Displaying 1-10 letters out of 408 published
Transparency grade reflects only one aspect of transparency
Imagine a drug company conducting 10 trials, 5 of which turn out positive and 5 negative. The 5 positive trials are, unsurprisingly, published. The 5 negative trials are subjected to outcome reporting bias, aka statistical alchemy and HARKing (hypothesizing after the results are known), so that they appear to be positive in the final publications. The downstream reader (or even systematic reviewer) is misled into believing that 10 of the 10 trials were positive. In my opinion, this hypothetical drug would deserve a transparency score of 50%, but according to the methods employed in this study, simply because all 10 were published, the drug would receive a score of 100%. To more fully assess for transparency, one must go beyond the fact of publication and verify the published results by checking for consistency with results reported in ClinicalTrials.gov and/or FDA Drug Approval Packages.
Conflict of Interest:
Separated cycling routes on cyclist safety
Globally, there is a strong movement to encourage active lifestyles in our communities. Clinicians, the public health community and policy makers around the globe are encouraging communities to engage in active lifestyles. Dr.Vivek Murthy, the US Surgeon General, is convinced of the benefits of active lifestyles in our communities and he encourages communities to embrace healthy behaviors (1). Furthermore, the Surgeon General aims to increase active lifestyles across the nation by calling for access to safe and convenient places to engage in these activities (1). Dr. Murthy rightly pointed out that it is not simply enough to encourage active lifestyles on its own without thinking about the importance of the built environment (1). A recent Canadian study by Teschke and her colleagues (2015) rightly recommend that the policy makers need to focus on bike infrastructure to promote safer cycling in our communities (2).
We also believe that the efforts to improve cyclist safety cannot succeed without making substantial improvements to the current built environment. Despite substantial evidence base documenting the efficacy of separated cycling routes on cyclist safety (3) there is a lack of funding or investment to implement this evidence based cyclist injury prevention strategy. Research also shows that safer cycling infrastructures increase cycling rates significantly (4).
(1) Step It Up! The Surgeon General's Call to Action to Promote Walking and Walkable Communities. Available online at: http://www.surgeongeneral.gov/library/calls/walking-and-walkable- communities/
(2) Teschke K, Koehoorn M, Shen H, Dennis J. Bicycling injury hospitalisation rates in Canadian jurisdictions: analyses examining associations with helmet legislation and mode share.BMJ Open. 2015; 2;5(11):e008052. doi: 10.1136/bmjopen-2015-008052.
(3) Reynolds CC, Harris MA, Teschke K, Cripton PA, Winters M. Environ Health. 2009 ;21;8:47.
(4) Winters M, Brauer M, Setton EM, Teschke K. Built environment influences on healthy transportation choices: bicycling versus driving. J Urban Health. 2010;87(6):969-93.
Conflict of Interest:
The effect of statins on average survival - revisited
The recent paper by Kristensen et al.  examined 11 randomized clinical trials (RCTs) of statins and concluded that they provide "a surprisingly small average gain in overall survival within the trials' running time", and that "The median postponement of death for primary and secondary prevention trials were 3.2 and 4.1 days, respectively."
This appears to question the value of statins, yet the information as presented in this paper should not dissuade their use. Because the study restricted its analysis to, "... within the trials' running time", it reflects a shortsighted view. This is compounded by the fact that many of these trials were stopped early. Importantly, patients on statins enjoy a much greater lifetime extension benefit than that noted within the limited timeframe of these trials, as we will show using a re-analysis of the Kristensen et al. data.
It's also important to note that one of the larger included trials lacking a mortality benefit, ALLHAT-LLA, had a between difference reduction in LDL-cholesterol of only 17% . This largely reflected the high rate of statin use in the usual care arm.
The authors used pixel count to estimate the area between the Kaplan-Meier curves for survival. This area equals the difference in mean survival between the two patient cohorts. But, the authors used only the initial part of the survival curves, during the trials' relatively short run-times, ranging from 2.0 to 6.1 years.
The true gain in mean survival should be based on the patients' entire lifetime. We attempted to estimate this quantity using the following method, which we demonstrate using the series  that the authors used for demonstration in their online Appendix . Figures are available from the authors upon request (ShuChing.Chang@providence.org).
1. 1. The original Kaplan-Meier curves from Figure 2A in the NEJM paper , were digitized using the Engauge Digitizer program  to convert the published curve segments into x and y coordinates.
