Infant mortality report neglects the most important detail
The new CDC report on infant mortality, Behind International Rankings of Infant Mortality: How the United States Compares with Europe, is an object example of how to deceive with statistics. It purports to be a detailed investigation of infant mortality, but it inexplicably fails to investigate the most important detail.
According to today’s article in the NYTimes:
High rates of premature birth are the main reason the United States has higher infant mortality than do many other rich countries, government researchers reported Tuesday in their first detailed analysis of a longstanding problem.
In Sweden, for instance, 6.3 percent of births were premature, compared with 12.4 percent in the United States in 2005, the latest year for which international rankings are available. Infant mortality also differed markedly: for every 1,000 births in the United States, 6.9 infants died before they turned 1, compared with 2.4 in Sweden. Twenty-nine other countries also had lower rates.
If the United States could match Sweden’s prematurity rate, the new report said, “nearly 8,000 infant deaths would be averted each year, and the U.S. infant mortality rate would be one-third lower.”
The use of this example highlights to disingenuousness of the authors. In their supposedly “detailed” report on infant mortality, they fail to analyze the most important detail: race. Unfortunately, African descent is a major risk factor for prematurity, and prematurity is a major cause of infant mortality. Therefore, it is hardly surprising that the US has a higher infant mortality rate than Sweden. The US has the highest proportion of women of African descent of any first world country. Sweden, of course, has virtually none.
The authors, however, seem more interested in jeering the US for its supposedly low standing in international comparisons than they seem in actually getting to the source of the problem. The report is filled with grim looking graphs that show how “poorly” the US fares when compared to other first world countries.
The first graph highlights the fact that the US is ranked 30th in the world for infant mortality. But the authors acknowledge that the US has a more comprehensive definition of infant mortality than other first world countries, many of which exclude the deaths of very premature infants even when they are born alive. The authors present a second graph adjusting for this discrepancy. In that more accurate graph, the US ranks 18th.
The authors mention the impact of race on prematurity, but they never adjust for it. The CDC Wonder website gives us access to the same database that MacDorman used in the study. Therefore, we can adjust for race. Doing so, would put the US 14th in the rankings.
The authors also mention assisted reproductive technology, but they don’t adjust for that either. The rates of twins, triplets and higher is greater in the US than in many European countries because of differing rates of assisted reproductive technologies and the difference in techniques.
The authors acknowledge that on an age specific basis, the US actually does better than almost all European countries. In other words, we are better at saving premature babies. Our relatively low ranking is the result of a higher rate of prematurity.
So our higher rate of infant mortality does not reflect poor medical care. It reflects factors beyond the control of doctors. Race is an uncontrollable factor; obstetricians and pediatricians have no control over assisted reproductive techniques. In fact, the data actually show obstetricians and pediatricians do a remarkable job of ensuring infant health.
Dr. MacDorman’s bias is most evident is her gratuitous swipe at obstetricians. According to the Times article:
Another factor in the United States, she said, is the increasing use of Caesarean sections and labor-inducing drugs to deliver babies early. The American College of Obstetricians and Gynecologists has guidelines stating that babies should not be delivered before 39 weeks without a medical reason, but doctors may be declaring a medical need more quickly than they did in the past.
“I don’t think there are doctors doing preterm Caesarean sections or inductions without some indications,” Dr. MacDorman said, “but there sort of has been this shift in the culture. Fifteen or 20 years ago, if a woman had high blood pressure or diabetes, she would be put in the hospital, and they would try to wait it out. It was called expectant management.
“Now I think there’s more of a tendency to take the baby out early if there’s any question at all.”
Dr. MacDorman neglects to mention that there is no evidence that such births are contributing in any way to the infant mortality rate. Indeed, the existing evidence suggests that these births actually save lives. During the time period when early deliveries increased, the rate of stillbirth dropped by 29%.
Infant mortality and prematurity are real and serious problems, and they won’t be solved by pretending they are simply medical problems. Infant mortality in general, and prematurity in particular, is the result of racial, social and economic disparities that must be investigated and addressed. MacDorman’s report risks obscuring this critical point in favor of castigating medical practitioners. Doctors are not responsible for the US ranking in infant mortality statistics, and therefore, they cannot fix it. If our goal is to prevent infant deaths, we must be honest about the real causes.
TIP:
Enter the amount, and click "Tip" to submit!
aliquot
Roger Fallihee
madcelt
Norwonk
John Moore
Esther AR
Salon.com
Comments
Anyway, great article!
