This protocol describes a pragmatic, 2-arm, individually randomized controlled trial (RCT) conducted in the Netherlands. The study population was recruited at 9 WBCs in urban and suburban regions with a population of predominantly low to middle socioeconomic position (SEP). Figure 1 shows a flow chart of the study’s overall design. This paper has been prepared following the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) reporting guidelines [31] and the Template for Intervention Description and Replication (TIDieR) [32].

Prior to study initiation, the research protocol was submitted to the Medical Ethics Committee of the University Medical Centre Utrecht for review and approval. The committee provided ethics clearance (NL60021.041.17; file number 17-133/D) and stated that the research complied with the applicable rules and requirements of the Medical Research Involving Human Subjects Act (Dutch abbreviation: WMO) and with the ethics code for the conduct of research as set out in the national Code of Conduct for Scientific Integrity in the Netherlands [33]. Participation was voluntary, and participants could withdraw at any moment with no consequences. Participants signed an informed consent form before enrollment. The trial is registered in the Netherlands Trial Register (NL8737).

This study includes parent-child dyads with children aged 6 to 12 months. The study excludes dyads with (1) children in need of specialized health care for a physical, mental, or medical condition; (2) parents with insufficient knowledge of Dutch; and (3) parents who expected to move outside the region of the WBC within the duration of the study. In the case of twins, only 1 child is enrolled in the study.

Previous studies show that the average number of caries lesions in Dutch children at the age of 5 years is 4.5 (SD 4.5) [12]. This information was used for study size estimation using G*power 3 [34]. This showed that the intervention group needed at least 149 participants to demonstrate a reduction in cumulative caries incidence of at least 30% with a statistically significant Mann-Whitney U test allowing for statistical power (1-β) of .8 and a 2-sided α of .05 as a threshold for statistical significance. With an anticipated 30% loss to follow up, the aim was to include at least 200 participants per group.

With a 1:1 ratio, participating parent-child dyads were individually allocated to the intervention or control group. This allocation of the TOHI as an add-on to usual care was stratified for each WBC (9 in total) and educational level of the mother (classed as low, medium, or high) as an indicator of SEP. A computerized random sequence generator [35] with a block size of 4 to ensure equally filled strata in each group was used to allocate the intervention status to predefined unique ID numbers. Accordingly, an intervention allocation file was prepared, including intervention codes with blocked series of ranked unique participant IDs for the strata.

Research nurses who were not involved in the intervention allocation determined the eligibility of parent-child dyads attending the selected WBCs. First, all eligible dyads received verbal and written information about what participation in the study entailed. Then, the parents or persons with parental legal authority signed the informed consent form to confirm voluntary participation. Upon receipt of this form, the principal investigator assigned a unique patient ID based on the WBC and education level of the mother. The corresponding intervention status was retrieved from the prepared allocation files of an independent data administrator. Next, the principal investigator contacted the parents to inform them about the allocated intervention and to ensure that they understood all the information about their participation.

The research staff involved in clinical assessments are kept blinded to the nature of the allocated treatment. Due to the nature of the intervention, blinding of participating parent-child dyads has not been possible. To maintain the blinding of outcome assessors, parents are repeatedly instructed not to talk about the intervention they received with the assessors. The success of the blinding of the outcome assessors is tested by asking them the following question: “To which group is this child assigned according to your opinion?” Possible answers are as follows: “Control group: I know for sure,” “Intervention group: I know for sure,” “Control group: I think so,” “Intervention group: I think so,” and “I do not know” [36,37].

The MRC guidance on developing and evaluating complex clinical health care interventions was utilized in development and evaluation of the TOHI [30]. Prior to this study, the feasibility of the TOHI was investigated in a small-scale feasibility study and evaluated through semistructured interviews with stakeholders, including oral health professionals, OHCs, youth health care professionals, and parents. The resulting findings were instrumental in the study design and were used to refine the TOHI and adapt it to participating WBCs in the trial. Textbox 1 presents the conceptual framework for the TOHI. A full description of the TOHI, following TIDieR reporting guidelines [32], is available in Multimedia Appendix 1. The different components of the TOHI are described below.

In both the intervention and the control group, all dyads received or are receiving preventive child health care at the WBCs, which are regulated under the Public Health Act 2008. When a baby is registered in a municipality in the Netherlands, parents receive an invitation from the nearest WBC for a series of at least 13 appointments between birth and the age of 4 years. The goals of this preventive health care program are to monitor growth and development, detect health and social problems (or risk factors) at an early stage, screen for metabolic conditions and hearing in the newborn, deliver the national vaccination program, and provide advice and information on health [14]. According to the Public Health Act, this advice includes brief oral health promotion, such as the use of fluoridated toothpaste (at 500 ppm-750 ppm) and the advice to visit an oral health professional within the first year of life (Dutch Public Health Act, Article 6, paragraph 1, Dutch Public Health Decree). Dyads enrolled in the intervention group receive the TOHI in addition to this preventive health care program. Dyads enrolled in the control group receive only the usual care that all children receive at all WBCs in the Netherlands.

