Body Composition Part 2: How It's Measured in Clinical Trials, and How That's Changing

This is part two of our three-part series on body composition. In part one, we established that two people can lose the same amount of weight and one loses muscle while the other loses fat—the scale shows the same number either way. In this post, we look at how body composition actually gets measured, from gold-standard clinical tools to the smartphone-based and at-home methods changing what's possible in decentralized research. Part three will show you how to put these measurement tools directly into an Alethios study.

In part one of our Body Composition blog series, we established that two people can lose the same amount of weight and one loses muscle and the other loses fat—the scale shows the same number. Body Mass Index (BMI) and weight don't give a comprehensive picture of health, weight fluctuation, or body makeup. Body composition, however, gives insight into the body's complexity.

Why Do People Use BMI and Weight?

While sometimes weight or BMI offers the information a researcher needs, often body composition would work better. So why do researchers still use BMI and weight? It's the same answer we see across health research; there's an accessibility problem.

The gold-standard method for measuring body composition, DEXA (dual-energy X-ray absorptiometry), requires specialized equipment and a clinic visit, which makes it expensive and slow to deploy at scale. That friction is a key challenge that keeps this kind of data out of decentralized studies and on-site studies alike.

"Smart scales," or other bioelectrical impedance devices, are cheaper and more portable but imperfect solutions that often trade away some accuracy, especially outside a controlled setting (Journal of Exercise Rehabilitation). The trade-off between accuracy and accessibility is exactly the problem that newer measurement approaches are bridging.

How Body Composition Has Traditionally Been Measured

Beyond DEXA, researchers have had a handful of other clinic- and lab-based tools for measuring body composition, each with its own tradeoffs between accuracy and practicality.

Hydrostatic (underwater) weighing, long considered a gold-standard method alongside DEXA, estimates body fat by comparing a person’s weight on land with their weight while fully submerged. This highly accurate approach requires a specialized tank and a willingness to go underwater, which limits its use primarily to research and university settings (hydrostatic weighing overview).

‍Air-displacement plethysmography, commercially known as the Bod Pod, works on a similar principle using air rather than water, and has been validated against other reference methods across a range of body types, offering DEXA-level rigor without the water (Bod Pod validation study; validation in overweight and obese adults). At the more accessible end, skinfold calipers—which estimate body fat by measuring the thickness of a pinched fold of skin at a few standard sites—are cheap and portable, but their accuracy depends heavily on the skill of whoever is doing the measuring, and studies comparing calipers to imaging methods like ultrasound have found meaningful discrepancies between technicians (skinfold caliper reliability study; caliper vs. ultrasound comparison). Each of these methods works well in the setting it was designed for—a lab, a university exercise physiology department, a specialist's office—but none of them were built with a decentralized study in mind.

Closing the accuracy-accessibility gap doesn't mean choosing between a $150 clinic visit and a bathroom scale that only tells you your weight. It means meeting participants where they already are—starting with the device already in their pocket.

Digital Methods Using a Smartphone

A regular smartphone camera has quietly become a legitimate body composition tool. Validated 2D photo-based models can estimate body fat percentage from just a couple of images, using deep learning trained against DEXA scans as ground truth (npj Digital Medicine, smartphone camera adiposity study; AI 2D-photo method, npj Digital Medicine). The technology has moved beyond flat photos, too: newer approaches reconstruct a 3D avatar from a short phone-camera video, which can estimate regional body composition—not just a single body fat percentage, but where that fat and lean mass are actually distributed (smartphone 3D imaging, PMC; mobile phone 3D body composition estimation, PMC). For a decentralized study, that means a meaningful body composition readout without shipping a single piece of equipment.

One company built specifically around this approach is Prism Labs, which uses a smartphone camera to generate a 3D body scan and estimate body composition without any specialized hardware. Prism Labs has positioned its technology as a scalable alternative to DEXA for research use specifically, arguing that consistent, repeatable smartphone scans can support body composition tracking across large study populations without the cost and scheduling burden of clinic-based imaging (Prism Labs, body composition research solutions; Prism Labs, DXA alternative white paper).

Smart Scales and Bioelectrical Impedance Devices

Bioelectrical impedance devices are already a familiar object in a lot of homes—a scale that sends a barely-perceptible electrical current through the body and estimates fat versus lean mass from how much resistance it meets. Newer multi-frequency BIA devices, tested against DEXA in healthy adults, show strong enough agreement to be useful research instruments rather than novelty gadgets (Frontiers, 2026 multi-frequency BIA validation), and portable bioimpedance analyzers built specifically for remote monitoring have held up under controlled clinical testing (portable bioimpedance analyzer study). Even the more basic function—weight—holds up well at home: in-home smart scale readings show high concordance with in-person clinical measurement (concordance study, PMC), which is reassuring baseline evidence that these devices behave consistently outside a lab.

Validation and Accuracy

None of these tools are trying to replace DEXA as the clinical gold standard—the question that actually matters for research is whether a given tool's error is small and consistent enough to trust for the specific thing being measured. A comparative study on BIA device reliability and validity found correlations strong enough to track meaningful within-person change over time, even where absolute values diverged somewhat from DEXA (reliability and validity of BIA measurement, PMC). That distinction is the important one for longitudinal research: a device doesn't need to nail someone's exact body fat percentage on day one if it can reliably and repeatably detect that the number changed by day ninety. For most research questions—did this intervention preserve muscle, did this program build it—trend accuracy is the accuracy that counts.

Bringing It All Together

This is the challenge Alethios is built to solve: combining digital, cutting-edge AI solutions and at-home devices to capture body composition data with the rigor a body-fat-versus-muscle question demands—without requiring a single clinic visit.

Curious what body composition data could add to your next study? Book a demo.