A Simulation Model for Designing Effective Interventions in
Early Childhood Caries
Gary B. Hirsch SM, Independent Consultant and Creator of Learning Environments
Wayland, MA GBHirsch@ comcast.net
Burton L. Edelstein DDS MPH, Columbia University, New York, and Children’s Dental
Health Project, Washington DC ble22@columbia.edu
Theresa Anselmo MPH, BSDH, RDH, San Luis Obispo County Health Agency.
(Formerly with the Colorado Department of Public Health and Environment, Denver CO)
tanselmo@ co.slo.ca.us
Marcy Frosh JD, Children’s Dental Health Project, Washington DC mfrosh@ cdhp.org
Abstract
Dental caries in primary teeth of children 5 years of age or younger is one of the major
health problems in the United States, especially for low-income children. This paper
presents a framework for assessing the impact of various programs designed to reduce
the prevalence and consequences of Early Childhood Caries. The paper describes a
System Dynamics simulation model of the population of children 0-5 years old in
Colorado. Results of simulations with a number of individual interventions and
combined strategies are presented and program costs and savings in treatment costs are
compared.
Introduction
Dental caries in primary teeth of children 5 years of age or younger is still
one of the major health problems in the United States, especially for low-
income children. This largely preventable disease continues to affect many
children in lower socioeconomic strata and many ethnic minorities. Poor
oral health leads to chronic pain that affects a child's ability to chew food,
thrive, and speak, as well as their psychological well being. One of the
measurable impacts of severe dental disease in young children is the general
medical condition referred to as " failure to thrive." Reports of children with
severe dental caries and inappropriately low body weight have been reversed
after completing dental care. (taken from C M Jones, et al, 2000)
This paper presents a framework for assessing the impact of various programs designed
to reduce the prevalence and consequences of Early Childhood Caries. The paper
describes a System Dynamics simulation model of the population of children 0-5 years
old in Colorado. The development and initial implementation of the model was a joint
effort of the Children’s Dental Health Project (CDHP) and the Oral Health Unit of the
Colorado Department of Public Health and Environment (CDPHE). The model is
designed to be generic and, with the appropriate data inserted, could represent any state
or large city, county, or metropolitan area.
A model such as this one is needed to provide a better idea of the long-term and
cumulative effects of different programs on a population of children. Interventions
implemented at the same point in time can have very different effects over time
depending on the age and income groups they are targeted at, their efficacy in reducing
prevalence, and inherent time delays before their impact is realized. Combinations of
interventions have even more complex effects that cannot be readily anticipated, but
often represent the most effective strategies.
This paper will present the model, describe the various data sources used in its
quantification, and present the results of simulations with a number of different
interventions and combinations of those interventions. The value of these simulation
results is not to provide forecasts, but to help compare interventions for their relative
impacts. In addition to calculating reductions in prevalence, dmf scores, and fraction of
children with untreated decay, the model also estimates reductions in restorative care
costs that may be possible with the application of preventive interventions.
System Dynamics (SD) has a long history of applications to health care delivery and
population health. (Homer and Hirsch, 2006) A comprehensive model of dental care and
oral health was developed in 1975 for the Division of Dentistry in the Bureau of Health
Manpower, USDHEW that projected dental manpower requirements and showed how
slightly higher levels of supply could encourage shifts in care-seeking behavior and
improve oral health. (Hirsch and Killingsworth, 1975; Pugh-Roberts Associates, 1975)
Later work applied the methodology to heart disease. (Luginbuhl et al, 1981) More
recent work has applied System Dynamics to developing strategies for dealing with
chronic illnesses such as diabetes (Homer et al, 2004; Jones et al, 2006) and
cardiovascular disease (Hirsch et al, 2010; Homer et al, 2010). The CDC has also
supported the development of a policy game called HealthBound that is based in an SD
model and helps people understand the importance of prevention and primary care
capacity in the context of health reform. (Milstein et al, 2010).
Structure of the Early Childhood Caries (ECC) Model
The basic structure of the ECC model emerged from a meeting of experts in various
aspects of children’s oral health in April, 2009 at Columbia University. The overall
structure of the model, shown in Figure 1, separates children by age and risk of
developing ECC. It was felt that separation by risk is important to characterize
differences in ECC prevalence in the population and also to provide options in the model
for allocating public health and dental resources to children at greatest risk. Furthermore,
socioeconomic status as measured by household income was decided to be the best
surrogate for risk, given the significant differences in ECC prevalence among children at
different income levels. (Edelstein, 2002) The model distributes Colorado’s population
of children ages 0 to 5 among these groups.
