What should we do when underperforming big ideas — breastfeeding saves lives — become entrenched?

Big Idea And Innovation Concept

I’m frustrated.

Every year tens of thousands of babies are hospitalized for breastfeeding complications at the cost of hundreds of millions of dollars. Meanwhile, despite 20 years of aggressive breastfeeding promotion and rising breastfeeding rates, the benefits we have been promised, from lives and money saved to conditions and diseases reduced, have failed to appear. Meanwhile researchers keep publishing papers claiming ever more arcane benefits even though their predictions consistently fail to come true.

[pullquote align=”right” cite=”” link=”” color=”” class=”” size=””]In breastfeeding research and practice, NO mechanisms exist to sunset failing initiatives; effort is devoted toward increasing funding for them.[/pullquote]

What should we do when underperforming big ideas — like the claim that breastfeeding saves lives and money — become so entrenched that they can’t be challenged, even by reality?

I found some answers in a paper by John Ioannidis and colleagues, What Happens When Underperforming Big Ideas in Research Become Entrenched?.

They were writing about different issues: gene therapy, stem cell therapy and electronic health records (EHRs):

For several decades now the biomedical research community has pursued a narrative positing that a combination of ever-deeper knowledge of subcellular biology, especially genetics, coupled with information technology will lead to transformative improvements in health care and human health…. [We] show that this approach has largely failed; and propose a wholesale reevaluation of the way forward in biomedical research.

They start with the primacy of the narrative:

In 2016 approximately $15 billion of the $26 billion of extramural research funding sponsored by the National Institutes of Health (NIH) could be linked to some version of search terms that include gene, genome, stem cells, or regenerative medicine. These topics have also increased geometrically in their representation among published articles… Apparently a large number of scientists either believe in the potential of these topics or feel compelled to work on them, recognizing that these topics constitute a major locus of important science, financial support, recognition, and prospects for a successful career.

Yet the extravagant predictions about curing most disease including cancer have not come to pass.

For example:

The complex and adaptive nature of most tumors thwarts the optimistic projections for molecularly targeted therapy for cancer. A randomized trial of targeted therapy based on molecular profiling … showed no improvement in progression-free survival… So far, just 2.5% of screened patients have been assigned to a trial intervention group… [T]he rarity of the targeted mutations means that this approach will help only a minority of patients with cancer.

This is hardly the only “big idea” whose benefits have failed to materialize. Stem cell research has been particularly disappointing, and the implementation of electronic health records (EHR) has had a disastrous impact physician morale without the promised improvement in patient care.

Sound familiar? The benefits of the big idea about breastfeeding — that increased rates will lead to diseases prevented and lives and money saved — has not merely failed to materialize, breastfeeding promotion turns out to have significant harms to infants, mothers and the bottom line. Demonstrating that breastmilk does reduce the risk of necrotizing enterocolitis (NEC) in preterm babies merely highlights the fact that breastfeeding can have substantial benefits in specific situations while simultaneously having no benefits for most.

How do those who promote big ideas respond when their claims cannot be substantiated?

They have two choices:

The first is to continue with calls for more funding, more complex measurements, and more sophisticated instrumentation. The second is to reevaluate and reset the current focus.

Thus far breastfeeding professionals continue with calls for more funding and more complex measurements, while claiming ever more arcane “benefits” (the microbiome! epigenetics!)

What if we were to re-evaluate?

When NIH funds translational or preclinical research with specific deliverables promised (as in the case of personalized medicine, and stem cell therapy), independent assessors should regularly appraise whether these deliverables were achieved and, if so, at what cost, and with what effect. Assessors must be objective, independent of the funding source, and have no professional stake in whether a particular line of research is deemphasized. The deliverable criterion should include public health benefit achieved by these initiatives (ie, measurable reductions in mortality and morbidity). Criteria such as number of publications, citations, prizes, and recognition are irrelevant as these are simply self-rewarding artifacts of the system…

How would that work for breastfeeding?

Independent assessor should regularly appraise whether the claimed benefits of increased breastfeeding rates — conditions prevented, lives and dollars saved — are actually achieved, at what cost, and with what iatrogenic complications for babies, mental health complications for mothers and costs to treat those complications.

Criteria such as the number of publications or the support of authoritative healthcare organizations should be recognized as irrelevant since they are simply self-rewarding artifacts of a system that continues to promote entrenched ideas long after they have been disproven.

The fundamental question:

Has aggressive breastfeeding promotion improved quality of life and life expectancy, by how much, for how many, and for whom?

Despite extravagant predictions, the benefits of breastfeeding have been limited to preterm babies. Breastfeeding cannot be shown to have prevented major disease, saved lives or extended life expectancy. And it can be shown to have caused tens of thousands of hospitalizations per year at the cost of hundreds of millions of dollars.

