UK Genome Wide Association Study (GWAS) project - draft website goes live, feedback sought on recruitment plan, and updates

The De Paul questionnaire assesses frequency and severity of symptoms.

i would recommend staying well away from Chalder fatigue score.

 

Ok, so I have had a think about it and I would like to propose selection of additional optional questionnaires, if possible, from the NIH Common Data Elements.

There would be no need to design or validate new questionnaires. There are quite a few validated instruments included in the Common Data Elements such as:

SF-36
EQ5-D
HRQoL-14
Neuro-Quol (may not be validated yet)
WHO Disability Assessment Schedule
DePaul Symptom Questionnaire
Composite Autonomic Symptom Scale
Moldofsky Sleep Assessment Questionnaire and several other sleep-related scales to choose from
Checklist for Individual Strength
Several fatigue instruments to choose from including the Fatigue Severity Scale, FACIT-Fatigue, Neuro-Quol Fatigue, PROMIS, MFIS or MFI.
Wood Mental Fatigue Inventory
Modifiable Activity Questionnaire (MAQ)

A few pain scales - I'm not sure which is best
Lots of cognitive and executive function instruments but I don't think these could be administered online/at a distance.

The Common Data Elements website also has several predesigned questionnaires that are aligned with the Common Data Elements e.g.:
Demographics
Employment and education history
Detailed personal and family health history Environmental exposure history
Symptom checklist
Neuroendocrine/hypothalamic symptom questionnaire
Orthostatic symptom grading scale
Post-exertional malaise assessment questionnaire
Medications and supplements taken
Activity and Symptom diary
Impact of symptoms on activities of daily living questionnaire
Bell CFIDS Disability Scale

While there are surely imperfections in some of these instruments and questionnaires, and self-report data has its limits, I think collecting this kind of data on such a large cohort in line with the common data elements that is linked to participants' genomic data, if possible and affordable, could be really valuable.

I am not suggesting all of the above questionnaires be used, but perhaps a well-chosen selection as optional extra data that participants can elect to provide if they wish and are able?

EDTA: link to Common Data Elements website where these instruments/questionnaires can be viewed:

https://www.commondataelements.ninds.nih.gov/Myalgic Encephalomyelitis/Chronic Fatigue Syndrome

 

I think it would be a case of looking for any statistically significant associations between specific mutations and severity across the whole cohort, rather than individual cases. Of course there are many factors that affect severity - especially environmental factors which wouldn't show up in a study looking at mutations (unless a mutation was found to be associated that clearly showed damage to a pathway that would be affect by a known environmental factor).

But people in the same family who might have almost the same mutations (e.g. a sibling) might not have ME at all, so it's not about looking for a deterministic association but rather any statistical significance that might suggest that certain mutation clusters predispose people to more or less severe ME.

Or, using different questionnaires, to different disease phenotypes in terms of symptom clusters.

The development of ME is very much likely to be an interaction between several SNPs and several environmental factors, which may change between people and over time.

So the questionnaires would be for looking for associations between e.g. genes, symptoms, severity and environmental factors (e.g. through the health history and environmental exposures questionnaires), not to seek deterministic predictors of cause and effect, but broad associations and predisposing factors on a larger scale than we have been able to do so far due to limited sample sizes.

This is unlikely to provide solid answers, and is limited due to its 'snapshot' rather than longitudinal design (but if participants were recontactable it could become longitudinal) but some patterns and associations might pop out of analysis of such a large cohort, which could guide further studies as to where to look in more detail. I.e. it could generate new hypotheses, or provide additional information for existing hypotheses (e.g. are specific mutations associated with risk of conversion from mononucleosis to ME? This could be looked at by recording onset types and triggers through questionnaires, for example).

Or, are specific mutations associated with ME with orthostatic symptoms Vs ME without orthostatic symptoms?

There are lots of questions which recording all of this data could help with.

 

A long list of questionnaires seems like a guaranteed way to reduce participation.

 

The point is the extra questionnaires would be optional. Would people be put off by optional questionnaires? If they don't want to do the optional ones, surely they just don't do them?

 

I see your point. I think, because as you say we don't know what will be included in the compulsory questionnaires, it would be best for the PIs to select whichever they feel are most appropriate, in the context of what they have already planned.

I agree that unnecessary extras would be a waste of energy and resources. I just think this is such a great opportunity (if funded) to gather any data on such a large cohort that could be helpful in advancing our knowledge of the disease, since we have never had a cohort this large.

