Abstract
Objectives
To investigate the relationship between caries in primary teeth and caries in permanent teeth among primary school students, and to identify potential effect modifiers of this relationship.
Methods
This retrospective cohort study utilized real-world caries examination data from the School Dental Care Service (SDCS) provided by the Hong Kong SAR government. Primary school students in grades 3 (P3) or 4 (P4) who participated in the SDCS were recruited from school dental clinics. The exposure was defined as any caries experience in primary teeth in Grade 1. To account for potential confounding factors, the marginal structural model with overlap weight was applied to estimate the average treatment effect for the overlap population.
Results
Among 825 children (mean [SD] age in P1: 6.2 [0.3]; boys 54.7%) without caries in permanent teeth (DMFT = 0) in P1, the incidence of caries experience in permanent teeth (DMFT > 0) increased to 5.7% and 10.5% in P2 and P3, respectively. Among those with caries experience in primary teeth (dmft > 0, n = 443), 15.4% had caries experience in permanent teeth in P3 compared to only 4.9% among those without (n = 382; adjusted risk ratio 2.8 [95% CI: 1.7-4.8; P < .001]; adjusted risk difference 9.6% [95% CI: 5.4-13.8; P < .001]). Sex and only-child status significantly modified this relationship, with stronger associations observed in boys and non-only children.
Conclusions
Caries in primary teeth increased the risk of developing caries in permanent teeth. This association was modified by sex and only-child status.
Clinical Significance
Controlling caries in primary teeth is essential for preventing caries in permanent teeth. Risk-stratified interventions targeting boys and non-only children may optimize preventive outcomes and resource allocation in oral health promotion programs.
Key words: Caries, Early childhood caries, Electronic dental records, Preventive dentistry, Dental public health, Causal inference
Background
Early childhood caries (ECC), or caries experience in primary teeth, can have significant consequences among children, and one of these consequences is an elevated risk of developing caries lesions in permanent teeth.1,2 ECC has been found to be predictive of caries in permanent teeth in several longitudinal studies.3, 4, 5, 6 A meta-analysis of longitudinal studies concluded that children with ECC were 3 times more likely to develop caries in their permanent teeth, despite inconsistencies between studies.2 Inequality in caries, whether in primary or permanent teeth, has been found to be associated with variations in sociodemographic status.7,8 In Hong Kong SAR, fluoridated water is supplied to the entire population, and a dental health care program known as the School Dental Care Service (SDCS) is offered to primary school children by the Hong Kong SAR government Department of Health, with an enrollment rate exceeding 90%.9,10 The extent of the association between caries experience in primary teeth and permanent teeth among children, given these oral health promotion programs, remains underexplored. According to the World Health Organization, it is recommended that primary teeth be examined at 5 years of age, and permanent teeth be examined at 12 years of age.11 This examination scheme has been adopted by the majority of existing longitudinal studies.2 However, the extended follow-up period makes it impossible to assess the impact of ECC on caries in newly erupted permanent teeth. In addition, current research primarily focuses on identifying risk factors for the development of dental caries in permanent teeth, with limited research on the effect modification of the association between caries experience in primary teeth and permanent teeth. Identifying this effect modification could help in identifying subpopulations that would benefit more from preventive measures and enhance the understanding of the biological, social, or other mechanisms.12
The SDCS includes annual oral examination, oral health education, preventive dental treatment, basic restorative dental treatment, and emergency dental service.10 Records of the participants’ oral health status, including their caries status, have been stored electronically in the SMILE (SDCS Management and Information in Linked Environment) system since their enrollment in the SDCS. Therefore, the SDCS enables the collection of data on the caries experience of primary school children from the time they enter primary school. Utilizing real-world data routinely collected in school dental clinics, rather than in specific research settings, would enhance the clinical relevance and external validity of the results.12
The objective of this study was to explore the relationship between caries in primary teeth and caries in permanent teeth among primary school students, utilizing real-world dental examination data obtained through the SDCS. Furthermore, this study sought to identify potential effect modifiers of this relationship.
Methods
Study design and data collection
This is a secondary analysis of another study, which aimed to investigate the long-term impact of a kindergarten dental outreach service among primary school children in Hong Kong.14 This retrospective cohort study utilized data collected via a cross-sectional survey and clinical oral examination database. Primary school children meeting the inclusion criteria were those in primary school grades 3 (P3) or 4 (P4) enrolled in the SDCS, and who had attended kindergarten in Hong Kong. Children were ineligible if they had caries experience in their permanent teeth in P1. Originally, the plan was to recruit children for the study from randomly selected primary schools in Hong Kong. However, due to the COVID-19 pandemic and the suspension of classes, the study protocol was amended. As a result, most participants were recruited from 4 school dental clinics between July 2022 and November 2022. For children recruited from primary schools, invitation links were sent to their parents, allowing interested parents to complete the questionnaires online using Qualtrics (Qualtrics, Provo, UT) after providing informed consent. For children recruited from dental clinics, research staff approached their parents/guardians sequentially and invited them to complete the parent questionnaire, either on paper or online, to collect sociodemographic information. Following the questionnaire survey, the children’s clinical examination records, fully matched by their identification information (ie, name, identification number, and date of birth), were retrieved and encrypted before being transferred from the Department of Health. The records included the number of decayed, missing due to caries, and filled teeth (dt score, mt score, and ft score), as well as the sum of dt, mt, and ft (dmft score) in the primary teeth. Furthermore, the number of decayed, missing due to caries, filled teeth (DT score, MT score, and FT score), and the sum of DT, MT, and FT (DMFT score) in the permanent teeth were also retrieved from the SMILE system by SDCS service year, which starts in November, to ensure consistency between the SDCS service year and the actual year.
