We use cookies to improve our service and to tailor our content and advertising to you.More infoYou can manage your cookie settings via your browser at any time. To learn more about how we use cookies, please see our cookies policy.
I read this paper with interest - I think there is a lot to be gained from studying the James Lind Alliance approach and the diversity of methods that have been applied by James Lind partnerships. I do think however that your search strategy (and approach to searching) was unnecessarily narrow and wonder why you didn't start with the James Lind PSP website (http://www.jla.nihr.ac.uk/priority-setting-partnerships/) which lists all the JLA PSPs undertaken and would have been the best place to start. Had you done this (and/or used Google) you would have identified more JLA PSPs, including ours on pressure ulcers (reported here http://www.jla.nihr.ac.uk/priority-setting-partnerships/pressure-ulcers/ and also [1, 2]. I bring this to your attention merely because we considered extensively some of the issues your raise regarding facilitating broad participation and we have also discussed other methodological aspects.
[1] Cullum N, Buckley H, Dumville J, Hall J, Lamb K, Madden M, Morley R, O'Meara S, Goncalves PS, Soares M, Stubbs N. Wounds research for patient benefit: a 5 year programme of research. NIHR Journals Library; 2016
[2] Madden M, Morley R. Exploring the challenge of health research priority setting in partnership: reflections on the methodology used by the James Lind Alliance Pressure Ulce...
I read this paper with interest - I think there is a lot to be gained from studying the James Lind Alliance approach and the diversity of methods that have been applied by James Lind partnerships. I do think however that your search strategy (and approach to searching) was unnecessarily narrow and wonder why you didn't start with the James Lind PSP website (http://www.jla.nihr.ac.uk/priority-setting-partnerships/) which lists all the JLA PSPs undertaken and would have been the best place to start. Had you done this (and/or used Google) you would have identified more JLA PSPs, including ours on pressure ulcers (reported here http://www.jla.nihr.ac.uk/priority-setting-partnerships/pressure-ulcers/ and also [1, 2]. I bring this to your attention merely because we considered extensively some of the issues your raise regarding facilitating broad participation and we have also discussed other methodological aspects.
[1] Cullum N, Buckley H, Dumville J, Hall J, Lamb K, Madden M, Morley R, O'Meara S, Goncalves PS, Soares M, Stubbs N. Wounds research for patient benefit: a 5 year programme of research. NIHR Journals Library; 2016
[2] Madden M, Morley R. Exploring the challenge of health research priority setting in partnership: reflections on the methodology used by the James Lind Alliance Pressure Ulcer Priority Setting Partnership. Res Involv Engagem 2016; 2:12. doi: 10.1186/s40900-016-0026-y.
Editor,
As corresponding author of the only paper out of 68 that "appeared to have undertaken a reassessment, which led to a correction" I have some sympathy for the authors of the other 67 papers. We decided that it was important that the problem due to the inclusion of subsequently retracted papers needed to be brought to the attention of readers, which we assumed would be relatively rapid. After agreement with my co-authors I emailed to the journal some text to highlight the problem and for readers to ignore the principle findings and rely on a secondary analysis which had excluded the suspect trials. The first email was sent on 18th August 2017. After receiving no response or acknowledgement a second email was sent on 13th November 2018, with a third on the 30th January 2018. A response was received on the 18th February 2018 and a correction published June 2018. Therefore it took the best part of a year to flag up an uncontested correction, which was a simple thing to do. It would be of interest to see if any of the authors of the remaining 67 papers had tried or are still trying to have a correction issues.
David Torgersoni
As readers will be aware, in the early years of the adverse mortality trends it was postulated that this may be due to UK government austerity. Using data at local authority level in England, Wales and Scotland this hypothesis has been questioned [1], and looks to have been a case of (inadvertent) correlation and not causation. This paper [1] also makes reference to the contribution from a series of influenza epidemics which coincidentally occurred in the austerity years.
It has also been pointed out that the use of calendar year data has been acting to conceal complex spatiotemporal patterns affecting deaths which resemble outbreaks of an infectious agent (influenza excluded as the causative agent) [2,3]. For this very reason the Office for National Statistics has begun publishing quarterly mortality data for England [4].
Since age standardized mortality is a single measure of mortality (with inherent assumptions), it is suggested that wider measures including single year of age patterns may need to be investigated.
As pointed out by the Scottish study, further research is indeed required.
1. Jones R. Austerity in the UK and poor health: Were deaths directly affected. British Journal of Healthcare Management 2019; 25(11): in press
2. Jones R. The calendar year fallacy: the danger of reliance on calendar year data in end-of-life capacity and financial planning. International Journal Health Planning Management 2019; doi: 10.1002/hpm.2838...
As readers will be aware, in the early years of the adverse mortality trends it was postulated that this may be due to UK government austerity. Using data at local authority level in England, Wales and Scotland this hypothesis has been questioned [1], and looks to have been a case of (inadvertent) correlation and not causation. This paper [1] also makes reference to the contribution from a series of influenza epidemics which coincidentally occurred in the austerity years.
