Methodology

Item Development

 

The items for the KIDSCREEN questionnaires were developed in three more or less parallel steps: literature research, expert interviews using the Delphi method, and focus groups with children and adolescents. Subsequently, the extremely extensive item pool was reduced for a first preliminary pilot version of the questionnaire using different techniques.

 

Literature review

 

First, a computer-assisted literature review was conducted with the aim of obtaining an overview of the relevant literature and existing instruments for assessing health-related quality of life in children and adolescents. Results from international as well as intercultural research were to be given special attention. The basis of this literature review was publications from the years 1985 to 2000, which were recorded in the MEDLINE database. A total of 9,029 abstracts met the previously defined search criteria. All publications were subjected to an internal project review, with the help of which relevant dimensions of health-related quality of life in children and adolescents were to be identified. 

 

Delphi-Study

 

What is the current state of quality of life research in children and adolescents? According to experts, which concepts and operationalizations are appropriate when measuring health-related quality of life in children and adolescents? To answer these and other questions, an international Delphi survey was conducted. The Delphi process included three waves of interviews involving a multidisciplinary group of 24 experts from 7 European countries. A consensus was reached that the new instrument should be a multidimensional profile with 30-49 items in 5 to 8 dimensions. The processing time should not exceed 10-15 minutes. The experts agreed on 8 specific dimensions to be covered by the questionnaire: psychological (well-being, self-esteem, body image, cognitive functions), physical (mobility, energy/vitality), social (social relationships, family/home) and other aspects (e.g. environment). 

 

 

Interviews mit Focus groups

 

In addition to expert opinion, the children's and adolescents' view of their health and quality of life played a major role in the construction of the KIDSCREEN questionnaires. What do children and adolescents imagine as a good life? What role does health play in this? What is important for children to feel good about themselves?

 

These and other questions were discussed within the focus groups, which were conducted in all participating countries. The focus groups, each with 4 to 6 children or adolescents, were divided by gender and age. In addition, focus groups were also conducted with parents and caregivers, as well as more in-depth interviews with individual children. To ensure comparability of results, the process of the focus group discussions was standardized to the greatest extent possible. Subsequently, the recorded discussions were transcribed and subjected to content analysis. The results of the focus group work were original statements by children and adolescents on health, well-being and quality of life, which could be classified into several dimensions and subcategories in terms of content analysis. In addition, the layout and response categories of existing quality-of-life questionnaires for children and adolescents were evaluated with the help of the participants.

 

Item Reduction

 

A total of 2505 statements on health-related quality of life in children and adolescents were formulated on the basis of the focus group discussions. This item pool was checked for redundancies and further reduced using a card sort method, so that in the end 185 items in 7 dimensions and 24 categories were included in the upcoming pilot test. The preliminary pilot questionnaire used a five-point response scale (frequency or intensity) and asked about the subjective well-being of the children and adolescents during the last week.

Translation and Harmonization

 

The participating project countries translated the items obtained with the help of the focus groups into the respective national language. Each country had to strictly adhere to a previously developed standardized translation methodology. This translation procedure followed international cross-cultural translation guidelines (WHOQol Group, 1993; IQOLA, 1991) and was intended to ensure cross-cultural harmonization of the items. The first step consisted of the actual translation. The original English items were translated into the respective national language by two independent translators (two national forward translations). Subsequently, both translations were compared item by item and aggregated to one version (national reconciled forward translations). This was then translated back into English (national backward translation) in order to compare it with the original English item. These comparisons were then used to refine the final national translations of the pilot questionnaire into the respective target languages (national final forward translation).

 

In order to be able to guarantee a cross-culturally uniform translation, the conceptual equivalence of the respective national translations was checked on an international basis. For this purpose, a telephone conference was held to resolve inadequate translations and discrepancies between alternative versions. The acceptance and applicability of the pilot questionnaire was subsequently tested by means of pretests and cognitive interviews in the respective countries. As a result of these tests, it was shown that children aged 8 years and older could read, understand and answer the items without any problems. The final formulation and structure of the pilot items was agreed upon during a meeting attended by all project partners, which also included the results of the national pre-tests.

Pilot Study

 

Study Design

 

The KIDSCREEN pilot questionnaire consisted of 159 items for children and 167 items for adolescents. In addition, there was a question on general health. These items formed the following 7 dimensions, which formed the theoretical KIDSCREEN Quality of Life model: Physical Well-Being (mobility, energy and relaxation, health and discomfort), Self Perception (body image, self-esteem, and self-worth), Psychological Well-Being (positive and negative feelings, worry and stress, life satisfaction, and optimism), Social Skills (autonomy, financial opportunities), Social Relationships (home and parents, friends, acceptance, and social support), Cognitive and School Skills (concentration and learning, school performance, and teachers), Personal Environment (environmental and surrounding conditions).

 

The pilot questionnaire was supplemented with items on socioeconomic status and health. The additional questions were taken from relevant international studies or selected based on recommendations or existing questionnaires. Adolescents also answered the Youth Quality of Life Instrument, YQOL (Patrick et al., 2002), and questions on the use of health care facilities. The order of the individual questions was the same in all countries.

 

With the exception of the Netherlands, which conducted a purely postal survey, data collection for the national pilot studies took place in public schools. Children and adolescents aged 8 to 18 years and their parents participated. The pilot sample included participants from different regions (metropolitan vs. small-town/rural) as well as different social backgrounds. The children and adolescents completed the questionnaires in the schools, where they were subsequently collected by the investigators and taken to the respective study center. In addition, one parent, either mother or father, answered the questionnaire at home and returned it to the respective study center free of charge. Special coding enabled the study centers to later merge parent and child questionnaires into one data set.

