The Relationship Between Somatic Symptoms, Health Anxiety, and Outcome in Medical Out-Patients

https://doi.org/10.1016/j.psc.2011.05.001Get rights and content

Introduction

Somatic symptom disorders, also known as “somatoform” disorders, are common, distressing to patients, families, and physicians but surprisingly understudied. The definitions of these disorders include impairment and/or treatment seeking, but the relationship between the symptoms and cognitive characteristics of these disorders and later outcome has not been assessed adequately in prospective studies. In particular, it is not clear whether somatic symptoms, the hallmark of somatization, and health anxiety (hypochondriasis) have similar relationships with health status, medical treatment seeking, and satisfaction with care. These relationships are important in clinical practice in order to decide which disorders will resolve spontaneously and which will require psychological treatment. They are relevant also to the examination of the somatoform disorders in preparation for DSM-V.1

One of the aims of the current revision of the DSM is to incorporate a dimensional approach to diagnosis in addition to the categorical one.2 This is useful for research and many of the findings in the field of psychosomatics indicate that the relevant variables, such as number of somatic symptoms, physiological measures, and outcomes are distributed as continuous variables.3, 4, 5, 6, 7, 8, 9 The clinical utility of the dimensional approach is a little less clear, however, as cut points on relevant scales have not been reliably established.

It is well recognized that blood pressure is distributed in the population as a continuous variable but there have also been established cut point values, above which complications are more likely. Similar cut points have not been established adequately in the field of somatic symptoms and health anxiety, although these are central to the somatic symptom disorders.1 This paper aims to assess whether it is possible to establish cut points on the dimensions of total somatic symptom count and health anxiety that are meaningfully related to outcome. The data come from a study which has been published previously10, 11, 12 but all the analyses in this paper are original. The current study also includes a broader range of outcomes than most previous similar studies; we have studied health status, frequency of medical consultations, and satisfaction.

Although somatic symptoms and health worries are ubiquitous, it is necessary to identify a threshold above which these phenomena are associated with impairment and frequent medical help-seeking. In primary care a threshold for the number of somatic symptoms which corresponds to a provisional diagnosis of “somatization” has been ascertained. Using the PHQ-15, Kroenke identified the top third and the top 10% of scorers to assess the relationship between a high number of bothersome somatic symptoms (PHQ-15) and impaired health status, time off sick, and frequent medical consultations.13 Scores of 5, 10, and 15 on PHQ-15 were said to correspond with a low, medium, and high somatic symptom severity. The data showed nicely that number of somatic symptoms and outcome are closely correlated and both appear to be continuously distributed. Barsky used a cut point on the PHQ-15 questionnaire to provide a ‘provisional diagnosis of somatoform disorder’ (approximately 20% of primary care patients) when demonstrating the association between total somatic symptom count with high health care costs and impaired function, after adjustment for confounders.14, 15

There are 2 important aspects of these studies that are relevant to this paper. First, the PHQ-15 is a self-administered questionnaire and measures all somatic symptoms; there is no attempt to identify “medically unexplained symptoms,” so the questionnaire yields a total somatic symptom score. This is important as the definition of somatoform disorders in previous versions of DSM has relied solely on ‘medically unexplained’ symptoms and less is known about the predictive power of total somatic symptom count.

Second, most previous assessments of the association between a high total somatic symptom count and impaired health status have used cross-sectional analyses preventing any conclusions about causality. Furthermore, the analyses of association with health care use have employed retrospective, not prospective, measures of health care, although we need to know the predictive value of somatic symptoms and health anxiety. The 2 studies that have used a prospective design have shown that, after adjusting for confounders: (a) somatic symptoms predicted subsequent impairment and (b) a combined measure of somatic symptoms and hypochondriacal cognitions predicted health care use.15, 16 The latter study did not identify the relative contribution of somatic symptoms and health anxiety to subsequent outcome; these are probably overlapping concepts.16, 17, 18

The study described here used prospective measures of health care use and health status. This is in line with the notion that treatment seeking and impairment should not be regarded as diagnostic criteria for somatoform disorders but they should be regarded as outcome measures.19

Although the dimensional approach is appropriate to establish relationships between variables in research, establishing appropriate thresholds is required to identify patients for treatment and to establish the prevalence of a disorder in a given population. There have been problems with determining prevalence because the threshold for diagnosing DSM-IV somatization disorder was too high for primary care or population-based studies.20 A variety of alternative measures has led to a very wide variation in the prevalence of somatoform disorders in primary care.21 Similarly, the use of a cut point on measures of health anxiety has led to widely differing estimates of the prevalence of hypochondriasis.18, 20, 22, 23, 24 The data from a clinic study, such as the one described here, cannot establish prevalence of a disorder, which requires a population-based study, but can contribute to our understanding of which clinic patients have clinically significant multiple somatic symptoms and health anxiety.

The present study concerned new patients at secondary and tertiary gastroenterology, cardiology, and neurology clinics where patients with persistent ‘medically unexplained’ symptoms are common as well as patients with serious organic disease. The first aim of this study was to assess whether cut points could be established on measures of total somatic symptom count and health anxiety measures that predicted subsequent outcome. The prospective design meant that we used number of somatic symptoms and health anxiety at first clinic visit as predictors of outcomes; the latter were health status 6 months later and number of medical consultations during the 6 months following first clinic visit. In a subsample we obtained measures of satisfaction 2 weeks after first clinic visit.12 We also compared whether there were differences between patients whose presenting symptoms could be attributed to well recognized physical diseases and those with ‘medically unexplained’ symptoms.