The TOHI builds on four components: (1) visiting an OHC at a WBC, (2) training OHCs to work accord to the NOCTP method, (3) using MI, and (4) using HAPA. The general understanding underlying the TOHI is that caries is a localized process and to a large extent can be prevented by brushing the teeth with fluoride toothpaste and lowering the frequency of sugar intake. With this understanding in mind, the OHC aims to create awareness among parents that through establishing adequate oral health behavior, they can effectively contribute to preventing caries in their children. The standardized protocol of the NOCTP method contributes to a uniform and transparent approach for OHCs and parents [19]. The method assigns 1 point based on parental cooperation and motivation, measured by the child’s oral hygiene and whether they follow the key oral health messages of the Ivory Cross national guidelines [38], and 1 point for active or incipient caries or lesions (Table 1). MI is used in the TOHI to elicit parents’ internal motivation and explore barriers and facilitators for desired oral health behavior. Finally, HAPA is a hybrid behavioral model that explicitly outlines 2 different phases of health behavior change: a motivational, intention-forming phase and a volitional phase in which translation of intention into action takes place [25]. Guided by HAPA, the OHCs focus on motivational and self-regulatory processes, such as self-efficacy, planning, and action control, which have been shown to be mediators in translating intentions into behaviors [29]. The OHC combines the NOCTP protocol with an intervention guided by HAPA and MI skills to determine which aspects of behavior need to be addressed for a tailored intervention.

Face-to-face, 10-to-20-minute appointments with an OHC were combined with the child health care appointments at the WBC at the ages of 6 or 8, 11, 15, 18, 24, 36, and 42 months. Every appointment with the OHC started with an assessment of the child’s caries risk, following the NOCTP protocol. Subsequently, the OHC adjusted the oral health promotion to this risk by targeting the mediating determinants according to the HAPA, depending on the parents’ health behavior phase. When the intention was formed, self-care agreements were made in the volitional phase that focused on planning and action control to translate intentions into behaviors. Additionally, ambivalent feelings and barriers or facilitators to self-care agreements were explored using MI to stimulate intrinsic motivation and reinforce parents in existing positive behavior. According to the NOCTP protocol, when an increased caries risk was detected, the OHC could decide to schedule an additional appointment. If incipient caries lesions were detected, parents were educated on how to clean the caries lesion, and the use of fluoride toothpaste was again emphasized. When dental treatment for caries was deemed necessary, the OHC referred the participants to a dental clinic.

A penlight and disposable materials (eg, cotton rolls to dry elements, dental mirrors, and wooden toothpicks to detect dental plaque on elements) were used for oral assessments by the OHC. In addition, the OHC used predefined paper patient records to ensure efficient and uniform reporting. These paper records contained age-related oral health topics to discuss, a caries risk assessment according to the NOCTP protocol, and an assessment of the parents’ oral health behavioral stage. The final caries risk and behavioral stage guided the OHC to follow-up actions, such as which mediating determinants from the HAPA should be addressed with the intervention (Multimedia Appendix 2). Parents in the intervention group received a specially developed TOHI booklet at baseline. This booklet was developed for 2 reasons. First, it contained oral health reports for each appointment with the OHC, which effectively communicated the child’s caries risk to the parents. These reports also covered agreed self-care actions to reduce caries risk. Together with the parents, the OHC examined and recorded the barriers to self-care actions and the mediating determinants of self-efficacy, action planning, and coping planning in these reports (Multimedia Appendix 3). Finally, the booklet covered essential oral health information and provided tips and tricks that the OHC could refer to.

Eligible OHCs for this study had a background in dentistry (they were dentists, dental hygienists, or dental assistants). They were prevention-minded, trained (or willing to be trained) in the NOCTP method, had good social communication skills, and had experience in pediatric dentistry. In addition, all OHCs received training in the TOHI delivered by a certified MI trainer and the principal investigator, who are both experienced in training oral health professionals and students in NOCTP, HAPA, and MI. A pediatric dietitian provided extra training on infant and toddler nutrition related to oral health. The first training session took place before the study started. Subsequently, training continued every 3 to 4 months for the duration of the study. During this training, the focus was on recognizing stages of health behavior, using HAPA, and developing advanced MI skills, such as recognizing and managing behavioral resistance and detecting ambivalence. Learning these skills takes time and is an ongoing process. Therefore, the training consisted of presenting practical examples of (successfully resolved) complex cases provided by the OHCs, role-playing, and peer feedback. In addition, the MI trainer provided individual feedback via self-recorded audio fragments that were uploaded by the OHCs. These audio fragments were assessed biannually with the Motivational Interviewing Treatment Integrity (MITI) [39] system by an independent MI expert. The results of the intervention fidelity will be analyzed and prepared for an original process evaluation article alongside the article reporting the results of the effectiveness study.

