A stadiometer only produces a reliable number if the patient can stand fully upright, unassisted, for the few seconds it takes to take the reading. For a meaningful share of the people dietitians actually assess — those with contractures, advanced scoliosis, lower-limb amputation, or who are simply too frail or acutely ill to stand — that condition is never met. Height still has to go on the chart, because it drives BMI, estimated energy and protein needs, and most malnutrition screening tools. The practical question isn’t whether to estimate height when a stadiometer can’t be used; it’s which validated proxy method to reach for, and how much error to build into the interpretation.
Why a Guess or an Old Self-Report Isn’t Good Enough
Self-reported height is a common workaround, but it drifts over time — older adults in particular tend to report a taller height than they currently measure, since it reflects their remembered adult peak rather than the height loss that comes with age-related vertebral compression. Skipping height entirely isn’t a safer default either: without it, BMI and ideal-body-weight calculations default to population averages that can misclassify a genuinely underweight or overweight patient. Validated proxy measurements exist precisely to close that gap with a documented, reproducible number rather than a guess.
Ulna Length: The Practical Choice for Contractures
Ulna length is measured from the point of the elbow (the olecranon) to the prominent wrist bone (the styloid process), then converted to an estimated height using published equations. It is the method most often reached for when contractures or scoliosis make the trunk and legs unreliable landmarks, because the forearm frequently retains enough range of motion to measure even when the rest of the body doesn’t. Published validation work reports a strong correlation with measured height (r around 0.96) and, in older adults, a small average overestimation — roughly half a centimeter in both men and women, with individual results varying by several centimeters either direction. In patients with contractures specifically, ulna-derived height has been shown to run about 2 cm short of true height on average, which is worth factoring in when the number sits right at a clinical threshold.
Knee Height: The Traditional Bedside Standard
Knee height is taken with a sliding broad-blade caliper, from the heel to the top of the kneecap, with the knee and ankle bent at roughly a 90-degree angle — a position that works whether the patient is supine in bed or seated at its edge. It has long been the default proxy for bedridden and wheelchair-using patients because the lower leg’s length stays relatively stable even as body composition changes. The trade-off is a wider error band than ulna length in practice: published error estimates for frail, older adults run from roughly 3 to 5 cm, and the standard equations (built on age, sex, and sometimes ethnicity) were derived from specific reference populations, so accuracy narrows or widens depending on how closely a given patient matches the group the equation was validated on.
Arm Span and Demi-Span: Fast, but Less Precise
Arm span — fingertip to fingertip with both arms extended horizontally — and demi-span, its half-body equivalent measured from the sternal notch to the fingertip and then doubled, are the quickest of the proxy methods and useful when a patient can sit up and extend at least one arm. Both correlate reasonably well with standing height, but published agreement in acutely ill older adults is weaker than the correlation numbers suggest: mean differences from actual height have run several centimeters in that population, generally overestimating rather than underestimating. Arm span in particular tends to exceed true height by an average of a few centimeters in adults, a gap that widens further with age-related height loss. These are reasonable methods for a quick screening estimate, but they carry the widest error margin of the group and are best treated as a fallback rather than a first choice when a more precise method is feasible.
Recumbent Length: Right for Infants, Riskier in Adults
Measuring length while the patient lies flat on a fixed board, with a headboard and a movable footboard, is standard practice for infants and children under two and produces results very close to eventual standing height in that age group. In adults, the same supine approach behaves differently: published comparisons show recumbent measurements running several centimeters longer than standing height on average, even though the two measurements correlate closely overall. Spinal decompression that happens naturally when a person lies down accounts for most of that gap, and the accuracy degrades further in patients with scoliosis, neuromuscular weakness, or contractures — the same conditions that make recumbent measurement necessary in the first place.
Choosing Between Methods
| Method | Best suited to | Typical error vs. true height |
|---|---|---|
| Ulna length | Contractures, scoliosis, limited mobility | Roughly 0.5–3 cm; tends to underestimate with contractures |
| Knee height | Bedridden or wheelchair-using patients generally | Roughly 3–5 cm in frail older adults |
| Arm span / demi-span | Quick screening when a patient can extend an arm | Often several centimeters, usually an overestimate |
| Recumbent length | Infants and children under 2 | Close to standing height in that age group; runs longer in adults |
No proxy method replaces a standing measurement when one is genuinely possible — a well-calibrated stadiometer is still the reference standard for a reason. But when it isn’t an option, the choice between ulna length, knee height, arm span, and recumbent length isn’t interchangeable guesswork. Each has a documented error profile, a population it was validated against, and a set of patient conditions it suits best. The habit worth building is simple: pick the method that fits the patient’s specific limitation, use the same method and equation at every follow-up so trends stay comparable, and note in the chart which proxy was used and why — so anyone reading the record later understands exactly how much confidence to put in the number.