Two people sit in an exam room. One has an illness that will end their life. The other knows. Neither wants to say it out loud, so neither does. The conversation drifts toward next steps (another scan, another line of therapy, another visit on the calendar), and both leave the room feeling that it went well.

This is not a thought experiment. In 2012, the New England Journal of Medicine published a study by Jane Weeks and colleagues that should have changed practice overnight. Among patients with metastatic lung or colorectal cancer (cancers for which chemotherapy is not curative), 69 percent and 81 percent respectively did not understand that their treatment would not cure them. That alone is damning. Worse: Patients who rated their physician’s communication as excellent were more likely to hold inaccurate beliefs about cure than those who rated it less highly. The better the conversation felt, the less the patient understood about what was actually happening.

That was 2012. Same finding turned up again recently: 986 advanced cancer patients, 74 percent still misperceiving their treatment’s intent. The pattern is alive and well.

The mechanism Weeks named was collusion. Patient and physician quickly move from prognosis to plan, generating a shared optimism with no anchor in the disease. Treatment becomes the instrument of misunderstanding. We don’t tell them the truth; we tell them the plan. Both parties feel better about it, right up until the plan stops working and there is no time left to do the things that mattered.

Two people in a room, neither wanting to hear the hard thing out loud, including the one wearing the embroidered Patagonia.

Start with us. Sixty percent of physicians say making survival estimates is stressful and difficult. Telling someone the time they have left is shorter than they think is grief work performed in real time. We dismiss it as a training deficit. It is really a signal that something hard is happening.

The system makes it harder. There is no billing code for sitting with a family for forty-five minutes and saying the hard thing clearly. Procedures get RVUs. Honesty does not. Ask oncologists when they would discuss hospice with a newly diagnosed metastatic patient. Only 14 percent say “now.” The rest wait.

Then there is the alibi we reach for to make the silence feel principled. I don’t want to take away their hope. They didn’t ask. I’m respecting their autonomy. Each has the texture of an ethical principle and the function of a shield. Autonomy without information is a fiction. You cannot exercise self-determination about a future you don’t know is coming. If I hand you a ballot with the candidates’ names blacked out, I have not respected your vote.

Now the patients. Most of them don’t ask either. They sense the room. They feel the rhythm of the visit shifting away from prognosis and toward plan, and they let it shift, because the plan is easier to hold than the prognosis. They are doing what we are doing, for the same reason. That is what makes the collusion stable: Both parties want to leave the room intact, and the easiest way is to not look at what is in it.

The most durable rationale for the silence is that telling the truth would crush hope. The data say otherwise. Researchers tested this directly. They followed patients with advanced cancer and tracked depression, anxiety, and suicidal ideation. Patients whose physicians discussed prognosis honestly fared no worse psychologically. They fared better practically: more advance care planning, more time at home, earlier hospice referral. Patients who knew what was coming organized their remaining time around it.

We are using one word to do the work of several. The hope that this cancer will respond to the next line of therapy belongs to a different category than the hope to live until a daughter’s wedding, the hope to die at home, the hope to mend a relationship while there is still time. The first is a clinical prediction. The others are lived priorities. The first is often false. The others can almost always be honored, but only if we know what the goals are, and we cannot know if we never said the time was short. When clinicians treat all hope as the first kind, we strip patients of the chance to invest in the rest. The mythology of preserving hope has often meant denying patients the information they need to direct what hope is left.

Ask the families of critically ill patients what they want. Researchers did: 179 surrogate decision-makers in intensive care units. Ninety-three percent said avoiding the conversation about prognosis is an unacceptable way to maintain hope. The people who will be left to make decisions are asking us to be specific.

A few honest qualifiers. The evidence here leans heavily on cancer. It applies less cleanly to advanced heart failure, dementia, and the long arc of multimorbidity. Cultural variation matters: Direct disclosure is itself a culturally specific value, and in some families the honest path runs through a different language and different people. And no clinician’s improved technique can fix a system that rewards procedures over presence. None of those qualifiers rescues the comfortable lie. They sharpen the question. They do not change the answer.

The silence is what humans do in a hard room. The response must be human too. We name the impulse to look away and push back gently and on purpose. We say the time is short, and we say it specifically: months when they have months, weeks when they have weeks. We treat the prognostic conversation as the clinical intervention it is, and resource it like one.

Patients are asking for honesty they can act on. The collusion serves us more than it serves them. We can stop.

Kyle Edmonds is a palliative care physician.