Introduction

Coeliac disease (CD) is a chronic immune-mediated disease in genetically susceptible individuals induced by ingested gluten causing an inflammatory response affecting the intestines and several organ systems with onset at all ages [1,2,3]. CD is a common food intolerance affecting 1% of the general population [4, 5], however with a prevalence of up to 3% in select populations such as Sweden [6]. The only available treatment of CD is a lifelong gluten-free diet (GFD) [2].

When following a GFD, most young patients recover and feel no physical impact of CD [7]. However, following a strict GFD diet may be demanding [8]. Children and adolescents with CD report that dietary restrictions have consequences for their emotional and social well-being [9,10,11,12,13,14,15,16,17].

Health-related quality of life (HRQoL) is the product of physical, psychological, and social well-being and perception of position in life compared to others [18,19,20]. HRQoL can be assessed with questionnaires that address physical, cognitive, behavioral, emotional, and social concerns [21]. Previous studies have provided conflicting results of HRQoL in children with CD [22]. Children’s HRQoL has been found to be similar to the healthy reference populations in some studies [22,23,24,25], whereas other studies have found their HRQoL to be lower [26,27,28]. Children who follow a GFD appear to have better HRQoL [17] indicating that HRQoL may be related to how well an individual adapts to having CD and following a GFD [29]. Studies among adults have found HRQoL differences in relation to gender [30,31,32]. We therefore hypothesized that the present study would find differences in HRQoL in relation to compliance and also in relation to age groups and gender.

Comprehensive, valid, and reliable disease-specific questionnaires are necessary to sensitively measure the impact of a condition on HRQoL [21, 22]. While some HRQoL instruments require different versions for age groups, other instruments cover children, adolescents, and adults in one version [33] thus allowing comparative research on differences and changes directly between age groups on the same items.

CD-specific QoL questionnaires have previously been developed [28, 29, 34, 35]. The 12-item CDDUX [28] and the 13- to 17-item CDPQOL [34] have been developed for children and adolescents and the 20-item CD-QOL for adults [35]. None of these questionnaires were available in Danish. Also, there was no questionnaire available that measures HRQoL across ages in persons with CD.

A CD-QoL questionnaire should be as short as possible, easily understandable, and capable of being self-administered by patients. The content of the questionnaire should be comprehensive in evaluating the impact of CD on a person’s life. Living with CD has been shown to give rise to different experiences: some felt burdened when receiving the CD diagnosis, while others felt relieved; some felt limited in one’s life due to managing dietary restrictions, while some felt competence in managing a GFD; some experienced difficulties in following a GFD, while others found it easy; some thought about CD constantly, while others only thought about it when confronted with gluten-containing food [36]. The scope of these experiences and concerns should be captured in a CD-QoL questionnaire. The aim of the present study was to develop the Coeliac Disease Quality of Life (CDQL) questionnaire—a comprehensive coeliac disease-specific HRQoL questionnaire for children, adolescents, and adults that can be completed directly or by proxy (i.e., parent or carer). The CDQL should be able to track and monitor the HRQoL on the same items and domains as the patients grow from children into adolescents and become adults.

Methods

Questionnaire design

The questionnaire was developed in three phases in accordance with best practice in the development of patient-reported outcome measures [37]. In the first of the three phases, focus group interviews were used to explore the impact on HRQoL of living with CD [36, 38, 39]. Each focus group lasted about 1 h. The focus groups included a total of 26 participants (15 children (nine girls, six boys), 8 adolescents (seven girls, one boy), and 3 adults (one woman and two men)). All participating children and adolescents had biopsy-proven CD with durations between 1 and 15 years. After seven focus group sessions the qualitative data were ‘saturated,’ meaning that no new themes were brought forward by participants in the last sessions [40, 41]. The qualitative data were analyzed using grounded theory [40] identifying patient concerns in living with CD [36]. From the patient concerns, a pool of questionnaire items was generated for the pilot version of the CDQL. In phase 2 the pilot-CDQL was administered to patients, followed by cognitive interviews [42] with children, adolescent, and adult respondents to ensure that all items were understood in the way they were intended and to amend or revise any items if necessary. In phase 3, final data collection was carried out and data were analyzed.

