Kim Johnson, a cancer epidemiologist at Washington University in St. Louis, begins her STAT First Opinion essay with a relative's death. Her aunt was 78 when she was diagnosed with stage 4 breast cancer, and she died three weeks later. She had lived in a long-term care facility since 2020, admitted for a sore on her foot and arthritis severe enough that she could barely walk. The cancer was found by accident, during a hospital stay for a urinary tract infection. She had felt a lump years earlier and never sought care, and after the diagnosis she blamed herself for not being screened and, in Johnson's account, promptly gave up on living.
The essay argues, correctly, that the nation's cancer screening guidelines have nothing to say about the setting where this happened. But the deeper problem is not that the guidelines are silent. It is what fills the silence. When no medical rule decides who gets screened, something else decides, and in long-term care the something else is logistics: who can arrange transportation, who has a family member to push, whether the facility has anyone whose job includes asking. The screening decision is being made, every day, by the ability to organize an off-site visit. That is not a medical standard. It is an eligibility rule written in scheduling.
The guidelines stop where the residents are
Start with what the major guidelines actually say. The U.S. Preventive Services Task Force recommends biennial mammography for average-risk women ages 40 to 74, and then stops: for women 75 and older, the task force concludes the evidence is insufficient to weigh the benefits and harms of screening. The American Cancer Society goes further in the other direction, advising that screening should continue as long as a woman is in good health and is expected to live 10 years or longer. The American College of Obstetricians and Gynecologists says the decision past 75 should rest on shared decision making between a woman and her clinician, informed by health status and longevity.
None of these documents is written for a resident of a long-term care facility, and the omission is not a technicality. The task force's guidance runs out at exactly the age, 75, that a meaningful share of facility residents have just entered. Johnson cites national data showing that 8.1 percent of breast cancer diagnoses in women 75 and older are stage 4 at detection, against 6.0 percent for women 40 to 74. The later the diagnosis, the fewer the options, and the residents who are oldest and least mobile are the ones the guidelines leave unaddressed.
The standard objection is that the median life expectancy after long-term care admission is short, roughly 2.2 years in the study Johnson cites, and screening a population with a two-year horizon invites overdiagnosis and treatment without benefit. The objection is real and must be taken seriously. It is also an average applied to a population that is not average. A resident admitted at 72 for arthritis has a different horizon than a resident admitted at 89 for end-stage heart failure, and the guidelines' silence treats them as interchangeable.
When there is no rule, the logistics become the rule
In the absence of a guideline, the screening decision defaults to whoever can make it happen. A mammogram for a facility resident means a referral, an appointment off-site, transportation, mobility assistance, often a family member or a staff member to accompany her, and follow-up for the results. Every one of those steps is a filter, and each filter selects for the same things: an engaged family, a facility with staffing to spare, a resident who can still travel. The woman who has none of those does not get a screening decision. She gets the absence of one, and the absence reads in the chart like a choice.
This is the administrative-friction problem in its purest form. The medical question of whether a given resident should be screened has a defensible answer in most cases, if anyone asks it. The reason nobody asks it is that the system has no slot for it: no guideline covers the population, no visit type is built for the conversation, no reimbursement rewards it. What the system does have is a process for the residents who can navigate it, and the process itself becomes the eligibility rule. The residents who are screened are not selected by risk or by preference. They are selected by transportation.
The essay's prescription follows from this. Johnson argues for guidelines specific to long-term care residents, for provider training in how to hold the conversation, and for decisions that are intentional, documented, and revisited as health status changes. That is the right ask, and it is worth noticing what it implicitly concedes: that the decision currently being made is none of those things. It is not intentional, it is not documented, and it is not revisited. It is whatever the calendar and the van schedule allow.
