Health and Quality of Life Outcomes BioMed Central

Background: It is known that there is an increase in the prevalence of allergy and that allergic diseases have a negative impact on individuals' health-related quality of life (HRQL). However, research in this field is mainly focused on individuals with verified allergy, i.e. leaving out those with self-reported allergy-like conditions but with no doctor-diagnosis. Furthermore, studies on food hypersensitivity and quality of life are scarce. In order to receive information about the extent to which adolescent females and males experience allergy-like conditions and the impact of these conditions on their everyday life, the present study aimed to investigate the magnitude of selfreported allergy-like conditions in adolescence and to evaluate their HRQL. Special focus was put on food hypersensitivity as a specific allergy-like condition and on gender differences. Methods: In connection with lessons completed at the children's school, a study-specific questionnaire and the generic instrument SF-36 were distributed to 1488 adolescents, 13–21 years old (response rate 97%). Results: Sixty-four per cent of the respondents reported some kind of allergy-like condition: 46% reported hypersensitivity to defined substances and 51% reported allergic diseases (i.e. asthma/ wheezing, eczema/rash, rhino-conjunctivitis). A total of 19% reported food hypersensitivity. Females more often reported allergy-like conditions compared with males (p < 0.001). The adolescents with allergy-like conditions reported significantly lower HRQL (p < 0.001) in seven of the eight SF-36 health scales compared with adolescents without such conditions, regardless of whether the condition had been doctor-diagnosed or not. Most adolescents suffered from complex allergy-like conditions. Conclusions: The results indicate a need to consider the psychosocial impact of allergy-like conditions during school age. Further research is needed to elucidate the gender differences in this area. A team approach addressing better understanding of how allergy-like conditions impair the HRQL may improve the management of the adolescent's health problems, both in health-care services and in schools. Published: 19 November 2004 Health and Quality of Life Outcomes 2004, 2:65 doi:10.1186/1477-7525-2-65 Received: 28 September 2004 Accepted: 19 November 2004 This article is available from: http://www.hqlo.com/content/2/1/65 © 2004 Marklund et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.


Background
About 20% of schoolchildren and adolescents in Sweden suffer from perceived hypersensitivity (e.g. allergy or intolerance) to defined food items [1,2]. The way such allergy-associated condition is manifested varies extensively depending on the specific foods and the type of reactions and symptoms that the child has. Adverse reactions to food can result in either immediate or delayed problems depending on the mechanism underlying the reaction. Food reactions can range from mild (e.g. mild itching or redness around the mouth), unpleasant and scary (e.g. intense stomach pains or asthma) to systemic reactions severe enough to be fatal (i.e. anaphylaxis) [3]. As there is no satisfying treatment for most kinds of food hypersensitivity individuals have to manage their problems by themselves and, if possible, try to avoid the offending food.
It is not uncommon that individuals with food hypersensitivity have other hypersensitivities or allergy-associated problems as well [4][5][6]. Food allergy is associated with other manifestations of atopy, such as urticaria, asthma, atopic eczema and hay fever [7]. Our previous study showed that 78% of food hypersensitive Swedish schoolchildren may experience such symptoms that can be related to asthma, eczema and/or rhino conjunctivitis [2]. A number of studies describe deterioration of the healthrelated quality of life (HRQL) in children and adults with the allergic diseases asthma [8][9][10][11], eczema [12][13][14][15] and rhinitis [16][17][18][19]. However, physical and organ-related measures and tests do not always correlate with HRQL scores [20][21][22]. Thus there is no absolute coherence between the biomedical severity of the allergy-associated condition and quality of life.
In recent years there have been a number of studies that show that food hypersensitivity impair the HRQL of individuals and even their families [23][24][25][26]. Our recent school survey [2] showed that those young individuals who, in addition to their food related problems suffered from other allergy-associated conditions, reported a greater impact on the physical life quality dimension than those without such chronic problems. The psychosocial impact of food hypersensitivity was however reported equally low regardless of other allergy-associated conditions [2]. Avery et al [24] have shown that children with peanut allergy have a poorer HRQL, apparently related to anxiety, compared with children with diabetes. Studies from Sicherer et al [25] and Primeau et al [26] have demonstrated a psychosocial impact of food allergy on quality of life on both the child and its family. It has also been shown that the parents of food allergic children experience more distress and worry, and also more interruptions and limitations in usual family activities, compared with a US population sample [25]. Furthermore, the parents of peanut-allergic children report significantly more disruption in the child's daily activities and in their familialsocial interactions than parents of children with rheumatological diseases [26].
It is first and foremost the children's parents who are responsible for not exposing their children to potentially dangerous dietary products, and food hypersensitivity in a child often leads to the whole family having to adapt to new food and eating practices. Yet, little is known of how child food hypersensitivity affects parents' HRQL and what aspects of the hypersensitivity condition relate to HRQL deterioration in the family. Thus the aim of the study was to investigate the parent-reported HRQL in families with a schoolchild considered to be food hypersensitive. The following research questions were addressed: • How do parents of a food hypersensitive child perceive the child's, their own and the family's HRQL?
• To what extent do allergy-associated parameters, age and gender relate to child, parent and family HRQL?
Allergy-associated parameters in this study include number of offending food items, adverse food reactions, additional hypersensitivities, allergic diseases and additional family members with food hypersensitivity.

