Michael P. Jones, M.D., Jason Bratten, B.S., and Laurie Keefer, Ph.D. Division of Gastroenterology, Northwestern University, Chicago, Illinois

Transcription

1 American Journal of Gastroenterology ISSN C 2007 by Am. Coll. of Gastroenterology doi: /j x Published by Blackwell Publishing Quality of Life in Patients With Inflammatory Bowel Disease and Irritable Bowel Syndrome Differs Between Subjects Recruited from Clinic or the Internet Michael P. Jones, M.D., Jason Bratten, B.S., and Laurie Keefer, Ph.D. Division of Gastroenterology, Northwestern University, Chicago, Illinois INTRODUCTION: The Internet is used increasingly to recruit patients for studies of gastroenterological disorders. The Internet allows access to potentially large study populations but no data exist comparing Internet-based populations with patients recruited from a clinical setting. We conducted an assessment of the quality of life (QOL) in patients with active inflammatory bowel disease (IBD) and irritable bowel syndrome (IBS) recruited from a gastroenterology clinic and the Internet. METHODS: RESULTS: CONCLUSIONS: Patients were recruited from a university gastroenterology clinic and several condition-specific support group web sites. The diagnosis of IBD and IBS was confirmed for clinic patients while Internet respondents were self-identified. Subjects completed both general (SF-36) and condition specific (IBS-QOL) measures. Subjects with IBD included 47 recruited from clinic and 96 recruited from the Internet. The IBS group included 147 patients recruited from clinic and 84 recruited from the Internet. Both IBS and IBD clinic and Internet IBD groups did not differ by. IBD Internet respondents were more likely to be women (clinic: 28F/19M, online: 80F/16M, P = 0.002) while IBS Internet respondents were more likely to be male (123F/24M) than IBS clinic patients (62F/22M) (P = 0.09). Compared with patients seen in clinic, both IBS and IBD Internet respondents had significantly poorer QOL as measured by both SF-36 and IBS-QOL. The greatest decrements in QOL occurred on SF-36 scales for physical and emotional roles and social functioning. On the IBS-QOL, the greatest decrement was seen on scales for dysphoria and life interference. Significant differences were maintained after controlling for gender. For both IBS and IBD, Internet-respondents had significantly poorer QOL than subjects recruited from clinic. These data demonstrate that subjects recruited from the Internet may represent a clinically distinct population and data obtained from online surveys may not generalize to broader clinical populations. Further study is needed to determine whether these differences reflect psychosocial characteristics of Internet responders or simply self-report behaviors in a relatively anonymous environment. (Am J Gastroenterol 2007;102: ) INTRODUCTION In 2004, Harris Interactive reported that 111 million adults accessed the Internet in search of health and medical information (1). While approximately half of these individuals are searching for information across many different sites, 23% went to sites that focus only on health-related topics. For individuals with certain conditions, the Internet serves as both a resource and a community (2). Internet-based support groups tend to arise for conditions that are chronic, disabling or poorly understood. These Internet-based support groups represent a concentration of individuals with a specific condition and such groups are a potentially fertile sampling ground for research recruitment. While recruitment through Internet-based support groups offers easy, low cost access to large study populations, concerns exist regarding the essential anonymity of subjects and potential sampling bias particularly related to gender, ethnicity and socioeconomic status (3 5). To date, the broadest use of the Internet as a research tool has been by communication scientists, sociologists, and psychologists. Clinical medicine, particularly oncology is beginning to make use of Internet-based studies. Almost no work has been done in gastroenterology. Soetikno et al. assessed QOL in 53 patients who had undergone surgery for ulcerative colitis (UC) recruited through the Internet compared with 47 similar patients recruited from a surgical clinic (6). The Internet-recruited patients were younger and had 2232

