Performance of the EQ-5D in Patients with Irritable Bowel Syndrome


Donald M. Bushnell, Health Research Associates, Inc., 6505 216th Street SW, Suite 105, Mountlake Terrace, WA 98043, USA. E-mail:


Objective:  The EQ-5D is a standardized, nondisease-specific instrument for evaluating patients’ preference-based valuations of health-related quality of life (HRQoL). This study's purpose was to determine the psychometric properties of EQ-5D in patients with irritable bowel syndrome (IBS).

Methods:  Data from four European IBS studies were assessed: UK (n = 161 and n = 297), Spain (n = 503), and Germany (n = 100). The EQ-5D is a five-item health state descriptive system used to develop health states (EQ-5DINDEX) and a visual analog scale (VAS) (0–100 from worst to best imaginable health state, EQ-5DVAS). Measures used with the EQ-5D included the SF-36, Irritable Bowel Syndrome—Quality of Life (IBS-QOL), and both subjective and clinical global assessments of IBS. Convergent validity was assessed using SF-36 and IBS-QOL data, discriminant validity using global ratings of IBS severity, and responsiveness by subjective and physician assessment of condition.

Results:  Moderate-to-high associations (r ≥ 0.33) were seen between the EQ-5DVAS and the SF-36 and IBS-QOL subscales. Mean response scores to EQ-5DINDEX dimensions and the EQ-5DVAS score were significantly better for control patients than for patients with IBS (all P < 0.01). The EQ-5DVAS was able to discriminate between levels of pain severity (quartiles, P < 0.001; mild/moderate/severe, P < 0.05) and general health severity (mild/moderate/severe, P < 0.001). The EQ-5DVAS and the EQ-5DINDEX were responsive in patients using both a self-perceived (Subject's Global Assessment) and physician-rated (Clinic Global Assessment) improvement.

Conclusions:  The EQ-5D performs well in comparison to general and disease-specific outcomes. It is a valid and responsive measure that can be used to generate preference-based valuations of HRQoL in patients with IBS and useful for comparisons in clinical and cost-effectiveness studies.


Irritable bowel syndrome (IBS) is a chronic gastrointestinal (GI) dysmotility disorder characterized by abdominal pain/discomfort, bloating and altered bowel habit (constipation, diarrhea or alternating periods of both) [1–3]. IBS has a broad spectrum of symptom severity, ranging from mild to severe and intractable. IBS is a common condition with an estimated symptom prevalence of 10% to 15% in Western countries [1–4].

Patients with IBS often experience symptoms for many years, with an average duration of 10 years or more [5]. IBS causes significant impairment to the quality of life (QoL) for those patients having this condition [6,7]. Patients with IBS experience physical role limitations, greater pain and a lower perception of their general health than those not having IBS. Epidemiological studies show that IBS accounts for 50% of visits to gastroenterologists in the United States and 41% of visits in the UK [8,9]. Nevertheless, only 30% of patients with IBS in North America and Europe seek medical treatment [8–13].

Most people can control their symptoms with diet, stress management, and medications prescribed by their physician. Nevertheless, traditional pharmacotherapies may not provide adequate relief of symptoms [14] because these treatments target only one of the multiple symptoms of IBS and may aggravate the others [15,16]. For some people, IBS has a major impact on QoL, affecting their work, social life and ability to travel even short distances.

The assessment of health-related quality of life (HRQoL) using patient-reported outcome (PRO) measures has evolved greatly over the last decade. One measure, the EuroQoL (EQ-5D), was developed jointly by a group of European-based researchers with the intent of constructing a simple, self-administered instrument that provided a composite index score representing the preference for a given health state [17]. The EuroQoL group designed the new instrument to be quick and easy to use, standardized and to be used alongside other measures of health status. This instrument provides both a health profile and index for individuals or groups that allow clinical and economic evaluation of medical interventions [18].

