The pathways from disadvantaged socioeconomic status in childhood to edentulism in mid-to-late adulthood over the life-course

Background This study aimed to examine the direct and indirect pathways from childhood socioeconomic status (SES) to the prevalence of edentulism in mid-to-late age Chinese individuals using structural equation modeling (SEM). Methods This study analyzed data from 17,032 mid- to-late age Chinese individuals in the 2014 and 2015 China Health and Retirement Longitudinal Study (CHARLS). Childhood SES was determined based on the parents’ education and occupation, financial situation of the family, primary residence, food availability, and medical convenience. Adulthood SES was established according to educational achievements of the individuals. Edentulism is defined as the loss of all natural teeth. SEM was used to examine the statistical significance of the association between childhood SES and edentulism, mediated by childhood health, adulthood SES, and adult health. Results Childhood SES had significant indirect (β = -0.026, p < 0.01), and total (β = -0.040, p < 0.01) effects on edentulism. It was determined that 65% of the total effect of childhood SES on edentulism was indirect, and mainly mediated by adult SES. Also, the goodness-of-fit indices of the best-fitting model were acceptable. Conclusion This study revealed that childhood health, adult health and adult SES are mediators that explain the relationship between childhood SES and edentulism. The global attention to alleviate the inequality in edentulism should focus on exploring recommendations and intervention strategies from childhood to adulthood, by considering adult SES, childhood and adult health.


Background
Oral diseases are a global public health issue, with particular concern about their increasing prevalence over the life course, associated with social and economic changes [1].They have substantial negative effects on the quality of life of individuals, due to pain, speech difficulties and self-esteem [2] in low-and middle-income countries (LMICs) [3].The current oral healthcare services are inequitable, and leave disadvantaged people with inadequate basic oral healthcare services [4].Inequalities in oral health are complicated by individual, social, economic, and environmental determinants that are globally acknowledged across the oral profession [5].
Edentulism is the condition of having lost all natural teeth [6].Due to its high prevalence [7], the age-standardized prevalence rates of edentulism, according to the Global Burden of Disease (GBD) 2015, was 4.1%, and edentulism was the leading cause of disability-adjusted life years (DALYs) [1].Edentulism is an irreversible condition, which reflects the lifetime outcome of oral diseases, absence of oral treatment, impaired masticatory function and unhealthy diet [8].Monitoring edentulism is a major determining factor in the assessment of the performance of the oral healthcare services [9], and a high priority for oral disease prevention efforts [10].Edentulism is significantly associated with disadvantaged socioeconomic status (SES) determinants [11,12], edentulism, being the cause and consequence of social inequality, represents oral healthcare services, social, and economic burden, being the source and consequence of social inequality [13].
The life course epidemiology theory proposes that social and economic exposures during the childhood developmental period [14], have long-term effects on health outcomes or disease risk in adulthood [15].Indeed, childhood SES inequalities in oral health remained with age [16], and showed a long-lasting relationship with childhood SES and oral health outcomes [17].Childhood SES was found to have long-term consequences on severe tooth loss [18].Disadvantaged childhood SES was found to have an effect on dental caries [19], and poor financial situation of the family in childhood was associated with the prevalence of unsound teeth in adulthood [20].Being closely associated to SES, oral diseases disproportionally affect poorer SES groups [3].Several studies have revealed that oral health inequalities are directly influenced by SES determinants [21][22][23], including education [24], and occupation [24].Childhood SES was assessed based on parental SES determinants, including parental education [25] and occupation [26], self-reported financial situation of the family [27],primary residence [28] and food availability [29].Disadvantaged childhood SES was also found to be related to oral healthcare service access [30,31], which may contribute to severe tooth loss [32].
Oral conditions disparity in childhood, is a lifelong condition that can be tracked across the life course [33], increasing evidence indicates a direct effect of childhood SES on oral health in adulthood after controlling for adult SES [34], as assessed by educational achievements [26].It has been reported that children who grew up in disadvantaged SES families had poorer oral health, with a threefold increase in low versus high SES families, in adult periodontal disease and dental caries [35].A recent study highlighted that the effect of SES inequalities on health outcomes is a significant contributor to inequalities in childhood health [36], and the mediating models suggested that childhood SES may influence adult health via education and occupational exposures [37].
Traditional regression models estimate independent direct effects, but there are methodological challenges to estimate direct and indirect effects of childhood SES without some degree of collinearity with oral health outcomes between intermediate confounding factors [38].Structural equation modeling (SEM) is a statistical approach to estimate the direct and indirect pathways of multiple variables that combines confirmatory factor analysis (CFA) and multiple regression analysis [39], and depends on the outcome variable to decompose effects that go through a chain, and its application in oral health has been well documented [40].

