Risk assessment models to predict caries recurrence after oral rehabilitation under general anaesthesia: a pilot study

Abstract

Objectives: Oral rehabilitation under general anaesthesia (GA), commonly employed to treat high caries-risk children, has been associated with high economic and individual/family burden, besides high post-GA caries recurrence rates. As there is no caries prediction model available for paediatric GA patients, this study was performed to build caries risk assessment/prediction models using pre-GA data and to explore mid-term prognostic factors for early identification of high-risk children prone to caries relapse post-GA oral rehabilitation. Methods: Ninety-two children were identified and recruited with parental consent before oral rehabilitation under GA. Biopsychosocial data collection at baseline and the 6-month follow-up were conducted using questionnaire (Q), microbiological assessment (M) and clinical examination (C). Results: The prediction models constructed using data collected from Q, Q + M and Q + M + C demonstrated an accuracy of 72%, 78% and 82%, respectively. Furthermore, of the 83 (90.2%) patients recalled 6 months after GA intervention, recurrent caries was identified in 54.2%, together with reduced bacterial counts, lower plaque index and increased percentage of children toothbrushing for themselves (all P < 0.05). Additionally, meal-time and toothbrushing duration were shown, through bivariate analyses, to be significant prognostic determinants for caries recurrence (both P < 0.05). Conclusions: Risk assessment/prediction models built using pre-GA data may be promising in identifying high-risk children prone to post-GA caries recurrence, although future internal and external validation of predictive models is warranted.

Introduction

Dental caries, the most common chronic childhood disease in the USA¹, affects nearly 50% children in the UK by age 5² and is reported to be the 10th most prevalent condition in the Global Burden of Disease 2010 study³. Various caries risk assessment (CRA) models have been proposed in the literature, such as CAT⁴, CAMBRA⁵, Cariogram⁶ and NUS-CRA⁷, with past caries status shown to be the single best predictor of future caries risk⁸. However, risk prediction in young children is still a challenge, with only one model (NUS-CRA) shown to achieve optimal accuracy in 3- to 4-year-old preschool children⁹.

Oral rehabilitation under general anaesthesia (GA) has been employed for treating very young children with rampant caries or those with cognitively, behaviourally and/or medically compromised conditions¹⁰. However, treating early childhood caries (ECC) under GA has frequently resulted in high hospital costs¹¹, familial and individual distress/anxiety¹² and, most discouragingly, high post-GA caries recurrence, of 33–53%, within 6–24 months13., 14.. The possible reasons speculated for the high caries recurrence are the lack of long-term change in the oral health behaviours and attitude of parent/child¹⁵ and/or failure to attend immediate follow-up appointments¹³. Furthermore, a repeat GA procedure was required in 11–76% of children16., 17.. Our previous study also reported very high caries recurrence rates (79.7%) within a year after GA¹⁸. Therefore, there is an urgent need to identify high-risk children susceptible for caries recurrence after GA.

Hence, the present study aimed to use baseline/pre-GA information to build a risk prediction model for early identification of high-risk dental caries recurrent children after GA intervention. A second aim was to characterise the 6-month post-GA changes in risk profiles among caries-free and caries-affected children, and to identify the potential prognostic factors for post-GA caries surveillance/prevention.

Materials and Methods

Study design

Ethical approval was obtained from the Institutional Review Board of Chang Gung Memorial Hospital (IRB no: 102-3921A3). This research was conducted in full accordance with the World Medical Association Declaration of Helsinki. Study participants, recruited from the Children’s Dental Clinic of Kaohsiung Chang Gung Memorial Hospital during the counselling session, underwent oral rehabilitation under GA and were followed up 6 months later. The inclusion criteria included clinical indication for GA [high-risk children with a decayed, missing and filled primary teeth (dmft) score of > 10] and/or unsatisfactory co-operation level, absence of any pre-existing medical condition, age <6 years and written informed consent from the primary caregiver. Sample size calculations were based on an estimated effect size of 0.3. Factoring in an attrition rate of 25%, a minimum of 89 subjects were required to detect an effect at 0.8 power and α = 0.05. Biopsychosocial information (including oral health status) obtained during the baseline visit was used to construct a risk-assessment model to predict caries recurrence at the 6-month recall. Moreover, changes in pre- and post-GA risk factors were compared to characterise the postoperative risk trajectory leading to caries recurrence after GA.