Study population and procedure
In May 2004, a postal questionnaire (see below) was distributed to parents of 220 schoolchildren with parentreported food hypersensitivity. The schoolchildren were pupils at the nine-year compulsory school and at the upper secondary school in a municipality in the south of Stockholm, Sweden. In 2002 this group of parents had participated in a previous study on allergy-associated conditions and health-care contacts of children with exclusion diets at school [27]. The previous study included parents of 230 schoolchildren, of whom 10 children were excluded from the present study as they had left school. The sample procedure for the previous study, and thus for the present study as well, has been described in detail elsewhere [27].
In the present study, twenty of the 220 questionnaires came in return with the information that the addresses were inaccurate. After two weeks a reminder letter was sent out to all the parents of the remaining 200 schoolchildren. A total of 147 questionnaires were sent back (74% answer rate). Of the 147 questionnaires eight parents notified that their child was no longer hypersensitive to food, and another five questionnaires were omitted for missing data. Thus 134 questionnaires provide the information that the following results are based on.

Questionnaires
Two questionnaires were used: the Child Health Questionnaire -Parent-completed Form 28 (CHQ-PF28) [28] and a study-specific questionnaire.

Child Health Questionnaire -Parent-completed Form 28
The generic instrument Child Health Questionnaire -Parent-completed Form 28 (CHQ-PF28) was used to measure HRQL in a child and its family. The CHQ-PF28 consists of 28 items, which refer to 13 health scales: nine scales measure the well-being of the child, two scales measure the impact of child's health on parents' HRQL and two scales measure the impact on the family. In order to make it possible to interpret the results across the scales, the raw scale scores are transformed to standardized 0 to 100 scores by using defined algorithms, with lower scores signifying lower HRQL. Ten of the 13 scales are summed up into two comprehensive health scales representing the child's physical and psychosocial well-being (see Table 1). These summary measures are standardized so that all scores above and below 50 are above and below the average, respectively, of a general U.S. child population [28]. One scale concerning the change of the child's health during the last 12 months is not used in the present study.
CHQ-PF28 is a well-validated and reliable measure of HRQL in children from the age of 5 and normative data are available for the general U.S. population [28] and 4-13 years-olds from the Netherlands [29]. The Swedish version of the CHQ-PF50, which is a 50-item version of the CHQ, has been found to be valid and reliable in a group of Swedish children aged 9-16 years [30].