Exposure
Exposure was defined as the presence of caries in the primary teeth (dmft score > 0) in P1. Children without any dmft in P1 (dmft score = 0) were categorized as controls.
Outcome
Caries incidence was defined as the development of new caries experience (DMFT > 0) in permanent teeth among children who were caries-free (DMFT = 0) at baseline (P1), represented by a binary outcome indicating whether a child transitioned from caries-free to having at least one decayed, missing, or filled permanent tooth by P2 or P3. Caries severity was measured by the DMFT score at P2 or P3, which quantified the number of decayed, missing due to caries, and filled permanent teeth.
Statistical methods
For caries severity, negative binomial regression models were used to account for the overdispersion of the count data (DMFT score). Log link was used to estimate the mean ratio (MR), which quantified the multiplicative association on the relative scale. This link function ensures that predicted counts remain positive and provide interpretable estimates of how many times the mean DMFT score in exposed children exceeds that in unexposed children. The identity link function was used to estimate the mean difference (MD), which quantified the absolute difference in mean DMFT scores between exposure groups. This additive measure is clinically intuitive, directly indicating the average number of additional affected permanent teeth attributable to caries experience in the primary teeth, which is valuable for clinical and public health decision-making. For caries incidence, binomial regression models were fitted. Specifically, a log link was used to estimate the risk ratio (RR), which indicated how many times more likely children with caries in their primary teeth were to develop caries in their permanent teeth. The identity link was used to estimate the risk difference (RD), which provides the absolute increase in incidence risk attributable to caries in primary teeth.
The analysis involved 3 main steps to estimate the causal effect of caries in primary teeth on the development of caries in permanent teeth. First, overlap weighting was applied to adjust for baseline confounders, focusing our inference on the population where exposed and unexposed children had the most comparable characteristics. Second, inverse probability of censoring weighting was applied to account for loss to follow-up. Finally, these weights were combined to estimate the causal risk ratio. This approach allowed us to obtain valid causal estimates while addressing both confounding and selection bias due to missing data.
The overlap weighting method was applied to adjust for potential confounding factors. The overlap weight (OW) is one of the balancing weights that balances the covariate distributions between the exposure and control groups toward a target population. The overlap weight defines the target population as the subpopulation with the highest degree of overlap in the measured confounders between the exposure and control groups by smoothly down-weighting the units with propensity scores close to 0 or 1. Thus, the target population created by overlap weighting emphasizes the comparison of individuals at clinical equipoise, making the target population of greater policy interest.15 In addition, OW leads to an exact balance of every measured covariate when the PS is estimated by logistic regression and can provide optimal precision of the measure of association. Operationally, the propensity score (PS) was estimated using logistic regression as the probability of each child having caries experience in primary teeth in P1 conditional on several potential confounders (sex, age entering primary school, year entering primary school, primary school location, birthplace, only child or not, parents’ higher education level, monthly household income, main caregiver, and exposure to the kindergarten dental outreach service provided by the University of Hong Kong [HKU]). These factors were identified using a directed acyclic graph. Subsequently, the OW for the exposure group was calculated as (1-PS), while the OW for the control group was calculated as PS. To account for loss to follow-up in P2 and P3, the inverse probability of censoring weight (IPCW) was applied under the missing at random (MAR) assumption. The IPCW for each individual was calculated as the inverse of the conditional probability of not being lost to follow-up in P2 or P3. These weights up-weight children who remained in the study but had characteristics similar to those who were lost to follow-up, effectively creating a pseudo-population that represents what would have been observed had no attrition occurred (under the assumption that missingness depends only on observed covariates). Consequently, children with examination records in P2 or P3 were assigned weights to ensure their representativeness within the entire study population during these time points. The final weights were calculated by multiplying OW and IPCW. By applying the overlap weight, the average treatment effect for the overlap population (ATO) was estimated. The standardized mean difference (SMD) was used to assess the balance between the groups before weighting.
To explore the effect modification of sex, only-child status, monthly household income, parents’ higher education level, and main caregiver on the association between exposure and outcome, several subgroup analyses were conducted. In these analyses, an interaction term between exposure and the modifier variable was incorporated into the model, in addition to the main effects. No adjustments were made to account for multiple comparisons, as the study was exploratory in nature.
