A research study assistant slid the informed consent document for the clinical trial across the desk to us.

My wife, Ruth, sitting next to me, signed it. She was in treatment at Memorial Sloan-Kettering Cancer Center, where I am also a doctor and cancer researcher.

Ruth had flipped through page after page of the informed consent forms. They were peppered with frightening words like “death” and “cancer recurrence,” and caveats discussing horrifying reactions to the experimental treatment. I winced each time I saw a scary term go by.

The clinical trial she was joining would test a new treatment for her breast cancer. Some women in the study would get it, while others would get the old standard. But during the trial, none of us, not even the doctors conducting the research, would know who was getting which drug.

It was, essentially, a coin toss.

The trial, which wasn’t designed to help her or to harm her, haunted both of us. The truth was that the new treatment could turn out to be worse than the standard approach, and we would not know in time to stop it.

This hard reality of cancer research is often tucked away out of view, but we both sensed it every time her nurse hung the mystery bag in the chemo suite. It might be the study treatment, or maybe it’s just saltwater. And then it would start dripping in.

Ruth woke me one night with a rhetorical barrage about the reasons for randomization in clinical trial research.

“Why do they need to randomize patients so some people get the new treatment but others only get the standard?” she said. “Don’t you guys know after all these years what happens so you could just do a study where everyone gets the new treatment and see if they do better? Aren’t you wasting time not just using it?”

Even half asleep I could explain that randomized trials are the best way of determining if one treatment is better than another. But as I began talking, I realized she had drifted back to sleep, just as my medical students do in class. Her chest rose and fell with her breathing. The scar from her cancer surgery was fully healed. The skin on her face was relaxed.

The truth is that even though we always expect the new treatment to be better, sometimes it is not. If there are benefits, they are usually so small that they are perceptible only when added up over hundreds or sometimes even thousands of patients carefully compared with similar patients getting the standard treatment.

This depressing fact meant that even if Ruth was getting the new treatment, and it proved to be effective, it would probably be only a hair better. Not enough to really matter much.

And in clinical trials, the new treatment sometimes proves worse. Every time this happens people are surprised and some doubt the results, even though it happens regularly. For instance, thousands of women with breast cancer were subjected to complications, and some killed, because doctors were sure that high-dose chemotherapy followed by bone marrow transplant would be more effective than conventional chemotherapy approaches. Randomized trials showed the opposite.

Most doctors know by rote some of the other examples of treatments that failed: an operation that removed diseased parts of the lung and left the healthy part behind, a drug that made the heart rhythm look better after a heart attack, a monitor placed inside the lung vessels that helped monitor patients who were critically ill. Each time we were sure we were doing the right thing for patients, until randomized trials showed we were killing people.

Scientists call our progress in cancer treatment a gradual accretion of knowledge, but put in regular English it is a slow, grueling slog guided by a poorly functioning compass. Most randomized trials like the one Ruth joined don’t even ask particularly novel questions, and their findings are not breakthroughs worthy of newspaper headlines.

The studies tweak what we do as doctors, sometimes finding a little more benefit achieved with a little less harm. More often than not, they just find out that the new approach did not work out.

Of course this reality — this arduous slog — doesn’t match how cancer research is presented to the public, which sees flashy headlines using words like “breakthrough” and “new hope.”

As I watched Ruth sleep, I thought about the hundreds of thousands of women who had volunteered over decades for clinical trials, those who did well and those who didn’t. Their privacy protected by laws, the depth of their experiences reduced by statisticians to just a few numbers and figures in the pages of a medical journal.

Each study was a small contribution, paving the way a few feet forward or shutting down a detour that went in the wrong direction. Over time, what they contributed has helped us get a little better each year at treating breast cancer.

Signing up for a clinical trial is a remarkable act of bravery. The day before Ruth’s diagnosis she did not think of herself as a patient. She had no doctor in her life except me. And yet, when Ruth’s oncologist asked for her to sign up to be randomized, to let Lady Luck have yet another go at her, she did so.

Ruth knew he was asking her to do something for others even when she was overwhelmed by what was happening to her, to us. But it was the obvious choice, and she made it without hesitation, just as hundreds of thousands of women have before her.

Now, as she slept through another fitful night, she was helping to answer a question that would not matter to her, but would someday matter to a woman she would never meet.

Dr. Peter B. Bach is an attending physician at Memorial Sloan-Kettering Cancer Center in New York City. His essays, about his wife’s breast cancer, appear Tuesdays on the Well blog.