Oral health behaviors and dental caries in low-income children with special health care needs: an observational prospective study

Summary

Background.

Dental caries is a significant public health problem for low-income children with special health care needs (CSHCN).

Aim.

We evaluated associations between oral health behaviors (e.g., diet, fluoride, dental care) and caries for CSHCN enrolled in Medicaid, a health insurance program for low-income populations that provides comprehensive dental coverage for children.

Design.

We recruited 116 CSHCN ages 7 to 20 years from Medicaid enrollment files in Washington state, USA. Caregivers completed a 166-item questionnaire and children received a dental screening. The outcome was dental caries, defined as total pre-cavitated, decayed, missing, or filled tooth (PDMF) surfaces. We ran log-linear regression models and generated prevalence rate ratios (PRR).

Results.

The mean age of study participants was 12.4±3.1 years, 41.4% were female, and 38.8% were white. The mean PDMF surfaces was 6.4±9.4 (range: 0 to 49). Only sugar-sweetened beverage intake was significantly associated with dental caries. CSHCN who consumed >4 sugar-sweetened beverages per week were significantly more likely to have tooth decay than those who consumed no sugar-sweetened beverages (PRR: 2.58; 95% CI: 1.37, 4.85; P<0.01).

Conclusion.

Sugar-sweetened beverages are an important target for future behavioral interventions aimed at preventing dental caries in low-income CSHCN.

Introduction

Dental caries is a significant public health problem for children with special health care needs (CSHCN).¹ A recent study from North Carolina showed that CSHCN have a higher caries risk and caries burden than healthy children.² Another study reported that among low-income Head Start enrollees, children with developmental delays had significantly higher caries prevalence.³ In terms of etiology, dental caries is a multifactorial disease linked to three main oral health behaviors: high carbohydrate diet, inadequate fluoride exposure, and poor access to dental care.

Dental caries is associated with added sugar intake. Children with autism spectrum disorders consume significantly more sugar-sweetened beverages servings per day (e.g., juice, punch, soda) than healthy children without autism (2.6 versus 1.7, respectively; P=0.03) and snack foods like candy, chips, cookies, and crackers (4.0 versus 3.0; P=0.01), and significantly fewer daily servings of fruits and vegetables (3.1 versus 4.4; P<0.01).⁴ Additional diet-related risk factors for CSHCN include delayed feeding skills and feeding problems, which can lead to increased intake of soft foods, nutritional supplements with high sugar concentration, and chronic use of medications sweetened with sugar.⁵

Inadequate fluoride exposure also contributes to caries risk. Fluoride is safe and effective.⁶ The most common sources of fluoride are tap water, fluoridated toothpaste, and fluoride varnish. These fluoride modalities are usually easy to incorporate into a child’s daily routines. However, many CSHCN have oral aversions, including hypersensitivities to textures, smells, and tastes, which can be barriers to toothbrushing, toothpaste use during brushing, and acceptance of fluoride varnish during dental visits. Furthermore, some caregivers of CSHCN have difficulties enforcing toothbrushing at home and may be concerned about the potential risks associated with ingesting fluoride. These latter concerns may prompt some caregivers of CSHCN to minimize fluoride exposure by using fluoride-free toothpastes, skipping toothbrushing, and refusing topical fluoride during dental visits.⁷

Inconsistent dental care use is the third behavioral risk factor for caries, mainly the consequence of delayed treatment and missed preventive care. Having dental insurance promotes access to dental care for children.⁸ Multiple studies have shown that CSHCN are more likely to have health insurance than children without special health care needs (SHCN) (82% versus 77%, respectively; P<0.001).⁹ However, findings are mixed regarding preventive dental care use for CSHCN. Some studies suggest that CSHCN use preventive dental care at equal or higher rates than children without SHCN (80% for CSHCN and 72% for children without SHCN).^10,11 A more recent study reported that Medicaid-enrolled CSHCN under age 6 years were significantly less likely to utilize preventive dental care than children without SHCN (prevalence rate ratio: 0.91; 95% CI: 0.88, 0.94; P<0.001).¹² This may be because of differential barriers to dental care encountered by CSHCN. Pediatric dentists are maldistributed and general dentists may not feel sufficiently trained to treat medically-complex CSHCN.^13,14 Furthermore, caregivers frequently have difficulties finding a dental office in which staff members and the dentist are comfortable treating CSHCN, especially if the child has difficulty cooperating because of behavioral diagnoses, an oral aversion, or dental anxiety and fears.¹⁵

