Determinants of orthodontic treatment need and demand: a cross-sectional path model study

Transcription

1 European Journal of Orthodontics, 2017, doi: /ejo/cjw020 Advance Access publication 15 March 2016 Original article Determinants of orthodontic treatment need and demand: a cross-sectional path model study Jari Taghavi Bayat 1, Jan Huggare 1, Bengt Mohlin 2 and Nazar Akrami 3 1 Division of Orthodontics, Department of Dental Medicine, Karolinska Institutet, Huddinge, Sweden, 2 Department of Orthodontics, Institute of Odontology, Sahlgrenska Academy, University of Gothenburg, Sweden and 3 Department of Psychology, Uppsala University, Sweden Correspondence to: Jari Taghavi Bayat, Division of Orthodontics, Department of Dental Medicine, Karolinska Institutet, SE Huddinge, Sweden. jari.taghavi@ki.se Summary Objectives: To put forward a model predicting orthodontic treatment need and demand. Furthermore, to explore how much of the variance in treatment demand could be explained by a set of self-assessed measures, and how these measures relate to professionally assessed treatment need. Subjects and methods: One hundred and fifty adolescents, aged 13 years, completed a questionnaire which included a set of self-assessed measures dealing with self-esteem, such as dental and global self-esteem, various aspects of malocclusion, such as perceived malocclusion and perceived functional limitation, and treatment demand. Treatment need was assessed by Dental Health Component of the Index of Orthodontic Treatment Need grading. Path analysis was used to examine the relations between the measures and if they could predict treatment need and demand. Results: The measures proved to be reliable and inter-correlated. Path analysis revealed that the proposed model had good fit to the data, providing a test of the unique effect of all included measures on treatment need and demand. The model explained 33% of the variance in treatment demand and 22% of the variance in treatment need. Limitations: The specific age group could affect the generalizability of the findings. Moreover, although showing good fit to data, the final model is based on a combination of theoretical reasoning and semi-explorative approach. Conclusions: The proposed model displays the unique effect of each included measure on treatment need and demand, explaining a large proportion of the variance in perceived treatment demand and professionally assessed treatment need. The model would hopefully lead to improved and more cost-efficient predictions of treatment need and demand. Introduction Deviation from an ideal occlusion, known as malocclusion, is common in a large proportion of children and adolescents (1 6). Many of these deviations are within the range of what is to be considered as normal biologic variation. Some deviations however may have negative influence on dentofacial development, contributing to impaired orofacial function (7) and/or tooth injuries (8 10). Apart from physical consequences locally, malocclusions could have a negative effect on patient s psychological well-being and quality of life including self-esteem and self-image (11 14). For adolescents and young adults, malocclusion can become a burden, leading to avoiding strategies to minimize the negative feelings associated with the condition (15). The main reason for seeking orthodontic treatment is dissatisfaction with dental aesthetics (16 18). For adolescents, the strongest motivation to undergo treatment seems to be the urge to The Author Published by Oxford University Press on behalf of the European Orthodontic Society. All rights reserved. For permissions, please journals.permissions@oup.com 85

2 86 European Journal of Orthodontics, 2017, Vol. 39, No. 1 fulfil existing social norms regarding dental aesthetics (19). Thus, it could be argued that apart from correcting malocclusions that could constitute a health risk in the oral environment (7, 17, 20), benefits from orthodontic treatment are mainly psychosocial (21). Therefore, it seems important to explore the psychosocial aspects of malocclusion and self-perceived need for treatment more widely. Understanding the role of psychosocial factors would in the long run enable improved assessments of treatment need. While previous research has been sufficient in identifying key psychological and social measures or variables related to perceived malocclusion and treatment demand, there are to our knowledge no studies on how these predictors perform in combination and if they can predict treatment demand or the outcome of professional assessment. For example, do any of these variables contribute uniquely in explaining treatment demand? Are they related to the treatment need assessed by professionals? To answer these questions, we examined the relations between core psychological and social measures (dental and global self-esteem and social influence) and malocclusion related measures (perceived malocclusion, functional limitation and prioritizing healthy and straight teeth) on one hand, and self-perceived treatment need (treatment demand) on the other. We also explored how these measures correlate with professionally assessed need of treatment (treatment need). The measures outlined above were chosen on basis of previous research. For example, the measures for dental self-esteem and social influence were based on findings from a preceding qualitative study on how adolescents are affected by their malocclusion in daily life (15). Other sources of