2. 2. Gompertz distributions were fit to the digitized points using the optim function in the statistical package R . The Gompertz distribution  is widely used to model survival, fits especially well for older populations  and fit these curves very well, as measured by the Kolmogorov-Smirnov tests (p ~ 1.00 for both Pravastatin and Placebo curves).
3. Once we had the model parameters for the Gompertz distribution, we extrapolated the curves out to the end of the patients' lifetimes. The median survival times for the fitted curves of Pravastatin and Placebo were 16.4 years and 13.3 years, respectively, and the difference of median survival times between these two groups was 3.1 years.
4. Finally, we computed the areas between the curves, using the Gompertz formulas. The extension of life out to the end of the RCT (6.1 years) was 22 days, which agrees with what the authors found using their pixel-counting method. But, the total savings in mean lifetime, shown in green, is 3.0 years.
To base clinical decision-making on the extended survival achieved during the relatively short time period of an RCT runtime is misleading; "extra lifetime during the RCT" is only one measure of efficacy. More pertinent to an informed discussion between physician and patient are the lifetime benefits of statins. In this example, the extension of life went from 22 days during the trial period only to 3.0 years for the patient's expected lifespan.
Just we were ready to submit our letter, we found the letter by Professor Marshall . We agree with all he said, and he even foretold the approach that we were taking: "It may be possible to improve on these estimates, scanning and digitizing the published survival curves and fitting the Weibull survival to the digital curves".
1. Kristensen, M. L., Christensen, P. M., & Hallas, J. The effect of statins on average survival in randomised trials, an analysis of end point postponement. BMJ open 2015; 5(9): e007118
2. ALLHAT Collaborative Research Group. JAMA 2002;288:2998-3007
3. [No authors listed]. Prevention of cardiovascular events and death with pravastatin in patients with coronary heart disease and a broad range of initial cholesterol levels. The Long-Term Intervention with Pravastatin in Ischaemic Disease (LIPID) Study Group. N Engl J Med 1998;339:1349-57.
4. http://bmjopen.bmj.com/content/suppl/2015/09/24/bmjopen-2014- 007118.DC1/bmjopen-2014-007118supp_appendix.pdf
5. Mitchell, M. (2002). Engauge digitizer 4.1. URL http://digitizer.sourceforge.net
6. R Core Team, R (2015). R: A language and environment for statistical computing. Vienna, Austria; 2014. URL http://www. R-project.org
7. Gompertz, B. On the nature of the function expressive of the law of human mortality, and on a new mode of determining the value of life contingencies. Philosophical transactions of the Royal Society of London 1825, 513-583
8. Kiche, J., Aduda, J., Athiany, K. H. O., BO, A., & Wanjoya, A. Comparison Of Survival Models And Estimation Of Their Parameters With Respect To Mortality in a Given Population. Mathematical Theory and Modeling 2014; 4(7): 164-176
Conflict of Interest:
Euthanasia for psychiatric patients: ethical and legal concerns about the Belgian practice
As Belgian professionals in mental health care and medical ethics, we felt the need to comments on the paper published in BMJ Open by Thienpont et al., entitled "Euthanasia requests, procedures and outcomes for 100 Belgian patients suffering from psychiatric disorders: a retrospective, descriptive study" (Thienpont et al., BMJ Open. 2015 Jul 7;5(7):e007454.) We argue that this study denies several important and unresolved issues in this ethically very delicate field.
First, we would like to emphasize that the paper represents the practice of a small minority of the Belgian psychiatric professionals. The 100 cases described in this study stem from one single psychiatrist, the first author. These 100 cases led to the actual performance of euthanasia in 35 patients between 2007 and 2011. The Federal Control and Evaluation Commission for Euthanasia (FCEC) does not provide specific data on euthanasia in psychiatric patients in the whole of Belgium during this period. However, the number of euthanasia cases linked to "neuropsychiatric disorders" (also including neurodegenerative disorders such as Alzheimer's and Huntington's disease) does not exceed 100 for the entire period 2007-2011 (see reports at http://www.health.fgov.be/eportal/Healthcare/Consultativebodies/Commissions/Euthanasia/Publications/). Hence, in Belgium a substantial proportion of euthanasia cases for psychiatric disorders - certainly more than 35% and probably close to 50% - has been approved by one single psychiatrist. The abstract suggests that out of 100 euthanasia requests submitted to this single psychiatrist, 48 got a positive advice, and 35 lead to the actual application of the euthanasia procedure. This is misleading. In fact, the paper states that 38 patients withdrew their request before any advice was formulated. This implies that the actual number of euthanasia requests approved in this study was 48 out of 62, or 77.4 %.
