Network structure of functional somatic symptoms, 2024, Fischer et al

[rvallee]

Network structure of functional somatic symptoms
https://www.sciencedirect.com/science/article/pii/S0022399924003805

Results
The final network had a relatively small number of edges, with small (46.5 %) or small- to medium-sized (47.1 %) correlations. Ten communities were identified: and cognitive problems/fatigue/depression, sensory problems, facial pain, head/neck/upper back pain, dizziness/nausea, throat pain/problems with swallowing, chest pain, widespread pain, abdominal pain/problems with digestion, and genital pain. The highest node strength in the network was found for the symptoms “tired”, “down, depressed, or hopeless”, and “tired after minimal exertion”.

Conclusions
The network analyses pointed to ten distinct groups of moderately associated symptoms in individuals with FSS. Fatigue and depression emerged as important symptoms connecting groups. Future studies should test whether (transdiagnostic) interventions specifically targeting these symptoms are particularly potent in alleviating FSS.

 

[Andy]

Full abstract.

Highlights

• The network structure of functional somatic symptoms (FSS) was estimated.
• Ten communities (i.e., groups of symptoms) emerged.
• Central symptoms connecting FSS were fatigue and depression.
• In addition, bridge symptoms connecting FSS and depression were identified.
• These included fatigue, balance, nausea, and pain/discomfort during digestion.

Abstract

Objective
The overlap among functional somatic syndromes (FSS) is substantial, which is why various empirical attempts at an improved understanding of related symptoms have been undertaken. Network analyses are particularly valuable from a clinical point of view, since they focus on the extent to which symptoms expression is co-dependent. The aim of this study was to provide the first estimation of the network structure of symptoms in 17 FSS.

Methods
N = 3054 young adults participated in an online survey. The Questionnaire on Functional Somatic Syndromes (FSSQ) was used to diagnose FSS and to assess related symptoms. The Patient Health Questionnaire (PHQ-9) was used to assess (comorbid) depression. Various R packages were used for network analysis, which yielded correlations between symptoms (edges), symptom groups (communities), and measures of centrality for individual symptoms (e.g., node strength).

Results
The final network had a relatively small number of edges, with small (46.5 %) or small- to medium-sized (47.1 %) correlations. Ten communities were identified: and cognitive problems/fatigue/depression, sensory problems, facial pain, head/neck/upper back pain, dizziness/nausea, throat pain/problems with swallowing, chest pain, widespread pain, abdominal pain/problems with digestion, and genital pain. The highest node strength in the network was found for the symptoms “tired”, “down, depressed, or hopeless”, and “tired after minimal exertion”.

Conclusions
The network analyses pointed to ten distinct groups of moderately associated symptoms in individuals with FSS. Fatigue and depression emerged as important symptoms connecting groups. Future studies should test whether (transdiagnostic) interventions specifically targeting these symptoms are particularly potent in alleviating FSS.

 

[rvallee]

If I understand this correctly, they did a bunch of graph math using questionnaires of symptoms, moods and other stuff, to find that people report those symptoms. So they basically identified the symptoms that patients report using a convoluted process. Somewhat. Genius stuff.

The conclusion is bizarre. They lump together a bunch of health issues unified by factors independent of the symptoms themselves, or the patients, rather it is medicine's lack of understanding that unifies them, then ponder if, and this has been done to death already, having lumped them together, perhaps they could transdiagnostically (i.e. generically) treat them specifically.

Unfortunately they used two useless questionnaires in the form of some psychosomatic questionnaire and the PHQ-9, which asks mainly generic questions where anyone ill without support would rate highly.

Just swinging randomly at things and shouting how they're hitting stuff very specifically, as they always intended to.

 

[Andy]

Introduction

Medically unexplained symptoms or functional somatic symptoms are highly abundant in the general population [1] as well as in primary [2] and secondary healthcare settings [3]. In specific constellations, they have received labels such as “chronic fatigue syndrome” [4], “fibromyalgia” [5], and “irritable bowel syndrome” [6], which can be found as exclusionary diagnoses in various somatic disease chapters within the International Classification of Diseases [ICD; 7]. However, the overlap among at least some of these functional somatic syndromes (FSS) is substantial [8, 9]. This has led some researchers to question the necessity of several distinct FSS [10], of which at least 17 have been described in the literature [11].

