The short version
- Accuracy is the settled part. In the knee, guided injections landed correctly 92.7% of the time versus 77.9% blind — meaning roughly one in five blind knee injections misses.
- Whether that changes outcomes is the genuinely contested part, and the honest answer is a modest yes for the knee on thin evidence.
- It matters more for some joints than others — deep, small, or hard-to-palpate targets gain most.
- A blind injection by someone who does hundreds beats a guided one by someone who does a handful. The operator is not a rounding error.
This is the page where I have the most obvious conflict of interest on the whole site. Ultrasound guidance is what I do, and thousands of times over. So the useful version is not an argument for it — it is what the literature actually shows, including where it is weaker than I would like.
How often does a blind injection miss?
More often than most patients assume, and the number depends heavily on the joint.
A review comparing guided and blind injection in the knee and shoulder found guided knee injections were significantly more accurate — 92.7% versus 77.9%. Put plainly, about one in five blind knee injections does not end up in the joint.
The knee is one of the easier targets. It is large, superficial, and has reliable landmarks. Joints that are deeper, smaller, or obscured by soft tissue are harder still, which is where guidance earns most of its keep.
Does better accuracy mean a better result?
This is the honest crux, and for a long time the answer was "not clearly."
A systematic review and meta-analysis of randomised trials comparing guided with landmark injection in osteoarthritis found significant improvements favouring ultrasound in procedural pain, pain at follow-up, and function.
The first of those is worth pulling out, because patients ask it and assume the opposite. Does a guided injection hurt more? On this evidence, no — procedural pain was lower with guidance, and it was the largest of the three effects. Watching the needle go where it is meant to go generally means less probing to find the target.
Read the size of that evidence base
That meta-analysis pooled four randomised trials and 338 patients, all with knee osteoarthritis. It could not answer the question for hip or hand at all, because the trials do not exist.
So the fair statement is: for the knee, on a small body of evidence, guidance appears to produce modestly better pain and function outcomes as well as better accuracy. For most other joints, the accuracy advantage is established and the outcome question is genuinely open.
The shoulder illustrates the gap. The same review found guided injection showed modest accuracy benefit in cadaver studies of the glenohumeral and acromioclavicular joints, but improved accuracy had not been demonstrated in clinical studies, with insufficient data for the subacromial bursa. A separate randomised trial in chronic subacromial bursitis found greater improvement in passive abduction and some quality-of-life measures with guidance — while both groups improved substantially on pain and range of motion.
Where it matters most
| Target | Accuracy without guidance | Does guidance change outcomes? |
|---|---|---|
| Knee | Reasonable — around 78% correct blind, rising to ~93% guided. | Modest benefit shown for pain and function, on four small trials. |
| Shoulder — subacromial | Variable; insufficient pooled data to state a figure. | One randomised trial found some benefit; both groups improved. |
| Shoulder — joint itself | Modest cadaver benefit; not demonstrated clinically. | Not established. |
| Hip | Deep and not reliably palpable — imaging is standard practice. | No randomised evidence either way. |
| Small joints, tendon sheaths, bursae | Hardest targets to hit blind. | Not formally studied; the anatomical case is strongest here. |
What guidance actually changes for me
The accuracy figures are the argument people cite, but the thing I would not give up is seeing what is there before injecting. An effusion I did not expect. A tendon sitting where the needle was going. A joint that turns out to be far more degenerate than the story suggested. Sometimes that changes the target; occasionally it means not injecting at all.
That is a judgment about how I work rather than a trial finding, and it is not something the accuracy literature measures. But it is the honest answer to why I use it for everything rather than only the difficult targets.
Should you seek out a guided injection?
Some practical framing, given that the answer is genuinely "it depends":
- For a straightforward knee injection by an experienced clinician, blind is reasonable and most will be accurate. Guidance improves the odds.
- For deep or small targets — hip, small joints, tendon sheaths, anything near a structure you would rather not hit — guidance is the sensible default.
- If a previous blind injection did nothing, a missed target is one plausible explanation among several, and worth raising before concluding that injections do not work for you. The other explanations are worth considering too.
- Volume matters more than equipment. An experienced clinician injecting blind will usually outperform an inexperienced one holding a probe. If you are choosing, ask how often they do the procedure — not just whether they use ultrasound.
That last point cuts against my own interest and is the one I would most want a patient to hear.
Sources
- Oo WM, Linklater J, Siddiq MAB, Fu K, Hunter DJ. Comparison of ultrasound guidance with landmark guidance for symptomatic benefits in knee, hip and hand osteoarthritis: systematic review and meta-analysis of randomised controlled trials. Australasian Journal of Ultrasound in Medicine 2024;27(2):97–105. Four RCTs, 338 patients, all knee OA. Retrieved via PubMed.
- Bookman JS, Pereira DS. Ultrasound guidance for intra-articular knee and shoulder injections: a review. Bulletin of the Hospital for Joint Disease 2014;72(4):266–270. Accuracy 92.7% guided versus 77.9% blind in the knee. Retrieved via PubMed.
- Hsieh LF, Hsu WC, Lin YJ, et al. Is ultrasound-guided injection more effective in chronic subacromial bursitis? Medicine & Science in Sports & Exercise 2013;45(12):2205–2213. Randomised trial, 92 patients. Retrieved via PubMed.