Conceptual Framework
A study has explored the association between childhood SES and edentulism in mid-to-late adulthood to understand the potential interventional policies using regression models [34].Edentulism develops via complex pathways over the life course, but there is little evidence on the direct and indirect pathways from childhood SES and edentulism.According to the 2010 report "A Conceptual Framework for Action on the Social Determinants of Health" by the World Health Organization (WHO) [41], a context-specific database, and literature review, the core components of the conceptual framework in this study include: (1) Childhood SES associated with parents' occupation and education, food shortage, medical convenience, primary residence, and financial situation of the family, which directly influence edentulism; (2) Childhood SES indirectly influence edentulism, via intermediate determinants including: adult SES, childhood health, and adult health.The multiple pathways from childhood SES to edentulism, as well as the statistical analysis strategy are shown in Fig. 1.

Study design and participants
This study used nationally representative data from the 2014 and 2015 waves of the China Health and Retirement Longitudinal Study (CHARLS) of Chinese residents aged 45 and above [42].The CHARLS includes a four-stage, stratified, probability sampling procedure covering 10,803 households from 450 villages/residential communities in 28 provinces across mainland China [42].The 2014 wave was a life history survey of 20,543 participants and retrospectively collected childhood SES [43].The 2015 wave was a regular follow-up survey of 21,095 participants and collected information on adult SES and health outcomes in mid-late adulthood.This study matched the participants of the 2014 and 2015 waves according to their IDs to trace the information on childhood SES and health, adult SES and health, and the prevalence of edentulism.The minimum sample size for SEM with more than 7 constructs was 500 [44].Participants aged under 45 years who had missing information on childhood SES and edentulism were excluded.After matching and exclusion, 17,032 participants were ultimately included in the current analysis.CHARLS was approved by the Ethical Review Committee of Peking University and informed consent was obtained from each participant at the time of participation [42].

Measures
The prevalence of edentulism was determined based on the response to the question: "Have you lost all of your teeth?" and dichotomized into "Yes" or "No".Demographic characteristics included sex (male or female) and age groups (45-59, 60-69, 70-79, and 80 or above).
Childhood SES was assessed by eight variables: education of parents, occupation of parents, food shortage, financial situation of the family, primary residence and medical convenience [34,43].Education level of parents was categorized into "illiterate", "primary school", "middle school" or "high school and above".Occupations of parents were categorized into "non-agricultural" or "agricultural".Food shortage was assessed based on the question: "When you were a child before the age of 17 was there ever a time when your family did not have enough food to eat?" and dichotomized into "Yes" or "No" [45].Financial situation of the family was assessed based on the question: "When you were a child before the age of 17, compared to the average family in the same community/village at that time, how was your family's financial situation?"and rated as "a lot better off than them", "somewhat better off than them", "same as them", "somewhat worse off than them" or "a lot worse off than them".Primary residence was determined based on the question: "Where did you mainly live before you were 16 years old? was it in a village or a city/town?[rural or urban (city or town)]" and dichotomized into "urban" or "rural".Medical convenience was determined based on the question: "Are you satisfied with the quality, cost, and convenience of local healthcare" and dichotomized into "Yes" or "No".
Childhood health was based on the question: "how would you evaluate your childhood health, up to and including age 15?" and categorized into "excellent", "very good", "good", "fair" and "poor".Adult SES was assessed by education [46].Adult education was determined based on the question: "What's the highest level of education you have attained now?", this study used the categories of "illiterate", "elementary school", "middle school" or "high school and above".Adult health was assessed by selfreported health.Self-reported health was based on the question: "how would you evaluate your adult health, after age 15?" and categorized into "excellent", "very good", "good", "fair" and "poor".

