Recent eLetters

Displaying 1-9 letters out of 287 published

  1. Correction for the reference 28

    The 28th reference cited in this article should be corrected as:

    28. Xie Y, Ho SC. Prenotification had no additional effect on the response rate and survey quality: a randomized trial. J Clin Epidemiol 2013;66:1422-6.

    Conflict of Interest:

    None declared

    Read all letters published for this article

    Submit response
  2. Stillbirth data seem incorrect

    "The registry is considered complete through the period" Hedegaard et al state in their registry study (1) without a supporting reference. However, there seem to be an issue with the quality of the data as the number of stillbirths deviate from the published official statistics. The authors in their online supplement report 539 stillbirths in 2009-10 decreasing to 489 in 2011-12. But according to the published official statistics the number of stillbirths were 260+255=515 in 2009-10 (2) decreasing to 274+236=510 in 2011-12 (3). Thus the seemingly dramatic decrease between the two time periods is reduced from 50 (=539-489) in the paper to 5 (=515-510) according to the official statistics. Even though we all hope for miracles, all too often they evaporate into thin air on closer inspection. I cannot see any other solutions to this conundrum than either the authors thoroughly document the validity of their data or they retract their paper.

    1) Hedegaard M, Lidegaard O, Skovlund CW, Morch LS, Hedegaard M. Reduction in stillbirths at term after new birth induction paradigm: results of a national intervention. BMJ Open. 2014 Aug 14;4(8)

    2) Sundhedsstyrelsen. Fodselsstatistikken 2011. Copenhagen, Sundhedsstyrelsen, 2012.

    3) Statens Serum Institut. Fodselsstatistikken - tal og analyser, 2012. Available at %20dansk/Sundhedsdata%20og%20it/NSF/Registre/Fodselsregisteret/f%C3%B8dselsstatistikken2012_vers%204.ashx. Accessed September 8th.

    Conflict of Interest:

    None declared

    Read all letters published for this article

    Submit response
  3. Response: Assessing the effect of an interactive decision-aid smartphone smoking cessation application (app) on quit rates: a double-blind automated randomised control trial protocol

    Author have come out with an interesting concept of using smart phone app in medical research. Smart phones are rapidly becoming omnipresent with greater processing power and memory and now are commonly used as first device to access information from or off internet. Thousands of medical apps are available for layman to get information regarding diseases. Also, several apps are available to physicians to quickly asses information from guidelines, algorithms and medical formulas making pocket books redundant. Increasingly new app are being developed to provide diagnostics at nominal cost.

    I think this is the first serious attempt to use phone app in medical research, Authors have meticulously prepared protocol and though some provision like locking the device to one particular group may have got frown from some ethical committee members but I think that was essential to have valid randomization.

    In my personal view this is well planned study on a very important socio-medical problem. I shall be very eagerly waiting for the study results and how this concept per se is taken by medical fraternity.

    Dr. Sandeep Tak

    Conflict of Interest:

    None declared

    Read all letters published for this article

    Submit response
  4. Does CAM training for GPs really reduce healthcare costs?

    I read with interest the recent contribution from Baars and Kooreman [1]. The present study is a replication of a previous study, using a different data set, as described by the authors in their introduction. The 2012 study by Kooreman and Baars [2] garnered criticism from me and others [3-5]. Furthermore, the new study was also presented in an earlier form [6], and this also faced criticism [7,8]. Baars and Kooreman provided response to both of these at the time [9,10].

    The authors have heeded some, if not all, of the criticisms directed at their previous work. In particular, the authors' reporting of their results is somewhat more measured and the limitations of the analysis are more thoroughly presented. It is likely that selection into the two groups at least partially (perhaps fully) explains the observed differences in costs, and the authors are right to identify this as a limitation. The new analysis of mortality data is a welcome improvement.

    In addition to the limitations already outlined by critics of the previous publication - and by the authors themselves - it is unfortunate that the new analysis did not account for the hierarchical structure of the data, with patients clustered by GP practice. It appears that the authors were able to identify this source of clustering in their data, so it is disappointing that a multilevel model was not used to account for this potential source of bias (which might also explain part of the observed difference). It is also regrettable that the results of the log-linear regression of costs appear to have been misreported in Table 4 (Appendix 2).