 

In the Netherlands, a representative sample was drawn based on data from the Leiden Regional Health Authority (GGD). The questionnaires were then sent to the children and their parents with a request to complete and return them independently. 

 

Sample

 

The international sample of the KIDSCREEN pilot test, which aimed to reduce the KIDSCREEN item pool, included data from 3,977 completed child and adolescent questionnaires. Between 48.1% and 89.3% of families contacted (cover letter and request for participation) agreed to participate in the study. The sample consisted of 1,437 (36.1%) children aged 8-11 years and 2,469 (62.1%) youth aged 12-18 years. The average age was 12.7 years with a standard deviation of 2.6 years. 1,942 (48.8%) of the children and adolescents were female, and 2,019 (50.8%) were male. The age and gender distribution was balanced and comparable between countries. Despite the similarity of data collection in all countries, there were large differences in response rates between countries, ranging from 48.1% (Germany) to 89.3% (Switzerland). The overall response rate was 56.7%. The response rate of the parent questionnaires varied even more compared to that of the children, ranging from 23.6% (United Kingdom) to 99.4% (Netherlands). 

 

Analysis

 

The statistical analyses performed to reduce the item pool can be divided into 4 different stages:

1. Item reduction using Item Response Theory (IRT)
2. Item reduction using Classiscal Test Theory (CTT)
3. Comparison of the results of both methods against the background of theoretical considerations and subsequent reduction of the combined version using Item Response Theory methods.
4. Improvement of predictive validity (validity for different age and gender groups and in the respective countries).

This multi-stage reduction process led to the construction of the multidimensional KIDSCREEN-52 research questionnaire, which consisted of 52 items in 10 dimensions.

Survey

 

The aim of the KIDSCREEN project was not only to develop a new instrument for assessing health-related quality of life in children and adolescents, but also to describe the relationships between the data collected with the help of the KIDSCREEN questionnaire and other health-related parameters. In addition to age and gender, the general health status as well as the mental health of the child or adolescent, the parent-child relationship, social support, the socioeconomic status of the family and the parental health status are frequently mentioned in the literature. In addition, the child's or adolescent's health-related quality of life has an impact on risk behaviors and health care utilization. The above determinants were included in the survey questionnaire and served to validate the KIDSCREEN questionnaire.

 

Instruments and Questionnaires of the KIDSCREEN Survey

 

The following instruments and questionnaires were used in the KIDSCREEN survey, which included determinants and variables relevant to validation and sample description:

• SES (sociodemographic and socioeconomic status)
• Health status (CSHCN Screener)
• Psychosomatic complaints (HBSC Symptom Checklist)
• Mental health (SDQ, Strength and Difficulties Questionnaire)
• Risk behavior (Smoking, Alcohol)
• Utilization of health Care Services (EUROHIS)
• Parent-child relationship and social support (SAS Social Support Scale, Oslo 3-item Social Support Scale)
• Emotional well-being (Items from CHQ, KINDL and SF-36)
• Cross-disease questions from the DISABKIDS questionnaire to assess the quality of life of children with specific chronic diseases
• YQOL (Youth Quality of Life Instrument)
• CHIP (Child Health and Illness Profile)
• CHQ PF 28 (Child Health Questionnaire Parent Form)
• Country-specific generic questionnaires (e.g. KINDL in AT, CH, DE; Cantril's Ladder in CZ, HU, PL)
• Parental health status (SF-12 Short Form).

 

Sampling and Implementation

 

The survey phase of the KIDSCREEN project aimed to collect reference data for health-related quality of life in all participating countries using nationally representative surveys.

 

Two different approaches to sampling were followed:

Six countries (Austria, Switzerland, Germany, Spain, France, Netherlands) initially contacted potential participants by telephone, in five countries initial contact was made through schools (Greece, Hungary, Ireland, Poland, Sweden), in the United Kingdom both methods were combined, and in the Czech Republic contact was made through home visits.

In Ireland, a preliminary study was conducted in two Health Board Areas and a national representative survey was planned at a later stage. For this reason, the data from Ireland could not be fully included in the international analyses.

 

In both approaches --- sampling by telephone and in schools --- a second questionnaire was sent to a sub-sample two to four weeks after the first survey. This questionnaire was used to check the retest reliability.

 

A total of 21,804 children and adolescents aged 8 to 18 participated in the KIDSCREEN survey. The age and gender distribution reflected the ratios of the underlying reference population. Compared to the reference population (Eurostat), mothers and fathers in the KIDSCREEN sample achieved a higher level of education. Similar results were found when comparing participants and nonrespondents. Participants in the KIDSCREEN survey were more likely to have a higher level of education, to rate their health better, to be married, and to live in cities. These results are consistent with other European studies.

Implementation

 

The members of the KIDSCREEN group have developed a comprehensive and sustainable set of activities to support the implementation of the KIDSCREEN questionnaires on a national and international level.  

These activities include the

• presentation of the instrument and first results at national and international conferences,
• publication of scientific articles and
• creation and publication of an international manual.

To gather further expertise, the KIDSCREEN questionnaires were made available relatively early on for various research purposes (e.g., clinical trials, evaluation studies, epidemiological studies, health surveys).