Assessment of health-related quality of life (HRQOL) is important for measuring the impact of chronic disease (1,2). It provides a clearer picture of health than do disease parameters alone and can be used to identify patient needs and to assess efficacy of treatment, as perceived by the patient. It is determined not only by physical well-being but also by psychological state, degree of social support, effects of treatment and complications (2). Assessment of HRQOL provides a global measure of the patient's perceptions, illness experience, and functional status (3). Assessment should encompass issues directly related to physical disease (pain, symptoms, complications) and qualities that are independent of the disease but may be affected by it (emotional state, personality).
There has been increasing emphasis on the assessment of HRQOL in patients with chronic illness, using generic or disease-specific measures. Disease-specific questionnaires are multidimensional instruments generated from patient-reported problems. Disease-specific instruments have the advantage of including only aspects of HRQOL relevant to the patient group and are more likely to be sensitive to detecting important change. The Inflammatory Bowel Disease Questionnaire (IBDQ) was developed through a rigorous process beginning with item generation using interviews with adult patients and clinician surveys (4). The IBDQ has been validated as an assessment tool that reflects the health status of adult patients and is itself a robust measure of therapeutic efficacy (5).
Inflammatory bowel disease manifests during childhood or adolescence in 20% to 25% of patients. Chronic and often unpredictable gastrointestinal symptoms and required treatments impose psychological and social stresses on young patients that may be different from those experienced by adults. Unique to pediatric patients is the potential for growth impairment and delay in pubertal development as a complication of the inflammatory process in Crohn's disease (CD) particularly (6-11). Teenagers affected by stunting and delayed physical maturation are at particular risk for difficulties with peer interactions and psychosocial trauma. Therefore, although many issues may be shared with adult patients, the effects of IBD on quality of life of children and teenagers should also be systematically explored.
Measurement of HRQOL in children and adolescents is more difficult than in adults because physical, intellectual, and emotional function are constantly changing with normal development. If an instrument is to be used among young patients of disparate age, the variation in disease-related quality-of-life concerns with age must first be explored. This article details the item-generation and item-reduction phases of our efforts to develop a multi-item measure to assess IBD-related quality of life of children and adolescents with CD and ulcerative colitis (UC).
Purpose of the Measure
The instrument is intended to be descriptive and evaluative. As a descriptive tool, it is hoped that a disease-specific quality-of-life measure may facilitate recognition in individual patients of disparity between apparent IBD activity and severity, organic disease-related phenomena, which the physician is accustomed to assessing, and emotional or functional disability. As an evaluative tool it will be incorporated as an outcome measure in clinical trials.
Population and Setting
The study was conducted through the IBD Program at the Hospital for Sick Children, Toronto, which provides follow-up care for children and adolescents with UC and CD. From January 1980 to May 1997, 952 patients (504 boys, 448 girls) were diagnosed with IBD (60.7% CD, 37.9% UC, 0.5% indeterminate colitis) and prospectively registered in a pediatric IBD database. Until very recently the practice of pediatric gastroenterology in the metropolitan Toronto area has been confined to The Hospital for Sick Children, and general pediatricians have not independently treated children with IBD. Therefore, patients registered in the database approximate a population-based cohort of early-onset IBD, encompassing a broad spectrum of disease severity (6). The mean age at diagnosis of CD was 12.7 ± 2.5 years and of UC 10.9 ± 3.9 years. All patients are transferred to the care of an adult gastroenterologist before their 18th birthday. Of the total 340 patients currently under the care of the pediatric IBD program, 38 (10 CD, 25 UC, 3 indeterminate colitis; 9%) are less than 10 years of age, and 15 (3 CD, 10 UC, 3 indeterminate colitis; 4.4%) are less than 8 years of age.
Children and adolescents with CD or UC diagnosed for at least 6 months were eligible to take part in the item-generation or item-reduction phases. The patients recruited represented consecutive patients attending the IBD clinic for follow-up and others selected from a listing of patients residing within the metropolitan Toronto area and therefore feasibly interviewed at the time of a home visit. Patients aged less than 8 years were excluded because of concern that systematic exploration of quality of life among very young children would require significantly modified methods.
The study was was approved by The Hospital for Sick Children Human Subjects Review Board. Patients and their parents gave written, informed consent.
To identify all factors that affect young patients' perception of quality of life and therefore potentially merit inclusion in a multi-item measure, interviews were conducted with individual patients. In calculating the sample size of patients to be interviewed, it is customary in the process of item generation to continue until it becomes apparent that very few new items are being gleaned. On average, this requires 75 to 100 interviews.
