Evaluating effects of animal assisted therapy on pediatric dental patients: a pilot clinical trial

Abstract

Background:

An estimated 6% through 22% of children have dental anxiety, which can contribute to disruptive behavior and dental care avoidance. Evidence from medical settings indicate reductions in pain and stress after therapy dog implementation. To identify a low-risk, non-pharmacological approach for anxiety and pain management in dentistry, a pilot prospective clinical trial was conducted to determine best practices for evaluating the efficacy of animal assisted therapy (AAT).

Methods:

The effects of AAT on pediatric patients were measured through physiologic, objective measures (i.e., heart rate, salivary cortisol and α-amylase, and video coding) and validated self-reported scales of anxiety, fear, and pain, during an invasive dental procedure. Children aged 7 through 14 years were consecutively enrolled into an AAT (n=18) or control (n=21) group. Participants underwent an operative or surgical (e.g., extraction) dental procedure. Descriptive and bivariate statistics were used.

Results:

Participants in the AAT group reported significantly less post-operative pain than those in the control group (p=0.001). The heart rates of AAT participants heart rates dropped after key events and had less variation than control participants. Objective video coding revealed that AAT participants had significantly longer durations of relaxed lower bodies than control participants (p=0.0204). No differences were seen in salivary cortisol and α-amylase. These physiologic and self-report measures are feasible for use in future studies.

Conclusions:

AAT may be an effective therapy for alleviating anxiety and pain in pediatric dental patients and warrants additional study using both objective, physiologic end points and subjective self-report measures.

Practical Implications:

Pending further research, canine therapy may be a valuable addition to dental clinics for anxiety and pain management. This clinical trial was registered at ClinicalTrials.gov. The registration number is NCT04708028.

Graphical Abstract

Introduction

Dental anxiety (DA) is an emotional state of apprehension or a reaction to feeling threatened in a dental setting, manifesting as disruptive behavior and distress.¹ DA is linked to avoidance and increased pain response during dental procedures.^2–4 Avoidant behavior ultimately can increase risk of developing infection and experiencing tooth loss and emergency visits.⁵ The global prevalence of DA in pediatric patients aged 3 through 18 years is 23.9%,⁶ and 16.4% of surveyed adults have DA, with 50.9% reporting anxiety onset in childhood due to early traumatic dental experiences.^7,8

Enhancing children’s treatment experience and preventing lasting fears of dentistry necessitate the application of behavioral management and anxiety-control techniques.⁷ The American Academy of Pediatric Dentistry outlined developmentally appropriate, evidence-based practices for managing pediatric DA, including use of basic and advanced behavior guidance techniques and non-pharmacological and pharmacological treatments. Rare but significant risks are associated with moderate to deep sedation (e.g., cardiovascular instability and impaired airway reflexes and patency) leading to occasional caregiver resistance.^8,9 General anesthesia for DA is safe but costly for the patient and the health care system, and expenses continue to increase.¹⁰ A low-cost, low-risk alternative to sedation is needed for patients who are anxious.

Animal assisted therapy (AAT) is a promising non-pharmacological intervention that uses a well-trained therapy animal to achieve treatment goals.¹¹ Dogs can sense elevated stress hormones and respond accordingly.¹² Data regarding AAT’s positive effects are encouraging and abundant in medicine, with brief AAT interventions decreasing stress hormones and increasing endorphins in adult patients and even health care personnel.^5,13

Although implementation of AAT in dentistry is in its early stages, initial data are promising.^14–16 Among children who verbalized distress, a therapy dog decreased their physiologic arousal while waiting for their dentist.¹⁴ After a randomized trial at a pediatric dental office, researchers reported a 15-minute interaction with a therapy dog reduced subjective stress in children aged 4 through 11 years.¹⁶ Acceptability of AAT has also been surveyed in pediatrics and orthodontics; 90% of caregivers indicated acceptability of a therapy dog in pediatric dental settings while 68% expressed a desire for a dog to be part of their child’s treatment.¹⁵ Orthodontic patients and their parents or guardians indicated broad interest in therapy dogs, with few to no safety concerns.¹⁷

Dental AAT studies have used skin temperature, heart rate (HR), blood pressure, and validated self-report scales to quantify the effects of interventions.^14,18–20 Although valuable, the literature lacks controlled trials that pair self-report and physiologic measures to evaluate impacts of AAT during invasive pediatric dental procedures. To address this gap, we conducted a pilot prospective controlled clinical trial to assess AAT’s effects on objective (e.g., salivary cortisol and α-amylase, video-coded behaviors, and HR) and subjective (e.g., self-report) measures of anxiety, fear, pain, and next visit expectations. We also aimed to identify feasible outcome measures and study recommendations for future, larger randomized trials. If efficacious, AAT will offer dental providers a non-invasive, low-risk approach to provide positive dental experiences, with the goal of reducing dental avoidance and its considerable health sequelae.