guiding knowledge were the systematic review presented by The Swedish Council on Technology Assessment in Health Care (17) and closely related research (11, 12, 22 26). Thus, the aims of the present study were to explore how much of the variance in treatment demand is explained by a set of selfassessed measures dealing with self-esteem and different aspects of malocclusion (questionnaire based) and how these variables relate to professionally assessed treatment need (based on data from dental records). More importantly, to put forward a model predicting professionally assessed orthodontic treatment need and self-perceived treatment demand. Subjects and methods Design and sample size This was a cross-sectional study where participants were randomly selected from a subpopulation of Swedish youths. Assuming a true correlation of 0.20 to be detected (average effect size in social psychological research), with a statistical power of 0.80 and a probability of Type I error (α) of 0.05, a sample size of 150 individuals was required (27). Also research suggest that the point of stability of a correlation is reached somewhere around a sample size of 150 (28). Ethical considerations All participants and their parents received written information about the study. The study protocol and informed consent procedure was approved by the Research Ethical Committee in Stockholm, Sweden (D.nr 2009/5:4 and 2014/ ). Subjects This research was conducted in Uppsala Municipality (> inhabitants) during the period of January to April Participants were recruited randomly from the population registry. The randomization process was performed by a third party (Evry Sweden, a consulting company). Initially 240 adolescents, aged 13 years, received written information about the study and were asked to participate by completing a questionnaire either online or in paper format. The adolescents were asked to fill in the questionnaire on their own, without assistance of their parents. Approximately 2 weeks after the first mailing a reminder was sent out. Due to insufficient response rate at 2 weeks after the reminder (92 responses, 38.3%), the procedure was repeated within a month including additionally 100 randomly selected adolescents. By the end of April 2015, 162 respondents had been listed. Because of incomplete responses, the final data set consisted of 150 participants. Measures The measures were selected from a questionnaire developed during , based on the findings from a qualitative study dealing with the impact of malocclusion on the daily life of adolescents (15). They were further weighed against the outcomes of previous methodological and quantitative studies (11, 12, 22 26). After theoretical processing, the language was adapted to suit the age group by a panel consisting of one child psychiatrist, one psychologist, and three orthodontists, and tested in minor pilot-studies concerning language, comprehension and the relevance of the items. The early pilot included six adolescents aged between 13 and 15. Two of the informants (aged 13) were then interviewed and asked to comment each question further and mark their relevance. The later pilot study was performed on nine individuals including a panel of experts (n = 4), dental personnel at an orthodontic clinic (n = 3) and young adults who had orthodontic treatment in their youth (n = 2, age = 30+). The questionnaire was then evaluated based on incoming comments and some minor adjustments were done. The measures used in this article are presented and described below. Basic statistics and reliability indices of the measures are presented in Table 1. Response to each statement was indicated on a five-point Likert like response scale ranging from 0 (Do not agree at all) to 4 (Agree fully). Reversed coded items were recoded and an index was created by averaging the responses across all items within each measure (scale). Dental self-esteem was measured by 8 items (3 reverse coded), item example: I am proud of (the appearance of) my teeth and I feel less unattractive because of my teeth (reverse coded). Higher scores indicating higher dental self-esteem. Global self-esteem was measured by 12 items (4 reverse coded) modified to fit the age group in the present study (29). Item example: I am happy being me and Sometimes I feel like I am not good enough (reversed coded). Higher scores indicate higher global self-esteem. Social influence was measured by 12 items (3 reverse coded) and included items measuring different aspects of influence (e.g., media, peers), item example: I am worried that people would comment on my teeth, To see advertising images with beautiful teeth do not bother me at all (reverse coded), and Sometimes I feel like an outsider because of my teeth. Higher scores indicate higher burden due to negative social influence. Perceived malocclusion was measured by 8 items (one reverse coded) taping a range of occlusal status, item example: My teeth are straight (reverse coded), I have crooked teeth and I have protruding front teeth (large overjet). Higher scores on the scale indicate higher perceived malocclusion. Perceived functional limitation was measured by 9 items that taped various aspects of limited oral function, item example: I have difficulties chewing, I feel fatigue in my jaw muscles and I bit