The questionnaire consists of three sections. The first section collects background information that is important for profiling and between subjects analysis: gender, age, area, CD duration, GFD-compliance, a measure of worries about CD-symptoms, and a measure of perceived CD-symptoms severity. Included also was a multiple-choice option to indicate if the respondent was assisted partly or if the questionnaire was answered completely on their behalf by, e.g., the mother, father, or another. The second section evaluates HRQoL using both generic and CD-specific items. The generic items used, one on ‘health’ and one on ‘quality of life,’ have been used previously in research in the WHOQOL-BREF both in English [20] and Danish [43]. The third section had an open question format so that respondents could comment freely.

Different types of response scales with 5 or 7 options and with or without text and/or smilies were presented to and evaluated by participants during the cognitive interviews. Response scales with smileys and 5 or 7 options were preferred because they were intuitively easy to interpret for all respondents, while allowing enough categories for discrimination [28, 44]. CD affected some differently than others, and the scale had to be able measure this. The scope of the scale therefore reached from positive to negative impact reflecting the differences in statements from the focus groups, for example, “I think it’s very difficult sometimes…” and “I feel no different […] I feel healthy just like everyone else.” [36].The final response scale chosen for the all CD-specific items was a 5-option Likert scale in the form: When I think of [a CD-specific issue] I feel … ‘Very unwell,’ ‘Unwell,’ ‘Neutral,’ ‘Well,’ or ‘Very well’ with each option supported by a smiley expressing the corresponding emotion. The CDQL pilot version was presented in both English and Danish, with 65 items including 17 background items, 2 generic HRQoL items, and 46 CD-specific QoL items. The CDQL was then reduced to 44 items: 12 background items, 2 generic HRQoL items, and 30 CD-specific QoL items. All items were formulated in the same direction; a higher score indicates better HRQoL.

Data collection

The pilot and the final data collection versions of the CDQL were programmed into a web-based questionnaire (WBQ). WBQs have been shown to have as good, if not better, psychometric properties than paper-and-pencil questionnaires [45,46,47]. WBQ is able to collect data from large populations, as well as from targeted, disease-specific populations [43,44,45, 48] with high levels of completion [49, 50]. Children aged 8–12 [51] and adolescents aged 13–17 [52] are able to use WBQ to self-report HRQoL. The pilot version of the questionnaire required 15–20 min to complete. The WBQs were pre-viewed on various units (computer, tablet, and phone) and web browsers to ensure uniformity in layout. The pilot version collected data in Danish and English languages from December 2012 to June 2013, and the final version collected data from January to April 2015.

Participants for the pilot version were identified and recruited through Odense University Hospital patient register and the Danish Coeliac Society (DCS) homepage, newsletter, Facebook page, and printed journal. The call for participation provided a description of the study and access to the questionnaire via link or QR-code. Irish participants were recruited through the Coeliac Society of Ireland through newsletter and Facebook page. Participants for the final data collection were recruited in collaboration with the DCS in Denmark. Membership of the respective Coeliac Societies was not required for participation. In order to follow questionnaire development guidelines and validate the questionnaire in one language and one cultural contexts at the time, the final data collection took place in Denmark only [37].

Cognitive interviewing

Cognitive interviewing is a test method aimed at assessing respondents’ understanding of questionnaire items, in order to identify confusing items that may need rewording, to ensure that respondents interpreted the items as intended by the questionnaire developers, and to ensure that the items address all relevant issues [53]. The revised questionnaire was handed out to children, adolescents, and adults attending a gluten-free camp hosted by the DCS. Upon completing the questionnaires, participants were asked about their experience with answering the questionnaire and to point out any confusing items or missing concerns. Children, adolescents, and adults participated in separate discussions. To take timid participants into consideration, the option was given to leave comments on the printed questionnaire and to contact the moderator alone afterwards.