2 QOL in Internet-Recruited Patients 2233 significantly poorer QOL as measured by the SF-36. Similarly, Hilsden et al. noted that complementary and alternative medicine used by patients with inflammatory bowel disease (IBD) as assessed via an Internet survey differed from results obtained from a tertiary care clinic population (7). Given the paucity of data related to Internet-based recruitment for studies of common digestive disorders, we undertook a preliminary survey of QOL among patients with IBD and irritable bowel syndrome (IBS) recruited both through a gastroenterology clinic and via the Internet. Our hypothesis, based upon existing studies and clinical experience, was that Internet respondents with IBD and IBS would have poorer QOL than patients encountered in the clinic and would thus represent a distinct study population. METHODS Subjects Subjects were enrolled between 27 March 2006 and 27 November Adult patients with IBS and IBD were recruited from the gastroenterology clinic at Northwestern University. The majority of patients (>85%) are referred from primary care while the remainder are either referred from other gastroenterologists (10%) or are self-referred. Gastroenterologists in the clinic identified patients with IBS and IBD who were potential subjects. One of the study investigators attended clinic daily during the recruitment phase of the study. To this end, recruitment of potential subjects was consecutive but it was not possible to insure that all patients with IBS or IBD were identified. Internet subjects were recruited through advertisements placed on several popular gastroenterology websites. These included IBS Self Help and Support Group ( ibsgroup.org/), HealingWell ( IBD Sucks ( WebMD ( and Northwestern Center for Functional Gastrointestinal and Motility Disorders ( We excluded any Internet respondents receiving care from the gastroenterology clinic at Northwestern University. IBS patients met Rome II criteria (8). This was determined by a diagnostic questionnaire administered both in the clinic and online. Online respondents endorsed screening items that acknowledged at least two of the three cardinal Rome II criteria (abdominal pain relieved with defecation; and/or onset associated with a change in frequency of stool; and/or onset associated with a change in form/appearance of stool). If they did not acknowledge these items accordingly, the online application would not allow access to the study documents. Patients with IBS were not further categorized based upon symptom predominance (9, 10). The diagnosis of symptomatic IBD was confirmed for all the clinic patients either by interview or chart review. Participation by Internet subjects required them to acknowledge that they had symptomaticibd, but the type and extent were not ascertained in this preliminary study. The presence of active inflammation was not required or ascertained for this preliminary study. Patients with proctitis were excluded. Targeted enrollment was 150 subjects in both the online and clinic groups. This sample size would allow us to detect a 3-point difference in the component summary scores of the SF-36 with 80% power at α = Participation by patients and controls was voluntary and no compensation was provided. The protocol was approved by the Northwestern University Institutional Review Board. Measures Patients with IBS and IBD completed both the IBS-QOL and Medical Outcomes Study SF-36 (SF-36). The IBS-QOL was developed by Patrick et al. as a condition-specific instrument to assess QOL in patients with IBS (11, 12). This is a 34-item self-report measure with high internal consistency (Cronbach s alpha = 0.95) and high reproducibility (ICC = 0.86). IBS-QOL scores have been shown by convergent validity and analyses to be more closely related to psychological well being (0.45) than to function (0.36) (11). The measure has also been shown to be responsive to treatment (12). Higher scores reflect poorer QOL. The decision to use the IBS-QOL for patients with IBD was based on a previous study by our group showing the IBS-QOL to be highly correlated with the Inflammatory Bowel Disease Questionnaire (IBDQ) in patients with active IBD (r = 0.83, P < ) (13). The IBDQ is a 32-item measure with four dimensions: bowel function, emotional traits, systemic symptoms and social function (14). There is relatively good correlation between the IBDQ and a widely used measure of disease activity, the Crohn s Disease (CD) Activity Index (r = 0.67, P < 0.001) (15). General QOL was assessed using the Medical Outcomes Study SF-36 (SF-36) (16 18). The SF-36 is a well-studied, valid instrument that has been used to determine general QOL in a variety of clinical settings including functional gastrointestinal disorders and IBD. This measure assesses eight health concepts: (a) limitations in physical activities