Articles

What’s the right C-section rate? Higher than you think.

Editor’s Note: Dr. Mark Crislip has been kidnapped by anti-vaccinationists. Fortunately, we have sent our black Illuminati, pharma-funded, vaccine-wielding helicopters to rescue him, but unfortunately, as a result of his trauma, his usual Friday post is likely to be delayed either until this afternoon or Saturday. In any case, fortunately for us our latest addition to the SBM crew, Dr. Tuteur, was willing to fill in on short notice; so here she is. Dr. Crislip will post by tomorrow. To whet your appetite for his patented sarcasm, let me just say that he will be having a little fun with a certain article from The Atlantic about flu vaccines. There, now doesn’t that make you want to check back tomorrow to find out what his take is on the article? I thought it would.

Buried in the midst of it new report, Monitoring emergency obstetric care; a handbook, the World Health Organization acknowledges what obstetricians have been saying for some time. The WHO’s goal of a 10-15% C-section rate lacks any empirical basis.

Earlier editions of this handbook set a minimum (5%) and a maximum (15%) acceptable level for caesarean section. Although WHO has recommended since 1985 that the rate not exceed 10–15%, there is no empirical evidence for an optimum percentage or range of percentages …

Of course, they’re not going to give up their recommendation simply because there is no science that supports it, insisting that “a growing body of research that shows a negative effect of high rates.”

Dr. Marsden Wagner, former head of the Perinatal Division of the WHO, appears to be responsible for the purported optimal C-section rate of 10-15%, the level at which both maternal and neonatal mortality rates are supposedly the lowest. Ironically, Dr. Wagner is a co-author of a recent study that actually demonstrates the opposite.

The paper is Rates of caesarean section: analysis of global, regional and national estimates (Paediatric and Perinatal Epidemiology, 2007; 21:98–113.) The article explicitly acknowledges that the 15% C-section rate recommendation was made without any data to support it:

Since publication of the WHO consensus statement in 1985, debate regarding desirable levels of CS has continued; nevertheless, this paper represents the first attempt to provide a global and regional comparative analysis of national rates of caesarean delivery and their ecological correlation with other indicators of reproductive health.

The data regarding C-section rates below 10% is stark:

…[T]he majority of countries with high mortality rates have CS rates well below the recommended range of 10–15%, and in these countries there appears to be a strong ecological association between increasing CS rates and decreasing mortality.

How about the data on C-section rates above 15%? The authors claim:

Interpretation of the relationship between CS rates and mortality in countries with low mortality rates is more ambiguous; nevertheless, the sum total of the evidence presented here supports the hypothesis that, as has been argued previously, when CS rates rise substantially above 15%, risks to reproductive health outcomes may begin to outweigh benefits.

Not exactly. Indeed, not even close. The data show that low maternal mortality and low neonatal mortality are associated almost exclusively with high and very high C-section rates.

The article contains a variety of charts that make this clear. Of note, rather than graphing C-section rates against mortality rates, the authors chose to graph the log (logarithm) of C-section rates against the log of mortality rates. A log-log graph has the advantage of exposing tiny differences when all the values are bunched close together, but all the values are not bunched together in this situation. C-section rates occur along a broad range, and mortality rates occur along a broad range. As a consequence, the log-log graph magnifies the effect of tiny differences and minimizes the effect of large differences. Therefore, you need to be very careful in interpreting the graphs.

This is an adaptation of the chart that appears in the paper comparing C-section rate to maternal mortality (the authors claim that graphing C-section rate against neonatal mortality produces a similar result). The area representing a C-section rate of 10-15% has been highlighted in yellow. The horizontal blue line represents a mortality rate of 15/100,000. Lower mortality rates are below the blue line and higher mortality rates are above the blue line.

The data themselves are quite clear. There are only 2 countries in the world that have C-section rates of less than 15% AND low rates of maternal and neonatal mortality. Those countries are Croatia (14%) and Kuwait (12%). Neither country is noted for the accuracy of its health statistics. In contrast, EVERY other country in the world with a C-section rate of less than 15% has higher than acceptable levels of maternal and neonatal mortality. There nothing ambiguous about that.

The authors claim:

Although below 15% higher CS rates are unambiguously
correlated with lower maternal mortality; above this range, higher CS rates are predominantly correlated with higher maternal mortality.

No, that’s not what it shows at all. It shows that only countries with high C-section rates have low levels of maternal and neonatal mortality. A high C-section rate does not guarantee low maternal and neonatal mortality because C-section rate is not the only factor. For example, Latin America (represented on the chart by open diamonds) has a high rate of C-sections performed for social reasons, but does not have a low level of maternal mortality.

The bottom line is this: The only countries with low rates of maternal and neonatal mortality have HIGH C-section rates (except Croatia and Kuwait). The average C-section rate for countries with low maternal and neonatal mortality is 22%, although rates as high as 36% are consistent with low rates of maternal and neonatal mortality.

The authors’ claims are not supported by their own data. There is simply no support for a C-section rate of 15%, since virtually none of the countries with low rates of maternal and neonatal mortality have a C-section rate of 15% or below, and most have rates that are far higher. There is also no support for the claim that “the sum total of the evidence presented here supports the hypothesis that … when CS rates rise substantially above 15%, risks to reproductive health outcomes may begin to outweigh benefits”. When C-sections are performed for medical indications, there is no evidence that rising C-section rates lead to rising rates of maternal or neonatal mortality.

The authors own data indicate that a C-section rate of 15% is unacceptably low, and that the average should be at least 22%, with rates as high as 36% yielding low levels of maternal and neonatal mortality.

Posted in: Obstetrics & gynecology, Science and Medicine

Leave a Comment (97) ↓

97 thoughts on “What’s the right C-section rate? Higher than you think.

  1. Noadi says:

    That some interesting data but as we skeptics are want to say: correlation does not imply causation.

    It very well could be that countries with c-section rates have better pre-natal and neo-natal health care. The rates are high because in these countries it’s safe to have elective c-sections. While in countries with higher mortality rates c-sections are riskier and therefore done less often.

    This doesn’t mean the WHO is right in their guidelines since they don’t have good data for them but your conclusions also don’t hold up.

  2. provaxmom says:

    From the frying pan into the fire, huh Dr. Amy?

    I think setting some arbitrary goal is not the answer. Nor is it a simple solution as in stating “Doctors perform them too often.” The fact that we are such a litigious society has much to do with it, I’m sure. And I believe I’ve read that countries with lower c/s rates often have socialized medicine, which indeed would mean a woman more likely to seek prenatal care.

  3. Draal says:

    Welcome Dr. Amy. I really appreciate your response rate to your prior post.
    A tiny weeny correction needs to be pointed out: The y-axis of the graph is a incidents per 100,000. so 15 incident per 100,000 is 15/100,000 = 0.00015 *100% = 0.015% not 15%. Imagine a neonatal mortality rate of 22%; 1 out of every 5 babies die.
    ————————
    So the MMR spread is between about 4 and 1050 (0.004-1.05%), which arguably appropriate to represent it as a log scale. The rate of C-section is between 0-100% which can easily fit into a linear scale without distorting the data.

  4. Noadi:

    “That some interesting data but as we skeptics are want to say: correlation does not imply causation.”

    That’s right, and nothing in this paper tells us what the safest C-section rate for the US would be.

    The WHO recommendation has carried tremendous weight among policy makers and “natural” childbirth activists. It is important for everyone to understand that the recommendation was simply made up without any empirical evidence, indeed without any study to find empirical evidence, to support it.

    The WHO recommendation was based on ideology, not on science.

  5. Draal:

    “A tiny weeny correction needs to be pointed out”

    My proofreading skills need improvement. That’s for pointing out the error. I fixed it.

  6. Draal says:

    no problem.

    In regards to higher MMR with higher CS rates above 15%, if you remove the data points from Central and S. America (diamonds), it VERY clearly shows that higher CS rates are associated with lower MMR.
    Donno about you, but I rather not have South America’s health system outcomes dictating N. American and European policies.