Psychometric analysis

Content validity was ensured by involving relevant parties—patients and clinicians—throughout the questionnaire development to ensure that all aspects of the CD-HRQoL construct were covered by the questionnaire [37]. The patient-generated items were reviewed by clinical experts on CD. The questionnaire commentary fields were carefully read and patients were interviewed upon completing the questionnaire. Statistical analysis was carried out using SPSS [54] and R [55].

The scale ‘Worries-about-symptoms’ was made from the twelve items addressing worries about CD-symptoms. The scale score was an average of the numeric values of the items. The scale ‘Symptoms’ was made from the twelve items addressing experienced severity of the same CD-symptoms. This scale’s score was also an average of the numeric values of the items. These two scales represented the perceived impact of CD-symptoms and we therefore hypothesized that scores on both scales would yield significant positive correlations with HRQoL-scores.

A factor analysis distributed the items with the highest item-total correlation on each scale. Factors were named as questionnaire subscales. Reliability to scale (Cronbach’s alpha) was computed for each scale. Reliability in deleting each item in turn was examined. Construct validity was assessed by correlating the scale scores with the generic health item and the generic QoL item of the WHOQOL-BREF [37]. Respondents were analyzed in two age groups (younger or older than 18 years) and also in three age groups (children 0–12 years, adolescents 13–17 years, or adults 18 years and older). Discriminant validity and sensitivity of the CDQL was examined by ANOVA.

Results

Pilot data collection

The online link to the pilot version of the CDQL was activated 1230 times and generated 475 complete questionnaires, whereof 196 where from Denmark and 270 were from Ireland. Only completed questionnaires were retained for analysis. The majority of the Danish respondents were younger than 18 years (n = 136), whereas the majority of the Irish respondents were above 18 years (n = 265). Of the respondents living in other countries (n = 9) one was under 18 years. CD-symptom-items that were identified as relevant by less than 5% of the respondents were omitted from the CDQL questionnaire, reducing the symptom list from 27 to 12 items. Participants’ statements in the commentary fields showed that patients felt that their concerns were covered by the items. The pilot version of the CDQL was found relevant and meaningful in two different populations as judged by respondents’ comments.

Cognitive interviewing

Seventy-six children and adolescents and six adults participated in cognitive interviewing in smaller groups moderated by a psychologist. The cognitive interviewing resulted in a rewording of the text introducing the CD-symptoms items, and a change from a 7-point to a 5-point Likert scale as participants found the 7-point scale cumbersome. Four CD-specific items were removed as participants found them confusing. No new HRQoL concerns were mentioned that were not already covered by the existing items. The cognitive interviewing ensured that items were relevant, comprehensive, understandable, and were interpreted as expected. The process supported the face and content validity of the CDQL. The pilot data collection and the cognitive interviewing demonstrated the feasibility of the questionnaire in different age groups, and reduced the number of items, and improved the understandability of the questionnaire as a whole.

Profile of the respondents

The final version of the CDQL returned 422 complete questionnaires of 510 initiated responses (17% dropout). See Table 1 for descriptive statistics of the respondents. As measured in the WBQ, the questionnaire took in average 10 min (651 s) to complete. The questionnaire was answered either solely by the person with CD (n = 380) or with the aid of another person (n = 42). In all these cases, the assisting person was the mother. Inspection of the respondents’ postal codes showed that the respondents represented all parts of the country.

Table 1 Characteristics of the 422 respondents with coeliac disease
Full size table

The modest sample size of male respondents (n = 43) in relation to female (n = 379) reflected the uneven distribution of CD between sexes. To examine whether the male sample size may have had limited statistical power in terms of the risk of type II errors, a post hoc power analysis was carried out revealing that on the basis of the mean between-groups comparison on the average score of the 30 CDQL items with the effect size (d = 0.20), a statistical power below the recommended 0.80 level [56] was present (1-β err prob. 0.23, critical t(420) = 1.97, α err prob. = 0.05) [57]. Given this small effect size, an n of approximately 394 in each group would be needed to obtain statistical power above the recommended 0.80 level on the average of all the CDQL items. The statistical power of the sample of respondents under 18 years (n = 77) in relation to adults (n = 345) was not satisfying either (d = 0.15, 1-β err prob. 0.21, critical t(420) = 1.97, α err prob. = 0.05). With results on the average of the 30 CDQL items being so close among the age groups, an n of 736 in each group would be needed to obtain statistical power above the recommended 0.80 to make sure that an actual difference between respondents under 18 and adults would not be missed. However, the intention of the present study was not so much to find differences on the average of all CDQL items, as it was to find differences on specific items, which succeeded.