because of health problems; (b) limitations in social activities because of physical or emotional problems; (c) limitations in usual role activities because of physical health problems; (d) bodily pain; (e) general mental health (psychological distress and well-being); (f) limitations in usual role activities because of emotional problems; (g) vitality (energy and fatigue); and (h) general health perceptions. These scales can be grouped into mental and physical composite scores. Lower scores indicate poorer QOL. Data Analysis Data were expressed as mean ± SDEV. Normality of all data sets was determined using the Kolmogorov-Smirnov (KS) test. For normally distributed data, comparisons were made between groups using unpaired t-tests. For non-normally distributed or categorical data, comparisons across groups were made using the χ 2 statistic. Statistical significance was set at

3 2234 Jones et al. Table 1. Study Population Demographics Significant Differences Existed for Gender. The Online IBS Respondents were more Likely than the Clinic Sample to be Male, While the Reverse was True for the IBD Group (χ 2 = 9.64; P = 0.002). Data Expressed As Mean ± SDEV IBS IBD Online Clinic Online Clinic N Age 37 ± ± ± ± 13 Gender 62F/22M 123F/24M 80F/16M 28F/19M P < Statistical calculations were made using GraphPad Prism version 4.00 for Windows (GraphPad Software, San Diego, CA, RESULTS Three hundred seventy-four subjects were evaluated (Table 1). Two hundred thirty-one subjects had IBS and 143 subjects had IBD. For subjects with IBS, 147 were recruited from the clinic and 84 were online respondents. These groups did not differ with respect to age. The clinic population was predominantly female while the online population contained a greater proportion of male respondents. The IBD population contained 47 clinic patients and 96 online respondents. These groups did not differ with respect to age. The groups did differ significantly with respect to gender. In contrast to the IBS population, the IBD group was characterized by fewer men in the online sample compared with clinic respondents Gender differences were significantly different across groups and location (χ 2 = 9.64, P = 0.002). Compared with the clinic population, online respondents in both the IBS and IBD groups had significantly lower scores for QOL as measured with both the SF-36 and IBS-QOL (Fig. 1). Patients with IBS seen in clinic had significantly higher scores for both the SF-36 Physical Component (Clinic: 45 ± 12, Online: 40 ± 12, P = 0.004) and Mental Component Summaries (Clinic: 35 ± 9, Online: 31 ± 10, P = ) than did online IBS respondents. Patients with IBD seen in the clinic also had higher scores for both the SF-36 Physical Component (Clinic: 51 ± 13, Online: 39 ± 12, P < ) and Mental Component Summaries (Clinic: 32 ± 8, Online: 28 ± 8, P = 0.004) than did online IBS respondents. Similarly, online respondents with IBS had higher condition specific scores (indicating poorer QOL) on the IBS-QOL than did clinic patients (Clinic: 79 ± 26, Online: 100 ± 31, P < ). The same effect was seen for patients with IBD (Clinic: 72 ± 29, Online: 96 ± 31, P < ). Analysis of variance including gender, location and condition as factors failed to demonstrate a significant effect for gender on the results. Similarly, interactions between condition, location and gender were not detected. Scores for subscales of the SF-36 are shown in Figure 2. Compared with clinic patients with IBS, online IBS respondents had significantly lower scores for emotional role (clinic: 70 ± 39, online: 44 ± 41, P < ), physical role (clinic: 58 ± 25, online: 35 ± 39, P < ), social function (clinic: 64 ± 26, online: 48 ± 25, P < ), general health (clinic: 55 ± 20, online: 45 ± 23, P = 0.001), and mental health (clinic: 64 ± 19, online: 58 ± 19, P = 0.03). Online IBD respondents had significantly lower scores on all SF-36 scales except general health. Similar to patients with IBS, the greatest differences were seen on scales for physical role (clinic: 64 ± 40, online: 31 ± 37, P < ), emotional Figure 1. General and condition specific quality of life (QOL) scores for clinic and online respondents with IBD and IBS. Online (OL) respondents had significantly poorer QOL than clinic patients (CL) on all measures. Note that for the IBS-QOL, poorer QOL is reflected by higher scores while for the SF-36, poorer QOL is reflected by lower scores. PCS = physical component summary; MCS = mental component summary. Data expressed as mean ± SEM. Comparisons between clinic and online IBS and IBD groups were made using unpaired t-tests.