Statistical analysis
Descriptive statistics, correlation analysis, and the calculations and estimates of the SEM were performed using the statistical software Stata 15.0 (Stata Corporation, College Station, TX, USA).No significant difference in the distribution of covariates was found between the final analysis sample and those with missing data in edentulism and childhood SES.Chisquared (χ 2 ) tests and Fisher's Exact tests were performed to compare the categorical variables including childhood SES, childhood health, adult SES and adult health between participants with and without edentulism.Correlation analyses were performed to assess the correlation of childhood SES, childhood health, adult SES and adult health, with edentulism.On the basis of the conceptual framework, this study developed an analytical model using the SEM approach to evaluate direct and indirect effects of childhood SES on edentulism.
CFA was used to examine the hypothesized measurement model by evaluating the relationships among observed and latent variables [47].The direct and indirect effects between outcome, observed and latent variables were determined by SEM.The total effect represents the sum of direct and indirect effects, mathematically expressed as follows [48]: c = c′ + ab, where c = total effect, c′ = direct effect, ab = indirect effect.Bias-corrected bootstrapping (2000 bootstrap samples) was then used to estimate the statistical significance of the direct and indirect effects of each pathway in the SEM.Multiple modification indices were used to adjust, modify and obtain the best-fit model [49], including the comparative fit index (CFI), incremental fit index (IFI), standardized root mean square residual (SRMR), and the root-mean-square error of approximation (RMSEA).The RMSEA and SRMR ≤ 0.06, CFI and IFI > 0.90 indicated an acceptable model [50].Although χ 2 values should be reported as one of the fit indices, they were highly sensitive to large sample sizes and thus were excluded [51].P < 0.05 indicated statistical significance.

Results
The results of descriptive statistics and univariate analysis are presented in Table 1.The average age of the participants was 63.74 ± 10.34 years, and 52% were females.Individuals whose parents' occupation was agricultural were more likely to have edentulism than those whose parents' occupation was non-agricultural (p < 0.05).Significant differences in the education level of the parents were determined between participants with and without edentulism (p < 0.05).Individuals with food shortage and medical inconvenience were more likely to have edentulism than those without food shortage and medical inconvenience during childhood (p < 0.05).Individuals whose primary residence was in a rural area were more likely to have edentulism than those whose primary residence was in an urban area (p < 0.05).Individuals who did not receive formal education were more likely to have edentulism than those with higher education level (p < 0.001).Individuals who self-reported poor health were more likely to have edentulism than those who did not self-report poor health (p < 0.01).

SEM results
The standardized path estimates of the pathways from childhood SES to edentulism in late-life adulthood are shown in Fig. 2, and the standardized estimates of the direct, indirect and total effects of childhood SES on edentulism, as well as the specific effects via multiple pathways of childhood health, adult SES and adult health are shown in Tables 3 and 4. The standardized factor loadings of the CFA model of childhood SES were acceptable.Childhood SES had significant indirect (β = -0.026,p < 0.01), and total (β = -0.040,p < 0.05) effects on edentulism, but the direct effect of childhood SES on edentulism was insignificant (β = -0.014,p = 0.135).It was determined that 65% of the total effect of childhood SES on edentulism was indirect, and mainly mediated by adult SES.Specifically, childhood SES negatively predicted edentulism in late-life adulthood via poor health in childhood with an estimated indirect effect of -0.001 (p < 0.01).Childhood SES negatively predicted edentulism in midto-late adulthood via disadvantaged adult SES, with an estimated indirect effect of -0.023 (p < 0.01).Childhood SES negatively predicted edentulism in mid-to-late adulthood via poor adult health, with an estimated indirect effect of -0.001 (p < 0.01).Childhood SES significantly predicted edentulism in mid-to-late adulthood via the sequential mediation of childhood health and adult SES, with an estimated indirect effect of higher than 0.001 (p < 0.01).Childhood SES significantly predicted edentulism in mid-to-late adulthood via the sequential mediation of childhood health and adult health, with an estimated indirect effect higher than 0.001 (p < 0.01).
Childhood SES significantly predicted edentulism in mid-to-late adulthood via the sequential mediation of childhood health, adult SES and adult health, with an estimated indirect effect higher than 0.001 (p < 0.01).
The goodness-of-fit indices of the best-fitting model were acceptable, RMSEA and SRMR < 0.06, and CFI and IFI > 0.90.

