Treatments for Tinnitus – What does success look like?

Guest Blog Post by Dr James Jackson, lecturer and chartered psychologist. James began his academic career by completing a PhD on cognitive performance decrement in tinnitus patients, and has continued researching tinnitus over the following decades. Here, he shares with us his thoughts on treatments for tinnitus.

"With so many drugs, fads, interventions and products out there, it’s hard – if not impossible – to accurately reflect on the efficacy of a particular treatment for tinnitus. Claims and counter claims about ‘cures tinnitus’ and providing ‘relief from tinnitus’ can often be found on boxes and packaging and webpages. First and foremost – there is no current cure for tinnitus – there is nothing out there, robustly tested, that works for everyone. There are so many different causes of tinnitus, and we are all individuals dealing with a chronic condition in our own separate ways, so how could it? But the claims remain.

Many products may have been tested, though few are tested reliably – and independently – and are published in peer-review academic journals. In most cases, it’s market research and feedback from those minded to feedback (which will only be a particular subset of users). Will these be the angry ones? The ones for whom the product didn’t work? Or the ones that gained the most, and are eager to tell their tale. (It’s probably the latter, but it’s impossible to say). There’s always a self-selection bias, so can the numbers be trusted?

The makers of a tinnitus ‘widget’ may turn around and say that 80% of their users report an improvement (or reduction) in tinnitus distress over time. Sounds good, but what’s happening here, and what is an ‘improvement’? Firstly, there is a world of difference between an improvement and a clinically significant improvement. Let’s consider a useful questionnaire for measuring tinnitus distress – the Tinnitus Functional Index, or TFI. If you complete the scale and have a score of ‘60’ before trying a new brand of herbal tea (for an example) and return after two drinks a day for eight weeks and complete the questionnaire again, you may score ‘56’ this time. That’s an improvement. Success!! Or is it? From day to day, tinnitus distress ebbs and flows, and there is a great deal of variability. Tinnitus spikes. Relaxing days off. etc. Is this ‘56’ an improvement, or were you caught on a good day? Or did you score that ‘60’ on a bad day. We have no idea. The academic literature currently agrees that a significant and clinical reduction in tinnitus distress requires us to move the TFI dial by fourteen points, at least. Going from ‘60’ to ‘56’ is an improvement in the technical sense, but it’s meaningless in reality. ’60’ to ‘50’ is meaningless too. As is ‘45’ to ‘33’. For an intervention to be successful, we need to get that ‘60’ down to ‘46’ and we need to keep it there. So if a product says that 80% of users report an improvement in their tinnitus distress, it’s a very high chance that most of these will not be referring to a clinically meaningful improvement.

For context and in my own research, I consider it good if 30% of participants report a clinically significant reduction over a period of several months. And of course, different interventions will work for different people. There is something out there for everyone. The trick is working out what this is and staying the course. To keep trying while other strategies and treatments fail to deliver.

The other thing we need to consider is participant adherence, or drop-out rate. How a study deals with this is critical to understanding the value of an intervention. Let’s say you start off with 400 people, but over time we see a typical drop-out rate of 25%. Where did these people go? Did they move away? Did they get bored of taking part? Did they fail to follow instructions (an honest mistake perhaps, or maybe too much was asked of them)? Did whatever they were doing not work? Or worse, did it antagonise their tinnitus and make it worse? In my research, unless a participant actively withdraws, I keep this data. So if someone completes the TFI and has a score of ‘75’ and for some reason, they never come back three months later, that is not a participant to be removed from the sample. It is a failed treatment. So I take that ‘75’ at baseline, keep that ‘75’ at the end of the study, and report no change in this participant. I’m guessing of course, but whatever I’ve asked them to do has put them off returning in some way, and we need to acknowledge this. While it’s a seemingly small change, it’s one with huge consequences. Let’s carry on with our example:

400 people, 100 drop-out and of the rest, let’s say that 40% of people see an improvement in their tinnitus distress. If we work it out, that’s 40% of 300 completes, which is 120 people. And let’s not forget, that’s 120 people with an improvement, only a small number of these people will have a clinically significant improvement. Let’s be generous and say that one third of these, 40 people – which is 13.3% of completes – gain clinically significant relief.

But, if we keep those who dropped out, and assume a failed intervention on their part, that’s 120 people out of the grand total of 400 participants who saw an improvement, 30% of our participants. If we again assume that one third of these saw a clinically significant improvement – and that’s a big ‘if’ by the way – then we can hope that 10% of the sample gain clinically significant relief.

What does this example tell us? For starters, even if we assume that some people gained far more than others, fiddling the numbers by looking at the drop-out rates in a different way, you can make a product look 33% more effective than it really is. And this is the problem. I’m not saying that tinnitus treatments are a con, but I am saying that some tinnitus treatments are considerably less effective than you could be led to believe. Watch out for statistical tricks, ask for clinically significant figures – if they are available - and assume that a drop-out is a participant who walked away for good reason rather than one who can be discounted. Look for sustained changes over time, not quick fixes that don't continue to deliver. These will be placebos, common not only in tinnitus research, but across the whole field of medicine and wellbeing. Caveat Emptor.

Dr James Jackson, C.Psychol. AFBPsS. SFHEA.