4 QOL in Internet-Recruited Patients 2235 Figure 2. SF-36 scale scores for clinic and online respondents with IBS and IBD. The greatest differences between online and clinic subjects with both IBS and IBD were seen for physical role (RP), emotional role (RE) and social functioning (SF). Differences on scales for vitality (VT) and general health (GH) were much less pronounced and did not consistently reach statistical significance between groups. PF = physical function; BP = bodily pain; MH = mental health. Comparisons between clinic and online IBS and IBD groups were made using unpaired t-tests. role (clinic: 75 ± 36, online: 43 ± 45, P < ), and Social Function (clinic: 75 ± 26, online: 53 ± 26, P < ) as well as Bodily Pain (clinic: 75 ± 22, online: 51 ± 24, P < ). Significant differences also existed between clinic and online groups with respect the scale scores on the IBS-QOL (Fig. 3). Compared with patients seen in the clinic, online respondents with both IBS and IBD had significantly greater scores on all scales (indicating poorer functioning) with the exception of body image for patients with IBS. DISCUSSION This study demonstrates that online respondents with either IBD or IBS had significantly poorer QOL compared with patients seen in a university gastroenterology clinic for the same conditions. This decrement in QOL among online respondents was not explained by condition, age or gender. Our findings support the results of a smaller, decade-old study demonstrating that patients with UC recruited through the Internet had poorer QOL than those recruited through the clinic (6, 19). Internet access and usage has evolved tremendously over the past ten years, and our findings have important Figure 3. IBS-QOL scale scores for clinic and online respondents with IBS and IBD. For online respondents with both IBS and IBD, the greatest differences occurred on scales measuring dysphoria and life interference. Note that in contrast to the SF-36, higher scores on the IBS-QOL indicate poorer QOL. Comparisons between online and clinic populations for IBS and IBD groups were made using unpaired t-tests. implications for subject recruitment, which increasingly uses the Internet as a recruitment tool (3 5). Importantly, the observed differences between clinic and online respondents appear greatest for certain QOL domains. Specifically, online respondents, whether having IBS or IBD, had the greatest decrements in quality life on scales reflecting dysphoria, life interference and physical and emotional roles. There are several possible explanations for this. First, the online respondents may be a more ill population. One of the limitations of this study is that we did not include either a direct measure of symptom severity or disease activity. QOL in patients with IBS has been shown to reflect gastrointestinal symptom burden (20) and we have shown that, in patients with IBD, scores on the IBS-QOL correlate highly with the IBDQ, which is well correlated with the CD Activity Index (r = 0.67, P < 0.001) (14, 15). These observations, coupled with the fact that clinic and online respondents did not differ greatly on measures of vitality and general health suggest that the groups did not differ greatly with respect to symptom severity. It seems more likely that the differences between online and clinic respondents with respect to overall QOL and scale

5 2236 Jones et al. scores for dysphoria, interference and impact on physical and emotional roles may reflect characteristics of the online respondent population. Quality of life in patients with IBS has been shown to be significantly predicted by extraintestinal symptom burden as well as a variety of psychosocial factors including patient-perceived severity (21, 22). Additionally, previous studies of Internet use have suggested that frequent Internet users may be more anxious and lonelier than those with less Internet use (23 26). It has also been shown that while online, individuals have a greater tendency to selfdisclose than when in person (27). Additionally, individuals display lower levels of social anxiety and social desirability when responding anonymously via the Internet than when using paper-based reporting methods (28). While we were able to exclude Internet respondents who received care through our clinic, many of these individuals were likely receiving care at other facilities. It is entirely possible that a patient with IBS or IBD will provide one set of responses regarding QOL while in clinic and another set of responses while responding anonymously online. This study has several weaknesses. While