As Dr. Fleishman points out, the increase in later (but still pre-term) c-sections is not as significant in the data as the high numbers for very premature births occuring. From the NY Times:
"These births — called “late preterm,” which occur after 34 to 37 weeks of pregnancy, instead of the normal 38 to 42 weeks — are the fastest-growing subgroup of premature births. A late preterm baby’s risk of dying is about three times that of a full-term infant. But late preterm babies are still far more likely to survive than very premature ones, and the very early babies account for much of the death rate, Dr. Fleischman said."
Europe is the United States’ much higher percentage of preterm births." The report makes no attempt to uncover the causes for these preterm births.
" 'The quality of neonatal intensive care is superb,' [Fleischman] said. 'We know how to rescue babies who are born very tiny, but what we don't do well is prevent prematurity.' Reasons for this, he said, include a lack of universal access to health care for women of childbearing age or pregnant women of any age. 'That's a tremendous difference with our European friends,' Fleischman said. There's also not enough attention on social and behavioral factors that affect prematurity, such as smoking and alcohol and drug use, he said."
There is indeed evidence. It would appear that Dr. MacDorman mentioned this particular factor to the NY Times because she has previous research experience in this area. She presented evidence in 2006:
"Neonatal mortality rates were higher among infants delivered by cesarean section (1.77 per 1,000 live births) than for those delivered vaginally (0.62). The magnitude of this difference was reduced only moderately on statistical adjustment for demographic and medical factors, and when deaths due to congenital malformations and events with Apgar scores less than 4 were excluded. The cesarean/vaginal mortality differential was widespread, and not confined to a few causes of death."
In speaking to the NY Times on that occasion, she did note:
"'This is nothing to get people really alarmed, but it is of concern given that we’re seeing a rapid increase in Caesarean births to women with no risks,' Dr. MacDorman said."
I understand why you might be suspicious, but reimbursement has little to do with the rising C-section rate. The C-section rate has risen dramatically in countries in which all doctors are on salary.
The article says that Native Americans also have higher rates - is this because they are Native American or because they, by and large, do not have access to quality health care?
I think health care is another common denominator here, along with ancestry.
The article also states that a program to provide home nursing visits to pregnant women in Kentucky reduced the number of premature births. Kentucky is only about 8% black, yet almost 17% of the population is poor. It would be helpful to know whether the premature births are happening predominantly among the poor blacks rather than poor whites, i.e., is the common denominator poverty or ancestry, or both?
Presumably if the factor in question is ancestry, visits from nurses would have no effect on the number of premature births. Nurse visits during pregnancy do not change the baby's ancestry.
Whether it is one, the other, or both, the fact is that we have much higher poverty rates and rates of people who can't afford health care than most of the countries ahead of us on that list.
Two groups who had the same outcome (death) were compared thus both birth and death certificates were available for these two groups. If pre-natal factors were under-reported in one group, they were almost certainly under-reported at an equal rate in the other group. In fact, one might guess that pre-natal factors would be reported at a higher rate in the group where c-sections were performed, because the reporting physician might wish to account for why the c-section was performed. Further, the removal of "deaths due to congenital malformations and events with Apgar scores less than 4" almost certainly reduced (from both groups) the number of cases where a pre-natal condition was the determining factor in mortality.
Dr MacDermon's work on mortality rates is a valuable contribution to the research in this field. Nonetheless, she notes, "This is nothing to get people really alarmed, but it is of concern given that we’re seeing a rapid increase in Caesarean births to women with no risks."
But, here's the deal... I have chronic high blood pressure. Last pregnancy I developed pre-eclampsia and HELLP... I had to have a c-section and had severe hemorrhaging afterwards (bp dropped to 50/20 and I had to have transfusions). OK, I didn't mean to plan an invisible violin solo, but I think following the advice of my doctor is the reasonable course here. Sorry to disappoint all my hippie friends who INSIST I need to do a VBAC (some say at HOME, no less - ACK!) I've heard this high infant mortality rate cited so many times, and I just had a hard time believing the statistic was saying what my friends were claiming it said. So THANK YOU so much for this perspective.
I'm looking forward to keeping up with your blog and learning more.
--Lily
Homebirth and "natural" childbirth advocates like to claim that Cesarean section dramatically increases the risk of neonatal death , based on MacDorman's findings. What they do not realize is that MacDorman herself has already revised this claim drastically downwards, and that even then, the conclusions are false.
