Irritable bowel syndrome and health-related quality of life : A population-based study in Calgary , Alberta

Departments of 1Community Health Sciences, 2Medicine and 3Psychiatry, University of Calgary, Calgary, Alberta Correspondence: Feng Xiao Li, Department of Community Health Sciences, University of Calgary, 3330 Hospital Drive NW, Calgary, Alberta T2N 4N1. Telephone 403-220-8843, fax 403-270-7307, e-mail fli@ucalgary.ca Received for publication September 24, 2002. Accepted February 24, 2003 FX Li, SB Patten, RJ Hilsden, LR Sutherland. Irritable bowel syndrome and health-related quality of life: A population-based study in Calgary, Alberta. Can J Gastroenterol 2003;17(4):259-263.

Although not a life-threatening illness, IBS can cause great distress to the patients.People with IBS may also experience other symptoms such as nausea, fatigue, headache, sexual dysfunction and urinary symptoms (eg, incomplete bladder emptying).These intestinal and extraintestinal symptoms can adversely affect the health-related quality of life (HRQOL) of people who suffer from IBS.To date, few studies, especially those using population-based data, were conducted that evaluated the HRQOL in people with IBS.There have been no population-based Canadian studies.
Most people with IBS do not seek medical care (4).Community-based studies show that only 25% to 60% of people with IBS present for medical care for evaluation or treat-ment of their symptoms (5).How frequently people with IBS seek medical help in Canada and how the help-seeking behaviour of people with IBS is affected by their HRQOL in general have not been well studied.As part of a multipurposed pilot study, the goals of the present study were to examine the impact of IBS on the HRQOL using a population-based urban sample and to compare the HRQOL of people with IBS who did and did not seek medical help.

METHODS
A random sample of the Calgary general population (18 years of age or older) was recruited during March and August 2001 using Random Digit Dialing (RDD).RDD is an established sampling procedure that, theoretically, provides an equal probability of reaching a household (6).The list of RDD numbers within the Calgary Census Metropolitan Area was generated by Canada Survey Sampler (a private firm) using traditional Mitofsky-Waksberg methodology (7).The list contained telephone numbers for all potential households on the list, and names and addresses for the households that were listed in telephone directory.A total of 1521 RDD numbers was used in this study.
The diagnosis of IBS was made using the Rome II Criteria, derived from the section on bowel disorders from the Rome II modular questionnaire, for which two algorithms were created (8) (Table 1).A study subject was determined as a case of IBS if "Q20 = Yes, plus two of (Q21 or Q22 or Q23) = Yes" or if "Q10 = Yes, plus two of (Q13 or Q14a or Q14b) = Yes".The Rome II Criteria are widely used by investigators and are considered to be valid and having good predictive value for IBS (9).All the individuals in the sample who met the Rome II criteria were considered as cases of IBS; otherwise, as control subjects.
The HRQOL was assessed using Medical Outcomes Study Short-Form 12-Item Health Survey, version 2 (SF-12v2) (10).SF-12 is a validated HRQOL measure (11) and can be used to compare patients with healthy individuals.SF-12v2 includes eight scale scores, ie, physical functioning, social functioning, role limitation-physical, role limitation-emotional, mental health, vitality/energy, bodily pain and general health.From the eight scale scores, two summary scores (physical component score and mental component score) are derived.All the scale and summary scores have a range of zero to 100 and are designed (using 1998 general United States population) to have a mean score of 50 and a standard deviation of 10 so that a high score indicates better health.For example, a mean score of 45 in the IBS group would mean that 69% of the general population has a better HRQOL than the average IBS patient group.
A questionnaire was developed that included the above standard forms and sections on demographics and history of healthrelated visitation.Based on the respondent's request, the questionnaire was either interviewer-administered (by telephone) or self-completed (through the web, by interactive voice telephony, or by mailed/emailed questionnaire).Before data collection commenced, each mode of the questionnaire administered was pilot-tested in five subjects and proved to be easy to answer.
The data collection involved three steps: recruiting households by mail or phone; selecting an individual from the households; and completing the questionnaire by the selected individuals.A selfaddressed stamped card with a cover letter was first mailed to the addressed households.On the card, the households could specify if they would be willing to participate in the study and which method they would use to complete the questionnaire.The card also specified a Web site, a 1-800 number and a mail and email address, to be used to complete the questionnaire.The households that did not respond to the mailing or did not have an address were called directly.The study was introduced verbally over the phone.The recruited households would then select an individual (18 years of age or older) using the 'last birthday method' (12) detailed in the cover letter or over the phone.This individual was allowed to decide which method to use to complete the questionnaire.
To analyze the data, the SF-12v2 scale and summary scores were first determined for each individual subject.To accomplish this, the SF-12v2 items were first scored in a computerized algorithm (10) for each individual subject.The means and standard deviations were then determined for each scale and summary score separately for people with and without IBS and for IBS cases with and without a physician-visiting history.The differences of the mean scores in each category and their 95% confidence intervals (CI) were determined as both crude and demographic-adjusted (for age, sex, education and household income).The adjustment for demographic variables was conducted using multiple regression analyses.