Second, we believe that the qualification of a patient's suffering being "unbearable" and "untreatable", a necessary condition for the approval of an euthanasia request according to the Belgian law, has not been sufficiently clarified in a psychiatric context. The author correctly states "...that the concept of 'unbearable suffering' has not yet been defined adequately, and that views on this concept are in a state of flux". But also the notion of a psychiatric disease being "untreatable" poses major problems. The most common psychiatric disorders diagnosed in the 100 cases described in the paper were major depression and (borderline) personality disorder. It is possible that in both disorders, the suffering is often perceived as unbearable by the patient, but that this state is transient in many cases. Depression is self-limiting in its natural course. Regarding borderline personality disorder (BPD), scientific evidence shows that this disease tends to resolve over time. A DSM diagnosis of BPD remits in 56% of patients after 2 years (Grilo e.a. 2004; Skodol e.a. 2005). According to Zanarini et al. (2011), 85% of BPD patients were in remission 10 years after the diagnosis was formulated.
Further, the application of the term "untreatable" in a psychiatric context poses fundamental problems. The act of diagnosing a psychiatric patient, then calling the disease "untreatable" and proceeding to a life- ending procedure, implies a narrow technical view on psychiatry. While a given disease might be resistant to classical treatments, the psychiatric patient as a person is never beyond the possibility of recovery. To consider a patient a "hopeless case" when evidence-based therapies do not lead to remission of the diagnosis, denies the fact that psychiatrists do not primarily treat diseases, but persons, who can always surprise by their capabilities to recover, with or without evidence-based therapies. This is demonstrated by the following paragraph in the paper "In the remaining 57 cases (12 men and 45 women), the patients or their practitioners were contacted and it was confirmed that these patients were still alive. In nine cases their requests were still in process and no decision had been reached. In 48 cases, their requests were on hold because they were managing with regular, occasional or no therapy."
The paper by Thienpont et al. also states that "all patients were legally competent under the Law" and that "the patient's capacity for discernment was evaluated during the process (...) according to explicit criteria described in the Belgian Legal Doctrine." Both from a clinical and ethical point of view this way of determining the accountability and moral autonomy of psychiatric patients by a merely legal procedure is questionable. In particular the determination of the capacity of discernment of psychiatric patients requires more clinical, therapeutic and ethical caution than a legal procedure can offer.
Summarizing, we believe that this paper reveals some alarming features of the Belgian law regarding euthanasia. The Law has been stretched out to psychiatry with insufficient consideration for the specificity of the field. We think that mental health care workers in Belgium should urgently start a fundamental debate on this matter. We also think that the current dangerous evolution in Belgium should be a warning for other countries considering to implement euthanasia legislation.
Grilo CM, Sanislow CA, Gunderson JG, e.a. Two-year stability and change of schizotypal, borderline, avoidant, and obsessive- compulsive personality disorders. J Consult Clin Psychol 2004; 72: 767-75.
Skodol AE, Gunderson JG, Shea MT, e.a. The collaborative longitudinal personality disorders study (CLPS): overview and implications. J Pers Disord 2005a; 19: 487-504.
Zanarini MC, Frankenburg FR, Hennen J, e.a. Prediction of the 10-year course of borderline personality disorder. Am J Psychiatry 2006; 163: 827- 32.
Conflict of Interest:
Assessing postponement time
Kristensen et al  have estimated "postponement" of death in trials due to statins. Their finding is that postponement is measured in days - median pooled 3 to 4 days - which suggests statins have scarcely any benefit. It may well be that their benefits are, indeed, minimal, but the point of this note is not to question efficacy of statins, rather it is to suggest that the statistical analysis presented in this study is somewhat misleading.
The method is based on estimating the difference in the area between survival curves of statin and placebo up to a cutoff in time. However, although the area under a survival curve equals average survival time when taken over the whole time span, its interpretation when taken up to a cutoff cannot be viewed in this way. Royston and Parmar  have referred to the area under a survival curve up to a cutoff as "restricted mean survival time" (rmst). In effect, rmst replaces survival times that occur after the cutoff with the cutoff value itself. For example in the authors' LIPID case study (Appendix B of ), which has a 6.1 year cutoff, it is tantamount to assuming that everyone who survived beyond 6.1 years actually "died" at 6.1 years. Since about 90% of people in the statin group and 85% in the placebo groups survived beyond 6.1 years, and are all regarded as having died at that same time, it is no surprise the mean difference is very small (22 days).
