“A System Dynamics Model for Long-term Aspect of Caries Preventive Interventions”

Introduction

Dental caries is still regarded as a burden of oral disease not only in Thailand but also around the world even though it is preventable [1]. Studies have shown a reduction in the prevalence of dental caries over decades in developed countries [2, 3]. However, in several developing countries, there has been an increasing trend in caries prevalence reported by some studies [3, 4]. The high prevalence of dental caries in developing countries is due to poor oral hygiene habits, low awareness of caries, growth in sugary consumption, insufficient access to preventive programs for dental health, and inadequate exposure to fluoride [3]. According to the 8^th National Oral Health Survey (NOHS) in Thailand in 2017, the prevalence of dental caries was still high at 52% in 12-year-olds, 62.7% in 15-year-olds, 91.8% in 35- to 44-year-olds, 98.5% in 60- to 74-year-olds, and 99.5% in 80- to 85-year-olds [5]. People having dental caries could suffer from pain, disturbance during eating, aesthetics, and functional problems [6]. Moreover, tooth loss due to caries was the most common consequence of caries [7].

Due to the burden of caries, the role of prevention and promotion of oral health is vital. Therefore, various preventive programs such as oral health education, toothbrushing, fluoride varnish, application of dental sealant, and so on have been designated for dental caries in both individuals and communities [8-11]. In Thailand, the Ministry of Public Health (MOPH) proposed preventive oral health programs for respective age groups including toothbrushing, fluoride varnish, dental sealant, oral health examination, and so on [12]. However, a long-term assessment of preventive interventions for conducting better decision-making and an alternative program is needed. Simulation models would help policymakers in making decisions for planning long-term processes. [13]. They have been widely used in health care and a study also reported models that were reliable to be used in dental caries interventions such as the Markov Model, System Dynamics Model (SDM), Microsimulation Model, and Decision-tree Model [14].

In this study, SDM was explored for caries preventive interventions intended by the Ministry of Public Health (MOPH) such as supervised toothbrushing (STB), sealant, and their combination. The interventions were primary preventive interventions for caries since they focused on preventing the occurrence of diseases in a healthy state before the process begins [15]. The System Dynamics Model (SDM) is a computer simulation to help the stakeholders in deciding the most admissible policy or intervention for the community [16]. It can be used as a cohort-based model and shows the interaction between the entities of the system by modeling the rate of change of the system [17, 18]. It also allows seeing feedback loops that emerge from the interaction [17, 18]. SDM should be considered as a cohort-based simulation model for the long-term aspect of the outcomes not only under individual preventive intervention but also the combination of the interventions.

Therefore, the objective of the study is to estimate the effectiveness of providing caries preventive interventions for 6- to 15-year-olds by the Ministry of Public Health (MOPH) namely supervised toothbrushing (STB), sealant, and combined STB+sealant, compared with the base case as a long-term aspect by conducting the System Dynamics Model (SDM).

Methodology

The modeling in this study focused on the permanent dentition age. The System Dynamics Model (SDM) involves two parts: qualitative and quantitative. The qualitative part showed the causal relationship of caries related and identifies the feedback loops among them by a causal loop diagram (CLD). The quantitative part translated the causal loop diagram (CLD) for evaluating the conditions quantitatively represented as stocks and flows diagram. Group model building (GMB) sessions including the researchers and experts in related fields were held three times to determine the final model. The experts are dentists with valuable experience and knowledge in dental public health and clinical fields. The activities such as identifying the time horizon, identifying the key variables, and their behaviors, formulating the causal loop diagram (CLD), and developing the stocks and flows diagram were involved in group model building (GMB) sessions. The setting interventions in the model were implemented interventions by the Ministry of Public Health (MOPH). Since the study was based on the effectiveness of the interventions, systematic reviews and meta-analyses of interventions were conducted first by the researchers for the effectiveness data to be used in the model. The Vensim DSS version 6.4 software was used for running the model. The model simulated the Thai population born in 2021 to 75 years old (average life span) to estimate the outcomes in age intervals under intervention scenarios. The total population was 678,243. The results are presented at the age of 15 years old when the interventions were discontinued and at 60 years old to see the long-term effect on the elderly.

Model structure

The causal loop diagram (CLD) presents the causal relationship of no caries, caries, related events of caries, and feedback loops among them, (Fig 1). A positive sign (+) in the diagram means that a variable adds to another one and a negative sign (-) that it subtracts from the other [21]. One balancing loop and two reinforcing loops were found in the diagram. A balancing loop means that a positive change in a variable lead to a pushback in the opposite direction and if the direction of variables keeps in the same direction, it is a reinforcing loop [21]. A balancing loop in the diagram shows that an increase in the no caries leads to an increase in caries and missing teeth; on the other hand, an increase in the missing teeth leads to a decrease in the no caries. When the intervention is given, it would reduce developing caries, and missing teeth, and on one side, the no caries would increase. Reinforcing loops show that an increase in caries tends to increase in the treated cases due to caries (restoration and endodontic) and recurrent caries from the treated cases. When the intervention is set, it would decrease developing caries, treated cases, and recurrent caries from the treated cases.