Routine health education before GA intervention

After each child participant and their primary caregiver were examined in the clinic and an accurate diagnosis was obtained, a 10-minute standard oral health counselling session was conducted to advise caregiver/child pairs on the aetiologies, prevention and management of caries; this included cleaning teeth thoroughly twice a day (each time for 3 minutes using fluoride-containing toothpaste), daily flossing, avoiding milk bottle-feeding immediately before or during sleep, reducing snack intake between meals and regular 6-monthly dental check-ups. Proper toothbrushing techniques were also demonstrated to parents.

Data collection

Questionnaire survey

Before and after GA, bio-psycho-behavioural data were collected from each child participant and their primary caregiver through an interviewer-administered questionnaire (Appendix) constructed to collect information on: (i) child’s demographic background (gender and date of birth); (ii) family socio-economic status (caregiver’s educational attainments); (iii) caregiver’s knowledge, attitudes and self-efficacy on oral health; and (iv) child’s oral health practices (dietary habits, oral hygiene measures, use of fluorides and dental attendance).

Microbiological profile

The salivary profile for two common caries-causing bacteria – mutans streptococci (MS) and lactobacillus (LB) – was characterised. Before performing the tests, all children were instructed to refrain from consuming food or drinks or from brushing their teeth, for at least 1 hour. Subsequently, each participant was asked to chew on a piece of paraffin for 5 minutes and the saliva was collected and spread over a Dentocult SM Strip (Orion Diagnostica, Espoo, Finland) and a Dentocult LB test slide (Orion Diagnostica) for determination of MS and LB counts, respectively. After incubation of MS and LB samples at 37 °C for 48 hours, the counts of colony-forming units per millilitre (CFU/ml) saliva were categorised into four levels, based on the manufacturer’s instructions and standard charts, by two calibrated examiners who were blinded to the participants’ caries status.

Oral examination

Caries detection was performed using modified International Caries Detection and Assessment System (ICDAS) criteria¹⁹ and the plaque scores were enumerated using the modified Silness-Löe Plaque Index²⁰ for six assigned teeth (1E, 1B, 2D, 3E, 3B and 4D). The dmft (decayed tooth) and dmfs (decayed surfaces) scores for primary teeth were derived from the ICDAS II cavitated scores (codes 3, 5, 6), which included cavitated lesions, ranging from localised enamel breakdown (code 3), to distinct cavity with visible dentin (code 5), to extensive distinct cavity with visible dentin (code 6). To overcome measurement bias, inter- and intra-examiner reliability were evaluated for the two paediatric dentists involved in caries estimation.

Full-mouth rehabilitation under GA

The same examiners performed full-mouth rehabilitation under GA, including caries removal under rubber dam isolation, pulp therapies, restorative treatments, crowning, and extraction. At the end of GA intervention, a topical fluoride gel (Scodyl®; Beauteeth Company Ltd, Taipei, Taiwan) containing 20,000 ppm fluoride in the form of sodium monofluorophosphate (11.4% w/w) and sodium fluoride (1.1% w/w), was applied on all tooth surfaces for 4 minutes.

Data analysis

All the data were analysed using the Statistical Package for the Social Sciences (SPSS, Version 24; IBM SPSS Statistics, Armonk, NY, USA). Intra- and inter-examiner reliability for oral examinations were evaluated using Kappa statistics based on scores derived from the ICDAS II cavitated stages (codes 3, 5, 6). Data on continuous variables were presented as mean ± standard error (SE), and data on categorical variables are given as N (%). The presence of new caries, defined as any decayed tooth with an ICDAS score of ≥3, at the 6-month recall visit (Δdmft6M-0M) was the main outcome, which was then categorised into two groups (as a binary outcome variable), namely caries-relapse (Δdmft6M-0M > 0) and caries-free (Δdmft6M-0M = 0). Multivariable logistic analysis was performed to build prediction models, with sample prediction performed using the same dataset as that used to derive predictive models, and using pre-GA data according to the type/simplicity of their source of information, namely (i) Level-I: questionnaire, (ii) Level-II: with additional clinical data for Level-IIa and the microbiological tests for II-b, and (iii) Level-III: with addition of oral examination (Table 1). Receiver–operating characteristic (ROC) analysis, quantifying the area under the curve (AUC), sensitivity (Se) and specificity (Sp), were performed to determine the predictive accuracy of the models. McNemar’s test and the paired sample t-test were performed to compare the changes in risk factors (pre- and post-GA intervention) for the whole sample and with subgroup analysis. For all analyses, statistical significance was noted when P < 0.05.

