Analysing UK Biobank Data Together

[Hazel]

Biobanks have been a hot topic of discussion recently. The BTA's recent political lobbying efforts called for the creation of a new tinnitus biobank. The just launched UNITI research consortium aims (amongst others) to aggregate data from existing biobanks.

All of the above is much needed, but it's also worth looking at the biobanks that already exist and may lead to new insights about tinnitus. These don't have to be tinnitus biobanks per se, as long as they include some relevant datapoints on tinnitus. Some countries have done large-scale population studies across many health-related indicators and aggregated such data in a biobank. A good example is the UK Biobank. It follows 500,000 UK citizens, who periodically provide their health information.

That's a wealth of data right there. The sample size is awesome, this is truly "big data." Plus, there's even longitudinal data, i.e. individuals are assessed at different points in time, so we can see how their situation changes or develops. The different types of data collected is also impressive: imaging data, genetic data, demographic data, physical measures, etc.

The UK Biobank includes only two datapoints related to tinnitus:

• "Do you get or have you had noises (such as ringing or buzzing) in your head or in one or both ears that lasts for more than five minutes at a time?"
• "How much do these noises worry, annoy or upset you when they are at their worst?"

So, these are obviously not very detailed indicators when it comes to tinnitus. But… the large sample size makes it interesting, as well as the fact that the biobank includes data on many other health indicators. So maybe we can try to unveil some trends or correlations that could lead to new avenues for curative research?

Jorge Simoes and Patrick Neff from Regensburg University – whom we've worked with before on this study – have proposed that we collaborate on a new project using the UK Biobank data, and they want to involve the Tinnitus Talk community. What we propose is the following:

1. You guys help us explore the health data points in the UK Biobank to identify a possibly interesting area of study. For instance, we could look at correlations between certain lifestyle factors and tinnitus, or tinnitus and genetic markers. You can explore the dataset here and come up with your own hypotheses for us to test!
2. We will then request the proposed data from the UK Biobank and analyse it. If you want to help out with this, let us know.
3. Finally, if our study does in fact yield some interesting findings, we could follow-up by gathering some data of our own, i.e. through a Tinnitus Talk survey. In the survey, we could ask more in-depth questions to corroborate our theories.

Alright then, we now need you to pitch in on step #1. Let's hear it!!!

 

[UKBloke]

@Hazel thank you for opening up another potentially useful avenue to this community.

The Biobank data thing I think is very interesting. There's one point I want to really try and understand at this junction, however. I've done a basic Google search of both Jorge Simoes and Patrick Neff, and it appears that the department they both operate out of at Regensburg is Psychiatry and Psychotherapy. Taking a quick flick over Jorge's blog posts I also see mention of phrases like "coping strategies" and "groups sessions".

The notion of embarking on an extensive project like a Biobank partnering with Psychiatry and Psychotherapy endpoints gives me a sinking feeling. We sufferers have been here too many times already. My view is that tinnitus is real - it's a physical thing that requires hard science to determine its causes and ultimately its treatment. I believe that tackling tinnitus will come from practical scientific intervention, not learning to cope or chatting about it in group therapy sessions.

At this point I don't have enough information about Jorge's and Patrick's intentions regarding the above, which is why I wanted to put this post out there for yours and hopefully others' comments.

 

[Hazel]

@Hazel thank you for opening up another potentially useful avenue to this community.

The Biobank data thing I think is very interesting.

Glad you like the project idea at least (in spite of any other misgivings), and would love to get your input on that! Let me first respond to your concerns though:

I've done a basic Google search of both Jorge Simoes and Patrick Neff, and it appears that the department they both operate out of at Regensburg is Psychiatry and Psychotherapy. Taking a quick flick over Jorge's blog posts I also see mention of phrases like "coping strategies" and "groups sessions".

We at Tinnitus Hub (and I hope you know this) are 100% about cure focussed research.