In the private interview with the IBD program nurse specialist (CS), research assistant (MM), or social worker (DW), the children and adolescents were asked to elaborate all the ways Crohn's disease or ulcerative colitis affected them and to respond to other open-ended questions. Patients were informed of the purpose of the research and consented to the interview in advance, so that they had time to reflect on their concerns. Interviews were conducted either in the clinic at the time of a routine visit, or in the patient's home. Parents were not present during the interview. Interviews were taped for later review and transcription.
Twenty-five pediatric gastroenterologists, who are considered experts in the field of IBD, were sent copies of the adult-generated IBDQ (4). They were asked to label the IBDQ questions as appropriate or inappropriate for children aged 8 to 12 years and separately for adolescents aged 13 to 18 years. Addition of other issues important to the quality of life of pediatric patients was also requested.
The transcripts of the interviews were reviewed by the investigators and consensus reached about which items were appropriate to combine and which represented distinct issues. The IBDQ and results of the survey of pediatric IBD practitioners were reviewed for additional items.
An item-reduction questionnaire incorporating the issues and concerns raised was constructed. The item-reduction questionnaire asked the patient to indicate on a 7-cm visual analog scale how important each item was for him or her (with "not important at all" at one end of the visual analog scale and "very important" at the other), and how often he or she worried or thought about that item (with "never" at one end of the visual analog scale and "always" at the other). "Importance" and "frequency" scores ranging from minimum 0 to maximum 7 were measured for each item. This item-reduction questionnaire was used to establish which concerns are most important to young patients and therefore merit retention in a pediatric IBD quality-of-life index.
Demographic information (type of IBD, duration of diagnosis, age, gender) was recorded for patients completing the item-reduction questionnaire. Each of the items in the questionnaire was classified by consensus of the investigators as pertaining to the domain of bowel symptoms, systemic symptoms, social impairment, functional impairment, treatment or investigation issues, body image concerns, or emotions.
Analysis of Data
Mean and median importance and frequency scores, and importance-plus-frequency scores were calculated for the entire patient group, and for patients in subgroups according to age (≤12 years, ≥13 years) and IBD type (CD, UC). Issues were ranked according to the order (greatest to least) of importance-plus-frequency scores for the entire group and within subgroups.
Eighty-two children and adolescents (36 girls, 45 boys) with IBD (61 CD, 20 UC) participated in the interviewing process between May 1996 and September 1996. Fifty-six patients were aged 13 years or more; 24 children were aged 12 years or less. After combining similar statements by consensus, 94 distinct patient-generated concerns remained. These encompassed the domains of bowel symptoms (14 items), systemic symptoms (4 items), emotions (37 items), social impairment (18 items), functional impairment (10 items), treatment or investigation issues (11 items), and body image concerns (4 items). (Four issues were assessed as pertaining equally to two domains. For example, "having to miss out on activities because of IBD" can be construed as either social or functional impairment.) Of the 32 issues contained in the adult IBDQ and others suggested by the survey of pediatric gastroenterologists, all but 3 were also raised by a child or adolescent during the interviewing process. The IBDQ concerns related to sleep disturbances (difficulty falling asleep and nighttime awakening) and limitation in sexual activity because of IBD were not elicited from the pediatric patients.
The item-reduction questionnaire included the 94 patient-generated items, plus the two adult IBDQ issues related to difficulty sleeping. A question specifically about limitation of sexual activity by IBD was not added because the consensus of pediatric gastroenterologists was that it was inappropriate for a pediatric questionnaire. Adolescents volunteered during the interviews that IBD interfered with dating; this concern was included in the item-reduction questionnaire.
The 96-item-reduction questionnaire was administered to 117 children and adolescents between April and August 1997. Patient characteristics are summarized in Table 1. Seventeen patients had also participated in the item-generating interviews 10 to 15 months previously.
The issues of greatest importance to children and adolescents with IBD are grouped within domains and listed in Table 2. Issues are ranked according to the magnitude of the mean frequency-plus-impact score incorporating responses of the entire group, recognizing that children and teenagers with CD were overrepresented in the sample of patients interviewed.
As is evident in Table 3, adolescents with CD shared many concerns with adolescents with UC, but other issues were much more troubling to one patient group than to the other. In Table 3 the designation of issues as "shared" or "disparate" is based on the assumption that the final questionnaire will contain roughly 30 items, to mirror the adult IBDQ. Therefore an issue ranking in the top third of concerns expressed by teenagers with UC and CD is considered shared. The teenagers with CD had greater concerns about height and appearance (body image domain) and about fatigue and low energy (systemic symptoms). Patients with UC were more affected by bowel symptoms, specifically diarrhea, gas, urgency, and bleeding. Concerns about having a flare-up, the lifelong nature of IBD, treatments and investigations, the effects of IBD on the family, and worries about future health problems were shared. Patients with UC expressed a broader range of emotions (anger, frustration, embarrassment about their bowel conditions), and indicated that they experienced them more intensely. Crohn's disease appeared to interfere with activities in more ways, affecting, for example, school attendance and performance in sports.