Ioannidis and colleagues declare:

Mechanisms should be in place to sunset underperforming initiatives.

In breastfeeding research and practice, NO mechanisms exist to sunset existing initiatives; instead massive effort is devoted toward increasing funding for them. Why? Because careers, reputations and economic support for lactation professionals and breastfeeding researchers rests on maintaining the fiction that breastfeeding has major benefits and no risks when the reality is that it has few benefits and significant risks.

The history of medicine is a history of a some excellent big ideas among a much greater number of underperforming big ideas. There’s no shame in the fact that many big ideas turn out to be bad ideas. The only shame is in refusing to recognize it.

10 Responses to “What should we do when underperforming big ideas — breastfeeding saves lives — become entrenched?”

  1. mabelcruet
    July 14, 2019 at 4:25 pm #

    Slightly OT-a lovely midwife story in the news recently:


    He’s picked up on how women have such a downer on themselves by not having what they think is ‘the perfect birth’. If more midwives had this sort of attitude, I think it would definitely benefit mums.

  2. CodeWench
    July 13, 2019 at 1:46 pm #

    What’s the problem with EHRs?

    • PeggySue
      July 13, 2019 at 3:25 pm #

      The products were not designed with any consideration of user work flows. The interfaces are in many cases inefficient and counter-intuitive. And there was a rush to produce these things to take advantage of incentives for their use. I used to work in software development. This ALWAYS results in products that are buggy and inconsistent. And the products don’t talk to each other. If you’re in a hospital that uses system X and are transferred emergently to one that uses system Y, your medical information is not easy to share. It goes either in poorly designed paper records or by hurried phone conversation. So the EHRs have not delivered on the promise of less errors or more effective information sharing. They’ve made charting more cumbersome and time consuming.

    • MaineJen
      July 15, 2019 at 8:52 am #

      Where do I start???

  3. swbarnes2
    July 12, 2019 at 3:33 pm #

    At least with the big genomics examples, the scientists are willing to admit “This panel of variants didn’t help us predict warfarin response” or whatever. Lactivists don’t seem to be willing to admit even that. As an example see this paper:


    This paper studies breastfeeding and behavior and cognitive functioning (7 different aspects). They tested at 3 years and 5 years. And they binned the babies by breastfeeding, not at all, 1 month, < 6 months, more than 6 months. So basically the odds of them not finding a single association at all was pretty low. And after correcting for everything, they found… exactly one. That disappeared at 5 years. But the paper authors could not bring themselves to say that they found nothing significant. They still insist that their paper is a positive finding for breastfeeding.

    • July 13, 2019 at 12:06 am #

      Yup, they found a weak correlation that kids who were breastfed exclusively for at least 181 days were less likely to be rated by their parents as hyperactive at age three.

      I’d be much more impressed if they could have corrected for maternal/paternal/sibling diagnosis of AD(H)D since that can run in families.

      Also – and I’m just spitballing here – might the ability of a parent to breastfeed a kid exclusively for 6 months be related to the energy level or personality traits of the kid? My son let us know when he was ready for baby foods, finger foods and drinking from a cup by making determined efforts to wrestle things away from us while we were eating. So…even if I could have EBF and wanted to EBF to 6 months, I’m pretty sure my son would have chomped down on something or started lapping up applesauce he snagged from my lunch while I was distracted.

      • GeorgiaPeach23
        July 13, 2019 at 4:55 pm #

        My baby has adhd in his genetics and is also a very distracted eater, which is one of the main reasons I weaned <6 months. As a data scientist I am compelled to note that breastfeeding duration and hyperactivity are likely not linearly independent variables.

  4. mabelcruet
    July 12, 2019 at 2:33 pm #

    Epigenetics and the microbiome are old hat now. The next big ideas that they are seizing on are molecular metabolite variations to prevent childhood obesity; and oligosaccharide research, particularly 2′-Fucosyllactose which has prebiotic, antimicrobial effects, affects gut maturation, and is responsible for immune modulation. Leaky gut is real-get with the programme!

    • Ruth Mayfly
      July 13, 2019 at 6:26 am #

      Oi vey.
      Can’t they just leap straight to ‘it’s MAGIC’ and spare the rest of us?

      • mabelcruet
        July 13, 2019 at 11:25 am #

        No, they love the long, sciency words like galactopoiesis and glactogenesis, because it makes them sound like they know what they are talking about and they can scare mums with it. So ‘oligosaccharide polymers’ and ‘immunodysregulation’ are up there with all the rest of the polysyllabic nomenclature to impress folk. If a mum is paying £100 an hour to a lactation consultant for advice, you have to throw in some long words in order to justify that rate.

Leave a Reply

You must be logged in to post a comment.