Is it considered good research practice to retract an article, change the title, the results and conclusion, and then re-publish it with a completely new ID-number? The criticism still stands and the article should be retracted.

By Nina E. Steinkopf

In April 2020, BMJ Peadiatrics published an article about a Norwegian study in which adolescents with chronic fatigue after EBV infection were treated with cognitive therapy combined with music therapy (1). In May 2020, I published an analysis of the article, and revealed that the article is full of contradictions and illogical conclusions, and essentially disproves the researchers’ own hypotheses and own conclusion.

Read the analysis here: Curing ME with music?

The study has garnered criticism both at home and abroad, and on 14^th July 2020, it became known that the journal was conducting a full review of the article. On October 21^st 2020, the study was retracted.

The journal justifies the retraction by saying that a reader pointed out that the study was presented as a feasibility study, while the original study design was for a fully powered trial. This is referred to as an editorial error, and it is emphasized that it is not the authors who have made a mistake. The article has now been re-published with a new ID-number: DOI: http://dx.doi.org/10.1136/bmjpo-2020-000797. (2)

I have taken a closer look at the new version and found that a number of significant changes has been made that are not mentioned by the journal or highlighted by the authors.

Alterations

The title has been changed from «feasibility study» to «randomized controlled trial». One may ask whether it really is a fully powered trial when the number of participants is 43. The authors themselves write in the abstract that “the study was underpowered regarding efficacy” and that “Power analyses suggested that 120 participants would be needed in order to detect a moderate effect size.”

The publication history states that the article was first received on 14th July 2020. The original article, which was first received on 3^rd December 2019, is not mentioned here. Revision history and changes are thus lost.

It’s not until at the very bottom, in a footnote under «Provenance and peer review», it is stated that this is an amended, republished version of a retracted paper.

Under “What this study adds”, one sentence is deleted:

“Combined CBT and music therapy may improve recovery and symptoms of postexertional malaise justifying a full-scale trial”

and replaced with:  

“This study provides important baseline data for a larger trial” and “Adolescents’ concerns regarding school absence might severely affect trial recruitment”.

Results

The original article states that the number of steps/day tended to decrease, whereas postexertional malaise tended to improve in the intervention group at 3 months. At 15 months’ follow-up, there was a trend towards higher recovery rate in the intervention group.

In the new version, the authors state that the primary endpoint (number of steps/day) did not differ significantly between the intervention group and the control group, and that secondary outcome  measures were also not significantly different among the two groups.

The number of steps per day is the primary endpoint of the study. The main hypothesis was that music therapy should increase the level of function, measured in the number of steps per day. The results and the illustration has been altered.

It was originally reported that the intervention group went from 7,998 steps per day at baseline to 7,165 at 15 months follow-up. The alterations reveal that the number of steps per day was reduced to 5,680. This shows that objectively measured activity level was reduced by 29%.

This also shows that the difference between the groups at 15 months was 1,908 steps, and not 874 as originally reported. Nevertheless, the authors write that the primary endpoint did not differ significantly between the intervention group and the control group.

Figures for Fatigue have been altered. The illustration for postexertional malaise (PEM) has been removed and the results changed. In fact, all the figures in Table 3: “Outcome of the mental training program intervention”, in the “Intention to Treat” columns at 3 months and 15 months are changed (except one). Some results have been adjusted upwards, others are downgraded.

The conclusion is changed

The original conclusion: «An intervention study of combined CBT and music therapy in postinfectious CF is feasible, and appears acceptable to the participants. The tendencies towards positive effects on patients’ symptoms and recovery might justify a full-scale clinical trial.”

has been changed to:

“An intervention study of combined CBT and music therapy in postinfectious CF is feasible. A fully powered trial is needed to evaluate efficacy; participants’ concern regarding school absence should be properly addressed to secure recruitment.”

Also: the new peer-review as mentioned in the footnote has not been published.

The article should be withdrawn

Compared to the original article that was retracted, the title, results and conclusion have been altered. The article has then re-published and given a completely new DOI. Although the conclusion in the new version is toned down, there are still a number of serious concerns that justify a withdrawal. Here are some key points:

This was approved as a CFS/ME study although only 14% of the participants in the intervention group met strict CFS/ME diagnostic criteria.

The participants’ activity level at baseline was 8,000 steps per day, close to twice as high as in the level of ME in mild degree.

The results of extensive clinical examinations and biomarkers at follow-up are not included in the article.

The results show a 29% reduction in objectively measured activity level. Nevertheless, the authors consider an intervention study with combined CBT and music therapy in post-infectious CF is feasible.

Participants and their parents were told that personality, thoughts and feelings influence the development of the disease. The «situation», which was a long-term fatigue after an infection, is described by the research team as negative illness behavior, disease attribution and «avoidance». The parents were involved in the therapy because the researchers believed that parents’ influence could strengthen negative disease behavior and attribution, which in turn could have a negative impact on the prognosis.

The therapists knew in advance the results of extensive psychological tests performed six months earlier. During the program, the focus was turned towards the children’s personality, their parents and the family dynamics. Both the participants and the parents were thus implicitly made co-responsible for the «situation» and for the outcome of the therapy.

The article defines «recovered» as a certain fall in subjectively reported fatigue. The results show that 8 participants in the intervention group were «recovered» – against 7 in the control group. At the same time, some participants in the same group became more depressed, more fatigued, had more pain and a lower level of function. 8 of the 21 participants in the intervention group withdrew along the way.

The mental training program that was applied has many similarities with the pseudo-scientific method Lightning Process. According to the researchers, representatives from Recovery Norway were involved in the design of the study design. The organization was formally founded in January 2018. The study design is dated June 2015. The organization was started by i.a. a group of Lightning Process instructors, and is marketing and lobbying for the method.

The results of the study mainly disprove the researchers’ own hypotheses and conclusions about the benefits of cognitive and music therapy – without this being commented on in the article. The main goal of recovery is the number of steps, and the intervention had the opposite effect. The researchers seek to confirm their own hypothesis by selectively selecting results, drawing unfounded conclusions and overselling them. This is cherry-picking.

An honest researcher would conclude that the intervention does not seem to have any positive effect and then proceeded with other hypotheses. This is the norm in the world of research; lack of results in a pilot study means that the project is shelved. Here, the logic is turned upside down, and lack of results is used as a justification for even more psychosomatic research. The researchers seem convinced that cognitive therapy works, regardless of whether their own research shows otherwise.

The study thus becomes another example of poor research that reflects the researchers’ predetermined psychosomatic point of view, rather than an honest scientific study.

Read my analysis of the original article here: Cure ME with music?

Written by Nina E. Steinkopf

Former HSE and Quality Director in international petroleum activities, now: ME patient and writer

NB: A short version of this text has been sent to the journal BMJ Peadiatrics as a commentary on the article.

References

1. Malik S, Asprusten TT, Pedersen M, et al Cognitive–behavioural therapy combined with music therapy for chronic fatigue following Epstein-Barr virus infection in adolescents: a feasibility study BMJ Paediatrics Open 2020;4:e000620. doi: 10.1136/bmjpo-2019-000620
2. Malik S, Asprusten TT, Pedersen M, et al Cognitive–behavioural therapy combined with music therapy for chronic fatigue following Epstein-Barr virus infection in adolescents: a randomised controlled trial BMJ Paediatrics Open 2020;4:e000797. doi: 10.1136/bmjpo-2020-000797

Revised 24.10.2020: At the same time, [some] participants in the same group became more depressed, more fatigued, had more pain and a lower level of function.