The evidence for stopping is thinner than it looks
The case for excluding older women from screening rests on modeling rather than trials, and the modeling has real gaps. No randomized controlled trial of breast cancer screening has ever included women 75 and older. The estimates come from cohort studies and simulation, and they carry the usual biases of both. One modeling exercise found that extending screening from age 74 to 79 would avert one additional breast cancer death per 1,000 women, at the cost of 67 to 172 additional false positives per 1,000. Another estimates that 47 percent of screen-detected cancers in women 75 to 84 are overdiagnosed. Competing mortality is the engine of the case: the 10-year risk of dying of something other than breast cancer rises from 14.5 percent at ages 66 to 74 to 65.4 percent at ages 85 to 94.
Those are honest, sobering numbers, and they are the reason the default for frail, very old residents should lean away from screening. But they do not license a blanket silence, and the essay is careful not to demand one. An ER-positive tumor like the aunt's is often slow-growing and can be managed with hormone therapy even after spread, and a lump felt and ignored for years is not the profile of a woman who would have been harmed by a conversation about it. The harms of screening are real. So is the harm of a treatable cancer found at stage 4 because nobody ever asked. The task force's "insufficient evidence" is an honest answer to a research question. It is a terrible answer to a specific woman, and that is the gap the facility setting widens.
The professional disagreement makes the friction worse. A motivated clinician in a facility has three documents pointing in three directions: the task force says the evidence will not tell her what to do after 75, the cancer society says keep going while the patient is healthy enough to benefit, and the obstetricians' group says work it out case by case with the patient. Each is defensible, and the disagreement means the screening conversation depends entirely on the individual clinician's willingness to have it. A guideline is supposed to be the thing that makes the decision consistent. Here the guidelines are themselves a source of variation, and the variation falls hardest on the patients least able to supply their own advocacy.
The family is where the information actually was, and the essay's account is specific about this. Johnson's aunt told the author's mother about the lump. The information existed, inside the family, for years. What never happened was the conversion: nobody with authority, in the facility or in a clinic, ever turned the knowledge that something was wrong into a screening decision, a referral, a conversation. A system that works converts signals into care. A system that works by logistics converts signals into care only when someone can drive the patient to the appointment, and the aunt could not drive and did not ask and blamed herself for all of it.
The excluded resident blames herself
The hardest detail in Johnson's essay is not the diagnosis or the timeline. It is that her aunt, after the diagnosis, blamed herself for not being screened and then, in the author's words, gave up on living. There is a specific cruelty in that sequence, and it deserves to be named. The decision not to screen was never hers. It was the product of a system with no guideline for her setting, no process for her situation, and no person whose job it was to ask. The system made the decision by default, and the woman on the receiving end experienced it as her own failure. That is what a hidden eligibility rule does at its worst: it excludes people and then lets them believe they excluded themselves.
Nothing about this analysis is an argument that every resident of every facility should be screened. It is an argument that the decision should be made deliberately, by a clinician and a resident and her family, against her actual health status and her actual wishes, and that the current default does not do that. The resident who wants screening and could benefit from it, like the resident who would decline it with good reason, deserves to be asked. The question costs one conversation. The van, the referral, and the family escort are only ever required for the women who say yes, and the point of a guideline is that the yes or the no should be hers, not the schedule's.
No one should reach the end of life blaming herself for cancer being detected too late. The blame belongs to whoever decided that a population of older, sicker, less mobile women could be handled with a silence in the guidelines and a gap in the system, and to the rest of us for letting the gap stand. A decision made by default is still a decision. It is just made by whoever declined to make it.
Primary sources
- Kim Johnson's STAT First Opinion essay for the account of her aunt's illness and death, the SEER stage-4 comparison, the 2.2-year life expectancy figure, the guideline summary, and the essay's prescriptions.
- The U.S. Preventive Services Task Force's 2024 breast cancer screening recommendation for the 40-to-74 recommendation, the insufficient-evidence statement for women 75 and older, the absence of trials in that age group, the modeling estimates, and the overdiagnosis and competing-mortality figures.
- The American Cancer Society's screening guidance for the 10-year life expectancy language, and the American College of Obstetricians and Gynecologists' bulletin for the shared decision-making position.