Suggestions welcomed for (not dear) activity monitor or step counter for research study

[Tom Kindlon]

(Not sure where to post this)

I have been asked for my opinion on an ME/CFS research proposal. I want to suggest an activity monitor or step counter as an outcome measure.

Unfortunately budget isn’t huge so would need to be two figures in US$/€/£ per unit. Suggestions welcomed.

Edited to add: To be exact, €60 per unit is the maximum that would currently fit but there could be some wiggle room with that.

 

[EndME]

That should be possible within that budget, I don't have a preference for a particular device as long as the limitations are understood. Fitbit seems to widely used in ME/CFS research so that might be an option. I think it is important to have an outcome measure that is blinded, i.e. patients and clinicans don't receive feedback on step count at any point in time, which will cause some restrictions (as it is used by some patients already in daily life).

 

[NelliePledge]

Noddy question here does it really matter what the device is? Isn’t the most important aspect to have data for say six months before any intervention is tested so there’s evidence of normal activity levels for each participant then the data following any intervention for comparison. So using people’s existing kit would be a valid option. Only providing step counters for anyone who doesn’t already use one. Do we even know what level of variation there is within specific products. If someone is recorded as able to move around 10% more or 10% less is what is needed.

 

[Yann04]

I was not particularly impressed with my Fitbit experience but I know there is a device called the Fitbit Air that has no screen, so that should be far easier to blind. I think it’s barely within the price range (maybe like 90£).

I also think there are research grade ones specifically designed for trials like ActiGraph and Axivity but they are almost certainly far out of range price wise.

 

[Tom Kindlon]

I was not particularly impressed with my Fitbit experience but I know there is a device called the Fitbit Air that has no screen, so that should be far easier to blind. I think it’s barely within the price range (maybe like 90£).

I also think there are research grade ones specifically designed for trials like ActiGraph and Axivity but they are almost certainly far out of range price wise.

Thanks. My message was copied from a short social media post so not as detailed as it should be. I think £90 would probably be too dear though I’m not definite on this (it would push the total cost above the limit so it would require other costs to be cut which they might not be willing to do).

To be exact, €60 per unit is the maximum that would currently fit but there could be some wiggle room with that.

 

[Trish]

If each person is to have their steps compared with their own previous steps, they may be able to use their own existing device such as a Fitbit, given that a lot of people already have some sort of step counter. That would mean the researchers only need to supply devices to those participants who don't have them. The researchers should be able to be given permission by the patient to access the data from their device directly. (I know this is possible with Fitbit, as I gave permission to an app developer we discussed on the forum to access my fitbit data directly)

 

[Tom Kindlon]

If each person is to have their steps compared with their own previous steps, they may be able to use their own existing device such as a Fitbit, given that a lot of people already have some sort of step counter. That would mean the researchers only need to supply devices to those participants who don't have them. The researchers should be able to be given permission by the patient to access the data from their device directly. (I know this is possible with Fitbit, as I gave permission to an app developer we discussed on the forum to access my fitbit data directly)

Have people seen this method (using different monitors) being used in published papers? I understand the logic but it can be easier to justify if has been used before, ideally on more than one occasion. I don’t want the proposal to fail to get funding due to one of my suggestions not being accepted by grant reviewers.

 

[Utsikt]

The Dara pilot used a fitbit charge 5. There’s also a new upcoming model called google fitbit air. Both might be slightly too expensive, although you might be able to get some discounts if you buy in bulk or tell them you’re using them for research.

I would stick to just one model. All of the models will will have biases, but that way you get the same kind for everyone.

If the participants receive a reward for participating, that could be replaced with letting them keep the device, which might increase the budget.

 

[Felixm99]

ChatGPT 5.6 recommends buying refurbished Fitbit Inspire 3 which shows good accuracy in people with impaired mobility

Frontiers | Comparison of daily step count between the Fitbit Inspire 3 and the activPAL 3 in adults with transtibial amputation

Physical activity has significant positive effects on health. Accelerometers can be used to track daily physical activity. The Fitbit Inspire 3 is a commerci...

www.frontiersin.org

 

[Tom Kindlon]

ChatGPT 5.6 recommends buying refurbished Fitbit Inspire 3 which shows good accuracy in people with impaired mobility

Frontiers | Comparison of daily step count between the Fitbit Inspire 3 and the activPAL 3 in adults with transtibial amputation

Physical activity has significant positive effects on health. Accelerometers can be used to track daily physical activity. The Fitbit Inspire 3 is a commerci...

www.frontiersin.org

Looks like it could be a good quality, cheap option. Anyone have any idea whether any of the phasing out of it could affect anything relevant? Also wonder whether any risks involved in using refurbished options? A new version might just be in budget.