Table 1.

Baseline risk-assessment models used to predict caries recurrence at 6 months post-general anaesthesia (GA)

Baseline risk-assessment models*	Outcome variable: caries recurrence at 6-month recall post-GA
	AUC (95% CI)	Se	Sp	Se + Sp	Accuracy
Level-I† Questionnaire	0.75 (0.64–0.85)	80	63	143	72
Level-IIa‡ Questionnaire Clinical examination	0.80 (0.70–0.89)	80	63	143	72
Level-IIb§ Questionnaire Microbiological tests	0.85 (0.75–0.94)	86	69	155	78
Level-III¶ Questionnaire Clinical examination Microbiological tests	0.87 (0.78–0.96)	91	72	163	82

95% CI, 95% confidence interval; AUC, area under the receiver–operating characteristics curve; Se, sensitivity; Sp, specificity.

^*Risk-assessment models were stratified based on the simplicity of obtaining the source of information.

^†Level-I: questionnaires included information on the age of the child, gender, mother’s education levels, father’s education levels, existing health problems, frequency of sweet snacks, toothbrushing time, average meal time, toothbrushing by whom, use of toothpaste, is putting baby to bed with milk bad, reason for tooth decay, frequency of toothbrushing; significant risk factors were high frequency of sweet snacks [relative risk (RR) = 1.86, 95% CI: 1.07–2.33; P = 0.033] and existing child health problem (RR = 1.58, 95% CI: 1.03–1.87, P = 0.039).

^‡Level-IIa: addition of clinical examination included caries detection and plaque index; significant risk factors were high frequency of sweet snacks (RR = 1.88, 95% CI: 1.06–2.35; P = 0.036), existing child health problem (RR = 1.61, 95% CI: 1.06–1.89; P = 0.032) and high baseline caries (RR = 1.83, 95% CI: 1.08–2.34; P = 0.028).

^§Level-IIb: addition of microbiological tests included oral bacterial counts [Streptococcus mutans (SM) and lactobacillus (LB)]; significant risk factors were high mutans streptococci (MS) count (RR = 4.45, 95% CI: 1.37–4.97; P = 0.026) and existing child health problem (RR = 1.73, 95% CI: 1.03–1.97, P = 0.042).

^¶Level-III: full-blown model, including questionnaires, biological tests and clinical examination; a significant risk factor was existing child health problem (RR = 1.78, 95% CI: 1.08–1.99; P = 0.032).

Results

Participant characteristics

Ninety-two participants were recruited at the start of the study, and 83 (90.2%) returned for the 6-month recall. No differences were noted between the study subjects and those lost to follow-up (P > 0.05), although the reason(s) for their non-participation was not known. Minimal missing data (<1%) were encountered during the analysis. At the time of undergoing GA, the age of the participants ranged from 27 to 71 months, with a mean age of 48.9 ± 10.6 months. At baseline, multiple caries lesions were diagnosed in all study participants (mean dmft = 12.09 ± 0.46). At the 6-month recall, 54.2% (45/83) had developed new caries; the mean increase in dmft was 1.22 ± 0.17. The inter- and intra-examiner reliability for caries detection, tested on 10% of participants, was 0.83 and 1.00, respectively.

Prediction models for caries recurrence after GA

In the Level-I model (AUC = 0.75; Se + Sp = 143), using only questionnaires, multivariable analysis demonstrated ‘high frequency of sweet snack intake between meals’ [relative risk (RR) = 1.86, 95% confidence interval (95%) CI: 1.07–2.33; P = 0.033] and ‘presence of existing health problem in the child’ (RR = 1.58, 95% CI: 1.03–1.87, P = 0.039) to be significant risk predictors for new caries development at the 6-month recall (Table 1).

In the Level-IIa model (AUC = 0.80; Se + Sp = 143), on adding data obtained from clinical information, ‘high frequency of sweet snack intake between meals’ (RR = 1.88, 95% CI: 1.06–2.35, P = 0.036), ‘presence of existing health problem in the child’ (RR = 1.61, 95% CI: 1.06–1.89, P = 0.032) and ‘higher baseline caries experience (dmft ≥ 12)’ (RR = 1.83, 95% CI: 1.08–2.34, P = 0.028) were shown to be significant risk predictors for caries recurrence (Table 1).