The outcome measures and additional data were collected with online questionnaires at baseline and with online questionnaires and clinical examinations when the children reached the ages of 24 and 48 months. Table 2 provides an overview of measurements and timepoints.

The clinical outcome measures were collected by trained and calibrated dentists or dental hygienists in an assessor-blinded manner. The assessors were experienced in the use of the International Caries Detection and Assessment System (ICDAS) codes [40] for teaching or research purposes. Therefore, the training consisted of the online International Caries Classification and Management System (ICCMS) core e-learning [45] system and an online interrater reliability test developed for this study. This test presented 20 expertly verified high-quality pictures with different merged ICDAS scores in random order. If necessary for calibration, additional norming meetings were organized to maintain the performance of the assessors for the duration of the study.

At 24 and 48 months, a knee-to-knee examination (Figure 2) is completed with the assistance of the parents or caregivers. The selected tooth surfaces are scanned from the incisal to the gingival edge with a toothpick to determine the presence of dental plaque. Next, the child’s teeth are cleaned and dried with cotton rolls or gauze to enable an adequate view of the tooth surfaces and systematic examination of each tooth for the presence of caries or incipient caries. Findings are recorded in an online case record form developed for this study by an assistant. If no assistant is available, an audio recording is used so that the assessor can work efficiently and speak their findings aloud, preventing recall bias in later reporting. In cases with findings such as dentine caries, caries-related inflammation, and eruption failure of the primary teeth, parents receive a letter with the advice to visit a dentist and the reasons for doing so. If necessary, the OHC provides a list of nearby dentists. After dental treatment, children in the intervention group return for follow up with the OHC.

The cumulative caries incidence at the age of 48 months will be assessed by merged ICDAS [40] scores. These merged scores differentiate teeth as being sound or having initial, moderate, or severe caries.

The average dmfs/dmft number and the caries incidence density (person-time at risk to the first cavity) will be calculated from the merged ICDAS scores. Consequences of untreated dental caries will be measured using the PUFA (visible pulpal involvement, ulceration caused by dislocated tooth fragments, fistula, and abscess) Index [46]. The presence of dental plaque will be determined with the 6-surface method of the Simplified Oral Hygiene Index [41]. Child cooperativeness during the outcome assessments will be documented by clinical assessors using the modified Frankl behavior scale [44].

Behavioral change over time will be assessed based on self-reported oral health behavior and psychosocial constructs from the HAPA, measured with online questionnaires at baseline, age 24 months, and age 48 months. The early childhood oral health–related quality of life will be measured using the Early Childhood Oral Health Impact Scale [43].

The demographic characteristics include the children’s age, sex, family composition, mother’s highest educational attainment according to the International Standard Classification of Education [47], and health literacy measured with the 16-item European Health Literacy Survey Questionnaire (HLS-EU-Q16) [48].

All analyses of the effectiveness of the TOHI will make use of the intention-to-treat principle. To prevent biased principal data analysis, concealment will be maintained during data analysis. The blinding of the study comparison will be broken only after the completion of data analysis for the evaluation of outcomes for children at the age of 48 months. After data cleaning and prior to data analysis, the amount and patterns of missing data will be explored. Sensitivity analysis will be used to explore the impact of missing data. The 3 following scenarios will be assumed: stability over time (last value carried forward or backwards to replace missing value), best case (largest positive change score observed to replace missing value), and worst case (largest negative change score observed to replace missing value).

The intervention effects for the cumulative caries incidence at 48 months and the average dmfs/dmft will be compared as the risk difference (additive scale), numbers needed to treat, and the relative risk (multiplicative scale), all with the 95% CI. Effects for subgroups (ie, WBC, education level of the mother, and health literacy) will be evaluated with interaction terms in a multivariate regression analysis. To determine incidence density and person-time at risk, the median and mean between-group differences in caries incidence (or the person-time to first caries lesion) will be based on a Kaplan-Meier analysis and incidence differences and ratios. In addition, when applicable, Cox proportional hazard regression analysis will be applied to explore relations with independent variables. Between-group differences at the age of 24 months will be assessed for the secondary outcomes, notably changes in the behavioral determinants (defined according to the HAPA), plaque, and self-reported oral health behavior, using linear regression analysis. Results from all analyses will be presented with the 95% CI. For all analyses, the SEP stratum will be used as the primary covariable, while sociodemographic characteristics for which a baseline difference is observed will be used as secondary covariables. Such postrandomization differences between groups will be explored using descriptive statistics. A difference of 20% for discrete variables, a full standard deviation (pooled) for normally distributed data, and a quartile for nonnormally distributed data will be used as thresholds for similarity.