904-7 [published Online First: 2023/07/12]
3. Bowe B, Xie Y, Al-Aly Z. Postacute sequelae of COVID-19 at 2 years. Nat Med 2023;29(9):2347-57. doi: 10.1038/s41591-023-02521-2 [published Online First: 2023/08/22]
4. Castanares-Zapatero D, Chalon P, Kohn L, et al. Pathophysiology and mechanism of long COVID: a comprehensive review. Annals of medicine 2022;54(1):1473-87. doi: 10.1080/07853890.2022.2076901 [published Online First: 2022/05/21]
5. Chen B, Julg B, Mohandas S, et al. Viral persistence, reactivation, and mechanisms of long COVID. eLife 2023;12 doi: 10.7554/eLife.86015 [published Online First: 2023/05/04]
6. Antonelli M, Pujol JC, Spector TD, et al. Risk of long COVID associated with delta versus omicron variants of SARS-CoV-2. Lancet 2022;399(10343):2263-64. doi: 10.1016/S0140-6736(22)00941-2 [published Online First: 2022/06/20]
7. Bragman W. New scientist group calling for pandemic answers is tied to rightwing Dark Money. OptOut, 2023; https://www.optout.news/newsletters/norfolk-group.
8. Prasad V. Do not report COVID cases to schools & do not test yourself if you feel ill. Vinay Prasad's Observations and Thoughts, 2023; https://vinayprasadmdmph.substack.com/p/do-not-report-covid-cases-to-sch....
I was surprised to see the discouraging New York Times headline: “What Works for Low Back Pain? Not Much, a New Study Says” [1]. This headline referred to a study by Cashin et al [2] that concluded that “The current evidence shows that only one in 10 non-surgical and non-interventional treatments for low back pain are efficacious, providing only small analgesic effects beyond placebo.” This conclusion surprised me because our research team had conducted randomized controlled trials evaluating various non-pharmacological treatments for chronic low back pain (CLBP) and found most of them more effective than routine care [3,4,5]. Our research, and that of many others, contributed evidence that led the American College of Physicians (ACP) to publish clinical guidelines recommending 13 evidence-based non-pharmacological treatments options for CLBP as first line treatments [6].
Show MoreSeveral explanations for the apparently conflicting conclusions of the Cashin review [2] and the ACP guidelines [6] come to mind:
1) Review [2] was restricted to placebo-controlled trials. As the authors acknowledged, identifying credible placebos for non-pharmacological treatments and interpreting the results of such trials is challenging. what are credible placebos for yoga, massage, or mindfulness? Placebo-controlled trials work well for evaluating the efficacy of medications, but have serious limitations for studying more complex treatments.
2) Review [2] was restricted to pain...
Author: Joe Ordia, MD.
The study “Analgesic effects of non-surgical and non-interventional treatments for low back pain: a systematic review and meta-analysis of placebo-controlled randomised trials” [1] which was published in this journal has generated a lot of comments. The authors performed a rigorous statistical analysis of several hundred trials.
While the publication has been widely covered, some media reports have implied that nothing can be done for low back pain. The publication specifically addresses "non-surgical and non-interventional treatments." Some media reports also overlooked the remark of the authors that “This study supports the efficacy of several non-surgical and non-interventional treatments for reducing pain intensity compared with placebo in low back pain.” The authors pointed out the limitations such as some treatments having only one trial with less than 100 participants in each group. They cautioned “There is a clear need for large, high-quality, placebo-controlled trials to reduce uncertainty in efficacy estimates for many non-surgical and non-interventional treatments.”
Given the widespread interest in the topic, many individuals will undoubtedly share news of the publication with their healthcare providers. It is our responsibility to keep the train on track and ensure that the conversation is focused, accurate, and compassionate.
Low back pain is a leading cause of years lived with disability. Chronic lo...
Show MoreDear Editor,
We read with great interest the article by MacCradden et al. titled “What makes a ‘good’ decision with artificial intelligence? A grounded theory study in paediatric care” published recently in BMJ Evidence-Based Medicine.1 The authors present a commendable effort to develop a framework for integrating machine learning (ML) models into clinical decision-making, particularly in the complex and high-stakes environment of paediatric intensive care. While the study offers valuable insights into the interplay between medical knowledge, contextual factors, and ML tools, we believe several underexplored aspects warrant further discussion to enhance the framework’s applicability and relevance to evidence-based medicine.