In September 2006, MacDorman et al. published (and heavily publicized with interviews to the lay press) a paper that purported to show that Cesarean section increases the risk of neonatal death almost 3 fold. MacDorman et al. compared outcomes of C-sections with "no indicated risk" (a blank space on the risk section of the birth certificate) with outcomes from vaginal deliveries with "no indicated risk" and found that the neonatal death rate was higher in the C-section group. However, MacDorman neglected to mention that it is well known that the risk section of the birth certificate is often left blank even when there are serious risk factors and complications. Indeed, in 50% or more of serious risk factors (heart disease, kidney disease, etc.) the space is left blank. So their assumption that "no indicated risk" means no risk is completely unjustified.
In the follow up paper Neonatal Mortality for Primary Cesarean and Vaginal Births to Low-Risk Women: Application of an "Intention-to-Treat" Model, MacDorman et al. use the same dataset, known to be flawed and incomplete, but applied a better form of analysis. Using this new, more accurate statistical analysis, MacDorman et al. went back and reviewed their incomplete, flawed dataset.
" ...In the most conservative model, the adjusted odds ratio for neonatal mortality was 1.6 (95% CI 1.35–2.11) for cesareans with no labor complications or procedures, compared with planned vaginal deliveries. Conclusions: The finding that cesarean deliveries with no labor complications or procedures remained at a 69 percent higher risk of neonatal mortality than planned vaginal deliveries is important, given the rapid increase in the number of primary cesarean deliveries without a reported medical indication."
MacDorman et al. adjusted their claim downward by a substantially amount. Instead of their original claim that C-section increases the risk of neonatal death by 200%, they now find that C-section increases the rate of neonatal death by only 69%. Yet even then, they failed to acknowledge the most significant flaw in the paper. MacDorman and colleagues assumed that a birth certificate that listed no risk factors for C-section indicated that there really were no risk factors. A large body of data on the accuracy of birth certificates had long ago shown that more than 50% of cases of major risk factors were never listed on the birth certificate and that, therefore, their assumption was completely unjustified.
In the January issue of Obstetrics and Gynecology there is a new paper that adds to the mountain of evidence demonstrating that it is impossible to determine C-section risk factors merely by looking at birth certificates. According to Cesarean Delivery Among Women With Low-Risk Pregnancies: A Comparison of Birth Certificates and Hospital Discharge Data:
"Among 40,932 women with primary cesarean deliveries and no risk indicated on the birth certificate, 35,761 (87.4%) had a risk identified in the hospital discharge data. The overall agreement between data sources on the presence of any risk indicator was low (κ=0.18). Among primary cesarean deliveries, the percentage without indicated risk was 58.3% when using birth certificate data alone and 3.9% when using hospital discharge data in combination with the birth certificate.
CONCLUSION: Using birth certificate information alone overestimated the proportion of women who had no-indicated-risk cesarean deliveries in Georgia. Evidence of many indications for cesarean delivery can be found only in the hospital discharge data. The construct of no indicated risk as determined from birth certificates should be interpreted with caution, and the use of linked data should be considered whenever possible."
In other words, virtually all women who had primary C-sections but had no risk factors on the birth certificate, actually did have risk factors. In the case of the MacDorman study, the authors reported that there were 469 deaths out of 271,179 births to women who had primary C-sections and no risks documented on the birth certificate. Yet if 87% of the birth certificates were inaccurate, that means that over 235,000 were wrongly placed in this category, thoroughly invalidating the results of the study.
The bottom line is that there is no evidence that C-section increases the risk of neonatal death in this study or in any other study to date.
"Do African-American women tend to receive prenatal care, in quality and quantity, equal to what non-African-Americans receive? Could this be a mitigating factor?"
Poverty and poor prenatal care are also risk factors, but it appears that race is an independent risk factor. There may be mechanisms or genetic issues faced by those of African descent that increase the risk of prematurity. Unless and until we recognize the issue of race as a risk factor, and explore the possible reasons why, we won't be able to solve the problem.
In addition to misrepresenting what is really going on with prematurity in the US, papers like MacDorman's elide the possibility that prematurity among African-Americans may have a different etiology and may require a different treatment.
Shouldn't that be "the most" not "most the" ?
How many degrees does it take to arrange the words in a sentence properly? With that sort of attention to detail it's no wonder you no longer practice medicine:)
But Dr. MacDermon did not rely solely on birth certificates, so how does this apply to either of her studies, both of which indicate an increased mortality in c-sections?