Study-specific questionnaire
A study-specific questionnaire based on relevant literature on similar subjects [2,31] was devised by the authors. It included 21 items and covered the following topics: food items avoided, adverse food reactions, additional allergy/ hypersensitivity besides food, and additional family members with food hypersensitivity. In order to get information about the prevalence of the allergic diseases asthma, eczema, and rhino conjunctivitis, the questions were formulated in a similar way as used in a multi-centred and international study on asthma and allergy, ISAAC [31]. The questionnaire can be obtained from the corresponding author on request. Prior to the data collection, a pilot test of the questionnaire was performed with nine parents, who were not included in the present study, and subsequently minor lexical adjustments were made. Some of the questions included were as follows: "Is your child allergic or hypersensitive to any of the following?" (Possible answers: furred animal, pollen, dust/ mite, food items, nickel, insects, drugs, detergents, other substances: Yes/No.) "If your child is allergic or hypersensitive to any food items, what reactions or symptoms does she/he perceive? (Possible answers: not allergic or hypersensitive, eczema, rash, eye-nose-symptoms, itchy mouth, breathing difficulties, vomiting-diarrhoea-stomach ache, allergic shock, other)." "Has anyone else in the child's family food allergy or food hypersensitivity? (Possible answers: Yes, one or both parents; Yes, one or more siblings; No, no one else in the family is food allergic/hypersensitive.) The following four questions were used to evaluate the prevalence of the allergic diseases asthma, eczema and rhino conjunctivitis. The wordings of these questions were derived from the ISAAC study [31].
"Has your child had asthma, wheezing or whistling in the chest in the past 12 months? (Yes/No)" "In the past 12 months, has your child had recurrent eczema or itchy rash for at least 6 months? (Yes/No)" "In the past 12 months, has your child had a problem with sneezing, or a runny, or a blocked nose when she/he did not have a cold? (Yes/No)" "In the past 12 months, has your child had a problem with itchy-watery eyes when she/he did not have a cold?
In the present study, celiac disease is regarded as a food hypersensitivity condition and not as an allergic disease.

Data analysis
The CHQ-PF28 and the quantitative data from the studyspecific questionnaire were analysed using the SPSS 11.0 program. CHQ-PF28 data was processed using SPSS syntax with calculations provided by the principal developer of the CHQ, Jeanne Landgraf [28]. Internal consistency of the CHQ-PF28 health scales was tested by Cronbach's alpha and ranged in this study between 0.74 and 0.86 except for the scales Mental Health (MH 0.68), General Health (GH 0.65) and Parental impact-Emotional (PE 0.65) (see Table 1).
To test differences in proportions the Chi-square test was used. The Mann-Whitney U-test was used to assess differences in CHQ-PF28 means of rank scores between two independent groups. The strength of the relationships between CHQ-PF28 scores and allergy-associated parameters were calculated using Spearman rank correlation coefficients. A p-value <0.05 was considered to be statistically significant.
Stepwise multiple linear regression analyses were performed to examine the contribution of each allergy-associated parameter, age and gender, to the predictions of CHQ-PF28 scores. The regression analyses were carried out manually as follows. Single regression analyses were executed for each variable that showed a statistically significant difference in CHQ-PF28 means of rank scores using the Mann-Whitney U-test, or a statistically significant relationship with the CHQ-PF28 scores using the Spearman rank correlation coefficients. One variable at a time was entered into the model, starting with the variable showing the highest t-value. The procedure was repeated until the addition of another independent variable did not increase the explained variation (adjusted R-square), so the final models included the variables that enhanced the degree of explanation and were statistically significant.
The statistical analyses aimed to find and asses the relationships between variables from the study-specific questionnaire (i.e. number of offending food items, type and number of adverse food reactions, food reaction in the past 12 months, number of allergic diseases, number of additional allergies/hypersensitivities besides food, additional family members with food hypersensitivity, age and gender) and HRQL as indicated by six of the CHQ-PF28 scales (i.e. PhS = Physical summary, PsS = Psychosocial summary, PE = Parental impact-Emotional, PT = Parental impact-Time, FA = Family Activities and FC = Family Cohesion). These particular CHQ-PF28 scales were chosen as they summarize the child's HRQL (PhS, PsS) and show the impact of the child's health problems on parents (PE, PT) and family (FA, FC).