The most prevalent comorbidities of the respondents were allergy (n = 36), asthma (n = 25), lactose intolerance (n = 21), metabolic disorders (n = 17), diabetes (n = 13), depression (n = 9), eczema (n = 5), anxiety (n = 5), and osteoporosis (n = 4). The most prevalent symptoms reported by the participants in case of gluten intake were abdominal pain, feeling bloated, and feeling tired. The same three symptoms were the ones participants had felt most often the past two weeks. While a large proportion knew one other person with CD (n = 78, 18%), two other persons (n = 71, 17%), or more (n = 133, 32%), there was a large group that did not know another person with CD (n = 140, 33%). There was no significant gender difference. In average, adolescents knew three people with CD; children knew two; while adults knew between two and three other people. However, this distribution was not statistically significant. More children and adolescents (78%) were members of the Coeliac Society than adults (59%) χ 2(2, N = 422) = 10.23, p < 0.01. The majority of the respondents had never participated in a CD event (n = 246, 58%), such as baking course, camp, meeting, or congress. More children and adolescents (53%) had participated in a CD event than the adults (39%) χ 2(1, N = 422) = 5.16, p = 0.02. The majority of respondents reported no dietary transgression in the past 14 days (n = 351, 83%). Eating gluten was reported on one occasion by 9% (n = 37), on two occasions by 4% (n = 15), and on three or more occasions by 5% (n = 19) of the respondents. There was no significant difference between female and male respondents’ dietary transgressions, nor between adults, adolescents, and children.

Development of the CDQL scales

A correlation matrix of the 30 CD-specific questionnaire items was inspected showing correlation-values and item-clusters in support of undertaking a factor analysis. Clustered items were explained as related to latent factors. Selecting the number of factors was based on the interpretation of the scree plot that each scale would have at least three items, that items preferably loaded >0.4 on only one factor with few cross-loadings, and that clustered items should be theoretically coherently covering a common distinctive construct (Table 2). Eight components with eigenvalues of 1 and higher were identified. The eight components cumulatively explained 67% of the variance (33.52, 6.67, 5.70, 5.54, 4.58, 4.29, 3.46, and 3.14%, respectively). Inspection of the scree plot showed no clear breaks to identify the number of subscales. Therefore, eight components were retained for exploratory factor analysis with solutions for 2, 3, 4, 5, 6, 7, and 8 factors using Varimax rotation. A solution with eight factors was chosen. No items compromised scale reliability. The CD-specific items in the QoL section of the questionnaire showed overall excellent internal consistency. Scales showed good to acceptable internal consistency (Table 3).

Table 2 Varimax rotation of factor loadings of the 30 CD-specific items of the CDQL (N = 422)
Full size table
Table 3 Descriptive statistics for the eight CD-QoL scales, the worries-about-symptoms scale and the symptoms scale (N = 422)
Full size table

The positive correlations between the worries-about-symptoms scale, the symptoms scale, and the CD-QoL scales showed that the lower the impact of the symptoms was (a higher score on the scale), the higher was the HRQoL score. All scales correlated positively with generic health and QoL items supporting construct validity of the CDQL (Table 4). Twenty-one differences (p < 0.05) were found between children, adolescents, and adults in ten CD-specific QoL items. Four significant differences were found in two scales. Differences were also found between female and male respondents on four CD-specific items, on the generic health item, and on one CD-scale (Table 5). Sensitivity and discriminant validity was shown by significant within-group differences.

