OHRQoL measures have been the target of investigations in the oral healthcare field and have proven valuable in assessing oral health needs, especially among the adult population . In recent years, OHRQoL measures have been designed for the child population . However, these instruments are not yet available in all countries or languages. Most questionnaires have been drafted in English-speaking countries and adapted for use in other countries . The translation and the testing of psychometric properties are important steps to ensuring the quality of a cross-cultural adaptation of an OHRQoL measure . Considering the differences between social, cultural and economic aspects, the availability of cross-culturally valid, multi-lingual versions of instruments is important to obtaining reliable, comparable data .
The present study determined the reliability and validity of the ECOHIS, which is a multidimensional assessment tool for measuring the negative impact of oral problems on quality of life among preschool children (0-5 years of age) that has been cross-culturally adapted to Brazilian Portuguese . In the first phase of the present study, both versions of B-ECOHIS (a version including the original item "missing preschool, daycare or school" and another replacing this item with "had difficulty doing daily activities") were tested with parents/caregivers of children who attended school and those who did not attend school. No differences in psychometric properties between versions were found. It is likely that the responses of the parents/caregivers of children who did not attend school/daycare for the item 'missed preschool, daycare or school' was 'never', which did not interfere in the psychometric analysis of the instrument. Therefore, the B-ECOHIS was maintained as originally drafted by Pahel et al. .
Assessment instruments should be reproducible over time, that is, they should produce similar results on two or more administrations to the same individual, provided that the general clinical state has not been altered. The analysis of test-retest reliability suggests the adequate stability of the instrument. The seven-day interval between interviews was important to diminishing the probability of systemic alterations in the clinical condition of the patient. It is recommended that the interval between measurements be long enough to reduce the effects of memory and short enough to diminish the likelihood of systemic alterations. Although the definition of this interval is arbitrary, a period of two to 14 days is considered adequate [21–23]. The B-ECOHIS was answered twice and demonstrated good stability during the preliminary and the field studies.
Internal consistency was calculated using Cronbach's Alpha Coefficient for the child and family impact sections and demonstrated adequate homogeneity (α ≥ 0.70) in the field study. As in the original version (α = 0.91 for CIS; α = 0.95 for FIS) , French version (α = 0.79 for CIS and FIS) , Chinese version (α = 0.87 for CIS; α = 0.85 for FIS)  and Farsi version (α = 0.89 for CIS; α = 0.85 for FIS) , the final version of B-ECOHIS performed reliably. One point that shoud be addressed regards the values of Cronbach's alpha in the preliminary study: 0.74 for the child section and 0.59 for the family section. Cronbach's alpha for the family section was low. Aspects such as the number of items on this section and sample size could have influenced this result. In the field study however, Cronbach's alpha for the family section was 0.76.
Correlations were obtained between the scores of the instrument and global measures of overall and oral health as well as on the child and family sections of the B-ECOHIS. The same was found with the original version of the ECOHIS . Measures concerning quality of life components should reflect the values of the subjects. However, lay people are sometimes asked to fill out questionnaires that do not reflect their real concerns, but rather the values of physicians, social scientists or other experts . We believe that items on the questionnaire concerning aspects of daily life could often reveal experiences that are relatively less important to the target population . This may explain the weak correlations found for convergent validity. In the dental field, it is common for lay people to consider the mouth as a focus unlinked to the individual as a whole, which could influence the comprehension of the relationship between oral health and quality of life.
Discriminant validity analysis is considered a useful method in the differentiation of groups that are known to be distinct [5, 18]. In the present study, the occurrence of oral problems implied limitations and difficulties, confirming the hypothesis formulated with regard to the construct validity of the B-ECOHIS: 1- individuals with one or more decayed, missing and/or filled teeth (dmft) obtained higher scores on the B-ECOHIS (indicating worse OHRQoL) than children without dental disease; 2- parents of children with dental disease and/or dental treatment experience obtained higher B-ECOHIS scores than those of children free of dental disease experience; 3- among children with dental caries experience, those with more severe dental disease obtained higher B-ECOHIS scores than those without dental caries and those with less severe dental disease; 4- children with discolored upper anterior teeth obtained higher B-ECOHIS scores than those without discolored upper anterior teeth.
Analyzing the distribution of items in both phases of the study, the most frequently reported items on the two sections of the scale were practically the same as those reported in previous validity studies carried out in Quebec, Canada  and Hong Kong, China . On the child impact section, the most prevalent items were related to "pain", "eating" and "irritation". On the family impact section, the most prevalent item was "been upset".
As done in the original version, the number and distribution of "I don't know" responses were taken into account. In the field study, 11.9% of subjects answered "I don't know" to one or more questions, which is somewhat higher than the 7% reported for the original ECOHIS , but similar to that found in the French ECOHIS . An "I don't know" response option is essential in studies that assess the participants' perceptions of health or quality of life of another individual, as it reflects a particular characteristic of the phenomenon under evaluation . Considering the management of this response option, Jokovic et al.  proposes the following: 1- exclude subjects with such responses; 2- use adjusted scores; or 3- drop items from the questionnaire that have high proportion of "I don't know" responses. As the present study was conducted on a large sample (population-based survey) and the proportion of such responses was low, the decision was made to exclude subjects with ≥ 2 and ≥ 1 "I don't Know" responses on the child impact and family impact sections, respectively (2.7% and 1.3% of total sample). Jokovic and colleagues  found that excluding subjects or using adjusted scores did not affect the validity analyses.
The present study has particular characteristics that should be recognized. In the preliminary and field studies most of the families were from less privileged economic classes. It should be pointed out that parents' perceptions regarding their child's oral health could be influenced by socioeconomic conditions . A recent Brazilian study confirmed that parents who earned a lower income were more likely to rate their child's oral health as 'poor' . We believe that this point does not affect the results of this validation study, as the prevalence of the negative impact was not high. Both studies were performed with different sample populations. The first was developed with a convenience sample, like the studies carried out in the United States of America, Canada, China and Iran [5, 17, 19, 24], whereas the field study was nested in an epidemiological survey. A large percentage of validation studies carried out in the field of dentistry use convenience samples. However, it is recognized that a large sample size leads to more accurate parameter estimates, which leads to a greater ability to meet the aims of the study . Thus, the decision was made in the present investigation to employ a population-based study with a large sample size in order to produce more accurate results. This strategy has been used in other fields of knowledge, such as medicine. The age of the children, place of recruitment and mode of administration of the B-ECOHIS were also different. Regarding the self-administration versus interviewer-administration of a measure, a number of studies have demonstrated that the mode of administration does not affect the performance of the measure [29, 30]. Moreover, the measure was administered in two different states in southeastern Brazil. Despite the methodological differences, the psychometric properties of the B-ECOHIS proved similar in both phases. This offers further evidence of the validity and reliability of the measure. However, it should be pointed out that, due to the cultural diversity of Brazil, using the B-ECOHIS in other regions may require some adaptation and such cases should involve further psychometric testing.