That's not to say that I can vouch for each and every research paper ever published by these guys, but I can certainly vouch for their good intentions. They want to work with this community to determine their research priorities, and I think that kind of openmindedness should be welcomed; not many researchers have the guts to do this. This is a genuine opportunity for us to help determine the future course of their research.

Honest question: Are you proposing that we not work with them? This is not a trick question, I really want to know!

I also think your perception of their work might be slightly off. For instance, I personally do not consider Jorge's work to fall under the banner of psychology or psychotherapy. Maybe you can read his recent article (based on data that we collected) and judge for yourself. The same goes for Jorge's blog post (I would encourage you to read it in detail, because I think you may have misunderstood its purpose?), which is not at all about coping strategies. As the title says, it's about subtyping, a concept that has been identified by many tinnitus researchers as key to finding a cure (or rather cures).

The phrases from the blog that you refer to are totally taken out of context; though I trust that was completely unintentional . Jorge analysed our data, from over 5,000 Tinnitus Talk members, which included what treatments people had tried — people chose from a comprehensive list of 25 treatments, which included psychological treatments, but also TMS, neuromodulation, and various medications. So it was not about promoting any one treatment (most of them barely work anyway), but using these datapoints, in combination with many other datapoints, to define specific groups of tinnitus patients.

My view is that tinnitus is real - it's a physical thing that requires hard science to determine its causes and ultimately its treatment. I believe that tackling tinnitus will come from practical scientific intervention, not chatting about it in groups.

100% agree. But we are not proposing any chat groups or wishy washy stuff here, we are proposing serious hardcore data analysis that could lead to valuable new insights about the nature of tinnitus, and thus potentially provide new avenues for curative research.

 

[UKBloke]

Honest question: Are you proposing that we not work with them? This is not a trick question, I really want to know!

No! Obviously the group as a whole will decide on that. On a personal note, I need to understand what I can, can't, wish to, and wish not to be involved with at this stage, and that very much depends on the profile of those handling the data and the kind of research they're involved in. Dr. Will Sedley I think is a good example of someone I'm really looking forward to Tinnitus Talk working with.

The same goes for Jorge's blog post (I would encourage you to read it in detail, because I think you may have misunderstood its purpose?), which is not at all about coping strategies, in any way, shape or form. As the title says, it's about subtyping, a concept that has been identified by many tinnitus researchers as key to finding a cure (or rather cures).

My main concern with the blog article is under the section titled, What the future awaits. Sure, subtyping, I get it. But when I see statement like:

Personality, coping strategies, stress levels, quality of sleep, life history, and mood represent some personal characteristics that can be further explored in the context of tinnitus subtyping (my emphasis)

My anti-Freudian radar goes on high alert. "Personality"??? Really??? This opens up an avenue that ends on a psychiatrist's couch with the notion that, Sorry, it's your personality. In other words, the reason you have tinnitus is your fault.

I'm not suggesting this is where it's headed. I'm trying to understand. And I know your intentions are good - I see that with all the time and effort you put into this site every time I log on.

 

[Hazel]

Personality, coping strategies, stress levels, quality of sleep, life history, and mood represent some personal characteristics that can be further explored in the context of tinnitus subtyping (my emphasis)

my anti-Freudian radar goes on high alert. "Personality"??? Really??? This opens up an avenue that ends on a psychiatrist's couch with the notion that, Sorry, it's your personality. In other words, the reason you have tinnitus is your fault

I will ask Jorge and Patrick to clarify further for everyone here (including myself) to better understand their vision. I don't want to make any assumptions on their behalf, and I agree it's important that we all understand their intentions before moving forward.

I'm also not a big fan of personality focussed research, but I don't per se agree that it's automatically bad news for us. If for instance it turns out that certain personality traits matter in tinnitus severity, that might provide leads for further neurological research or imaging studies perhaps? [And in fact, the UK Biobank does include imaging data.] Or it could imply that when new treatments come to market, they should be tested on those who are most prone to severe tinnitus, based on certain personal characteristics; because if it doesn't work for them then what's the point?