we were able to verify the diagnosis of IBS and IBD for clinic patients, we relied upon patient self-report for online respondents. Generally speaking, a major limitation of online recruiting and data collection is that data is less verifiable and potentially more variable. We also did not further characterize patients with IBD with respect to CD or UC. Previous literature has suggested that patients with CD report poorer QOL than their UC counterparts (Drossman, 1994) but this seems to be mediated in part by symptom severity (29, 30). An additional limitation is that we did not obtain socioeconomic or educational data from respondents and it is possible that Internet respondents differed from clinic patients in this regard. While previous reports suggested that Internet access was greatest in white populations and higher socioeconomic status (31), these disparities appear to be rapidly changing. Presently, 77% of all Americans aged 12 or greater have Internet access (32). This is further supported by a recent primary care study evaluating an urban, largely Hispanic population of lower socioeconomic status (33). Seventy-seven percent of respondents had accessed the Internet at least once and 79% used the Internet to find health-related information with 73% using online information to make a healthrelated decision. While future studies need to address these variables, Internet access has become widespread and our belief is that we are querying similar populations who respond differently based upon the venue in which they are approached. In summary, this study demonstrates that patients with IBS and IBD will report significantly poorer QOL when responding online than when seen in a university clinic. Further study is warranted to better understand the characteristics of the online respondent population. Specifically, it is important to determine whether this population is psychosocially distinct or whether patients simply disclose more freely when in an anonymous Internet environment. In the meantime, investigators need to be aware that responses gathered via the Internet (and perhaps even patients recruited via the Internet) may not be representative of the general clinical population. ACKNOWLEDGMENT The authors are grateful to Ms. Tiffany Havlicek for creating the online Internet application used in this study. STUDY HIGHLIGHTS What Is Current Knowledge The Internet has become an increasingly important tool in health education, clinical care and research. Patient access and use of the Internet has tremendously broadened over the past decade. It is currently unknown whether study subjects recruited via the Internet are comparable to patients recruited from a clinical setting. What Is New Here For both patients with irritable bowel syndrome (IBS) and inflammatory bowel disease (IBD), subjects recruited via the Internet had poorer quality of life (QOL) than subjects recruited from a clinical setting. The observed differences between clinic and online respondents were greatest for QOL dimensions reflecting dysphoria, life interference and impairment of physical and emotional roles. Further study is needed to determine the role of both socioeconomic and psychosocial factors as well as response characteristics in a clinical setting compared with the relatively anonymous Internet environment. Reprint requests and correspondence: Michael P. Jones, M.D., 251 East Huron St, Galter Pavilion 4-104, Chicago, IL Received January 3, 2007; accepted May 23, REFERENCES 1. Greenspan R. Net attracts health-seeking surfers: ClickZ Network; page= Accessed: July 2, Ziebland S, Chapple A, Dumelow C, et al. How the internet affects patients experience of cancer: A qualitative study. BMJ 2004;328: Im EO, Chee W, Tsai HM, et al. Internet communities for recruitment of cancer patients into an internet survey: A discussion paper. Int J Nurs Stud [Epub ahead of print]. 4. Lenert L, Skoczen S. The Internet as a research tool: Worth the price of admission? Ann Behav Med 2002;24: Eysenbach G, Wyatt J. Using the Internet for surveys and health research. J Med Internet Res 2002;4:E Soetikno RM, Mrad R, Pao V, et al. Quality-of-life research on the Internet: Feasibility and potential biases in patients with ulcerative colitis. J Am Med Inform Assoc 1997;4:

