Patient registries

I have a feeling that patient registries built on voluntary membership via the internet are only going to be of any value for conditions where you can establish features with 90-95% objective accuracy. Rare genetic diseases or conditions like relapsing polychondritis might apply. Otherwise the cohort will inevitably be skewed out of recognition, unless there is some very clear focus on a research aim like DecodeME.

 

I was surprised (having given it zero thought) that DecodeME couldn't recruit via GP surgeries doing a 10-second computer search of their patients' records and finding the ones coded for ME/CFS. Do GPs ever recruit for research studies by doing such searches and reaching out to patients?

I'd thought that in the UK, we didn't have our own NHS disease code ('SNOMED' code) but according to MERUK, we do in England, at least.

 

From what I have seen, there are almost no established routes for studies to recruit participants via the NHS in any form. Setting aside whether there is any ethical issue with GPs searching their records to find potential research participants, a research study would have to communicate directly with every GP surgery in the UK, requesting that someone at the surgery to find and communicate with those potential participants, assuming that they are coded correctly. I wouldn't expect this to have much success, GPs don't have enough time to see all the people who want to see them, let alone be proactive to find participants for somebody else's research project.

 

Good points! But could some central NHS organisation pull the info from surgeries' databases and do that? Or are the surgeries' data siloed off? I haven't followed all the data-sharing, Palantir stuff.

 

Another biobank and registry..

https://www.metrodorainstitute.co/metxstudy
Contribute to the Biobank

• Biofluids
  
  
• Solid tissue
  
  
• Pathology blocks & images

The Metrodora biobank is a secure facility housing patients’ most valuable assets. Our team treats every specimen with respect knowing it represents a human life impacted by chronic illness.

Donating specimens can be done through various studies. If a sample is required as part of a study in which you participate, our team will help facilitate collection.

Contributing specimens through research studies helps build out Metrodora’s rich dataset to more quickly identify underlying causes. Download the app to get started.

edit to add:
Google reviews are dreadful for this center and one person states that “there is no biobank”. Other people state they’ve laid off 80% of staff and “restructured” so not sure what’s going on here or w this project.

 

I don't know that the data at GP surgeries are deliberately siloed or not but in Unequal access to diagnosis of myalgic encephalomyelitis in England, 2024, Ponting and Samms the authors used hospital data as that is what is available.

However that data will be anonymised so of no use for recruitment.

 

I wonder if the fact that they're using ICD codes that don't really identify us cleanly means this wouldn't be useful anyway.

 

Health registries can backfire. There’s one under development in Norways that doesn’t include PEM and where you can’t say that you believe an infection caused you ME/CFS.

Essentially, they are trying to get rid of the entire diagnosis and label everything as MUS, FND, etc.

 

Thank you for livening up this thread, Jonathan, you must have a few more followers than me

The bolded bit is precisely where I keep getting stuck. Any ideas what such a focused research aim might be?

My interest in a registry is actually to a significant degree due to the potential flow-on effects of a well designed and executed registry. For example, it could help portray ME/CFS as a serious research topic. It could also be used by advocates as the base of an educational campaign to GPs: refer your ME/CFS patients to the registry as something patients and doctors can feel positive about doing, oh and btw here are some good clinical guidlines to diagnosing and managing ME/CFS

But all of this is predicated on the registry being well designed and executed. And for that you need at the core of the registry some clearly articulated research questions that 1) a registry can answer despite all the diagnostic inconsistency and inevitable recruitment bias, and 2) a registry can answer better or more cheaply than other methods

• Anyone with any ideas for some clear research questions that fit the bill? Talking about bill, think relatively low cost, no fancy stuff like attached biobanks etc
• And what would a minimal dataset be to make a registry worthwhile?
• Or is it an impossible enterprise for now? In which case, what would need to change before a useful registry can be established?

 

I've just learned about this preliminary study getting underway in NZ: Determining a minimum dataset for a proposed ME/CFS Registry

It would be good if we could feed them some ideas, starting with answers to the questions in the previous post

https://www.hrc.govt.nz/resources/r...ining-minimum-dataset-proposed-mecfs-registry

 

You can opt in to be part of research in the NHS app. I haven’t, so I can’t see the screens. I don’t want to be bombarded with stuff.

 

Ignoring, as I point out, that the data being anonymised makes it useless for recruitment, use of ICD codes wouldn't necessarily be an issue for recruitment as studies should screen potential participants so that they conform to the criteria that the research is using.

My guess would be that takes you to UK: Be Part of Research registry. When you sign up you can choose to be notified when studies are researching particular conditions, so I obviously have signed up to be notified when ME/CFS studies are recruiting. So far in two years I have been notified of one study as well as receiving quarterly (I think) general newsletters.

 

These would be useful things to study. Recruitment bias would preclude quantitative interpretations but it would still deepen our understanding of the range and nature of disease courses experienced, a worthwhile undertaking even if we can't put nice % figures beside them

A new registry could also be an opportunity to add some new questions about things we've been discussing here but that remain underexplored elsewhere, like the need to lie flat. Still ask the same questions other registries are asking about OI, for comparison, but add some more fine-grained ones to help interpret the whole OI thing better

In a similar vein, other registries ask about PEM, at least I hope they do, but as discussions here show respondents likely mean all manner of different things when they tick yes. I like the idea of graphical representations of PEM and PEM-like time courses, similar to what one of the Norwegian surveys, I forget which one, did for overall illness progression. It may be more informative to ask people to choose the graphic/s that best represent/s their experience than trying to explain the darn thing in words

These are just examples, for general discussion on lying flat please go to this thread and for general discussion on describing PEM please go to this thread

Some of this sort of information could be gained from one-off surveys, too, but would be easier as an add-on question in a registry. Plus a registry would facilitate checking if there's consistency or change over time

 

One problem with a registry looking at natural disease course is continuity of funding. Well, it's a problem for any sort of registry but even more so for one specifically looking at long-term outcomes. In NZ you'd have to be extraordinarily lucky to secure 5 years, 3 years is probably more realistic and even that would be difficult. Then the first year would be spent setting the whole thing up

I guess if you can show you've got a good system set up generating good data and good research you may have a better chance to get more funding later to keep it going but the uncertainty is a worry

 

Could a registry be a vehicle for collecting reports of harm following specific interventions?

I'm not sure it's particularly helpful at this stage to ask people about everything they believe helped or harmed, that would just create an unmanageable cacophony of noise. But maybe something more limited could be worthwhile, e.g. asking about significant changes in overall function over time after therapies advised by doctors (the cynic in me thinks this might even help with getting funding if presented as a way to identify useless 'therapies' that can be cut to save money, by the same token you'd have to stay mum about the possibility that your data might show the saved money should instead be spent on home help services to assist people with effective pacing)