While studies have examined the aforementioned behavioral risk factors individually, no studies to date have examined how oral health behaviors collectively influence dental caries prevalence in low-income CSHCN. This critical gap in the literature limits our ability to develop relevant interventions aimed at addressing sub-optimal oral health behaviors as a strategy to prevent caries. In this study, we evaluated associations between oral health behaviors and dental caries in low-income CSHCN. We tested the hypothesis that factors related to diet, caregiver views on fluoride, and dental care use would be associated with dental caries. This study is an important next step in designing behavioral interventions aimed at preventing dental caries and improving the oral health of low-income CSHCN.

Materials and methods

Study design and participants.

This was a prospective observational study of CSHCN ages 7 to 20 years who were enrolled in the Washington state Medicaid program. Medicaid provides eligible low-income children up to age 20 years with no-cost health insurance that covers comprehensive health services, including dental care. We obtained Medicaid enrollee data from the Washington State Department of Social and Health Services. Special health care needs were identified from medical claims data using previously published Clinical Risk Grouping methods.¹⁶ The study focused on CSHCN living in one of three counties in Washington state (King, Pierce, and Snohomish counties). We contacted caregivers whose home address was located in one of these counties and for whom there was a telephone number in the enrollment files.¹⁷ Families were contacted by study staff who introduced the study, described the procedures, and recruited caregiver-child dyads for an in-person study visit. Interested dyads were then scheduled for a study visit that took place at a local dental office from December 2016 to March 2017. Written consent was obtained from each child’s primary caregiver and participants were verbally assented. A total of 922 households were contacted, 211 households were recruited, and 116 children participated.¹⁷ The study was approved by the Washington State Institutional Review Board. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement was used to report study findings.

Study procedures.

Each study visit took about one hour. Caregivers completed a 166-item paper questionnaire that was pilot tested with caregivers, modified, and then finalized. This is the same questionnaire that was used for a previously published study.¹⁸ While caregivers completed the questionnaire, children received a dental screening. The screening was conducted visually with a mouth mirror only after brushing the teeth with a dry toothbrush. At the end of the visit, caregivers were given a $40 gift card and children a $20 gift card as a thank you for participation.

Outcome.

The outcome was dental caries prevalence. Each primary and permanent tooth surface was classified as pre-cavitated (e.g., demineralized white spot lesion), decayed, missing, or filled (PDMF) using National Institute for Dental and Craniofacial Research (NIDCR) Early Childhood Caries Collaborative Centers (EC4) Criteria, which is based on the World Health Organization criteria.¹⁹ The setting was a regular reclined dental chair with an overhead light. We calculated the total PDMF surfaces for all erupted tooth surfaces. Each participant was screened by a trained and calibrated pediatric dentist or dental hygienist. Intraexaminer reliability was good (kappa coefficient = 0.97).

Predictor variables.

We organized predictor variables into three groups: diet, fluoride, and dental use. The first group consisted of dietary factors measured with a validated beverage and snack questionnaire.²⁰ We asked caregivers about the frequency with which their child consumed various foods (e.g., chips, candy, chocolates, donuts, ice cream, vegetables, and fruits) and beverages (e.g., water, milk, fruit juice, sweetened beverages, soda) each week at specific locations (home, school, or friend’s home). There were seven options ranging from “Never or <1 per week” to “4 or more per day”. If a caregiver was unsure about an answer, they were free to solicit input from the child. We aggregated across the three locations for each food and beverage. We combined candy, chocolates, donuts, and ice cream into a single category called sweet snacks. Fruit juices, sweetened beverages (e.g., Tang, Kool-Aid, Hawaiian Punch, Sunny delight, lemonade, Snapple), and soda were combined into a category called sugar-sweetened beverages. The remaining items were retained as separate categories (vegetable, fruit, milk, and water). Each food or beverage was modeled as a categorical variable consistent with a previous study based on the same questionnaire.¹⁸

The second group of predictor variables was related to fluoride. There were three items. The first was whether the child received topical fluoride at their last dental check-up (yes, no, I don’t know). The second was how concerned caregivers were with fluoride safety (somewhat or very concerned versus not at all or not too concerned). The third item was how okay the caregiver was with topical fluoride (strong okay and okay versus somewhat okay and not at all okay).