It seems more meaningful to consider postponement of death as the area between placebo and statin groups taken over the full time span. Without raw data and complete follow-up, this cannot be done. However an attempt can be made to model survival from the reported Kaplan-Meier plots to extrapolate the full survival. For example, in Appendix B of , a model which allows an increasing hazard may be appropriate (given the apparent increasing slopes). A Weibull model could be considered . It has two parameters and a relatively simple mathematical form for its survival function. The model can be fitted from any two points on a survival curve, and solving for the two parameters. For the LIPID study I took time points at t=6.1 and t=4 years and estimated the survival, S(t), as best I could from the published graphic, for each of the two curves, giving S(6.1)=0.888 and S(4)=0.935 for the statin group, and 0.854 and 0.919 for placebo. The fitted models give mean survival 27.1 and 19.2 years, with mean difference 7.9 years (calculation details available from me).
This analysis, which assumes that the benefit of the drug continues beyond the duration of the trial, is admittedly rather crude. First, it assumes a Weibull model with parameters estimated, rather inefficiently, from two points on a survival curve and read off presumably with some inaccuracy from the ordinates of a zoomed printed graphic. It may be possible to improve on these estimates, scanning and digitizing the published survival curves and fitting the Weibull survival to the digital curves.
However, the difference is unlikely to be appreciable; the main point is that postponement of death is more realistically measured in years if survival curves are extrapolated rather than a few days (7.9 years for the LIPID trial against 22 days as reported in ). Further, given that there is a different cutoff for each of the trials analysed in  (ranging from 2 to 6.1 years), the reported postponements are not comparable so that pooling them, as the authors have done, is questionable. It would be interesting to analyse each of the trials, as suggested here for the LIPID trial, to obtain a more constructive overall measure.
1. Kristensen ML, Christensen PM, and Hallas J. The effect of statins on average survival in randomized trials: an analysis of end point postponement. BMJ Open 2015;5:e007118.doi:10.1136/bmjopen-2014-00718
2. Royston P, Parmar MKB. Restricted mean survival time: an alternative to the hazard ratio for design and analysis of randomized trials with a time-to-event outcome. BMC Medical Research Methodology 2013,13:152 http://www.biomedcentral.com/1471-2288/13/152)
Conflict of Interest:
Government cancer research spend recorded in the NCRI Cancer Research Database
We read with interest this article by Luengo Fernandez et al. and commend the team on exploring a number of important areas in health research funding, especially given that the collection of data across a number of years in a bespoke manner such as this is very challenging.
While the article highlights some very interesting trends, one trend that has not been observed in the National Cancer Research Institute (NCRI) Cancer Research Database (http://www.ncri.org.uk/what-we- do/research-database), in the instance of cancer, is the sharp decline in government spend between 2008 and 2012. The government spend recorded in the database in 2008 was 173 million pounds while in 2012 it was 194 million pounds (unadjusted for inflation). This would suggest there is room for further analysis to understand why this inconsistency has occurred.
The NCRI collects data on the peer-reviewed cancer research funded by its Partners annually (which includes BBSRC, Chief Scientist Office, Department of Health, ESRC, Health and Care Research Wales, HSC Public Health Agency (R&D division) and Medical Research Council). Partners submit information for each cancer-relevant grant that is in progress on the 1st April that year. The data includes the award title and abstract as well as financial information allowing analysis of spend by Partner, research category or cancer site. In this way we create a snapshot of work in progress on the census date of the 1st April (it should be noted that this is not the same as actual expenditure over any given 12-month period). The analysis of research category (www.icrpartnership.org/cso) and cancer site is replicated internationally through the ICRP whose members (including the NCRI) communicate regularly to combine resources to conduct regular international analyses of cancer research funding.
Of note, the UK Clinical Research Collaboration (UKCRC) has also recently published a UK Health Research Analysis, looking at health research portfolios of government and charity organisations funding health related research in the UK. This covers a range of disease areas including cancer, and adds complementary detail to better understand the UK picture.
Conflict of Interest:
Response: Assessing knowledge and attitudes regarding HPV vaccination among those who have never heard of it
Thank you for your detailed reading of our paper, we appreciate you bringing this concern to our attention. As described previously, respondents who had not heard of the HPV vaccine were briefly educated. Specifically, we read them the following statement: "Human Papillomavirus (HPV) can cause cervical cancer. HPV is not the same as HIV. Vaccines are medications given to prevent the development of disease or illness." Respondents were also encouraged to ask questions during the interview. We believe this was a sufficient amount of information in order for the referenced questions to be asked. The interpretations in Table 4 are valid.