    Baars and Kooreman's statement that they intend to carry out further analyses on this or similar data with more appropriate estimation strategies is encouraging. Such analyses are crucial due to the limitations of this study. The implementation of better estimation strategies, such as propensity score matching or (as the authors suggest) use of instrumental variables, will be an important part of this. I look forward to reading these potentially enlightening studies once completed.


    [1] Baars EW, Kooreman P. A 6-year comparative economic evaluation of healthcare costs and mortality rates of Dutch patients from conventional and CAM GPs. BMJ Open 2014;4:e005332. doi: 10.1136/bmjopen-2014-005332

    [2] Kooreman P, Baars EW Patients whose GP knows complementary medicine tend to have lower costs and live longer. European Journal of Health Economics 2012;13(6):769-76. doi: 10.1007/s10198-011-0330-2

    [3] Sampson CJ, Whitehurst DGT, Street A. Do patients registered with CAM-trained GPs really use fewer health care resources and live longer? A response to Kooreman and Baars. Eur J Health Econ (2012). 13:769-776. European Journal of Health Economics 2013;14(4):703-5. doi: 10.1007/s10198-013-0466-3

    [4] van Erp P. Alternatieve huisartsen werken 15 procent goedkoper? Een verzinsel!. [Blog] Kloptdatwel? 2013. Available at: [Accessed 3 Sep. 2014].

    [5] Sampson CJ, Whitehurst DGT, Street A. Bad science in health economics: complementary medicine, costs and mortality. [Blog] The Academic Health Economists' Blog 2013. Available at: [Accessed 3 Sep. 2014].

    [6] Kooreman P, Baars EW. Complementair werkende huisartsen en de kosten van zorg. Economisch Statistische Berichten 2014;99(4678):90-2.

    [7] Pomp, M. Reactie op: Complementair werkende huisartsen en de kosten van zorg. Economisch Statistische Berichten 2014;99(4679):125.

    [8] van Erp P. Verzekerde zorgkosten van pati?nten bij alternatieve huisarts. [Blog] Kloptdatwel? 2014. Available at: [Accessed 3 Sep. 2014].

    [9] Kooreman P, Baars EW. Do patients registered with CAM-trained GPs really use fewer health care resources and live longer? A reply to Christopher James Sampson. European Journal of Health Economics 2013;14(4):707-8. doi: 10.1007/s10198-013-0475-2

    [10] Kooreman P, Baars EW. Reactie op: Complementair werkende huisartsen en de kosten van zorg (Naschrift). Economisch Statistische Berichten 2014;99(4679):125.

    Conflict of Interest:

    None declared

    Read all letters published for this article

    Submit response
  5. Re: Herbal medicine (Gan Mai Da Zao decoction) for depression: a systematic review protocol

    This is a very interesting research proposal to evaluate a formula originated from Chinese traditional medicine, Gan Mai Da Zao (GMDZ) decoction. The protocol for a systematic review is generally well written; however, there are several points that need to be addressed from the standpoint of Japanese traditional Kampo medicine, which incorporated the Gan Mai Da Zao decoction as kambakutaisoto.

    Kampo medicine originated from ancient Chinese traditional medicine, but developed distinctively [1]. Kampo medicine was the orthodox medicine in Japan until the 19th century when modern Western medicine took over. Nevertheless, some Kampo formulae are still officially registered in the Japanese Pharmacopoeia [2]. Since 1967, Kampo formulae have been covered by National Health Insurance, and currently 148 Kampo formulae are approved for ethical use [3]. Although Kampo extracts are crude drugs derived from plants, animals, and minerals, their quality is strictly controlled in accordance with the Japanese Pharmacopoeia by quantitative analysis of marker components using high-performance liquid chromatography.