Age 0-6 Months, > Age 6-24 Months, be Age 2-5 Years,
High Risk High Risk High Risk
Age 0-6 Months, . Age 6-24 Months, > Age 2-5 Years,
Moderate Risk Moderate Risk Moderate Risk
Age 0-6 Months,
_ Age 6-24 Months, > Age 2-5 Years,
Low Risk Low Risk Low Risk
Figure 1: Overview of ECC Model Structure
In simulations with this model, children naturally age over time with births introducing
new children and others aging out as they reach their sixth birthday. There is also the
possibility of moving between risk categories if, for example, preventive programs result
in a different set of circumstances for some children in lower income (higher risk) groups
that are less conducive to ECC development and help to promote better oral health.
There are also important things going on within each of the boxes in Figure 1: the
progression of ECC. The stages of the disease process, as represented in the model, are
shown in Figure 2. Over the course of a simulation, children move from left to right as
they develop ECC. Children start initially with No Caries Activity (NCA) and many
remain in this category throughout their early childhood. However, some develop caries
at rates tied to their age and risk groups and to various other factors that may be affected
by preventive interventions. The experts at the April, 2009 meeting urged that the model
make an important distinction between caries, that is any presence of the disease, and
cavities where the disease creates measurable depressions in teeth. Children who move
from the No Caries Activity (NCA) box to the one second from the left labeled Caries are
ones who have developed pre-cavity lesions (e.g., white spots), but do not yet have
measurable cavities. The purpose of adding this stage to the model is to provide an
additional (critical) point at which to test interventions in the ECC process.
Treated Caries Treated Cavities | og.
Treating Symptomatic
Cavfties
Developit here i
Teme ule
Treating ‘Treating
Caries Y Cavities
ies Activit Untreated Caries Untreated ‘Symptomatic
No Caries Activity| ——x——> =| Cavities = oo ie
Developi Developing Developing avities
Untreatet Untreat ‘Symptomatic
Caries Cavities Cavities
Figure 2: ECC Disease Stages Reflected in the Model
Without treatment or preventive activities, children with Caries develop Cavities that are
initially Untreated. Some fraction of these are discovered and Treated during the course
of regular dental visits. Others become Symptomatic and require Treatment on a more
urgent basis. Some fraction of those children who have had Cavities Treated develop
recurrent cavities that are initially Untreated. At each point in a simulation, children are
moving in at least two directions as they age and also move through the sequence of
stages in ECC and potentially in a third direction if preventive interventions also enable
them to move among risk groups.
The model is initially set up in equilibrium and will continue to reflect the initial
distribution of children among age and risk groups and distribution among these groups
by disease stage in the absence of any new programs. Interventions change rates of flow
from one box to another and, over time during a simulation, yield very different patterns
of ECC prevalence. The model calculates a number of summary variables (e.g., overall
fraction of children with cavities, cumulative cost of restorative care) that enable users to
evaluate the potential impacts of different interventions and combinations of programs.
The next section describes how the computer model was quantified.
Quantifying the Early Childhood Caries (ECC) Model
Quantifying the ECC model for Colorado required data on prevalence of ECC relative to
different demographic and behavioral characteristics such as household income. These
data can be obtained for the entire US through the National Health and Nutrition
Examination Survey (NHANES), but are not typically available at the state level.
Fortunately, it was possible to access data from the Colorado Child Health Survey done
as an adjunct to the Behavioral Risk Factor Surveillance System (BRFSS). The Child
Health Survey has several oral health questions as well as others about access to medical
care and insurance and behaviors such as consumption of sugary drinks. (See
http://www.cdphe.state.co.us/hs/yrbs/child_health_questionnaire 2004 1.pdf fora
complete questionnaire for this survey.)
The first step in quantifying the ECC model was to choose the income ranges for the
High, Moderate, and Low Risk groups. Because there were no data on the fractions of
children with cavities in Colorado to explore this question, fractions from the Child
Health Survey with a positive answer to the question: Pain\ Cavities\ Broken or Missing
Fillings\Teeth Pulled Because of Cavities? were used as a surrogate. These data revealed
a pattem with a consistently high prevalence of these decay-related problems in income
levels going up to 200% of FPL (averaging 18.6%) with some decline from 200% to
300% (15%), and a much lower level for income levels greater than 300% of FPL (8.4%).
