In this study comprising 1,516 young adults, the 1-year prevalence of hand eczema was more than 15%. One third of these individuals also had 1-year prevalence at the baseline 1995. The 1-year prevalence, and not the point prevalence, was used in all calculations because it better reflects the persistency, the relapsing course and the seasonal variations of the disease [2, 19]. The increase in the one-year prevalence between the two occasions is in accordance with previous large Swedish cross-sectional studies with respect to the age groups [3–5, 22].
The estimated incidence of hand eczema in our study was 11.6 cases per 1000 person-years, 14.3 among females and 5.2 among males. Our figures are in the upper amplitude compared to an earlier population based study from Sweden, which showed between 11.4 and 3.7 cases/1000 person-years among 20–29 year-old females and males, respectively . One explanation could be that our study is prospective, and underreporting is to be expected in retrospective questionnaire studies . Based on 7 European hand eczema studies performed among 16–77 years-olds, the median incidence rate of hand eczema was 9.6 cases/1000 person-years (range 4.6–11.4) among women and 4.0 cases/1000 person-years (range 1.4–7.4) among men , which is also slightly lower than our current findings, probably due to age-differences. To the best of our knowledge there are no comparable studies of the cumulative incidence in this age group. The cumulative incidence of hand eczema in our study across 13 years was 15.1% (18.6% for females and 10.7% for males). This can be considered to be a high proportion . When using a questionnaire for estimating the true occurrence of a disease it is important to know the sensitivity and specificity of the question used. The question on 1-year prevalence of hand eczema underestimates the occurrence. . However, regarding childhood eczema the occurrence has been found to be overestimated especially if the true prevalence is low [5, 19]. Based on prevalence as well as incidence, the occurrence of hand eczema is approximately twice as common among females compared to men, which is similar to other population-based studies [1, 26, 27].
The advantage of a longitudinal cohort study compared with a cross-sectional study is that it enables the estimation of both cumulative incidence and incidence rate. Another advantage of performing a follow-up study is the possibility to compare the development of hand eczema over time in relation to different risk factors.
The four groups (HX9508, HX95, HX08 and NoHX) were used to investigate the relationship between childhood eczema and the incidence of hand eczema. The assumption was that a smaller proportion of individuals who had hand eczema in 2008 but not in 1995 reported childhood eczema. However, there were no significant differences between the three hand eczema groups concerning childhood eczema. Furthermore, it was found that a higher proportion of individuals who had hand eczema at both occasions reported childhood eczema.
Thus, in this cohort childhood eczema was the most important predicting factor regardless of the debut of hand eczema. In 2008, around 30% of our sample reported childhood eczema (females 36%, males 20%). In a large population-based Swedish study performed from 2002–2003, among 21–30 years-olds, childhood eczema was reported by 30.1% of females and 20.8% of males, [4, 28, 29]. The corresponding figures in the 31–40 year-olds were 21.8% and 16.2% . Thus, in our study, the prevalence of childhood eczema was higher. Similar to other studies, the relationship between having had hand eczema and reporting childhood eczema was highly significant . The agreement in self-reports of childhood eczema at the two occasions was high. This high reliability over time in this age-group can be useful to know when hand eczema is diagnosed. However, the lower rate of reported childhood eczema in 2008 can be explained by recall bias as was found in a study comprising respondents aged 31 to 42 years . For the individuals who reported only rhino-conjunctivitis, there was no significant association with one-year prevalence of hand eczema. Also, there was no association with asthma only, but there were very few respondents. Thus in our study no additional information concerning risk for hand eczema was obtained by asking about asthma or rhino-conjunctivitis. These results are in accordance with Meding et al. who showed that asthma and rhino-conjuntivitis in adults were only associated with hand eczema at an age below 30 years ; in another study, including adolescents, a marginally significant association with inhalant allergy was found .
Analyses of exogenous factors showed that the individuals with hand eczema only in 2008, reported a significantly higher frequency of hand washing compared to the individuals without hand eczema.
Females with hand eczema spent significantly more time doing household activities than men with hand eczema (Table 3). Hand washing was more frequent among females with hand eczema than females without hand eczema as well as compared with men with hand eczema. In the multinominal regression analyses hand washing in the group HX08 was the only significant exogenous risk factor associated with hand eczema. In the majority of hand eczema studies hand washing is found to be the most significant risk factor for developing hand eczema . In our cohort, other exogenous risk factors such as cooking, washing and cleaning and taking care of young children did not have any significant association with hand eczema. Furthermore, female gender was not a significant risk factor. However, it is well known that females have hand eczema more often than men. This can be explained by the high exposure to water and other skin irritants. Experimental as well as epidemiological studies [14, 35] have demonstrated that female skin is not more sensitive to irritants than male skin  which is in line with our findings.
An interesting finding was the high odds-ratio in daily use of moisturisers in the two groups with current 1-year prevalence of hand eczema (HX9508 and HX08). This pattern was not seen in the group having had hand eczema in 1995 (HX95).
When self-administrated questionnaires are used, it is important for the results to be adjusted based on sensitivity and specificity of validated questions. This is especially important in diseases that are common and affect the general health and well-being of individuals, such as hand eczema. The development of specific instruments like questionnaires implicates problems. In this case the questions regarding childhood and hand eczema were not validated in 1995 but 2,535 of the 2,572 pupils (98.6%) were clinically examined, and the sensitivity of 73% and the specificity of 99% were found . The question regarding the 1-year prevalence of hand eczema, which was used in the present study and in the first study, was previously validated . Thus, the true one-year prevalence of hand eczema can be estimated from our data and is 20.6% for all; 26.8% among females and 12.5% among males.
The answers to the open questions on occupation as well as work tasks gave no further information regarding risk factors for developing or maintaining hand eczema. This circumstance seems to be a common problem in questionnaire studies . In a study regarding occupational exposure to water as a risk factor for hand eczema, it was found that the title of an occupation gave misclassified results; exposure time and frequency of water use were more appropriate measures . For result validity, it is important to have high response rates in general population studies [37–39]. The response rate in this study was almost two thirds of the individuals who received a questionnaire in the mail. Females were significantly more willing to participate than the males. There were, however, no significant differences within the female or the male groups regarding having had 1-year prevalence of hand eczema at the two occasions. The response rate was similar to the annual national public health questionnaire performed by Swedish National Institute of Public Health .