Pacing with a heart rate monitor for people with [ME/CFS]. and [LC]: a feasibility study, 2025, Clague-Baker, Davenport, Bull et al.

[SNT Gatchaman]

Pacing with a heart rate monitor for people with myalgic encephalomyelitis/chronic fatigue syndrome and long COVID: a feasibility study

Spoiler: Authors

N Clague-Baker; T E Davenport; B Wickens; H Leeming; K Dickinson; E McBurney; K Leslie; M Bull; N Hilliard

BACKGROUND
People living with ME/CFS and LC frequently live with post-exertional malaise (PEM), which is associated with impairments in aerobic metabolism. They often use pacing with a heart rate monitor (HRM) to minimize time spent above the anaerobic threshold; however, there is limited research on the feasibility and efficacy.

OBJECTIVE
To establish the acceptability, adherence, outcomes, and adverse events associated with pacing with an HRM for a future definitive study.

METHODS
After informed consent and baseline measurements (including 10 min stand test, 5 questionnaires, accelerometry, heart rate variability, and lactate), participants were randomized into a control or intervention group using simple randomization and sealed envelopes. The intervention group used a heart rate monitor with weekly online HRM pacing advice (how to use the HRM, problem solving), and the control group received weekly online pacing advice (how to pace, problem solving). Follow-up measures were repeated, and semi-structured interviews were conducted at two and six months post-enrolment.

RESULTS
47 participants were recruited; however, recruiting people with LC was difficult due to wanting to use/already using HR monitoring. The interviews identified that the procedure was acceptable, and the majority of the participants completed the outcome measures. There were some changes from baseline to follow-up in all the outcome measures except the 10-minute stand test and accelerometry. There were no serious adverse events. Follow-up interviews identified 89% continued using HRM at 8 weeks and 66% after 6 months.

CONCLUSIONS
Studies of HRM are feasible and acceptable for ME/CFS and LC, although recruitment strategies should be reviewed for LC.

Clinical Trial registration number: ISRCTN10554129.

Web | PDF | Fatigue: Biomedicine, Health & Behavior | Open Access

 

[Hutan]

At the end of the eight weeks, the PI returned to repeat the baseline measures. The control group then received an HRM to try, and the intervention group had the choice of continuing to use the HRM or to stop using it.

This was a feasibility study. Followup was at 8 weeks. Controls were given access to the heart rate monitor at 8 weeks. There were also interviews at 6 months.

Participants in the control group accepted that they would not use an HRM because they knew they would receive a device when the intervention period ended.

That's a pretty major participation bias - people agreeing to be controls because they would get access to a device after 8 weeks. Given you really need longer than 2 months, probably 6 months, in order to work out if the heart rate monitoring approach is more useful than just muddling through, I'm not sure how feasible the study approach really is. It is likely to be hard to recruit and keep people in the control arm for 6 months if they join the study with the aim of getting access to a heart rate monitor and support for that monitoring.

 

[Hutan]

Given the short period before assessment, probably not much can be concluded. The preliminary data here doesn't seem to suggest that heart rate monitoring is really worth the effort even with the likely bias of the participants in favour of the approach e.g.

Table 3 shows a decrease in physical function from baseline to follow-up for both groups; however, both groups showed a positive increase in physical health, emotional health, and energy levels, with the intervention group showing greater changes. Both groups showed a positive increase in emotional well-being, with the control group greater than the intervention and the control group had a greater reduction in pain. The intervention group demonstrated a positive change in social functioning compared to the control group. Both groups showed a positive change in general health.

The PROMIS physical function and fatigue questionnaires were the quickest and easiest of the questionnaires to complete, although there were four missing values in addition to the dropout missing data. The intervention group showed no change in physical function but a decrease in fatigue, whereas the control group showed an increase in physical function but no change in fatigue (see Table 4).

There's a good discussion of issues with future studies. e.g.

When recruiting people with LC with PEM, it was very difficult to recruit through social media and local support groups. It appeared that people with LC were reluctant to get involved, either because they were already using an HRM or they wanted to use it and, therefore, did not want to be randomized into a control group. Since this study was conducted, Visible has further developed its app and device and published data reporting data on over 25,000 people who are now using them worldwide [Citation69]. This means finding people with LC and ME/CFS who are not using an HRM may become more difficult over time, especially recruiting through social media.