The next step was to distribute the population of children in each age and risk group into
the stages in the ECC disease process shown in Figure 2, beginning with the fractions of
children in each group with cavities. As indicated above, the percentages of children
with a positive answer to the question Pain\ Cavities\ Broken or Missing Fillings\Teeth
Pulled Because of Cavities? in the Child Health Survey to establish the pattem of relative
prevalence of cavities among the age and risk groups. However, these pattems reflected
self-report by parents rather than the more rigorous identification of cavities by
examination that is done as part of the NHANES survey. The fractions in the Child
Health Survey data do not, for example, include children whose cavities are
asymptomatic and have not come to the attention of parents. Therefore, to get
comparable fractions of children with cavities, the fractions with positive answers to this
question in the Child Health Survey (about 12.75% overall) were inflated to reflect cavity
prevalence of 23% reported for 1999-2002 in the NHANES survey for 2-5 year olds,
when adjusted for Colorado’s income distribution. (See Beltran-A guilar et al, 2005).
The NHANES data also provided fractions of children 2-5 with untreated cavities, by
income level. The fraction of children with symptomatic cavities came from data on 2-5
year olds in “urgent need of treatment” in a GAO report derived from NHANES. (GAO,
2008).
There are no readily available data on the prevalence of what we are calling Caries in our
model, pre-cavity conditions such as white spots. Fractions with caries (but not cavities)
therefore had to be derived using percentage changes in cavity prevalence between the
age groups (6-24 months>2-5 years from the Child Health Survey; 2-5 years>6-11
years from the NHANES data), assuming that children developing cavities in a particular
age group who didn’t have them before would be likely to have a “pre-cavity” condition.
Table 1 shows the fractions of children in the different age/risk groups at different stages
in the disease process. Transition rates (flows between the boxes in Figure 2 expressed in
terms of children per month) were estimated initially based on increases in prevalence in
the various ECC stages between one age group and the next (6-24 months>2-5 years
and 2-5 years>6-11 years). The model was then used to more finely calibrate these rates
(children per month moving from one stage to the next). As indicated earlier, the purpose
of this calibration was to create a model in equilibrium that would make it possible to see
the incremental effect of any interventions.
One check on the calibration resulted in an additional adjustment. Treatment rates for
cavities generated by the model (numbers of children moving from the Untreated Cavities
to Treated Cavities boxes) were compared to numbers of children who could be expected
to have a restorative procedure during a year based on data from the Medical Expenditure
Panel Survey (MEPS) (Manski and Brown, 2007). This comparison revealed that the
numbers being generated by the model were too low and they were increased
accordingly.
Age 6-24 Months
Fraction with
NoCaries Untreated = Untreated Treated Symptomatic Fraction with Untreated
Activity Caries Cavities Cavities Cavities Cavities Cavities
Low Risk 0.85 0.08 0.03 0.03 0.01 0.07 0.63
Moderate Risk 0.73 0.14 0.08 0.03 0.02 0.13 0.74
High Risk 0.67 0.17 0.10 0.04 0.03 0.16 0.76
Age 2-5 Years
Low Risk 0.76 0.09 0.07 0.06 0.03 0.15 0.64
Moderate Risk 0.57 0.16 0.16 0.07 0.04 0.27 0.74
High Risk 0.46 0.20 0.20 0.08 0.06 0.34 0.76
Table 1: Fractions of Children at Various Stages of ECC Development by Age and Risk
The model produces restorative visit rates that fall within the range suggested by the
MEPS data. (Manski and Brown, 2007) The cost of restorative care for the 0-6
population is also calculated by the model. Cumulative costs for restorative care are a
useful metric for comparing simulations and estimating potential savings on restorative
care that might offset programmatic costs for implementing various interventions. There
are two components to this cost calculation: 1) conventional care in the dental office and
2) care under anesthesia in hospital ORs or ambulatory surgical centers for very young
children and others who require it.
The model also includes fractions of children in the different age/risk groups with
detectable levels of s. mutans bacteria, a prime causal agent in the ECC disease process.