Cardiovascular and oncology are two areas where the EQ-5D has been used quite often, but new studies in musculoskeletal, respiratory, and GI diseases included the EQ-5D as generic questionnaire [19]. Furthermore, the EQ-5D is one of a handful of measures recommended for use in cost-effectiveness analyses by the Washington panel on cost-effectiveness in health and medicine [20].

Although one published study has evaluated IBS populations using the EQ-5D [21], it did not specifically address whether the EQ-5D is an appropriate and meaningful tool to use in patients with IBS. Using data from four separate studies of IBS (two in the UK, one in Spain, and one in Germany), psychometric properties were assessed to determine the validity (focusing on convergent and discriminant validity) and responsiveness of the EQ-5D in patients with IBS.


Data Sources

UK1 (n = 161) [21].  A study to identify the impact of IBS on HRQoL, time off work, and the utilization and cost of health services took place from February to November 1998. A sample of patients with IBS diagnosed by their general practitioners (GPs) was recruited from six practices in the Trent Region of the UK. The GP practices were chosen to be representative geographically and on the basis of deprivation level and social class. The GPs selected patients from their practice lists based on those prescribed antispasmodics or bowel regulators and/or diagnosed with IBS. The project researcher applied the Rome I criteria [22] to diagnose patients with IBS and only those meeting the criteria were included in the study in the IBS group.

UK2 (n = 297) [23].  A follow-up study to the UK1 was conducted, from May to September 2001, to further assess the quality of life (QoL) and resource use by patient-perceived IBS severity. The same enrollment, inclusion, and exclusion criteria were used. IBS patients were identified from GP records from 13 practices in the Sheffield area of the UK. In addition to the IBS sample, using the same methodology, 326 control subjects were recruited, case matched, by their GPs, on age, sex, and social characteristics. The IBS sample consisted of male or female patients aged 20 to 65 years who were currently on antispasmodics/bowel regulators, and/or had been diagnosed with IBS.

Spain (n = 503) [24].  This nationwide, multicenter observational, prospective, cohort study, initiated in October 2000, compared the impact of IBS on patients with the three IBS subtypes: constipation, diarrhea or alternating between constipation and diarrhea. The study assessed GI symptoms, QoL, psychological well-being, use of resources and environmental stress. Eighty-five gastroenterologists and 25 primary care doctors were involved. Subjects meeting Rome II criteria [25] were recruited from primary health-care centers, external hospital clinics, specialized medical centers, and other outpatient care sites.

Germany (n = 100) [26].  This observational study aimed at assessing the impact of IBS on QoL and the utility benefits of tegaserod began in 1999. This study used the EQ-5D and the disease-specific QoL questionnaire (IBS-QOL) to design a decision model to link the QoL benefits of tegaserod based on IBS pain severity. Patients with a diagnosis of IBS, and who had been treated for IBS for at least 12 months prior to enrollment, were interviewed by a random selection of private practice physicians.

QoL Assessment (Table 1)

Table 1.  Instruments used in the quality of life questionnaires
InstrumentNumber of questionsIssues coveredMeasurement
  • *

    The EQ-5D includes the EQ-5D visual analog rating scale.

  • EQ-5D, EuroQOL instrument; HRQoL, health-related quality of life; IBS-QOL, Irritable Bowel Syndrome—Quality of Life questionnaire; QoL, quality of life; SF-36, Medical Outcomes Study 36-item Short Form health survey.

SF-3636Eight dimensions: physical functioning, role limitations because of physical health, role limitation because of emotional health, social functioning, bodily pain, mental health, vitality or energy and general health perceptionResponses to each question within a dimension are combined to generate a score from 0 to 100, where 100 indicates good health and 0 indicates poor health
EQ-5D* 5Mobility, self-care, usual activities, pain, and anxiety−0.6 to 1 scale that was multiplied by 100 for compatibility with the SF-36 and IBS-QOL: −60 to 100. A value of 100 indicates good health, 0 death and negative values health states worse than death
IBS-QOL34Overall score, 8 subscale scores including: dysphoria, interference with activity, body image, health worry, food avoidance, social reaction, sexual, and relationships100 indicates highest possible QoL and 0 indicates worst possible HRQoL