    Although the paper attempts to correlate CS rates with economics, they conveniently forgot to compare it to MMR.

  7. TimMills says:

    I am disappointed at the rather sloppy scholarship evidenced by this review. While there are shortcomings to the study, I don’t think you have fairly represented either the authors’ position or the data presented.

    You are right that there are other factors involved in maternal and infant moratlity rates, besides simply the CS rate. In fact, it seems to me (having skimmed the original paper) that the authors acknowledge this too. From the abstract: “The analysis suggests a strong inverse association between CS rates and maternal, infant and neonatal mortality in countries with high mortality levels. There is some suggestion of a direct positive association at lower levels of mortality.”

    They note that the graph which you present shows different tendencies over different ranges – the slope of the regression line changes from negative at one end of the maternal mortality range to positive at the other.

    For countries with very low CS rates, there is a negative correlation: more CS correlates with less maternal and neonatal deaths. For countries whose CS rate is above 15%, there is a positive correlation: more CS correlates with more maternal and infant deaths. This is clearly illustrated in Figure 3 from the paper, which looks just at those countries with CS>15%. (I’ll leave it to you to decide whether to reproduce this figure for your readers who do not have access to the original article.)

    As a research scientist, I have to disagree with your characterization of log-scaling. “A log-log graph has the advantage of exposing tiny differences when all the values are bunched close together”. Log-scaling is generally used to transform data with lots of high outliers into a distribution that is more normal (in the formal, statistical sense of the word). The distributions of datapoints shown in Figure 1 in the paper suggest that log-scaling CS rates is appropriate, because it gives them a more normal distribution. Non-normal distributions render most statistical analyses invalid. They do not present non-log-scaled graphs of mortality rates, so we cannot comment on whether that log-scaling is appropriate or not. In fact, I would prefer that the authors had taken the time to motivate their log-scaling with before-and-after tests of normality. I don’t like having to rely on my own visual interpretation of distributions from graphs like these.

    Overall, I think your critique would be more compelling if you were to acknowledge that there probably is a level beyond which CS rates are likely to be more harmful than beneficial. In the paper, they point out that “According to data from the United Kingdom Confidential Enquiry into Maternal Deaths, an elective CS with no emergency presents a 2.84 times greater chance of maternal death than a vaginal birth, suggesting that, when population CS rates rise beyond medically necessary levels, risks may outweigh benefits.” (p104) There is inevitably some point where the increasing risk of death from surgical complications begins to counterbalance the decreased risk of death from CS-preventable obstetric problems. And it is worth trying to determine where that balance point is, so that you can have some idea if you’ve gone too far.

    The study conducted in this paper uses very blunt measures: national and regional CS rates and maternal and infant mortality. So any conclusions drawn from the correlations they found must be tentative pending more specific and powerful studies. The study is consistent with, but does not necessarily justify, the existing WHO recommendations. (In particular, I suspect that provaxmom is on the nose. Perhaps someone should study litigiousness in socialized vs privatized medicine with respect to CS and birth outcomes.)

    Though there are valid objections to be raised to the paper, I think you have done a poor job of interpreting the data they present, and you are drawing conclusions that are largely unsupported.

  8. Dr Benway says:

    Latin America …has a high rate of C-sections performed for social reasons…

    Such as?

  9. lizdexia says:

    “”Latin America …has a high rate of C-sections performed for social reasons…”

    Such as?”

    Elective C-section has been on an upswing in Latin America – the speculation is that it’s the effect of increasing affluence with some Latin machismo – women have reported that latin american men want their women to have tight vaginas after giving birth.

  10. provaxmom says:

    Oh my. That’s lovely. Guess the gut full of sutures doesn’t bother them?

  11. TimMills:

    “Overall, I think your critique would be more compelling if you were to acknowledge that there probably is a level beyond which CS rates are likely to be more harmful than beneficial. In the paper, they point out that “According to data from the United Kingdom Confidential Enquiry into Maternal Deaths, an elective CS with no emergency presents a 2.84 times greater chance of maternal death than a vaginal birth, suggesting that, when population CS rates rise beyond medically necessary levels, risks may outweigh benefits.”

    I am not suggesting that any C-section rate, no matter how high, is acceptable. There is a point of diminishing returns.

    What I want to highlight, however, is that “findings” in regard to C-sections have been subject to tremendous ideological bias. There is a sense among some that the ideal C-section rate is between 10-15%, but there is simply scientific evidence to support drawing the line at that point.

    There’s actually no good data to show that C-sections increase the risk of maternal death, despite a number of studies that make that claim. It stands to reason, of course, that surgery has more risks than not doing surgery. However, in many situations, a C-section is not optional. The mother or the baby or both would die without one.

    Moreover, the risk factors for C-section, such as pre-existing medical conditions, pregnancy complications, maternal hemorrhage, pre-eclampsia complicated by DIC, advanced maternal age and multiple pregnancy among others, are all risk factors for maternal death. Therefore, the relationship between C-section and maternal death is rather difficult to tease apart.

  12. Kylara says:

    I have a related question about C-sections and the practice around them, which perhaps you or someone else would be interested in taking up at some point.

    I gave birth by C-section a little while back (my spawn had turned breech in such a way that even “trial labor” was contraindictated, and when they got in there they found him so stuck in my pelvis they had a heck of a time pulling him out).

    Anyway, what I was interested in was — two of my female friends about the same age (early 30s) and the same health (good) had abdominal surgery at around the same time I did. One was a hysterectomy; the other was … I forget what, exactly, but pretty “minor.” All were incision, not laproscopic.

    Both of them were given very slow recovery plans, resting and not moving around a whole lot for at least six weeks, reminded repeatedly they’d just had major abdominal surgery and needed to rest quietly, eat boring diets, etc. (One even got longer leave from work, since “abdominal surgery” was more serious than “having a baby (via abdominal surgery)”!)

    I, on the other hand, was urged to get out of bed and start walking around as soon as the catheter was removed, and while I was urged to heal quickly … all while breastfeeding an insatiable baby ever 2 hours, which puts significant demand on one’s system.

    I wondered why there was such a difference in our post-surgical orders. I find it hard to believe that after a significantly larger incision, while recovering from pregnancy and the crazy hormones and all the body changes, while breastfeeding constantly and of course having to lift the baby, that I was magically recovering faster than my friends who had much smaller incisions, weren’t recovering from pregnancy, and didn’t have baby care thrust upon them. (Not that I at all belittle what they went through and, after all, surgery isn’t fun, but at least my surgery ended with a baby.) We all thought it was strange that six weeks later I was supposed to be basically recovered while they were supposed to be just starting to move around more actively and feel like themselves.

    Is it just practicality, that a new mother CAN’T rest that much for six weeks? Or is it an attitude towards women’s medicine that childbearing pain is unimportant that carries over to C-sections? Or is it just a total lack of communication between obstetrical surgery and other abdominal surgery on recovery techniques? It was so odd.

  13. windriven says:

    @Kylara

    I am not a physician but can tell you that the approach to recovery has changed dramatically over the years. Today the trend seems to be to get patients up and moving as soon as possible after surgery – even major surgery. I am told that this reduces the likelihood of a variety of complications not least of which is pneumonia.

    I wonder if the difference in post-op instructions that you ask about is simply related to the standing orders of the various physicians managing each case.

  14. Harriet Hall says:

    The statistics would be more meaningful if we could separate out C-sections done for medical indications from those done for convenience or at patient request. Also if we could compare different specific medical indications.

    Of course, the real question is not the C-section rate, but whether an individual C-section is likely to do more harm than good.

    I wonder how much the US C-section rate is influenced by lawyers? You don’t want to be on a witness stand explaining why you failed to do a C-section in a case with poor outcome. I’m not suggesting that obstetricians consciously consider this, but there is a climate favoring the precautionary principle.