So it's not necessarily about having the "wrong" personality. I do however understand where this sensitivity comes from, as tinnitus sufferers have been told too often and for too long that their attitude, and not the tinnitus, is the problem. And I understand that there is a risk, no matter what the intentions, that certain research outcomes can be misappopriated by people with a vested interest in pushing psychological therapies. That's obviously something we want to avoid.

Having said all that, I don't want this research project to focus on personality traits. There is for instance a lot of biomarker data in the UK Biobank and I'd love to explore whether we can do anything new in correlating those with tinnitus. I dunno, but I'd really only want to proceed with this if we can come up with a genuinely innovative idea on how to use this data. I don't want to spend hundreds of hours analysing data just because we can.

But... that's just my viewpoint. I obviously cannot speak for Jorge and Patrick, and I think it's only fair to give them right of reply here, and for you (and others with similar concerns) to learn more before deciding whether you want to collaborate on this project. And of course it's very much your right if you don't want to take part. I appreciate you bringing this to light in a constructive way rather than assuming bad intent

 

[UKBloke]

I will ask Jorge and Patrick to clarify further for everyone here (including myself) to better understand their vision. I don't want to make any assumptions on their behalf, and I agree it's important that we all understand their intentions before moving forward.

I'm also not a big fan of personality focussed research, but I don't per se agree that it's automatically bad news for us. If for instance it turns out that certain personality traits matter in tinnitus severity, that might provide leads for further neurological research or imaging studies perhaps? [And in fact, the UK Biobank does include imaging data.] Or it could imply that when new treatments come to market, they should be tested on those who are most prone to severe tinnitus, based on certain personal characteristics; because if it doesn't work for them then what's the point?

So it's not necessarily about having the "wrong" personality. I do however understand where this sensitivity comes from, as tinnitus sufferers have been told too often and for too long that their attitude, and not the tinnitus, is the problem. And I understand that there is a risk, no matter what the intentions, that certain research outcomes can be misappopriated by people with a vested interest in pushing psychological therapies. That's obviously something we want to avoid.

Having said all that, I don't want this research project to focus on personality traits. There is for instance a lot of biomarker data in the UK Biobank and I'd love to explore whether we can do anything new in correlating those with tinnitus. I dunno, but I'd really only want to proceed with this if we can come up with a genuinely innovative idea on how to use this data. I don't want to spend hundreds of hours analysing data just because we can.

But... that's just my viewpoint. I obviously cannot speak for Jorge and Patrick, and I think it's only fair to give them right of reply here, and for you (and others with similar concerns) to learn more before deciding whether you want to collaborate on this project. And of course it's very much your right if you don't want to take part. I appreciate you bringing this to light in a constructive way rather than assuming bad intent

I agree, I think we've just spent too long where the leaning has been towards psychotherapists telling us to accept our symptoms rather than bio-mechanics coming up with a cure. We need to move on - there's so much fantastic technology out there and we very well may be approaching a paradigm shift. If we can feed data into that then I'm all for it. If you are able to have some further clarification from Jorge and Patrick that would be incredibly useful. I also agree wholeheartedly with the notion of not spending hundreds of hours analysing data just because we can.

It'd be great if others can start feeding into this discussion: @brokensoul I've really enjoyed reading some of your posts recently. Any thoughts on this thread?

 

[Badger19]

I LOVE this idea. But, is there a possibility to include hyperacusis?

 

[Markku]

But, is there a possibility to include hyperacusis?

Nope because the UK Biobank doesn't have data on hyperacusis.

 

[Jorge P. Simoes]

Thank you for the chance to present ourselves and to explain some of our work and some of our ideas.

The notion of embarking on an extensive project like a Biobank partnering with Psychiatry and Psychotherapy endpoints gives me a sinking feeling.

I don't know if that will bring any consolation, but Patrick and I are members of the E-Health group of the department of Psychology and Psychiatry: we are not clinical psychologists (or represent/speak on their behalf), and we are not advocates of a particular treatment. Instead, our group is interested in using Big Data to generate knowledge that would be otherwise impossible to obtain (hence our interest in the BioBank). Our work focuses on apps such as the Track Your Tinnitus and developing databases for multicenter studies.

