In this study we investigated the illness representation in IBS affected individuals, different aspects of their use of the health care system, and the relation between both. The perceived consequences and the emotional representation as subdomains of the illness perception seemed to be of primary relevance. Cure and personal control played a minor role. The perceived consequences predicted the use of the conventional health care system, as well as the use of treatment approaches with lacking or weak evidence regarding effectiveness in IBS.
Study population and characteristics of the sample
In accordance with the well-established association of IBS and female gender [
1], the vast majority of the participants of our study were women. The median age was surprisingly low in comparison to that one of IBS-affected participants in other population-based studies [
34]. This might be caused by our recruitment strategy via study announcements mainly in cities with large universities and partly using mailing lists of the university. The recruitment strategy might also contribute to our finding, since the proportion of participants with a university entry diploma or university degree was high compared to the general population in Germany [
35], in line with findings that a low educational level may act as barrier for study participation [
36]. However, an analysis of nonresponders was not possible due to the completely anonymous recruitment.
Almost two third of the participants indicated the presence of a mental or gastrointestinal diagnosis, which is in an expectable range [
1]. About 20% reported a history of depression, a similar proportion a history of anxiety. Multiple answers were possible. The prevalence of depression and anxiety was lower than known from the literature: A recently published analysis based on routine data from a German health insurance plan found 58% of patients with IBS with “neurotic, stress-related and somatoform disorders”, and 42% with “affective disorders” (mainly depression) [
37]. The same study group reported 45% with at least one further diagnosis affecting oesophagus, stomach and/or duodenum [
37].
About two third of the study participants met the Rome IV criteria. This was less than expected and might be due to the fact that Rome IV is more restrictive than Rome III, resulting in a reduction of up to 50% of diagnosed cases [
38]. However, this gap between the illness representation of the individual (“being affected with IBS”) and the diagnostic Gold standard (“fulfilling Rome IV criteria”), as well as the gap between the Rome III and Rome IV criteria underline the relevance of our research question, the individual conceptualisation of IBS.
Participants fulfilling the Rome IV criteria reported IBS to be more often chronic and recurrent, with more consequences in daily life, and with less cure and personal control than patients not meeting Rome IV criteria. Additionally, they indicated a higher emotional strain by IBS. With regard to illness perception, it is noticeable that the differences between Rome IV positive and negative study participants were small, but applied to almost every subscale of the IPQ-R. This may be seen as indicating consistently more dysfunctional illness perception in individuals fulfilling the Rome IV criteria as compared to their counterparts and it raises the question, whether the perception of the symptoms initiates the dysfunctional illness representation, or the other way around. Both – the gastrointestinal symptoms influence the illness representation or the representation influences the symptom perception – would be thinkable. Also, a bidirectional relationship with the development of a vicious circle between dysfunctional illness representation and perceived symptom severity would be possible. However, these hypotheses need further research activity and empirical validation.
Doctor visits, examinations, and medication
In our sample nine out of ten patients had seen a health care professional during the last year before the survey, which was high in comparison to other studies: An Australian working group found 51.3% of IBS affected community subjects seeking conventional care due to bowel problems during the last year [
40]. More than one third of our sample was classified as “high utiliser” with six or more doctor visits due to IBS during the last year. This classification may be arbitrary to some extent, however we consider it as a suitable and rather conservative approach, as we will specify in detail below. A recently published French study [
41] based on 30.000 records from the national health data system examined health care use five years before and one year after a hospital admission due to IBS and found > 80% of patients meeting their family doctor at least three times per year independent of the time (before/after admission). The authors also reported high rates of gastroenterological visits (approximately 80% of patients within the year before hospital admission), as we did (85.4% consulted an internal medicine specialist). However, our question was not restricted to the last year, and the investigated samples were different (hospital patients vs. an unselected sample in our case). Both studies found a more intensive health care use in comparison to an American study group, which reported 50.1% of participants visiting a gastroenterologist in a two years’ time period around the IBS diagnosis [
42]. Moreover, the American researchers found a considerable heterogeneity between different US states. This underlines the role of insurance status as enabling factor in the sense of the Andersen model [
4] as described in the introduction section.
Predicting high frequency of doctor visits from the individual disease representation via a logistic regression analysis, we found age and the perceived consequences of IBS as key factors. It is conceivable, that growing older with IBS may go along with better working or adjusted coping strategies, but this hypothesis could not be verified with our dataset because of lacking information about the duration of the disease.
The significant effect of the perceived IBS consequences in our regression model might be understood as encouragement for caregivers to focus more on individual functional capability in clinical practice, instead of the number or the intensity of the reported symptoms. This approach is already adopted in the German medical Guideline [
29], which defines IBS not only as positive diagnosis with a set of symptoms, but also requires health care seeking and/or the fact, that the patient is worried about the symptoms and experiences an impairment of his or her quality of life.
The effect of the perceived consequences on health care use seemed to be independent of mental and sociodemographic aspects (PHQ-4, age, gender), but literature regarding the relationship of mental disorders and/or personality traits and health seeking behaviour is conflicting: Some researchers found associations between personality traits, mental and/or psychosocial factors and health care use [
7,
43] – others did not [
6].