In the Level-IIb model (AUC = 0.85; Se + Sp = 155), which included microbiological factors along with questionnaire, ‘presence of existing health problem’ (RR = 1.73, 95% CI: 1.03–1.97; P = 0.042) and ‘high Streptococcus mutans (SM) counts’ (RR = 4.45, 95% CI: 1.37–4.97; P = 0.026) were shown to be significant risk predictors (Table 1).

On including all the variables in the model (Level-III), a clinically satisfactory predictive value was demonstrated (AUC = 0.87; Se + Sp = 163), with ‘presence of existing health problem’ (RR = 1.78, 95% CI: 1.08–1.99; P = 0.032) being a significant risk predictor for new caries development within 6 months after GA (Table 1).

Risk profiles after GA intervention

Overall, significant changes in behavioural and biological factors were noted after GA intervention. There was a significant decrease in the proportion of children with a prolonged meal time (from 55.4% to 41%; P = 0.031), together with a longer toothbrushing time (an increase from 16.9% to 27.7%; P = 0.049), a greater frequency of toothbrushing (an increase from 56.6% to 71.1%; P = 0.038) and more self-toothbrushing by the child (an increase from 19.3% to 44.6%; P < 0.001) after GA intervention. Furthermore, a significant decrease in SM (from 85.7% to 57.5%; P < 0.001) and LB (from 67.2% to 11.1%; P < 0.001) counts was shown, in addition to a decrease in plaque index (P < 0.001) after GA (Table 2).

Table 2.

Comparison of caries risk factors among all study children before and after general anaesthesia (GA)

Caries risk factors	Baseline [N (%)]	6-month recall [N (%)]	P
KAP-related
High frequency of snacks (≥2 times/day)	65 (78.3)	64 (77.1)	1.00
Long toothbrushing duration (≥2 minutes)	14 (16.9)	23 (27.7)	0.049
Long meal duration (≥0.5 hour)	46 (55.4)	34 (41.0)	0.031
Child self-toothbrushing	16 (19.3)	37 (44.6)	<0.001
Use of toothpaste	60 (72.3)	67 (80.7)	0.167
Is putting baby to bed with milk bad?	77 (92.8)	77 (92.8)	1.00
Reason for tooth decay – bacteria and sugar	43 (51.8)	51 (61.4)	0.170
High frequency of toothbrushing (≥2 times/day)	47 (56.6)	59 (71.1)	0.038
Biological
High SM count (≥105)	60 (85.7)	42 (57.5)	<0.001
High LB count (>104)	43 (67.2)	8 (11.1)	<0.001
Plaque index (mean ± SE)	2.44 ± 0.05	1.24 ± 0.06	<0.001*

McNemar’s test was used to compare the risk profiles of caries-free and caries-relapse groups before and after GA.

KAP, Knowledge Attitude Practice; LB, lactobacillus; SE, Standard Error; SM, Streptococcus mutans.

^*Paired sample t-test.

Comparison of risk trajectory between caries-free and caries-affected children

In both caries-free and caries-affected children at the 6-month recall visit, increased self-toothbrushing by the child and reduced SM counts and plaque index were shown after GA intervention (all P < 0.05) (Table 3). However, increased toothbrushing time (from 10.5% to 26.3%; P = 0.031) and a reduction in the length of meal time (from 57.9% to 36.8%; P = 0.021) was observed only in caries-free subjects at the 6-month recall, with no change in caries-affected individuals (P > 0.05) (Table 3).

Table 3.

Change of risk profiles in caries-free and caries-relapse children after general anaesthesia (GA)

Caries risk factors	Caries-free children			Caries-relapse children
	Baseline(%)	6-month recall(%)	P-value	Baseline(%)	6-month recall(%)	P-value
KAP-related
High frequency of snacks (≥2 times/day)	71.1	73.7	1.00	84.4	80.0	0.774
Long toothbrushing duration (≥2 minutes)	10.5	26.3	0.031	22.2	28.9	0.549
Long meal duration (≥1 hour)	57.9	36.8	0.021	53.3	44.4	0.454
Child self-toothbrushing	18.4	47.4	0.013	20.0	42.2	0.013
Use of toothpaste	76.3	84.2	0.508	68.9	77.8	0.344
Putting baby to bed with milk is bad	92.1	94.7	1.00	93.3	91.1	1.00
Reason for tooth decay – bacteria and sugar	57.9	60.5	1.00	46.7	62.2	0.118
High frequency of toothbrushing (≥2 times/day)	63.2	81.6	0.065	51.1	62.2	0.332
Biological
High SM count (≥105)	74.2	40.0	0.013	94.9	73.7	0.021
Plaque index (mean ± SE)	2.34 ± 0.08	1.18 ± 0.09	<0.001*	2.53 ± 0.06	1.29 ± 0.09	<0.001*

McNemar’s test was used to compare the risk profiles of caries-free and caries-relapse groups before and after GA.