To assess the variation in the impact of IBD with age, responses of patients with CD and those with UC to the item-reduction questionnaire were analyzed within two age categories. The majority of issues impacting most on the lives of adolescents with CD were also ranked of greatest concern by the younger patients. The exceptions are listed in Table 4. (As previously, the designation disparate is applied when an issue was ranked highly enough to merit inclusion in a 30-35 item questionnaire for one patient subgroup but not the other.) Concerns about height, weight, appearance, giving up or missing out on activities, taking medications, worries about having a flare-up and about future health, low energy, concern about effects on the family, feelings of unfairness and frustration were all shared by young and older patients. As shown in Table 4, treatment issues (hospital, dietary restrictions) were more a focus of younger children.
The cohort of young patients with UC completing the item-reduction questionnaire was too small to make similar comparisons within that disease type. Furthermore, it was observed that young patients with UC in complete clinical remission for many months had difficulty recalling what had bothered them when their colitis was active.
In this study, we systematically explored the concerns of pediatric patients with IBD. Although the children and adolescents participating in the item-generation and item-reduction phases were not randomly selected from a truly population-based sample, they were drawn from a large pediatric IBD cohort that has been well characterized and which, because of referral patterns, encompasses the spectrum of disease severity (6). In the next stage of instrument development (item selection), the issues ranking of greatest importance to young patients will be formatted into a questionnaire. Overlapping and closely related concerns will be combined into a single question in the item-selection phase.
The process of developing a pediatric IBD disease-specific measure of quality of life has hitherto mirrored the methods employed by Guyatt et al. (4) in developing the adult IBDQ. In that adult study, 150 items generated largely through patient interviews were reduced to the 32 items most frequently cited and importantly ranked by a second cohort of 100 patients. The final IBDQ includes these 32 questions encompassing four domains: bowel symptoms (10 questions), systemic symptoms (5 questions), emotional function (12 questions), and social function (5 questions). The response options for each question are framed as a seven-point scale in which 7 represents best function and 1 represents worst function. Scores range from 32 to 224; higher scores mean better quality of life.
There is merit to keeping the format and scale of a new pediatric index comparable, so that the significance of an individual numerical score can be more readily appreciated. A recent review of the interpretation and conduct of randomized controlled trials in IBD argued for consensus on a standardized group of outcome measures to facilitate between-study comparisons and meta-analysis (12).
The results of the item-reduction questionnaire suggest that CD and UC affect the lives of young patients differently. Differences between patients categorized by IBD type were more frequent and more extreme than between patients in subgroups according to age but with the same diagnosis. We did not attempt, however, to study the impact of IBD on very young patients, who comprise a unique but rare subset of our entire IBD population. Clearly, issues of major concern to all subgroups of older children and adolescents surveyed in the item-reduction phase will be retained in the pediatric IBDQ. Disparate concerns could be handled by means of a modular approach, whereby a core of shared concerns is supplemented by questions in the appropriate age and disease module. For example, the colitis module would pose questions about the specific bowel symptoms so important to patients with UC, but having a comparatively limited effect on the quality of life of most patients with CD. The adolescent age module would include issues of concern to the older pediatric patients but not relevant or ranked as important to younger children. A modular questionnaire would have greater content validity. By including fewer irrelevant issues for any individual patient, it would more likely capture the impact of disease on his or her quality of life.
An alternate approach involves the construction of a single questionnaire for patients with UC and those with CD in late childhood and adolescence. The issues selected would include those of major concern to any one subgroup as well as shared concerns. Each subgroup of patients would have to be represented by comparable numbers of questions relevant to them. The lowest likely total scores (worst quality of life) would therefore be higher than with the modular approach because of inclusion of more issues irrelevant to individual patients.
In the subsequent process of instrument development, a questionnaire based on each of these two models will be formatted and pretested with respect to wording and ease of administration. There is no precedent for the modular approach, but it may offer a means of addressing the variation in issues of concern with age and disease type and thereby facilitate assessment of disease-specific quality of life in pediatric chronic illness.
Acknowledgements: This study was supported by a grant from The Hospital for Sick Children Foundation New Initiatives Fund.