"have more than once cited the U.S. "insanely high" infant mortality in their arguments."
When it comes to evaluating obstetrical care, the correct parameter is not infant mortality, but perinatal mortality. Infant mortality is deaths from birth to 1 year; perinatal mortality is deaths from 28 weeks gestation to 1 month. The US has a very low perinatal mortality rate, lower than Denmark, the UK or The Netherlands.
There are quite a few studies that demonstrate that C-sections are safer for babies than vaginal birth.
Baskett TF, Allen VM, O’Connell CM, et al. Fetal trauma in term pregnancy. Am J Obstet Gynecol 2007;197:499.e1-499.e7 is a comprehensive study of over 100,000 consecutive deliveries of singleton term neonates in the vertex position with no congenital anomalies:
"The potential study population in Nova Scotia for the 14-year period (1988-2001) was 153,209. Twenty-two percent of the population was excluded based on the inclusion criteria of singleton (1%), term (7%), no major anomalies (3%), no fetal deaths (0.1%), and in vertex presentation(11%) to give the actual study population of 119,432."
Deliveries were grouped as follows:
"... mode of delivery was defined as vaginal or cesarean, and method of delivery was defined as spontaneous vaginal, assisted vaginal (vacuum, forceps), ... and cesarean delivery (with and without labor)..."
The definitions of trauma were:
"Fetal trauma was considered major trauma if 1 or more of the following were present: depressed skull fracture, intracranial hemorrhage, or brachial plexus palsy. Minor trauma was considered if 1 or more of the following were present: linear skull fracture, other fractures, facial palsy, or cephalhematoma."
The results of the study showed that the risk of major and minor trauma was dramatically reduced by Cesarean section, particularly if the Cesarean was performed before labor began. If the risk of major trauma in a vaginal delivery is described as 1.0, the relative risk of major trauma is reduced by 82% by a C-section done in labor (RR 0.18), and reduced by 88% by a C-section done before labor begins (RR 0.12).
Yes, she did. Read the paper.
The complex nature of the etiology is discussed rather interestingly in Preterm Birth: Causes, Consequences, and Prevention.
This paper, as you can tell from the title, is about international comparisons. But the comparison she draws is disingenuous since the rankings reflect the proportion of women within the population who are of African descent. Furthermore, rather than addressing the critical issue of prematurity among African-Americans, the paper virtually ignores it.
I am quite familiar with Dr. MacDorman's work. She has an ax to grind and she's entitled to do so, but she is not entitled to confuse her opinion with the scientific evidence.
"In the linked birth and infant death data set the information from the death certificate is linked to the information from the birth certificate for each infant under 1 year of age who dies in the United States, Puerto Rico, The Virgin Islands, and Guam."
Dr. MacDermon uses linked birth and death certificate data, not solely birth certificate data, thus the point that "the accuracy of birth certificates had long ago shown that more than 50% of cases of major risk factors were never listed on the birth certificate" is simply not relevant.
I wondered why I could find no "jeering" or "swipe"-taking in the NY Times article, where you found them. Talk about disingenuous.
This paper is presented as an in depth view of the problems of infant mortality and premature birth. How can it be an intellectually honest look at the problem when it fails to discuss what may be the most important risk factor?
No, it isn't.
Have you read it?
So then you know that it is an attempt to understand why the US ranks behind so many European countries. They address the fact that almost half the discrepancy comes from differing methods of counting infant deaths. Yet the fail to address the fact that a good portion of the rest of the discrepancy is the results of differing risk levels among populations.
Either they want to explain the disparity or they don't. If they want to explain it, they can't in good conscience leave out one of the most important explanations.
Your very first comment on this thread is a swipe at obstetrics, implying that the high C-section rate is a cause of infant mortality. You were quick to conclude that this report is an indictment of obstetric practice. That's because you didn't understand the contribution of various risk factors. I've explained it to you, but Dr. MacDorman should have explained it in the paper.
She left you with the wrong impression, so why are you defending her when I point out that she is deliberately leaving people with the wrong impression?
No, it wasn't. See above.
AmyTuteurMD wrote: "You were quick to conclude that this report is an indictment of obstetric practice."
No, I didn't. See above.
AmyTuteurMD wrote: "She left you with the wrong impression."