Ethical considerations
The postal questionnaire that was distributed to the parents included an information letter about the purpose of the study and pointed out that participation was voluntary. The questionnaire was anonymous, i.e. no inquiries were made about personal data and the respondents could not be identified from the questionnaires.
In connection with a previous study [27] the parents agreed to be contacted again later on for further questions. The present study, as well as the previous one [27], was approved by the Director of School Administration in Tyresö municipality and ethical permissions for both studies were obtained from the research committee at Huddinge University Hospital (Dnr 122/03 and Dnr 404/ 1).

Results
The results are based exclusively on information from parents of 134 schoolchildren with parent-reported food hypersensitivity.

Allergy-associated parameters
The 134 children were reported to be hypersensitive to 1-6 food items each, with a median of two food items. The five most commonly reported food items were fruit/berries (34%), nuts (34%), peanut (33%), lactose (24%) and tomato (22%). The following offending food items were reported for 16-9% of the children: almond, egg, fish, carrot, soy, vegetables, gluten, milk and shellfish.
The children were reported to have 1-6 different types of adverse food reactions, with a median of two types of reactions per child. The most common type of food reaction reported was gastro-intestinal symptoms (50%), followed by OAS (Oral Allergy Syndrome, 47%). Of all 134 children, 45% (n = 60) had experienced at least one food reaction during the past 12 months ( Table 2). The occurrence of food reactions during the past year differed according to what kind of reaction the child was reported to have, with the highest frequency (55%) among children with gastro-intestinal symptoms and the lowest frequency (36%) among those with food-induced breathing difficulties.
At least one of the allergic diseases asthma, eczema and rhino conjunctivitis was reported for 75% (n = 100) of the children (Table 2). Forty-six per cent (n = 61) of the children suffered from more than one of these diseases. According to what kind of food reaction the child experienced or was at risk of, the occurrence of allergic disease ranged between 70% (children with gastro-intestinal symptoms) and 87% (children with breathing difficulties).
No statistically significant gender differences were found in relation to the allergy-associated parameters, except for lactose. A higher prevalence of lactose intolerance was reported for girls (31%) than for boys (16%), p < 0.05.

Health-related quality of life
Parents of food hypersensitive children with allergic diseases reported significantly lower HRQL in eight of the 14 CHQ-PF28 scales used in this study, compared with those with no allergic disease (Table 3)

HRQL and the relationship with allergy-associated parameters, age and gender
Significant rank mean score differences in six CHQ-PF28 scales (PhS, PsS, PE, PT, FA and FC) in schoolchildren with or without reported allergy-associated parameters and gender differences are shown in Table 4. There were significantly lower PhS scores reported for children with food-induced difficulty breathing and male gender, indicating that these conditions involved the experience of physical limitations for the child. Table 5 demonstrates that the larger the number of offending food items, food reactions, allergic diseases, additional allergies/hypersensitivities besides food the lower the PhS scores, indicating decreased physical well-being.
To be the only child in the family with food hypersensitivity implied better psychosocial well-being (PsS) for the child than if she/he had sibling(s) with food hypersensitivity as well (Table 4).
There were significantly lower PE scores reported by the parents to children with gastro-intestinal symptoms, indicating emotional worry or concern as a result of the child's health problems (Table 4). The higher the number of allergic diseases and additional allergies/hypersensitivities besides food, the lower the scores on the PE and PT scales, i.e. the more worry and concern (PE) and limitation in time for personal needs (PT) were reported by the parents (Table 5).
There were significantly lower FA scores reported by the respondents who had additional family members with food hypersensitivity, indicating a negative impact on family activities (Table 4). Furthermore, the higher the number of offending food items, allergic diseases and additional allergies/hypersensitivities, the lower the FA scores and the younger the child the lower the FA scores, i.e. more disruption in family activities due to the child's health (Table 5).

Allergy-associated parameters and HRQL enhancement
There were significantly higher scores on the PsS scale reported for the children who were reported to react to food with OAS, difficulty in breathing and/or anaphylaxis, indicating better psychosocial well-being for the children with these conditions (Tables 4). Moreover, the higher the number of offending food items and different types of food reactions the higher the PsS scores, indicat-ing better psychosocial well-being for children with complex food hypersensitivity (Table 5).  Finally, there were significantly higher FC scores (families' ability to get along) reported by parents of children with food-induced breathing difficulties compared with the remaining children (Table 4). The higher the number of types of food reactions, the better the families ability to get along ( Table 5).