The third group of variables consisted of dental and oral health-related factors. There were three items. The first was whether the child has a personal dentist (yes, no). The second was toothbrushing frequency (twice daily versus once daily). The third was caregiver-reported oral health status of the child (excellent and very good versus good versus fair and poor).

Demographic factors.

There were two sets of demographic items: sociodemographic and household factors. Sociodemographic factors included child’s age, sex, race (corresponding to the U.S. Census Bureau categories), Hispanic ethnicity (no, yes), and severity of special health care need (measured using 3M Corporation Clinical Risk Grouping methods with three categories of increasing condition severity: episodic, life-long, catastrophic/malignancy).¹⁶ The two household factors were annual income and food insecurity, the latter of which was measured using the two-item food insecurity screener.²¹

Data analyses.

Descriptive statistics were generated on all variables. To examine bivariate associations between each of the demographic, household, and behavioral factors and dental caries prevalence (total PDMF surfaces), we ran log-linear regression models with the natural log of total number of tooth surfaces as an offset. The models generated unadjusted prevalence rate ratios (PRR) and corresponding 95% confidence intervals (CIs). Each variable that was statistically significant in the bivariate analyses (α=0.05) was included in the final multiple variable log-linear regression model. Individuals with missing data were excluded from the final model. As a sensitivity analysis, we ran the final model with an outcome that excluded pre-cavitated caries. Because the findings were similar, we only report data from the original outcome. All data were analyzed using Stata SE 15.1 (College Station, TX).

Results

Descriptive statistics.

Of the 116 participants, the mean age was 12.4±3.1 years, 41.4% were female, 57.8% were white, and 31% were Hispanic (Table 1). In terms of SHCN severity, 69% had an episodic chronic condition, 23.3% a life-long chronic condition, and 5.2% a catastrophic chronic condition or malignancy. One-in-three children were from households with an annual income of less than $20,000 and a similar proportion of children were from food insecure homes.

Table 1.

Unadjusted prevalence rate ratios (PRR) from log-linear regression analyses of tooth decay prevalence for low-income children with special health care needs ages 7 to 20 years (N=116)