Table 4 Correlation matrix, Pearson correlations, pairwise two-sided p values, df = 375
Full size table
Table 5 Health-related quality of life results of children, adolescents, and adults with CD assessed by the CD-specific CDQL and the two generic items of WHOQOL-BREF
Full size table

Coeliac disease quality of life results

Most participants rated their quality of life ‘good’ or ‘very good’ (69%) on the generic QoL item (item GQ1). On the generic health item, more participants were ‘satisfied’ or ‘very satisfied’ with their health (50%) than dissatisfied or very dissatisfied (32%). Respondents younger than 18 years rated their satisfaction with their health in relation to having CD higher than adults did (item CQ28). This positive evaluation was also found on their generic rating of health and QoL (items GQ1 and GQ2) (Table 5).

Participants with diseases in addition to CD scored lower on the generic QoL item (M = 2.45, SD = 0.98), than participants with no other diseases than CD did (M = 2.97, SD = 0.90) t(415) = −5.56, p < 0.001. This difference was also found between the total CDQL scale score for participants with additional diseases (M = 1.68, SD = 0.57) and those with no other diseases than CD (M = 1.80, SD = 0.59) t(375) = −2.00, p < 0.05. Also, on the scale addressing worries about symptoms, a difference was found between participants with additional diseases (M = 2.26, SD = 0.80) and no other diseases (M = 2.47, SD = 0.88) t(274) = −2.03, p = 0.04. Finally, the symptoms scale showed the same pattern of respondents with additional diseases scoring lower (M = 3.08, SD = 0.73) than respondents with no other diseases than CD (M = 3.36, SD = 0.65) t(365) = −3.83, p < 0.001. The results suggest that CD patients with comorbidities perceived a lower QoL than patients with no other diseases than CD.

Respondents younger than 18 years reported higher satisfaction on the scale addressing contact to health care. The scale showed lower reliability score than other scales (see Discussion).

Negative correlations were found between the number of dietary transgressions (non-compliance with GFD) and both the total CDQL scale score r(375) = −0.13, p = 0.01 and the total CDQL item score r(375) = −0.14, p < 0.001. Respondents who complied with the GFD scored higher (i.e., better) on the symptoms scale (M = 3.34, SD = 0.68) than participants who reported dietary transgressions (M = 2.80, SD = 0.62) t(365) = −5.75, p < 0.001. The results suggested that CD-QoL was associated with dietary compliance.

Adolescents in particular were more satisfied with the quality of GF food than adults were (item CQ13).

Children scored lower than both adolescents and adults on the scale ‘Having coeliac disease and following a gluten-free diet,’ indicating that children experience a larger overall burden in managing CD and GFD. Children also scored lower on item CQ9 about one’s ability to find out whether the food contained gluten; however, the result was not statistically significant.

Children were more negatively impacted by following a GFD (item CQ7) and maintaining a GFD when eating out than adolescents and adults were (item CQ24). Interestingly, adolescents, on the other hand, rated their satisfaction with eating out higher than adults did, although this finding was not significant (p = 0.07). Respondents younger than 18 years reported less impact of CD-symptoms than adults did. This finding could be interpreted as children being less impacted by symptoms, but more troubled by managing CD and GFD than adults.

Positive correlations were found between the number of persons the respondent knew and both the total CDQL scale score r(375) = 0.19, p < 0.001, and the total CDQL item score r(375) = 0.19, p < 0.001. Members of the Coeliac Society did not score significantly higher on the QoL total scores. However, positive correlations were found between the number of CD-related events the respondent had participated in and the total CDQL scale score r(375) = 0.18, p < 0.001, the total CDQL item score r(375) = 0.17, p < 0.001, and the score on both the worries-about-symptoms scale r(274) = 0.19, p < 0.01, and the symptoms scale r(365) = 0.18, p < 0.001. The result suggests that being a member of the CD society alone did not improve QoL; however, participating actively in CD-related activities such as meetings, camps, or baking courses coincided with better QoL.

Children were more troubled by the fact that there are not many peers with CD than were adults (item CQ25). Similarly, respondents younger than 18 years were more negatively impacted than adults when thinking about food that was restricted for them (item CQ8). Most impacted of all were children, who felt worse than both adolescents and adults when thinking about seeing others eat food that they themselves could not eat (item CQ23). Respondents younger than 18 felt more negatively than adults when they thought about telling someone about CD (item CQ20). These results suggest that CD-specific social support and interventions targeting improved coping strategies may benefit CD-QoL.