On top of this line of research, Patrick also has published several articles in acoustic stimulation, sound therapy and neurofeedback, he is also a promoter of citizen science workshop at in two TRI meetings (including this year's upcoming event in Vancouver); I am a co-author in an upcoming paper about the effects of a drug for tinnitus patients (sorry I can't go into too much details as the paper is still in preparation). In other words, we are not part of a cabal of psychologists and/or psychiatrists. We have seen with our own eyes that some people respond to some treatments magnificently, to the point that some of them start to cry from joy. Unfortunately, those responses are not universal, and we (still) don't know why. The BioBank is indeed an extensive project, and one that we would take willingly. And more importantly, we would like to open channels of communication with the Tinnitus Hub community, so you are also part of this project.

Hazel's question was not mere rhetoric: the question wasn't opened to the community just to be ignored and for us to do the research our own way: we never proposed investigating personality or psychological phenomena. The whole reason why Patrick and I are continuously reaching out to Tinnitus Hub (we have regular meetings with Markku, Hazel, and Steve) is to bridge the gap between scientists and the community. Our goal is to get away from our "academic ivory tower" and answer questions that are meaningful to those that matter the most: tinnitus sufferers. Although we can't change the entire field to do the same, we can at least start with our own attitude by doing more participative research.

My view is that tinnitus is real - it's a physical thing that requires hard science to determine its causes and ultimately its treatment. I believe that tackling tinnitus will come from practical scientific intervention, not learning to cope or chatting about it in group therapy sessions.

We both agree 100% with that statement. Like previously said, we are not raising the banner of treatment A or B. We want to use state-of-the-art methods and data to answer questions pertinent to tinnitus patients. On top of that, we would like to make this a collaborative effort between researchers and the community. This is an extremely rich dataset, and many different analyses can be done. Why not let tinnitus sufferers decide what they want to be investigated?

Personality, coping strategies, stress levels, quality of sleep, life history, and mood represent some personal characteristics that can be further explored in the context of tinnitus subtyping (my emphasis)

My anti-Freudian radar goes on high alert. "Personality"??? Really??? This opens up an avenue that ends on a psychiatrist's couch with the notion that, Sorry, it's your personality. In other words, the reason you have tinnitus is your fault.

The highlighted statement was inspired by a paper recently published in Lancet. The authors investigated insomnia, a condition that overlaps with tinnitus, and also seems to be heterogeneous in both etiology and manifestation. The authors investigated all the aforementioned traits and fed all the information to a cluster algorithm. In turn, the algorithm identified five subgroups of patients. Critically, the differences across subgroups were observed not on the behavioral level (that is, insomnia patterns), but also on the neuronal level (figure 4C): these findings suggest that different groups of insomnia patients have different neuronal substrates that weren't identified until then. Since its publication, a tsunami of papers have been published further exploring the underpinnings of those neuronal differences and its potential clinical implications.

Personally, I (Jorge) would like to try something similar in the tinnitus field. In a way, this is like looking at a problem from a different angle (let's say "from outside in", instead of "inside out"). Given the slow progress in tinnitus research, I think that fresh ideas, especially those that are successful in other fields, should, at least, be tried. I don't know what role personality plays in the "tinnitus jigsaw puzzle", maybe it does not play any role whatsoever, but a critical aspect of research is to ask questions that don't have an answer yet. In the aforementioned study, before running the first analysis, the authors ran a review with all the traits (30ish, if my memory is correct) they suspected would be related to insomnia based on previous literature. Unfortunately, we are far away even from that preliminary goal. We aren't even sure about what heterogeneity consists of! Hazel's comment on this type of analysis, with its pros and cons, is spot on.

Patrick and I can honestly say that we reach out to the Tinnitus Hub without ulterior motives. After realizing the colossal gap between academics, clinicians and tinnitus patients, the next question we asked was how to address this issue. The BioBank project was only a suggestion, and we never proposed investigating personality or other psychological phenomena. I certainly have some ideas of what could be done with this dataset (noise maps, biomarkers, neuroimaging, diet, just to name a few), but the idea was always to hear from you what is most relevant to you.