I wouldn't want to see that opportunity missed, considering how hard it is to obtain funding for large studies.

But I completely agree that questions/questionnaires used should be selected judiciously. My long list of suggestions was intended as a pool to choose from to save time in developing/validating new questions (and to align with CDEs to aid comparability with other studies), rather than 'let's use all of these'!

 

I worry about the use of most of the currently available questionnaires. Many of us have had to fill in questionnaires that were not specifically designed for ME patients and then had results misinterpreted and used against us.

In the future this information will presumably be shared anonymously. Which may mean that researchers who like to torture the data till it give the answer preferred by them will have access. In my view many of the questionnaires would facilitate such endeavours. I won't be a part of that.

 

I would be extremely wary about mentioning anything about more questionnaires beyond the absolute minimum necessary to make a diagnosis for this study. The key thing is recruitment. As soon as potential recruits think they may have to wade through 'paper' either straight away or later they are likely to close the link. I suspect a large number of potential recruits will ask themselves 'do I really want to get involved' and anything that puts them off is likely to be a disaster.

In general my view is that if anything is not absolutely essential to a research question it should be cut out.

 

I agree.

At the top of anyone's mind when considering possible recruitment will be "Do these people understand what it is like to have ME/CFS?" If they perceive the answer as likely to be 'no, then they will rapidly lose confidence in such a study, and not partake. A key indicator of whether recruiters understand what it is like to have ME/CFS, will be how onerous - or not - the recruitment process itself is.

I think it will be essential for recruiters to give clear signals they really do understand what it is like to suffer from ME/CFS. Not by saying it, but by example.

 

But I have said Multiple times that any additional questionnaires beyond those essential for the GWAS would be optional, and my idea would be that people, if they choose to complete them, would be given e.g. a couple of years to complete them.

Does that not demonstrate an understanding of what it's like to have ME? I.e. an acknowledgment that firstly, people might not be able to complete additional questionnaires and therefore they should be optional, and secondly, that if they choose to complete them, they will need to pace themselves and may need a very extended period of time in which to do so.

My suggestion is not loading on additional compulsory questionnaires that people would need to complete in order to participate in the study.

My proposal is that people register for the study using the method already suggested by the PIs (minimal questionnaires) and send off their saliva sample.

Then once they have done this they could be given the OPTION to complete additional questionnaires, at their own pace, if they wish to and feel able.

Would this really put people off? I am not convinced by the counter-arguments given so far.

 

But then the population completing the questionnaires is self-selecting and not representative of the cohort participating in the GWAS.

 

Yep, I think very definitely from my own little focus group investigation

I think this is the general idea but because of the specific question being asked (are there linked genes) some slight adjustments are likely to be sensible. For instance it would be important not to miss people with a past history of ME who are now doing OK and not 'significantly disabled'.

 

I would much rather the focus of the research remained tight - gather the genetic data and see what comes out of it.

Spending money gathering data from questionnaires - especially ones not particularly suitable for ME patients that can lead to misinterpretation seems like a massive waste of resource (apart from being personally unacceptable) to me.

However, once the data has been gathered we may learn the right questions to ask - not necessarily in terms of patient questionnaires but in terms of the direction the next research steps should take.

Having focussed questions or hypothesis to research with some solid background proof via the genetic data is more likely to receive funding.

In throwing the kitchen sink at it with questionnaires, which may not be at all relevant, we may be wasting resources that could be used at a later date & in a more targeted way once the GWAS data has been examined.

 

Thanks, and thanks for the helpful questioning of this estimate of the patient community.

In terms of estimating the patient community, the key question is probably "how big is the 'silent group' that is linked to the patient community but not counted by likes of either of the two main charities' Facebook pages?"

Perhaps I should clarify that my current estimate does already include quite a few of the silent group. More details are in the box below, but in summary:

• The basic figures of 40k for the patient community = 30K from the Facebook estimate (75%) +10K (25%) who are part of the patient network in other ways (and so part of that silent group).
• However, If the overlap between AfME and ME Association Facebook likes is 70% , rather than 50%, the figures give a smaller Facebook estimate of 26K, and a bigger one for others 14K (35%).

