Cross-cultural adaptation and psychometric evaluation of the early childhood oral health impact scale (ECOHIS) in chilean population

Oral diseases are highly prevalent in children worldwide despite the improvement in oral health indices initiated in the last decades [1‐3]. It is well known that their consequences on children are serious and can affect their quality of life [4‐8]. Early childhood caries continues to be a serious public health problem in Chile, with a prevalence that can reach 80% at 4 years of age [9‐11]. Oral Health-Related Quality of Life (OHRQoL), together with clinical indicators, can jointly provide a more comprehensive assessment of the patient’s oral health [12]. The OHRQoL has been defined as a multidimensional concept which includes a subjective evaluation of the individual’s oral health, functional well-being, expectations and satisfaction with care, and their sense of self [12].

The knowledge of the OHRQoL might help to improve the development of effective oral health programs and services because it permits the assessment of young children’s perceived needs, and treatment strategy effectiveness [13]. This can contribute to the identification of groups with a higher level of need, to prioritize public health programs for care of children and adolescents, and to improve access to care [14]. The use of OHRQoL as an outcome measure is consistent with patient-centered care, being crucial in understanding the effectiveness of treatment from the patients’ perspective [12].

Several instruments have been developed to assess the OHRQoL, yet few of them have been specifically designed for preschoolers. The first OHRQoL questionnaire for this age group was the Michigan Oral Health-Related Quality of Life (Michigan OHRQoL) in 2003 [15]. Subsequently, the Early Childhood Oral Health Impact Scale (ECOHIS) was developed in 2007 [16], the Pediatric Oral Health-Related Quality of Life (POQL) in 2011 [17] and the Scale of Oral Health Outcomes for 5-year-olds (SOHO-5) in 2012 [18]. POQL and the ECOHIS measure the OHRQoL impact of dental diseases not only on the children, but also on their families. It is important because oral health conditions have an indirect impact on parents and family members, because they result in lost workdays or in having to spend time and money on dental care [19, 20]. The ECOHIS demonstrated high reliability [21, 22], good validity [23, 24] and responsiveness [25, 26], and it has been adapted into about 10 languages and countries [21‐24, 27‐31], including Spanish for Argentina [32].

Culture is an important factor that can influence a person’s activities, thinking and behavior. As countries differ regarding public health strategies, attitudes, socioeconomic conditions and other factors, the expression of their culture can change between populations [33], and instruments to measure Health Related Quality of life (HRQoL) should go through a cultural adaptation process before being used in a different country. Therefore, even among Spanish speaking countries it is usual to develop country-specific versions of instruments measuring HRQoL [34‐36]. Even when the translation is performed with great precision, cultural factors may not be accurately conveyed. In order to study the health care needs of people with diverse cultural backgrounds, research instruments must be reliable and valid in each culture studied [37].

The aim of this study was to develop a Chilean Spanish version of the Early Childhood Oral Health Impact Scale (ECOHIS) that is conceptually equivalent to the original and to assess the acceptability, reliability and validity of this version in the preschool population of Chile.

The study was performed in two phases. In the first phase, the scale was translated into Spanish and adapted to the Chilean culture. In the second phase, the psychometric properties were tested among a sample of parents of preschool children. The Ethics Committee of the Universidad de La Frontera approved the study protocol (resolution n° 061/2015).

We used a standard cross-cultural adaptation process to develop the Chilean version of the ECOHIS, which demonstrated good acceptability by parents; high reliability and good construct validity. The results are consistent with those obtained for the original ECOHIS and suggest that the Chilean version is conceptually and metrically equivalent.

“Don’t know” and/or missing responses may reflect comprehensibility problems [24]. In our sample, only one parent left some missing items and only 19 (6.3%) responded “Don’t know”, similarly to the original ECOHIS study (7%) [16]. However, other studies have shown higher “Don’t know” percentages [19‐22]. The low percentage of “Don’t know” supports that the mode of administration (proxy-report) is not a limitation for the ECOHIS Chilean version. According to the ECOHIS proxy-report design [16], in our study most parents completed it during the school meeting, and those who did not attend it completed the questionnaire at home. No interview administration was needed, and no one required assistance to self-complete the questionnaire. Self-administration presents advantages, such as lower cost, preservation of participant’s anonymity, and reduction of interviewer bias [48]. Furthermore, studies with other OHRQoL instruments showed that administration mode (interview versus self-administered) does not influence the instruments’ scores [48‐50]. On the other hand, evidence shows that parents underestimate the impact of children’s oral health problems, since they have a different perspective and limited knowledge, particularly related to social and emotional well-being [51]. Indeed, oral health problems directly observable by parents, such as physical complaints and functionality, concur better with children’s perceptions [52, 53]. However, in this age group due to their cognitive immaturity, limited social experience and continued dependency, parents are the best source of their child’s oral health [54]. As in the original version, we included parents with “Don’t know” responses in the analysis because a “Don’t know” response reflects an essential characteristic of the phenomenon under evaluation, rather than errors by the respondents [55].