  15. daedalus2u says:

    Kylara, all that nest-making and nurturing that new mothers do causes the release of oxytocin which accelerates wound healing.

    http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0005523

    This response is very likely mediated through nitric oxide. Isolation reared rats have reduced numbers of NO releasing neurons, and so are likely more sensitive to the effects of NO on stress.

    Inhibiting nitric oxide synthase in ewes blocks maternal bonding; doing things that raise NO would likely increase bonding and also accelerate healing.

    I think if there was a technique to raise NO levels, that your friends would heal as quicklly as you did.

  16. David Gorski says:

    As a research scientist, I have to disagree with your characterization of log-scaling. “A log-log graph has the advantage of exposing tiny differences when all the values are bunched close together”. Log-scaling is generally used to transform data with lots of high outliers into a distribution that is more normal (in the formal, statistical sense of the word). The distributions of datapoints shown in Figure 1 in the paper suggest that log-scaling CS rates is appropriate, because it gives them a more normal distribution. Non-normal distributions render most statistical analyses invalid. They do not present non-log-scaled graphs of mortality rates, so we cannot comment on whether that log-scaling is appropriate or not. In fact, I would prefer that the authors had taken the time to motivate their log-scaling with before-and-after tests of normality. I don’t like having to rely on my own visual interpretation of distributions from graphs like these.

    Being a research scientists as well, in addition to being a surgeon, I’m going to have to agree with Tim here, I’m afraid. The reason to do log transformations is to make data fit a normal distribution better or to linearize data, not to “magnify the effect of tiny differences and minimize the effect of large differences.” It is particularly useful when data varies more than two orders of magnitude in order to induce symmetry in the data. It’s hard to tell if log scaling was appropriate in this case, but, whether it was or not, the reason is to provide a more normal distribution upon which to apply statistical tests.

    BTW, I note that you appear to have changed your mind:

    http://www.kevinmd.com/blog/2006/03/kate-steadman-thinks-defensive.html#comment-60172

    Is this study what led you to do so?

    Welcome to the SBM collective. I do tend to challenge even our own SBM bloggers from time to time, just to keep ‘em on their toes. :-)

  17. Harriet Hall:

    “I wonder how much the US C-section rate is influenced by lawyers?”

    It’s certainly influenced by fear of malpractice litigation, but that is compounded by the high false positive rate of electronic fetal monitoring.

    If you are looking at a reassuring fetal tracing, you know everything is fine. However, if you are looking at a non-reassuring tracing, you don’t know exactly what is going on. At that point, you have two choices, hope that the baby is okay and wait to see what happens next, or deliver the baby by C-section and insure that it will not suffer anoxic brain damage. When the baby’s life or brain function is on the line and there is doubt, the doubt is usually resolved in favor of delivering the baby.

  18. lkw says:

    I do not have access to the paper. What regression technique was used to draw the line on the graph?

  19. daedalus2u says:

    Dr Gorski, I am sure you noticed that both references that Dr Tuteur cited are dated after the comment that you linked to. Revising an opinion in the light of new data is to be encouraged.

    I think the data on the correlation between CC rate and MMR are too sparse to draw firm conclusions as to treatment for an individual patient.

  20. David Gorski says:

    Dr Gorski, I am sure you noticed that both references that Dr Tuteur cited are dated after the comment that you linked to. Revising an opinion in the light of new data is to be encouraged.

    Your point?

    And do you really think I didn’t notice that the comment was a couple of years old? That’s the reason I asked if this study is what swayed Amy to change her mind, not as any sort of “gotcha!” moment. It’s just me being my curmudgeonly but lovable self.

  21. David Gorski:

    “The reason to do log transformations is to make data fit a normal distribution better, not to “magnify the effect of tiny differences and minimize the effect of large differences.”

    I am a very cynical person, and I know that the people who wrote the paper wanted very much to show that C-sections are dangerous. It was Marsden Wagner himself, in his time at the WHO, who pushed through the 10-15% recommendation and he speaks and writes widely to promote that rate.

    I graphed the data on a non-log scale to see what it looked like. It’s not nearly so impressive looking when it is graphed like that. I suspect (and I’m willing to admit that I am prejudiced in this regard) that the authors liked the way the log-log scale looked and that’s why they used it.

    “I note that you appear to have changed your mind”

    I like to say that my thinking on the subject is “evolving.” That sounds better, doesn’t it?

    Actually, I “feel” that the C-section rate should be in the range of 15%, but I have to admit that there is no data to support that contention. Furthermore, the ACOG position on VBAC (vaginal birth after cesarean) has made it almost impossible to allow women to attempt a VBAC. That has forced the C-section rate up considerably.

  22. David Gorski says:

    I graphed the data on a non-log scale to see what it looked like. It’s not nearly so impressive looking when it is graphed like that. I suspect (and I’m willing to admit that I am prejudiced in this regard) that the authors liked the way the log-log scale looked and that’s why they used it.

    Fair enough as far as it goes for the study in question, but Tim was correct in noting that you misstated the purpose of log transformation in making your suspicion known, in the process appearing to impose your bias about these researchers onto a legitimate practice in statistical analysis in general. In brief, your original language was…imprecise.

  23. storkdok says:

    @Kylara

    Congratulations on the birth of your baby!

    I am wondering if the difference in recommendations might have been from different OB/GYNs for each of you? My personal recommendations for my patients have been more consistent with any abdominal surgery with a large incision, including C/S. I always told my patients it takes a minimum of 2 months to recover most function and energy, but it is harder with a baby and the sleeplessness that goes with it. I did recommend taking 2 months off after C/S if possible. No heavy lifting (>15lbs), which can be hard if your baby is big :0) I did crack the whip and get all my patients moving after surgery so that they decreased the risks of deep vein thrombosis and pneumonia, but I did not advocate exercising until after at least 6 weeks in either case, and then to start light with easy walking.

    Having had a C/S twice, I was glad to have maternity leave the first time of almost 3 months, since my baby decided to come 3 weeks early, and I really didn’t have the energy to get back into full time OB/GYN schedule at 6 weeks. The extra month was great!

    I also advise my new moms to concentrate on just what needs to be done for them and their baby, not try to clean the house from top to bottom and do a lot of extras. The best things family and friends can do is to offer to help out with housework, laundry or cooking, so you can concentrate and enjoy your baby.

  24. Robin says:

    You guys have vaccine wielding helicopters? The next time I get in a woo argument, I’m calling your dispatcher. Thanks for the laugh!

  25. SF Mom and Scientist says:

    Thank you for this post. I have always wondered about that 10% to 15% ideal C-section rate. I always hear people say we need to lower this rate, yet I haven’t seen good examples of instances where C-section should have been avoided. I do hear general comments about doctors scheduling C-sections for their own convenience, but I have yet to hear an actual story of this happening.

    To Harriet Hall – from what I understand (and I admit I my numbers could be way off) the number of women who actually elect to have a C-section that has no medical indication is very low. Depending on where I look, it seems to be around 1% to 3% of all births. Since C-sections in the US are overall around 30%, I would think that eliminating elective C-sections would not change the results all that much. If someone would like to correct me on this, please do.

  26. Dr Benway says:

    …women have reported that latin american men want their women to have tight vaginas after giving birth.

    Wow, Latin America. You are one twisted mofo.

  27. Basiorana says:

    I’ve read the Latin America demand for elective c-sections also has to do with fears of incontinence and sexual dissatisfaction afterwards, as well as concern about the pain levels. Many elective C-sections are for second-time moms who had a bad experience the first time around.

    Also, by our standards Latin American hospitals are atrocious to mothers. Dunno if it’s gotten better, but I’ve read horror stories where it’s basically a row of drugged-up women strapped to bed with curtains between them, and the doctors and nurses wandering by with no privacy. With that scenario as the alternative, I would probably choose a simple C-section too.

  28. Prekky says:

    One thing to consider when looking at C-section rates is not only the outcome of the current pregnancy, but all future pregnancies. The scar tissue left by having a C-section increases the chance of placenta previa (necessitating another C-section), placenta accreta (greatly increasing the risk of maternal hemorrhage), and uterine rupture.