3 J. Taghavi Bayat et al. 87 Table 1. Basic statistics for the questionnaire based (Q) and dental record based measures (R). Measure M SD No. of items α Inter-item r Psychological and social (Q) Dental self-esteem Global self-esteem Social influence Malocclusion related (Q) Perceived malocclusion Perceived functional limitation Prioritizing healthy and straight teeth Treatment demand and need Treatment demand (Q) Treatment need (R) M, mean item score; SD, standard deviation within item scores; α, internal consistency reliability; inter-item r, average correlation among the items within each measure. N varies between 148 and 150 as a function of outliers. Response range for all scale scores ranged between 0 and 4 except for treatment need ranging from 1 to 5 (DHC grading). myself in the palate (gum tissue) when I bite together. Higher scores indicate higher perceived functional limitation. Prioritizing healthy and straight teeth was measured by three items (1 reverse coded), item example: It is more important to have healthy teeth than white teeth and Having white teeth is more important to me than having straight teeth (reverse coded). A higher value indicates higher prioritizing of healthy and straight teeth appose to white teeth. Treatment demand was measured by 11 items (1 reverse coded), measuring demand for orthodontic treatment, item example: I have longed for braces for a long time, I would like to have braces even though it could be painful initially, and I do not want to have braces if I have to pull teeth (reverse coded). Higher scores indicate higher treatment demand. Treatment need was assessed by the Dental Health Component of the Index of Orthodontic Treatment Need, IOTN-DHC (30). The index was developed in United Kingdom at the end of 1980s and is based on the Swedish Dental Board index of treatment priority. It was initially put forward to improve communication concerning orthodontic treatment need between dentists and specialists in orthodontics. IOTN is primarily a morphological index, consisting of two components: aesthetic component (AC) and dental health component (DHC). The DHC deals with different occlusal traits and is categorized into five grades: 1 5. The IOTN was considered since it is the official index for prioritizing subsidized orthodontic care in the area where this research was conducted. In order to retrieve the information on DHC grading, a dental record screening was conducted in July 2015 where data was collected retrospectively from the dental records of each participant. A database was set up including professional assessments and diagnosis/notes on occlusal status, treatment need evaluation and IOTN-DHC grading. The treatment need measure was based on retrieved data from dental records, where DHC-grading 4 or 5 is usually noted in case of treatment need. In case of missing DHC-grading, the first author (an orthodontist and experienced user of IOTN) interpreted the professionals notes and converted them into corresponding DHC-grading. This was done in 28 out of 46 cases with confirmed treatment need and repeated twice for better intra-examiner reliability. This data collection was done in order to enable matching of the findings from the self-assessed measures (questionnaire based) with professionally generated data from dental records. Any of the measures presented or underlying materials can be obtained upon request. Statistical analysis For basic statistical analysis, the mean scores and standard deviations of the measures were calculated as well as the Cronbach s alpha coefficient to evaluate the reliability/internal consistency of the measures. Further the Pearson product-moment correlations were calculated to study the relations between measures. Path analysis was conducted using Mplus (31). Path analysis is a powerful method to simultaneously study, or model, the relations among different measures including more than one dependent variable and can be used instead of regression analysis where only a single dependent variable is used. Path analysis is a special form of structural equation modelling, SEM, where the relations between latent variables are used. Model testing in path analysis is straightforward. A theoretical model is proposed and tested. The fit between the theoretical model and the data-generated models is expressed by a variety of indicators. The chi-square test (χ 2 ) indicates the overall test of difference between observed and expected relations. A chisquare closer to zero indicates a better fit between the data and the theoretical model. However, since chi-square is dependent on sample size, supplementary indicators of fit are recommended to evaluated models (32). The root mean square error of approximation, RMSEA, is used to measure the discrepancy between the observed and the hypothesized relations per degree of freedom. Outcomes near or lower than 0.05 indicate a satisfactory fit. The RMSEA is presented together with the 90% confidence interval, CI. The standardized root mean square residual, SRMR, is the square root of the discrepancy between data and the theoretical model. SRMR ranges from 0 to 1, a value of 0.08 suggesting an acceptable model fit. The final indicator used, the comparative fit index, is a relative measure of fit only interpretable when comparing two theoretical models. CFI values range from 0 to 1, with larger values indicating better fit. A CFI value 0.90 is considered to indicate acceptable model fit (33). Results Preliminary analyses The final data set comprised of complete responses from 150 individuals (56% girls). The prevalence of professionally assessed treatment need, based on IOTN-DHC grades 4 and 5, was 28 percent.

