Effectiveness of a cognitive behavioural intervention for patients with undifferentiated somatoform disorder - 2019 Sitnikova et al. (CIPRUS trial)

Published in the Journal of Psychosomatic Research

Abstract
Objective: To examine the effectiveness of a cognitive behavioural intervention delivered by mental health nurse practitioners (MHNPs) to patients with undifferentiated somatoform disorder (USD), compared to usual care.

Methods: We conducted a cluster randomized trial among primary care patients with USD comparing the intervention to usual care. The intervention consisted of six sessions with the MHNP. Primary outcome was physical functioning (RAND-36 physical component summary score). Secondary outcomes were the RAND-36 mental component summary score and the eight subscales; anxiety and depression (Hospital Anxiety and Depression Scale) and somatic symptom severity (Patient Health Questionnaire-15). Outcomes were assessed at baseline, 2, 4 and 12 months. We analysed data using linear mixed models by intention-to-treat, and investigated effect modifiers.

Results: Compared to usual care (n = 87), the intervention group (n = 111) showed an improvement in physical functioning (mean difference 2.24 [95% CI 0.51; 3.97]; p = .011), a decrease in limitations due to physical problems (mean difference 10.82 [95% CI 2.14; 19.49]; p. = 0.015) and in pain (mean difference 5.08 [95% CI 0.58; 9.57]; p = .027), over 12 months. However effect sizes were small and less clinically relevant than expected. We found no differences for anxiety, depression and somatic symptom severity. Effects were larger and clinically relevant for patients with more recent symptoms and fewer physical diseases.

Conclusion: The cognitive behavioural intervention was effective in improving pain and physical functioning components of patients' health. It was particularly suitable for patients with symptoms that had been present for a limited number of years and with few comorbid physical diseases.

Trial registration: The trial is registered in the Dutch Trial Registry, www.trialregister.nl, under NTR4686.

Keywords: Cognitive behavioural therapy (CBT); General practice; Medically unexplained physical symptoms; Primary care; Problem-solving treatment; Undifferentiated somatoform disorder.

https://pubmed.ncbi.nlm.nih.gov/31285038/

 

Seems like another null result that is spinned into something positive.

The authors write:

"Overall we found statistically significant effects of our intervention, but effect sizes were small (d = 0.22 for RAND-36 PCS, d = 0.33 for role functioning/physical, and d = 0.23 for bodily pain), and lower than we aimed for (0.4 sd for the RAND-36 PCS). Also, the effect on the primary outcome was not clinically relevant (difference of 2.24 whereas a difference of 3–5 is considered clinically relevant [34]). However, effect sizes for the primary outcome were substantially higher in patients with a duration of symptoms shorter than the median (0.39) and with < 3 comorbid physical diseases (0.36). These are considered small, clinically relevant effect sizes.

Surprisingly, none of the variables that we hypothesized to be potential mediators actually mediated the effect on patients' well-being and symptoms."
​

They did not correct for multiple comparisons and only warn to interpret the p-values cautiously:

"For secondary outcomes, p-values should be interpreted cautiously due to multiple statistical comparisons, unless highly significant (e.g. p < .01)."​

 

If they trusted in their own research and the value of this methodology, this would disprove the hypothesis they are working with. If they can show improvements in physical function but no change in those, even though they are circularly defined using symptoms, then they are not causative factors. But of course it does neither, as this is not serious research, has been overdone to death and serves no purpose other than denial.

It's really disturbing how effectively these people can lie to themselves. They basically take the same approach as people who do hard tricks hundreds of times over and post the one take they managed it.

Like this one. This is a real take, no CGI or trickery. They just tried it hundreds of times and used the good one in the movie.

Spoiler: Animated GIF of a Spiderman movie

 

The following paper which includes SHarpe, Stone, Carson saying that the PHQ 15 doesn't identify people with unexplained symptoms better than chance I thought I'd do a quick google in case there were papers for which this reference might be relevant:

Somatic symptom count scores do not identify patients with symptoms unexplained by disease: a prospective cohort study of neurology outpatients - PubMed (nih.gov)

Conclusions: Self-rated symptom count scores should not be used to identify patients with symptoms unexplained by disease."

To discuss this paper, go to this thread:
Somatic symptom count scores do not identify patients with symptoms unexplained by disease: a ... study of neurology outpatients, 2015, Carson, Sharpe