Conflict of Interest:
There are some corrections to be made to this article.
Issue 1: Page 4, Table 2 footnote
*x(squared) =0.35 p=0.557. (dagger)x(squared) =0.88 p=0.348. (double dagger)x(squared) =0.70 p=0.403. (section sign)x(squared) =2.90 p<0.088
*x(squared) =0.35 p=0.557 (dagger)x(squared) =0.88 p=0.348 (double dagger)x(squared) =0.70 p=0.403 (section sign)x(squared) =2.90 p=0.088
Issue 2: Page 4
Comparison of drug requirement among study participants
Using the WHO guidelines for CVD risk assessment and management, (2) estimates of proportions
Comparison of drug requirement among study participants
Using the WHO guidelines for CVD risk assessment and management, estimates of proportions
Issue 3: Page 5, Table 4 footnote
*Without cholesterol versus single risk factor, threshold greater than or equal to 30%: x(squared) =6.32, p<0.012 (dagger)Without cholesterol, threshold greater than or equal to 30% versus without cholesterol, threshold greater than or equal to 20%: x(squared) =10.06, p<0.002. (double dagger)Without cholesterol versus single risk factor, threshold greater than or equal to 20%: x(squared) =0.45, p<0.500. (section sign)With cholesterol versus single risk factor, threshold greater than or equal to 30%: x(squared) =123.9, p<0.001. (paragraph sign)With cholesterol, threshold greater than or equal to 30% versus with cholesterol, threshold greater than or equal to 20%: x(squared) =8.77, p<0.003. **With cholesterol versus single risk factor, threshold greater than or equal to 20%: x(squared) =74.5, p<0.001.
*Without cholesterol versus single risk factor, threshold greater than or equal to 30%: x(squared) =6.32, p=0.012 (dagger)Without cholesterol, threshold greater than or equal to 30% versus without cholesterol, threshold greater than or equal to 20%: x(squared) =10.06, p=0.002 (double dagger)Without cholesterol versus single risk factor, threshold greater than or equal to 20%: x(squared) =0.45, p=0.500 (section sign)With cholesterol versus single risk factor, threshold greater than or equal to 30%: x(squared) =123.9, p<0.001 (paragraph sign)With cholesterol, threshold greater than or equal to 30% versus with cholesterol, threshold greater than or equal to 20%: x(squared) =8.77, p=0.003 **With cholesterol versus single risk factor, threshold greater than or equal to 20%: x(squared) =74.5, p<0.001.
Conflict of Interest:
Study still dead after resurrection attempt fails
The August 11 response of Dr. Elard Koch and several co-authors fails to rectify the two major flaws in their study that I exposed in my April 9 rebuttal, and in fact, further confirms those flaws. Their very lengthy response also ignored many of my arguments and repeated many of their disputed points. This reply explains why those two major errors remain fatal to their study's conclusions. I then address the question of anti-abortion bias, which I believe played a key role in the flawed design of the study. To be fair, I also address my own biases. Finally, I deal with several ancillary issues arising from Koch et al.'s response (for those interested to read on).
FATAL FLAWS ARE STILL FATAL
My rebuttal had explained that abortion law was not related to abortion practice in Mexico, and that this invalidates the study's methodology and conclusion. That's because you can't hypothesize a possible effect on abortion mortality based on the existence or absence of various legal exemptions across Mexican states, when those exemptions don't even work. In practice, few Mexican women can access abortions under the legal exemptions. In response, Koch et al. confirmed this flaw by saying: "...between states, legal permissiveness may be the same, but accessibility may differ by multiple unmeasured factors. We do not say nor have we stated that we are using this term as a proxy for greater or less accessibility to pregnancy termination..." But their failure to account for the difference between law and practice IS THE FLAW. Their methodology implicitly assumed that abortion restrictions reflect abortion practice, but since they now concede that's not the case, their conclusion that maternal mortality is unrelated to the presence or absence of certain abortion restrictions is meaningless.
