This survey was conducted in June 2005 in the UK, France, and Germany by internet; in Italy mainly (89%) by internet but also by telephone; and by telephone in Spain where internet usage rates are low. The survey was piloted as a paper version administered in-person in the USA. The survey was then translated by native speakers of the target languages, and the translations were checked for equivalence to the English by qualitative researchers from each of the countries.
The survey consisted of a screening section with the exclusion and inclusion criteria, two sociodemographic items, and four content items, and then the main questionnaire with 38 content items and seven sociodemographic items; many items contained a few subquestions. These questions were designed mostly as multiple choice and yes/no lists, with some quantification questions and Likert-scale rating questions. The content items of the survey covered topics such as: clinical manifestations of AR, treatment behavior patterns, and medication preferences. In the present paper, we focus selectively on the subset of data revealing a widespread, multistage pattern of undertreatment of rhinitis symptoms.
Survey participants were recruited as follows. For internet-based interviews, ‘panels’ of people in each of the countries were used. ‘Panelists’ were initially recruited through a variety of online advertisement channels and through affiliated companies and were invited to join the panel only to participate in online surveys. When they first registered, the panelists had to provide profiling information about themselves and they agreed to periodically participate in a variety of online surveys. The participation level and data quality from individual panelists are monitored when they participate in surveys. Panelists who are inactive or identified as providing dubious survey answers are deregistered. Survey invitation was aimed at obtaining a mixture of gender, age, and geographic region to match the census data of the country. Within those three parameters, panel members were randomly selected. The sampling procedure for telephone surveys proceeded by dialing random numbers from a purchased list of residential numbers, distributed by area code, originally drawn from telephone books. Telephone surveys were conducted on weekday evenings (17:00–21:00) and weekends (11:00–20:00). If no one was reached or available at a number, it would be called again 24–48 h later, up to seven times. If the person answering was under 18, the surveyor would ask to speak to someone else.
Respondents were informed that the survey was about ‘allergies, including hay fever, pet allergies, dust, and dust mite allergies’. The survey began with screening questions. The exclusion criteria were: (i) occupation of the respondent or any immediate family member in marketing, market research, advertising, health-care products, pharmaceuticals, or health-care professions; (ii) under 18. The inclusion criterion was: suffers from outdoor allergies, indoor allergies, pet allergies, or hives as the most frequent form of allergy suffered from. After being asked simply whether or not they suffer from allergies, affirmative respondents were asked, ‘What kind of allergies do you suffer from: (a) outdoor allergies (hay fever, pollen, trees, grass, weeds); (b) indoor allergies (mold, dust, dust mites); (c) hives, allergy rash (urticaria); (d) pet allergies; (e) insect allergies (such as bites and/or stings); (f) medication allergies; (g) food allergies; (h) none of the above’ and were allowed to check all that apply. The following question asked, ‘Which do you suffer from most frequently? (select only one)’ and presented all the options they had just affirmed. So those who chose a, b, c, or d from the list above as the most frequent form of allergy suffered from were included in the study, while all others were not included.
The following numbers of people contacted either declined to participate or met one of the two exclusion criteria: UK = 1610, F = 1410, G = 2073, I = 879, S = 514. The following numbers of people contacted did not meet the inclusion criteria: UK = 612, F = 911, G = 2279, I = 970, S = 1373. The following numbers of qualified participants withdrew before the survey was completed: UK = 195, F = 342, G = 278, I = 168, S = 140. The raw data were weighted to match census data for the parameters of sex, age, and income, specific to each country.
The survey was completed by 2966 adults with AR in the EU (UK: n = 585, France: n = 577, Germany: n = 654, Italy: n = 575, Spain: n = 575). After data weighting, this transforms to a total survey population of 3036 (UK: n = 615, France: n = 595, Germany: n = 676, Italy: n = 575, Spain: n = 575). Although the tables and figures present the data from the five countries differentially, the study is not meant to be comparative. For the most part, the variations between the countries are not large enough to be meaningful, thus demonstrating that the problem is pan-European. Although the survey did not cover all EU countries, it is most likely applicable to all western European countries.
There is one last important point to make about the subjects of this survey. The survey relies upon respondents’ self-reports about their allergies but never clinically confirms their reports. A previous study has shown that respondent self-reports and clinical diagnoses do sometimes diverge (2). However, it is safe to say that the vast majority of these survey respondents did in fact have AR. As the survey results found (see below), half to two-thirds of the respondents had seen a doctor within the past year for the their allergies, so we may presume that they have already had a skin-prick or blood test from their doctor, confirming the allergies they reported in the survey. Also, only about 10% of the respondents had never seen a doctor for their allergies, but they cannot all be wrong about their condition. So only <10% of the respondents may be mistaken about whether or not they have allergies – and precisely because they do not seek medical diagnosis for their symptoms, which is the issue discussed here. So finally, it should be emphasized that this paper is not about the epidemiological prevalence or clinical symptomology of AR itself. This survey is about the health behaviors of people in the general population who acknowledge (rightly or wrongly, but almost all rightly) that they have allergies.