EQ-5D.  The EQ-5D consists of two parts: the health states descriptive system and the visual analog rating scale (VAS). The descriptive system records the level of self-reported problems on each of the five dimensions of the classification (mobility, self-care, usual activities, pain/discomfort, anxiety/depression). For each dimension the respondent is asked to choose between three options: no problem, some/moderate problems, or extreme problems/unable. Health states defined by the five-dimensional descriptive system can be converted into a weighted health state index by applying scores from value sets elicited from general population samples [27]. Respondents then describe their own health status using a VAS. A 20-cm vertical VAS has become the standard means of obtaining valuations for health states. The endpoints of the VAS are labeled “best imaginable health state” and “worst imaginable health state” anchored at 100 and 0, respectively. Respondents are asked to indicate how they rate their own health state by drawing a line from an anchor box to that point on the VAS that best represents their own health on that day [18]. UK English, Spanish, and German versions have been adapted culturally and translated [28].

SF-36.  The Medical Outcomes Trust Short Form 36-Item Health Survey (SF-36) is a generic measure of functional status and well-being [29]. It contains 36 questions that measure health across eight dimensions—physical functioning (PF), role limitations because of physical health (RP), role limitation because of emotional health (RE), social functioning (SF), bodily pain (BP), mental health (MH), vitality or energy (VT), and general health perception (GH). Responses to each question within a dimension are combined to generate a score from 0 to 100, where 100 indicates “good health” and 0 indicates “poor health.” The SF-36 survey was included only in the UK studies presented in this article.

IBS-QOL.  The Irritable Bowel Syndrome—Quality of Life questionnaire (IBS-QOL) is a 34-item condition-specific instrument allowing the evaluation of QoL of patients with IBS by means of eight domains (dysphoria, interference with activity, body image, health worry, food avoidance, social reactions, sexual, and relationships). This questionnaire has proved to be valid and reliable for IBS patients [30,31] and has been used in a wide range of clinical trials.

Clinical Variables

To aid the evaluation of validity of the EQ-5D in IBS patients, both subjective and clinical (physician) global assessments of pain were used. In the UK studies, a 0- to 100-point VAS item was given to the patient to rate their assessment of IBS-related pain severity on abdominal pain and discomfort (VAS-ADP). In the Spanish study, a similar VAS item was given to both subjects and physicians to rate IBS severity. In the study from Germany, the severity was defined by subject's rating of severity as mild, moderate or severe.

Data Analyses

Participants who completed the QoL assessments at baseline in all studies were included in the analysis. EQ-5D scores from each study were described by their mean, standard deviation, median, interquartile range, and the percent of scores at the floor and ceiling of the distribution. All analyses were conducted using spss® for Windows, Release 11.5 [32].

The validity of the EQ-5D was tested by means of comparisons with validated instruments used to measure various characteristics of health status. Construct validity of the five individual EQ-5D items and EQ-5DVAS was tested using Spearman correlations with the SF-36 domains and the IBS-QOL subscales and total score. Using a correlation between 0.3 and 0.5 as the definition for moderate correlation and correlations of at least 0.5 for strong correlations [33], the following relationships were hypothesized:

  • 1Correlations between EQ-5D and the SF-36 would be stronger than those between EQ-5D and IBS-QOL because the EQ-5D and the SF-36 are both generic measures of health status and the IBS-QOL is an IBS-specific measure.
  • 2EQ-5D mobility would be strongly associated with the SF-36 physical function domain and moderately correlated with the SF-36 social functioning domain.
  • 3EQ-5D pain/discomfort would be strongly correlated with SF-36 bodily pain domain and moderately correlated with the SF-36 physical functioning domain.
  • 4EQ-5D anxiety/depression item would be strongly associated with the SF-36 mental health domain and moderately correlated with the SF-36 vitality domain.
  • 5EQ-5DVAS would be moderately to-strongly correlated with the SF-36 general health domain.
  • 6EQ-5D pain/discomfort would have stronger associations with the IBS-QOL domains and total score than the other EQ-5D items.