    I feel any doctor considering a C-section for a patient should strongly take into account her future maternity plans.

  29. sowellfan says:

    @Kylara:

    My wife had a total hysterectomy last year, when she was about 35 years old, and her doctor definitely encouraged her to walk around a lot. If I recall correctly, she was doing laps of that particular floor the day after surgery, albeit slowly. Now, she had restrictions on how much she could lift, and on doing exercises in a gym, etc., but the doc definitely encouraged her to be active. On the other hand, she was one of the only people on that Ob/Gyn floor to be moving around anywhere close to that much (her doc did notice, and he was quite happy with her).

  30. SF Mom and Scientist:

    “the number of women who actually elect to have a C-section that has no medical indication is very low.”

    It is extremely low. Despite that, it has generated controversy among “natural” childbirth advocates.

    The same people who believe strongly in choice of birth place, refusing interventions, etc. are completely opposed to C-section for maternal request. Since this is difficult to reconcile with choice of homebirth they are reduced to arguing that women who request C-section are suffering from a form of false consciousness. They “think” they want a C-section, but they have been led astray by social conditioning and therefore, the requests are not “authentic” and should not be honored.

  31. marilynmann says:

    I thought Amy was simply making the point that there is no empirical support for the WHO 10-15 percent goal for C-sections. I feel that some of the commenters have read more into her post than she intended.

    In addition, by reason of my history of posting references to peer-reviewed research on SBM, I feel I have earned the right to tell a personal anecdote. Namely, that I had the worst of both worlds. I went all the way through labor, including pushing for hours with the epidural turned completely off, until it became totally clear that my daughter was “stuck” and was never going to come out. Only the top of her head was showing, so forceps were not an option. Hence, both of us would have died without a C-section.

    In addition, when the anesthesiologist juiced up my epidural to get me ready for the C-section, there was a little problem. Namely, that when my OB starting cutting, I could FEEL EVERYTHING. I started screaming and the anesthesiologist came running and said “You have pain? I put you to sleep.” The next thing I remember is waking up on a gurney and the nurse saying “I *told* them you weren’t ready.”

    My C-section obviously was medically necessary, so why did I have a feeling for a long time afterward that I had somehow failed? I think it is partly because childbirth classes, or at least the one my husband and I attended, do not get you ready for what might not go according to plan.

  32. TimMills says:

    @lkw: They used lowess to provide a locally-weighted regression line that can have different slopes at different parts of the graph.

    @Amy Tuteur: For the record, I don’t think the data were presented well in the paper, and if I were a reviewer I would have asked that they change the following things before publishing:

    1. As I mention above, they don’t actually demonstrate that log-scaling is necessary. Though the distributions in their first graph suggest to me that they made the right choice, my intuition doesn’t justify their omission: when making a quantitative inference, you must make sure that it’s backed up.

    2. The graph that you present in your post is presented “on its side” in the paper – they have CS rate on the vertical axis and maternal mortality on the horizontal. This may, in fact, be a terminal fault – depending on the (under-reported) details of how exactly they did the statistics. Generally, you put your independent (manipulated) variable on the horizontal and the dependent on the vertical. Though they did not manipulate anything (they’re simply looking for correlations), their conclusions do suggest recommendations that do (or could) directly affect CS rates. I’m not inclined to get into an extensive statistics lecture here, but I am surprised that you didn’t mention this as a red flag.

    You are right that this paper has shortcomings. But in your review of it, (a) you missed the actual statistical gotchas that were there, and (b) key points you did raise were off the mark, as far as the actual content of the article goes (see my points in an earlier comment on log-log scaling and the discussion of Figure 3 in the paper).

  33. TimMills,

    One of the reasons I jumped at the chance to be part of SBM is because it is an opportunity to present my views to a very sophisticated audience … and, inevitably, have my posts torn apart and all flaws identified.

    Your admonitions are the first, but I’m sure not the last. Thanks for taking the time to set me straight.

    I hope, at least, that the premise of the post survived unscathed: this paper fails to provide an empirical basis for claiming that an ideal C-section rate is 10-15%.

  34. TimMills says:

    Amy, that’s very gracious of you, given my admonitions.

    Yes, I agree: this paper fails to provide and empirical basis for claiming that an ideal C-section rate is 10-15%.

  35. Mark P says:

    Perhaps someone should study litigiousness in socialized vs privatized medicine with respect to CS and birth outcomes.

    New Zealand has effectively no medical malpractice suits. While not strictly speaking illegal, you cannot win any damages, so they are almost never pressed.

    We also have a high C-section rate. I suspect it is because we can afford it, and don’t like to see unnecessary suffering.

    Like another commenter, I hear a lot about “unnecessary” C-sections, but have yet to meet anyone who has had one, other than women choosing to have a second or third baby which they know will likely be by C-section because an earlier one was.

  36. momkat says:

    As an L&D nurse for over 20 years. I’ve seen a lot of changes in the practice of obstetrics, some good, some not so good. It is becoming more commonplace in my area (the deep south) to have women scheduling primary C/Ss purely for convenience. My favorite reason is “I don’t want to hurt through labor!” They seem to think the following few weeks are going to be painfree as well. I am also a women’s health nurse practitioner so I get to follow these same women later in life when they have health complications because of those major abdomenal surgeries. All of this is really tangental to the topic of your post, but thanks for letting me put this observation on paper, so to speak. I look forward to your future posts.

  37. DownWithWoo says:

    re “unnecessary” CS….

    Purely anecdotal, but here are some numbers of babies born in my family:

    From my grandparent’s generation: 14 babies born (0 deaths, 0 birth-injured, 0 born by CS or other instrumental delivery)

    From my parent’s generation: 11 babies born (0 deaths, 0 birth-injured, 0 born by CS or other instrumental delivery)

    From my generation (so far): 7 babies born (0 deaths, 0 birth-injured, 5 born by CS)

    All 5 CS were “emergency” CS (an emergency CS is any unplanned CS). Interestingly, the 2 that were born vaginally were both planned homebirths (1 born after non-emergency transfer to hospital after failing to progress, 1 born at home). Two of the emergency CS were for fetal distress following routine inductions at 40+7. In the 2 previous generations there was only one induction done (at 42+3).

    Were all of these caesareans necessary, or did they become necessary as a consequence of changing hospital birth culture and practices? The same question could be asked where the population at large sees a rising caesarean rate, if adverse outcomes aren’t reduced as a consequence.

    The response that I’ve heard from the OBGYN community is that caesareans are safer now, so it’s not a big deal that we have more of them. It is a big deal to have major abdominal surgery requiring two months recovery. It also increases risks in future pregnancies.

  38. Coqui says:

    Miami’s C-section rate is about 50%, an estimate from the admittedly not-always-accurate Miami Herald. In comparison, arguing over the WHO’s 15% goal seems … like a minor quibble. Even irrelevant.

    I suppose there is probably no evidence of increased mortality with a 50% C-section rate. Now, I’m all for science-based medicine – don’t get me wrong – but I think 50% is just too damn high. It’s not just a question of mortality, it’s an increase in costs (for the patient as well as insurance), recovery time, time lost from work, time away from the baby in post-op, future childbearing….

    As to why we have such a high rate, there is a culture of elective C-sections here, driven in part by recent arrivals from certain South American countries (Brazil, Venezuela, Colombia). I don’t know about this hypothesis of tight vaginas and machismo. I would say instead that vaginal deliveries are simply viewed as barbaric and disgusting. But there also seems to be a culture of “elective” C-sections among OB/GYN doctors that seems to go beyond fear of litigation. It’s almost considered the norm now. Why would anyone bother with a vaginal delivery?

  39. katie b says:

    I’ve had two C-sections. They both went well, and I’m sure if I hadn’t had them, either I, or the baby (or both), might have died. I was encouraged to get up and about as soon as possible afterwards, and it was all fine.
    I’ve occasionally met people who seem to think the C-sections are the work of the devil, and are unnecessary, etc, but I have 2 friends whose babies died as a result of very traumatic deliveries, when in both cases a C-section would have saved them.