Variable	N (%) or mean±SD	Unadjusted PRR (95% CI)	p-value
Outcome
Total pre-cavitated, decayed, missing, and filled tooth surfaces (PDMFS)	6.4±9.4	-	-
≥1 PDMFS	73 (62.9%)	-	-
Child Demographics
Age (years)	12.4±3.1	1.07 (0.95, 1.20)	0.25
Sex
Female (reference)	48 (41.4%)	reference	-
Male	68 (58.6%)	1.00 (0.58, 1.71)	0.99
Race
Non-white (reference)	45 (38.8%)	reference	-
White	67 (57.8%)	1.23 (0.69, 2.19)	0.48
Missing	4 (3.5%)	n/a	-
Hispanic ethnicity
No	80 (69.0%)	reference	-
Yes	36 (31.0%)	0.84 (0.46, 1.54)	0.57
Severity of special health care need
Episodic	80 (69.0%)	reference	-
Life-long	27 (23.3%)	1.64 (0.86, 3.14)	0.14
Catastrophic or malignancy	6 (5.2%)	1.39 (0.46, 4.21)	0.56
Missing	3 (2.6%)	n/a	-
Household Factors
Annual income
< $20,000	38 (32.8%)	1.24 (0.63, 2.45)	0.53
$20,000 - $29,999	21 (18.1%)	1.24 (0.51, 3.04)	0.64
$30,000 - $49,999	26 (22.4%)	1.28 (0.59, 2.75)	0.53
$50,000 +	18 (15.5%)	reference	-
Missing	13 (11.2%)	n/a	-
Food insecurity
No	78 (67.2%)	reference	-
Yes	37 (31.9%)	1.48 (0.81, 2.69)	0.20
Missing	1 (0.9%)	n/a	-
Dietary Factors (servings per week)
Sweet snacks
0	7 (6.0%)	reference	-
> 0 to 8	62 (53.5%)	1.04 (0.32, 3.36)	0.94
> 8	47 (40.5%)	1.00 (0.30, 3.34)	0.99
Vegetables
0	41 (35.7%)	reference	-
> 5 to 7	15 (13.0%)	1.08 (0.52, 2.25)	0.83
> 7 to 18	36 (31.3%)	0.80 (0.37, 1.73)	0.58
> 18	23 (20.0%)	1.66 (0.84, 3.29)	0.14
Fruit
0 to 4	26 (22.4%)	reference	-
> 4 to 9	35 (30.2%)	1.11 (0.49, 2.49)	0.81
> 9 to 18	23 (19.8%)	0.71 (0.28, 1.76)	0.45
> 18	32 (27.6%)	1.05 (0.45, 2.43)	0.91
Sugar-sweetened beverages
0	43 (37.1%)	reference	-
1 to 4	41 (35.3%)	0.82 (0.47, 1.43)	0.48
> 4	32 (27.6%)	2.40 (1.30, 4.44)	<0.01
Milk
0 to 8	43 (37.1%)	reference	-
> 8 to 24	39 (33.6%)	0.77 (0.44, 1.37)	0.38
> 24	24 (20.7%)	1.52 (0.73, 3.17)	0.26
Water
0 to 11	24 (20.7%)	reference	-
> 11 to 27	33 (28.5%)	1.24 (0.58, 2.65)	0.58
> 27 to 46	39 (33.6%)	0.93 (0.44, 1.99)	0.85
> 46	20 (17.2%)	0.83 (0.38, 1.81)	0.64
Fluoride Factors
Child received topical fluoride at last checkup
Yes	59 (50.9%)	reference	-
No	39 (33.6%)	1.23 (0.67, 2.26)	0.51
I don’t know	14 (12.1%)	1.10 (0.34, 3.75)	0.86
Caregiver concern about topical fluoride safety
Somewhat or very concerned	84 (72.4%)	reference	-
Not at all or not too concerned	26 (22.4%)	2.06 (1.03, 4.09)	0.04
Missing	6 (5.2%)	n/a	-
How okay is caregiver with topical fluoride
Strongly okay or okay	81 (69.8%)	reference	-
Somewhat okay or not at all okay	29 (25.0%)	0.66 (0.34, 1.30)	0.23
Missing	6 (5.2%)	n/a	-
Dental and Oral Health-Related Factors
Child has a personal dentist
No	30 (25.9%)	reference	-
Yes	86 (74.1%)	1.04 (0.52, 2.05)	0.92
Child toothbrushing frequency
Twice daily	55 (47.4%)	1.01 (0.57, 1.81)	0.97
Once a day	58 (50.0%)	reference	-
Missing	3 (2.6%)	n/a	-
Caregiver-reported oral health status of child
Excellent or very good	43 (37.1%)	reference	-
Good	51 (44.0%)	1.47 (0.87, 2.46)	0.15
Fair or poor	22 (19.0%)	1.95 (0.90, 4.22)	0.09

Oral health behaviors.

In terms of diet, 40.5% of CSHCN consumed eight or more sweet snacks and 27.6% consumed >4 servings of sugar-sweetened beverages per week (Table 1). Nearly 36% consumed no vegetables and 22.4% consumed ≤4 servings of fruit per week. Regarding fluoride factors, over 50% of children received topical fluoride at their last dental check-up and 33.6% did not. Over 70% of caregivers were somewhat or very concerned about topical fluoride and 25% were somewhat okay or not at all okay with topical fluoride. Finally, in terms of dental and oral health-related factors, 74.1% of children had a personal dentist, 47.4% reported brushing twice daily, and 81.1% of caregivers reported their child’s oral health status as excellent, very good, or good.