• Download figure
• Open in new tab

Fig 1. Causal loop diagram (CLD)

CLD was translated into the stocks and flows diagram, Fig 2, to evaluate the outcomes quantitatively. The square blocks in the diagram are denoted as stocks and the arrow symbols represent the flows. The stocks are defined as states or levels of variables moving through the system [13]. The flows control the movement of stocks or determine the changes in the stocks by in and out valves of the stocks [13]. The quantitative outcomes are attributed to the stocks in the model. The fractions affect the rate of flows, for example, the caries development fraction influences the rate of developing caries, and the fraction of restorative treatment affects the rate of untreated caries to restoration. It is assumed that when the intervention is set, the interventions could reduce the caries development fraction tending to slow the rate of developing caries as shown in Fig 2.

• Download figure
• Open in new tab

Fig 2. Stocks and Flows diagram

Results

Table 3, the population with no caries under combined STB+sealant is the highest, 153,042 (22.56%) than in other scenarios: dental sealant, 141,097 (20.80%), supervised toothbrushing (STB), 133,306 (19.65%) and base-case, 112,405 (16.57%) at the age of 15 years. It increased by 36.2% under the combined STB+sealant, 25.5% in sealant, and 18.6% in STB compared to the base case. The untreated caries population is the lowest under combined STB+sealant, 257,655 (37.99 %) followed by sealant, 264,507 (38.99%) and STB, 270,870 (39.94%) whereas the base-case is the highest, 280,244 (41.32%). Compared to the base case, it decreased by 8.1% in combined STB+sealant, 5.6% in sealant, and 3.3% in STB. The population with missing teeth and populations with restoration and endodontic treatment are also the highest in base-case and lowest in combined STB+sealant followed by sealant and STB. Compared to the base case, the population with restoration decreased by 6.1%, 4.4%, and 3.8% while the population with endodontic decreased by 4.7%, 3.3%, and 3.0% in combined STB+sealant, sealant, and STB respectively. The missing teeth population decreased by 6.7% in combined STB+sealant, 4.8% in sealant, and 4.3% in STB compared to the base case.

Table 4, at the age of around 60 years, the no caries population is the lowest population among other populations. Even though there is no significant difference between the scenarios, the highest, 808 (0.12%) is in combined STB+sealant, followed by sealant, 496 (0.073%), and supervised toothbrushing (STB), 311 (0.046%) compared to the base case, 127 (0.019%). Whereas the untreated caries population increased by 4.9% in combined STB+sealant, 3.2% in sealant, and 2.1% in STB compared to the base case. For the population with restoration, it increased by 1.6%, 1.1%, and 0.9% while the population with endodontic decreased by 0.66%, 0.43%, and 0.28% under combined STB+sealant, sealant, and STB respectively compared to the base case. The population with missing teeth decreased by 1.2% in combined STB+sealant, 0.79% in sealant, and 056% in STB from the base case.

Figure 3 shows the population with no caries of permanent dentition age, from 6 years old to elderly age. It seems that the no caries population is the highest under combined STB+sealant followed by sealant and supervised toothbrushing (STB) and lowest under base-case. It is still higher under intervention scenarios than the base case until the adult age and the differences can be seen prominently in the graph before the age of 30 years. Nevertheless, it cannot be seen differently between interventions and the base case after the age of 40 years. Besides, it is seen that the population with no caries significantly decreases over time.

• Download figure
• Open in new tab

Fig 3. Population with no caries

Figure 4 presents the untreated caries population from 6 years old to elderly age. It shows that it is the lowest under combined STB+sealant followed by sealant and STB then base-case. It can be seen clearly in the graph until the age of 20 years. However, at ages over 20 years, the caries population under interventions is not lower than the base case. Further, the graph shows that the caries population is slightly decreased over the years.

• Download figure
• Open in new tab

Fig 4. Population with untreated caries

Discussion

When establishing the interventions in the model, the effective rates of the interventions were identified and adjusted from the coverage rates in Thailand. Depending on the coverage rate, the effective rate of supervised toothbrushing (STB) was only 9.5%. Although the coverage rate is high, the effective rate was not high, and the caries development rate was not significantly reduced from the base case. While the sealant and combined STB+sealant were quite good in effective rates, the coverage rates were not satisfactory, and their adjusted effective rates were 15.7% and 19.8% respectively. It is considered that the effective rates were not very high due to quite low coverage rates. Therefore, the reduction of caries development rate under these two interventions was also not significantly different from the base case.