  40. DownWithWoo:

    “From my generation (so far): 7 babies born (0 deaths, 0 birth-injured, 5 born by CS)”

    A timely C-section should always yield a healthy baby. So it is impossible to determine whether a C-section is necessary (even an emergency C-section) by the fact that the baby is healthy. Indeed, a C-section that yields a brain damaged or dead baby is a C-section that should have been done sooner.

  41. TimMills says:

    “A timely C-section should always yield a healthy baby.”
    “A C-section that yields a brain damaged or dead baby is a C-section that should have been done sooner.”

    Amy, you seem to be in danger of setting up a non-disprovable (and therefore scientifically useless) attitude toward C-sections. I suspect this impression, on my part, is inaccurate: perhaps you have simply been imprecise in your language.

    What sort of evidence would convince you that either (a) a specific C-section is/was unnecessary or (b) an overall (regional or national) C-section rate is too high?

  42. TimMills:

    “What sort of evidence would convince you that either (a) a specific C-section is/was unnecessary or (b) an overall (regional or national) C-section rate is too high?”

    A C-section is unnecessary if it is done without an appropriate medical indication. Many C-sections done for medical indications may turn out to be unnecessary in retrospect, but that does not mean that the C-section itself was unnecessary.

    C-sections are a form of preventive medicine. When we see a baby is in trouble, we get the baby out. Could the baby have tolerated more labor? Maybe, maybe not. There’s no way to know. One could easily claim that the millions of negative Pap smears done each year are unnecessary in retrospect, but that doesn’t mean that they shouldn’t have been done, either.

    There is no question that when it comes to regional or national rates, there is a point of diminishing returns. If, for example, the maternal death rate begins to rise at a certain point, the C-section rate would be unjustified.

    Any birth has risk. That is inescapable. Childbirth is inherently dangerous. The only question is who is going to carry the risk. In a vaginal delivery, the baby carries most of the risk. In a C-section, the mother takes most of the risk on herself. Most women prefer to take the risk onto themselves.

    The solution, of course, is to improve our diagnostic abilities and that will involve new technologies, not less intervention.

  43. lkw says:

    Amy Tuteur, MD: “There’s no way to know.”

    Isn’t figuring out how “to know” exactly what science-based medicine is about?

  44. lkw:

    “Isn’t figuring out how “to know” exactly what science-based medicine is about?”

    Of course. We have a much better idea of which babies are in trouble than we ever did, but our methods are not perfect.

    “Natural” childbirth advocates prefer to believe that the rising C-section rate reflects benefits to doctors, whether in money or convenience. That’s not why the C-section rate has gone up. It has gone up because American society has demanded it.

    Our society expresses some of its most important values through the legal system. One of the values we appear to hold dear is that every baby should be born healthy with its full intellectual potential intact. As a corollary of this value, we also appear to believe that when a baby is injured or dies during birth, the doctor presiding over the birth should bear personal responsibility for that outcome.

    The only effective legal defense it to show that you did a C-section and you did it as soon as there was any evidence that the baby might be compromised. Obstetricians are well aware that our diagnostic technologies are imperfect. The courts don’t care about that. As far as the court system is concerned, any evidence of fetal compromise necessitates an immediate C-section. If the obstetrician doesn’t do the C-section, the obstetrician is punished.

    If women want to lower the C-section rate, step 1 is to look in the mirror.

  45. lkw says:

    I’m sorry, I thought we were talking about science-based medicine. Apparently, we’re talking about the law.

  46. Harriet Hall says:

    No we’re not just talking about the law. We’re talking about trying to do what is best for each individual baby. Science-based medicine can tell us that X% of babies in a given situation will do better with a C-section; it can’t say which ones. We don’t have a crystal ball. When we don’t know ahead of time which baby will suffer from avoiding a C-section, prudence demands that we do one. The law reflects that.

  47. lkw says:

    “We don’t have a crystal ball. When we don’t know ahead of time which baby will suffer from avoiding a C-section.”

    That’s exactly my point. We don’t need a crystal ball. We need science.

  48. wertys says:

    I am very pleased to have read a post which tends to confirm one of my pet biases ! I have heard the 5% figure quoted dogmatically by many midwives and some obstetricians but I have never been able to find any reasonable evidence on which this might be based. It is virtually an article of faith for registration as a midwife in Australia to believe that too many CS’s are done, but nobody can say how many would be right, or even come up with some matrix of variables into which data for a region or health service could be fed that would generate an acceptable range. There is even a strong movement within our Government-funded state hospitals for funding to be linked to such an ‘acceptable’ range of CS being performed…..all without any evidence to support such belief.

  49. lkw:

    “We need science.”

    We have science; we don’t have perfect science. I’m not really sure why you would think that we did.

    In the past 100 years neonatal mortality has dropped 90% and maternal mortality has dropped 99%! Prior to fetal monitoring, many babies died during labor. Now intrapartum mortality is rare. Monitoring and C-sections have made the difference. Have unnecessary C-sections been performed? Of course.

    There are only two choices available: unnecessary C-sections or preventable neonatal deaths. Which do you prefer?

  50. Harriet Hall says:

    Ikw,
    “We need science”

    Science tells us that a certain percentage of patients with pneumonia will recover without antibiotics. It can’t tell us which ones. Science tells us how many radioactive atoms will decay in a given time period, but it doesn’t tell us which ones. Predictions have limits due to chance, probability, chaos theory, etc.

  51. lkw says:

    It just sounds like you’re throwing in the towel. “It can’t tell us which ones.” Shouldn’t that say “yet” or have some qualifier to it? These are the answers which we are pursuing but do not yet have. That’s the science. The pursuit. Not the answers.

  52. lkw says:

    By the way, comparing the inability to perfectly predict the need for a c-section (yet) with radioactive decay is not accurate. Radioactive decay is governed by an inherently random process at the quantum level. I think science can work out a few more reliable indicators for c-section and antibiotic treatment for pneumonia before the Heisenberg uncertainty principle comes into play!

  53. Pattoye says:

    lkw:
    I think you’re missing the point. Of course we should (and almost undoubtedly will) be able to determine how necessary a c-section is with greater accuracy.
    To imply that we should be able to “know” ahead of time with certainty whether or not it is necessary is contrary to how both science and medicine work. In nearly every aspect of medicine there is a significant element of chance because we’re dealing with a very large number of variables that are difficult or impossible to measure. Bodies are complicated.
    Even if we could say “There is a 99% chance a baby in this condition will die without a c-section,” there is no way to determine in retrospect if the baby would have been that 1 in 100 that would have lived and it doesn’t mean that doing it for 100% of babies with that condition means that 1% of them were prospectively “unnecessary,”

  54. Calli Arcale says:

    It just sounds like you’re throwing in the towel.

    No, she’s not throwing in the towel. It’s just that although someday we may have the ability to judge which babies need to be born by c-section, we don’t now, and they’re not going to wait for the science to come in. Babies don’t wait for much of anything, as all obstetricians are painfully aware. So you have to make decisions in the meantime. Dr Tuteur is not a research scientist; she is an obstetrician. So she’s speaking from that in-the-trenches perspective, not about what things should, in a perfect world, be like, but about what they *are* like in this world, now.

  55. MOI says:

    Dr. Tuteur said – “Prior to fetal monitoring, many babies died during labor.”

    Are you talking about electronic fetal monitoring? I’ve heard many state that EFM has not done much in preventing deaths. Many feel it is unnecessary. You have stated that the issue with fetal monitoring is the number of false positives. I understand the need to err on the side of caution but I’m wondering if there are stats available that can accurately address this question. How does one determine if the false positive “really needed the cs”?

    “Continuous cardiotocography during labour is associated with a reduction in neonatal seizures, but no significant differences in cerebral palsy, infant mortality or other standard measures of neonatal well-being. However, continuous cardiotocography was associated with an increase in caesarean sections and instrumental vaginal births. The real challenge is how best to convey this uncertainty to women to enable them to make an informed choice without compromising the normality of labour.”
    http://www.cochrane.org/reviews/en/ab006066.html

    It’s not the best but it’s what I found.

