Adherence to a Gluten-free Diet in Celiac Disease
Adherence to a Gluten-free Diet in Celiac Disease
When measuring adherence to a GFD, factors such as a protracted learning curve or patient's perception of increased burden of treatment could render initial measurements of adherence during the first years after diagnosis unrepresentative of lifelong adherence. Therefore, measuring adherence in the long-term is likely a more representative marker of ongoing and future adherence. No known previous US study has directly addressed the question of long-term GFD adherence, specifically, more than 5 years after CeD diagnosis.
Our study is, to the best of our knowledge, the largest long-term GFD adherence study in the USA. We found an adequate adherence rate of 75.5% with 24.5% not adhering closely to the diet. Our adequate adherence rate is consistent with rates of 75.8% in a prior study with our institution's Celiac Center patients. Canadian population studies in adult patients with CeD have found adherence rates between 68% and 90%, consistent with our study's findings. However, in the literature, the adherence rates vary from 33.7% to 95%. Possible reasons for this variability include the use of different methodologies to determine adherence, differential response bias, the time since diagnosis in the patient population and the nature of the study population itself (community/country, reference centre patients, support group patients, Internet-recruited patients). Therefore, comparability between studies is limited and population-specific data are necessary.
One of the strengths of our study is the novel use of the previously validated test to determine CeD adherence (CDAT) yielding both quantitative and qualitative measures of adherence to a GFD. Using the CDAT, this study was able to correlate key patient factors with GFD adherence. Another novel aspect of this study was the incorporation of median annual household income data as correlated with adherence.
This study explored a number of key factors correlated with GFD adherence. The issue of cost as a barrier to GFD adherence has been described previously. A US study in 2007 found that on average a standard gluten-free 'product basket' was 240% more expensive than its gluten-containing counterpart. A similar study in the UK in 2011 found that 11 of 19 gluten-free products studied were significantly more expensive than their gluten-containing counterparts. Also, the price of the gluten-free products was at least 2% but as high as 518% more expensive than their gluten-containing counterparts.
Our patient population is almost evenly divided on whether they consider cost to be an obstacle to adherence. The cohort of respondents that considered cost to be a limiting factor for adherence have significantly higher CDAT scores and lower rates of adequate adherence. Nevertheless, respondents with an adequate adherence did not have significantly different median annual household income as compared to respondents with inadequate adherence. This finding raises the question of whether perception of cost vs. actual economic limitation is an obstacle to GFD adherence.
Another key factor that was explored in this study was education and the perception of knowledge of the gluten-free diet as factors in adherence. A higher education level was correlated with better adherence independent of median household income. A higher education level was also associated with better self-perceived knowledge of the GFD. Nonetheless, respondents' perceived knowledge of the GFD remains independently correlated with better adherence even after correcting for education level. These findings may translate into clinical practice by providing patients with self-perceived poor or fair knowledge of the GFD with increased dietary counselling regardless of the patient's education level.
Our findings of a higher proportion of serological normalisation in patients with adequate adherence as compared to those with inadequate adherence are consistent with previous findings that support that adherence to the diet is key for serological normalisation or clinical improvement. However, serology cannot be reliably used as a marker of adherence to a GFD in patients with CeD.
Our study has some limitations that are important to note. As in any survey study, there will be a group of nonrespondents for whom it is impossible to rule out the possibility of different adherence behaviour. However, we assessed the respondent and nonrespondent group for differences in gender, age, location, time on a GFD and median household income and found no differences. We did find a significant difference in the availability of tissue transglutaminase IgA samples in the period around the administration of the survey. This likely reflects a lesser involvement with healthcare and use of health services in the nonrespondent subpopulation rather than a different clinical behaviour as supported by the absence of difference in the proportion of patients with serological normalisation. A previous study with a mail-in survey for functional gastrointestinal disorders having a 52% response rate reviewed medical records of a subset of nonrespondents and found no significant difference in gastrointestinal symptoms or diagnoses but did find differences in healthcare usage behaviour concluding that a nonresponse bias seems to be nonsubstantial in this type of study.
In summary, adequate long-term adherence to a GFD was observed in the majority of CeD patients that responded to our survey but one quarter of those who responded reported poor adherence. Factors associated with inadequate adherence were a lower education level and a negative perception of patients' own knowledge of the GFD as well as of their own self-effectiveness in following the diet. Factors such as patient age, gender, time on a GFD and median household income were not associated with GFD adherence. These findings bring to light the challenge of designing proven, cost-effective strategies to improve long-term GFD adherence in the poorly adherent population. Future studies should focus on developing and testing interventions for this non-adherent group. Such interventions might target increased knowledge of the GFD and perceptions of the importance of close adherence.