We hope this explains our intentions. Publishing stuff just for the sake of publishing sounds like a lot of wasted time, money, and opportunity, and we are the first ones to say that we don't want any of that.

 

[UKBloke]

Jorge,
Thank you for taking the time to write this post. There's certainly enough here to help me understand a little bit more about the project.

we are not part of a cabal of psychologists and/or psychiatrists.

This is good to hear. Hopefully you understood my initial concerns, particularly in the context of the Lancet inspired article.

Why not let tinnitus sufferers decide what they want to be investigated?

I completely agree with this and think it's the next logical step for researchers to take in the Information Age. As you're no doubt aware this community has already cast some votes in another thread for research we'd like to see go ahead. Dr Will Sedley in the UK has also reached out to Tinnitus Talk and looks set to begin what seems like quite an exciting phase of his own research. The time is definitely ripe to build these kind of relationships.

I was going to raise the following point in another thread but just to close off this post I feel it's probably more pertinent to raise it here.

I feel like we're in a kind of closed-loop with tinnitus at the moment and I want to quickly try and explain why. As a UK citizen, the main go-to tinnitus charity here is the BTA. The BTA is lobbying parliament, and I think this represents a significant move forward for our cause, especially as the BTA states that it wants to see the community move to finding a cure. I think that we would all agree this is a noble position to take, however, if one peruses the treatments section of the BTA, which is a section that uses a kind of traffic-light system to identify which treatments are available, the only with two green lights, and as such the one the BTA in effect recommends is, CBT.

The problem with CBT is that it's not a cure. And yet in the absence of a cure I can see that it makes absolute sense (although I'm sure it'll come as no surprise that I'm not an advocate of psychological therapy for tinnitus) to engage in coping mechanisms. The glaring problem with this, however, is that people new to tinnitus, and I daresay those in the political arena whom we as a community need to influence (and perhaps even some medics), could be excused for thinking, why bother looking for a cure for tinnitus if we can treat it? And thus we end up in the often confusing and contradictory cycle that we seem to be in.

Apologies for rambling on about this but I wanted to try and ensure that you understood my concerns regarding the psychotherapy side to tinnitus treatment aren't just born out of an irrational fear of this discipline. I actually believe at this point in time that the contradictory nature of this situation represents a hurdle towards finding a cure. If I can see a project that sets itself, in the main, as something separate from that, I'll be 100 percent behind it.

 

[brokensoul]

Hi @Jorge P. Simoes,

It is really great that researchers are reaching out to the biggest tinnitus patient community. Thank you so much for doing so. It is truly the right mindset to understand the need to engage with patients who actually have the condition. This is really positive. I very much appreciate this mindset.

When I look at this specific idea, I'm highly pleased on a conceptual level, but less on a practical level in regards to delivering valuable qualitative tinnitus research. Conceptually this is really positive and I believe that analysing big data sets is essential in moving tinnitus research forward. I'm a strong proponent of building a global tinnitus patient (big data) database with a huge datamodel containing all possible tinnitus datapoints (and longitudinally to see evolution).

However I can't help thinking this research (analysis) proposal in this case highly depends on the quality of the dataset in the UK biobank, by which I don't mean that the biobank has low quality info in general, but there is a significant concern here in my opinion in regards to tinnitus. I'm sure I'm not the only one thinking that analysing this dataset imposes a significant problem or challenge.

The information in the biobank directly indicating who is a tinnitus patient or not, is not very convincing and this will surely lead to skewed results. How would you deal with the fact that the primary datapoint only clarifies whether someone has had "tinnitus" for longer than 5 minutes (in the last year)? From my point of view the problem starts there. The second question is probably not even that relevant anymore. Fleeting tinnitus is not really tinnitus and everyone experiences this, but I must admit it usually does not last 5 minutes, more like 30 seconds or so. Still, I feel the problem really starts with this datapoint and I suspect too many assumptions will be made.