All statistical analyses and plots were performed using SAS 9.4 (SAS Institute) and R (version 4.3.1). A P < .05 was considered statistically significant.
Results
A total of 850 children had examination records in P1. Among them, 25 (2.9%) with caries in permanent teeth in P1 were excluded. Out of the 825 analyzed children, 698 (84.6%) and 790 (95.8%) had follow-up data in P2 and P3, respectively, while 16 children (1.9%) had no follow-up data in either P2 or P3. In total, 443 (53.7%) children who had caries experienced in their primary teeth in P1 were classified into the exposure group, while the remaining 382 (46.3%) were classified into the control group (Table 1). Children in the exposure group had a mean (SD) dmft score of 4.1 (2.9) and a median (Q1, Q3) dmft score of 3.0 (2.0, 6.0). The mean (SD) age of the study children was 6.2 (0.3) years at the time of entering primary school. Boys comprised 54.7% of the sample, with more than half (52.1%) having started primary school in 2019, and the remainder in 2018. The majority (74.6%) attended primary schools in the Tuen Mun district, followed by those in the Fan Ling district (19.4%) and the Kowloon district (6.1%). Most children (72.4%) were non-only children, while 45.6% of their parents had received a tertiary education or above, and more than half (50.7%) had household incomes exceeding HK$30,000 per month. Seventy percent of these children were primarily cared for by their mothers, and nearly one-fifth had been exposed to dental outreach services provided by the University of Hong Kong during their kindergarten years. Imbalances with estimated absolute SMDs >0.1 were observed between groups in terms of primary school location, birthplace, parents' higher education level, and monthly household income before weighting. Generally, children from the exposure group were more likely to be from primary schools located in the Kowloon district, to have been born in mainland China or other places, to have parents with lower education levels, and to have lower monthly household incomes. After applying overlap weighting, a perfect balance between groups regarding the adjusted confounders was achieved due to the property of overlap weighting. The distribution of the OW-weighted sample was found to be similar to that of the original sample with respect to these confounders.
Table 1.
Characteristics of the study children by exposure status before and after adjustment
| Unadjusted |
Overlap weight adjusted |
|||||||
|---|---|---|---|---|---|---|---|---|
| Characteristics | Total | Control | Exposure | SMD | Total † | Control | Exposure | SMD |
| n (%) | 825 | 382 (46.3) | 443 (53.7) | 387.9 | 193.9 (50.0) | 193.9 (50.0) | ||
| Sex | ||||||||
| Male, n (%) | 451 (54.7) | 216 (56.5) | 235 (53.1) | 0.070 | 210.8 (54.3) | 105.4 (54.3) | 105.4 (54.3) | 0 |
| Female, n (%) | 374 (45.3) | 166 (43.5) | 208 (47.0) | 177.1 (45.6) | 88.6 (45.7) | 88.6 (45.7) | 0 | |
| Age entering primary school, mean (SD) | 6.2 (0.3) | 6.2 (0.3) | 6.2 (0.3) | 0.083 | 6.2 (0.2) | 6.2 (0.2) | 6.2 (0.2) | 0 |
| Year entering primary school | ||||||||
| 2018, n (%) | 395 (47.9) | 185 (48.4) | 210 (47.4) | 0.021 | 186.3 (48.0) | 93.1 (48.0) | 93.1 (48.0) | 0 |
| 2019, n (%) | 430 (52.1) | 197 (51.6) | 233 (52.6) | 201.6 (52.0) | 100.8 (52.0) | 100.8 (52.0) | 0 | |
| Primary school location | ||||||||
| Fan Ling, n (%) | 160 (19.4) | 67 (17.5) | 93 (21.0) | 0.087 | 74.0 (19.1) | 37.0 (19.1) | 37.0 (19.1) | 0 |
| Kowloon, n (%) | 50 (6.1) | 14 (3.7) | 36 (8.1) | 0.188 | 19.4 (5.0) | 9.7 (5.0) | 9.7 (5.0) | 0 |
