Reply to Maureen Minchin’s non-responsive piece

Reply isolated on cyan blue keyboard button abstract illustration

Maureen Minchin has refused to abide by impartial debate rules so she has forfeited. She’s still forging ahead any way, posting what amounts to an opening statement.

At no point does she provide any evidence that breastfeeding has been shown to actually have the benefits claimed; she simply repeats the claims. On her Facebook page she reports that writing the piece was “a lot of work and great fun, too.”

[pullquote align=”right” cite=”” link=”” color=”” class=”” size=””]Maureen Minchin’s piece — verbose and self-promoting — is non-responsive. She has yet to show that breastfeeding has the benefits claimed; she simply repeats the claims. [/pullquote]

Here’s my response:

Are the benefits of breastfeeding real and clinically relevant or merely theoretical and not reproducible in large populations?

Public health initiatives, by definition, are meant to improve public health.

They are usually based on solid scientific evidence, their implementation saves thousands if not millions of lives, and they pay for themselves many times over in lives saved, earnings preserved and medical expenditures averted.

Consider the classic public health campaigns to promote vaccination and to reduce tobacco smoking.

This graph shows the dramatic drop in incidence of vaccine preventable disease after the introduction of the vaccine for the specific disease:

C370D66E-28BD-4F1D-933B-A2933BD5770F

Notice that the y-axis is logarithmic, which means that the actual changes were far more dramatic than a glance at the graph would indicate. For example, there were approximately a one hundred thousand cases of smallpox per year prior to the introduction of the vaccine. In 2012 there were no cases at all. For each and every vaccine, the number of cases decreased by several orders of magnitude after the introduction of the vaccine.

The public health campaign to reduce tobacco smoking has had similarly spectacular results.

379811B8-793F-4A1E-9F71-E5CF44DFFA0B

This graph originally published in the National Cancer Institute Bulletin shows that in the wake of the Surgeon General’s report of 1964 warning about the link between smoking and lung cancer, per capita cigarette consumption dropped dramatically. After a lag period, lung cancer deaths began to drop dramatically, too.

We have spent millions of dollars promoting vaccination and reducing smoking and it has paid off in both lives and money saved.

How about breastfeeding?

An entire industry, the lactation industry, has arisen to promote and profit from efforts to increase breastfeeding rates. They’ve claimed a myriad of benefits for breastfeeding and predicted that an increase in breastfeeding rates would produce a decrease in infant mortality as well as reductions in a variety of diseases and conditions.

Breastfeeding initiation rates have risen in response. They have nearly quadrupled since 1970 rising from 22% to over 83% today. But the breastfeeding rate appears to have had no impact on the infant mortality rate. The graph below illustrates the steep drop in infant mortality over the course of the 20th Century. I’ve added markers for the breastfeeding rate at various points. As you can see, the precipitous drop in breastfeeding rates did not have an impact on infant mortality and the rising rate of breastfeeding initiation does not seem to have an impact, either.

EDD0D051-7FAC-4B8D-8799-EBA3A3BA0ABD

Breastfeeding was supposed to prevent obesity, but obesity rates continue to rise. It was supposed to prevents asthma but rates have continued to rise. In fact, nearly all the predictions that flow from claims about the benefits of breastfeeding have failed to come to pass.

That raises the question: Are the benefits of breastfeeding real? It also raises a follow up question. If the benefits of breastfeeding are shown to be real, are they clinically relevant?

What do the questions mean?

Before we can answer those questions, we need to understand what they mean. I’m going to use a simple, silly example to explain.

Imagine a study that looked at the average body temperature of children in different grades. The investigator went to a local elementary school and took the temperatures of all 189 first graders in multiple different classes and all 193 second graders in multiple different classes.

He found that the average temperature of first graders was 98.7 and the average temperature of second graders was 98.9; he concluded that getting promoted from first to second grade raises body temperature.

What do we mean if we ask if that is a real result? We aren’t questioning whether the investigator took temperatures properly or whether he accurately wrote them down and accurately averaged them. When we ask if a result is real, we’re asking (among other things) whether it is reproducible. Would the results be the same if the investigator repeated the investigation a month later? Would the results be the same if the investigator repeated the investigation in a different school? Would the results be the same if the investigator repeated the results using 1000 first graders and 1000 second graders?

We can’t consider the results real unless the same results occur repeatedly.

In reaching his conclusion that promotion to second grade raises body temperature, the investigator assumed that first graders and second graders were otherwise the same in all respects except for body temperature. But what if he had measured the temperature of the first graders before recess and the second graders after recess? The extra physical activity of the second graders have been responsible for their higher average body temperature.

Recess in this example is what is known as a confounding variable. It isn’t the promotion that caused the second graders’ average temperature to be higher, it’s the physical activity that occurred before their temperatures were measured.

We can’t consider results real unless they have been corrected for confounding variables.

What do we mean when we ask if a result is clinically relevant?

In the example of the first and second graders, the second graders had an average temperature of 0.2 degrees higher than the first graders. Even if it were a real result (reproducible and corrected for confounding variables), it isn’t a clinically relevant result. Both groups of children had average body temperatures well within the normal range. It makes no difference that the average temperature is higher in second graders; both groups are healthy.

What do the questions mean in the context of breastfeeding?

When I ask whether the benefits of breastfeeding are real and clinically relevant, I’m asking whether they have been reproduced, whether they have been corrected for all confounding variables and whether they make a meaningful difference to the health of babies and mothers.

What won’t answer the questions?

There’s a long list of things that might at first sound impressive, but don’t really answer the questions.

The statements of authority figures or organizations don’t answer the questions.
Scientific citations of studies that found effects in small groups don’t answer the questions.
The naturalistic fallacy (“if it’s natural it must be good”) does not answer the questions.
Personal beliefs and personal anecdotes don’t answer the questions.
Mathematical models based on extrapolation of small studies don’t answers the questions.

What would answer the questions?

As with any public health measure, the proof of the pudding is in the eating. In other words, the benefits must be measurable.

Would we believe that a vaccine was worthwhile if we gave it to millions of children but we could not find a measurable impact on the incidence of the disease it was supposed to prevent? No.

Would we believe that stopping smoking reduces the risk of lung cancer if millions of people stopped smoking and the rate of lung cancer remained the same? No.

Should we believe that breastfeeding has a myriad health benefits for term babies including saving lives if no one can show that any lives have been saved? No. Should we believe that breastfeeding has a myriad of health benefits if incidence of the diseases that breastfeeding was supposed to prevent remained unchanged or even rose? No.

Maureen Minchin’s piece — verbose and self-promoting — is non-responsive. She has yet to show that breastfeeding has the benefits claimed; she simply repeats the claims.

Why? Because she couldn’t find any data that shows that the benefits claimed for breastfeeding term infants are real or clinically relevant.