The presence of s. mutans colonization as a discrete element in the model will enable us
to test the effects of various interventions such as reducing the transmission of s. mutans
from caregiver to child, education to reduce the consumption of sugary drinks and use of
baby bottles to put children to sleep, and direct administration to children of substances
such as xylitol that reduce s. Mutans colonization.
Simulations with Different Interventions
The model supports a number of possible interventions. Simulations with the model over
a ten year period can project changes in fractions of children ages 0-6 with cavities and
with untreated cavities and symptomatic cavities, dft scores, and costs of restorative care.
These results can be weighed against estimated program costs to get a rough idea of cost-
benefit ratios for different interventions. Interventions can be applied to the entire
populations or to particular age and/or risk groups. Possible interventions include:
e Educational programs that reduce the consumption of sugary drinks, use of baby
bottles at night, and other harmful practices that contribute to the growth of s.
mutans and the ECC disease process.
e Programs aimed at reducing the transmission of s. mutans from parents and other
caregivers to children using xylitol gum, chlorhexidine, or other substances.
e Use of xylitol products directly with older children.
e Aggressive screening for and treatment of caries (pre-cavities) to reduce
progression to cavities.
e Expanded use of fluoride varnish.
e Focused preventive care and education for children who already have cavities to
reduce recurrence rates.
e Rigorous tooth brushing programs with fluoride toothpaste.
e Expansion of Community Water Fluoridation (CWF) to the entire population.
¢ Motivational interviewing with a strong educational component.
The following section describes a number of simulations done with the model and
presents their results. Results of each simulation are shown at the end of ten years, once
any new interventions have had their full effect, and compared to the results of a
“baseline” simulation in which no new interventions are assumed and the model remains
in equilibrium.
Community Water Fluoridation and A pplication of Topical Fluorides
Assumptions:
In the first simulation (only), Community Water Fluoridation (CWF) is extended to
the 24.6% of the population in Colorado not currently covered.
We started with the assumption that initiation of CWF in a population could reduce
measured caries by 50.7% based on post exposure measurements of concurrent
comparison groups (range: 22.3% to 68.8%) (CDC Task Force on Community
Preventive Services, 2002). However, based on expert input, the assumed impact of
CWF was reduced by half since many of those children not currently covered by
fluoridated water systems might already be getting fluorides from toothpastes, foods
and vitamins. (Maas, 2010a) A cost of 50 cents per person is added for the entire
additional population covered (including adults) since CWF is applied to everyone.
(US Centers for Disease Control and Prevention, 2009)
In the second simulation, fluoride varnish is applied to all children in age groups 2
and 3 and, in the third simulation, only to children in the high risk groups. A fourth
simulation applies fluoride vamish to children in all risk groups, but only those in age
group 3.
Application of fluoride varnish reduces dmfs in deciduous dentition by 33% based on
a pooled estimate from Cochrane database review. (Marinho et al, 2002) A range of
27% to 44% was reported in another article. (Kanellis, 2000)
A cost of $16 per child is assumed for fluoride vamish application. Two applications
per year are provided for all children (in the second simulation) and three per year
when the high risk group is the focus in the third simulation. (Maas, 2010b)
Results are shown in Table 5 on the next page.
Extending CWF to all children has only a limited impact on the prevalence of ECC
because such a large fraction of the population is already covered by community-level
fluoridation. However, it is a good investment since $6 million in program costs can
potentially buy a $14 million reduction in restorative care costs over the 10 year period.
And that does not include reduced treatment costs for older children who also benefit
from the fluoridated water.
Difference in
Cumulative
Overall Cumulative Cost of
Overall Percentage Cost of Restorative Cumulative
Percentage — with Untreated Restorative Care Relative Program Cost
with Cavities Cavities Overall dft © Care ($Mil) toBase($Mil) —_{$ Mil)
Base 18.2% == 71.39% 265,923 208 0 0
Extending CWF to Everyone 17.0% 70.80% 249,042 194 14 6
Fluoride Vamish, All Kids, Age Groups 2 and 3, 12.4% 67.33% 182,196 143 65 118
Fluoride Varnish, High Risk Only, Age Groups 2 and 3, 15.7% 68.68% = 214,281 174 33 56
Fluoride Vamish, All Kids, Age Group 3 Only 16.0% 67.79% 233,723 181 27 85
Table 5: Results of Simulations Extending CWF and Applying Fluoride Varnish
Providing topical fluoride varnish for all children in age groups 2 and 3 yields a more
significant impact on ECC prevalence. Starting with younger children helps produce this
larger impact because it allows this intervention to reach kids early before cavities
develop. ECC prevalence is considerably lower and savings on restorative care costs
have “paid back” more than half the program costs incurred. Reductions in restorative
care cost understate the benefit from this and other interventions because it doesn’t
include many other benefits of improved oral health.