KAP, Knowledge Attitude Practice; SE, Standard Error; SM, Streptococcus mutans.

^*Paired sample t-test.

Discussion

This was the first study, among paediatric GA patients, to establish risk-assessment/prediction models with clinically satisfactory predictive accuracy²¹ to identify high-risk children before GA intervention. With proper application and prevention, these models may help may help to determine the various risk factors for dental caries recurrence after costly GA intervention and possibly the need for a repeat surgery in this future. This can be helpful in countries that have seen ever-increasing numbers of paediatric GA patients in the last 10 years².

Based on simplicity of obtaining information, we categorised the risk-assessment models into three levels, with the use of questionnaires being the simplest. Level-I may be suitable for those clinics and hospitals with limited resources, to select GA patients at high risk of caries recurrence for timely and intensive preventive strategies. Interestingly, addition of clinical examination data (Level-IIa) did not increase the prediction accuracy, suggesting that past caries experience may no longer be a critical risk predictor after GA intervention, especially in high-risk children. Although further addition of microbiological tests to the above two models (Level-IIb and III, respectively), did improve the accuracy of the models, the difference was not statistically significant, highlighting the need to determine alternative prognostic factors to predict caries recurrence, especially after clinical intervention.

An increase in child’s self-toothbrushing was seen in all study subjects, reflecting positive reinforcement after oral health education. Furthermore, there was an increase of 15.8% (from 10.5% to 26.3%) in the proportion of children in the caries-free group who brushed their teeth for longer than 2 minutes in the 6 months after treatment under GA (P = 0.031) but this difference was not statistically significant in the caries relapse group (P = 0.549). Although toothbrushing duration of at least 2 minutes is usually recommended and shown to result in substantial plaque reduction (>67%) in adults²², studies have focussed on toothbrushing frequency, instead of its duration, as a potential ECC risk determinant in young children²³. In contrast, our findings show that the duration of toothbrushing may be more critical in preventing post-GA caries recurrence compared with frequency of toothbrushing, especially in high-risk children.

Furthermore, previous studies have largely focussed on the type of meals/sugars²⁴ and/or frequency of between-meal snack intake²⁵ on ECC development, without any data on the duration of the meals. In the present study, a significant reduction in the percentage of children with prolonged meal duration (from 57.9% to 36.8%) was demonstrated only in the caries-free group after intervention, suggesting that meal duration may be an important risk determinant for caries recurrence in high-risk children in Asia where force feeding for infants and young children is culturally rooted²⁶. Longer duration of meals may result in a prolonged acidogenic phase and reduced salivary clearance, leading to the persistence of higher levels of certain carbohydrates, even 1 hour after ingestion of meals²⁷ and thus increasing the chance of demineralisation/cavitation.

One of the limitations of the study is the convenience sample, as in most prospective studies involving subjects undergoing GA. Moreover, internal validation could not be performed because of the limited sample size. Further studies investigating a randomised sample, with a larger sample size, may incorporate internal/external validation of prediction models and thus improve the generalisability of the study findings.

In conclusion, risk prediction models, with promising clinical accuracy, have been constructed using pre-GA data and may be further developed for early identification of high-risk children prone to caries relapse after oral rehabilitation under GA.

Acknowledgements

This study was financially supported by Kaohsiung Chang Gung Memorial Hospital, Taiwan (grant #CMRPG8C1131) and partially assisted by National Medical Research Council (NMRC/CIRG/1341/2012; grant #R221000059511), Singapore.

Conflicts of interest

The authors declare no potential conflicts of interest.

Appendix. Oral Health Questionnaire

• 1Name of child: ____________________
• 2Gender: (1) Male (2) Female
• 3Date of birth: ___________ (dd/mm/yy);
• 4Number of siblings: ____; Order among siblings: ___________
• 5Questionnaire completed by: (1) Mother (2) Father (3) Other guardian: ______________
• 6(A) Highest education of father: (1) No education (2) Primary (3) Secondary (4) High school; (5) Polytechnic or Bachelors’ degree (6) Masters or Doctorate degree (7) Others: ______________;

  (B) Highest education of mother: (1) No education (2) Primary (3) Secondary (4) High school; (5) Polytechnic or Bachelors’ degree (6) Masters or Doctorate degree (7) Others: ______________;

  (C) Nationality of mother: (1) Taiwan; (2) Foreign countries: ______

• 7Who take(s) care of the child outside school/kindergarten hours?