Known-groups construct (discriminant) validity of the EQ-5DVAS (all studies) and the EQ-5DINDEX (in UK studies only) was examined by testing the hypothesis that as IBS severity increases (or worsens), the EQ-5D scores decrease (health status also becomes worse). Results from the subject's global assessment and VAS-ADP item (available in the UK2 study) were divided into tertiles and used to graph the mean EQ-5DVAS scores. Results were investigated for the German study by graphing the EQ-5DVAS mean scores against the subject's response of mild, moderate, and severe IBS pain; and for the Spanish study using the subject's response on a seven-point severity of IBS scale (very good to very bad). No assessment of severity was available in the UK1 study. Analysis of variance was used to evaluate differences between groups. Group differences were identified using Scheffe post hoc procedures.

To assess responsiveness, data from Spain were used because this study included a 1-year follow-up measurement. Mean change scores (between baseline and 1 year) were calculated for both subjective and physician global ratings of IBS severity. Based on these change scores, subjects’ IBS was identified as having improved, remained the same, or declined in severity. EQ-5DVAS and EQ-5DINDEX scores were then compared across the change in IBS severity. Responsiveness was reported in terms of the effect size statistic (mean change score divided by the standard deviation of baseline score) [34]. This effect size identified the differences in EQ-5D scores associated with the change in IBS severity over time.


The study samples included: UK1, 161 subjects; UK2, 297 subjects; Spain, 503 subjects; and Germany, 100 subjects. The majority of subjects were female in all studies (Table 2) with a mean age of between 43 and 54 years. As seen in Figure 1, the mean scores for each of the five EQ-5D dimensions were very similar across the different studies. Median scores were 1.00 for mobility, self-care, and usual activities, and 2.00 for pain/discomfort and anxiety/depression in all studies. On individual questions, fewer people reported problems with EQ-5D self-care (8.1% in UK1, 12.8% in UK2, 5.2% in Spain, and 0% in Germany). The majority of subjects reported experiencing pain/discomfort (83.2% in UK1, 84.2% in UK2, 77.0 in Spain, and 93.0% in Germany). The mean EQ-5DVAS score ranged from 55.3 in Germany to 62.4 in UK1 and the EQ-5DINDEX scores ranged from 0.62 in UK2 to 0.70 in Spain (Fig. 2). Figure 2 also includes mean EQ-5D scores for the control group (n = 326) included in the UK2 study. The scores (VAS = 81.0, INDEX = 0.84) indicate that the EQ-5D is able to detect differences between individuals with IBS and controls (P < 0.001).

Table 2.  Sample characteristics
SampleDescriptionAge (year) (mean, SD)Sex (% female)
  1. HRQoL, health-related quality of life; IBS, irritable bowel syndrome.

UK1 (n = 161)To identify impact of IBS on HRQoL, work, and utilization47.1 (11.9)86.3
UK2 (n = 297)To compare HRQoL, utilization, and cost by self-perceived IBS severity48.2 (11.6)83.9
Spain (n = 503)Observational study to assess impact of IBS in Spain42.7 (15.2)75.9
Germany (n = 100)Observational study to assess impact of IBS in Germany53.8 (15.6)80.0
Figure 1.

EQ-5D descriptive system mean scores. Note: Higher scores indicate greater negative impacts.

Figure 2.

EQ-5DVAS and EQ-5DINDEX scores by study. Note: EQ-5DINDEX scores have been multiplied by 100 for comparison; Error bars represent standard error of the mean. VAS, visual analog rating scale.