Details:

And perhaps 35% is high enough? See below:

Age is clearly a factor. ONS stats show internet use falls with age.
16-54: 99%
55-64: 93%
65-74: 83%
75+: 47%

Ofcom stats indicate that FB is the top social media platform for older users, but social media profiles are less common for older users (and they go online to look at social media less often)
55-64: 72%
65-74: 51%
75+: 38%

Though this separate 2018 survey indicates higher FB use than this:



What the figures show is that fewer people are online over age 65; they still use some social media, mainly Facebook. It is quite likely that a good number of older people who are connected to patient online networks are included in the Facebook estimate (and some of the others will be included in the 25%-35% "other" category included in the 40k patient community estimate.

A big question is, can we reach older people outside the network, and if so, how?

I don't know if they tend to belong to forums or follow blogs, but suspect not.

You have previously suggested people asking of personal networks. That is a great idea, and I think will be critical in reaching beyond the limited online patient community deeper into the whole patient population. We should also encourage people to recruit/publicise in their communities, e.g. by writing to the local paper, posting on local Facebook groups, putting posters/flyers in their local GP surgeries.

We will also ask everybody who signs up to the study to recruit one more person to the study from their personal network.

Current recruitment plans include online advertising (e.g. on Google, but we are waiting to hear from professionals if is this is viable on a sufficient scale).

This is hugely important. There will be a PR launch, hopefully lots of publicity, plus a fired up patient community should lead to a big bang start. With luck (planning, patient support) there will be 5k or more patients signed up to receive notifications the moment the project launches.

That would be huge.

 

There is another reason to avoid lots of questionnaires: the BPS people will spin a BPS narrative out of the data if at all possible.

When the study is published and genes are found, it could be a significant setback for them. Meaning they would be in a situation where need something to not lose ground to biological explanations.

The Science Media Center will help the BPS side as usual.

 

I'm afraid I'm unconvinced. Getting one's head around multiple options is itself cognitively demanding. I'm not the one with ME/CFS (my wife is), and when she is struggling, then any kind of additional decision making is just not viable. And as @Sarah94 said, self selecting which questions you want to answer would, I believe, be too bias-prone for reliable results.

 

That had passed me by, but you are right - those people would be especially important to not overlook gathering data from.

 

Just spitballing but one way of maximizing questionnaire returns would be make it a two-step process that uses the spit-and-post test as the required step and allows for near indefinite, as practical as possible, participation in the second step of however many questionnaires people are able to answer. Likely somewhere around after a year further submissions would change little to final analysis but the value of having a base from which to survey the community is almost as significant as the GWAS itself.

Near indefinite could be something like a year at first, this kind of reanalysis should be redoable, at least the final calculation part, and at some point we would observe no significant change and could call it a day. I know the programming aspect of this is doable, even rather simple. At some point you can trunk, to borrow a software development concept, the current research tree and do an analysis on a snapshot of current data. In development terms it would represent no less than 5% of the total effort, to be able to recalculate from a new snapshot in time of data, including newly added records.

So basically participants send in their spit-and-post and receive a unique URL that allows them to answer questionnaires in the future (with additional two-step validation, say from a code given in answer by letter and email). They may answer them at their leisure and analysis could be done periodically until it obviously shows no difference as additional submissions basically make no statistical difference. If it's none it's too bad but the most severe ones will not be able to answer any, meaning the step of answering questionnaires unfortunately has to be optional.

This would even allow to add future questionnaires and direct community surveys. This could additionally serve as a research outpost, able to answer future questions and many unanswered ones about the efficacy of the current services and paradigm. It could answer such basic questions as the actual disease burden, the number who returned to work, duration, etc.

Overall, it could be thought as a significant way to increase participation by turning it into a patient registry, which can be further refined in a future step if it becomes valuable. Participation in this study would mean essentially declaring oneself to be significantly impaired by this disease and willing to participate in the solution, even if only to be counted. The message this would send cannot be overstated. Be heard.


Bit off-topic but my favorite marketing approach would be a "Take on ME" campaign, using the song of the same title by A-ha (or the Reel big fish one ). It's sad but presentation matters and this could ring in some nostalgia, sending the message that most people probably either know someone or of someone, by a single link such as knowing someone who is the sibling, child or parent of a ME sufferer. Maybe a colleague. A cousin. I doubt that more than 2 degrees of separation captures 90% of the population. Videos like that are cheap and easy to produce using freelancing platforms.

 

I'm not sure about that. Doesn't that risk including people who had something else which was misdiagnosed as ME?