When asked how participants imagined it will be in the future to have CD, children reported more negative outcomes than both adolescents and adults (item CQ26). Children were more negatively affected than adults by the fact that having CD cannot be changed (item CQ27).

Males report better HRQoL than females in all cases where significant gender differences were found. Males were more satisfied with the time it took to be diagnosed (item CQ1) and less worried about what may happen if they eat food containing gluten (CQ5). They were more satisfied with the quality of GF food (item CQ13), and they scored higher on the scale about GF food supply (Scale 7). They were also more satisfied with their health in relation to having CD (item 28) and their health in general (item GQ2). Males also scored higher (better) on the scales addressing symptoms and worries about symptoms.

Males also reported higher satisfaction on how others take one into account in relation to GF food (item CQ17, albeit p = 0.06), on the scale on GF food in relation to one self (Scale 5, albeit p = 0.09), on the scale about contact to health care (Scale 8, albeit p = 0.08), and on generic QoL (item GQ1, p = 0.08). Gender specific differences have previously been reported in other food-related chronic diseases [58].

Discussion

Respondents reported being mainly satisfied with their QoL and their health when these concepts were assessed by generic items. The impact of comorbidities on QoL pointed to the needs of patients that must manage different diseases. While patients shared concerns and burdens, they also represented a heterogeneous group. The detailed disease-specific items of the CDQL may therefore be more responsive in evaluating these concerns in CD than generic items [59, 60].

The perceived larger burden of following a GFD on children may be due to the limited scope children have in choosing and buying food. Competence in GF management may grow as the person becomes accustomed to following a GFD as well as a result of normal cognitive development.

Children and adolescents were more negatively affected by thoughts of desired gluten-containing food and by thoughts of others eating it, when they cannot. Children and adolescents are more affected by feelings of exclusion or difference from peers, than are adults, and find it more challenging to regulate their affect including feelings of envy. Young people are typically more dependent on friends as support in their psychological and social development. Having CD is likely to set children and adolescents apart from peers, giving rise to a danger of being excluded from peer groups [13].

Adolescents’ higher satisfaction with GF food and eating out may point to the fact that young people may have different standards and preferences in relation to food than adults have [61]. This may also be due to the differences in eating situations that children, adolescents, and adults face: children are presented with prepared food at, for example, birthday parties and may experience limited food-choices; adults may eat out at restaurants where the menu may require querying the staff; while adolescents may tend to eat fast-food that is now more available in GF versions than ever before.

Children were more troubled by not knowing others with the same dietary constraints. This may be due to children’s wishes not to be different from others [13, 36], the need for peer-support, and to the need for assistance in choosing GF food. The social support and dietary advice of others that can be pursued through participation in CD-related activities may be an important factor in improving QoL in CD.

Imagining the future is a complex cognitive task that develops with age. Children’s’ negative perceptions of the future in relation to having CD may represent a concrete issue for support: children may need to be assured that the future with CD will be brighter as more GF products become available and the knowledge of CD in the general population increases. Psychological support could also be aimed at the acceptance of having CD in children in order to increase dietary compliance and QoL.

Having CD and managing a GFD impacted more on females than males. This may be due to women valuing more highly the socialization aspect through shared meals more [8, 30, 32]. However, this may also be due to females experiencing more severe impact of CD-related symptoms than males, as female respondents reported higher impact of CD-symptoms than males did. A similar pattern has been found in relation to other food-related chronic diseases and in the general population [58]. More research on the interaction of gender in reporting and in experiencing having CD is warranted.

Respondents were analyzed in two and three age groups. Analyses of other age groups could also be interesting to discover the patient-reported impact of CD across the lifespan. The age group 18–25 years may constitute a period with specific developmental tasks in relation to identity exploration and growing independence [62]. This group faces other challenges than adolescents and adults, and may therefore experience a different impact in disease management that may become apparent in differences in item or scale scores.