Our previous study indicated somewhat different relationships between somatic symptoms and health anxiety with outcome. That association was linear for somatic symptoms and outcome. However, for health anxiety, it was only the highest scores (top 10%) that were associated with poor outcome.10 We used cut points derived from that study in the present one, but we performed more detailed analyses that adjusted for relevant baseline scores of the outcome measures in addition to other confounders. Our previous study suggested that the relationship between somatic symptoms, health anxiety, and outcome was similar for symptoms explained and not explained by organic disease. We checked this further in the present study as this is implicit in the proposed DSM-V complex somatic symptom disorder.1

Our analyses aimed to answer the following questions: (1) Is the correlation between number of somatic symptoms and outcomes similar to that between health anxiety and outcomes and, in addition, are these similar for patients with explained and unexplained symptoms? (2) Do the patients who score above the cut points on the somatic symptom and health anxiety scores have worse outcomes, and is this regardless of whether the symptoms were explained? (3) Do the patients who had both high number of somatic symptoms and marked health anxiety have the worst outcomes of all? (4) Are somatic symptoms score and health anxiety score independent predictors of outcome when adjustment is made for confounders, including anxiety and depression?

Section snippets

Method

The sample comprised consenting new patients attending the neurology, gastroenterology, and cardiology out-patient departments of 2 large hospitals in the UK.10, 11 These clinics receive referrals from primary care and other out-patient clinics, ie, secondary and tertiary referrals. We approached all new patients at these clinics if they were: aged 18–75 years, physically and mentally able to complete questionnaires in English, and clearly symptomatic; we excluded asymptomatic patients (eg,

Results

Of the 383 patients approached at their first clinic visit, 292 (76.2%) joined the study. The mean age of the sample was 47.2 (SD = 13.6) years; 55% were female, 58% married or cohabiting, and 21% separated, widowed or divorced. Forty-six percent were of lower socio-economic status and 91.9% were white Caucasians. Of the 292 participants 113 (38.7%) had medically unexplained symptoms.

Cut Point Scores and Prediction of Outcome

A cut point score of 9 on the somatic symptom scale identified 24.7% of the sample. This group had significantly more medical consultations during the subsequent 6 months than the remainder and had significantly lower (more impaired) scores on both physical and mental summary scores at follow-up even after adjustment for relevant baseline values (Table 2, top half).

Ten percent of the sample scored 30 or more on the Health Anxiety Questionnaire and these participants also had significantly

Discussion

In relation to our first aim, we have shown that, among patients attending medical out-patient clinics, identification of those who score above cut point points for total somatic symptoms count and for health anxiety produces clinically meaningful groups with worse outcomes than the remainder. These findings are robust in the sense that many of the associations between the predictors and poor outcomes remained significant after controlling for the baseline value of the relevant outcome measure

Summary

Our findings suggest that cut point scores on scales of somatic symptoms and health anxiety have clinical utility as they predict outcome. Our data also suggest that the relationship between these predictors and outcomes hold for both medically explained and unexplained symptoms. We did not find evidence that general anxiety and depression were responsible for these predictor-outcome relationships. Further work is needed to test the optimal cut point scores in primary and secondary health care

Acknowledgments

This work was funded by the Central Manchester NHS Trust R&D Directorate and the Functional Gastrointestinal Disorders Working Group.

First page preview

First page preview
Click to open first page preview

References (48)

  • F. Creed et al.

    A systematic review of somatisation and hypochondriasis

    J Psychosom Res

    (2004)
  • S. Reid et al.

    Recognition of medically unexplained symptoms–do doctors agree?

    J Psychosom Res

    (1999)
  • I. Bjelland et al.

    The validity of the Hospital Anxiety and Depression ScaleAn updated literature review

    J Psychosom Res

    (2002)
  • S. Rosenzweig et al.

    Mindfulness-based stress reduction for chronic pain conditions: variation in treatment outcomes and role of home meditation practice

    J Psychosomatic Research

    (2010)
  • F. Creed et al.

    The cost-effectiveness of psychotherapy and paroxetine for severe irritable bowel syndrome

    Gastroenterology

    (2003)
  • P.K. Myint et al.

    Relation between self-reported physical functional health and chronic disease mortality in men and women in the European Prospective Investigation into Cancer (EPIC-Norfolk): a prospective population study

    Ann Epidemiol

    (2006)
  • M.C. Sha et al.

    Physical symptoms as a predictor of health care use and mortality among older adults

    Am J Med

    (2005)
  • C. Dickens et al.

    Negative illness perceptions are associated with new-onset depression following myocardial infarction

    Gen Hosp Psychiatry

    (2008)
  • L. Stafford et al.

    Are illness perceptions about coronary artery disease predictive of depression and quality of life outcomes?

    J Psychosom Res

    (2009)
  • M.C. Juergens et al.

    Illness beliefs before cardiac surgery predict disability, quality of life, and depression 3 months later

    J Psychosom Res

    (2010)
  • W. Katon et al.

    The association of depression and anxiety with medical symptom burden in patients with chronic medical illness

    Gen Hosp Psychiatry

    (2007)
  • D.A. Regier

    Dimensional approaches to psychiatric classification: refining the research agenda for DSM-V: an introduction

    Int J Methods Psychiatr Res

    (2007)
  • F.S. Luppino et al.

    Symptom dimensions of depression and anxiety and the metabolic syndrome

    Psychosom Med

    (2011)
  • J. Jackson et al.

    Differences between out-patients with physical disease and those with medically unexplained symptoms with respect to patient satisfaction, emotional distress and illness perception

    Br J Health Psychol

    (2004)
  • Cited by (0)

    The author has no conflicts of interest to disclose.

    View full text