Predictors of HRQL outcome
In order to find factors that predict impact on HRQL, stepwise multiple linear regression analyses were carried out. The six CHQ-PF28 scales were used as dependent variables. The parameters in Tables 4 and 5 shown to be statistically significant were used as independent variables.
Three variables explained 30.8% of the variance of the PhS scale: number of allergic diseases, number of additional allergies/hypersensitivities and male gender (Table 6). Thus, the higher the number of allergy-associated conditions in addition to food hypersensitivity, the lower the physical HRQL. Also, male gender predicted lower physical HRQL than female gender. Regarding the PsS scale, 9% of the variance was explained by two variables: if the child had food-induced breathing difficulties it predicted higher psychosocial HRQL, while having sibling(s) with food hypersensitivity predicted lower psychosocial HRQL for the child of interest (Table 6).
Fourteen per cent of the variances of the two scales concerning impact on parents' HRQL (PE = emotional impact and PT = time impact) were explained by two variables each. A high number of allergic diseases predicted low HRQL in both these scales. Moreover, the variable foodinduced gastro-intestinal symptoms predicted low HRQL in PE and the variable sibling(s) with food hypersensitivity predicted low HRQL in PT (Table 6).
Twelve per cent of the variance of the FA scale was explained by two variables: number of allergic diseases and additional familymember(s) with food hypersensitivity -both predicting lower family HRQL, i.e. more disruption in family activities due to the child's health. Finally, 4% of the variance of the FC scale was explained by the variable food-induced breathing difficulties, i.e. this kind of food reaction in the child predicted higher family HRQL concerning the family's ability to get along (Table  6).

Discussion
This paper focuses on parent-reported health-related quality of life (HRQL) in families with a food hypersensitive schoolchild and the relationships between allergy-associated parameters and HRQL scores.
The children in the present study were hypersensitive to 1-6 defined food items and the most common food reac-tion was gastro-intestinal symptoms. The majority (75%) of the children also suffered from allergic disease(s). The presence of allergic disease was found to be an important predictor of low child, parental and family HRQL, i.e. they were reported to have lower HRQL scores on several of the CHQ-PF28 scales compared with children with no such disease in addition to their food hypersensitivity.

Children's physical and psychosocial HRQL
The group of children suffering from allergic disease in addition to food hypersensitivity was reported to have lower physical HRQL compared with those without such disease. Also Sicherer et al have shown that atopic disease additional to food allergy has an impact on the General Health scale [25]. This is in concordance with our previous study [2] showing lower HRQL on the SF-36 scales Bodily Pain and General Health for the food hypersensitive children with additional allergic disease. This is not surprising as the allergic diseases asthma, eczema and rhino conjunctivitis to a great extent are physical disorders with somatic symptoms.
Looking at the food hypersensitive group without allergic diseases, the physical HRQL scores (CHQ-PF28) in this study did not seem low compared with a general U.S. population of children 5-18 years of age [28] and a general Netherlands population of children 4-13 years of age [29]. Thus, the present study does not suggest any physical HRQL deterioration among food hypersensitive children as long as there is no additional allergic disease.
The psychosocial HRQL scores in the present study showed no significant differences between those with and those without an allergic disease in addition to their food hypersensitivity. In comparison with the U.S. [25] and the Netherlands [29] general populations mentioned above, three of the four psychosocial HRQL scales for the child (REB, BE and MH) seem to show lower scores for the food hypersensitive children, suggesting psychosocial HRQL deterioration for food hypersensitive children whether additional allergic disease is present or not. This is in concordance with the results from our previous study [2], showing lower psychosocial HRQL (SF-36) for the hypersensitive adolescents compared with "healthy" adolescents, regardless of additional allergic disease. From this one may draw the conclusion that food hypersensitivity by itself is associated with psychosocial HRQL deterioration despite the consequences of additional allergic disease.
Avery et al [24] have shown that food allergic children can experience fear of adverse events and anxiety about eating.
In the present study more than half (55%) of the children had not experienced any food reactions in the past 12 months. Still, all the children were reported to have lower psychosocial HRQL than the general population samples mentioned [25,29]. This may suggest that it is rather the risk of food reactions and measures to avoid them that is associated with lower HRQL than the clinical reactivity induced by food intake.
It is not unusual that children's hypersensitivity will get less severe over the years and that they even can grow out of it, as they grow older. This might explain the association found between age and HRQL, i.e. the lower the age the lower the scores on the FA (Family Activities) scale. However, in the regression analysis age did not turn out to be significantly associated with any HRQL scale.
Male gender implied significantly lower physical HRQL compared with female gender but female gender involved lower scores on the Mental Health scale. This finding differ partly from the results of our previous study [2], in which the food hypersensitive adolescent females to a greater extent reported lower HRQL in both the physical and the mental SF-36 scales. This inconsistency may be due to dissimilar responding groups (parents and adolescents, respectively). It is known that parental and child reports on HRQL can differ [32][33][34]. For example, parents may have limited knowledge concerning their children's HRQL, especially their psychosocial well-being [32]. As Williams [35] suggests, girls incorporate a chronic condition with their social identity and boys tend to diminish the importance of the condition to avoid stigmatisation. Alternatively, parents may overlook more physical limitations in girls than in boys. Anyhow, these findings require further exploration.

