Discussion
We report there are significant patient-centered barriers that impede serologic screening and contribute to the delayed detection and diagnosis of celiac disease. The diagnosis of celiac disease is delayed from 5.8 to 11 years on average after the onset of symptoms. This is the case despite readily available inexpensive serologic screening tests of high sensitivity, specificity and an increased positive predictive value when applied to populations having an increased prevalence of celiac disease.
We surveyed a group of volunteers from the general population who met study inclusion criteria for symptoms, co-morbidities and/or family history placing them clinically in populations with an increased prevalence of celiac disease. Participants had a sex ratio that paralleled that found among diagnosed celiac disease patients, symptoms and co-morbidities typical of those found in diagnosed celiac disease patients and a broad range of income and education levels.
We reasoned a priori that key patient-centered barriers would relate to a lack of knowledge about celiac disease and serologic testing. This proved to be the case as the most significant barriers among the participant group were, "I did not know I was at risk for celiac disease" and "I did not know where to get tested". Although our study was powered to find significant barriers that affect the entire participant group, we cautiously point out several barriers that were found in some of the demographic subgroups and may indicate worthwhile avenues for further research or outreach. Regarding access to healthcare, for example, not having a doctor was a significant barrier among those with annual incomes less than $25,000 and those in the younger half of participants. Consistent with this, not having health insurance strongly correlated with participants not having a doctor. It remains to be seen to what extent recent changes in health care laws will increase the number of Americans with both insurance and a doctor, and if this will lead to a higher frequency of testing among individuals in those demographic groups.
Lack of knowledge among the public about celiac disease, its symptoms, and risk factors represented an important barrier to diagnosis. More than 50% of participants reported not knowing anything about celiac disease until recently or not knowing that they were at risk for celiac disease. As a result, at-risk patients seeing a doctor for their symptoms would be unlikely to question their physicians about celiac disease or request serologic testing. Interestingly, more than a third of participants did not think their doctor would know which test to order, suggesting low patient confidence in physician knowledge about celiac disease. Consistent with that, primary care doctors, who form the front line for detection of probable celiac disease, were in fact found to have limited knowledge of the symptoms, natural history, and methods of testing for celiac disease, implying that greater education of health care professionals about celiac disease would be beneficial. To address both problems, the National Institute of Health launched a campaign in 2006 to increase celiac disease awareness among health-care professionals and the public. In addition, the growing presence of gluten-free food products in grocery stores, gluten-free options on restaurant menus, and articles about celiac disease and non-celiac gluten sensitivity in the lay press also may help to increase public awareness and the patient-physician conversation on this topic.
A lack of motivation to get tested for celiac disease was self-reported in more than 40% of participants and was the most common barrier cited by those in the highest income group. Nonetheless, each participant possessed sufficient motivation to contact the Warren Medical Research Center for Celiac Disease after seeing our posted flyers, engage in a telephone screening process, and, if qualified, complete the study. Study completion required the time and the cost of travel to the UCSD campus, the time taken to complete a 20–30 minute questionnaire, and the donation of a blood sample for celiac serology. This level of motivation likely exceeds that in the general population.
Individuals with celiac disease present with a spectrum of intestinal and extraintestinal symptoms. Many of these symptoms are not debilitating, and patients often learn to "live with" their symptoms despite a decreased quality of life. This is consistent with 85% of the participants rating their overall health as good to excellent and may help explain a lack of motivation for testing. Although 10 participants claimed to not have symptoms at the time of the screening telephone interview, each of them reported symptoms consistent with celiac disease on our questionnaire.
Three of 119 (2.5%) of participants enrolled in the study, or roughly 1:40, tested positive for IgA tTG. This is consistent with a prior study in which 1 in 56 (1.7%) symptomatic individuals tested had positive serology, compared to 1:133 (0.7%) of the general population. The fact that 2/33 (6%) of our participants who indicated they had a first degree relative with celiac disease tested positive for IgA tTG is consistent with the reported prevalence of celiac disease in 5–10% of first-degree relatives of celiac disease probands. The 3 serologically positive individuals in our study would be classified as presumptive celiac disease according to AGA criteria, since small intestinal mucosal biopsy was not included as part of this barriers research study. We do not know whether any of the participants with symptoms and negative IgA tTG values have non-celiac gluten sensitivity.
AGA, ACG, and NIH recommendations favor active case-finding, rather than general population screening. This strategy is more cost-effective as it increases the pretest probability and therefore the rate of positive testing, leading to fewer unnecessary procedures. However, such guidelines leave the question as to whom to screen open to clinical discretion as many common symptoms and co-morbidities are associated with an increased prevalence of celiac disease. For example, a study using an active-case finding strategy to screen for celiac disease reported that 64% of survey respondents met the criteria for celiac screening. This strategy will also miss many individuals with asymptomatic celiac disease and those with atypical symptoms. Notably, the case-finding strategy thus far has failed to detect and diagnose a large majority of individuals with active celiac disease. However, it is also controversial as to whether celiac disease is an appropriate disease for mass screening consistent with World Health Organization criteria. Nonetheless, it is clear that new improved testing strategies for celiac disease appropriate for application in the general or select populations are required.
Several limitations of this study are worth noting. We powered the study to find significant barriers overall, and this limited the ability to definitively delineate some of the barriers within distinct demographic subgroups. Participants were not a random sample from high celiac disease prevalence populations. We designed the study such that participants were individuals from the general population who met criteria for being at a higher risk for celiac disease than those in the general population at large. Further, the study did not accurately reflect the ethnic populations residing in San Diego County as it included 13.4% of participants who self-identified as Hispanic, compared to the 33% of residents of San Diego County who reported as Hispanic by census. Finally, we note that large prevalence studies have reported roughly equal numbers of men and women with celiac disease. Nonetheless, women are more frequently diagnosed with celiac disease than men at a rate of roughly 3:1, and the population of women and men who volunteered for this study was consistent with that ratio. This female predominance may reflect a greater likelihood for women to seek medical attention for their symptoms or press their physicians for celiac testing. In this study, our advertising strategy (e.g. in grocery stores) may have reached more women than men.