The advantage of starting early is evident from comparing the results of the second
simulation with that of the last one in which fluoride varnish is only applied to children in
age group 3, missing the opportunity for an early preventive effect. Though the program
cost is lower in that final simulation, the benefit in terms of reduced restorative care costs
is proportionately smaller per dollar spent.
Providing this intervention for the highest risk (lowest income group) also has a smaller
overall impact, as might be expected, but slightly a higher ratio of benefit (in terms of
reduced restorative care cost) to program cost. This suggests that, with limited funding,
priority be given to children at highest risk.
Treatment of Mothers with Xylitol to Prevent Transmission of S. Mutans
Assumptions
- 88% reduction in age group 2 and 64% reduction in age group 3 in S. Mutans
colonization for children whose mothers were treated with Xylitol based on one study
that produced a 0.2 RR at age 2 and 0.42 RR at age 3, increased to produce better
match with colonization prevalence data in older age group. (Soderling et al, 2001).
Similar reductions in colonization were found by in children whose mothers went
through another preventive program. That program used a combination of education
and chlorhexidine, but illustrated what could be accomplished with s. mutans
reduction in children when mothers have been treated. (Kohler et al, 1983)
10
- 73% reduction in development of caries in children without S. Mutans colonization
based on the Xylitol study cited above that produced a 0.27 RR in 3-5 year olds who
did not have S. Mutans colonization as two-year olds. (Isokangas et al, 2000). The
earlier study cited above (education+chlorhexidine) found that children who were not
colonized by age 2 were much less likely to have had caries by age 4 (25% vs. 89%
for those who were colonized at age 2). (Kohler and Andreen, 1994)
- Applied to mothers of children in age groups 1 and 2 only; delayed effect on children
in age group 3 as children who did not benefit from treatment age out and those
whose mothers were treated age into that group; 12 month delay assumed between
time mother’s S. Mutans level is lowered and time transmission would have occurred
- $100 average one-time cost per mother
Difference in
Cumulative
Overall Cumulative Cost of
Overall Percentage Cost of Restorative Cumulative
Percentage with Untreated Restorative Care Relative Program Cost
with Cavities Cavities Overall dt Care ($ Mil) toBase($ Mil) —($ Mil)
Base 18.2% 71.39% — 265,923 208 0 0
Xylitol Moms, All Kids 10.8% 68.22% 160,609 152 56 79
Xylitol Moms, High Risk Only 15.0% 69.27% 201,765 180 28 25
Table 6: Results of Treating Mothers with Xylitol to Reduce Transmission of S. Mutans
A significant (40%) reduction in fraction of children with cavities is projected to occur
when mothers are treated with X ylitol gum to reduce transmission of S. Mutans bacteria.
Again, concentrating on the highest risk children will yield a smaller over all effect, but
larger reduction in restorative care cost per program dollar spent. Treating mothers with
Xylitol has its most direct effects on the youngest children and more delayed effects on
those in the oldest age group. As with fluoride vamish, the reduction in restorative costs
pays back a substantial portion of the program costs. The leverage provided by this
intervention, when applied to the highest risk (lowest income) group may pay back all or
most of the program’s costs.
Figure 5 on the next page contrasts the benefit over time from this intervention with the
one in the previous set using fluoride varnish with all children in age groups 2 and 3. The
graph shows the effect on the number of children in age group 3 with cavities and how
treating mothers with X ylitol is slower to have an impact, but eventually can have a
greater effect than the fluoride vamish.
11
Total with Cavities (Age Group 3)
80,000
70,000
60,000
50,000
40,000
0 12 24 36 48 60 72 84 96 108 120
Time (Month)
Total with Cavities in the Different A ge Groups(Age3] : xylitol moms, all risk
Total with Cavities in the Different A ge Groups[Age3] : fluoride vamish all, ages 2 and 3
Total with Cavities in the Different Age Groups[A ge3] : new base 10-3
Figure 4: Results over Time Comparing Fluoride Vamish with
Xylitol Treatment of Mothers
Motivational Interviewing
Assumptions
- Motivational interviewing, with appropriate follow-up, can reduce cavity prevalence
by 63% (Weinstein, 2004). Simulations were done applying this intervention for all
children and for those in the highest risk group only.