| Tuen Mun, n (%) | 615 (74.6) | 301 (78.8) | 314 (70.9) | −0.182 | 294.5 (75.9) | 147.2 (75.9) | 147.2 (75.9) | 0 |
| Birthplace | ||||||||
| Hong Kong, n (%) | 773 (93.7) | 368 (96.3) | 405 (91.4) | 0.203 | 367.5 (94.8) | 183.8 (94.8) | 183.8 (94.8) | 0 |
| Mainland China or other, n (%) | 52 (6.3) | 14 (3.7) | 38 (8.6) | 20.4 (5.3) | 10.2 (5.3) | 10.2 (5.3) | 0 | |
| Only child | ||||||||
| Yes, n (%) | 228 (27.6) | 108 (28.3) | 120 (27.1) | 0.026 | 106.0 (27.3) | 53.0 (27.3) | 53.0 (27.3) | 0 |
| No, n (%) | 597 (72.4) | 274 (71.7) | 323 (72.9) | 281.9 (72.7) | 140.9 (72.7) | 140.9 (72.7) | 0 | |
| Parents’ higher education level | ||||||||
| Middle school or below, n (%) | 129 (15.6) | 46 (12.0) | 83 (18.7) | 0.185 | 57.0 (14.7) | 28.5 (14.7) | 28.5 (14.7) | 0 |
| High school, n (%) | 320 (38.8) | 140 (36.7) | 180 (40.6) | 0.082 | 152.2 (39.3) | 76.1 (39.3) | 76.1 (39.3) | 0 |
| Tertiary or above, n (%) | 376 (45.6) | 196 (51.3) | 180 (40.6) | −0.215 | 178.7 (46.1) | 89.3 (46.1) | 89.3 (46.1) | 0 |
| Monthly household Income | ||||||||
| HK$10,000 or below, n (%) | 42 (5.1) | 15 (3.9) | 27 (6.1) | 0.099 | 18.7 (4.8) | 9.4 (4.8) | 9.4 (4.8) | 0 |
| HK$10,000 - $19,999, n (%) | 155 (18.8) | 59 (15.5) | 96 (21.7) | 0.160 | 70.0 (18.0) | 35.0 (18.0) | 35.0 (18.0) | 0 |
| HK$20,000 - $29,999, n (%) | 209 (25.3) | 87 (22.8) | 122 (27.5) | 0.110 | 99.8 (25.7) | 49.9 (25.7) | 49.9 (25.7) | 0 |
| HK$30,000 - $39,999, n (%) | 138 (16.7) | 60 (15.7) | 78 (17.6) | 0.051 | 66.1 (17.1) | 33.1 (17.1) | 33.1 (17.1) | 0 |
| HK$40,000 - $59,999, n (%) | 152 (18.4) | 86 (22.5) | 66 (14.9) | −0.197 | 72.7 (18.7) | 36.4 (18.7) | 36.4 (18.7) | 0 |
| HK$60,000 or above, n (%) | 129 (15.6) | 75 (19.6) | 54 (12.2) | −0.206 | 60.6 (15.6) | 30.3 (15.6) | 30.3 (15.6) | 0 |
| Main caregiver | ||||||||
| Mother, n (%) | 578 (70.1) | 260 (68.1) | 318 (71.8) | 0.081 | 272.2 (70.2) | 136.1 (70.2) | 136.1 (70.2) | 0 |
| Father, n (%) | 35 (4.2) | 13 (3.4) | 22 (5.0) | 0.078 | 14.2 (3.7) | 7.1 (3.7) | 7.1 (3.7) | 0 |
| Grandparents, n (%) | 111 (13.5) | 56 (14.7) | 55 (12.4) | −0.066 | 53.1 (13.7) | 26.6 (13.7) | 26.6 (13.7) | 0 |
| Others, n (%) | 101 (12.2) | 53 (13.9) | 48 (10.8) | −0.093 | 48.4 (12.5) | 24.2 (12.5) | 24.2 (12.5) | 0 |
| Exposure to the HKU kindergarten outreach | ||||||||
| Yes, n (%) | 153 (18.6) | 66 (17.3) | 87 (19.6) | 69.1 (17.8) | 34.6 (17.8) | 34.6 (17.8) | 0 | |
| No, n (%) | 672 (81.5) | 316 (82.7) | 356 (80.4) | 0.061 | 318.8 (82.2) | 159.4 (82.2) | 159.4 (82.2) | 0 |
SMD = standardized mean difference. The overlap weight was adjusted for the child’s sex, year of entry into primary school, birthplace, only child status, parents’ level of higher education, monthly household income, main caregiver, and exposure status to the kindergarten dental outreach service provided by the Faculty of Dentistry of the University of Hong Kong.
In P2, 5.7% (95% CI: 4.1-7.7) of the study children developed caries in their permanent teeth (Table 2). The mean DMFT score was 0.08 (95% CI: 0.05-0.11) (Table 3). In the exposure group, the incidence was 7.7% (95% CI: 5.2-10.8), which was nearly double the incidence of 3.5% (95% CI: 1.5-5.5) in the control group (adjusted RD 3.6% [95% CI: 0.3-7.0; P = .033]; adjusted RR 2.0 [95% CI: 1.0-4.1; P = .047]) (Fig. 1A). Similarly, the mean DMFT score was 0.11 (95% CI: 0.07-0.14) in the exposure group, which was nearly double the mean of 0.04 (95% CI: 0.01-0.08) in the control group (adjusted MD: 0.05 [95% CI: 0.00-0.09; P = .040]; adjusted MR 2.1 [95% CI: 1.0-4.4; P = .055]) (Fig. 1B).
Table 2.