Dr. MacDermon left me with the following impression: preterm births are responsible for a lot of infant mortality in the US compared to Europe. Nothing more, nothing less. Because that is what the data indicated. The report did not attempt in any way, shape, or form, to explain the reasons for preterm births!
And twice I quoted Dr. MacDermon's statement regarding late-term c-sections with no known medical cause: "'This is nothing to get people really alarmed." More indictment to you?
Everywhere you look, you see indictment, jeering, and swipe-taking. The common link is you.
But the CDC study you link to clearly states:
"Infant mortality rates for infants born at 37 weeks of gestation or more are higher in the United States than in most European countries."
Plus where are these studies that show that you can separate race from poverty and teen pregnancy (which I'm sure counts for more than a few premature births.)? I'd like to read them.
That doesn't tell us anything about the impact of induction for convenience since most of those births are spontaneous or induced for serious medical indications.
How did you perform the adjustments?
"These births — called “late preterm,” which occur after 34 to 37 weeks of pregnancy, instead of the normal 38 to 42 weeks — are the fastest-growing subgroup of premature births. A late preterm baby’s risk of dying is about three times that of a full-term infant."
Your stillborn statistic doesn't seem to hold up to that evidence since infant mortality includes live births.
Finally, how do you address the fact that U.S. is still behind European countries in infant mortality of babies delivered at term? Is this due to African descent as well?
I guess the question I'd like to ask is, do you think that U.S. social policies, health care policies and medical practice are in NO way responsible for this discrepancy. And should the U.S. be entirely ignoring this report and blaming it on genetics?
Or is there more gray here than you're prepared to admit?
And until you come up with some evidence that these problems are genetic not social, I'm going to maintain that the biggest factor is reporting.
Did you read the report?
Did you read the report?
The statements are true-true and unrelated.
A late preterm baby's risk of dying is higher because most late preterm infants are spontaneously born early because of congenital anomalies, multiple pregnancy or a serious medical problem requiring preterm delivery. In all three of those groups, the death rate is going to be higher almost by definition.
What we really want to prevent are preterm deliveries for convenience. Those deliveries can lead to more short term self limited medical issues, but they don't lead to more deaths.
"Your stillborn statistic doesn't seem to hold up to that evidence since infant mortality includes live births."
That's why perinatal mortality is a much better measure of obstetric care, and that's why the WHO recommends using perinatal mortality for international comparisons. Late pre-term births that are induced for medical reasons are induced to prevent stillbirth and they do.
"Finally, how do you address the fact that U.S. is still behind European countries in infant mortality of babies delivered at term?"
Infant mortality extends up to one year of life. It includes causes such as motor vehicle accidents, child abuse, etc. There are wide variations in causes of infant deaths from country to country.
It is possible that US medical care is the cause of the higher infant mortality rate, but that is only one of many possible causes. The others, including poverty, lack of access to medical care, and poor parenting are more likely to be the reason.
"Do people of african descent have a higher incidence of premature birth, or do they just have a shorter gestation period?"
That's a great question. If they naturally had a shorter gestation period we would expect that infant mortality rates would be similar, but in fact, the mortality rates for black infants are very high.
Larry Lundgren - Linköping, Sweden (no I am not Swedish)
Do you have some data that demonstrate that all late preterm infants fall into one (or more) of these three categories? If not, then this statement is clearly incomplete. And the claim that "Those deliveries can lead to more short term self limited medical issues, but they don't lead to more deaths." is not proven at all.
Dr. MacDorman has studied linked data such as that recommended by Kahn, et al.,). Those studies have (respectively) shown:
1) that among babies who die when delivered in late preterm: "Neonatal mortality rates were higher among infants delivered by cesarean section (1.77 per 1,000 live births) than for those delivered vaginally (0.62)".
2) that if one examines all births (as opposed to only those who die) and categorizes them according to the delivery method: "In the most conservative model, the adjusted odds ratio for neonatal mortality was 1.69 (95% CI 1.35–2.11) for cesareans with no labor complications or procedures, compared with planned vaginal deliveries."
The two studies are different and the second does not "adjust their claim downward." In fact, MacDorman, et al., make no "claim," they report the statistical contents of the data sets they examine.
In the study published today, Dr. MacDorman's analysis demonstrates that: "The main cause of the United States’ high infant mortality rate when compared with Europe is the very high percentage of preterm births in the United States."
In none of Dr. MacDorman's published research is there an indication that she has an "ax to grind" or is doing anything other than acting as any responsible statistician should: reporting on the contents of the data available.