  56. Harriet Hall says:

    A better analogy might be with weather forecasting. We can improve our accuracy but it will never be perfect. Science has its limits.

  57. MOI:

    “Are you talking about electronic fetal monitoring? I’ve heard many state that EFM has not done much in preventing deaths. Many feel it is unnecessary.”

    This is a common misconception (pardon the pun). EFM is no better than rigorously done intermittent fetal monitoring. Indeed, the Cochrane review you linked to is about comparing EFM with intermittent auscultation (either using a doppler or a fetoscope). The protocol for auscultation is quite rigorous and requires one on one nursing care for the bulk of labor. That is simply not possible in today’s economic climate.

    The real problem is that the method is indirect at best. Imagine if the only way a doctor could assess your health was by listening to your heartbeat. You could be close to death from any of a variety of causes before your heart rate would change in response.

    What we really want to know is how well the baby’s blood is oxygenated. A lot of work has been done on a fetal O2 saturation monitor (similar to the clip that you wear on your fingernail while undergoing anesthesia). Everyone had high hopes for it, but the first large scale clinical trial was thoroughly disappointing.

  58. lkw says:

    “Dr Tuteur is not a research scientist; she is an obstetrician. So she’s speaking from that in-the-trenches perspective, not about what things should, in a perfect world, be like, but about what they *are* like in this world, now.”

    Fair enough, we’re talking about the application of current scientific knowledge in medical practice, not about the ongoing research. I just wish that people would stop saying things like “there’s no way to know” and “it can’t tell us which ones” … if that were the case, then there would be no need to fund further research. Science *can* answer some of these questions and *can* tell us which ones. And we need money for it ;-)

    I’d like to see more emphasis on the continuing refinement of the knowledge through research and the potential for continuing improvements in the applied science. Maybe I’m looking for more *hopeful* words. And definitely not for law discussions ;-). Surely there are other blogs for that.

  59. lkw says:

    Weather forecasting *is* a much better example of a complex system that we don’t yet understand, but have a need to predict. And it’s even better in the sense that it is one that *everyone* has experienced it.

  60. AlexisT says:

    The Miami Herald linked the high CS rate there to the fact that (supposedly) most South Florida OBs are practicing without malpractice insurance.

    I have known of unnecessary CS–several for “big babies” where ultrasound estimated the baby to be in the 9 lb range. I’ve also heard of ones where the OB called “failure to progress” after only a few hours.

    The “Vaginal Preservation Society” issue could be viewed as consumer choice, but there are a few OBs telling their patients they should really consider CS for this reason, which is another matter.

  61. Plonit says:

    The protocol for auscultation is quite rigorous and requires one on one nursing care for the bulk of labor. That is simply not possible in today’s economic climate.

    +++++++++++++

    That statement has nothing to do with science-based medicine.

    CEFM using CTG provides a *recording* of fetal heartrate and uterine activity. It cannot *interpret* the trace in clinical context (not even the fancy computerised CTGs which have come in since Dr Amy was in practice). Nor can it make decisions or plan care based on that interpretation. To do that requires regular nursing care, even 1-2-1 nursing care if you factor in the other needs of labouring women beyond interpretation of the CTG trace.

    When looking at studies comparing cEFM with intermittent auscultation with doppler, it is important to know how frequently the trace was being looked at and systematically interpreted. If normal practice varies considerably, then the ‘real world’ results for EFM are likely to be very different (possibly worse) than those under trial conditions. As it goes, several of the largest studies of cEFM stipulated one-to-one care in active labour in the trial protocol.

    It is not true to say that “[c]EFM is no better than rigorously done intermittent fetal monitoring.” It would be more accurate to say that a policy of routine use of cEFM is worse than IA, since it offers no proven benefits, but is associated with increased rates of instrumental and caesarean delivery, procedures which have certain risks.

  62. “To do that requires regular nursing care, even 1-2-1 nursing care if you factor in the other needs of labouring women beyond interpretation of the CTG trace.”

    No, it doesn’t. It only requires one person sitting at a central monitoring station, watching multiple tracings at once.

    To my knowledge, there has been no study comparing EFM to intermittent auscultation under real world conditions. The only studies compare EFM to special auscultation procedures used only in research.

    The issue has more to do with the nature of the methods than anything else. EFM has a very low false negative rate and a high false positive rate. Intermittent auscultation has a high false negative rate and a high false positive rate, is profoundly dependent on attendant and technique and provides no permanent record for later analysis.

    Because EFM has a high false positive rate, and because obstetricians have a low threshold for acting on negative results, the C-section rate rises. However, the intrapartum death rate is almost non-existent. It simply never happens that a baby drops into the obstetricians hands already dead.

    On the other hand, homebirth midwives, who by their very nature provide 1 on 1 care, and use intermittent auscultation (supposedly rigorously) have a dramatically higher rate of intrapartum death. Homebirth midwives may have no idea that a baby is in trouble until it is born dead.

    It is far too late to abandon EFM. It is both the medical and the legal standard. What we need is a better method of evaluating fetal oxygenation.

    C-section activists ignore the complexities of the issue by making blanket statements that too many “unnecessary” C-sections are being done. At the same time they offer no alternative to the existing situation.

    Everybody knows too many unnecessary (in retrospect) C-sections are being done. No one knows in advance which ones they are. Until C-section activists can offer a reliable method for predicting which C-sections are necessary in advance (a method with a low false positive rate and a low false negative rate) they bring nothing useful to the table. One really has to wonder whether they want to solve the problem, or merely complain about it.

  63. Emma B says:

    lkw, some of the birth problems we’re discussing are actually really technically difficult, and solutions belong to the realm of Star Trek-style science fiction medicine.

    Consider shoulder dystocia, where the baby’s shoulders get stuck on the mother’s pubic bone on the way out, and the umbilical cord gets compressed the way a hose does under a car tire. It’s one of the most frightening delivery complications and major causes of intrapartum deaths, because by the time it begins to happen, it’s too late to do a CS to fix it — the baby’s head is already out, and you’ve got two to four minutes before the baby is irreversibly brain-damaged.

    We know that there are a couple of risk factors for shoulder dystocia: it’s more likely to happen to large babies, or to mothers with very narrow pelves, or to women who have had previous shoulder dystocia. However, most such pregnancies will be perfectly fine; for example, even among 10-lb babies, only 5-10% will have SD. If you preemptively CSed all 10-lb babies (and if you could accurately identify all 10-lb babies, which we cannot currently do), 90% of those CS would be unnecessary — but you would never know which infants would have had SD, because you couldn’t have predicted how the baby would have positioned itself over the course of labor.

    To really, truly predict an individual’s likelihood of SD, you would need to be able to get some very accurate imaging of the fetal skull and shoulders, and of the mother’s pelvic outlet, via something like CT scan/MRI. Further, you would have to do it during the pushing phase of labor, since the baby’s position changes as it begins to move down the birth canal (it kind of corkscrews), but before the shoulder reaches the pubic bone. For that matter, you would have to know when the shoulder is approaching the pubic bone, which isn’t something you can see or palpate, so we’re probably talking about continuously scanning the mother. You would also need to monitor the mother’s position continuously, since the pelvic contours change slightly as she moves around, and mothers aren’t noted for remaining still while pusing out babies.

    Then, you would have to feed in all that data and run a computer simulation which models all the possible ways that baby could progress from its current position. It would be on the order of millions of scenarios, given the large number of variables at work, and the number-crunching would need to be constantly updated with each contraction and maternal position shift. That’s how computers play chess, by looking at the convergence of probabilities, and that’s how SD modeling would have to work. You would then count up how many of those paths resulted in SD, and if the number reached some predetermined threshold, you would need to conduct an IMMEDIATE CS, because labor can move very quickly at the end, and the baby would pass the point of no return. You couldn’t transfer to the OR or wait for anesthesia to arrive, and you’d need to be making the incision in no more than two or three minutes after making the decision to CS, before the mother has another contraction.