Sample characteristics
Of the 1521 RDD numbers used, 951 households were contacted successfully, with the remaining being nonresidential or nonworking numbers (note: RDD contains not only residential telephone numbers but also business and not-in-service telephone numbers).Of the 951 households, 590 (62%) were willing and eligible to participate, of which 437 subjects (74%) completed the questionnaire.The age distribution of the 437 subjects was similar to that of the Calgary general population for both men and women (Figure 1).The authors identified 110 individuals who met the Rome II criteria and 327 individuals who did not.Table 2 shows the demographic characteristics of individuals who did and did not meet the Rome II criteria.*Often means that the symptoms were present during at least three weeks (at least one day in each week) in the last three months people without IBS.The eight scale scores were systematically lower in people with IBS than in those without.Adjustment of demographic variables (ie, age, sex, education and household income) of the study subjects did not change the differences either in direction or in significance.As with the scale scores, the two mean SF-12v2 summary scores were also significantly lower in people with IBS than in those without, with or without adjustment for demographic variables.While the mean scale and summary scores for people without IBS were close to the normal standard for the United States general population (ie, 50, Figure 2), they were all below 50 for people with IBS.

HRQOL for IBS patients with and without seeing a physician
Of the 110 subjects with IBS, 44 (40%) visited a physician for their gastrointestinal problems in the previous three months.
The mean SF-12v2 scale and summary scores for those who did and did not visit a physician are presented in Figure 3 and Table 4.As shown, all the mean scale and summary scores were lower than the normal standard of 50, whether or not the subjects visited a physician.There were, however, differences in the mean scores between the two groups.As indicated by the two summary scores, people with IBS who sought medical help reported worse physical health than those who did not, but their mental health was no different.This is also reflected by the difference in the scale scores.The scales with a significant difference between the two groups include general health, social functioning and role limitation due to physical health.While the former two scales included the physical components, the latter directly indicated reduced physical well-being.
No scale score on mental health showed a significant difference between the two groups.The difference remained unchanged in direction or in significance after adjusting for age, sex, education and household income.