4 88 European Journal of Orthodontics, 2017, Vol. 39, No. 1 Initially, we conducted a series of basic statistical analyses examining the properties of our measures and their interrelations. The first series of these analyses, presented in Table 1, showed that the reliability of the scale scores was good except for prioritizing healthy and straight teeth (α = 0.68) which can be considered satisfactory. Further analyses examined the correlations among all measures (Table 2). These correlations varied between 0.63 (social influence with dental self-esteem) and 0.64 (treatment demand with treatment need). Notably, the correlations among variables measuring similar or theoretically related constructs were high, supporting the construct validity of the measures. For example, global and dental self-esteem and perceived malocclusion and perceived functional limitations were highly correlated. Main analyses Path analysis was conducted using Mplus (31). With insights from the correlation analysis (Table 2) and theoretical elaborations guided by previous research (15, 19), we arrived at a model including the following relations. First, we reasoned that individuals perception of malocclusion is affected by their tendency to prioritize healthy and straight teeth and their perception of functional problems/limitations. Thus, we included a path indicating that perceived malocclusion is affected by prioritizing healthy and straight teeth and perceived functional limitations. Second, we reasoned that dental self-esteem should be affected by individuals global self-esteem, their sensibility to social influences as well as their perception of functional problems in everyday life. Responding to this, we included paths from global selfesteem, social influence and perceived functional limitation to dental self-esteem. Third, we reasoned that the correlation between an individual s dental self-esteem and perception of malocclusion might reflect a two-directional (reciprocal) relation with different effect sizes (Table 2). Therefore, we included reciprocal paths between perceived malocclusion and dental self-esteem. Finally, examining to what extend self-perceived treatment demand and professionally assessed orthodontic need of treatment can be predicted, we included paths from perceived malocclusion and dental self-esteem to treatment need and treatment demand. The model incorporating the relations outlined above forms our theoretical model presented in Figure 1. Testing this model, the path analysis revealed a reasonable fit to the data, χ 2 (10) = 21.15, P = 0.02, CFI = 0.96, RMSEA = 0.09, 90% CI = , SRMR = However, the paths from prioritizing healthy and straight teeth to perceived malocclusion and dental self-esteem to treatment need were not significant. In the next step, we removed these non-significant paths and run the model again. The removal of these paths provided an improvement of model fit, χ 2 (8) = 17.42, P = 0.03, CFI = 0.97, RMSEA = 0.09, 90% CI = , SRMR = Further, modification indices provided by Mplus suggested a few changes that would possibly improve model fit. For example, Mplus suggested a path from treatment need to dental selfesteem. However, these suggestions seemed unwarranted, theoretically as well as empirically. Indeed, adopting the suggested modifications would result in a model with significantly poorer fit. The final model Table 2. Intra correlations between measures. Correlations (r) Measure Dental self-esteem 2. Global self-esteem 0.39* 3. Social influence 0.63* 0.41* 4. Perceived malocclusion 0.50* 0.21* 0.29* 5. Perceived functional limitation 0.18* 0.38* 0.29* 0.46* 6. Prioritizing healthy and straight teeth * 0.19* Treatment demand 0.44* * 0.55* 0.25* Treatment need (DHC) 0.26* * 0.17* * r, Pearson product-moment correlation coefficient. *P < Figure 1. Theoretical path model explaining orthodontic treatment need and treatment demand.

