Do Ultrasound-Guided Treatments Actually Work for Shoulder Pain?

In the age of high-definition imaging and low tolerance for uncertainty, imaging often gets thrown into the clinical mix early, especially for shoulder pain. The logic feels solid: identify the pathology, tailor the treatment, improve the outcome. Clean. Linear. Evidence-based… right?

Well, not exactly.

Let’s discuss the Ottenheijm et al. (2016) RCT, which is one of the few trials that actually tested this scan-and-treat hypothesis in real-world primary care.

The Setup

Adults aged 18–65 with acute (<3 months) unilateral shoulder pain suggestive of a subacromial disorder. All patients received an ultrasound and two weeks of standard guideline-based care, including analgesics, reassurance, and activity advice. This qualification period aimed to filter out quick responders and avoid overtreatment. Those still symptomatic after two weeks were randomised into one of two groups:

1.  Ultrasound-Tailored Treatment: Clinicians saw the scan and matched treatments to the findings. Bursitis? Inject it. Tendinopathy or partial tear? Refer to physio. Full-thickness tear? Refer to ortho.
2.  Usual Care: Clinicians had no access to the scan. They treated based on symptoms and function. They used their brains and clinical judgement.

The Results

At 12 months:

•  Pain scores (SPS): Ultrasound group = 11.1 | Usual care = 11.5
•  Disability (SDQ): Ultrasound group = 24.3 | Usual care = 31.0
•  Recovery: 72.5% in the ultrasound group vs 60% in the usual care group (not statistically significant)

No meaningful differences. No clinically relevant effect. The ultrasound added complexity and cost but not outcomes.

The Hook

This trial, like others, shows a persistent theme: knowing the pathology and tailoring treatment to the pathology doesn’t mean you can change the trajectory. Ultrasound may improve diagnostic confidence, but that doesn’t automatically lead to better clinical decisions or patient outcomes.

Remember: structural findings on imaging are common, inconsistent, and often asymptomatic. Matching treatments to them assumes a clean cause-effect relationship that doesn’t hold up under scrutiny.

So, What Do We Do?

Simple:

•  Scan findings can help, but they rarely guide treatment
•  Be led by the patients’ symptoms and goals and devise a treatment plan to suit
•  Use imaging only when you suspect it may alter management (e.g., suspected full-thickness tear with significant functional loss)

Bottom Line

Imaging might help reveal the shoulder to the clinician in great detail. But it doesn't help the shoulder recover