 

[Andy]

"In brief, N=3,054 young adults residing in Switzerland were recruited via college and university mailing lists between 2009 and 2010 and participated in an online survey on functional somatic syndromes. Mean age was 24.6 ± 5.6 years, 2,042 (73.4%) of the participants were women and 812 (26.6%) were men. In total, n=289 individuals fulfilled research diagnostic criteria for at least one of 17 FSS, including chronic fatigue syndrome, fibromyalgia, and irritable bowel syndrome "

Fukuda was used to diagnose "CFS".

 

[NelliePledge]

I wonder if the participants were made aware this was based on a psychosomatic view.

 

[alktipping]

Should be as yet medically unexplained .When their entire attitude is not to spend any effort into finding an explanation for persistent symptoms the rank amateurs will carry on invading this space since they have to bear no consequences for gross incompetence or outright fraud.

 

[Nightsong]

This kind of network analysis is a bit outside my experience, but a few comments:

The network had a relatively small number of edges, with small (46.5%) or small- to medium-sized (47.1%) correlations. Twenty-one edges (1.2%) were greater than 0.5 (see Supplement 1)

Network stability was sufficient for edge weights (0.52) and node strength (0.52), but stayed below the cut off of 0.5 for closeness (0) and betweenness (0.05). As the network contained nodes with no edges, closeness contained no variance (0)

I think a bit greater caution is needed in reporting the findings given both the small edge weights and the fact that multiple centrality measures are below cut-off. The network analysis alone can't establish predictive relationships and they group together very broad categories ("tiredness", "pain") without differentiating out (say) acute vs sporadic/intermittent vs chronic pain. I don't think there's great value in aggregating symptoms into communities in this way.

The cohort (n=3054) was recruited "via college and university mailing lists between 2009 and 2010 and participated in an online survey" (original paper using this dataset can be found here); being of younger and more educated adults, it is not representative of the general population. The authors say that 289 individuals (9.5%) met "research diagnostic criteria for at least one of 17 FSS", but there was no clinical evaluation; this is how they did it:

The Questionnaire on Functional Somatic Syndromes [FFSS; see 23 for the full German version] was used to diagnose FSS, which comprises of five subsequent sets of questions that are connected via a complex algorithm. First, screening questions on 51 somatic symptoms are presented (see also Table 1), which are rated according to their frequency of occurrence (“never/rarely”, “frequently”, “almost always/always”). If a participant endorses cardinal symptoms of a specific FSS (e.g., if they report frequent fatigue for chronic fatigue syndrome), they are forwarded to the next set of questions covering the research diagnostic criteria of chronic fatigue syndrome (in this case the Fukuda criteria [4]). These questions are built on established instruments to diagnose FSS wherever they exist and are scored according to existing instructions. In the case that no established instruments exist, each diagnostic criterion is answered on a dichotomous scale (“not present” vs. “present”). If a participant fulfils research diagnostic criteria, they are subsequently surveyed about health care visits (e.g., ‘Have you ever visited a doctor about your fatigue/post-exertional malaise?’). If a participant responds with “yes”, they are consequentially directed to a list of items addressing frequent medical exclusionary diagnoses (“What diagnosis did your doctor give you?”). Finally, if a participant does not have any somatic diseases which might account for their symptoms (e.g., meningitis), they are considered fulfilling criteria for a specific FSS (e.g., chronic fatigue syndrome).

They go on to say:

Current evidence-based treatment approaches for chronic fatigue syndrome, which prominently features this symptom, include graded exercise [47, 48] and cognitive behaviour therapy [48, 49]

Reference 47 is the Larun et al Cochrane exercise therapy review and reference 49 is the Price et al Cochrane CBT review.

Also, despite the use of "post-exertional malaise" once, the authors then consistently use "post-exertional tiredness", indicating that they don't understand the symptom.