5 J. Taghavi Bayat et al. 89 with standardized path coefficients is presented in Figure 2. This model provides a test of the unique contribution of all included measures on both treatment need and treatment demand. However, there were two critical questions that needed to be elaborated on. The first question was how would a model dealing with survey data only fit the data? To test this we simply run a model after excluding treatment need. This model showed excellent fit to the data χ 2 (4) = 3.71, P = 0.45, CFI = 1.00, RMSEA = 0.001, 90% CI = , SRMR = 0.02, explaining 33% of the variance in treatment demand. The second question concerned the relation between treatment need and treatment demand. There were no suggestions regarding which of these two measures would function as the antecedent of the other. Responding to this question we examined two models, one where perceived malocclusion and dental self-esteem predicted treatment demand which in turn predicted treatment need and another where the last path was exchanged with treatment need predicting treatment demand. These models, would arguably examine the causal order between treatment demand and treatment need. A reasonable prediction was that the model where treatment demand precedes and predicts treatment need is the better one, both statistically and theoretically. Supporting our prediction, the results showed that the first model had significantly (P < 0.01) better fit [χ 2 (9) = 24.29, P = 0.004] than the second [χ 2 (9) = 45.03, P < 0.001]. Discussion The findings proved that the constructed measures were reliable and related to treatment need as well as treatment demand. The results also revealed that treatment demand and treatment need were highly correlated. More importantly, the results showed that our proposed model had good fit to the data. The model showed that treatment demand is largely explained by perceived malocclusion and dental self-esteem, which in turn are linked together and explained by perceived functional limitation, social influence and global self-esteem. Further, treatment need was predicted by perceived malocclusion which in turn was predicted by perceived functional limitation and dental self-esteem. Our model illustrates that the measures presented are not only linked together but also significantly linked to treatment need as assessed by orthodontists. While the general conclusion is that the model provides good fit to data and seems theoretically sound, some aspects need to be discussed in more detail. One of these aspects is the relation between treatment need and treatment demand. Here, we found a high zeroorder correlation between the two measures (r = 0.64). Correlation of such a magnitude shows that professionally assessed orthodontic treatment need is predicted well by self-assessed treatment demand. This suggestively emphasizes the need for reflecting patients perspectives in similar area of research, as recently discussed (34). The current results are mainly based on the responses of one hundred and fifty 13-year-olds. This age-group was considered adequate since orthodontic treatment with fixed appliances is often implemented at the age of years, when all the permanent teeth has erupted. But also since a certain level of psychological maturity is needed when responding to items, for instance, related to psychological variables. Earlier studies also indicate that children usually need to have reached an age of years in order to achieve adequate reasoning about aesthetics in relation to orthodontics (35, 36). The results are also linked to dental record data concerning IOTN, being the official index for assessment of need for subsidized orthodontic treatment in the county of interest. The prevalence of treatment need, IOTN-DHC grades 4 and 5, in our study population was 28 percent. This could be compared to prevalence rates of percent reported in numerous European countries (30, 37 40). A similar variation has been reported within the county councils of Sweden concerning commenced subsidized orthodontic treatment (17). Further, the proportion of individuals receiving orthodontic treatment based on dental records was 21%. Analyses showed that the proportion of participants receiving treatment in this study was not significantly different (z = 1.13, P = 0.26) from figures presented earlier, 24%, concerning Uppsala County (17). Given the above mentioned factors, the study population is considered to be representative and the treatment need figures to be fairly reliable. However, when it comes to IOTN as well as other morphological indices for treatment need prioritization, a systematic review in 2005 concluded that evidence for conclusions concerning their validity are lacking (17). In 2014, another systematic review concluded that most of the outcomes used in orthodontic research are concerned with measuring morphologic changes of treatment and do not reflect patient perspectives (34). Together with our findings this highlights that orthodontic treatment need assessments should be based on the consequences of malocclusion for the individual. Therefore, we suggest that, as a complement to the clinical evaluation and diagnosis, validated self-assessment measures ought to be used in a much higher degree than is the case today. Figure 2. Path model explaining orthodontic treatment need and treatment demand. All standardized coefficients are significant (P < 0.05, all P-values are twotailed).