I don't think they have given enough attention to the issue of controls not wanting to wait the necessary number of months before trying an approach that is being promoted as useful. They talk about recruiting participants from clinics as these people probably won't have worked out their approach to pacing. But, these people won't be wanting to spend 6 months with an approach that they see as sub-optimal.


Interesting to see this, given Sarah Tyson's strong views here that wearables aren't feasible for monitoring people with ME/CFS

This demonstrates the feasibility of recruiting people with a range of disease severities if the study is designed to support them, following the recommendations for patient inclusion made by Professor Tyson [Citation68].

I'm not sure that lactate levels have been shown to be higher in people with ME/CFS

These measures could be continued into the definitive study to help determine if HRM can impact lactate levels and physiological stress, both of which have been shown to be increased in people with ME/CFS.

Measuring change in people with ME /CFS and LC is difficult due to the fluctuating nature of the condition [Citation74,Citation75]. Lessons from research in other fluctuating conditions, like human immunodeficiency virus, which highlight the importance of patient-reported measures [Citation76] and the need to measure health utility alongside the physical, psychological and social impacts of an intervention [Citation77].

I don't think the answer to the reality of ME/CFS being a fluctuating condition is patient-reported measures. I was surprised to see that suggestion.

Reading this paper, I am left feeling that doing a good job of an assessment of heart rate monitoring or using heart rate variability in pacing is actually really difficult. Perhaps more effort needs to be put to actually demonstrating the biological rationale for these approaches, before adding the complication of operationalising training methods (or, for that matter, promoting these techniques as useful) .

 

[Jonathan Edwards]

Reading this paper, I am left feeling that doing a good job of an assessment of heart rate monitoring or using heart rate variability in pacing is actually really difficult. Perhaps more effort needs to be put to actually demonstrating the biological rationale for these approaches, before adding the complication of operationalising training methods (or, for that matter, promoting these techniques as useful) .

Thanks for the analysis.

It brings me back to the thought that the only way to make studies of this sort interpretable is to use a 'dose-response' type format so that you do not have all the problems of unblinded controls. It may be useful to know that you can get people to wear devices for a study but we know that some people wear them anyway. More might wear them if usefulness was demonstrated.

The biological rationale is unclear but it may only emerge from detailed study data. I would suggest that a study should compare between two and four data-driven pacing strategies. Everyone would see themselves as trying out a potentially valid strategy but should probably not be told what its putative rationale is.

 

[MSEsperanza]

It may be useful to know that you can get people to wear devices for a study but we know that some people wear them anyway.

I would suggest that a study should compare between two and four data-driven pacing strategies. Everyone would see themselves as trying out a potentially valid strategy but should probably not be told what its putative rationale is.

I see they also used accelerometers. Could accelerometers also be used as such a monitoring device -- if all trial participants wear them anyway, could the data be used to develop different sensible 'potentially valid pacing strategies' that could be used as controls in addition to different strategies informed from heart rate monitoring?

 

[Jonathan Edwards]

Having as many different sorts of data as possible seems a good idea.

 

[Adrian]

The biological rationale is unclear but it may only emerge from detailed study data. I would suggest that a study should compare between two and four data-driven pacing strategies. Everyone would see themselves as trying out a potentially valid strategy but should probably not be told what its putative rationale is.

I suspect you need different data as it may be just encouraging people to actively think about pacing helps.

It brings me back to the thought that the only way to make studies of this sort interpretable is to use a 'dose-response' type format so that you do not have all the problems of unblinded controls.

Probably unethical but could you have everyone using devices but somedays they give a random signal (or a bad signal) to users others they give accurate information. Then see how people report their pacing abilities on different days.

I see they also used accelerometers. Could accelerometers also be used as such a monitoring device -- if all trial participants wear them anyway, could the data be used to develop different sensible 'potentially valid pacing strategies' that could be used as controls in addition to different strategies informed from heart rate monitoring?

I would have thought one of the interesting questions would be around correlations between different types of monitoring devices (also looking for lag in correlations). If multiple devices have signals useful for pacing then there should be evidence of correlation.

 

[ME/CFS Science Blog]

The outcome measures are hard to interpret because the authors don't report variability or statistical tests.

 

[Jonathan Edwards]

I suspect you need different data as it may be just encouraging people to actively think about pacing helps.

The idea would be that all groups are encouraged equally and one is only looking for effects specific to individual strategies.