Outcome.

Of the 116 CSHCN, most participants (62.9%) had any PDMF surfaces while 16% had any untreated decay. The mean PDMF surfaces was 6.4±9.4 (range: 0 to 49).

Bivariate statistics.

Only two variables were associated with dental caries prevalence: consuming sugar-sweetened beverages >4 times a week (PRR: 2.40; 95% CI: 1.30, 4.44; P<0.01) and caregiver concern with fluoride safety (PRR: 2.06; 95% CI: 1.03, 4.09; P=0.04) (Table 1).

Final model.

In the multiple variable log-linear regression model that included sugar-sweetened beverage consumption and caregiver concern about topical fluoride, only the former variable was associated with dental caries. CSHCN who consumed >4 sugar-sweetened beverages per week were significantly more likely to have dental caries than those who consumed no sugar-sweetened beverages (PRR: 2.58; 95% CI: 1.37, 4.85; P<0.01) (data not shown). There was no significant difference in caries prevalence for children who consumed no sugar-sweetened beverages and those who consumed one to four per week (P=0.45). Caregiver concern about topical fluoride failed to reach statistical significance (P=0.05).

Discussion

This is one of the first known studies to simultaneously assess the behavioral risk factors for dental caries in Medicaid-enrolled children with special health care needs. We specifically examined dietary, fluoride, and dental use factors, and found that sugar-sweetened beverage intake was the only variable significantly associated with dental caries in low-income CSHCN.

Sugar-sweetened beverage intake is associated with dental caries.²² The high concentration of sugar in these popular beverages serves as a nutritional reservoir for acidogenic bacteria that promote dysbiosis and demineralize tooth structure. Our study definition of sugar-sweetened beverages was broad and included 100% fruit juices, fruit drinks sweetened with added sugar, and sodas. While numerous studies demonstrate a link between beverages sweetened with added sugar and dental caries, studies on 100% fruit juices have led to mixed findings.²³ Nevertheless, recent recommendations by the American Academy Pediatrics are the most stringent to date and call for limiting even 100% fruit juice intake in children.²⁴ There is a need for research on ways to help families of CSHCN to minimize their child’s exposure to sugar-sweetened beverages.

Strategies to address sugar-sweetened beverage intake range from individual-level education-based efforts to upstream policies aimed at reducing demand. There are no published large-scale interventions that have addressed sugar intake in CSHCN. However, there is a growing literature on interventions that target children and adolescents more broadly. Approaches based solely on health education are likely to fail because knowledge is necessary but insufficient in promoting health behavior change. This also explains why chairside efforts to educate caregivers about sugar-sweetened beverages and caries implemented in dental office settings are largely ineffective. Furthermore, even when health education is accompanied by concomitant environmental change, like providing individuals with access to healthier alternative beverages, behavior change is typically not sustained.²⁵ The likely reasons is that such interventions do not target the behavioral determinants of health, including self-efficacy, response efficacy, perceived susceptibility to caries, and perceived severity of caries.²⁶ Efforts are currently underway to implement and evaluate a community-based behavioral intervention to reduce sugared fruit drink intake in Alaska Native communities.²⁷ Policies like regulating serving sizes, menu labeling, and taxes are promising system-level adjuncts to behavioral interventions.

Our findings indicate that some caregivers of Medicaid-enrolled CSHCN are concerned about the safety of topical fluoride. While topical fluoride concerns were significantly associated with dental caries in the bivariate analyses, this variable was not significant in our final model. Statistical significance aside, the direction of the prevalence rate ratio was in an unexpected direction. It is reasonable to assume that caregivers who are concerned about topical fluoride would be more likely to skip fluoride treatments, thereby increasing the child’s caries risk and prevalence. However, our findings were in the opposite direction: caregivers who were not concerned about topical fluoride had children with higher caries prevalence. There are two potential explanations. First, scientific evidence is unequivocal that topical fluoride is safe and effective⁶ but caregivers who are concerned about topical fluoride may be more extra vigilant about enforcing other behaviors, including diet and brushing, to offset the increased caries risk associated with skipping topical fluoride treatments. However, our data indicate that among concerned caregivers, 46% of their children did not get fluoride at their last dental checkup, 38% had more than eight sweet snack per week, 27% had more than four sugar-sweetened beverages per week, and 35% brushed in the morning and evening. This means that children of concerned caregivers were less likely to get topical fluoride or brush twice daily, but had similar sugar intake as other CSHCN, which would collectively increase caries risk. However, caries prevalence for these CSHCN was lower. Future studies are needed to validate these paradoxical findings regarding reduced fluoride exposure, compensatory behaviors, and caries prevalence.