Caries incidence in permanent teeth by exposure status among primary school children in P2 and P3
| Adjustment | Grade | Group | Proportion (95% CI) | RD (95% CI) | P value | RR (95% CI) | P value |
|---|---|---|---|---|---|---|---|
| Not adjusted | P2 | Control (reference) | 3.5 (1.5, 5.5) | 4.2 (0.9, 7.6) | .014 | 2.2 (1.1, 4.4) | .021 |
| Exposure | 7.7 (5.2, 10.8) | ||||||
| Overall | 5.7 (4.1, 7.7) | ||||||
| P3 | Control (reference) | 4.9 (2.9, 7.6) | 10.5 (6.4, 14.6) | <.001 | 3.1 (1.9, 5.2) | <.001 | |
| Exposure | 15.4 (12.1, 19.2) | ||||||
| Overall | 10.5 (8.5, 12.9) | ||||||
| Overlap weight adjusted | P2 | Control (reference) | 3.5 (1.9, 6.4) | 3.6 (0.3, 7.0) | .033 | 2.0 (1.0, 4.1) | .047 |
| Exposure | 7.2 (5.0, 10.3) | ||||||
| Overall | 5.3 (3.9, 7.3) | ||||||
| P3 | Control (reference) | 5.2 (3.3, 8.3) | 9.6 (5.4, 13.8) | <.001 | 2.8 (1.7, 4.8) | <.001 | |
| Exposure | 14.8 (11.8, 18.7) | ||||||
| Overall | 10.0 (8.1, 12.4) |
RD = risk difference. RR = risk ratio. The overlap weight was adjusted for the child’s sex, year of entry into primary school, birthplace, only child status, parents’ level of higher education, monthly household income, main caregiver, and exposure status to the kindergarten dental outreach service provided by the Faculty of Dentistry of the University of Hong Kong.
Table 3.
Caries severity in permanent teeth by exposure status among primary school children in P2 and P3
| Adjustment | Grade | Group | Mean (95% CI) | MD (95% CI) | P value | MR (95% CI) | P value |
|---|---|---|---|---|---|---|---|
| Not adjusted | P2 | Control (reference) | 0.04 (0.01, 0.08) | 0.06 (0.01, 0.11) | .019 | 2.4 (1.2, 5.0) | .019 |
| Exposure | 0.11 (0.07, 0.14) | ||||||
| Overall | 0.08 (0.05, 0.11) | ||||||
| P3 | Control (reference) | 0.07 (0.02, 0.12) | 0.15 (0.08, 0.22) | <.001 | 3.2 (1.9, 5.4) | <.001 | |
| Exposure | 0.22 (0.17, 0.26) | ||||||
| Overall | 0.15 (0.12, 0.19) | ||||||
| Overlap weight adjusted | P2 | Control (reference) | 0.04 (0.02, 0.07) | 0.05 (0.00, 0.09) | .040 | 2.1 (1.0, 4.4) | .055 |
| Exposure | 0.10 (0.06, 0.13) | ||||||
| Overall | 0.07 (0.05, 0.09) | ||||||
| P3 | Control (reference) | 0.07 (0.04, 0.10) | 0.13 (0.07, 0.19) | <.001 | 2.9 (1.7, 5.0) | <.001 | |
| Exposure | 0.21 (0.15, 0.26) | ||||||
| Overall | 0.13 (0.11, 0.17) |
MD = mean difference. MR = mean ratio. The overlap weight was adjusted for the child’s sex, year of entry into primary school, birthplace, only child status, parents’ level of higher education, monthly household income, main caregiver, and exposure status to the kindergarten dental outreach service provided by the Faculty of Dentistry of the University of Hong Kong.
Fig. 1.
Severity and incidence of caries in permanent teeth over time among primary school students. dmft = decayed teeth, missing teeth due to caries, or filled teeth in primary teeth. DMFT = decayed teeth, missing teeth due to caries, or filled teeth in permanent teeth. Figures 1A andB are for the outcomes of caries incidence and caries severity, respectively. Overlap weighting was used to adjust for the child’s sex, year of entry into primary school, birthplace, only child status, parents’ level of higher education, monthly household income, main caregiver, and exposure status to the kindergarten dental outreach service provided by the Faculty of Dentistry of the University of Hong Kong.
In P3, the incidence of caries experience in permanent teeth and the mean DMFT score increased to 10.5% (95% CI: 8.5-12.9) and 0.15 (95% CI: 0.12-0.19), respectively, among the study children, both of which nearly doubled compared to P2 (Tables 2 and 3). In the exposure group, this incidence increased almost linearly to 15.4% (95% CI: 12.1-19.2), which was nearly 3 times that of 4.9% (95% CI: 2.9-7.6) in the control group (adjusted RD 9.6% [95% CI: 5.4-13.8; P < .001]; adjusted RR 2.8 [95% CI: 1.7-4.8; P < .001]). Similarly, the mean DMFT score was 0.22 (95% CI: 0.17-0.26) in the exposure group, which was nearly 3 times higher as that of 0.07 (95% CI: 0.02-0.12) in the control group (adjusted MD 0.13 [95% CI: 0.07-0.19; P <0.001]; adjusted MR 2.9 [95% CI: 1.7-5.0; P < .001]).