Discussion
To the best of our knowledge, this is the first study to explore the direct and indirect pathways from childhood SES to edentulism using a SEM approach in a LMIC.The findings of this study support the conceptual framework    Due to the existence of multivariable and complicated mediators between childhood SES and edentulism, the SEM approach used in this study provided a clearer picture of the direct and indirect pathways from childhood SES to edentulism.This study identified an unexpected direct association between childhood SES and edentulism, which showed the importance of studying pathways from childhood SES to oral health outcomes.The findings of this study are consistent with those of previous life-course epidemiological studies [52,53].Two main chain risk models from childhood SES to edentulism were validated in this study: (1) Childhood SES was indirectly associated with edentulism mediated by childhood health; and (2) childhood SES was indirectly associated with edentulism mediated by adult SES.
Untreated dental caries in adulthood were negatively associated with childhood SES, both adults from LMICs and adults with disadvantaged SES in high-income countries (HICs) have life course history of oral disease [3].Childhood SES affected tooth loss mediated by adult SES in Southern Brazil [54].This study indicated that the indirect pathway from childhood SES to edentulism via adult SES.This is in agreement with previous study, in which the effect of disadvantaged childhood SES on oral health in adulthood was found to be mediated by adult SES [55].SES has a cumulative effect on oral diseases, resulting in a pervasive SES over the life course [16].Disadvantaged adult SES is associated with a higher risk of having dental lesions or dental experience [56], a severe dental caries [57], untreated dental caries lesions or lesion caries experience [56].The findings in this study showed that SES has a cumulative effect on edentulism over the life course, thus highlighting the importance of childhood SES as an indirect effect on edentulism in adulthood.An advantaged early life is important to minimize the influence of chain risk and accumulative risk, but it is necessary to decrease SES inequalities by oral health recommendations during the whole life course in prevention of edentulism [4].
Accumulating evidence supports the suggestion that there is a relationship between oral and general health [58], which is agreement with the finding of this study, childhood SES significantly predicted edentulism in midto-late adulthood via the sequential mediation of childhood health and adult health.This is also in agreement with another previous study [59], supporting the accumulative risk model over the life course theory [54].An earlier study also highlighted that edentulism-related health loss in childhood in LMICs was significantly associated with psychological challenges mediated by childhood health [60].Swedish longitudinal surveys showed that financial hardship in childhood affects psychological distress in adulthood, and higher education in adulthood decreases the psychological distress [61].The chain risk model is supported by the finding in this study indicating that childhood SES significantly indirectly predicted edentulism via the sequential mediation of childhood health and adult SES, the sequential mediation of childhood health and adult SES, and the sequential mediation of childhood health, adult SES and adult health.