Adding the variable „Rome IV “ in a further regression model (model 1b), the analysis yielded similar results than model 1a. This underlines the relevance of the assessment of subjective consequences in daily life (as assessed by e.g. IPQ-R) instead of an assessment of “pure” symptomatology (e.g. Rome IV) or focusing on symptom severity only. This finding is also in line with a study which examined the individuals’ complaint-related cognitions and the utilisation of the health care system: Van Dulmen and colleagues [
21] conducted a two time-point follow-up study in the Netherlands and found attention paid by the professional to the patients’ disease representation to be associated with lowered health care use. Additionally, they reported somatic attribution of symptoms in patients not only related to intensive health care use, but also they proposed a link between somatic attribution in patients and referral behaviour of their GPs: “the more the GP referred to internal medicine, the more their patients attributed their abdominal complaints to a somatic abnormality” [
21]. This underlines the need for clearly reflecting the indication for somatic diagnostics and for refraining from repeated investigations – not only with regards to increasing costs [
44], but also because of its negative impact on the progression of functional diseases such as IBS.
Not only over-, but also substantial underuse was found: In our study, only around one third of IBS affected women reported a gynaecological examination due to IBS, as it is recommended in the German guideline as first line diagnostic procedure [
29]. In addition, further recommended examinations such as full blood count [
29] were omitted in a significant proportion of patients, whereas other examinations such as gastroscopy and colonoscopy – useful, but not first line – were reported to be performed in over 2/3 and 3/4 of participants, respectively. Based on our data it would be daring and unreliable to label this as overuse. However, many participants in our sample reported colonoscopy due to IBS – which is associated with certain risks – but not risk-free urine examination. This may be due to a recall bias, but also limited adherence to guidelines would be a possible cause and has been described in IBS [
45], as well as in other functional gastrointestinal conditions [
44].
Limited guideline adherence could also contribute to the high proportion of medication use due to IBS: Almost one third of the study sample reported the use of NSAR because of IBS, which is against the recommendation of the German guideline [
29]. It is reasonable to assume that many patients take NSAR without prescription and as over-the-counter medication, but it may also be understood as indicating an insufficient or ineffective provision of adequate therapy in the health care system.
Treatment approaches with lacking or weak evidence regarding effectiveness in IBS
In our study, more than half of the patients indicated to use treatment approaches with lacking or weak evidence because of their IBS, and most of them were females. Female gender is also the strongest predictor in regression model 2. Both – the widespread use of evidence lacking methods, as well as the female predominance – are in line with other studies [
46,
47]. The most prevalent method reported was homeopathy, followed by yoga and meditation. Whereas there were single studies which supported yoga as complementary therapy option in IBS [
48], a recent published Cochrane review [
49] found no sufficient evidence for a recommendation for homeopathy in IBS. An updated meta-analysis regarding mind–body interventions in IBS reported no effects of meditation as treatment [
50], but considerable heterogeneity for meditation therapy, so no definitive recommendation was given.
Similarly to the regression models 1a und 1b with regard to conventional health care use, we found perceived consequences in daily life as second predicting factor for the use of evidence lacking methods; however the effect size was small.
The widespread use of evidence lacking methods in our sample may point towards a lack of trust in conventional medicine. Indeed we did not collect data regarding this question, but there is evidence suggesting that patients did not discuss the use of alternative medicine with their physician [
46]. For health professionals it is important to know, if the patient uses over-the-counter medication and/or alternative medicine – not only because of the risk of medication interaction [
47], but also to identify patient frustration with current treatments, which is an important reason to modify the current therapy.
Limitations
To the best of our knowledge, there are only few studies examining the individual representation of IBS and the utilisation of the health care system. Doing so, not only the mainly public health insurance covered sector of conventional medicine was investigated, but also the use of treatment approaches with lacking or weak evidence for IBS treatment. The large database of over 500 participants is exceptional; however, this study has a few limitations:
First, it is important to clarify how to deal with the fact, that many of the reported effects in our study were of small or intermediate effect size – both in the group comparisons (Pearson’s r) and in the regression analysis (explained variance/Nagelkerke R
2). However, considering the wide range of features predicting health care use in the Andersen model [
4] (needs, predisposing and enabling factors) and having in mind that the individual illness perception reflects only one of these aspects, we did not expect large effect sizes. Indeed there is evidence for the influence of a positive modification of illness representation during a psychotherapeutical process on patients’ quality of life [
51]. Therefore, we consider our findings as clinically relevant despite low effect sizes.
Second, a selection bias caused by our recruitment strategy seems likely: We assume an over-representation of academics and participants with higher mental strain, because this subgroup is more motivated to take part in a study like ours. This may have impact on our results.
Third, no objective validation of the diagnosis IBS or the information about health care use and other features was possible because of the nature as an internet-based questionnaire study. However, there is evidence that only about ten percent of patients with a positive diagnosis of IBS (according to Rome III) had a structural bowel lesion detectable with colonoscopy at a subsequent visit [
52]. Applying Rome IV criteria this proportion is probably even lower, because of the more restrictive criteria [
38]. Also, validity of self-reported medication and health care use seems to be satisfying [
53].
Fourth, our classification of “high” versus “normal” with a cut off at five contacts to health professionals may be seen as arbitrary: We established above, that this cut off was chosen given the fact that up to five appointments may be needed for a full diagnostic assessment of IBS complaints according to the German medical guideline [
29]. However, from a medical point of view, an optimally managed irritable bowel syndrome does not require five mandatory appointments per year. Therefore, our cut off can be considered as a rather conservative approach. This underlines the evidence for the postulated inappropriate health care use in patients suffering from IBS.
Fifth, due to the design of our study as cross-sectional study, causal relationships between illness perception and health care use cannot be drawn, thus our results have to be interpreted carefully and need further replication in longitudinal studies.