    Unlike China and Korea, traditional medicine in Japan is used in a Western-style medical system, and is prescribed by medical doctors, educated in Western medicine [3], who are required to have a basic knowledge of Kampo formulae. Reflecting the needs for establishing evidence-based medicine in Kampo, the Japan Society for Oriental Medicine has run a project from 2001 to gather comprehensive data on the randomized controlled trials (RCTs) of Kampo formulae in Japan, and to compile structured abstracts with critical appraisal. This report is published annually as Evidence Reports of Kampo Treatment (EKAT) in Japanese and English.[4] In the current version EKAT 2013, 402 RCTs of Kampo formulae are included.

    The first point to be addressed is that the systematic review protocol developed by Dr. Jun et al. lacks searches of articles written in Japanese; thus, it may result in a biased review. The authors are encouraged to search in Japanese database, Ichushi [5], although no RCTs of kambakutaisoto (Japanese name for GMDZ decoction) seem to be reported in EKAT 2013. Secondary, the authors mention that they include forms of GMDZ such as extracts, tablets, capsules, pills, powders or injections. However, unlike Japanese Kampo extracts, the quality and ingredients of the formulae such as tablets, capsules, or crude powders are not strictly controlled, especially in the older articles, which may result in a bias. Thirdly, although the authors are planning to search the Kampo formula by kambakutaisoto or kam baku tai soto, older articles seem to use the word kanbaku-taiso-to to express the formula, and this word should be included in the search. The structured notation of kambakutaisoto is not quite correct; it should be Kambaku-taiso-To.

    1. Matsumoto M, Inoue K, Kajii E. Integrating traditional medicine in Japan: the case of Kampo medicines. Complement Ther Med 1999;7(4):254-5 2. Kawashima N, Deveaux TE, Yoshida N, et al. Choreito, a formula from Japanese traditional medicine (Kampo medicine), for massive hemorrhagic cystitis and clot retention in a pediatric patient with refractory acute lymphoblastic leukemia. Phytomedicine 2012;19(12):1143-6. 3. Motoo Y, Arai I, Tsutani K. Use of Kampo diagnosis in randomized controlled trials of kampo products in Japan: a systematic review. PLoS One 2014;9(8):e104422. 4. Task Force for Evidence Reports/Clinical Practice Guidelines (ER/CPG- TF) Special Committee for Evidence-based Medicine (EBM), The Japan Society for Oriental Medicine (JSOM). Evidence Reports of Kampo Treatment (EKAT) Appendix 2012. Retrieved September 4, 2014, from 5. Tomizawa Y. Ichushi, Japanese medical bibliography. Kyobu Geka 2010;63(7):585-9, in Japanese

    Conflict of Interest:

    None declared

    Read all letters published for this article

    Submit response
  6. Reply to Clausen and Rydahl.

    Thanks to Jette Clausen (JC) and Eva Rydahl (ER) for their considerations on our study on stillbirths.

    JC and ER suggest our interpretation being "overly optimistic" due to our observational design. But again, observational studies have identified a long list of causes of diseases, be it lung cancer among smokers, venous thrombosis among users of hormonal contraception, greenhouse gas emissions as a cause of global warming, etc. Not only do we identify and quantify causes by observations, we also act according to the knowledge they bring. By combining different observational strategies we may actually achieve evidence of causation, amounting almost to certainty.

    So a well confounder controlled study may actually bring new knowledge, which should - of course - be confirmed by other independent studies, but which in the end may bring progress in our clinical practice. So much in general on interpreting observational studies.

    And no, we have examined a long list of fetal and neonatal outcomes. Our first publication, however, focused on stillbirths. We have submitted the next paper assessing the frequency of asphyxia, cerebral palsy, neonatal deaths, Apgar score, referral to neonatal intensive care units, shoulder dystocia, and peripheral nerve injuries with the more proactive induction practice, and look forward to continue the discussion with these new data.

    The stochastic year-to-year variations in stillbirth rates don't change the convincing and substantial declining trend in stillbirth rates by time, a decline which was steeper after 2009 than before 2009.