    So, to review, you’d need to develop: 1) continuous, unobtrusive CT scanning of women in late-stage labor; 2) enough computing power (likely involving a distributed cluster of many computers) to perform millions of simulations in a matter of seconds; 3) nearly instantaneous anesthesia and surgery prep in the delivery room. And STILL, you wouldn’t know with certainty that the baby would have suffered a SD — you just would have known that the probability of SD is likelier than not. Even with all that mythical technology, you’d still be performing some “unnecessary” CSes on women who would have successfully delivered vaginally, and you still wouldn’t be able to identify them in retrospect.

    Science and math are necessarily non-deterministic. The best they’ll ever be able to do is to make more or less accurate generalized observations about groups, and they cannot tell us what an individual’s clinical course will be. You can refine those observations to some degree, but 100% accuracy is mathematically impossible. It’s extremely improbable that you’ll get anywhere close to it, even with thousands of person-years of engineering.

  64. Plonit says:

    Intermittent auscultation has a high false negative rate and a high false positive rate, is profoundly dependent on attendant and technique and provides no permanent record for later analysis.

    +++++++++

    If that were the case, then you would expect studies to show IA to have worse outcomes than cEFM. Given that studies do not show that, I’m not sure what the evidence base is for your assertion.

  65. Plonit says:

    The above comment refers to your assertion about high false negative rate.

    cEFM is also profoundly dependent on attendant (in fact, the systematic interpretation of cEFM is a skill that needs to be learnt). Listening in is a clinical skill – but then we expect midwives/nurses/doctors to learn and practice clinical skills, so why not this one. The permanent record is the documentation of fetal heart rate heard, and is no different in essence from any other observation that is made clinically and then documented.

  66. Plonit says:

    On the other hand, homebirth midwives, who by their very nature provide 1 on 1 care, and use intermittent auscultation (supposedly rigorously) have a dramatically higher rate of intrapartum death. Homebirth midwives may have no idea that a baby is in trouble until it is born dead.

    +++++++

    Again with the sloppy thinking. All women who go into spontaneous labour at low-risk of complications receive intermittent auscultation in the UK. The standard of care is no different at a home birth and in the hospital.

  67. Plonit says:

    It is far too late to abandon EFM. It is both the medical and the legal standard. … One really has to wonder whether they want to solve the problem, or merely complain about it.

    ++++++++++++

    Is this blog called ‘science-based medicine’? Surely it is never to late to change standard of care in the direction that is dictated by the science. In the UK changes were made in practice to massively restrict the use of cEFM in response to the scientific evidence. Everyone knows the problems with cEFM, the issue is whether you plan to solve the problem (by providing the resources and context that support IA) or whether you merely want to shrug your shoulders and resign yourself to greater harm to women with no corresponding benefit for babies.

  68. Plonit:

    “Is this blog called ’science-based medicine’?”

    And that’s exactly what has been presented.

    1. All the existing scientific literature to date shows fetal monitoring (whether by EFM or auscultation) improves neonatal outcomes.

    2. The existing scientific evidence shows that EFM does not improves outcomes beyond rigorous protocols for auscultation requiring one on one nursing care.

    Since one on one nursing care is simply not possible given current economic constraints, EFM is what is used.

    “Natural” childbirth advocates like to fling around charges that EFM is not evidence based. That, like most of their charges directed at obstetric practice, is simply untrue.

  69. Plonit says:

    2. The existing scientific evidence shows that EFM does not improves outcomes beyond rigorous protocols for auscultation requiring one on one nursing care.

    +++++++++++++++

    The existing scientific evidence shows that cEFM has worse outcomes in comparison with intermittent auscultation (i.e. outcomes are the same for babies, but worse for mothers).

    Can you cite any health economic studies that show that cEFM is either absolutely cheaper, or more cost effective than IA. Any health economic study would need to take into account not only the level of care required to provide IA, but also the reduced rate of spontaneous vaginal delivery with cEFM (which would increase the costs of cEFM). Do you have access to any such health economic analysis? Have you compared the costs of maternity care in systems that do not use cEFM routinely in all woman with the cost of maternity care in the US. What is the basis of your claim that a move to IA is not possible “given current economic constraints”?

  70. Plonit:

    In what objective way are outcomes worse? Do you have any citations to support that claim?

  71. AlexisT says:

    Ah, the Amy we know and love: when confronted with contradictory evidence, she resorts to a pointless aside about “homebirth midwives”. To add to Plonit’s comments, IA for low risk labors is recommended in the official NICE intrapartum guideline.

  72. Plonit says:

    Yes, I do. The Cochrane systematic review has already been cited in previous comments. And as a commentator on the science-based medicine in the field of obstetrics, you should certainly be aware of it. But for your benefit the reference is

    Alfirevic Z, Devane D, Gyte GML. Continuous cardiotocography (CTG) as a form of electronic fetal monitoring (EFM) for fetal assessment during labour. Cochrane Database of Systematic Reviews 2006, Issue 3. Art. No.: CD006066. DOI: 10.1002/14651858.CD006066.

    http://mrw.interscience.wiley.com/cochrane/clsysrev/articles/CD006066/frame.html

    The outcomes are objectively worse, becaues the rate of spontaneous vaginal delivery is lower in labours monitored by cEFM than in labours monitored by IA. All other things being equal, spontaneous vaginal delivery is a better health outcome for mothers than instrumental delivery or caesarean delivery. I am shocked that as an obstetrician you are unaware of this systematic review and its included studies.

  73. Plonit:

    “I am shocked that as an obstetrician you are unaware of this systematic review and its included studies.”

    Of course I am aware of it, and other studies as well.

    The expected and inevitable result of a screening test with a high false positive rate is that diagnostic or treatment procedures will be performed that are unnecessary in retrospect. We know that. That’s the definition of a high false positive rate.

    That does not mean that the outcomes are objectively worse, since the end point is mortality and morbidity rates of neonates and mothers.

    What you need to provide is data showing that these endpoints are worse.

    Moreover, you need to provide an alternative. Obstetricians and midwives work in the real world, not in highly controlled studies. It is not enough to complain that EFM is not perfect. You must offer an alternative if you expect practitioners to do something different.

  74. AlexisT says:

    We’ve been over this before on her old blog, Plonit, and given her this citation before. As well as the NICE guideline, I’m sure. This is when she reverts to “it’s not economic” and having 1 nurse doing central monitoring is an adequate standard of care, plus the additional criticism of “natural birth advocates” (of which I am not one, particularly) not having any valid criticisms of obstetric practice. For those new to Dr. Amy, be warned: any non-obstetrician daring to criticize obstetrics is not tolerated.

  75. AlexisT says:

    How is the NHS not a real world alternative? Are you going to resort to claiming it just doesn’t have the malpractice issues the US does?

  76. Plonit says:

    Unlike you, I do work in a very cost-conscious real world, where registered midwives use intermittent auscultation routinely. This is in accordance with the evidence, national and professional guidelines.

    Are you seriously questioning that instrumental and caesarean delivery cause increased morbidity in women? Perhaps you need to go back to your basic medical textbooks.

  77. Plonit says:

    I know Alexis, I can’t help but bite. Nice to see you here!

  78. Plonit:

    “Are you seriously questioning that instrumental and caesarean delivery cause increased morbidity in women?”

    Again, you seem to think this is dispositive, when it is merely the definition of a test with a high false positive rate. Mammograms lead to many unnecessary (in retrospect) breast biopsies. Does that mean we should stop doing mammograms?

  79. Plonit says:

    That would depend on whether the screening test had a benefit. In the case of cEFM (vs. IA) where is the proven benefit?

  80. Plonit says:

    Actually, the correct analogy would be this:

    You have two forms of initial screening for breast cancers. Screening A results in many more women undergoing breast surgery but no fewer cancer deaths. Screening B results in fewer women undergoing breast surgery but no more cancer deaths. Which form of screening should you use?