It has also been pointed out that the use of calendar year data has been acting to conceal complex spatiotemporal patterns affecting deaths which resemble outbreaks of an infectious agent (influenza excluded as the causative agent) [2,3]. For this very reason the Office for National Statistics has begun publishing quarterly mortality data for England [4].
Since age standardized mortality is a single measure of mortality (with inherent assumptions), it is suggested that wider measures including single year of age patterns may need to be investigated.
As pointed out by the Scottish study, further research is indeed required.
1. Jones R. Austerity in the UK and poor health: Were deaths directly affected. British Journal of Healthcare Management 2019; 25(11): in press
2. Jones R. The calendar year fallacy: the danger of reliance on calendar year data in end-of-life capacity and financial planning. International Journal Health Planning Management 2019; doi: 10.1002/hpm.2838
3. Jones R. Unexplained periods of higher deaths contribute to marginal changes in health care demand and health insurance costs: International perspectives. International Journal Health Planning Management 2019; in press.
Allina Health is a Minnesota based not-for-profit healthcare system caring for patients across our 90+ clinics, 12 hospitals, 15 pharmacies, specialty care centers and specialty medical services. This includes our Penny George Institute for Health and Healing, which brings together physical therapy, chiropractic care and acupuncture alongside our traditional healthcare system to provide whole-person-care. As one of the largest healthcare systems in the United States, we are deeply committed to making a positive difference in the opioid crisis by including access to all effective modalities of care.
We found your recently published paper (by Kazis et al [BMJ Open 2019;9e028633] regarding their retrospective study of the association between initial health provider for low back pain with subsequent opioid use) to be a very valuable study with the potential to further shape healthcare’s response to the opioid epidemic. Unfortunately, some of the pertinent concluding evidence has been left out of the abstract - specifically acupuncture effectiveness. We are reaching out to request that you address this in order to maximize the acceptance of all effective treatments in responding to our opioid crisis.
In the body of the study, data on three types of conservative therapists as well as various types of physicians are included. It showed that there was a decrease in both early and long-term opioid use among patients with either physi...
Allina Health is a Minnesota based not-for-profit healthcare system caring for patients across our 90+ clinics, 12 hospitals, 15 pharmacies, specialty care centers and specialty medical services. This includes our Penny George Institute for Health and Healing, which brings together physical therapy, chiropractic care and acupuncture alongside our traditional healthcare system to provide whole-person-care. As one of the largest healthcare systems in the United States, we are deeply committed to making a positive difference in the opioid crisis by including access to all effective modalities of care.
We found your recently published paper (by Kazis et al [BMJ Open 2019;9e028633] regarding their retrospective study of the association between initial health provider for low back pain with subsequent opioid use) to be a very valuable study with the potential to further shape healthcare’s response to the opioid epidemic. Unfortunately, some of the pertinent concluding evidence has been left out of the abstract - specifically acupuncture effectiveness. We are reaching out to request that you address this in order to maximize the acceptance of all effective treatments in responding to our opioid crisis.
In the body of the study, data on three types of conservative therapists as well as various types of physicians are included. It showed that there was a decrease in both early and long-term opioid use among patients with either physical therapists (PTs), chiropractors or acupuncturists. While all three of these conservative therapies and results are discussed in the paper, only chiropractic and physical therapy results are mentioned in the abstract - the acupuncture results are missing.
We are respectfully requesting that the abstract be reviewed and revised to include the acupuncture results. Exclusion of these results may diminish an effective alternative form of therapy we hope to leverage in addressing the opioid crisis.
Thank you for your consideration,
John S. Misa, MD
Vice President, Medical Operations
Allina Health
Attempting to follow it gives the message: "Sorry - we can't find the page you are looking for"
This is the reference which provides the justification for what seems to be the extremely high carbon footprint for each inhaler. Without this, it is difficult to have any faith in the conclusions of this paper.
In general, it is bad practice to provide a citation to a non-permanent web resource.
We read with interest the recent article by Garcia Rodriguez et al. (2019) [1] “Appropriateness of initial dose of non-vitamin K antagonist oral anticoagulants in patients with non-valvular atrial fibrillation in the UK”. We have concerns that there are significant methodological flaws with the study. Consequently, we suggest that the inferences raised in the Discussion with respect to dosing of apixaban are speculative and the conclusions are not based on a considered evaluation of evidence.
Bias resulting from misclassification of renal function
In the study by Rodriguez, the authors state, “For each NOAC cohort, we calculated the percentage of patients appropriately dosed, both overall and according to whether the daily dose of the index NOAC was a standard or reduced dose. For this calculation, patients with missing data on renal function were assumed to have normal renal function.” To assume a ‘missing’ value as ‘normal’ is scientifically unsound, leading to significant misclassification bias which is likely to over-estimate the proportion of patients inappropriately prescribed the lower dose. It is important to note that the level of renal function recording in the THIN and CPRD databases is sub-optimal [3, 4]. We suggest that it would have been methodologically apposite to define the proportion of patients where renal function was unknown and perform a ‘known parameters’ analysis to validate the assumption and extrapolation to the total population....