Recommendations for practice
Regardless of the childhood SES, the optimal oral health policy may overestimate the effects of other factors [34].The current oral health policy lacks an analysis focused on pathways from childhood SES to oral health outcomes, and thus cannot completely tackle the global burden of oral diseases [62].This study highlighted the urgent need to address childhood SES inequality as an oral health priority in LMICs [3].A different strategy for the prevention of edentulism is needed to tackle the challenge, from a life course perspective, which considers how such an oral health SES policy can be implemented from childhood to mitigate the prevalence and development of edentulism in mid-to-late adulthood [4].
The implementation of intervention policies to address SES inequalities has been slow, the oral health community has been advocating the importance of integrated upstream and community based interventions, but oral healthcare services still operate as a non-integrated downstream oral program [4].The prioritization of downstream interventions, such as clinical treatment services is an effective approach, that may decrease SES inequalities in oral health [63].Oral policy makers rely on simplistic downstream interventions, due to the challenges of efficient evidence for the complex upstream interventions [3].It is necessary to change the downstream interventions that cannot effectively achieve oral health gain or tackle inequalities, by creating resilient approaches for the prevention of upstream oral health [3].Integrated and coordinated strategies targeting upstream prevention are required to tackle the underlying SES inequalities in oral diseases [4].A reform of oral healthcare services is needed for cohesive, comprehensive, and integrated interventions, to address the neglect of oral health inequities in disadvantage childhood SES [4].
In LMICs access to dental prosthetic care may be significantly lower than in HICs [64].In China, in the past, tooth extraction was the main treatment for severe oral problems, which may be associated with edentulism [34].
In LMICs, the supply of medical resources, and investment on oral healthcare services are very limited, thus oral health practitioners may be insufficient, unavailable, and unaffordable for the public, particularly the disadvantaged SES population [4].In China, due to the lack of access to primary and basic oral healthcare services for children and elders, such as hygiene guides, and oral health education, it is commercial pressures and treatment drives rather than oral health promotion and prevention [4].The oral intervention approaches need to be effectively tailored to the needs of communities and individuals, to ensure oral health equity [4].The distribution of medical care is a primitive outdated social form, and any return to it would exacerbate the maldistribution of medical resources [65].Inequitable distribution of oral healthcare services may limit access to high quality and affordability of oral healthcare and preventive services, children with disadvantaged SES may become vulnerable to edentulism in mid-to-late adulthood, which could be avoided [34].Households that pay the whole oral healthcare services have a higher risk of disadvantaged SES population, when spending a larger portion of disposable income [66].It should be a priority to provide whole insurance coverage to support oral healthcare services for disadvantaged SES population from childhood [4].

Strengths and limitations
This study has a number of limitations.First, edentulism was a self-reported outcome variable, which may potentially introduce reporting bias, although the self-reported oral health outcome was valid and reflected the real oral health status [67,68].Second, childhood SES data was represented by the parents' SES were retrospective data, and although these variables were validated, memory bias cannot be ignored.To be considered for this, childhood SES was a latent variable, including education and occupation of the parents, food shortage, financial situation of the family, primary residence and medical convenience.Third, the prevalence of edentulism (5.8%) in this Chinese group was lower than that in the WHO Study on global AGEing and adult health (SAGE) (8.0-9.0%)[69].The report of edentulism in the CHARLS database may suffer from survival bias.According to the CHARLS data, about 70% of participants did not report income in midto-late adulthood, the reason could be that 91.2% of them were living in rural areas, and did not receive a stable income.Therefore, this study did not use the variable of income to represent adult SES.Thus, this study was limited by the available variables and participants' responses in the existing CHARLS database.Despite these limitations, this study has some strengths.For instance, this study used a unique and nationally representative data from a LMIC.The statistical analysis was conducted using SEM based on conceptual framework, which is an advanced statistical approach to estimate direct and indirect effects from exposure to outcome including latent variables.This is the first study using SEM to analyze the pathways from childhood SES to edentulism via childhood health, adult health and adult SES over the life course.This study is beneficial for addressing inequality in oral health from a life course perspective among the disadvantaged SES.

Conclusions
This study contributed to the evidence on the life course of the direct and indirect pathways from childhood SES to edentulism using SEM, and showed that childhood health, adult health and adult SES are mediators that explains this relationship in China.It suggested that it is necessary to explore integrated recommendations and intervention strategies from childhood to adulthood, considering the mediators of adult SES, childhood and adult health to mitigate the effects of childhood SES inequality on oral health outcomes.Currently, there is health inequality in access to oral healthcare services among the disadvantaged SES population that contributes to the inequality of oral health outcomes in LMICs.Primary oral healthcare services should consider that childhood health, adult health and adult SES may attenuate the effect of childhood SES inequality on the prevalence of edentulism in mid-tolate adulthood.

Fig. 1
Fig.1Conceptual model of the relationship between childhood SES and edentulism in mid-late adulthood.SES, socioeconomic status

Table 1
Descriptive statistics of the participants with and without edentulism (N = 17,032)

Table 3
Standardized regression weight of variables in SEM model (N = 17,032)

Table 4
Standardized indirect, direct and total effects of childhood SES on edentulism (N = 17,032) that the pathways from childhood SES to edentulism proceed via childhood health, adult SES, and adult health, which are targeted as oral health interventions to reduce the prevalence of edentulism and oral health inequities across the life course of individuals.