6 QOL in Internet-Recruited Patients Hilsden RJ, Meddings JB, Verhoef MJ. Complementary and alternative medicine use by patients with inflammatory bowel disease: An Internet survey. Can J Gastroenterol 1999;13: Thompson W, Longstreth G, Drossman D, et al. Functional bowel disorders and functional abdominal pain. In: Drossman D, Talley N, Thompson W, Whitehead W, Corazziari E, eds. Rome II : Functional gastrointestinal disorders: Diagnosis, pathophysiology, and treatment, 2nd Ed. McLean, VA: Degnon Associates, Inc.; 2000: Walter SA, Skagerstrom E, Bodemar G. Subgroups of irritable bowel syndrome: A new approach. Eur J Gastroenterol Hepatol 2004;16: Mearin F, Baro E, Roset M, et al. Clinical patterns over time in irritable bowel syndrome: Symptom instability and severity variability. Am J Gastroenterol 2004;99: Patrick DL, Drossman DA, Frederick IO, et al. Quality of life in persons with irritable bowel syndrome: Development and validation of a new measure. Dig Dis Sci 1998;43: Drossman DA, Patrick DL, Whitehead WE, et al. Further validation of the IBS-QOL: A disease-specific quality-oflife questionnaire. Am J Gastroenterol 2000;95: Jones MP, Wessinger S, Crowell MD. Coping strategies and interpersonal support in patients with irritable bowel syndrome and inflammatory bowel disease. Clin Gastroenterol Hepatol 2006;4: Guyatt G, Mitchell A, Irvine EJ, et al. A new measure of health status for clinical trials in inflammatory bowel disease. Gastroenterology 1989;96: Irvine EJ, Feagan B, Rochon J, et al. Quality of life: A valid and reliable measure of therapeutic efficacy in the treatment of inflammatory bowel disease. Canadian Crohn s Relapse Prevention Trial Study Group. Gastroenterology 1994;106: Ware JE, Jr, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care 1992;30: McHorney CA, Ware JE Jr, Lu JF, et al. The MOS 36- item Short-Form Health Survey (SF-36): III. Tests of data quality, scaling assumptions, and reliability across diverse patient groups. Med Care 1994;32: McHorney CA, Ware JE Jr, Raczek Ae. The MOS 36-Item Short-Form Health Survey (SF-36): II. Psychometric and clinical tests of validity in measuring physical and mental health constructs. Med Care 1993;31: Soetikno RM, Provenzale D, Lenert LA. Studying ulcerative colitis over the World Wide Web. Am J Gastroenterol 1997;92: Lackner JM, Gudleski GD, Zack MM, et al. Measuring health-related quality of life in patients with irritable bowel syndrome: Can less be more? Psychosom Med 2006;68: Hahn BA, Kirchdoerfer LJ, Fullerton S, et al. Patientperceived severity of irritable bowel syndrome in relation to symptoms, health resource utilization and quality of life. Aliment Pharmacol Ther 1997;11: Spiegel BM, Gralnek IM, Bolus R, et al. Clinical determinants of health-related quality of life in patients with irritable bowel syndrome. Arch Intern Med 2004;164: Black DW, Belsare G, Schlosser S. Clinical features, psychiatric comorbidity, and health-related quality of life in persons reporting compulsive computer use behavior. J Clin Psychiatry 1999;60: Carden R, Rettew S. Internet chat room use, satisfaction with life, and loneliness. Psychol Rep 2006;98: Lavin MJ, Yuen CN, Weinman M, et al. Internet dependence in the collegiate population: The role of shyness. Cyberpsychol Behav 2004;7: Scealy M, Phillips JG, Stevenson R. Shyness and anxiety as predictors of patterns of Internet usage. Cyberpsychol Behav 2002;5: Suler J. The online disinhibition effect. Cyberpsychol Behav 2004;7: Joinson A. Social desirability, anonymity, and Internetbased questionnaires. Behav Res Methods Instrum Comput 1999;31: Casellas F, Alcala MJ, Prieto L, et al. Assessment of the influence of disease activity on the quality of life of patients with inflammatory bowel disease using a short questionnaire. Am J Gastroenterol 2004;99: Casellas F, Lopez-Vivancos J, Badia X, et al. Influence of inflammatory bowel disease on different dimensions of quality of life. Eur J Gastroenterol Hepatol 2001;13: National Telecommunications and Information Administration. Falling through the net: toward digital inclusion. US Department of Commerce; Available at: Accessed March 28, Center for the Digital Future. Online World As Important to Internet Users as Real World? USC-Annenberg Digital Future Project Finds Major Shifts in Social Communication and Personal Connections on the Internet. Available at: report.asp?int- GlobalId = 19; Accessed March 28, Liszka HA, Steyer TE, Hueston WJ. Virtual medical care: How are our patients using online health information? J Community Health 2006;31: CONFLICT OF INTEREST Guarantor of the article: Michael P. Jones, M.D. Specific author contributions: Michael P. Jones: Study design, data analysis, and manuscript preparation. Jason Bratten: Recruitment, database management, data analysis. Laurie Keefer: Data analysis, manuscript preparation. Financial support: None. Potential competing interests: None.