We read with interest the recent article by Garcia Rodriguez et al. (2019) [1] “Appropriateness of initial dose of non-vitamin K antagonist oral anticoagulants in patients with non-valvular atrial fibrillation in the UK”. We have concerns that there are significant methodological flaws with the study. Consequently, we suggest that the inferences raised in the Discussion with respect to dosing of apixaban are speculative and the conclusions are not based on a considered evaluation of evidence.
Bias resulting from misclassification of renal function
In the study by Rodriguez, the authors state, “For each NOAC cohort, we calculated the percentage of patients appropriately dosed, both overall and according to whether the daily dose of the index NOAC was a standard or reduced dose. For this calculation, patients with missing data on renal function were assumed to have normal renal function.” To assume a ‘missing’ value as ‘normal’ is scientifically unsound, leading to significant misclassification bias which is likely to over-estimate the proportion of patients inappropriately prescribed the lower dose. It is important to note that the level of renal function recording in the THIN and CPRD databases is sub-optimal [3, 4]. We suggest that it would have been methodologically apposite to define the proportion of patients where renal function was unknown and perform a ‘known parameters’ analysis to validate the assumption and extrapolation to the total population.
The authors neither discuss the existence of this misclassification bias, nor make any attempt to quantify its impact via sensitivity analysis.
Key dose reduction variable (bodyweight) missing from results
Explicit in the abstract is the mention of ‘bodyweight’ as a key variable of interest, yet, surprisingly, this is not disclosed in the Results section of the manuscript. Body weight too is often missing from electronic health records [5] and cannot be assumed to be normal in elderly frail patients or patients with type II diabetes. Certain patient characteristics (including older age and renal disease) are associated with an increased risk of stroke and major bleeding, and specific dose-reduction criteria for apixaban are based on a patient’s age, bodyweight, and serum creatinine [2]. Furthermore, without bodyweight data, it is not possible to assess renal function using the standard Cockcroft-Gault method for estimation of creatinine clearance and drug dose adjustment for both rivaroxaban and dabigatran.
Finally, Rodriguez and colleagues suggest that there is under-dosing in a significant proportion of those treated with apixaban. A sub-group analysis of the ARISTOTLE study previously demonstrated that 2.5 mg is appropriate in certain NVAF patients in reducing stroke [2]. The inappropriate dosing of patients with any NOAC is of concern but the analysis in this study does not permit these inferences to be made.
References
1. García Rodríguez LA, Martín-Pérez M, Vora P, et al. Appropriateness of initial dose of non-vitamin K antagonist oral anticoagulants in patients with non-valvular atrial fibrillation in the UK. BMJ Open 2019;9: e031341. doi:10.1136/bmjopen-2019-031341
2. Granger CB, Alexander JH, McMurray JJ, Lopes RD, Hylek EM, Hanna M, et al. Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med. 2011 Sep 15;365(11):981-92. doi: 10.1056/NEJMoa1107039. Epub 2011 Aug 27.
3. McDonald HI, Shaw C, Thomas SL, Mansfield KE, Tomlinson LA, Nitsch D. Methodological challenges when carrying out research on CKD and AKI using routine electronic health records. Kidney Int. 2016; 90(5); 943-949.
4. Ramagopalan SV, Stamp E, Sammon CJ, Besford M, Carroll R, Mehmud F, Alikhan R. Renal function and oral anticoagulant treatment of incident non-valvular atrial fibrillation: a retrospective study. Future Cardiol. 2019 Jul;15(4):301-309. doi: 10.2217/fca-2019-0012. Epub 2019 May 29.
5. Petersen I, Welch CA, Nazareth I, Walters K, Marston L, Morris RW, Carpenter JR, Morris TP, Pham TM. Health indicator recording in UK primary care electronic health records: key implications for handling missing data. Clin Epidemiol. 2019 Feb 11; 11:157-167. doi: 10.2147/CLEP.S191437. eCollection 2019.
The study by Jensen et al. aimed “to examine the degree of residual confounding” in the association between childhood vaccinations and subsequent morbidity and mortality [1]. Jensen et al. conclude that “our study also suggests that residual confounding may have been present in previous Danish register-based studies investigating the impact of MMR [Measles-mumps-rubella vaccine] on the morbidity outcomes respiratory syncytial virus and hospitalisation for infection until 2 years of age” [1].
Our previous Danish register-based studies referred to in this conclusion examined the effect of having MMR (recommended at 15 months of age) compared with the 3rd dose of diphtheria– tetanus–pertussis–polio–H. influenzae type b (DTP, recommended at 3,5 and 12 months of age) as the most recent vaccine on subsequent infectious disease hospitalisations (adjusted incidence rate ratios [IRR], 0.86; 95% confidence interval [CI], 0.84-0.88) [2] and respiratory syncytial virus infections (adjusted IRR, 0.78; 95% CI, 0.66-0.93)[3].
There are fundamental methodological differences between our previous studies and the study by Jensen et al. Therefore, the results from Jensen et al. cannot be generalized to these previous studies.