Since the effective rate of supervised toothbrushing (STB) is not very satisfactory, this must be emphasized in the intervention. A study indicated that STB has a favorable effect on preventing caries [25]. In this study, the children received an intensive preventive program including the Bass method of brushing and finger flossing monitored by dentists. During the program, daily practicing toothbrushing with fluoride toothpaste in addition to flossing was carried out after lunch on every school day under the supervision of school nurses for one semester (20 weeks). This program lasted about 10 years. Therefore, the type of resource used, and the scheme would be crucial for the intervention, and a comprehensive approach to toothbrushing should be contemplated for more beneficial effects. In the aspect of sealant, applying the resin-based sealant on the pit and fissure permanent teeth has a valuable effect on caries. However, more resources would be needed than for STB, and the availability of resources needs to be of concern.

Based on the simulation, combined STB+sealant is better than each of these assigned separately. Supervised toothbrushing (STB) is proposed not only to prevent the occurrence of dental caries but also to develop oral hygiene habits [9] while the dental sealant is especially focused on molar teeth with a quite good preventing effect on caries [11]. Therefore, the combination of the interventions would be better intervention. It can be seen that at the age of 15 years, under the combined intervention, the no caries population is the highest, and the populations with untreated caries, restoration, endodontic treatment, and missing teeth are the lowest compared with each intervention administered alone, as shown in Table 3. It also seems that the no caries population remains the highest under the combined intervention among the adult and elderly ages.

Even though the no caries population is still higher than in the base case until the elderly age under intervention scenarios, the untreated caries population gradually increased in intervention scenarios over the base case scenario after 20 years old. It is regarded that the interventions that are allocated among the younger ages could no longer reduce the untreated caries population in the adult ages. Therefore, the interventions have a caries reduction effect for only about 15 years from when they are started. It shows that the consequences of caries such as treated cases due to caries and missing teeth under intervention scenarios are not very different from the base case in the elderly ages since the interventions have no positive effect on caries from the adult age.

A study by Splieth et al., (2008) [26] developed the System Dynamics Model (SDM) and reported the combination of fluoride toothpaste, fluoride salt, and fluoride gel was the most effective among fluoride regimes for caries prevention and was also cost-effective. Another study (Hirsch et al., 2012) [27] reported that the SDM estimated the combined early childhood caries (ECC) interventions were the most significant approach for caries reduction and costs. The present study is consistent with these previous studies in that combined interventions are preferred based on SDM simulation. Two studies (Hirsch et al., 2012, Edelstein et al., 2015) projected caries reduction for 10 years by simulating the SDM [27, 28]. It is analogous to this study where the reduction of caries persists for about 15 years. Nevertheless, the present study focused on the long-term simulation of more than the projected 10 years. Furthermore, the previous studies mentioned only the effective rates of the interventions whereas this study concerned not only the effective rates of the interventions conducted by meta-analyses but also the coverage rates in Thailand.

In this study, the data for the cause of missing teeth due to dental caries or periodontal disease was impossible to separate depending on the available data source, Health Data Centre (HDC). Therefore, the data of all causes of missing teeth from HDC were used and it was assumed that the missing teeth may be mostly due to caries among the younger ages, and from both caries and periodontal diseases during the adult and elderly ages [29, 30]. Therefore, if the interventions focus not only on caries but also on the periodontal disease with huge effectiveness, the number of the population with missing teeth could be reduced more than that indicated in this simulation study. In contrast, it would not be different if the focus were to be on caries alone. Moreover, considering each socioeconomic factor related to dental caries as a variable in the model is complicated. It is assumed that developing caries is prevented by reducing the risk factors that could be influenced by interventions. Due to the above limitations, this study may require accomplishing a more distinct quantitative model prediction.

Since it has been allowed that combined STB+sealant is more effective to implement, expanding the coverage rate of the intervention must be considered. This is because the effective rate of the combined intervention is quite low after adjusting the coverage rate although the combined intervention is good in effect. Besides, since the interventions afforded among younger ages have meager effectiveness in adult age and elderly age, furnishing some effective interventions beyond the younger ages should be considered.

The System Dynamics Model (SDM) may be used for long-term assessment in analyzing the effectiveness of interventions when projecting the interventions for improving oral health. Moreover, the SDM may help in comparing the cases as “what if” scenarios to anticipate the best case and poor case of other services or policies. Since the study did not encompass the costs, cost-effectiveness, or cost-saving, these aspects should be appraised in further studies.

Conclusion

Among the interventions provided by the Ministry of Public Health (MOPH) during the age of 6-15 years, the combined STB+sealant is the best intervention compared with each applied separately. It is supposed that the interventions given between the ages of 6-15 years old can reduce caries for 15 years from when they are started but have no beneficial effect on the population during the elderly age. Moreover, System Dynamics Model (SDM) would help decision-makers anticipate the best interventions or policies by simulating scenarios.

Supporting information

S1 Fig 1. Causal loop diagram (CLD).

S2 Fig 2. Stocks and Flows diagram.

S3 Fig 3. Population with no caries.

S4 Fig 4. Population with untreated caries.

S5 Table 1. Uncertainty analysis of the population with no caries at the age of 15 years

S6 Table 2. Uncertainty analysis of the population with no caries at the age of 60 years

S7 Table 3. Simulation results when the age of 15 years

S8 Table 4. Simulation results when the age of 60 years