The hypothesized relationships were confirmed between the EQ-5D and the SF-36 (Table 3) and IBS-QOL domain scores (Table 4). The correlations between the EQ-5D and SF-36 were generally stronger than those found between the EQ-5D and IBS-QOL. The EQ-5D mobility item correlated strongly with the SF-36 physical function domain (−0.70 and −0.71). The EQ-5D pain/discomfort item correlated strongly with the SF-36 bodily pain domain (−0.63 and −0.66). The EQ-5D anxiety/depression item was strongly correlated with the SF-36 mental health domain (−0.65 and −0.60). The EQ-5DVAS item was most strongly associated with the SF-36 general health domain (−0.71 and −0.67). Correlations between the EQ-5D items and the IBS-QOL total score were all moderate (−0.30 to 0.51) in the UK2 study. In the Spanish study, the EQ-5D usual activities, pain/discomfort and anxiety/depression items had moderate correlations with the IBS-QOL (−0.31 to 0.47), but the EQ-5D mobility and self-care items were smaller, but still significant (−0.12). Associations in UK1 and German studies were similar (Table 4).

Table 3.  Correlations between the EQ-5D and the SF-36
EQ-5DSF-36 domains
Social functionRole—emotionalMental healthVitalityPhysical functionRole—physicalBodily painGeneral health
  • *

    Correlation is significant at the 00.05 level (2-tailed).

  • Correlation is significant at the 00.01 level (2-tailed).

  • VAS, visual analog rating scale.

UK1 (n = 161)
 Usual activities−0.593−0.240−0.230−0.491−0.575−0.402−0.606−0.588
 EQ-5DVAS 0.562 0.135 0.352 0.676 0.550 0.330 0.534 0.669
UK2 (n = 297)
 Usual activities−0.608−0.461−0.361−0.448−0.554−0.683−0.565−0.567
 EQ-5DVAS 0.613 0.485 0.477 0.536 0.588 0.565 0.609 0.710
Table 4.  Correlations between the EQ-5D and the IBS-QOL
DysphoriaInterference with activityBody imageHealth worryFood avoidanceSocial reactionSexualRelationshipOverall score
  • *

    Correlation is significant at the 00.05 level (2-tailed).

  • Correlation is significant at the 00.01 level (2-tailed).

  • Cannot be computed because the “Self-Care” item is constant (all responses are “1”).

  • IBS-QOL, Irritable Bowel Syndrome—Quality of Life questionnaire; VAS, visual analog rating scale.

UK1 (n = 161)
 Usual activities−0.383−0.460−0.371−0.379−0.266−0.401−0.260−0.323−0.442
 EQ-5DVAS 0.391 0.406 0.325 0.423 0.263 0.347 0.305 0.369 0.470
UK2 (n = 297)
 Usual activities−0.311−0.340−0.246−0.257−0.214−0.240−0.325−0.248−0.335
 EQ-5DVAS 0.536 0.532 0.392 0.425 0.401 0.420 0.396 0.433 0.519
Spain (n = 503)
 Usual activities−0.385−0.332−0.266−0.247−0.255−0.330−0.385−0.310−0.408
 EQ-5DVAS 0.460 0.304 0.319 0.318 0.349 0.360 0.373 0.364 0.466
Germany (n = 100)
 Mobility−0.149−0.154 0.075−0.143−0.152−0.071−0.136−0.134−0.150
 Self-care 0 0 0 0 0 0 0 0 0
 Usual activities−0.319−0.269−0.166−0.201*−0.303−0.124−0.322−0.149−0.313
 EQ-5DVAS 0.396 0.188 0.192 0.322 0.279 0.079 0.220* 0.145 0.322

Figure 3 shows that the predicted relationship between the EQ-5DVAS score and the VAS-ADP (defined in quartiles) was confirmed in the UK2 study. EQ-5DVAS scores decreased as levels of subjective severity increased (F = 22.1; P < 0.001). Similar results were observed in the German study where EQ-5DVAS scores decreased as subject's assessment of their IBS pain increased (F = 4.4; P < 0.05) (Fig. 4) and in Spain where EQ-5DVAS scores decreased as IBS severity increased (F = 16.8; P < 0.001) (data not shown). The EQ-5DINDEX discriminated between tertiles of subjective pain severity, decreasing as levels of severity increased (data from the UK2, F = 29.3; P < 0.001) (Fig. 5).

Figure 3.

EQ-5DVAS by subject's global assessment (SGA) of pain severity (VAS scale). Data source: UK2 (n = 297). CI, confidence interval; VAS, visual analog rating scale.