First, to enhance exchangeability, we restricted the study population to children, who had received the first two DTP doses before 11 months of age [2, 3], to exclude children who did not receive vaccines in a timely fashion, which could be relate...
The study by Jensen et al. aimed “to examine the degree of residual confounding” in the association between childhood vaccinations and subsequent morbidity and mortality [1]. Jensen et al. conclude that “our study also suggests that residual confounding may have been present in previous Danish register-based studies investigating the impact of MMR [Measles-mumps-rubella vaccine] on the morbidity outcomes respiratory syncytial virus and hospitalisation for infection until 2 years of age” [1].
Our previous Danish register-based studies referred to in this conclusion examined the effect of having MMR (recommended at 15 months of age) compared with the 3rd dose of diphtheria– tetanus–pertussis–polio–H. influenzae type b (DTP, recommended at 3,5 and 12 months of age) as the most recent vaccine on subsequent infectious disease hospitalisations (adjusted incidence rate ratios [IRR], 0.86; 95% confidence interval [CI], 0.84-0.88) [2] and respiratory syncytial virus infections (adjusted IRR, 0.78; 95% CI, 0.66-0.93)[3].
There are fundamental methodological differences between our previous studies and the study by Jensen et al. Therefore, the results from Jensen et al. cannot be generalized to these previous studies.
First, to enhance exchangeability, we restricted the study population to children, who had received the first two DTP doses before 11 months of age [2, 3], to exclude children who did not receive vaccines in a timely fashion, which could be related to unmeasured health, social, and vaccine attitude confounders.
Jensen et al. in contrast included children with major deviations from the recommended vaccination schedule, e.g. children who had received zero or very few vaccines by age 16 months[1]. This increases the likelihood of unmeasured confounding.
Second, we investigated the association between MMR and different types of infections and found that the effect was most pronounced for respiratory tract infections[2], an observation which have been supported by studies from the US and The Netherlands [4, 5]. It is hard to imagine an unmeasured confounder that would impact the association between MMR and different types of infections differently and due to the consistency in the pattern across countries this could indicate that not all of the association can be explained by unmeasured confounding.
The study by Jensen et al. did not consider type of infection; important insights might be missed by grouping all infections together in a combined outcome.
Third, while our main cohort was children, who had received 3 doses of DTP, the study also included a small cohort of children who had received 2 doses of DTP followed by MMR. In that cohort, getting the 3rd DTP was associated with an increased rate of infectious disease hospitalisations compared with MMR as the most recent vaccine (adjusted IRR, 1.62; 95% CI, 1.28-2.05)[2]. If there were unmeasured confounders related to getting an additional vaccine, then getting the 3rd DTP should have been associated with a reduced rate of infectious disease hospitalisations in this cohort. This was not the case. Similar patterns were derived from The Netherlands [6]. This informative analysis was not undertaken by Jensen et al; in fact Jensen et al did not consider the sequence of vaccines, which has been shown to be a major determinant of their overall health effects[7].
Fourth, Jensen et al. emphasize their use of the negative control outcome “hospitalisations for accidents” to assess the risk of residual confounding. The purpose of a negative control is “to reproduce a condition that cannot involve the hypothesized causal mechanism but is very likely to involve the same sources of bias that may have been present in the original association.”[8]. Jensen et al find an association between vaccination status and the control outcome and use that finding to infer that other studies also have residual confounding: “Since the negative control outcome, as already mentioned, points toward residual confounding, our study also suggests that residual confounding may have been present in previous Danish register-based studies”. Jensen et al seemingly overlooked that we also had a negative control outcome: MMR as the most recent vaccine was associated with an adjusted IRR of 0.97 (95% CI, 0.95-0.99) [2] and 1.02 (95% CI, 0.98-1.06) [3] for “emergency room visits with accidents”. In comparison, Jensen et al. found that 3DTP+MMR was associated with an adjusted HR of 0.83 (95% CI, 0.80-0.85) for “hospitalisations due to accidents” [1]. These divergent results probably reflect that there was more residual confounding in Jensen et al.’s study than in our studies. This is not surprising, given that Jensen et al. do not include the same methodological considerations to enhance exchangeability. But it can obviously not be used to infer that our studies had the same residual confounding.
Fifth, Jensen et al. emphasize their findings of an association between childhood vaccines and childhood mortality. However, it is unlikely that vaccination status and childhood mortality should be linked among Danish children and therefore we did not examine this association. Non-specific effects of vaccines are related to an impact on the children’s immune system[9]. This will be measurable as “mortality” in areas where deaths are caused by infections, such as in Africa, but not in Denmark, where very few deaths are due to infections. Based on the Danish death registry from 2012-2014, only 6 (5.6 %) of 107 deaths among children aged 1-4 years of age were caused by an infection. The predominant causes of death were accidents, death not caused by medical conditions, conditions in the nervous system, cancer, and congenital malformations (71.9 %)[10], conditions where no vaccine effects, specific or non-specific, would be anticipated.