  (1) Mother (2) Father (3) Grandparents (4) Nanny/domestic helper (5) Others: ____________

• 8At what age did the child stop breastfeeding? ______ years ______ months
• 9In the last 6 months, how did the child go to sleep/take a nap?

  (1) Nothing in the mouth (2) With bottle of water (3) With pacifier only (4) Nursing at mother’s breast (5) With bottle of milk/formula/juice (6) With something sweet in the mouth

• 10How many meals and snacks does your child eat in a day?

  (1) 3 times a day (2) 4–5 times a day (3) 6–7 times a day (4) More than 7 times a day

• 11How long is the meal/dinner time?

  (1) Within 30 minutes, (2) 1 hour; (3) More than 1 hour

• 12How often do you give sweet snacks or drinks to your child between meals during a day?

  (1) None (2) Once (3) 2–3 times (4) 4–5 times (5) More than 5 times

• 13Does your child take sweet snacks and then sleep without brushing the teeth?

  (1) Never (2) Occasionally (3) Frequently (4) Almost every night

• 14Who brushes the teeth of your child most frequently?

  (1) Maid (2) Child’s parents (3) Child himself/herself (4) Child’s grandparents (5) Others ________

• 15How many times a day do the child’s teeth get brushed?

  (1) None (2) Once/day (3) Twice/day (4) 3 times/day (5) More than 3 times/day

• 16For how long are your child’s teeth brushed each time?

  (1) <1 minute (2) 1–2 minutes (3) More than 2 minutes, less than 3 minutes (4) ≥3 minutes

• 17Do you agree with the statement ‘I can do a good job brushing my child’s teeth each day thoroughly even when I am very busy’?

  (1) Strongly Agree (2) Agree (3) Neutral (4) Disagree (5) Totally Disagree

• 18Is fluoride-containing toothpaste used for brushing your child’s teeth?

  (1) Yes (2) No (3) Don’t know

• 19Have you or the dentists ever used, for your child, any fluoride-containing products (e.g. fluoride-containing mouthrinse, gel, tablets, varnish)?

  (1) Yes (2) No (3) Don’t know If YES, how frequent? ________________________

• 20Does your child visit the dentist at least once a year?

  (1) Yes (2) No (*if YES, ignore question #21)

• 21If your child does not receive dental check-up every year, what is the main reason?

  (1) No Money (2) No Time (3) Difficult transportation (4) Fear of drills, injection and dentists (5) His/her teeth do not bother him/her (6) Others _____________

• 22At what age do you think is appropriate for your child to start dental check-ups?

  _______ years old

• 23How many decayed teeth does your child have in his/her mouth?

  (1) None (2) 1–2 teeth (3) 3–4 teeth (4) More than 4 teeth (5) Don’t know

• 24Do you believe that putting baby to bed with a milk bottle is bad for his/her teeth?

  (1) Yes (2) No

• 25What do you think is the main reason for tooth decay?

  (1) Tooth worms (2) Heatiness (3) Ineffective tooth brushing (4) Sugar (5) Bacteria (6) Sugar + Bacteria

• 26Does your child have any health problems (e.g. asthma, eczema, food allergy, obesity)?

  (1) Yes (2) No If YES, what are the problems? __________________________

• 27Is your child taking any medication regularly?

  (1) Yes (2) No If YES, what are the medications? _______________________________

  *********************************************************************************************

  *The following questions were included only in the questionnaire for the follow-up visit:

• 28Has your child been seen by other dentist during this study period?

  (1) Yes (2) No (skip the next 2 questions)

• 29If yes, please specify the location:

  (1) Kaohsiung Chang Gung Memorial Hospital, (2) Near home, (3) Others: ____

• 30Please specify the treatment received

  (1) Restoration, (2) Extraction, (3) Fluoride treatment, (4) Sealants, (5) Others: _____

• 31Besides fluoride toothpaste, is there any other method used by your child to prevent caries?

  (1) Yes

  (1a) Toothpick, (1b) Dental floss, (1c) Interdental brush, (1d) Electric toothbrush, (1e) Mouthrinse, (1f) Others: _______

  (2) No