  81. Plonit:

    “Screening B results in fewer women undergoing breast surgery but no more cancer deaths. Which form of screening should you use?”

    Tell it to the people who run the hospitals. Obstetricians have no control over nursing staffing. We must work in the situations that exist, not the ones we wish for.

    The main point is this: obstetricians are practicing according to the scientific evidence and in line with the economic limitations in contemporary practice. If you have any better ideas on what we should do, by all means, share them.

  82. Plonit says:

    Tell it to the people who run the hospitals.

    +++++++++++

    So actually we are having a discussion about the irrationalities of hospital practice, rather than the evidence and science base for particular screening modalities? I’m glad we’ve cleared that up. I bet SBM are now regretting taking you on baord.

    Why do you think UK practice been able to change in response to evidence (despite economic limitations)?

  83. “Why do you think UK practice been able to change in response to evidence”

    They haven’t gotten rid of EFM; it is still an integral part of care during labor.

    As with any screening test, the expectation is that reserving that test for patients at risk will have a lower false positive rate than using it for all patients.

    According, the UK recommendation is now that women without risk factors should be monitored with auscultation. If any risk factors are present, or develop during labor, an immediate switch should be made to cEFM. These risk factors include:

    • meconium

    • abnormal FHR detected by intermittent auscultation (less than 110 beats per minute
    [bpm]; greater than 160 bpm; any decelerations after a contraction)

    • maternal fever

    • fresh bleeding developing in labour

    • oxytocin use for augmentation

    • the woman’s request

    and, of course, would include pre-existing medical conditions, and complications of pregnancy.

    It will be interesting to see what happens with these guidelines. But you need to understand what is going on here. No one is condemning EFM. No one is saying it is useless. They are acting in accordance with everything I have said. EFM has a high false positive rate. The false positive rate of a diagnostic test is usually reduced by restricting it to a population at risk instead of the entire population. That’s what they are doing.

  84. Plonit says:

    Yup, I know all that already – I am quite familiar with those guidelines – we were talking about routine use of cEFM regardless of risk.

    To quote myself upthread

    “All women who go into spontaneous labour at low-risk of complications receive intermittent auscultation in the UK. The standard of care is no different at a home birth and in the hospital.”

    Let me repeat “at low-risk of complications”.

    and “In the UK changes were made in practice to massively restrict the use of cEFM in response to the scientific evidence.”

    Again, let me repeat “massively restrict” (i.e. there are still circumstances in which cEFM is quite rightly used).

    Now, can you stop changing the subject?

  85. “massively restrict”

    Really? What percentage of patients go through labor without any EFM?

  86. Plonit says:

    In my own unit, at least half of all women who go into spontaneous labour (i.e. excluding elective CS and inductions) have only intermittent auscultation. If your comparison is 100% use of cEFM…I think “massive restriction” is fair, but pick your own term if you like.

    By the way, pretty much all auscultation is done electronically (using a doppler) just intermittently, hence specifying cEFM.

    We do collect the data, so I can find out my own unit’s latest figures and post it here at the weekend.

  87. “We do collect the data”

    What is the percentage of all women who don’t have any cEFM?

    How much has your C-section rate dropped? How much money have you saved?

    The overall UK C-section rate has continued to rise. Have there been any cost savings overall from the guidelines?

  88. Plonit says:

    A bit over a third.

    It’s not necessarily the case that the CS rate will go down as a result of not using cEFM universally – for instance, it may increase more slowly than it would do if we were using cEFM universally. There are lots of variables that are hard to control for in real life, which is why we turn to RCTs to give us answers on these things.

    The NICE intrapartum care guidelines on cEFM are not that different from previous UK practice in a lots (though not all) of units. So, again lots of variables, lots of noise – and very difficult to attribute cost savings to one particular item in a set of guidelines that covers every aspect of intrapartum care.

  89. Plonit says:

    Incidentally, for the purpose of health economic arguments, 1-2-1 care in active labour is standard regardless of the method used for fetal monitoring. Most UK units do not have central monitoring, so the midwife is required in the room for interpretation of the trace.

  90. lkw says:

    Emma B:

    I understand your point. Really I do.

    But, no one is arguing that the total elimination of post de facto “unneccessary c-sections” is possible. However, the purpose of the scientific method is to answer unanswered questions. We can then apply the answers to those questions to the improvement of medical care. My only point was that some improvements are possible through science and that I prefer when people use language that indicates as much. I think this would help the general public understand that science is not some static set of facts, but a continuing endeavor.

    “Science and math are necessarily non-deterministic.”

    Some systems may, in fact, be non-deterministic, but this a not a characteristic of “science” in general and it makes no logical sense to say that “math” is non-deterministic.” Again, statements like this leave people with a very wrong impression of what science is. In the modern sense, science is the *process* through which we make new discoveries. The body of knowledge we might refer to as “science” is constantly being refined.

  91. lkw:

    “But, no one is arguing that the total elimination of post de facto “unneccessary c-sections” is possible.”

    You haven’t explained how ANY C-sections that turn out to be unnecessary in retrospect can be eliminated, let alone all of them.

  92. wales says:

    Elsewhere here the question was raised as to the impact of malpractice suits on US cesarean rates. This article points to a JAMA study that found a positive association between medical malpractice claims risk and the rate of cesarean delivery http://www.medscape.com/viewarticle/702712

    I do not see any mention here of ACOG’s position on the subject, but perhaps I missed it. This 63 page ACOG report issued in 2005 is entitled “Evaluation of Cesarean Delivery”.

    http://images.ibsys.com/2005/0504/4450560.pdf

    This report states “A Department of Health and Human Services expert working group on cesarean delivery rates, which included ACOG representatives, discussed the Healthy People 2010 objectives and developed evidence-based cesarean delivery rate goals for the year 2010.” The report goes on to say “The expert working group proposes the following cesarean delivery rate benchmarks:

    1. Nulliparous women at 37 weeks of gestation or greater with singleton fetuses with vertex presentations: The national 1996 cesarean delivery rate for this group was 17.9%; the expert working group goal at the 25th percentile for this group is 15.5%.

    2. Multiparous women with one prior low-transverse cesarean delivery at 37 weeks of gestation or greater with singleton fetuses with vertex presentations: The national 1996 VBAC rate for this group was 30.3%; the expert working group goal at the 75th percentile is 37%.”

    The reality is that there are different acceptable cesarean rates for different types of patients.

    Plonit said “I am shocked that as an obstetrician you are unaware of this systematic review and its included studies.” It appears from a cursory investigation that Amy Tuteur retired from the practice of medicine around 2003. I am unable to verify whether or not she was board certified by ACOG.

  93. wales says:

    Here’s an ACOG report from 2005 on the subject of cesarean delivery with recommended benchmark cesarean rates http://images.ibsys.com/2005/0504/4450560.pdf

    Plonit said “I am shocked that as an obstetrician you are unaware of this systematic review and its included studies.” It appears from a cursory investigation that Amy Tuteur retired from the practice of medicine around 2003. I am unable to verify whether or not she was board certified by ACOG.

  94. wales says:

    Elsewhere here the question was raised as to the impact of malpractice suits on US cesarean rates. This article points to a JAMA study that found a positive association between medical malpractice claims risk and the rate of cesarean delivery http://www.medscape.com/viewarticle/702712. This topic is also discussed in the 2005 ACOG report “Evaluating Cesarean Delivery”.

  95. storkdok says:

    Wales said, “It appears from a cursory investigation that Amy Tuteur retired from the practice of medicine around 2003. I am unable to verify whether or not she was board certified by ACOG.”

    So what? What difference does that make to the discussion? People who retire from active practice don’t check their brains at the door. This is an obvious attempt to discredit Dr. Tuteur.

    This is 2 out of 3 posts by Dr. Tuteur that you have used this tactic on.

  96. storkdok says:

    Dr. Tuteur, very interesting post and discussion! Glad to have you here!

Comments are closed.