6 90 European Journal of Orthodontics, 2017, Vol. 39, No. 1 Limitations The present study was performed in a specific age group, which could affect the generalizability of the findings. Further, it should be mentioned that our path modelling, although showing good fit to the data, is a result of semi-explorative approach based on a theoretical ground with many variables. The validity of the model needs to be confirmed in future studies as well as by examining factors outside the model and possible mediators when it comes to the relation between for example perceived malocclusion and treatment need and demand. One candidate variable could for instance be the access to subsidized dental care. Conclusions The proposed path model displays the unique effect of each included measure on treatment need and demand, explaining a large proportion of the variance in perceived treatment demand and professionally assessed treatment need. This indicates that orthodontic treatment need assessments would benefit from including measures enabling assessment of patients perceptions. The model would hopefully lead to improved and more cost-efficient predictions of treatment need and demand. Funding The County Council of Uppsala (FTVLUL to J.T.B.) and Uppsala Public Dental Service. Acknowledgements The authors would like to thank all the participants. The authors also like to thank professors Pia Gabre and Frank Lindblad for their support in the process of data collection. Moreover, thanks to Pia Pettersson and Helene Lindroth for their administrative help. References 1. Helm, S. (1968) Malocclusion in Danish children with adolescent dentition: an epidemiologic study. American Journal of Orthodontics, 54, Thilander, B. and Myrberg, N. (1973) The prevalence of malocclusion in Swedish schoolchildren. Scandinavian Journal of Dental Research, 81, Hannuksela, A. (1977) The prevalence of malocclusions and the need for orthodontic treatment in 9-year old Finnish schoolchildren. Proceedings of the Finnish Dental Society, 73, Brunelle, J.A., Bhat, M. and Lipton, J.A. (1996) Prevalence and distribution of selected occlusal characteristics in the US population, Journal of Dental Research, 75, Thilander, B., Pena, L., Infante, C., Parada, S.S. and de Mayorga, C. (2001) Prevalence of malocclusion and orthodontic treatment need in children and adolescents in Bogota, Colombia. An epidemiological study related to different stages of dental development. European Journal of Orthodontics, 23, Dimberg, L., Lennartsson, B., Arnrup, K. and Bondemark, L. (2015) Prevalence and change of malocclusions from primary to early permanent dentition: a longitudinal study. The Angle Orthodontist, 85, Shaw, W.C., O Brien, K.D., Richmond, S. and Brook, P. (1991) Quality control in orthodontics: risk/benefit considerations. British Dental Journal, 170, Järvinen, S. (1979) Traumatic injuries to upper permanent incisors related to age and incisal overjet. A retrospective study. Acta Odontologica Scandinavica, 37, Forsberg, C.M. and Tedestam, G. (1993) Etiological and predisposing factors related to traumatic injuries to permanent teeth. Swedish Dental Journal, 17, Burden, D.J. (1995) An investigation of the association between overjet size, lip coverage, and traumatic injury to maxillary incisors. European Journal of Orthodontics, 17, Helm, S., Kreiborg, S. and Solow, B. (1985) Psychosocial implications of malocclusion: a 15-year follow-up study in 30-year-old Danes. American Journal of Orthodontics, 87, Kenealy, P.M., Frude, N. and Shaw, W. (1989) An evaluation of the psychological and social effects of malocclusion: some implications for dental policy making. Social Science and Medicine, 28, Liu, Z., McGrath, C. and Hagg, U. (2009) The impact of malocclusion/ orthodontic treatment need on the quality of life. A systematic review. Angle Orthodontist, 79, Dimberg, L., Arnrup, K. and Bondemark, L. (2015) The impact of malocclusion on the quality of life among children and adolescents: a systematic review of quantitative studies. European Journal of Orthodontics, 37, Taghavi Bayat, J., Hallberg, U., Lindblad, F., Huggare, J. and Mohlin, B. (2013) Daily life impact of malocclusion in Swedish adolescents: a grounded theory study. Acta Odontologica Scandinavica, 71, Shaw, W.C., O Brien, K.D. and Richmond, S. (1991) Quality control in orthodontics: factors influencing the receipt of orthodontic treatment. British Dental Journal, 170, SBU The Swedish Council on Technology Assessment in Health Care (2005) SBU Report: malocclusions and orthodontic treatment in a health perspective a systematic review. Elanders Infologistics Väst AB, Mölnlycke, Sweden. 