- The estimated cost is $100 per child.
Difference in
Cumulative
Overall Cumulative Cost of
Overall Percentage Cost of Restorative Cumulative
Percentage — with Untreated Restorative Care Relative Program Cost
with Cavities Cavities Overall dt © Care ($ Mil) toBase ($ Mil) ($M)
Base 18.2% 71.39% 265,923 208 0 0
Motivational Interviewing, All 65% = 59.54% 95,434 86 122 M1
Motivational Interviewing, High Risk Only 12.9% 65.99% 159,020 144 64 35
Table 9: Results with Motivational Interviewing
12
Motivational interviewing can have a significant impact on reducing the fraction of
children with cavities and the costs of restorative care. As indicated earlier, these cost
savings are only a fraction of the benefit that would be derived from such a large
reduction in cavities. Using this technique with only the high-risk group will, as with the
other interventions, produce a smaller overall impact, but a larger benefit per dollar spent
on the program. Concentrating on the highest risk group also helps to improve equity
among the different risk groups.
Combined Interventions
The following interventions were combined in each of the simulations whose results are
shown in the table below:
1. Fluoride vamish for all children in age groups 2 and 3 together with screening and
treatment for caries (pre-cavity lesions) using lower impact assumption.
2. Interventions in #1 together with secondary prevention aimed at children who have
had restorative care for cavities with the assumption that recurrence rates are cut by
50%.
3. Interventions in #2 together with motivational interviewing benefiting all children.
The overall effect of this combination is to reduce cavities by 75% as a result of the
fluoride varnish (33%) and the motivational interviewing (an additional 63%) and a
further amount due to the effect of screening and treating for caries. Program costs
are $16 per child per application for the fluoride varnish, $100 for the motivational
interviewing, and $242 for the caries (pre-cavity) treatment.
Overall
Percentage
with Cavities
Base 18.2%
Combined 1 9.1%
Combined 2 9.1%
Combined 3 3.8%
Overall
Percentage
with Untreated
Cavities
71.39%
66.39%
57.38%
42.44%
Overall_dft
265,923
136,005
136,005
56,158
Difference in
Cumulative
Cumulative Cost of
Cost of Restorative Cumulative
Restorative Care Relative Program Cost
Care ($ Mil] to Base ($ Mil) ($ Mil)
208 0
121 87
108 100
59 149
Table 10: Results with Combined Interventions
The results in this table show that combining the various interventions can have
cumulative and complementary effects. Combining several interventions can produce a
smaller fraction of children with cavities than any of the interventions can individually.
Adding secondary prevention for children who already have cavities can reduce the
fraction with untreated cavities and the cost of restorative care.
0
147
147
245
13
Conclusions
This paper has described a model of Early Childhood Caries in a population of children
aged 0-5 and presented results of simulations with a variety of interventions designed to
reduce the prevalence and consequences of ECC. The following general conclusions can
be drawn from the simulation results presented in Tables 5 through 10:
e Interventions aimed at the youngest children will take longer to affect the entire
population, but will ultimately have a more profound effect in reducing prevalence as
the impact percolates into the older groups as children age.
e Interventions limited to the highest risk (lowest income) groups of children will have
the greatest impact per dollar spent because of the greater relative risk of ECC in that
population. Limited budgets are best spent on these groups.
e Combined interventions that target ECC at several stages of development in the
disease process are likely to have the greatest impact. Primary prevention provides
the greatest leverage, but it is also productive to limit disease progression by, for
example, screening for and treating caries before cavities form.
These conclusions are likely to hold for any population of children, but the impacts of
particular interventions will differ somewhat depending on the composition of the
population and prevalence of ECC. With data similar to that used to quantify the model
for Colorado’s children, the model can represent any other state, large city or county, or
Metropolitan area population and provide simulation results more closely tailored to that
area’s population of young children and their oral health needs.
