Why did you decide to undertake this study? Did it build on any previous work in particular?

This study directly built on one we published in 2021 investigating whether people’s management preferences were influenced by different diagnostic labels for rotator cuff disease.

Rotator cuff disease is an umbrella diagnosis that also goes by subacromial pain syndrome, rotator cuff-related shoulder pain and other names.

The previous study randomised 1308 people with and without shoulder pain to read one of six hypothetical scenarios describing a person with rotator cuff disease who visits their health professional.

The only difference between the scenarios was the diagnostic label used by the health professional to describe the person’s condition.

We found that using the label ‘rotator cuff tear’ increased people’s perceived need for shoulder surgery compared to using ‘bursitis’.

Shoulder surgery is commonly provided for rotator cuff disease, yet on average it has limited clinical benefit.

In our study, a rotator cuff tear was described as a tear in one of the shoulder tendons and bursitis was described as inflammation of a fluid-filled sac in the shoulder.

As part of the scenario, participants were encouraged to stay active and given positive prognostic information (guideline-based advice).

We realised that this advice might not reflect what patients with rotator cuff disease receive in the real world.

Therefore, for the follow-up study, we wanted to explore whether different advice also influences people’s perceived need for shoulder surgery and whether there is an interaction (a ‘multiplying’ effect) between labels and advice.

What diagnostic labels and advice did participants receive? How did you choose them?

All participants were randomised to receive one of four sets of labels and advice in their hypothetical scenario—a two-by-two factorial design. The scenario was similar to the previous study.

The labels were ‘rotator cuff tear’ and ‘bursitis’ because they previously influenced people’s perceived need for surgery the most (but in opposite directions).

They were described to participants in the same way as in the earlier study.

Dr Josh Zadro.

One message of advice stressed that treatment was needed for people to recover (treatment recommendation).

We chose this message because it was from a reputable source (a Harvard Medical School webpage) and presented contrasting advice to our other message (guideline-based advice).

How did you measure the perceived need for shoulder surgery and other secondary outcomes?

After reading the scenario, participants were asked, ‘Do you think you need surgery to relieve your symptoms?’

Response options ranged from 0 (‘definitely do not’) to 10 (‘definitely do’). Secondary outcomes were assessed using a similar approach.

They included the perceived need for imaging, an injection, a second opinion or seeing a specialist and the perceived seriousness of the condition, recovery expectations, impact on work performance and need to avoid work.

What was the effect of diagnostic labels and advice on people’s perceived need for shoulder surgery? What about the secondary outcomes?

As in our previous study, using the label ‘rotator cuff tear’ increased the perceived need for surgery compared to using ‘bursitis’. The effect size was slightly smaller in this study (0.5 vs 0.7 on a 0–10 scale).

Describing the condition as a ‘rotator cuff tear’ also decreased the perceived need for imaging and seeing a specialist and the perceived seriousness of the condition and need to avoid work.

Compared to a treatment recommendation, guideline-based advice—including encouragement to stay active and positive prognostic information—decreased the perceived need for surgery.

It also decreased the perceived need for imaging, an injection, a second opinion or seeing a specialist and the perceived seriousness of the condition and recovery expectations.

The effect of guideline-based advice was around two to three times stronger than the effect of labelling for most outcomes (eg, 1.0 vs 0.5 for perceived need for surgery).

There was no interaction effect for any outcome. This suggests that labels and advice likely have an additive effect on people’s management preferences.

What is your takeaway message? Were the effects clinically important?

Appropriate advice and labelling from health professionals appear to be important for people with rotator cuff disease and could reduce the overuse of shoulder surgery.

Although we do not have data on how to interpret the effect size for our outcomes, avoiding potentially harmful advice and labels is free of cost, risk and effort.

Therefore, even small effects could be worthwhile on a population level.

What does research in this field need to look at next?

Future research is needed to investigate whether diagnostic labels and advice influence clinical outcomes (eg, pain, function, satisfaction and self-efficacy) and real-world healthcare use. It may also be valuable to test labels and advice different from the ones we chose.

>> Dr Josh Zadro APAM is a physiotherapist and a National Health and Medical Research Council-funded Research Fellow investigating strategies to improve access to and uptake of high-value care for musculoskeletal conditions. He has published over 80 research articles in physiotherapy and medical journals.