In spite of the disparate methodologies used in these Danish studies, Jensen at al. make several bold and unfounded conclusions. First, they suggest as a major conclusion in the abstract that their study “did not support the hypothesis that DTP is detrimental, since the group of children with fewer DTP vaccinations experienced increased mortality”. Given the high risk of confounding by indication and the doubtful link between vaccination status and overall mortality in the Danish setting this is an unwarranted deduction. Second, Jensen et al. stress that observational studies on the association between MMR and childhood mortality are likely hampered by residual confounding and “therefore non-specific effects of vaccines are best examined in randomised trials”[1]. Randomised controlled trials of non-specific effects have previously been requested in the WHO’s review of non-specific effects of vaccines [7]. However, randomised controlled trials of already recommended vaccines are hampered by ethical constraints, which limits the possibility to investigate non-specific effects. With reference to the concepts of causal inference [11], observational studies incorporating innovative methods to enhance exchangeability between exposure groups and control for confounders may be the only way to investigate health effects of an already implemented health policy such as vaccination programmes. Triangulation of results from different settings and different methodologies can further strengthen the evidence for causal inference[11] and elucidate non-specific effects of vaccines [12].
References:
1. Jensen, A., P.K. Andersen, and L.G. Stensballe, Early childhood vaccination and subsequent mortality or morbidity: are observational studies hampered by residual confounding? A Danish register-based cohort study. BMJ Open, 2019. 9(9): p. e029794.
2. Sorup, S., et al., Live vaccine against measles, mumps, and rubella and the risk of hospital admissions for nontargeted infections. Jama, 2014. 311(8): p. 826-35.
3. Sorup, S., et al., Measles-mumps-rubella vaccination and respiratory syncytial virus-associated hospital contact. Vaccine, 2015. 33(1): p. 237-45.
4. Bardenheier, B.H., et al., Risk of non-targeted infectious disease hospitalizations among U.S. children following inactivated and live vaccines, 2005-2014. Clin Infect Dis, 2017.
5. Tielemans, S., et al., Non-specific effects of measles, mumps, and rubella (MMR) vaccination in high income setting: population based cohort study in the Netherlands. Bmj, 2017. 358: p. j3862.
6. Benn, C.S., S. Sørup, and P. Aaby, Re: Non-specific effects of measles, mumps, and rubella (MMR) vaccination in high income setting: population based cohort study in the Netherlands. BMJ, 2017. 358.
7. Higgins, J.P.T., et al., Association of BCG, DTP, and measles containing vaccines with childhood mortality: systematic review. BMJ, 2016. 355.
8. Lipsitch, M., E. Tchetgen Tchetgen, and T. Cohen, Negative controls: a tool for detecting confounding and bias in observational studies. Epidemiology, 2010. 21(3): p. 383-8.
9. de Bree, L.C.J., et al., Non-specific effects of vaccines: Current evidence and potential implications. Semin Immunol, 2018. 39: p. 35-43.
10. Jarlbæk, L., Dødssted og dødsårsager for børn og unge under 19 år i Danmark, 2012-2014. 2017.
11. Hernan, M.A. and J.M. Robins, Causal inference: What If. 14 October 2019, Forthcoming ed. 2020: Boca Raton: Chapman & Hall/CRC.
12. Benn, C.S., et al., How to evaluate potential non-specific effects of vaccines: the quest for randomized trials or time for triangulation? Expert Rev Vaccines, 2018. 17(5): p. 411-420.
As nearness to death is an important contributor to hospital utilization, especially in the last six months of life, is it possible for the data to be reanalysed in relation to nearness to death in this cohort?
We read with interest the recent paper by Kazis et al [BMJ Open 2019;9e028633], an observational retrospective study of the association between initial health provider for low back pain with subsequent opioid use.
This included data on three types of conservative therapists as well as various types of physicians. Initial treatment from any of the former, namely physical therapists (PTs), chiropractors and acupuncturists, was associated with substantially decreased odds for both early and long-term opioid use. However, these results, though tabulated, described and discussed in the main text of the paper do not transfer in the same form to the abstract. Specifically, the acupuncture results have been removed.
The benefits of both PT and chiropractic are stated in two places in the Results section and once in the Conclusions, but there is no mention of acupuncture. Although we understand that the sample of acupuncturists was relatively small, acupuncture reduced short and long term exposure to opioids by 75% to 90% compared to the same patient starting with a PCP, and the confidence interval was significant and similar to PT.
Given that many people will look initially or only at the abstract, this omission misleads the reader into inferring that acupuncture did not have the same benefits as the other two conservative treatments. The implications for health care policy/access/delivery are of concern. Does a...
We read with interest the recent paper by Kazis et al [BMJ Open 2019;9e028633], an observational retrospective study of the association between initial health provider for low back pain with subsequent opioid use.
This included data on three types of conservative therapists as well as various types of physicians. Initial treatment from any of the former, namely physical therapists (PTs), chiropractors and acupuncturists, was associated with substantially decreased odds for both early and long-term opioid use. However, these results, though tabulated, described and discussed in the main text of the paper do not transfer in the same form to the abstract. Specifically, the acupuncture results have been removed.