While the model is specific to Early Childhood Caries, it also demonstrates how this
approach can be applied to many other oral health problems as well as the linkages
between oral health and other chronic illnesses. For example, there is an effort currently
underway to develop an oral health workforce model for the State of Colorado. This
model will project the oral health status and needs of the state’s population, both for
adults and children, and potential impact of different levels and mixes of personnel as
well as other policies. Another model being contemplated is one that represents the
relationships between oral health chronic illnesses such as cardiovascular disease and
diabetes. That model would help to examine how improved oral health care can serve as
a leverage point in reducing the consequences of those common chronic problems.
References
Beltran-A guilar E et al, Surveillance for Dental Caries, Dental Sealants, Tooth Retention,
Edentulism, and Enamel Fluorosis --- United States, 1988--1994 and 1999—2002
MMWR August 26, 2005 /54(03);1-44
CDC Task Force on Community Preventive Services. Reviews of Evidence on
Interventions to Prevent Dental Caries, Oral and Pharyngeal Cancers, and Sports-
Related Craniofacial Injuries. Am J Prev Med 2002; 23(1S) 24-31
14
Edelstein B, Dental Care Considerations for Y oung Children. Spec Care Dentist 2002;
22(3):118-25S
Government Accountability Office (GAO). MEDICAID: Extent of Dental Disease in
Children Has Not Decreased, and Millions Are Estimated to Have Untreated Tooth
Decay. 2008, Table 5
Hirsch GB, Killingsworth WR, A New Framework for Projecting Dental Manpower
Requirements. Inquiry. 1975; 12: 126-142
Hirsch GB, Homer JB, Evans E, Zielinski A, A System Dynamics Model for Planning
Cardiovascular Disease Interventions. American J ournal of Public Health. 2010;
100:616-622.
Homer JB, Hirsch GB, System dynamics modeling for public health: background and
opportunities. American Journal of Public Health. 2006; 96: 452-458
Homer J, Hirsch G, Minniti M, Pierson M. Models for collaboration: how system
dynamics helped a community organize cost-effective care for chronic illness. System
Dynamics Rev. 2004; 20:199-222.
Homer J, Milstein B, Wile K, Trogdon J, Huang P, Orenstein D. Simulating and
evaluating local interventions to improve cardiovascular health. Prev Chronic Dis.
2010:7(1).
Jones AP, HomerJB, Murphy DL, Essien JDK, Milstein B, Seville DA. Understanding
diabetes population dynamics through simulation modeling and experimentation.
American Journal of Public Health 2006; 96(3): 488-494.
Luginbuhl W, Forsyth B, Hirsch G, Goodman M. Prevention and rehabilitation as a
means of cost-containment: the example of myocardial infarction. J Public Health
Policy.1981; 2: 1103-1115.
Maas B, 2010a, Pew Center on the States (Formerly, Director, Division of Oral Health,
CDC), Personal Communication, May 28, 2010
Maas B, 2010b, Personal Communication, Sept. 28, 2010
Manski, RJ and Brown, E, Dental Use, Expenses, Private Dental Coverage, and
Changes, 1996 and 2004. Rockville (MD): Agency for Healthcare Research and
Quality; 2007. MEPS Chartbook No.17.
http://www.meps.ahrg.gov/mepsweb/data_files/publications/ch17/cb17.pdf
Marinho V, Higgins J, Logan S, Sheiham A. Fluoride varnishes for preventing dental
caries in children and adolescents (Review). Cochrane Database of Systematic
Reviews 2002, Issue 1. Art. No.: CD002279
Milstein B, Homer J, Hirsch G. Analyzing National Health Reform Strategies with a
Dynamic Simulation Model. American Journal of Public Health. 2010; 100:811-819
Pugh-Roberts Associates Inc. Examining Alternatives for Improving the Nation’s Oral
Health: A System Dynamics Model of the Dental Care Delivery System. 1975
Available from the National Technical Information Service (www.ntis.gov )
Document number PB254772
Tolliver, R, Colorado Child Health Survey data, Colorado Department of Public Health
and Environment, Personal Communication, November 6, 2009.
US Centers for Disease Control and Prevention, Cost Savings of Community Water
Fluoridation, 2009, available at http://www.cdc.gov/fluoridation/fact_sheets/cost.htm
Weinstein P, Harrison R, Benton T. Motivating parents to prevent caries in their young
children: One-year findings. J Am Dent Assoc 2004; 135: 731-738