    And no, we did not restrict the study window to the period 2009 to 2012 but to the 13-year period 2000-2012, and we acknowledged that several factors in addition to the more proactive induction contributed to the fall in stillbirths. About preeclampsia and intrauterine growth restriction, please see our reply to Carbillon, Hosseraye & Mekinian.

    We have as JC and ER also noticed a still very low stillbirth rate in 2013, actually the next lowest ever measured. Isn't that reassuring?

    Conflict of Interest:

    See original paper.

    Read all letters published for this article

    Submit response
  7. Re:Foetal growth restriction, induction of labour and reduced still birth rate at term

    Thanks to Lionel Carbillon, Claire de la Hosseraye & Arsene Mekinian for their interest in and comments to our paper on stillbirth reduction in Denmark with a more proactive induction practice.

    No doubt, that the improved ultrasound monitoring of pregnancies at term is expected to have decreased the stillbirth rates all over the industrialised world. It is also true, that the stillbirth rates may be higher in women with preeclampsia and intrauterine growth restriction.

    The number of deaths after 37 weeks associated with IUGR was around 6 per year without any consistent trend during the study period. The adjusted hazard ratio of stillbirth among women with IUGR was 1.85 (1.4- 2.5). And the mean gestational age at delivery among women with IUGR was stable during the study period (39 weeks +/- 1 day). Thus, IUGR had a relatively little share of the total number of fetal deaths from 37 weeks, and did not play any confounding role on the influence of inductions.

    The number of women with preeclampsia (or more correctly with a diagnosis of preeclampsia) increased through the study period, but stillbirths in women with preeclampsia were slightly decreasing from 6 to 4 deaths per year. The HR of stillbirths among women with preeclampsia was not significantly elevated. The average gestational age at delivery among women with preeclampsia was (surprisingly) stable through the study period; 39 + 1-2. The confounding influence from preeclampsia on the influence of induction was not significant.

    Our point was and still is that the more intensive surveillance of women with IUGR or preeclampsia did not differ from our neighbour countries. While the stillbirth rates have been rather stable in these countries, our stillbirth rates were substantially reduced. Therefore, we don't think that these two conditions or the handling of them, can explain the substantial reduction in stillbirths in Denmark, especially not the decrease after 40 weeks, at which time the majority of women with these two conditions have delivered. Earlier induction in women having passed 40 weeks is certainly not the only contributing factor for the decrease in stillbirths in Denmark. We already had a low rate 10 years ago, but have experienced a further substantial reduction with the new induction practice, a reduction not seen in other countries. Thus, we still think that the more proactive induction practice has the main responsibility for this recent decrease.

    Conflict of Interest:

    See original paper.

    Read all letters published for this article

    Submit response
  8. Infections in rheumatoid arthritis - roles for herpes viruses

    This is a fascinating study into infections during onset and progress of RA.

    I have recently been investigating potential roles for herpes viruses, especially cytomegalovirus (CMV), in a series of infectious-like outbreaks seen within the UK over many years. These outbreaks lead to increased deaths and hospital admission for a range of medical conditions which include allergy[1-9]. There are several areas of potential cross- over with this study.

    Firstly, a review of the role of CMV in autoimmune diseases has suggested that this immune modulating virus may act to exacerbate symptoms [10].

    CMV has also been circumstantially implicated in a range of respiratory symptoms which appear to be associated with the outbreaks seen in the UK [11] and noted in this study.

    Hence would it be possible for the authors to re-evaluate their data to see if the infectious episodes cluster in time around 2002, 2007 and 2012 (not included in the study which is however presumably on-going) - or at least roughly five years apart? My own studies indicate that the 2012 outbreak can be seen in over 9 European countries (unpublished) and Russia is unlikely to be an exception.

    High fever of uncertain origin (as reported in the study) is one potential red flag for an active CMV infection.

    The study may well be too small, but could the authors also look and see if there is a disproportionate incidence of appendicitis, since the outbreaks in the UK also appear to be in some way linked to surges in the incidence of appendicitis [12].


