If I could sum up my breast-feeding ordeal in one image, it would be me sitting in a hospital bed with one of my newborn sons cradled in my arms.

Above me hovers the lactation consultant, her complexion dewy, her curls buoyant. She clearly hasn’t endured a 38-week twin pregnancy — that’s 14 pounds of baby, baby — followed by a cesarean and 48 hours of unsuccessful breast-feeding, and if she’s trying to hide her impatience with my mental and physical state, she’s not succeeding.

I was supposed to leave the hospital an hour ago, but I can’t go, in the opinion of the lactation consultant, because, up to this point, I’ve done just about everything wrong.

I now realize my not leaving was a little extreme — not to mention spineless. But I was acting on my absolute conviction, inspired and repeatedly reinforced by pediatricians, authors and the U.S. government, that breast is best by a wide and indisputable margin. Now it’s starting to look like I, and millions of others who made significant sacrifices based on that assumption, may have been wrong.

No one is saying that breast-feeding isn’t perhaps, by some measures, a little bit better for your baby’s health, but the argument coming from critics such as Joan Wolf, an associate professor of women’s studies at Texas A&M University, is that the actual scientific research doesn’t say much beyond that. In her 2011 book, “Is Breast Best?” Wolf repeatedly cites studies with conflicting outcomes, important methodological problems, and results far less dramatic than we’ve been led to believe.

“My biggest concern is that we have blown infant feeding completely out proportion,” Wolf says. “I’m certainly not against breast-feeding in any way, and I think it’s a perfectly viable choice for women who want to do it. But I think it’s very hard right now for a woman to decide whether to breast-feed — there are so many external pressures telling them that it’s the only thing to do, the only reasonable thing to do if you care about your baby. And that concerns me — that we have skewed the alternatives here in ways that are completely inaccurate.”

Breast-feeding advocates acknowledge that the data isn’t perfect. Because scientists can’t ethically perform experiments in which women are randomly assigned to breast-feed or bottle-feed, they can’t be completely sure that better outcomes are actually caused by breast-feeding, as opposed to the type of mothering associated with breast-feeding, according to Dr. Richard Schanler, chairman of the American Academy of Pediatrics’ section on breast-feeding.

But Schanler points to an analysis of existing studies in an AAP policy statement that he co-authored. In the analysis, more breast-feeding was associated with better health outcomes.

“When you can’t do a randomized trial, one of the next things you’d like to do is a dose response: Do you get more effect with more breast-feeding? And you see that with a lot of the outcomes that we’ve measured,” he says. “To me, that’s very telling. But can I swear that this is causation, or just association? I can’t, but I have a feeling a lot of this is causation.”