How do you interpret the first datapoint? Do you consider a confirmation on that question to be someone with tinnitus or not?

In essence you would probably have to limit the dataset to be analysed to people who answered 'yes' on the first question and indicated they were effectively impacted by it on the second question to have a better idea of who could possibly really be a tinnitus patient in the biobank. The data is longitudinal and that is of great value of course, so maybe you would need to limit the data to people who at least answered 'yes' on the first question for several years in a row. Even so this could still give a wrong impression.

Additionally people are unique and you'll end up wondering what exactly the correlation could be between these two questions. Did someone who answered 'yes' on the first question and stated that they were significantly impacted by it, have loud chronic tinnitus or is it a person who reacts very upset to a short lasting tinnitus episode and so on. You will never know and you'll end up speculating. The highly limited and not perfectly clear two datapoints highly complicates your ability to draw any meaningful conclusion from this data. Any meaningful correlation with other data in the biobank will be difficult I suspect.

Therefore I would rather see an effort in expanding the biobank with a detailed tinnitus questionnaire that will provide an excellent base for tinnitus researchers in the future. This would surely have much more impact in the future instead of analysing this dataset now.

One clear question to know if people have chronic tinnitus (e.g. lasting for two years minimum) would already make a huge difference. Any research done on this dataset would be of much higher quality as you would only look at actual tinnitus patients and not a dataset with people who may have had or have regular fleeting tinnitus episodes.

I'm quite convinced for example that by looking at the current dataset in the UK biobank, you would come to the conclusion that tinnitus only affects a very very small minority of the full tinnitus population. In reality chronic tinnitus affects everyone who has it. Loud or intrusive tinnitus deeply affects the strongest of us all. Those who have mild tinnitus will still be affected in a minor way, but will probably be able to deal with it without significant impact on their lives.

If you would take anyone in this dataset who answered 'yes' on question one, you will think that an incredibly high percentage has absolutely no impact whatsoever. Of course, because they don't have tinnitus! In my opinion this leads to wrong conclusions and wrong perceptions of the life changing condition that tinnitus really is.

I also suspect that this type of data is what is behind the numbers that state that 15% of the population has tinnitus. I don't believe this is true, but I could be wrong. It's quite possible I suspect that those numbers are conflated with people who have experienced a brief tinnitus episode and those who really have chronic tinnitus. The latter is probably a much lower percentage of the total population. Unfortunately I believe such conflation leads to wrong perceptions of what impact tinnitus really has.

So I don't want to discourage you in any way. Just a critical thought and I'm interested to hear how you would or will deal with this?

My opinion is mainly that it would be better to lobby with the UK biobank to expand their dataset with extensive datapoints on tinnitus. People who have tinnitus would gladly provide in-depth data on their situation knowing how important it is to drive research forward and others obviously would not have to provide any of those datapoints. This could really push tinnitus research forward.

Best regards,

 

[Jorge P. Simoes]

Hi @brokensoul,
Thank you very much for the kind words, and thank you very much for all the thought put into this.

Here's some food for thought:

The information in the biobank directly indicating who is a tinnitus patient or not, is not very convincing and this will surely lead to skewed results. How would you deal with the fact that the primary datapoint only clarifies whether someone has had "tinnitus" for longer than 5 minutes (in the last year)? From my point of view the problem starts there. The second question is probably not even that relevant anymore. Fleeting tinnitus is not really tinnitus and everyone experiences this, but I must admit it usually does not last 5 minutes, more like 30 seconds or so. Still, I feel the problem really starts with this datapoint and I suspect too many assumptions will be made.

I agree with you with regards to question #1 (especially if the goal is to use it as way to cluster people with/without tinnitus), but question #2 can be quite informative. Indeed, the distress of someone suffering for years is (I would think) different from someone who developed tinnitus for a couple of weeks. However, the large sample should be enough to account for that difference, and to allow us to model factors associated with distress.