    1. Jones R (2013) Could cytomegalovirus be causing widespread outbreaks of chronic poor health. In Hypotheses in Clinical Medicine, pp 37-79, Eds M. Shoja, et al. New York: Nova Science Publishers Inc. Available from:

    2. Jones R (2013) Recurring outbreaks of a subtle condition leading to hospitalization and death. Epidemiology: Open access 4(3): 137.

    3. Jones R (2013) Do recurring outbreaks of a type of infectious immune impairment trigger cyclic changes in the gender ratio at birth? Biomedicine International 4(1): 26-39.

    4. Jones R (2013) A recurring series of infectious-like events leading to excess deaths, emergency department attendances and medical admissions in Scotland. Biomedicine International 4(2): 72-86.

    5. Jones R, Goldeck D (2014) Unexpected and unexplained increase in death due to neurological disorders in 2012 in England and Wales: Is cytomegalovirus implicated? Medical Hypotheses 83(1): 25-31.

    6. Jones R (2014) Unexpected single-year-of-age changes in the elderly mortality rate in 2012 in England and Wales. British Journal of Medicine and Medical Research 4(16): 3196-3207.

    7. Jones R (2014) Infectious-like Spread of an Agent Leading to Increased Medical Admissions and Deaths in Wigan (England), during 2011 and 2012. British Journal of Medicine and Medical Research 4(28): 4723- 4741.

    8. Jones R (2014) Trends in admission for allergy. British Journal of Healthcare Management 20(7): 350-351.

    9. Jones R (2014) Infectious-like spread of an agent leading to increased medical hospital admission in the North East Essex area of the East of England. Biomedicine International 5(1): in press

    10. Jones R (2014) Roles for cytomegalovirus in infection, inflammation and autoimmunity. In Infection and Autoimmunity, 2nd Edition, Eds: N Rose, et al. Elsevier: Amsterdam. (in press)

    11. Jones R (2014) A Study of an Unexplained and Large Increase in Respiratory Deaths in England and Wales: Is the Pattern of Diagnoses Consistent with the Potential Involvement of Cytomegalovirus? British Journal of Medicine and Medical Research 4(33): 5179-5192.

    12. Jones R (2014) An unexpected increase in adult appendicitis in England (2000/01 to 2012/13): Could cytomegalovirus (CMV) be a risk factor? British Journal of Medicine and Medical Research (in press)

    Conflict of Interest:

    None declared

    Read all letters published for this article

    Submit response
  9. Re:Re:Reply to "Reduction in stillbirths at term after new birth induction paradigm: results of a national intervention"

    A call for humility on the Reduction in stillbirths at term after new birth induction paradigm

    Hedegaard et al state in their recent article: Reduction in stillbirths at term after new birth induction paradigm: results of a national intervention, that they 'see no reason why a similar, more proactive induction paradigm could not be implemented in other countries' (1). Their observational study design limits drawing conclusions about causality, and their overly optimistic interpretation of the data provide good reason to proceed with caution.

    Is stillbirth the best and only outcome measure?

    All interventions carry risks for mother and unborn. Although the authors mention risks associated with inductions they do not provide any data on maternal risk and only limited data on risk to the foetus.

    Hedegaard et al base their conclusion mainly on differences in stillbirth rates and deaths the first week. The overwhelming importance to parents is whether they can celebrate future birthdays with their healthy child or not, not that they only reached full term with a live foetus. Neo -, peri- or infant mortality rates provides essential information to monitor delayed complications of induction these measures is however not used in this article. According to the official published Danish statistics on birth outcome there has not been any significant change from 2009-12 in the overall perinatal mortality (2,3).

    What is the best way to make conclusions about causality from observational data?

    When reviewing observational data, the conclusions can radically change based on broadening the time frame, or by excluding or including "anomalous" data points.