Parental and family HRQL
Bollinger et al [36] have shown that food allergy has a significant impact on daily activities of food allergic children and their families. Moreover, Sicherer et al [25], who also used the CHQ-questionnaire, have shown that parents of food allergic children experience worry and concern (PE) and limitations in family activities (FA) due to their child's health problem. None of these studies, however, found that additional allergic disease had any significant impact on parents and families. Yet, the results from the present study indicate that additional allergic disease do have a negative impact on parental and family HRQL, i.e. on worry and concern (PE), limitation in time for personal needs (PT), and disruption in family activities (FA). When comparing our results with those of Sicherer et al different conclusions were drawn on whether additional allergic disease impairs the parents' and families' HRQL. This difference might be explained by the fact that Sicherer et al [25] in addition to their significance level of p < 0.05 also used the criterion of a 10-point difference in CHQ scores for statistical significance. In the present study, the PE-scale showed a 10-point difference and therefore fulfils also this criterion. The HRQL scores in both studies show a similar pattern and it seems likely that caring for a food hypersensitive child with somatic health problems may involve both emotional and time-consuming efforts. Still, the divergent results emphasise the need for further investigations on the significance of allergic diseases for families with food hypersensitive children.
Studies have shown that in families with a food hypersensitive child, it is common that more than one family member suffers from adverse food reactions [27,37]. In the present study, 58% of the families had more than one family member with food hypersensitivity and the regression analyses showed that having more than one family member with food hypersensitivity involved increased family strain, i.e. negative impact on the child's psychosocial HRQL (PsS), parents' time for personal needs (PT) and family activities (FA). The research on risks for the sibling of a chronically ill child is comprehensive [38][39][40][41], but the linkage between psychosocial HRQL deterioration and being more than one child in the family with a chronic disorder has, to our knowledge, not previously been reported. Although further research is needed to verify the coherence in these findings, it should be noticed that having more than one family member with food hypersensitivity is associated with HRQL deterioration.