How do you interpret the first datapoint? Do you consider a confirmation on that question to be someone with tinnitus or not?

I believe that the question refers to tinnitus, but unfortunately it is an unspecific question. It is not possible to disentangle acute and chronic tinnitus from it, which is, as you rightly put, a major shortcoming of this dataset. Much more interesting is to have a dataset with a clear inclusion criteria for tinnitus duration.

Therefore I would rather see an effort in expanding the biobank with a detailed tinnitus questionnaire that will provide an excellent base for tinnitus researchers in the future. This would surely have much more impact in the future instead of analysing this dataset now.

I don't know how lobbing for a new questionnaire in the UK Biobank works (or how long it would take to do so successfully), but I would imagine that research using the database would favor such endeavour. Even though the formulation of the two questions are not ideal, one can deliver a powerful message depending on the research question.

Instead of developing a research question around biomarkers (because probably question 1 encompasses too much heterogeneity to define a precise distinction of people who actually have tinnitus), I would propose a research question around the implications of tinnitus distress.
This paper using the UK Biobank is a great example of what I would propose. The authors were able to link pulmonary obstruction disease with environmental factors such as greenness of the area where people live. I know that the UK has data on sound pollution, so one could do a similar analysis on the relation between distress and noise.

And if we really want to have results specific to the chronic tinnitus population, we could have a survey with members of the tinnitus talk: an exploratory analysis at the societal level, and confirmatory at the community level (same/ questions but with a more specific inclusion criteria).

Such a paper would have enough merits on its own, but with 1) participative development, 2) a project with policy implications, 3) big data, and 4) an exploratory/confirmatory I believe this type of research could also be used as a way to lobby for improvements in the current biobank.

I hope this this example illustrates my point: although the Biobank is not perfect, some really meaningful research can be done with it. Of course this is just one way of (partially) circumventing some of the limitations you mentioned, and I am sure there are other ways to do so. Personally, I wouldn't wait until we have a dataset that is just right. Every study has its own limitations, the question is always whether the insights/implications of the study outweighs its limitations.

Looking forward to hear (read?) your input.

Best wishes,
Jorge

 

[Frédéric]

Paradoxically, to understand what UK biobank is, I had to go to Wikipedia:
https://en.wikipedia.org/wiki/UK_Biobank
And it seems that this constitutes a very large study, initiated by the British National Health Service, a top down process in a way, with a hard core of tests done in the same way on everyone, with very strict protocols, to get new data.

So I browsed this database in order to understand how it works. I began with the item 4814 Tinnitus annoyance: we can see that 2187 persons are severely annoyed by their tinnitus.

Questions, in this biobank:
- is it possible to get access to each individual medical folder of these persons? Like their tonal audiogram for example...? I was not able to do that, maybe I missed something.
- I saw that there is an item related in a tab: hearing test. But it seems that it gives the average hearing test of all persons who were asked to do it, including healthy persons. So there is no way to extract tonal audiograms from tinnitus sufferers and to compare to those of healthy persons. And I guess it is the same problem if you want to compare genome, MRI imaging (size of thalamus for example...) and so on.

So to sum up, I don't think one can (unless maybe if there is a restricted access for researchers):
- get brut data of each person,
- make extracts and operations on data for comparison according to a criteria (tinnitus annoyance for example).

More over, unfortunately for us, there are only 2 items about tinnitus (presence and annoyance) and 2 items about hearing loss (hearing tests).

I would be tempted to say that our problem is the opposite, we have to engage a bottom up process. A priori, for those who went to tinnitus clinics, we did batteries of test or treatments, but each time the nature of the tests was according to the appreciation of each practitioner and not according to a universal flow chart, with different protocols, and generating (for the technical part) different types of files (imagery, examination reports ...). The goal is to recover all these existing data from these clinics and try to order them.