It is also worth noting that Dr. MacDorman has never made any statement regarding home births. The fact that home-birthers like to refer to her studies can hardly be her fault.
Further, your characterization of her NY Times comments (made to the press, not published in research journals) is incorrect. Dr. MacDorman does not "jeer" nor does she "gratuitous[ly] swipe" at obstetricians. When called upon to speculate on the reason for the very high percentage of preterm births, Dr. MacDorman mentions one factor with which she has any research experience, leaving the rest of the speculation to others with research experience in the relevant areas.
I am indeed defending Dr. MacDorman. All the evidence indicates that she is a good statistician doing what good statisticians do, all for the reward of a government salary, an occasional call from the New York Times, and a hard time from you.
[My previous posting of a nearly identical comment got mangled by my sloppy use of html. Apologies.]
I'm not sure why you are still arguing with me. There is no question that race is a major risk factor for premature birth. There is no question that MacDorman did not mention it. There is not question that it must be discussed in any review of international comparisons of infant mortality.
But we still disagree.
You state: "There is no question that race is a major risk factor for premature birth. There is not question that it must be discussed in any review of international comparisons of infant mortality."
Uh, no, again.
The "review" is not even a "review." It is a brief. Are you familiar with academic publishing? Reviews tend to be long, comprehensive and wide-ranging. This one is clearly not and never purports to be.
This brief makes absolutely no claims concerning which risk factors do or do not contribute to preterm birth! The study in no way claims to explain what risk factors are responsible for the high rates of preterm birth in the US! The study merely demonstrates that the difference in preterm birth rates accounts for much of the difference between the US and other nations.
Coming from Canada, and having worked in the health policy area, my biggest question (not wanting to chew through the data myself): how much of this variance can be explained by "race" as you label it, and how much by social inequalities.
All the population health literature that I've read points out the sharp health gradients across all populations, whether you do rankings by income, education, social status or some other measure of hierarchy. I would have thought that the impact of social inequalities was waay bigger than the residual left by differences in genetic endowment alone. Your policy recommendation -- that people should therefore seek to understand the impact of this difference in genetic endowment -- makes no sense if the impact from factors related to social inequalities is far bigger.
The "leap to it" conclusion that most of us make on these kind of comparisons, looking at them from outside the U.S., is that the impact of social inequality in the U.S., exacerbated by how that's linked to access to health services in the absence of some form of universal coverage, explains most of why the U.S. indicators lag. The counter example of Cuba (developing nation overall, but constrained levels of social inequality and significant investments in health services for all) seems instructive.
I've run into a similar debate at a conference where Robert Putnam presented his famous "bowling alone" data -- he pointed to ethnic diversity as being a key factor in reduced social cohesion, with the surprising exception of Seattle; I pointed out that in all of his other examples, ethnic groups were strongly associated with differences in income, and that maybe it was _class_ that he should be looking at instead of race.
Right on!!
This was my thought after reading the title of your post, without even having read the post yet.
This is a major issue in science and medical fields right now, and one that intersects with policy - which means it is the business of everyone. On election day, its fitting to think about educated voters - and our choosing of educated senators and representatives.
Rated.
How to Lie with Statistics by Darrell Huff and Irving Geis
I give three examples and think it likely that others who have commented have made similar points. Each of the many examples that could be presented illustrate a disconnect between Dr. Tuteur’s assertions and the actual content of the document.
What is important is to get to questions that Dr. Tuteur seems unwilling to approach at the level one might have hoped for of a physician.
1) She writes that “…the new CDC report...is an object example of how to deceive (my emphasis) with statistics.”
I cannot see any reason at all for this assertion. The document is a very short summary of some findings on infant mortality, which, by using comparisons with Sweden, notes how a change in one element of the record in the USA could bring the two countries close together.
2) She follows that assertion with another: “It (a CDC Brief consisting of 6 pages of text and figures) purports to be a detailed investigation of infant mortality but it inexplicably fails to investigate the most important detail.”
The report in now way “purports to be a detailed investigation” as is evident from the Document classification NCHS Data Brief – No. 23 – November 2009. A data brief is exactly what the name indicates.
3) And last, but not least, we read the following: “The authors, however, seem more interested in jeering the US for its supposedly low standing…”
This is indeed a disturbing departure from what one would expect of a medical professional. There is not a single sentence in the report that uses phrasing that might be interpreted as “Talking Heads ridicule” instead of straightforward presentation. Coupled to this phrase is the phrase “grim-looking graphs”, which are in fact the simplest possible bar graphs presenting data that she does not question.