    We planned to add figure 1. here, however the BMJ comment system does not allow us to upload pictures. You can access the relevant figures here:

    Figure 1. contains crude data from the Medical Birth Register (including all Danish birth from 1996-2013 from 37+0 weeks and onwards). The red curve illustrates stillbirth rates. The authors highlight that: 'The fall [in stillbirth] was steepest from 2009-2010 to 2011-2012' and they link the steep decrease to the change in the 2009 guidelines about induction of labour. However, the rate of stillbirths actually increased in 2010. The authors' selective reporting of annual data does not help explain the trends and variations in incidences, and it paves the way for a conclusion that is unfounded. One might even notice a slight increase in stillbirth in 2013 which is after the study period (4).

    Stillbirth is the one side of the equation on the other side is inductions; let's now take a look at this (Figure 2).

    Figure 2. Induction rates 1997-2023 in %. You can access it here:

    Figure 2. shows a rather steep increase in number of inductions 1997- 2001, a stable period 2002-2007, and a steep increase 2008 -2013. However, the stillbirth rate appears to decline continuously over time and if we overlay the two figures, it does not mirror the change pattern in induction rates. Hedegaard et al. utilize a narrow time frame between the years 2009-12 to support an argument in favour of early induction.

    However one peak or decrease in the dataset of an observational study is not sufficient to draw strong conclusions. The Cochrane handbook recommends that studies like Hedegaard et al. which utilize a single intervention or studies which do not clearly define a period of time course through which the intervention was carried out should adhere to an additional level of scrutiny. In particular, The Cochrane Collaboration recommends at least three data points prior to and three after the intervention, in order to strengthen conclusions from these weaker study designs. Reviewing Hedegaard et al with this in mind, it is unclear if the authors are utilizing one or two interventions (the guidelines or inductions). And because the intervention itself is unclear it is difficult to determine the time course through which the intervention was carried out.

    Does the data support induction in gestational week 41 +3 to reduce stillbirths?

    We acknowledge, as we mentioned earlier, the decrease in stillbirth throughout the study period. We are however not convinced that 'the striking decreases in risk of late foetal deaths is likely primarily to be due to the earlier and increased induction rates' (our emphasis). Numerous additional factors are still not accounted for by the authors. Others rightfully highlighted that the study did not control for a range of relevant factors, such as birth defects, IUGR, and socio-economic status to name only a few. Also, the presented data shows that the stillbirth decreased in all gestational weeks after 37 weeks. This decrease in stillbirth in week 37+0-6 weeks is curious, especially if we are supposed to conclude that induction accounts for the decrease in stillbirth from 37 weeks. It has not been common practice in Denmark to induce in the 37+0-6 gestational week except in rare cases, and this decrease suggests that factors other than induction of labour, such as improved surveillance, could significantly contribute.

    This study raises many methodological questions, we have pointed to a few of them. We believe that the conclusions drawn from this study are too far reaching.

    Will the correct intervention please step forward?

    Hedegaard et al suggests that other countries should implement a proactive induction regime including early induction of labour for low risk women. We suggest a more humble approach that acknowledges multiple factors interacting in this significant decrease in stillbirths beyond the 37. gestational week over the last decades. Also we have significant concerns about unleashing a technology intensive solution, such as induction of labour, especially on women with normal and uncomplicated deliveries. We see no reason why this more proactive induction paradigm should be exported to other countries as the authors do not present data that support such a strong conclusion.

    1) Hedegaard M, Lidegaard O, Skovlund CW, M?rch LS, Hedegaard M. Reduction in stillbirths at term after new birth induction paradigm: results of a national intervention. BMJ Open. 2014;4:e005785.

    2) Sundhedsstyrelsen. F?dselsstatistikken. 2011; tabel 2.2. http://sun3) Accessed August 26th

    3) Statens Seruminstitut. F?dselsstatistik tal og analyser. 2012; table 2.2. %20dansk/Sundhedsdata%20og%20it/NSF/Registre/Fodselsregisteret/f%C3%B8dselsstatistikken2012_vers%204.ashx. Accessed August 26th

    4) Statens Serum Institut. The online Medical Birth Registry. Accessed August 26th

    Conflict of Interest:

    None declared

    Read all letters published for this article

    Submit response

Don't forget to sign up for content alerts to receive selected information relevant to your specialty interests and be the first to know when the latest research is published.