The benefits of both PT and chiropractic are stated in two places in the Results section and once in the Conclusions, but there is no mention of acupuncture. Although we understand that the sample of acupuncturists was relatively small, acupuncture reduced short and long term exposure to opioids by 75% to 90% compared to the same patient starting with a PCP, and the confidence interval was significant and similar to PT.
Given that many people will look initially or only at the abstract, this omission misleads the reader into inferring that acupuncture did not have the same benefits as the other two conservative treatments. The implications for health care policy/access/delivery are of concern. Does access to “conservative therapies” change one’s likelihood of using opioids? The answer is “yes.” Many patients don't want opioids or other analgesics, but they are not given reasonable access to “conservative therapies” due to issues of availability (we don't offer acupuncture, here) or cost (acupuncture is not covered by insurance so you must pay out of pocket and cannot afford it).
We kindly ask that this omission be corrected.
Richard Harris, PhD, LAc. Society for Acupuncture Research
Mark Bovey, MSc. British Acupuncture Council
David Mills, MD, LAc American Society of Acupuncturists
Freda Dreher, MD American Academy of Medical Acupuncture
We read with interest this paper which asserts that recent increases in infant mortality have affected poorer areas of England disproportionately. In our analyses of largely post-neonatal unexplained infant deaths we too have found a proportional increase of poorer families; markers of deprivation including maternal age, education, parity and smoking status will all impact on infant mortality. (1-3) However, it is perhaps worth pointing out that this recent increase in infant mortality in England & Wales from 2014 to 2017 is limited to neonatal deaths, the rate of post-neonatal deaths (4 weeks to 1 year old) has flat-lined in these 4 years from 1.09 deaths per 1000 live births in 2014 to 1.08 deaths per 1000 live births in 2017. Furthermore, in the data release by the Office for National Statistics (ONS) for 2017, (4) they report a 23.5% reduction in infant mortality in the most deprived areas over the last 10 years, compared with a 10.0% reduction in the most affluent areas. Using the data provided in that release on age at death we can see that the increase in overall infant mortality is driven by very early neonatal deaths of infants under 1 day old, from 1.3 per 1000 live births in 2014 to 1.5 in 2015, and 1.6 in 2016 and 2017. Mortality in all other age groups within infancy from one day to one year show reductions over that time. (4) The major contributory causes to early neonatal mortality are prematurity, congenital anomalies and infections, with a small but...
We read with interest this paper which asserts that recent increases in infant mortality have affected poorer areas of England disproportionately. In our analyses of largely post-neonatal unexplained infant deaths we too have found a proportional increase of poorer families; markers of deprivation including maternal age, education, parity and smoking status will all impact on infant mortality. (1-3) However, it is perhaps worth pointing out that this recent increase in infant mortality in England & Wales from 2014 to 2017 is limited to neonatal deaths, the rate of post-neonatal deaths (4 weeks to 1 year old) has flat-lined in these 4 years from 1.09 deaths per 1000 live births in 2014 to 1.08 deaths per 1000 live births in 2017. Furthermore, in the data release by the Office for National Statistics (ONS) for 2017, (4) they report a 23.5% reduction in infant mortality in the most deprived areas over the last 10 years, compared with a 10.0% reduction in the most affluent areas. Using the data provided in that release on age at death we can see that the increase in overall infant mortality is driven by very early neonatal deaths of infants under 1 day old, from 1.3 per 1000 live births in 2014 to 1.5 in 2015, and 1.6 in 2016 and 2017. Mortality in all other age groups within infancy from one day to one year show reductions over that time. (4) The major contributory causes to early neonatal mortality are prematurity, congenital anomalies and infections, with a small but significant number of deaths relating to intrapartum events, especially in babies of low birth weight or short gestation. (5) We therefore feel that any analysis looking at trends over time should be applied to early neonatal deaths, as this is clearly the age range in which something has changed. Efforts to understand the very real inequalities in infant health outcomes for families experiencing poverty are vital if we are to identify how best to overcome them, but we must do so with accuracy and a determination to engage with the complexity within which they operate. It is likely that strategies to reduce early neonatal mortality – which will need to target services directed to pregnant women – will be different to those aimed at reducing later infant mortality.
References
1. Blair PS, Sidebotham P, Berry PJ, Evans M, Fleming PJ. Major epidemiological changes in sudden infant death syndrome: a 20-year population-based study in the UK. Lancet (London, England). 2006;367(9507):314-9.
2. Blair PS, Sidebotham P, Evason-Coombe C, Edmonds M, Heckstall-Smith EM, Fleming P. Hazardous cosleeping environments and risk factors amenable to change: case-control study of SIDS in south west England. BMJ (Clinical research ed). 2009;339:b3666.
3. Fleming PJ, Blair PS, Ward Platt M, Tripp J, Smith IJ. Sudden infant death syndrome and social deprivation: assessing epidemiological factors after post-matching for deprivation. Paediatric and perinatal epidemiology. 2003;17(3):272-80.