According to the subgroup analyses, it was found that 2 factors, namely sex and only-child status, significantly modified the association between caries experience in primary teeth and caries experience in permanent teeth in P3 (Fig. 2). For boys in P3, the adjusted incidence of caries in permanent teeth was 16.9% in the exposure group, which was 15.4% (95% CI: 10.1-20.7) higher than that of 1.5% among those in the control group. In contrast, the corresponding RD among girls was only 2.7% (95% CI: –3.9 to 9.3). The interaction term for RD was found to be statistically significant (P = .004), as were those for RR (P = .002), MD (P = .033), and MR (P = .001).
Fig. 2.
Subgroup analysis among the weighted sample in P3. Figures 2A and B are for the caries incidence and caries severity, respectively. Overlap weighting was used to adjust for the child’s sex, year of entry into primary school, birthplace, only child status, parents’ level of higher education, monthly household income, main caregiver, and exposure status to the kindergarten dental outreach service provided by the Faculty of Dentistry of the University of Hong Kong.
Among only children, the incidence was 5.7% in the exposure group and 5.5% in the control group. However, among non-only children, 18.3% of those in the exposure group developed caries in their permanent teeth in P3, compared to only 5.1% in the control group. The estimated RD was 13.1% (95% CI: 8.0-18.3) among non-only children, which was significantly higher than that of 0.2% (95% CI: –6.5 to 6.9) among only children (P = .003). Although the interaction term was not statistically significant for RR (P = .069), it was found to be significant for both MR (P = .044) and MD (P < .001).
Discussion
In this study, it was found that primary school children with caries experience in their primary teeth in P1 were associated with a nearly twofold and threefold risk of developing caries in their permanent teeth after 1 year and 2 years, respectively. A stronger association was observed in boys than in girls, and in non-only children compared to only children.
The early years of primary school play a crucial role in the development of caries in permanent teeth. The study results revealed a nearly linear trend in both the incidence and severity of caries in permanent teeth from P1 to P3, showing an annual increase of approximately 5%. This finding is consistent with a study conducted in Malaysia, where the prevalence of DMFT increased from 4.6% to 10.3% to 15.1% from P1 (age 6 years) to P2 to P3, with an annual increase of around 5.%16 A significant association was found between caries in primary teeth and caries in permanent teeth on both the additive (difference) and multiplicative (ratio) scales. After adjusting for the measured confounding between the exposure and control groups, children with caries in primary teeth were 2.0 and 2.8 times as likely to experience caries in permanent teeth after 1 year and 2 years, respectively. Therefore, it is imperative to implement effective preventive measures for tooth decay among preschool children in order to mitigate the risk of later caries in their permanent teeth.1
Surprisingly, the study findings revealed that only 1.5% of boys without caries experience in primary teeth developed caries in their permanent teeth in P3, which is notably lower compared to the rest of the study population. It is acknowledged that sex-based differences in tooth eruption timing and caries susceptibility are important considerations. While it has been reported that females are more susceptible to caries than males from childhood, the association between sex and caries among children remains inconclusive.5,6,17,18 Sex is often studied more as a potential risk factor for caries rather than as an effect modifier for the association between caries in primary teeth and permanent teeth. Despite the later eruption of permanent teeth in boys, leading to a shorter exposure of their teeth to the cariogenic oral environment, 16.9% of boys with caries in their primary teeth still developed caries in their permanent teeth, which was even greater than that in girls.19 A potential explanation is that boys who have ‘survived’ caries attack in primary teeth may be less susceptible to caries due to unobserved confounders (eg, behavioral factors, bacterial load, or genetic differences). Alternatively, it was observed that 9.6% of girls developed caries in their permanent teeth even after ‘surviving’ the caries attack on their primary teeth, indicating that girls may develop caries in permanent teeth regardless of their caries experience in the primary teeth. In other words, if both boys and girls had no caries experience in their primary teeth, girls are still more likely to develop caries in their permanent teeth than boys.
A larger family size has been identified as a risk factor for caries among preschool children.20, 21, 22 In this study, the incidence of caries in permanent teeth was 18.3% among children with siblings and previous experience of caries in primary teeth in P3, which is higher than the range of 5% to 6% observed in the rest of the study population. This disparity may be attributed to a lack of attention from caregivers towards these particular children with siblings. Additionally, it is possible that an increase in family size could lead to a greater likelihood of transmission of mutans streptococci among family members.23,24 Further research is needed on potential effect modifiers of the association between caries in primary teeth and caries in permanent teeth, allowing for more targeted preventive measures to be implemented.