Firstly, while the study acknowledges the importance of contextual factors, it does not sufficiently address the dynamic and evolving nature of clinical contexts. The hypothetical case of Siri, a 4-month-old with trisomy 18, provides a useful starting point, but the framework’s generalisability to other clinical scenarios remains unclear. For instance, how might the framework adapt to decisions in outpatient settings, where longitudinal patient interactions and resource constraints differ significantly from the ICU? Additionally, the study’s reliance on a single hypothetical case may limit the transferability of its findings.2 Future iterations of the framework could benefit from incorporating a broader range of clinical scenarios, includ...
Show MoreIt is misleading (as stated in “what this study adds”) to described this trial as placebo controlled. Although dummy capsules were used to blind participants to which combination of curcumin or omeprazole they were receiving, no group received placebo only. One interpretation of the findings therefore remains that they are due to a placebo effect. It is unfortunate that this misrepresentation of the study design has already been picked up by a UK National newspaper (Guardian 12 September)
Dear Editorial Office,
I want to express my concern regarding “Curcumin and proton pump inhibitors for functional dyspepsia: a randomised, double blind controlled trial” by Kongkam et al(1). It was published against the journal’s editorial policy and has serious issues with reporting and interpretation of results.
The article shouldn’t have been published in the first place. It lacks prospective registration, which directly contradicts the BMJ Evidence-based medicine editorial policy stating that a prospective registration is mandatory for any clinical trials(2). The Thai Clinical Trials Registry(3) registration TCTR20221208003 is retrospective which is clearly stated in the registry. The registration was submitted on 07 December 2022, just before a preprint was posted on medRxiv on 09 December 2022, while the study was completed on 30 April 2020.
On top of that, there are serious issues with the reporting and interpretation of results.
According to the authors an equivalence design was used with the equivalence margin of 2 points in the SODA score. Nine comparisons of SODA scores in the curcumin plus omeprazole (C+O), curcumin only (C), and omeprazole only (O) groups were reported. For three of those confidence intervals include equivalence margin. The only available interpretation here is that the trial failed to demonstrate equivalence. To demonstrate equivalence the confidence intervals should be between the two equivalence margins rath...
Show MoreHealth & healthcare is a human rights issue. Persons with intellectual disabilities use disproportionately more health care resources than the population without intellectual disabilities. In spite of this, they experience poorer health outcomes and they and their carers are significantly less satisfied with the quality of care provided to them by a variety of healthcare personnel. This can include doctors, pharmacists, nurses and other personnel.
The right to health contains freedoms. These freedoms include the right to be free from non-consensual medical treatment, such as medical experiments and research or forced sterilization, and to be free from torture and other cruel, inhuman or degrading treatment or punishment.
The right to health also contains entitlements. These entitlements include among others: The right to a system of health protection providing equality of opportunity for everyone to enjoy the highest attainable level of health; The right to prevention, treatment and control of diseases; Equal and timely access to basic health services; The provision of health-related education and information; Equal and timely access to basic health services etc.
Human rights are interdependent, indivisible and interrelated. This means that violating the right to health may often impair the enjoyment of other human rights, such as the rights to education or work, and vice versa. Persons with disabilities face various challenges to the enjoy...
Show MoreFebruary 28, 2024
To: Juan VA Franco, MD
Editor-in-Chief
British Medical Journal, Evidence-Based Medicine
Dear Editor:
The study on cutaneous melanoma overdiagnosis attempts to tackle an important issue. However, we wish to address several methodological concerns that may warrant a critical evaluation of its conclusions.
First, one key study assumption is that the overdiagnosis of melanoma is due to over-screening by clinicians, including dermatologists. However, whether the patients were actually screened by clinicians is unknowable with the current study design. Thus, the lack of direct evidence to support this key assumption limits the study's capacity to attribute melanoma diagnoses to the prevalence of screening.