 

[Felixm99]

Looks like it could be a good quality, cheap option. Anyone have any idea whether any of the phasing out of it could affect anything relevant? Also wonder whether any risks involved in using refurbished options? A new version might just be in budget.

It still get‘a software updates and is usable with the new Google Health app. I also can‘t imagine refurbishing it would have any impact on the sensors. Probably mostly about changing batteries or screens

 

[EndME]

It seems to me that Fitbit offers very little in terms of running a "blind mode". Perhaps someone at the same institute has some ActiGraphs lying around? I do think the issue of blinding is something one should think about.

 

[Tom Kindlon]

It seems to me that Fitbit offers very little in terms of running a "blind mode". Perhaps someone at the same institute has some ActiGraphs lying around? I do think the issue of blinding is something one should think about.

They are lab scientists really, so not sure about what they would have access to but all thoughts welcome.

 

[Utsikt]

It seems to me that Fitbit offers very little in terms of running a "blind mode". Perhaps someone at the same institute has some ActiGraphs lying around? I do think the issue of blinding is something one should think about.

Why would that be an issue? People usually have a feel for how much they are doing, I don’t think knowing your step count would introduce much bias?

 

[Yann04]

Tangentially related, I’m probably in a miniscule minority so this probably doesn’t affect the trial. But just in case, personally I would think twice before accepting a study that has a wearable that sends all my health data to google (ie. fitbit).

 

[Eleanor]

Snap - I was just going to say the privacy concerns around Fitbit and Google would put me off participating.

If it's just a step counter that's needed, you can get them for a few quid each in sports and outdoors/hiking stores.

 

[Kitty]

These used to be called pedometers, and only counted how many times your feet hit the ground. it would be hard to find mechanical ones now, but basic digital ones are less than £20.

Might less complicated = more consistency between individuals? If you want simple metrics, I guess the less algorithmic interpretation is added by smart devices, the better. With such low prices it might also be feasible to buy two or three of the same unit and test them first, including whether or not fidgeting whilst sitting records steps.

One drawback could be that the memory is likely to be limited. If you needed participants to wear them for weeks, you'd have to ask them to make notes.

 

[EndME]

Why would that be an issue? People usually have a feel for how much they are doing, I don’t think knowing your step count would introduce much bias?

One problem I see is that "pacing strategies" with data from wearable devices have become quite popular. The other problem, which I'm not quite sure about myself, is that if step count is an outcome measure for a trial that has a short timeframe, then I would think having access to this data could alter behaviour (from what I’ve seen discussed in PACE people were supposedly doing more of certain exercises (as per the programme) and in turn did less elsewhere, I can imagine the opposite scenario in terms of step count can also take place.). Possibly that plays no role if the timeframes are very long, but for something short I can imagine the impacts could possibly be bigger. I think it's something worth discussing but in any case what would the benefit be of giving participants and clinicans access to such data during a trial?

 

[Utsikt]

One problem I see is that "pacing strategies" with data from wearable devices have become quite popular.

As in people staying at the same step count for pacing reasons, even if they could do more? I can see how that might blunt the measured effect. Although I’m not too concerned about it - we saw the opposite in the dara pilot. If people really do get better, I think it will show, unless they are already at 8-10k steps a day for some or the mildly affected.

The other problem, which I'm not quite sure about myself, is that if step count is an outcome measure for a trial that has a short timeframe, then I would think having access to this data could alter behaviour (from what I’ve seen discussed in PACE people were supposedly doing more of certain exercises (as per the programme) and in turn did less elsewhere, I can imagine the opposite scenario in terms of step count can also take place.). Possibly that plays no role if the timeframes are very long, but for something short I can imagine the impacts could possibly be bigger. I think it's something worth discussing but in any case what would the benefit be of giving participants and clinicans access to such data during a trial?

I think there are all kinds of ways step data could be biased. But as long as you treat it as «step data» and not a direct proxy for health or activity levels, it’s probably fine.

I think step data is mostly useful when you’d expect a large change, at least one severity level. It’s an easy way of objectively documenting that these people are actually walking quite a lot more.

 

[Kitty]

I think step data is mostly useful when you’d expect a large change, at least one severity level. It’s an easy way of objectively documenting that these people are actually walking quite a lot more.

True.

Could also be important to keep following it for quite some time afterwards. I guess that's an argument for using whatever devices people already have, which would record data before and during the trial, and keep doing so after any effects on behaviour resulting from being a trial participant have faded away.

If the intervention provides useful benefit they'll still tend to do more—or maintain activities for longer periods—than in the months running up to the trial. And whatever the quirks of devices or software, changes should be apparent when individuals are compared to themselves.