This study possesses several strengths. By utilizing clinical examination data derived from a real-world setting within a causal inference framework, the findings of this study are more relevant to a broader population compared to evidence generated from traditional specialized research settings.13,25 Additionally, the target population created by overlap weighting might be of greater policy interest. This is because overlap weighting places a stronger emphasis on comparing children who are at clinical equipoise.26,27
This study has several limitations. First, the sample was not randomly selected and thus might not be representative. Clinicians and policymakers should evaluate whether this study population's characteristics align with their target populations when considering the applicability of the study results. Second, while previous studies have reported that the location, severity, and depth of caries lesions in primary teeth were associated with caries development in permanent teeth, these variables were not collected and evaluated in this study.6,28 Furthermore, even though our analysis adjusted for some family-level factors, certain specific variables, such as parental marital status and detailed living arrangements (eg, co-residence with grandparents), were not collected. However, our included covariates can serve as proxies for family-level heterogeneity. Third, there may be measurement errors in caries, as the examinations were primarily used for treatment purposes rather than research purposes. Although calibration between examiners was conducted, inter-examiner reliability was not reported in practice settings. Fourth, while only-child status was identified as an effect modifier, detailed birth order information was not collected in our study. Although only-child status captures a fundamental dimension of birth order and family resource allocation, future studies with larger sample sizes and detailed birth sequence data are needed to elucidate the mechanisms underlying sibling-related effect modification further. Fifth, it is important to note that the incidence of caries in permanent teeth might be affected by restricted access to dental care for children due to the COVID-19 pandemic. Lastly, the follow-up only extended to P3, which may be considered too early for all permanent teeth to have erupted in primary school children. However, a shorter time interval might reduce the likelihood of being influenced by time-varying confounding factors (eg, new caries in permanent teeth may be impacted by caries prevention measures, which are also influenced by caries experience in primary teeth).11,29
Conclusions
Primary school children who have experienced caries in their primary teeth during P1 are nearly 2 to 3 folds in likelihood to develop caries in their permanent teeth in P2 and P3, respectively. This association is particularly strong among boys and non-only children. Preventing caries in primary teeth is crucial for avoiding caries in permanent teeth. Risk-stratified preventive approaches targeting boys and non-only children may optimize intervention outcomes in disrupting the link between caries in primary teeth and caries in permanent teeth.
Author contributions
Sicheng Wu: conception and design of the study, acquisition of data, analysis and interpretation of data, drafting the article and revising it critically for important intellectual content. Edward Chi Man Lo: conception and design of the study, analysis and interpretation of data, and revising the article critically for important intellectual content. Chun Hung Chu: conception and design of the study, revising the article critically for important intellectual content. May Chun Mei Wong: conception and design of the study, analysis and interpretation of data, and revising the article critically for important intellectual content. All authors gave final approval of the version to be submitted.
Conflict of interest
None disclosed.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the General Research Fund, Research Council of Hong Kong SAR, China; Project No. 17100220.
References
- 1.AAPD Policy on early childhood caries (ECC): consequences and preventive strategies. Ref Manual Pediatr Dent Chicago, Ill: Am Acad Pediatr Dent. 2021:81–84. [Google Scholar]
- 2.Lam P.P.Y., Chua H., Ekambaram M., Lo E.C.M., Yiu CKY. Does early childhood caries increase caries development among school children and adolescents? A systematic review and meta-analysis. Int J Environ Res Public Health. 2022;19(20) doi: 10.3390/ijerph192013459. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Al-Shalan T., Erickson P., Hardie N. Primary incisor decay before age 4 as a risk factor for future dental caries. Pediatr Dent. 1997;19:37–41. [PubMed] [Google Scholar]
- 4.Alm A., Wendt L-K, Koch G., Birkhed D., Nilsson M. Caries in adolescence–influence from early childhood. Commun Dent Oral Epidemiol. 2012;40(2):125–133. doi: 10.1111/j.1600-0528.2011.00647.x. [DOI] [PubMed] [Google Scholar]
- 5.Du Q., Yu M., Li Y., Du H., Gao W., Mei H., et al. Permanent caries experience is associated with primary caries experience: a 7-year longitudinal study in China. Commun Dent Oral Epidemiol. 2017;45(1):43–48. doi: 10.1111/cdoe.12257. [DOI] [PubMed] [Google Scholar]
- 6.Li Y., Wang W. Predicting caries in permanent teeth from caries in primary teeth: an eight-year cohort study. J Dent Res. 2002;81(8):561–566. doi: 10.1177/154405910208100812. [DOI] [PubMed] [Google Scholar]
- 7.Manton DJ. Child dental caries – a global problem of inequality. eClinicalMedicine. 2018;1:3–4. doi: 10.1016/j.eclinm.2018.06.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Wen P.Y.F., Chen M.X., Zhong Y.J., Dong Q.Q., Wong HM. Global Burden and Inequality of Dental Caries, 1990 to 2019. J Dent Res. 2021;101(4):392–399. doi: 10.1177/00220345211056247. [DOI] [PubMed] [Google Scholar]
- 9.Chen K.J., Gao S.S., Duangthip D., Lo E.C.M., Chu CH. Early childhood caries and oral health care of Hong Kong preschool children. Clin Cosmetic Investig Dent. 2019;11:27–35. doi: 10.2147/CCIDE.S190993. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Introduction of the School Dental Care Service: Department of Health of the Government of the Hong Kong SAR; Available from: https://www.schooldental.gov.hk/wsmile/en/aboutus_intro. Accessed July 1, 2024.