This is where I will see the role of UNITI, to put pressure on all these clinics, that they agree to share their data (because I have the impression that this is far from being the case). For example concerning France, I do not know if the various members of AFREPA like IMERTA share their data, obviously they do not share it with TIN ACT which is however their partner. I will be able to inquire and if necessary put pressure on them, but this is where I think that UNITI would have more weight.*

* EDIT: I went to my appointment with TIN ACT yesterday. The PhD student told me that she obtained the list of the patients of IMERTA without any problem. The problem is that when she called patients to warn them about her clinical trial, almost everyone told her that they were not interested taking part in it though they have severe tinnitus and live close to the clinical trial site. Little bit of out topic, that corroborates the apathy of the tinnitus sufferers for the BTA petition.

 

[Jorge P. Simoes]

Questions, in this biobank:
- is it possible to get access to each individual medical folder of these persons? Like their tonal audiogram for example...? I was not able to do that, maybe I missed something.

From what I understand, it is possible to request the individual medical folder of each person in the UK biobank. One has to make a formal request, select the desired items and pay an administration fee. The information in the website is a description of what they have available (and for how many people).

- I saw that there is an item related in a tab: hearing test. But it seems that it gives the average hearing test of all persons who were asked to do it, including healthy persons. So there is no way to extract tonal audiograms from tinnitus sufferers and to compare to those of healthy persons. And I guess it is the same problem if you want to compare genome, MRI imaging (size of thalamus for example...) and so on.

Indeed, it seems that they don't have an audiogram, but only two 5-item questions on whether the person has hearing problems or not.

 

[mikki_]

To me this seems like useful data. Tinnitus prevalence at ~10% (tinnitus most of the time/all of the time + a lot of the time).

I'd like to feed the tinnitus responses into a huge regression model and see what predictors surface on the other side. Most likely hearing loss but what is there beyond that. To my understanding not all people with hearing loss have tinnitus.

I tried to sign up to download the data, but it resulted in Internal Server Error. Has anyone else been able to download the data?

https://bbams.ndph.ox.ac.uk/ams/signup

 

[Frédéric]

I found this publication today using the UK Biobank about tinnitus:

Understanding Factors That Cause Tinnitus: A Mendelian Randomization Study in the UK Biobank

 

[Frédéric]

Genetic predisposition to tinnitus in the UK Biobank population

We identified 17 suggestive single nucleotide polymorphisms (p < 1e−5) spanning 13 genes in two sex-separated cohorts reporting chronic, bothersome tinnitus (control males n = 7,315, tinnitus males n = 226, control females n = 11,732, tinnitus females n = 300).

 

[Harder Calm]

I'm new to the forum (even though lurking around a bit longer), so I discovered this thread only after it popped up again recently.

Are there other interesting open questions with respect to this, is it maybe possible to share some code etc., are there possibly other interesting datasets that would be explored?

 

[Nick47]

Some new findings from the UK biobank. Very large sample size. Recently published findings.

Causal Associations of Genetically Determined Tinnitus With Neuroimaging Traits: Evidence From a Mendelian Randomization Study

 

[Nick47]

Not from the UK Biobank but biomarkers that are encoded by genes have been found. A very large sample size of over half a million people.

Highlights:

From the new research, multiple genes were found to relate to synaptic areas—the connections where electrical signals move from one nerve cell to another.

A gene called GRK6 that controls "GABA receptors" emerged as a potential key player. GABA (gamma-aminobutyric acid) is a kind of brake pedal for cell signaling; when GABA is present, it's like applying the brakes on a car, inhibiting the speed and preventing neurons from firing too rapidly or excessively. When GABA levels are reduced, the "release of inhibition" contributes to increased spontaneous activity, bursting, and enhanced sound-evoked response. The GRK6 gene is associated with reduced GABA levels and reduced inhibitory neurotransmission in the auditory cortex and thus may play a role in tinnitus.

The study also suggests dopamine receptors, which are also controlled by the GRK6 gene, may provide new therapeutic avenues.

→ Genetic architecture distinguishes tinnitus from hearing loss

 

[Nick47]

Ben Shapiro Award:

UCL Ear Institute study awarded UK prize for tinnitus research | Brain Sciences - UCL – University College London