Gastro-intestinal symptoms, breathing difficulties and HRQL
From the statistical regression analyses it can be concluded that there were two kinds of food reaction with significant impact on HRQL, i.e. gastro-intestinal symptoms and breathing difficulties. Gastro-intestinal symptoms were significantly associated with parents' worry and concern. Such hypersensitivity reactions are often diffuse and enduring and the causal connection can be hard to con-firm. Moreover, there is almost no medication existing for these kinds of symptoms. Consequently, parents may find it difficult to reduce the risk of inconvenience and to mitigate the child's pain, and this may account for the association between a child's gastro-intestinal symptoms and deterioration of parent's emotional HRQL.
In contrast to gastro-intestinal symptoms, the presence of food-induced breathing difficulties seems to improve two aspects of HRQL, i.e. the psychosocial well-being for the child (PsS) and family cohesion (FC). These somewhat puzzling results are partly in line with the findings of Rydström et al [42], who showed that Swedish children with asthma reported less impairment of quality of life in the domain of emotions than in activities. Concerning families' ability to get along, Case-Smith [43] has shown that although a child with a chronic medical condition is a strain to the whole family, this may strengthen the family cohesion. Also Sicherer et al [25] showed that family cohesion was stronger in families with peanut allergic children compared with a population sample. According to Reichenberg and Broberg [44], in families with asthmatic children the parent's ratings of family cohesion is clearly related to their child's psychological adjustment, i.e. high family cohesion is related to high scores on global self-worth for the child. In contrast to the gastro-intestinal symptoms, food induced breathing difficulties can be life threatening, the triggers are usually simple to identify and there are medical resources to treat the symptoms. Even though the risk of an acute reaction might cause worry and fear one knows what to do when it occurs. It may be speculated that individuals and families with this kind of hypersensitivity problem develop coping strategies built on cooperation and communication with those around and between family members. Also LeBovidge et al [45], who assessed parental response to children's food allergies, have highlighted the significance of parents' agreement and support from spouses and extended families. Such a hypothesis should be tested, and may generate valuable knowledge of how best to support families with food hypersensitive children.

Methodological considerations
In the area of social and behavioural sciences, the possibility to predict the variance in the dependent variable is generally low, and an explained variance of 25% is considered high [46]. In the present study, accuracy of prediction as determined by adjusted R-square values was between 30.8% and 4.1%. There are of course predictors of importance for HRQL in food hypersensitivity missing in our regression models. Two questionnaires for measuring the burden in families with a child with food allergy have recently been developed [45,47]. These questionnaires, not available when the present study was conducted, point out some issues of importance for the food hypersensitive family, such as parental frustration over others' lack of appreciation for the seriousness of the hypersensitivity condition and parental coping. By addressing such issues in a future study perhaps the percentage of the variance explained would increase. Furthermore, although statistical significances are strong, some of the results show small score differences between groups and one can discuss the question as to at what level the differences are of importance. It is also important to discuss the clinical relevance of the results and look into this issue more closely in future studies.
The choice to use the short version of the CHQ (i.e. CHQ-PF28) instead of the longer CHQ-PF50 may be a subject of debate. According to Landgraf [28], CHQ-PF28 was derived from CHQ-PF50 to be used in larger populationbased studies (N > 60). However, a Netherlands study [29] has shown that the summary measures in CHQ-PF28 and in CHQ-PF50 were comparable, but not each separated CHQ scale. A shorter questionnaire is preferable for many reasons: thus we chose to use the CHQ-PF28 and in the regression analyses the two summary measures for the children's quality of life were used, together with the scales that specifically measure the parents' and the families' HRQL. Subsequently, the results from the parents' and families' HRQL scales should be interpreted with some caution as these scales are not yet fully validated.
The scales MH, GH and PE showed Cronbach's alpha <0.70 but as they still show medium internal consistency (0.65 -0.68) they were included in the analyses. According to Nunnally and Bernstein [48] comparisons between groups can be performed with lower internal consistency (0.50-0.70) than in comparisons between individual scores. Furthermore, two of the scales (MH and GH) are included in the summary measures PsS and PhS, which are used in the main part of the analyses.

Conclusion
The results from this study are based on parent-reported HRQL in families with food hypersensitive schoolchildren. The variance in the child's physical HRQL was to a considerable extent explained by the presence of allergic disease. However, food hypersensitivity by itself was associated with deterioration of child's psychosocial HRQL, regardless of additional allergic disease. The results suggest that it is rather the risk of food reactions and measures to avoid them that are associated with lower HRQL than the clinical reactivity induced by food intake. Therefore, food hypersensitivity must be considered to have a strong psychosocial impact.