I had shown that it was arbitrary and erroneous to select, in particular, the legal exemption for abortion in cases of genetic or congenital malformation as the criterion for assigning Mexican states to the categories of "more permissive" or "less permissive" abortion legislation, depending on whether the state had that exemption. The rarity of abortion for fetal abnormality makes this variable useless as a proxy, because the miniscule numbers of maternal deaths that may result would be impossible to detect statistically in order to compare states with and without the exemption. The authors' only response is to argue that such abortions may be slightly less rare than I claim, based on the low prevalence of Down???s syndrome at birth in several other countries. But that has nothing to do with Mexico or the fact that abortions for any kind of serious fetal abnormality are in the range of 1% of all abortions in countries where abortion is widely legal. So their error remains, especially since abortions due to fetal abnormality are likely even rarer in Mexico than in western countries because of its restrictive laws, inaccessibility of legal abortion under the exemptions, profound stigma, and physician refusals.
The authors say that exploratory analyses are "valid and valuable tools to avoid an arbitrary categorization." But somehow, that's exactly what they ended up with. The authors confirm their flaw by explaining that they selected the legal exemption for fetal abnormality as the main variable because it was the only one that yielded an association with maternal mortality, as well as the only one where the 32 Mexican states had a roughly equal distribution in terms of whether they had the exemption or not. In other words, the variable was arbitrarily selected on the basis of convenience and because it happened to yield an association that I've already shown is meaningless. (Correlation does not equal causation.) A random association is easy to find if you conduct a fishing expedition for one amongst a host of variables, but one should not then base an entire study on it and draw conclusions from it.
Nine out of ten authors of the BMJ Open study had an undisclosed bias because they are signatories to the "Dublin Declaration." (http://www.dublindeclaration.com/signatories/) This document denies the need for legal abortion even to save a woman's life. Notably, the peer review process for their BMJ Open study failed to catch the substantive errors in the study because neither peer reviewer had expertise in the subject matter, and one (R. Lieva) appears to hold the same position as the authors. (http://www.thelancet.com/journals/lancet/article/PIIS0140- 6736%2810%2961251-2/abstract) The study contains repeated citations (over 30) of the authors' own past research on abortion and lists over a dozen references from Koch and various co-authors. The effect is to create a false picture of scientific confusion and conflicting data in the abortion field. This same pattern is repeated in their reply to me, even though several of their studies have been criticised. (Here's a compendium of rebuttals to their work: http://choice-joyce.blogspot.ca/2012/08)) Koch et al. have gone to great lengths (in this and other studies) to try and show that factors such as access to emergency obstetric care have the greatest apparent impact on maternal mortality, not abortion laws. But their focus on maternal deaths, while important, obscures all the other suffering and harms that criminal abortion laws cause to women, including high complication rates from unsafe abortion (about 159,000 women are hospitalized annually in Mexico: http://www.guttmacher.org/pubs/Unintended -Pregnancy-Mexico.html). The implication is that we should accept this human cost as long as good health care systems can save women's lives in the end. Presumably the authors would disagree, and hopefully their implication was unintentional.
Given the overwhelming global and historical evidence of the danger posed to women by criminal abortion laws, (http://www.who.int/reproductivehealth/publications/unsafe_abortion/9789241501118/en/ and http://www.ncbi.nlm.nih.gov/pubmed/1949105 and numerous other sources) any study that concludes that restrictive abortion laws do not contribute to maternal mortality should be treated with suspicion.
But let me turn now to the question of my own biases, since Koch et al. would presumably claim that I'm the one who's biased, as well as unscientific. This is apparent from the authors' near ad hominem treatment of my criticisms, dismissing them as "based largely on personal opinions or speculative assumptions," and "not scientifically based." I'm a writer and pro-choice activist, not a scientist. I do not apologize for my impassioned defense of women's rights and lives. I've been monitoring the anti-choice movement for 25 years and Koch's work for 3 years.
For the record, I did not have "discomfort" with the study's findings, because I recognized the study methodology and conclusions were deeply flawed. However, I did feel offended by the study. I object to its very premise and even that it was published in a reputable scientific journal where it does not belong. Please allow me to explain.
I have a bias in favour of the belief that women deserve equality, human rights, and dignity. I also have a bias against criminalizing life- saving healthcare that only women need. Should such biases, if expressed, lessen my credibility or weaken my arguments? If so, it must be because it's still up for debate whether women deserve human rights and freedoms, including the fundamental right to control their fertility.