Figure 4.

EQ-5DVAS by subject's global assessment of pain (Categorical scale). Data source: Germany (n = 100). CI, confidence interval; VAS, visual analog rating scale.

Figure 5.

EQ-5DINDEX by subject's global assessment (SGA) of pain severity (VAS scale). Data source: UK2 (n = 297). CI, confidence interval; VAS, visual analog rating scale.

Figure 6 shows the responsiveness of the EQ-5DVAS score for different levels of subjective and physician global assessments of IBS severity from the data collected in Spain. Subjects who reported a decline or worsening in IBS severity also experienced a decline in EQ-5DVAS scores. Subjects for whom IBS severity remained the same experienced slight improvements in EQ-5DVAS mean scores, whereas those who improved reported large improvements in EQ-5DVAS scores. The clinician rating of severity resulted in smaller amounts of patient-reported levels of improvement and decline in EQ-5DVAS scores. Effect size statistics were 0.64 for VAS-ADP and 0.48 for the Clinician's Global Assessment (CGA). Mean group comparisons showed each of these differences to be significant (P < 0.001). Similar trends were seen using the EQ-5DINDEX score with effect size of 0.33 and 0.29, respectively.

Figure 6.

EQ-5DVAS response using global assessment of IBS severity (after 1 year). Data source: Spain (n = 503). Effect size = change score divided by standard deviation at baseline. Note: Error bars represent standard error of the mean; number discrepancies relate to patients lost to follow-up. ES, effect size; IBS, irritable bowel syndrome; VAS, visual analog rating scale.


Any PRO measure, such as the EQ-5D, should demonstrate satisfactory properties before being used in studies of specific populations, such as patients with IBS. Although there have been studies exploring the health status of IBS patients [7,21,35], there is a lack of information regarding the applicability and sensitivity of these measures in determining the impact of IBS on patients.

The present study has certain limitations that need to be taken into account when considering the study and its contributions. In this research, data were obtained from several studies to provide evidence to support the use of the EQ-5D in an IBS population. Inclusion and diagnostic criteria varied across studies. For example, in the UK1 study, patients meeting the Rome I criteria for IBS were enrolled and the German sample consisted of constipation-predominant or alternating IBS patients). Another limitation in this analysis includes the variation in measures used. Although the data were collected from studies with divergent methodologies in several European countries, the EQ-5D scores remain similar with self-care having the least impact and pain/discomfort having the greatest effect on patients’ QoL.

From the studies in the UK, we observed that each dimension of the EQ-5D was most highly associated with the corresponding dimension measured concurrently by the SF-36. The strongest relationships were seen between mobility (EQ-5D) and physical functioning (SF-36); the pain/discomfort dimension (EQ-5D) and bodily pain (SF-36); and anxiety/depression (EQ-5D) and mental health (SF-36).

The ability to convert EQ-5D scores to utility measures is advantageous because it provides a quantitative expression of an individual's preference for, or the desirability of, a particular state of health under conditions of uncertainty. These utility scores can then be used as the quality weights when computing a quality-adjusted life-year. This is a universal health outcome measure that is applicable to all individuals and all diseases, and enables comparisons across diseases and across programs. This feature of the EQ-5D provides greater support in cost-effectiveness studies where more simple descriptive measures may not. It is noted that the VAS scale should not be used to measure utilities directly, rather health state evaluations are more relevant to obtain utility measures [36].

The EQ-5D performs well in comparison to general and disease-specific outcomes and data presented in this article demonstrate that it is a valid and responsive PRO measure that can be used to generate preference-based valuations of HRQoL in patients with IBS and useful for comparisons in clinical and cost-effectiveness studies.

This research was funded by a research contract from Novartis Pharma AG, Basel, Switzerland to Health Research Associates, Inc. Donald Bushnell and Mona Martin have no financial stake in Novartis Pharma AG. This article was presented orally at the 6th Annual European meeting of the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) on November 11, 2003 in Barcelona, Spain.

Source of financial support: This is a Novartis-funded research study.