Second, the study's ecological methodology does not sufficiently account for variables that could affect melanoma diagnosis and mortality rate over time, such as advancements in diagnostic technologies, treatments, public awareness, and healthcare access, all factors that dermatologists have worked to improve over time. These factors could independently influence trends in melanoma incidence and mortality. This limitation is critical as it underlines the difficulty in drawing definitive conclusions from the ecological data presented.
Thirdly, the choice to manually input annual data from the SEER program into the DevCan software, deviating from the standard 3-year data aggregatio...
Show MoreDear Editor,
we are writing in response to the article titled " How methodological pitfalls have created widespread misunderstanding about long COVID”(1).
We agree with the authors, that the existing epidemiological research on long COVID has suffered from overly broad case definitions and a striking absence of control groups, which may have led in an overestimation of risk.
Show MoreIt is important to acknowledge that Long- and Post-COVID syndrome are heterogeneous conditions, likely comprising different pathomechanistic groups such as autoimmunity, mitochondrial dysfunction, and virus persistence (2). This complexity, coupled with the lack of routine biomarkers, makes it difficult to accurately define and study this condition. Høeg et al et al. therefore raise some relevant points regarding the challenges faced in studying Long- and Post-COVID syndrome, particularly the need for properly matched control groups and internationally-established diagnostic criteria. Regarding the latter, the authors of the article themselves fail to use definitions accurately, particularly in distinguishing between the now consented WHO definitions of Long-COVID and Post-COVID (WHO/2019-nCoV/Post_COVID-19_condition/Clinical_case_definition/2021.1).
It is true, that initial studies depicted a high prevalence of Post-COVID syndrome (PCS). However, more recent population-based studies present a different perspective. In assessing the clinical picture, the primary focus isn...
Dr Juan Franco
Editor-In-Chief
BMJ Evidence Based Medicine
BMA House
Tavistock Square
London WC1H 9JP
UNITED KINGDOM
31 October 2023
Dear Editor-In-Chief,
We read with interest the recent article by Høeg and colleagues that describes how methodological limitations in long COVID research distort risk and overestimate prevalence.[1]
The authors propose criteria to improve epidemiological research of long COVID. We write in support of these criteria, and to suggest two additions. We recently compared outcomes three months after PCR-confirmed COVID-19 infection with PCR-confirmed influenza infection, and found no difference between these illnesses.[2] Our comparative observational study had limitations (which we acknowledged) but was noteworthy because it was conducted in an Australian population that was primarily exposed to the Omicron variant after achieving high vaccination rates (>90%).
As a result, our two proposed additions to Høeg et al’s criteria relate to the exposed population which, as they suggest, should have diagnostic evidence of infection.
The first addition is to document the COVID variant to which this population was exposed. Recent data from Sweden shows a progressive (and substantial) decrease in the risk of long COVID from the wild type to the Omicron variant.[3] In addition, the type and frequency of symptoms has changed as the virus evolves.[4] This inclusion would improv...
Show MoreWe were surprised that BMJ Evidence Based Medicine chose to publish the flawed article by Høeg and co-authors on methodological limitations of research on long COVID (1). This piece appears to be a ‘Trojan Horse’ article where a scientifically dubious proposition escapes proper scrutiny because it is cloaked in otherwise plausible research commentary.
As the authors state, we need well designed studies to provide a valid measure of the long-term effects of acute COVID-19 infection (Long COVID). Such studies require robust case definitions, adequate duration of follow-up, and suitable comparison groups.
But in a section titled “The most well-designed studies provide reassuring estimates”, the authors include just two studies to support that sweeping statement. This highly selective ‘mini meta-analysis’ subverts the very purpose of evidence-based medicine. The main message of the Høeg paper appears to be that there is a negligible risk of long COVID, based on the selection of papers they have cited. That message does not fit with the actual body of scientific evidence (2). There is now overwhelming research that SARS-CoV-2 infection carries a significant risk of long-term effects over and above the generic effects of post-ICU syndrome and pneumonia (3).
The evidence of long-term effects comes from multiple sources, including epidemiological studies and basic science research looking at the severe and lasting pathological changes that occur in some pati...
Show MorePages