ADDENDUM: ANCILLARY ISSUES
Unsafe Abortion as a Contributing Factor to Maternal Mortality: The authors protest my "unsubstantiated and dissociated" suggestion that they are using other contributing factors to maternal mortality as a smokescreen to cover up the effect of unsafe or illegal abortion. Their objection is largely answered by my points above on author bias. Further, I believe the authors are being disingenuous when they say they want to emphasize the other factors that contribute to women suffering and dying in childbirth from preventable causes. Rather, they appear to be primarily concerned with abortion. Their study's title begins with "Abortion legislation" as the key factor. The study objective is: "To test whether there is an association between abortion legislation and maternal mortality outcomes after controlling for other factors thought to influence maternal health." Their press release for the study (http://www.melisainstitute.org/pr-18022015-english.html) opens with: "Laws protecting the unborn and therefore, less permissive in regard to abortion, bear a negative ...connotation, because induced abortion in clandestinity might increase maternal deaths. However, a new study conducted in 32 Mexican states ...challenges this notion..." And most of their other research focuses on abortion.
Of course, many factors contribute to maternal mortality, and many effective ways exist to address it. But as I said in my rebuttal, this range of other factors could swamp the effect of unsafe abortion on maternal mortality rates and make it harder to detect statistically, particularly if a country's mortality rate from unsafe abortion is relatively low compared to other countries. The study methodology the authors employed was almost guaranteed to find no meaningful association anyway, which allowed them to present their foregone conclusion that abortion laws don't affect maternal mortality. Clandestine abortion is still a major cause of maternal mortality in many parts of the world, particularly in Africa and Asia. (http://www.prb.org/pdf11/abortion-facts-and-figures-2011.pdf) It is unusual for developing countries with strict bans on abortion to have relatively low maternal mortality rates. In such countries (like Chile and a handful of others in Latin America), other factors indeed play an important role including widespread use of misoprostol - however, maternal mortality would almost certainly be even lower if abortion was legal, safe, and accessible. Regardless, complications from unsafe abortion remain high throughout Latin America and the Caribbean, with over one million women hospitalized annually - 159,000 in Mexico alone during 2009 (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3557184/ and http://www.guttmacher.org/pubs/Unintended-Pregnancy-Mexico.html). This demonstrates that restrictive abortion laws continue to pose a great danger to women in Mexico, even if fewer are dying than in the past.
Issues with Vital Statistics for Abortion Deaths: The authors' arguments that Mexico does not have a problem with underreporting or misclassification of abortions and associated mortality are unpersuasive. They put a lot of faith in vital statistics from the Mexican government - their only source - but it's unlikely that government statistics are as robust and accurate as they claim. The authors can only cite their own research as support. I found mentions in several studies (listed below) about the problems of misclassification and underreporting associated with the use of vital statistics when it comes to abortion, even in countries like Mexico that have otherwise high-quality records for maternal mortality.
In settings like Latin America where abortion is mostly illegal and highly stigmatized, women and their families would be more likely to not report or misreport an abortion, and health care personnel would be more likely to misclassify causes of death out of ignorance, fear, or compassion. For example, even with the existence of specific ICD codes for "Abortive Outcomes" (ICD-10 O00 to O08), why should we exclude the possible misclassification of unsafe abortion deaths under the codes for sepsis or hemorrhage in early pregnancy (ICD-10 O20 and O23)?
Vital statistics should not be relied upon alone to determine maternal mortality from unsafe abortion. Reputable studies employ a variety of sources and methods, including studies, hospital data, various types of surveys, and more. For no valid reason, Koch et al. neglect and dismiss these additional methodologies, while their own narrower methodology for estimating abortion incidence has been questioned. (http://www.guttmacher.org/media/resources/response-to-methodology- critique.pdf)
Recent studies on maternal death from unsafe abortion do not support Koch et al's claim that the rate for Mexico is as low as 3% of all maternal mortality. Before looking at these other figures, it's important to understand that studies and official statistics on maternal mortality rates and causes do not always distinguish adequately between deaths from induced abortion and spontaneous abortion (miscarriage), or even other "abortive outcomes" such as ectopic or molar pregnancies and other "unspecified" complications. However, we know that maternal deaths from miscarriage, legal abortion, and molar pregnancy are all very rare, at least in settings with reasonable access to health care. While ectopic pregnancies carry a higher risk of death, they are far less common than abortion or miscarriage. Therefore, the majority of officially-recorded abortion deaths in countries with restrictive laws are plausibly due to unsafe induced abortion, including many coded under other "Abortive Outcomes."
The following studies provide rates of maternal mortality due to abortion in Latin America and/or Mexico. The data for Mexico appear to be sourced from vital statistics only, except for #3, which also cites the WHO Mortality Database (in the study's Appendix).
