An outcome measure based on (reported) changes in activity in multiple domains

Split thread

There's actually a lot to be said for a modified version of what Collin & Crawley 2017 did - see tables 3 and 6 in this paper: https://pmc.ncbi.nlm.nih.gov/articles/PMC5513420/

[This may merit a different thread if people think it's worth discussing.]

So something like:

For each of these domains [see note at end that we could also include domains for basic ADLs]:

• Paid work
• Education
• Unpaid work including childcare
• Domestic tasks
• Social and leisure

For each domain, you would then answer the following with yes/no:

Since starting the intervention:

• I do the same amount of paid work.
• I do more paid work because of improvement in ME/CFS.
• I do more paid work for other reasons.
• I do less paid work because of ME/CFS.
• I do less paid work for other reasons.

Since starting the intervention:

• I do the same amount of social/leisure activities.
• I do more social/leisure activities because of improvement in ME/CFS.
• I do more social/leisure activities for other reasons.
• I do fewer social/leisure activities because of ME/CFS.
• I do fewer social/leisure activities for other reasons.

And the same for the other domains.

I divided Collin & Chalder’s category of unpaid work and domestic tasks (which included childcare) because someone’s ability to do domestic tasks could improve if a relative could suddenly look after their children.

Collin & Chalder had “for other reasons” only for a reduction in activity, whereas I think it should be there for an increase in activity too.

I removed "ability" and "able" so that it's just about what you do, not what you perceive you're able to do.

Then to score it, you could have Kitty’s criterion of improving in at least 2 domains and make sure that’s because of ME/CFS, not for other reasons.

I would add another criterion that the person has not reduced activity in any domain because of ME/CFS or other reasons. Because if they've stopped work then domestic and social could well go up without any improvement in ME/CFS.

So:
Number of domains the person has increased activity in because of ME/CFS?
Number of domains in which the person has reduced activity because of ME/CFS?

We would need to add domains for severe and very severe that are about basic ADLs like feeding, washing, communicating with carers, toileting.

Worth a thread maybe?

 

yup.

 

Maybe I’m misinterpreting but I think funcap does this well.

Although it’s more about what you think you can do / consequences for doing something will be

 

Thanks to moderators for creating the thread. I know there was a lot happening the day I requested it.

Yeah, @Kitty had suggested this as an adjunct to a wearable:

I had asked that Kitty’s post be copied over to the new thread so that that context would be there, but it didn’t make it.

So this whole thread is about something that would be an adjunct to a wearable.

What I thought was great about Kitty’s idea was that it acknowledged the fallibility of wearables, and suggested that pairing them up with another measure could fill the gap. If an intervention directly encourages people to be more active, at least some will be able to do that for at least a period of time. Their step count might be higher, but what if they’re doing that at the expense of another domain. For example, let’s say someone starts going for a walk every day and building that up to a longer walk, but they’re only able to do that by having a relative take the kids every morning, or by stopping the few hours of paid work they were doing from home. Or someone’s step count goes down but actually they’ve started working from home, and maybe they can work more. The wearable does not tell the whole story.

The other issue with wearables is that I think the currently available ones aren’t accurate down at the very severe and lower end of severe. For example, my FITBIT does not accurately measure the number of steps I take between bed and bathroom. But if I lost the ability to take those steps, it would be a very meaningful deterioration. For people who are bedbound, a FITBIT is unlikely to tell us much of use. There is a lot of room for wearables that are more suitable for all of us, measuring things like whether our head is up and our feet are down, but I don’t think we’re there yet. And again, they won’t capture things that are key for the very severe. Let’s say someone is now sitting up for 30 seconds twice a day, but.was able to use a bedpan/bottle independently before an intervention, and needs assistance to do that after the intervention. To say they've improved would not make much sense.

I like FUNCAP, with the exception of

• the concentration/cognitive section which I think could be used to suggest that someone who can send a text message, read an A4 page and do something cognitive for 2 hours from bed, could work
• the fact that it misses the cumulative nature of it all, i.e. that PEM is different depending on what else you’re doing that week/month

What’s good about FUNCAP is that it asks people about their PEM. What I think is different about what we’re considering here is that it would ask people what they are doing, not what they report the consequence of doing something is. The latter can be dismissed as a perception, and if an intervention targets perceptions, then we can get false signals.

Yep, that's why I suggested

Regarding vagueness of domains, Collin and Crawley described each domain a bit in the tables.

Well I think what the measure would be trying to capture is no change vs increase vs reduction in activity, and so it's up to the individual to decide what constitutes more or less or no change.

If we look at what people are doing in the context of their whole lives, they we should be able to see if a rise in step count truly does indicate an increase in their overall activity level. And if we look at that over time, maybe we can see if it's a sustainable overall increase in activity or not.

 

Yep. The comment was in the context of mildly affected people who're well enough to take part in trials.

I still think a direct comparison to healthy people might be useful for them. Some are still able to work, for instance, but researchers need to be able to pick up whether they're only managing it by stopping nearly everything else. This phenomenon can all too easily be "disappeared" in poor quality trials and manipulative questionnaires, and even participants may fail to realise how much hidden pacing they're doing if they haven't had ME/CFS long.

But it will tend to break down for everyone else. The point about multiple domains still holds, but participants need to be compared to how they were pre-intervention, not healthy people or others with milder/more severe ME/CFS.

 

Yeah, I agree. I think that's where these two components of the proposed measure would help

• @Kitty 's idea of "at least two" domains having to increase
• Collin & Crawley's idea of distinguishing between "because of ME/CFS" and "for other reasons" (tweaked by me)

I think when we have no idea whether our health is better, worse or about the same, it's probably because it's about the same. When my health truly improved and truly deteriorated, I knew.

 

There’s a dedicated thread for FUNCAP so I’m not going to derail this thread, but I want to address these two points.

I don’t think that anyone serious would suggest that someone who can only use their head 2 hours a day can work from their bed. Because you need to use your head to do other activities that are required to live.

I also think it’s a slight misconception that FUNCAP misses the accumulative effect of exertion. It ask something like «on a medium day the last month - what would be the consequence of doing X?»

If X1 results in reduced ability to function for two days, and X2 results in reduced ability to function for three days, it follows that doing both X1 and X2 in quick succession would potentially lead to a worse outcome than doing either in isolation. Or that doing X1 and another thing they didn’t ask about is worse than just doing X1.

FUNCAP also doesn’t require you to get symptoms in order to not be able to do your usual activities, because you can answer that you have to do less if that is a consequence of having to rest more to avoid getting PEM symptoms.

So FUNCAP never claims that you should be able to do X1 or X2 or Xn based on a given score. All it says that is that «everything else equal, this is the estimated consequence of doing Xn».

FUNCAP is essentially a questionnaire to assess the marginal cost of an activity.

 

I understand that, and agree. But once we're testing drugs that modify the underlying cause of ME/CFS, the changes should be more visible.

We're used to discussing treatments with such marginal effects that it can be hard to know whether positive changes are due to expectation bias, natural fluctuation in symptoms, the treatment, or a shift in routine. Positive effects are worth having, but if they're only small improvements it's a sign the problem isn't really being addressed.

It shouldn't be so difficult with a disease modifying drug, as there ought to be some marked changes even if it's not fully effective. It might include the easing of symptoms and discomforts that people were barely aware of having until they stopped—that rarely gets a mention in questionnaires and outcome measures, but it's an important indicator of feeling genuinely better.