DISCUSSION
Previous studies have demonstrated the negative effect of IBS on HRQOL of its sufferers (13)(14)(15)(16)(17)(18).Most of these studies were conducted in hospital settings.Much less is known about the HRQOL of nonclinical samples of people with IBS.Using population-based data in Alberta, we examined the impact of IBS on HRQOL as measured by the SF-12v2.We found that: 1.People with IBS experienced significant impairment in HRQOL compared with that of the population-based controls; 2. The HRQOL of the randomly selected control subjects was close to the normal standard (ie, 50) of the United States general population; 3. People with IBS who sought medical help had reduced physical but not mental health compared with those who did not seek medical help; and 4. Forty per cent of people with IBS seek medical help in the Alberta context.
Our finding that people with IBS experienced significant impairment in HRQOL is consistent with that of the previous studies.Using a hospital-based sample of patients and the SF-36, Gralnek et al (14) found that IBS patients had significantly worse HRQOL than that of the general United States population.The reduced HRQOL covered each of the eight scales and the two summary components.Similar findings were reported by other investigators in the United States and the United Kingdom (13,19).One difference between the present study and the previous studies is that the previous studies used published general population data as the source for controls.The present study collected population-based asymptomatic individuals as control subjects.The HRQOL of the randomly selected control subjects in this study, as determined by the eight scale scores (ranging from 49.0 to 52.6) and the two summary scores (separately 51.3 and 50.2), is, however, close to that of the United States general population.
A large proportion of people with IBS do not seek help from health professionals.Our finding that 40% of people with IBS have consulted health professionals for evaluation or treatment of their illness fits the previously reported 25% to 60% range (5).Previous studies found that HRQOL affects health  $65,000 to $75,000 $75,000 or more care-seeking behaviour.Whitehead et al (13) reported that the SF-36 scores were significantly lower in IBS patients who sought medical help than in those who did not, on the general health, role physical, social functioning, vitality and bodily pain scales.Our study found that people with IBS who sought medical help had reduced scores on general health, role physical and social functioning, which are consistent with Whitehead's findings.Although there is no significant difference on the scores of vitality/energy and bodily pain (between those who did and did not seek medical help) in our study, both studies indicated that the physical health components rather than the mental health components of HRQOL were associated with IBS patients' health care-seeking behaviour.
The increased prevalence of depression and anxiety in people with IBS has been hypothesized to play a role in health care seeking behavior (20)(21)(22).Our study, however, did not provide evidence that supports this hypothesis, but rather reinforced Whitehead's findings.
To our knowledge, this is the first study of the HRQOL of people with IBS that used SF-12v2 as measurement tool.SF-12 is a shorter version of SF-36 that has been translated into 40 languages and has been used in many health outcome studies (23).Both instruments are generic measures of HRQOL and allow for the comparison of patients with healthy individuals.SF-12, rather than SF-36, was used in this study because its reduced administration time helps to improve the response rate for this pilot study.SF-12 is also a clinically-validated set of HRQOL measure (11).The two instruments have the same eight scale measures (ie, physical functioning, social functioning, role limitation-physical, role limitation-emotional, mental health, energy/vitality, bodily pain and general health perception) and two summary components (ie, physical component score and mental component score).As a result, the findings based on the two instruments can be directly compared with each other.
The use of the population-based IBS patients and control subjects is the major strength of this study.Studies that use hospital-based IBS patients are likely to over-sample the more severe cases.As a result, the reduced HRQOL based on such cases may not be generalizable to all IBS patients.Similarly, hospital-based or institutional controls may represent people  There are many possible explanations for the difference in the prevalence estimates between the two studies.For example, regional variation in IBS prevalence is one potential explanation (ie, the different prevalence estimates resulting from the differing target populations).The low response rate in the previous Canadian study may be another explanation.The 'worst case' scenario for our study, however, would be selected completion of the survey questionnaire by those who had IBS symptoms (ie, selection bias).Because the HRQOL of the control subjects is close to the US normal standard, selected completion, if it exists, should have minimal impact on the study results.

Figure 1 )
Figure 1) Age distribution of the sample and Calgary general population by sex

Figure 2 )Figure 3 )
Figure 2) Comparison of mean Medical Outcomes Study Short-Form 12-Item Health Survey, version 2 scores for people with and without irritable bowel syndrome (IBS).BP Bodily pain; GH General health; MCS Mental component score; MH Mental health; PCS Physical component score; PF Physical functioning; RE Role-emotional; RP Role-physical; SF Social functioning; VT Vitality/energy

Table 3 ,
the mean SF-12v2 scale scores ranged from 43.7 to 47.0 for people with IBS and from 49.0 to 52.6 for Can J Gastroenterol Vol 17 No 4 April 2003 260

TABLE 3 Comparison of mean Medical Outcomes Study Short-Form 12-Item Health Survey, version 2 scores for people with and without irritable bowel syndrome (IBS)
Difference of mean scores between people with and without IBS, without adjusting for demographic variables; † Difference of mean scores between people with and without IBS, adjusting for age, sex, education and income; ‡ 95% CI that does not overlap zero suggests a statistical significance or P<0.05 with certain health problems, with potentials of having reduced HRQOL.Unless specifically designed to compare IBS patients and other patients, using such controls will reduce the significance of difference in HRQOL (eg, at subscale level) between IBS patients and the controls.Our study, by using population-based IBS patients and controls, has overcome these potential problems.This study also has certain limitations.The response rate for this study (62% for the households and 74% for the individual household members) is suboptimal. *

TABLE 4 Comparison of mean Medical Outcomes Study Short-Form 12-Item Health Survey, version 2 scores for people with irritable bowel syndrome (IBS) who did and did not visit a physician
Difference of mean scores between IBS cases who did and did not visit a physician, without adjusting for demographics; † Difference of mean scores between IBS cases who did and did not visit a physician, adjusting for age, sex, education and income; ‡ 95% CI that does not overlap zero suggests a statistical significance or P<0.05 *