Discussion
When measuring adherence to a GFD, factors such as a protracted learning curve or patient's perception of increased burden of treatment could render initial measurements of adherence during the first years after diagnosis unrepresentative of lifelong adherence. Therefore, measuring adherence in the long-term is likely a more representative marker of ongoing and future adherence. No known previous US study has directly addressed the question of long-term GFD adherence, specifically, more than 5 years after CeD diagnosis.
Our study is, to the best of our knowledge, the largest long-term GFD adherence study in the USA. We found an adequate adherence rate of 75.5% with 24.5% not adhering closely to the diet. Our adequate adherence rate is consistent with rates of 75.8% in a prior study with our institution's Celiac Center patients. Canadian population studies in adult patients with CeD have found adherence rates between 68% and 90%, consistent with our study's findings. However, in the literature, the adherence rates vary from 33.7% to 95%. Possible reasons for this variability include the use of different methodologies to determine adherence, differential response bias, the time since diagnosis in the patient population and the nature of the study population itself (community/country, reference centre patients, support group patients, Internet-recruited patients). Therefore, comparability between studies is limited and population-specific data are necessary.
One of the strengths of our study is the novel use of the previously validated test to determine CeD adherence (CDAT) yielding both quantitative and qualitative measures of adherence to a GFD. Using the CDAT, this study was able to correlate key patient factors with GFD adherence. Another novel aspect of this study was the incorporation of median annual household income data as correlated with adherence.
This study explored a number of key factors correlated with GFD adherence. The issue of cost as a barrier to GFD adherence has been described previously. A US study in 2007 found that on average a standard gluten-free 'product basket' was 240% more expensive than its gluten-containing counterpart. A similar study in the UK in 2011 found that 11 of 19 gluten-free products studied were significantly more expensive than their gluten-containing counterparts. Also, the price of the gluten-free products was at least 2% but as high as 518% more expensive than their gluten-containing counterparts.
Our patient population is almost evenly divided on whether they consider cost to be an obstacle to adherence. The cohort of respondents that considered cost to be a limiting factor for adherence have significantly higher CDAT scores and lower rates of adequate adherence. Nevertheless, respondents with an adequate adherence did not have significantly different median annual household income as compared to respondents with inadequate adherence. This finding raises the question of whether perception of cost vs. actual economic limitation is an obstacle to GFD adherence.
Another key factor that was explored in this study was education and the perception of knowledge of the gluten-free diet as factors in adherence. A higher education level was correlated with better adherence independent of median household income. A higher education level was also associated with better self-perceived knowledge of the GFD. Nonetheless, respondents' perceived knowledge of the GFD remains independently correlated with better adherence even after correcting for education level. These findings may translate into clinical practice by providing patients with self-perceived poor or fair knowledge of the GFD with increased dietary counselling regardless of the patient's education level.
Our findings of a higher proportion of serological normalisation in patients with adequate adherence as compared to those with inadequate adherence are consistent with previous findings that support that adherence to the diet is key for serological normalisation or clinical improvement. However, serology cannot be reliably used as a marker of adherence to a GFD in patients with CeD.
Our study has some limitations that are important to note. As in any survey study, there will be a group of nonrespondents for whom it is impossible to rule out the possibility of different adherence behaviour. However, we assessed the respondent and nonrespondent group for differences in gender, age, location, time on a GFD and median household income and found no differences. We did find a significant difference in the availability of tissue transglutaminase IgA samples in the period around the administration of the survey. This likely reflects a lesser involvement with healthcare and use of health services in the nonrespondent subpopulation rather than a different clinical behaviour as supported by the absence of difference in the proportion of patients with serological normalisation. A previous study with a mail-in survey for functional gastrointestinal disorders having a 52% response rate reviewed medical records of a subset of nonrespondents and found no significant difference in gastrointestinal symptoms or diagnoses but did find differences in healthcare usage behaviour concluding that a nonresponse bias seems to be nonsubstantial in this type of study.
In summary, adequate long-term adherence to a GFD was observed in the majority of CeD patients that responded to our survey but one quarter of those who responded reported poor adherence. Factors associated with inadequate adherence were a lower education level and a negative perception of patients' own knowledge of the GFD as well as of their own self-effectiveness in following the diet. Factors such as patient age, gender, time on a GFD and median household income were not associated with GFD adherence. These findings bring to light the challenge of designing proven, cost-effective strategies to improve long-term GFD adherence in the poorly adherent population. Future studies should focus on developing and testing interventions for this non-adherent group. Such interventions might target increased knowledge of the GFD and perceptions of the importance of close adherence.
Source...