18. Feldens, C.A., Nakamura, E.K., Tessarollo, F.R. and Closs, L.Q. (2015) Desire for orthodontic treatment and associated factors among adolescents in Southern Brazil. The Angle Orthodontist, 85, Trulsson, U., Strandmark, M., Mohlin, B. and Berggren, U. (2002) A qualitative study of teenagers decisions to undergo orthodontic treatment with fixed appliance. Journal of Orthodontics, 29, Mohlin, B. and Kurol, J. (2003) To what extent do deviations from an ideal occlusion constitute a health risk? Swedish Dental Journal, 27, Benson, P.E., Javidi, H. and DiBiase, A.T. (2015) What is the value of orthodontic treatment? British Dental Journal, 218, Jokovic, A., Locker, D., Stephens, M., Kenny, D., Tompson, B. and Guyatt, G. (2002) Validity and reliability of a questionnaire for measuring child oral-health-related quality of life. Journal of Dental Research, 81, Dunlow, N., Phillips, C. and Broder, H.L. (2007) Concurrent validity of the COHIP. Community Dentistry and Oral Epidemiology, 35, O Brien, C. and Benson, P.E. (2007) Evaluation of a quality of life measure for children with malocclusion. Journal of Orthodontics, 34, Agou, S., Locker, D., Streiner, D.L. and Tompson, B. (2008) Impact of selfesteem on the oral-health related quality of life of children with malocclusion. American Journal of Orthodontics and Dentofacial Orthopedics, 134, Bernabé, E., de Oliveira, C.M. and Sheiham, A. (2008) Comparison of the discriminative ability of a generic and a condition-specific OHRQoL measure in adolescents with and without normative need for orthodontic treatment. Health and Quality of Life Outcomes, 6, Faul, F., Erdfelder, E., Lang, A.G. and Buchner, A. (2007) G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39, Schönbrodt, F.D. and Perugini, M. (2013) At what sample size do correlations stabilize? Journal of Research in Personality, 47, Forsman, L. and Johnson, M. (1996). Dimensionality and validity of two scales measuring different aspects of self-esteem. Scandinavian Journal of Psychology, 37, 1 15.

7 J. Taghavi Bayat et al Brook, P.H. and Shaw, W.C. (1989) The development of an index of orthodontic treatment priority. European Journal of Orthodontics, 11, Muthén, L.K. and Muthén, B.O. (2012) Mplus User s Guide. 7th edn. Muthén & Muthén, Los Angeles, CA, USA. 32. Browne, M.W. and Cudeck, R. (1993). Alternative ways of assessing model fit. In Bollen, K.A. and Long, J.S. (eds.), Testing structural equation models. Sage, Newbury Park, CA, USA. 33. Kline, R.B. (2010). Principles and practice of structural equation modeling. 3rd edn. Guilford Press, New York, USA. 34. Tsichlaki, A. and O Brien, K. (2014) Do orthodontic research outcomes reflect patient values? A systematic review of randomized controlled trials involving children. American Journal of Orthodontics and Dentofacial Orthopedics, 146, Shaw, W.C. (1981) Factors influencing the desire for orthodontic treatment. European Journal of Orthodontics, 3, Birkeland, K., Katle, A., Løvgreen, S., Bøe, O.E. and Wisth, P.J. (1999) Factors influencing the decision about orthodontic treatment. A longitudinal study among 11- and 15-year-olds and their parents. Journal of Orofacial Orthopedics, 60, Burden, D.J. and Holmes, A. (1994) The need for orthodontic treatment in the child population of the United Kingdom. European Journal of Orthodontics, 16, Souames, M., Bassigny, F., Zenati, N., Riordan, P.J. and Boy- Lefevre, M.L. (2006) Orthodontic treatment need in French schoolchildren: an epidemiological study using the index of orthodontic treatment need. European Journal of Orthodontics, 28, Josefsson, E., Bjerklin, K. and Lindsten, R. (2007) Malocclusion frequency in Swedish and immigrant adolescents influence of origin on orthodontic treatment need. European Journal of Orthodontics, 29, Perillo, L., Masucci, C., Ferro, F., Apicella, D. and Baccetti, T. (2010) Prevalence of orthodontic treatment need in southern Italian schoolchildren. European Journal of Orthodontics, 32,