It appears to me that Dr. Tuteur has two concerns that may account for some of the intemperate and unscientific phrasing evident throughout. One is that she seems to find in the report some reason to infer that the authors are stating that “…the higher rate of infant mortality …(I invert her “does not reflect” to “reflects”) reflects poor medical care. The other is that the report fails to take “race” into account.
On the first count the assertion is simply unjustified. Dr. Tuteur has invented this position for reasons known only to her. The second count – that “race” is a factor – is potentially more scientifically interesting, but since Dr. Tuteur does not explain for us just what the term or concept “race” stands for I will write little more here but hope to take this up in my blog www.only-neverinsweden.blogspot.com.
If by a “race effect” she means that African American women have genomes that, for example, trigger pre-term delivery then let us hear more about the possibility for investigating this. If, on the other hand, she means that a greater percentage of African American women face to a greater extent than “whites” a variety of economic and societal hindrances and a greater degree of exclusion from medical care then she should say so. These issues are not a matter of “race” but rather a matter of societal and economic problems.
To conclude, as a first approximation, it is very likely that the low infant mortality in Sweden is first and foremost a result of the fact that 99% of pregnant Swedish women, whatever their “race”, enter the maternal care system very early in their pregnancies. This is to say that if she had given the matter a moment’s thought she would have realized that the authors would not for a moment suggest that American maternal-care professionals are any less competent than their Swedish counterparts but rather that only some fraction of American women get to benefit from those professionals from start to finish.
Larry Lundgren - Linköping, Sweden
only-neverinsweden.blogspot.com
Open Salon usalars
"I would have thought that the impact of social inequalities was waay bigger than the residual left by differences in genetic endowment alone."
But other minorities face social and economic inequities and some have better perinatal mortality statistics than white women. In addition, African women who are well educated and have high incomes nonetheless have an increased rate of prematurity and perinatal mortality.
Check it out: http://www.sciencebasedmedicine.org/?p=2507
"sloppy scholarship"
"a poor job of interpreting the data they present, and you are drawing conclusions that are largely unsupported"
"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"
"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)."
"Unfortunately, we don’t know what MANA’s data actually confirms or contradicts. Many of us in the skeptical community would jump on such a purely speculative assertion if tendered by the woo-crazies."
"Again, you might not prefer to look at the issue that way, but science doesn’t and can’t say that one view is “right” and one is “wrong.” You have a habit – I’ve gone and read some of your blog – of assuming that those who disagree with you can only be doing so because they are mentally or emotionally inferior in some way. That habit will help you fit in well at this site, but it won’t help you deal with potential patients, who know at some level that they are being gamed and resent it."
"Good job of sneering, but a very poor job of interpreting scientific papers."
"Your personal re-analysis of the study has been widely disseminated on the internet, but has never been peer-reviewed or accepted for publication in an appropriate scholarly environment. Your analysis and presentation of statistics extracted from the CDC database also does not meet the standards for publication in a scientific journal."
"I really don’t think this is fair or appropriate for this website"
"As others have noted, this post seems far less even-handed than most of the others on SBM."
I have one final piece of advice AmyTuteurMD. Get yourself "educated" in statistics or they are going to hand you your ass.
Way to go demonstrating how those who don't understand science misrepresent science. You cut and pasted the comments that you prefer, but you neglected to mention that most people strongly agreed with my conclusions: homebirth is not safe, and there is no empirical evidence to support an "ideal" C-section rate of 10-15% advocated by the WHO.
By the way, thanks for publicizing my new gig. I am very happy that I was invited to contribute to the Science-Based Medicine blog and I invite those who are interested in rigorous scientific discussion to check out the blog and its many contributors.
overworkedtiredandnumb, Ph.D., physics
However, if such studies as you have reviewed here were adjusted for race and economic factors, evidence that the U.S. is failing those populations would be unfairly obscured, and we would continue to keep our heads in the sand about the need for comprehensive health care reform that addresses the issues of all of our citizens. The data should pose an ethical question, and yes, to the extent that doctors can be vocal (as here) in answering or at least pondering the question, the issues are within their control.
Unless you truly believe that African Americanism is a pre-existing medical condition, you are doing a great injustice to minorities in this country by declaring that their unique medical needs do not belong in the mainstream of results.