select search filters
roundups & rapid reactions
before the headlines
Fiona fox's blog

a controversial treatment for children with chronic fatigue syndrome – The Lightning Process

In a new study, and despite their strong initial scepticism, researchers publishing in the Journal of Archives of Disease in Childhood report that a training course alongside specialist care appears to be effective treatment for children with chronic fatigue syndrome (CFS)/myalgic encdephalomyelitis (ME).

A roundup accompanied this analysis.


Title, Date of Publication & Journal

Title: ‘Clinical and cost-effectiveness of the Lightning Process in addition to specialist medical care for paediatric chronic fatigue syndrome: randomised controlled trial’ by Crawley et al.

Published: Wednesday 20th September

Journal of Archives of Disease in Childhood


Study’s main claims – and are they supported by the data

The paper (rather than the press release which is not currently available) does support:

  • the main outcome that physical function, as measured by the SF36-physical function subscale, was better in those in the Lightning Process (LP) group at both 6 and 12 months
  • that fatigue and anxiety were improved at 6 months in the LP group
  • that fatigue, anxiety, depression, and school attendance were improved at 12 months in the LP group.

The methods used and the caveats/limitations given by the authors are appropriate.

Participants were randomised to receive either standard medical care (SMC) or SMC+LP.   So everyone received SMC, which was defined in the same way for all participants.   This means (and the authors do point this out) that the conclusion can only be that LP is effective in addition to SMC.  The paper cannot – and does not – suggest that it is effective on its own.  Nor can it explain which aspect or aspects of LP would work to improve outcomes.

The groups were reasonably balanced at baseline.  There is some loss to follow up in both groups, with slightly higher drop out in the SMC-only group.  Loss to follow up can introduce bias.  However, the authors report that those who did not complete follow up were similar to those who did and also used an appropriate statistical method to estimate the missing values, which are still statistically significant.

It is possible that the results might vary depending upon whether participants completed all of the LP course.  To ensure this was not biasing the results, the authors undertook a CACE analysis, which is an appropriate statistical approach, and the results still showed greater improvements in the SMC+LP group than the group with SMC only.




The study is randomised so the groups were similar at baseline and can be compared to explore whether the treatment is effective, without having to worry too much about confounding.

Although the study is not large, it does have a sample size calculation, which can be replicated, and the study did recruit the required number of participants.  Loss to follow up is slightly higher than anticipated but analyses have been done in imputed data as well, which give the same result, which is reassuring.

The study has been reported and analysed in line with its own published protocol and statistical analysis plan, which suggests that there is less likely to be bias in the reporting of its results.  All the methods used are appropriate to a trial of this type.


Although SMC is defined in the study protocol, it is not always possible for researchers to ensure that this usual care is delivered in the same way to all participants.  Moreover, SMC is in fact a number of different practices that are patient specific – the number and timing of sessions varied by need and some will have chosen graded exercise therapy or received CBT.  So although everyone received SMC, the actual content of that SMC may have differed.

The study uses a patient reported outcome.  Whilst this can be a strength, because they measure whether the patient feels they have improved, as the authors point out in this study, the participants were not blinded and this may have biased the outcome.  Participants knew whether they were in the LP group and may have completed their outcome questionnaires accordingly.  LP can be costly (the introduction suggests £620 per participant) and may be seen as desirable so it is possible that the outcome measures may have been completed by patients in a way that were biased in its favour.

The authors also rightly point to the low uptake of the study amongst those who were eligible.  This study may, therefore, be a self-selected group and its results may not apply to the population of young people with CFS more widely.

The study experienced relatively high loss to follow up and quite a large proportion of participants did not complete the full LP course.

The authors have used appropriate statistical methods throughout to deal with these issues and provide estimates.  However, results based on models are never quite as good as, for example, not losing people to follow up.  It is not clear why there was attrition from both arms of the study, nor why some people who were randomised to LP did not complete all their sessions. We do not know the implications for the results, however, this might have implications if this treatment were to be rolled out more widely.



SMC (Specialist Medical Care) – Given to all patients

LP (Lightning Process) – The experimental treatment being tested


* ‘Clinical and cost-effectiveness of the Lightning Process in addition to specialist medical care for paediatric chronic fatigue syndrome: randomised controlled trial’ by Crawley et al. published in Journal of Archives of Disease in Childhood on Wednesday 20th September.


Before The Headlines is a service provided to the SMC by volunteer statisticians: members of the Royal Statistical Society (RSS), Statisticians in the Pharmaceutical Industry (PSI) and experienced statisticians in academia and research. A list of contributors, including affiliations, is available here.

in this section

filter Headlines by year

search by tag