How the Osteoporosis Screening Recommendation Was Built

The short answer#

In January 2025 the US Preventive Services Task Force (USPSTF) reaffirmed a B recommendation for osteoporosis screening in all women 65 years and older, and a B for postmenopausal women under 65 who carry added risk. A B grade means moderate certainty of moderate net benefit. For men the Task Force issued an I statement, meaning the current evidence is not enough to weigh benefits against harms. The statement appeared in JAMA (2025;333(6):498-508). The letter on the page is short, but it sits on top of a long argument, and that argument is the interesting part of the whole thing.

Key points#

• A B grade is a statement about confidence and size of benefit, not a marketing endorsement.
• The argument runs through three questions: can a scan sort people by risk, do calculators add much, and does screening then treating actually prevent fractures.
• Central DXA and the FRAX calculator each help, but neither is decisive on its own.
• Most fragility fractures happen in people who fall short of the osteoporosis threshold, which is why the guideline reads as triage rather than prophecy.
• Men get an I statement because the trials were run almost entirely in women.

What the grade is actually promising#

It helps to be clear about what a USPSTF letter is promising you. The question behind it is not "is this test good" but something more demanding: across a whole population of people who feel well, does routine screening leave the average person better off once every benefit and every harm is tallied. A B grade answers that with moderate certainty of moderate net benefit.

That word certainty is worth pausing on. It describes how solid the underlying evidence is, which is a separate matter from how large the payoff might be. A practice can genuinely help people and still earn a cautious grade if the studies backing it are few, indirect, or drawn from a narrow set of patients. For bone health the payoff that matters most is a prevented fragility fracture, above all a hip fracture, which for an older adult can be the event that ends independent living. So the real question is narrow and practical: does measuring bone in people without symptoms, and then acting on what the measurement shows, prevent enough of those fractures to justify the effort and the downsides.

Three questions behind the letter#

The cleanest way to read the recommendation is as a chain of three linked questions. Each link is graded for what it can and cannot carry.

Can a scan sort people by risk#

The measurement at the center is central dual-energy x-ray absorptiometry (DXA) of the hip or lumbar spine. Osteoporosis is defined by a bone mineral density at least 2.5 standard deviations below the young-adult reference, a T score of -2.5 or below. The evidence review found that DXA does predict fractures, with discrimination (area under the curve) for major osteoporotic fracture landing somewhere between roughly 0.60 and 0.80 depending on the study and the outcome measured.

Think of that range the way you would a weather forecast. An AUC near 0.60 is a forecast that beats a coin toss but not by much, while a value close to 0.80 is genuinely useful for planning. Bone density is a real signal, but it is one reading among several, and by itself it separates future fracture patients from everyone else only imperfectly.

Do the risk calculators add much#

This is where FRAX comes in. FRAX blends age, sex, prior fracture, family history, smoking, glucocorticoid use, and other factors, and it can run with or without a bone density value plugged in. In the review, FRAX for ten-year hip fracture risk in women reached AUCs of about 0.74 to 0.77 without a DXA number and 0.76 to 0.79 with one. Feeding the bone density value into the calculator nudged the accuracy up only a little.

That small gap shapes the advice for younger women. The Task Force lays out a two-step path: first check whether a postmenopausal woman under 65 has one or more risk factors, then run a clinical risk tool to decide whether a DXA scan is worth ordering. FRAX is one option, with instruments like OST and ORAI alongside it. The statement is candid that several of these tools were built on small or fairly uniform groups of people, which limits how well they travel to everyone. Calculators frame the decision; they do not close it.

Does screening then treating prevent fractures#

Here is the link that carries the most weight and, it turns out, the least direct evidence. A scan can be accurate and a calculator can be sensible, yet the whole point stands or falls on whether the full sequence, find low bone density and then treat it, actually lowers fractures. The review found three randomized trials that pitted an organized screening program against usual care with fracture as the outcome: SCOOP, ROSE, and SOS, all run in older women in Europe. Pooled together, they pointed toward fewer hip fractures, with a relative risk near 0.83, and an absolute difference on the order of a handful fewer hip fractures per thousand women screened over several years.

Those effects are real but modest, and they come from a fairly narrow slice of the world. That is precisely why the recommendation lands at B rather than A. The direction of benefit is consistent, but the direct trial base is small, geographically limited, and does not speak evenly to every group of people.

The counterintuitive part#

One finding is easy to misread and deserves its own spotlight. Most fragility fractures do not happen in people who meet the T score definition of osteoporosis. The reason is arithmetic rather than biology: far more people carry moderately low bone density than carry severe deficits, so the larger raw number of fractures comes out of that broad middle group. A strict density threshold flags the individuals at highest personal risk while still missing much of the total fracture burden across a population.

The practical lesson is that a screening program built around a cutoff is a sorting tool, not a personal forecast. It tells you who sits in the higher-risk bucket; it does not promise that you will or will not break a bone. The 2025 statement is careful to claim no more than the numbers allow.

Why men receive an I statement#

The I statement for men is often misread as "screening cannot help men." It says something narrower and more honest: the evidence has not been generated. The predictive data and nearly all of the treatment trials came from women, and the Task Force declined to assume a net benefit in men that had never been directly tested. An I statement is the system reporting a blank where a blank exists, rather than filling it in with assumption.

How to read a recommendation like this#

The 2025 statement rewards being read as an audit trail rather than a verdict. Every link is on the page and labeled for how much it can bear: the accuracy of DXA, the modest lift from FRAX, the small and regionally limited treatment trials, and the paradox of where fractures actually fall. Nothing is smuggled in. That transparency is what a well-built recommendation looks like, and it is also why the letter grade should never be the end of the conversation. Whether you are screened still belongs in a discussion with a clinician who knows your details.