4. Office for National Statistics. Child and infant mortality in England and Wales: 2017. 2019.
5. Lehtonen L, Gimeno A, Parra-Llorca A, Vento M. Early neonatal death: A challenge worldwide. Seminars in fetal & neonatal medicine. 2017;22(3):153-60.
I read this paper with interest - I think there is a lot to be gained from studying the James Lind Alliance approach and the diversity of methods that have been applied by James Lind partnerships. I do think however that your search strategy (and approach to searching) was unnecessarily narrow and wonder why you didn't start with the James Lind PSP website (http://www.jla.nihr.ac.uk/priority-setting-partnerships/) which lists all the JLA PSPs undertaken and would have been the best place to start. Had you done this (and/or used Google) you would have identified more JLA PSPs, including ours on pressure ulcers (reported here http://www.jla.nihr.ac.uk/priority-setting-partnerships/pressure-ulcers/ and also [1, 2]. I bring this to your attention merely because we considered extensively some of the issues your raise regarding facilitating broad participation and we have also discussed other methodological aspects.
[1] Cullum N, Buckley H, Dumville J, Hall J, Lamb K, Madden M, Morley R, O'Meara S, Goncalves PS, Soares M, Stubbs N. Wounds research for patient benefit: a 5 year programme of research. NIHR Journals Library; 2016
Show More[2] Madden M, Morley R. Exploring the challenge of health research priority setting in partnership: reflections on the methodology used by the James Lind Alliance Pressure Ulce...
Editor,
As corresponding author of the only paper out of 68 that "appeared to have undertaken a reassessment, which led to a correction" I have some sympathy for the authors of the other 67 papers. We decided that it was important that the problem due to the inclusion of subsequently retracted papers needed to be brought to the attention of readers, which we assumed would be relatively rapid. After agreement with my co-authors I emailed to the journal some text to highlight the problem and for readers to ignore the principle findings and rely on a secondary analysis which had excluded the suspect trials. The first email was sent on 18th August 2017. After receiving no response or acknowledgement a second email was sent on 13th November 2018, with a third on the 30th January 2018. A response was received on the 18th February 2018 and a correction published June 2018. Therefore it took the best part of a year to flag up an uncontested correction, which was a simple thing to do. It would be of interest to see if any of the authors of the remaining 67 papers had tried or are still trying to have a correction issues.
David Torgersoni
As readers will be aware, in the early years of the adverse mortality trends it was postulated that this may be due to UK government austerity. Using data at local authority level in England, Wales and Scotland this hypothesis has been questioned [1], and looks to have been a case of (inadvertent) correlation and not causation. This paper [1] also makes reference to the contribution from a series of influenza epidemics which coincidentally occurred in the austerity years.
It has also been pointed out that the use of calendar year data has been acting to conceal complex spatiotemporal patterns affecting deaths which resemble outbreaks of an infectious agent (influenza excluded as the causative agent) [2,3]. For this very reason the Office for National Statistics has begun publishing quarterly mortality data for England [4].
Since age standardized mortality is a single measure of mortality (with inherent assumptions), it is suggested that wider measures including single year of age patterns may need to be investigated.
As pointed out by the Scottish study, further research is indeed required.
1. Jones R. Austerity in the UK and poor health: Were deaths directly affected. British Journal of Healthcare Management 2019; 25(11): in press
2. Jones R. The calendar year fallacy: the danger of reliance on calendar year data in end-of-life capacity and financial planning. International Journal Health Planning Management 2019; doi: 10.1002/hpm.2838...
Show MoreTo the editors of the BMJ:
Allina Health is a Minnesota based not-for-profit healthcare system caring for patients across our 90+ clinics, 12 hospitals, 15 pharmacies, specialty care centers and specialty medical services. This includes our Penny George Institute for Health and Healing, which brings together physical therapy, chiropractic care and acupuncture alongside our traditional healthcare system to provide whole-person-care. As one of the largest healthcare systems in the United States, we are deeply committed to making a positive difference in the opioid crisis by including access to all effective modalities of care.
We found your recently published paper (by Kazis et al [BMJ Open 2019;9e028633] regarding their retrospective study of the association between initial health provider for low back pain with subsequent opioid use) to be a very valuable study with the potential to further shape healthcare’s response to the opioid epidemic. Unfortunately, some of the pertinent concluding evidence has been left out of the abstract - specifically acupuncture effectiveness. We are reaching out to request that you address this in order to maximize the acceptance of all effective treatments in responding to our opioid crisis.
In the body of the study, data on three types of conservative therapists as well as various types of physicians are included. It showed that there was a decrease in both early and long-term opioid use among patients with either physi...
Show MoreProbably the most important reference cannot be reached, number 23:
https://www.greatermanchester-ca.gov.uk/download/meetings/id/2423/enviro...
Attempting to follow it gives the message: "Sorry - we can't find the page you are looking for"
This is the reference which provides the justification for what seems to be the extremely high carbon footprint for each inhaler. Without this, it is difficult to have any faith in the conclusions of this paper.
In general, it is bad practice to provide a citation to a non-permanent web resource.