The high floor effect observed in the total score (41.6%) and domain or section scores (ranging 49–92%) is congruent with the clinical characteristics of our participants, since over 40% of the sample was free of oral conditions. Although these results are similar to those obtained in other studies, which have also shown a strong floor effect for ECOHIS total score (ranging 20–54%) [16, 24, 29] they could indicate a limitation of the instrument. The ECOHIS Chilean version needs to be tested in a population with more oral problems to assess adequately the instrument’s more severe response levels. The ECOHIS has shown an excellent reliability, both in its internal consistency and its reproducibility, since its coefficient values were over 0.8 allowing to use its scores for the comparison between groups [56]. Our result of internal consistency for the child section (Cronbach’s alpha coefficient = 0.88) was similar to the 0.91 reported by the original English questionnaire, but it was lower for the family impact section (0.80 vs. 0.95). However, with exception of the original version, the family section usually shows a lower internal consistency (Cronbach’s alpha ranging 0.59–0.85) than the child impact section (ranging 0.74–0.92) [21, 22, 29], which may be due to the lower number of items rather than a lower consistency. In the test-retest reliability, the ICC for total score was the same as reported in the original version (0.84), but lower than reported in the French (0.95) [24] and Brazilian versions (ranging 0.94–0.99) [22, 57]. Despite this, the ICC value shows that the Chilean version of ECOHIS has an excellent test-retest reliability in which it is able to produce reproducible scores when it is administered at two different times [43].

The good results on equivalence with the original ECOHIS shown by its comparison with the back-translation of the Chilean-adapted version support the content validity of this new country version. The higher difficulty of the back-translation compared to the forward one, observed in our adaptation process, has been also described for other adapted instruments [58, 59]. As the first translation seeks conceptual equivalence, and the second one seeks a literal translation of the expressions, this back-translation can often be harder to carry out.

To the best of our knowledge, there is no previous publication describing the factor structure of the ECOHIS. Our results confirm the two- section structure proposed by the developers (child and family impact sections), as well as that correlations between them can be explained by the second order model representing the global OHRQoL. The confirmation of this measurement model in other country versions of the ECOHIS would be recommendable.

For construct validity, the Chilean version of the ECOHIS scale showed significant differences among groups defined by the children’s dental health status as reported by parents. These findings were consistent with previous studies where parents who perceived their child’s oral health as poor had significantly higher mean ECOHIS scores [16, 21, 24, 27]. Our results showed higher ECOHIS scores among those with more than 6 decayed teeth, compared to those who had 1–5 decayed teeth or to those who were caries-free. The large effect size in children with poor oral health status reported by parents and who have high severity of caries supports the parents’ recognition of oral health problems when they become evident, or when it manifests in the form of pain [60]. However, the ECOHIS was not able to discriminate presence or absence of malocclusion or type of traumatic dental injuries. Although the ECOHIS was originally developed to assess the impact of dental caries, it has been widely used to evaluate several oral pathologies [4, 61], but only few studies have validated this application: Peker et al. only found a moderate correlation with gingival index [21], and Scarpelli et al. showed a statistically significant association with discolored upper anterior teeth [22]. This is important because the ECOHIS has been used to measure OHRQoL in patients with traumatic dental injuries or malocclusion, not detecting any impact on the children [4, 62]. Further research is needed to explore whether this absence of impact can be due to the inability of the instrument to discriminate between certain degrees of these pathologies.

The poor correlation between ECOHIS and PedsQL™4.0 Generic Core scale suggests that ECOHIS captures additional information, which is not covered by instruments measuring HRQoL. This is in line with results reported by Lee et al., showing that the ECOHIS is more sensitive than PedsQL™4.0 measuring the impact of oral problems on preschool children [63]. As expected, a high correlation was found with the Oral Health scale of PedsQL™4.0 because it also could be considered specific for measuring OHRQoL [64, 65]. The moderate correlation between the child and the family impact sections of the scale found in our sample (r = 0.57) was similar to results reported in previous studies ranging 0.36–0.68 [16, 21, 27]. The correlation in the original English questionnaire between both sections was the lowest (Spearman’s r = 0.36, p ≤ 0.001), and the Turkish version the strongest (Spearman’s r = 0.68, p ≤ 0.001). Although child and family sections assess different aspects of child’s OHRQoL, both sections are related with the underlying construct.

The main limitation of this study was the homogeneity of the sample studied, since only preschoolers from public schools were included. Nonetheless, our sample is representative of children between 2 and 5 years old attending public preschools, and these children are the main target of Oral Health Policies in Chile. Another limitation was that information regarding the parents, such as age, gender, and educational level, was not registered. Finally, the responsiveness was not assessed; therefore, future studies are necessary to evaluate the capacity of the ECOHIS Chilean version to detect changes over time in a clinical or public health intervention.

The Chilean version of the Early Childhood Oral Health Impact Scale was valid and reliable for assessing the OHRQoL in preschool children through proxy. The comparison with the original U.S. version shows similar results in reliability and validity, suggesting that the cross-cultural adaptation method followed has yielded an equivalent Chilean version.

Researchers and clinicians now have at their disposal an OHRQoL instrument for use in Chilean preschool children to assess the impact of oral disorders on them and their families, and also to facilitate the identification of groups at a higher risk of dental health inequity to improve their access to oral health care services.