- 11.Organization WH. Oral health surveys: basic methods: World Health Organization; 2013.
- 12.Hernán M.A., Robins JM. CRC Boca Raton; FL: 2010. Causal inference. [Google Scholar]
- 13.Sherman R.E., Anderson S.A., Dal Pan G.J., et al. Real-world evidence—what is it and what can it tell us. N Engl J Med. 2016;375(23):2293–2297. doi: 10.1056/NEJMsb1609216. [DOI] [PubMed] [Google Scholar]
- 14.Wu S, Lo ECM, Chu CH, Wong MCM. Long-term Impact of Kindergarten Dental Outreach Service on School Students. J Dent Res. 2025;105(2):217–225. doi: 10.1177/00220345251347943. [DOI] [PubMed] [Google Scholar]
- 15.Li F., Morgan K.L., Zaslavsky AM. Balancing covariates via propensity score weighting. J Am Stat Assoc. 2018;113(521):390–400. [Google Scholar]
- 16.Masood M., Yusof N., Hassan M.I., Jaafar N. Longitudinal study of dental caries increment in Malaysian school children: a 5-year cohort study. Asia Pacific J Public Health. 2014;26(3):260–267. doi: 10.1177/1010539511420704. [DOI] [PubMed] [Google Scholar]
- 17.Sæthre-Sundli H.B., Wang N.J., Wigen TI. Do enamel and dentine caries at 5 years of age predict caries development in newly erupted teeth? A prospective longitudinal study. Acta Odontologica Scandinavica. 2020;78(7):509–514. doi: 10.1080/00016357.2020.1739330. [DOI] [PubMed] [Google Scholar]
- 18.Lukacs JR. Sex differences in dental caries experience: clinical evidence, complex etiology. Clin Oral Investig. 2011;15:649–656. doi: 10.1007/s00784-010-0445-3. [DOI] [PubMed] [Google Scholar]
- 19.Lukacs J.R., Largaespada LL. Explaining sex differences in dental caries prevalence: saliva, hormones, and “life-history” etiologies. Am J Human Biol. 2006;18(4):540–555. doi: 10.1002/ajhb.20530. [DOI] [PubMed] [Google Scholar]
- 20.Mansbridge J.N. The influence of family size on the prevalence of dental caries in children. Arch Oral Biol. 1960;2(3):209–214. doi: 10.1016/0003-9969(60)90051-0. [DOI] [PubMed] [Google Scholar]
- 21.Wellappuli N., Amarasena N. Influence of family structure on dental caries experience of preschool children in Sri Lanka. Caries Res. 2012;46(3):208–212. doi: 10.1159/000337399. [DOI] [PubMed] [Google Scholar]
- 22.Fontana M., Jackson R., Eckert G., et al. Identification of caries risk factors in toddlers. J Dent Res. 2011;90(2):209–214. doi: 10.1177/0022034510385458. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Doméjean S., Zhan L., DenBesten P.K., Stamper J., Boyce W.T., Featherstone J.D. Horizontal transmission of mutans streptococci in children. J Dent Res. 2010;89(1):51–55. doi: 10.1177/0022034509353400. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Manchanda S., Sardana D., Liu P., Lee G.H., Lo E., Yiu CK. Horizontal transmission of streptococcus mutans in children and its association with dental caries: a systematic review and meta-analysis. Pediatr Dent. 2021;43(1):1E–12E. [PubMed] [Google Scholar]
- 25.Schwendicke F., Krois J. Data dentistry: how data are changing clinical care and research. J Dent Res. 2022;101(1):21–29. doi: 10.1177/00220345211020265. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Li F., Thomas L.E., Li F. Addressing extreme propensity scores via the overlap weights. Am J epidemiol. 2019;188(1):250–257. doi: 10.1093/aje/kwy201. [DOI] [PubMed] [Google Scholar]
- 27.Thomas L.E., Li F., Pencina MJ. Overlap weighting: a propensity score method that mimics attributes of a randomized clinical trial. JAMA. 2020;323(23):2417–2418. doi: 10.1001/jama.2020.7819. [DOI] [PubMed] [Google Scholar]
- 28.Skeie M., Raadal M., Strand G., Espelid I. The relationship between caries in the primary dentition at 5 years of age and permanent dentition at 10 years of age–a longitudinal study. Int J Paediatr Dent. 2006;16(3):152–160. doi: 10.1111/j.1365-263X.2006.00720.x. [DOI] [PubMed] [Google Scholar]
- 29.Mansournia M.A., Etminan M., Danaei G., Kaufman J.S., Collins G. Handling time varying confounding in observational research. BMJ. 2017;359 doi: 10.1136/bmj.j4587. [DOI] [PubMed] [Google Scholar]