We read with interest the recent article by Garcia Rodriguez et al. (2019) [1] “Appropriateness of initial dose of non-vitamin K antagonist oral anticoagulants in patients with non-valvular atrial fibrillation in the UK”. We have concerns that there are significant methodological flaws with the study. Consequently, we suggest that the inferences raised in the Discussion with respect to dosing of apixaban are speculative and the conclusions are not based on a considered evaluation of evidence.
Bias resulting from misclassification of renal function
Show MoreIn the study by Rodriguez, the authors state, “For each NOAC cohort, we calculated the percentage of patients appropriately dosed, both overall and according to whether the daily dose of the index NOAC was a standard or reduced dose. For this calculation, patients with missing data on renal function were assumed to have normal renal function.” To assume a ‘missing’ value as ‘normal’ is scientifically unsound, leading to significant misclassification bias which is likely to over-estimate the proportion of patients inappropriately prescribed the lower dose. It is important to note that the level of renal function recording in the THIN and CPRD databases is sub-optimal [3, 4]. We suggest that it would have been methodologically apposite to define the proportion of patients where renal function was unknown and perform a ‘known parameters’ analysis to validate the assumption and extrapolation to the total population....
The study by Jensen et al. aimed “to examine the degree of residual confounding” in the association between childhood vaccinations and subsequent morbidity and mortality [1]. Jensen et al. conclude that “our study also suggests that residual confounding may have been present in previous Danish register-based studies investigating the impact of MMR [Measles-mumps-rubella vaccine] on the morbidity outcomes respiratory syncytial virus and hospitalisation for infection until 2 years of age” [1].
Our previous Danish register-based studies referred to in this conclusion examined the effect of having MMR (recommended at 15 months of age) compared with the 3rd dose of diphtheria– tetanus–pertussis–polio–H. influenzae type b (DTP, recommended at 3,5 and 12 months of age) as the most recent vaccine on subsequent infectious disease hospitalisations (adjusted incidence rate ratios [IRR], 0.86; 95% confidence interval [CI], 0.84-0.88) [2] and respiratory syncytial virus infections (adjusted IRR, 0.78; 95% CI, 0.66-0.93)[3].
There are fundamental methodological differences between our previous studies and the study by Jensen et al. Therefore, the results from Jensen et al. cannot be generalized to these previous studies.
First, to enhance exchangeability, we restricted the study population to children, who had received the first two DTP doses before 11 months of age [2, 3], to exclude children who did not receive vaccines in a timely fashion, which could be relate...
Show MoreAs nearness to death is an important contributor to hospital utilization, especially in the last six months of life, is it possible for the data to be reanalysed in relation to nearness to death in this cohort?
Dear BMJ Open Editor Adrian Aldcroft,
We read with interest the recent paper by Kazis et al [BMJ Open 2019;9e028633], an observational retrospective study of the association between initial health provider for low back pain with subsequent opioid use.
This included data on three types of conservative therapists as well as various types of physicians. Initial treatment from any of the former, namely physical therapists (PTs), chiropractors and acupuncturists, was associated with substantially decreased odds for both early and long-term opioid use. However, these results, though tabulated, described and discussed in the main text of the paper do not transfer in the same form to the abstract. Specifically, the acupuncture results have been removed.
The benefits of both PT and chiropractic are stated in two places in the Results section and once in the Conclusions, but there is no mention of acupuncture. Although we understand that the sample of acupuncturists was relatively small, acupuncture reduced short and long term exposure to opioids by 75% to 90% compared to the same patient starting with a PCP, and the confidence interval was significant and similar to PT.
Given that many people will look initially or only at the abstract, this omission misleads the reader into inferring that acupuncture did not have the same benefits as the other two conservative treatments. The implications for health care policy/access/delivery are of concern. Does a...
Show MoreWe read with interest this paper which asserts that recent increases in infant mortality have affected poorer areas of England disproportionately. In our analyses of largely post-neonatal unexplained infant deaths we too have found a proportional increase of poorer families; markers of deprivation including maternal age, education, parity and smoking status will all impact on infant mortality. (1-3) However, it is perhaps worth pointing out that this recent increase in infant mortality in England & Wales from 2014 to 2017 is limited to neonatal deaths, the rate of post-neonatal deaths (4 weeks to 1 year old) has flat-lined in these 4 years from 1.09 deaths per 1000 live births in 2014 to 1.08 deaths per 1000 live births in 2017. Furthermore, in the data release by the Office for National Statistics (ONS) for 2017, (4) they report a 23.5% reduction in infant mortality in the most deprived areas over the last 10 years, compared with a 10.0% reduction in the most affluent areas. Using the data provided in that release on age at death we can see that the increase in overall infant mortality is driven by very early neonatal deaths of infants under 1 day old, from 1.3 per 1000 live births in 2014 to 1.5 in 2015, and 1.6 in 2016 and 2017. Mortality in all other age groups within infancy from one day to one year show reductions over that time. (4) The major contributory causes to early neonatal mortality are prematurity, congenital anomalies and infections, with a small but...
Show MorePages