A second explanation is that caregivers of CSHCN without concerns about topical fluoride who accept fluoride may believe their child is protected against caries, thereby leading to more permissive snacking and sugary beverage behaviors. There is a need to communicate to all caregivers that topical fluoride is effective in prevention but that high sugar intake negates the benefits of fluoride. Topical fluoride-related behaviors are not fully understood, but previous work indicates that 14.4% of dentists believe that having a child with special needs was associated with caregivers refusing topical fluoride.²⁸ Future research should continue to identify the determinants of topical fluoride hesitancy and develop strategies aimed at helping dentists manage refusal behaviors in clinic especially for CSHCN at increased risk for caries.

No dental and oral health-related factors were significantly associated with dental caries. One interpretation is that other behaviors like diet are more important determinants of dental caries for Medicaid-enrolled CSHCN. There is also the possibility of selection bias that limited variation across dental and oral health-related factors, and predisposed our study population to be healthier than other CSHSCN in Medicaid. For example, our participant recruitment strategy required families to have a working telephone number and to make an in-person study visit, which could have inadvertently biased our population toward resourced households. In fact, 75% of caregivers reported their children having a personal dentist and nearly all children reported brushing at least once per day. These rates are higher than corresponding published rates for healthy child populations.²⁹ The characteristics of our study population lend support to the possibility of selection bias.

In terms of caregiver-reported oral health status, there was a trend in which the unadjusted prevalence rate ratios increased linearly as caregiver-reported oral health of the child got worse, which suggests that caregivers of CSHCN are accurate in assessing their child’s oral health status. This relationship could be function of caregivers of CSHCN in our study having a good overall sense of their child’s dental health status and needs. However, our finding is inconsistent with past research indicating weak correlations between caregiver-reported oral health and dental disease in young children as measured by a dentist.³⁰

Our study had three main limitations. First, our findings are generalizable to Medicaid-enrolled CSHCN from three counties in Washington state. The distribution of special health cared severity was similar to another study that used the Clinical Risk Grouping methods.¹¹ Larger epidemiologic studies are needed that examine behaviors related to dental caries for publicly-insured CSHCN. Second, all of our behavioral measures were self-reported, which introduces the potential for recall and social desirability biases. Objective measures of behaviors could overcome the limitation of self-reported measures and reduce the possibility of spurious findings because of misreporting. Third, our overall dental caries prevalence may be an underestimation given the potential for selection bias. In comparing caries prevalence from our study to the national prevalence for children ages 7 to 20 years with public insurance in the 2011–2012 U.S. National Health and Examination Nutrition Survey (NHANES), the overall caries prevalence in our study was slightly lower (63% current study versus 68% for NHANES) as was the prevalence of untreated decay (16% versus 22% for NHANES). Children in the NHANES data had a mean decayed, missing, and filled surface (DMFS) score of 6.2, while the mean DMFS score (excluding white spots) in our study was lower at 5.3. Future studies that enroll CSHCN from diverse income levels would help to clarify how the determinants of caries prevalence differ by socioeconomic status.

Bullet points.

• Behaviors are important determinants of dental caries in CSHCN, but few large-scale interventions focus on improving behaviors.

• Sugar-sweetened beverage intake was the only significant behavioral risk factor associated with dental caries in our population of Medicaid-enrolled CSHCN.

• There is a need for evidence-based behavioral interventions that help families of CSHCN improve dietary intake in children, which is a promising strategy to prevent dental caries, improve quality of life, and reduce oral health inequalities for low-income CSHCN.