1. The Schiavon et al. estimate for Mexico was 7% for 2008. (http://www.ncbi.nlm.nih.gov/pubmed/22920626)
2. A May 2014 global analysis by the World Health Organization (WHO) (http://www.thelancet.com/pdfs/journals/langlo/PIIS2214-109X%2814%2970227- X.pdf) provided an estimate of 9.9% for Latin America and the Caribbean. For Mexico, the study shows a rate of almost 8% as denoted on a bar graph in the Appendix (page 36). (http://www.thelancet.com/cms/attachment/2014592567/2036073535/mmc1.pdf)
3. A September 2014 study by IHME, the Institute for Health Metrics and Evaluation, shows in bar graphs an average rate of about 17.5% for Latin America and the Caribbean in 2013 (Figure 6, http://www.thelancet.com/action/showFullTextImages?pii=S0140- 6736%2814%2960696-6), and a rate of about 11% for Mexico (Appendix, page 127: http://www.thelancet.com/cms/attachment/2019762887/2039744952/mmc1.pdf).
Why are these data from Mexico (7%, 8%, and 11%) higher than Koch et al's estimate of 3%? It appears it's because Koch et al. excluded deaths coded under ICD-10 O00, O01, O02, and O08 (respectively: ectopic pregnancy, molar pregnancy, other abnormal products of conception [including missed abortion], and complications following abortion and ectopic and molar pregnancy). Excluding these codes could omit many unsafe abortion deaths that were misreported or misclassified.
GLOBAL RATES: The oft-cited 2008 WHO figure of 13% for global maternal mortality due to unsafe abortion (http://apps.who.int/iris/bitstream/10665/44529/1/9789241501118_eng.pdf) should be retired due to the new WHO analysis (#2 above). That study provides a figure of 7.9% globally for 2003 to 2009, with the reduction due to the use of recent data and improved methods. I would suggest that the reduced figure may also reflect the increased use of safer medical abortion by women in some regions. Using a somewhat different methodology, the IHME study (#3 above) yields an estimate of about 15% global maternal mortality from abortion in 2013 (see Table 2, page 995: http://www.thelancet.com/pdfs/journals/lancet/PIIS0140-6736%2814%2960696- 6.pdf). The variance with the WHO figure of 7.9% is unexplained.
INCIDENCE OF UNSAFE ABORTION: This report: http://www.guttmacher.org/pubs/Unintended-Pregnancy-Mexico.pdf was the source for my figures on the significant number of unsafe abortions and complications in Mexico (over a million abortions a year, and 159,000 hospitalizations). These numbers are higher than the estimates from Koch et al. because the latter are based only on vital statistics data, which is insufficient.
Medical Abortion: The authors state: "It is self-contradictory to say that unsafe abortions have increased substantially over the last decades while observing substantial reductions in deaths from this cause." Unsafe abortions have increased globally in recent years but not substantially (from 19.7 million in 2003 to 21.6 million in 2008). But in Latin America, unsafe abortions DECREASED by a third between 1990 and 2008, and maternal mortality went down 70%. (http://apps.who.int/iris/bitstream/10665/75173/1/WHO_RHR_12.01_eng.pdf) This dramatic reduction in maternal deaths coincides with the widespread availability of medical abortion (using misoprostol) in Latin America over the last two decades. In contrast, maternal mortality from unsafe abortion remains high in regions where misoprostol is not as easily attainable, such as Africa. This suggests that misoprostol plays a significant role in reducing women's deaths.
Several studies and reports also attribute the reduction in maternal deaths in Latin America to the increased use of medical abortion (such as: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3557184/). This method has largely replaced more dangerous traditional methods, although it can still lead to a high rate of mostly non-lethal complications when used without medical supervision.
Koch et al. are unsatisfied however, and want to see an epidemiologic study with evidence that misoprostol decreases unsafe abortion mortality "independently of other major factors such as access to emergency obstetric care." This seems to be an unreasonable expectation given the swamping effect those factors would have on the detection of unsafe abortion mortality, especially if the study plugs in artificially low death rates from abortion.
Conflict of Interest:
Erratum: Equation (2) Women, Regression Coefficient Estimate for (age) squared, Page 4
For equation (2) women, on page 4, the regression coefficient estimate for (age)2 is a typographical error. The correct value is -0.0226. Please note that this typo is limited only to the text on page 4. There is no impact on the tables and figures, which were generated using the correct value of -0.0226.
Conflict of Interest:
Nancy Wojcik is currently employed by ExxonMobil Biomedical Sciences, Inc. Robert Dobie is a paid consultant.