There were limitations in the development of the CDQL. The initial formulation of the questionnaire items was based on focus groups with children and adolescents, and only few adults. The items were, however, formulated in a non-age-restrictive manner. The pilot data collection and cognitive interviews involved all age groups ensuring that adult’s concerns were covered in the items, and demonstrated that the questionnaire could be meaningfully answered and completed by all ages.

The lower reliability of scale 8 may reflect differences in patients’ use of health care. The diagnosis may have happened recently or many years ago. Items on diagnosis-experiences were included as patients described it as a life-altering event. Health checks (item CQ3) are routinely offered to children and adolescents. Adults are expected to maintain a GFD and are not offered checks. This may lead to differences in the frequency of contact to health care in relation to having CD. The scale was kept as it addressed important aspects of being diagnosed with CD and the transition from pediatric to adult care.

Three items cross-loaded on two factors, each indicating associations with two latent concepts (Table 2). In closer inspection this proved meaningful in relation to patients’ accounts. The marking on GF foods (item CQ10) was related to both ‘Knowing about CD and GF food’ (scale 6) and to ‘GF food supply’ (scale 7). Being offered food that the person knew contained gluten (item CQ14) was related to ‘Confronting gluten-containing food’ (scale 5 at 0.447) and to ‘Having CD and following a GFD’ (scale 1 at 0.446). How others took one into account in relation to GF food (CQ17) was associated with ‘Others’ handling of the person’s’ CD (scale 4) and with ‘Communicating about CD and GF food’ (scale 2). Cross-loadings may be expected when two issues are formulated together in one item. The items were formulated on the insight that social situations involving food are central to CD-QoL [8, 13, 14, 29, 36]. The items were kept because they were regarded as central and valuable to measuring CD-related QoL by patients and clinicians. On the basis of focus group interviews and the definition of HRQoL, the items on labeling, availability, and cost of GF food were seen as formative of HRQoL. Following this, it can be hypothesized that HRQoL will change if the labeling or availability of GF food or the cost of living on a GFD changes. The CDQL will be tested in Ireland to validate the English version and for cross-cultural comparison.

The aim of this study was to develop the Coeliac Disease Quality of Life (CDQL) questionnaire—a comprehensive CD-specific HRQoL questionnaire that applies to all age groups and can be answered directly by the patients or by proxy. The CDQL is comprehensive with 12 profile items, 2 generic HRQoL items, 30 CD-specific items, and a commentary field. CD-specific QoL seems to revolve around the social nature of food: not just wanting to eat a particular food, but to eat the same as others. The items of the questionnaire correspond to the experiences of having and coping with CD [13, 14, 29, 32, 36] supporting face and content validity of the CDQL. Patients’ evaluation of their QoL in relation to their perceived impact of gluten intake, emotional and social support, and communicating dietary needs when following a GFD, is measured. The CDQL can be used for patient-reported CD-QoL evaluation across age groups allowing direct comparison on the same items and scales. Future research will examine and validate the CDQL’s performance on different age groups.

Monitoring and discussing HRQoL have been shown to increase awareness of emotional and psychosocial issues [63] and improve psychosocial well-being in some cases in adolescents with type 1 diabetes [64] and in adults with CD [65]. The lasting improvement in well-being may however be dependent on ongoing HRQoL assessment [66]. A multidimensional and sensitive age appropriate tool can improve patient-clinician communication and allow for the clinician to select interventions that are appropriate and targeted specifically to the patient. HRQoL and other patient-reported outcome measures may prove useful as tools to focus communication on coping with food intolerance or difficulties in managing risk. Coping skills training and techniques and communication skills may be particularly useful. Future research should also explore the possible impact of age, culture, and gender-related variables. Further evidence-based research aimed at exploring the effects of interventions for CD on HRQoL-outcomes is needed. Given that this study was cross-sectional, the long-term stability of these findings requires further inquiry. In the future, the CDQL will be used to monitor changes in HRQoL before and after supportive CD-specific interventions focusing on problem-based learning, social support, and well-being.