Questionnaires to Assess Atopic Dermatitis Prevalence May Be Problematic

Questionnaires to determine history of atopic dermatitis may carry risk of bias and misclassification, notes a new study from Denmark.

A team of investigators, led by Lea Nymand, MSc, of the University of Copenhagen, examined the accuracy of questionnaires used to test the prevalence of atopic dermatitis and psoriasis in adults.

Questions that inquire about previous diagnosis of either condition have shown to have moderate-to-high sensitivity and specificity, according to other validation studies.

“Importantly, while studies have validated these questions as a snapshot in time, to our knowledge no study to date has assessed whether adult patients change their responses over time,” Nymand’s team wrote.

“Understanding the potential inconsistencies in responses is important because they may explain some of the fluctuations and changes in prevalence that may be reported.”

Validating Questionnaires

The investigators conducted a nationwide population-based cohort study where they assessed patient responses from the Danish Skin Cohort, a prospective cohort designed to evaluate the natural history and disease course of psoriasis and atopic dermatitis in Denmark.

Of the entire cohort, 2333 and 2312 patients had responded to questionnaires on psoriasis and atopic dermatitis, respectively, in both 2018 and 2020. They indicated whether a dermatologist and/or a non-specialist has ever diagnosed them with either condition. The same questionnaire was administered in the exact same format at the 2-year follow-up.

Responses were linked to individual patients using unique identification numbers. Nymand and colleagues then used Cohen κ to assess test-retest reliability, and results were interpreted according to Landis and Coch (≤0 = no agreement; 0.81-1.00 = almost perfect agreement).

The team visualized participants responses with Sankey diagrams, and they calculated prevalence as positive responses divided by total responses.

The Results

“Among participants reporting a history of psoriasis, agreement between individual responses was high (κ = 0.7558); however, among those reporting a history of atopic dermatitis, agreement was low (κ = 0.4395),” the investigators reported.

Prevalence of psoriasis among patients changed from 7.8% to 8.0%, and prevalence of atopic dermatitis decreased from 8.2% to 7.6%.

Even more, of those who reported in 2018 having been diagnosed with atopic dermatitis by dermatologist (n = 84), 36.9% indicated in their 2020 questionnaire responses that they never had it.

Among those patients who initially reported they were diagnosed by a nondermatologist-physician (n = 60), 66.7% changed their answer to never having atopic dermatitis.

And finally, the investigators’ analysis showed substantial agreement for psoriasis across all age age groups. On the other hand, for atopic dermatitis, κ declined with increasing age, with 0.2613 being associated with those ≥65 years old. 

“In general, assessment of atopic dermatitis in adults using questionnaire data is problematic,” Nyman and colleagues wrote.

“Similar to the unreliability found in the present study,” they continued, “we have previously shown that applying the United Kingdom Working Party criteria in adult populations can identify patients with atopic dermatitis with a sensitivity of 0.71 and a specificity of 0.96 in the general population but may also incorrectly diagnose atopic dermatitis in a large proportion of patients with other skin diseases, such as psoriasis.”

The study, “Limitations of Using Questionnaires for Assessing the Prevalence of Psoriasis and Atopic Dermatitis Among Adults,” was published online in JAMA Dermatology.