Methods

Study design

Details of informed consent, assent, institutional review board approval (#19-1911), and clinical trial registration are included in Supplemental Methods (Appendix). Although our study is not a randomized trial, we followed the Consolidated Standards of Reporting Trials when applicable (Supplementary Material 1).²¹ Participants were recruited from the graduate pediatric dental clinic at the Adams School of Dentistry, University of North Carolina at Chapel Hill, Chapel Hill, NC; pediatric patients were scheduled semi-randomly throughout the week, with no specific allocation for weekday or provider. The parents or guardians of participants (N=39) aged 7 through 14 years who were scheduled to receive an invasive dental procedure (i.e., restoration or extraction, with or without nitrous oxide) were contacted via telephone for screening and enrollment (Appendix Table 1, Figure 1A).

Figure 1. Trial designs and study visit summary.

A. Consolidated Standards of Reporting Trials²¹ flowchart for enrollment and data collection. Flowchart includes the number of patients screened, enrolled, and assigned to the 2 groups (AAT, control). The flowchart shows patients excluded during the screening process before group assignment (n=66) as well as those excluded due to altered treatment plan (n=1). B. Study visit timeline and summary. Continuous video filming with a Hero9 camera (GoPro) and heart rate (HR) recording. Baseline pain and anxiety scales and saliva (S1) were obtained. Participants interacted with a therapy dog for 2 minutes (AAT group), colored a picture of a dog for 2 minutes or sat quietly for 2 minutes. A second saliva sample was obtained 8 minutes post-intervention (S2). Participants then walked to the clinical treatment room, where the dental procedure began; 8 minutes after injection or insertion of the dental handpiece (in participants who did not receive an injection), the third saliva sample (S3) and second pain scale score were obtained. For participants receiving local anesthetic (n=34, 87%), topical anesthetic was first applied, followed by injection of 2% lidocaine with 1:100,000 epinephrine. The procedure ended and the fourth saliva sample (S4) and third pain scale score were obtained 8 minutes later, along with the fear and anxiety scales and Next Visit Expectations questionnaire results.

The pilot trial overview is presented in Figure 1A–B. Participants were allocated to the control group or AAT intervention group, depending on the appointment’s weekday, which was determined on the basis of the therapy dog and handler’s availability. On arrival, participants provided consent in a private consultation room, followed by a calming period of 7 through 9 minutes. Then, continuous video and HR recording began. Anxiety, fear, and pain assessments were completed at the visit start, in addition to collection of an unstimulated saliva sample (S1) using saliva collection Salivette swabs placed into the buccal vestibule. Thereafter, the AAT intervention or control period took place for 2 minutes in the private consultation room. The AAT group interacted with the therapy dog using a reproducible, standardized method; the handler introduced the dog, asked for oral permission to approach the child, and invited participants to pet the dog, shake her paw, and give her treats. The handler and dog then left the room. Participants assigned to the control group colored a picture of a dog. At the end of the intervention, a second saliva sample (S2) was collected. The participant was then escorted into the treatment clinic for the procedure for either a restoration (i.e., composite restoration, sealants or stainless-steel crown) or extraction (Figure 1B).

Pediatric dentistry residents completed all treatments. Per clinic policy, parents or guardians were invited to be in the clinic. The procedure was attended by the patient, dental care provider, assistant and 2 study personnel (N.G., E.N.); a parent or guardian elected to be present in 64% of visits.

The rehabilitation facility dog was a female golden retriever trained to assist in therapeutic interventions within clinical settings alongside health care professionals. Working under the guidance of a trained pediatric dentist dog handler, the rehabilitation facility dog provided support for patients with physical or neurologic disabilities, leveraging more than 2 years of specialized training to manage high-stress environments and establish rapport with multiple patients.

Physiologic measurements

Measurement of salivary cortisol and α-amylase:

Unstimulated saliva samples were collected and processed using saliva collection Salivette swabs according to the manufacturer’s instructions (SalivaBio Oral Swab, Salimetrics) at specified events and stored at −80°C until analysis (Figure 1B, Appendix Table 2). Salivary cortisol was quantified in duplicate using a Salivary Cortisol ELISA Kit (Salimetrics) per the manufacturer’s instructions. a-Amylase activity was quantified in singles with a Salivary Alpha-Amylase Enzymatic Kit (Salimetrics). Details are in Supplemental Methods (Appendix).

Video data:

Video data were captured using a wireless Hero9 camera (GoPro). Videos were coded according to the Paediatric Dental Pain, Anxiety and Fear Coding Approach,²² a software-based observational coding method, in Observer XT software (Noldus). Details are in Supplemental Methods (Appendix).

HR:

Data were collected using a GSR+ unit (Shimmer Sensing) in 32 patients and a Polar Verity Sense device (Polar Electro) in 7 patients at visit start until conclusion. Details are included in Supplemental Methods (Appendix). Missing data due to technical malfunction or inability to collect data are provided in Appendix Table 3; missing data points were omitted from statistical analyses.

Self-reported measures

Self-reported dental fear was measured with the Children’s Fear Survey Schedule-Dental Subscale (CFSS-DS) (Appendix Table 4) and self-reported dental anxiety was measured using the Modified Child Dental Anxiety Scale (MCDAS) (Appendix Table 4).^19,23,24 The CFSS-DS was developed to assess dental fear in children; it is a revised form of the Fear Survey Schedule for Children including dental fear items as a subset (e.g., “the dentist drilling”).²⁵ MCDAS has been utilized in children 8 through 15 years with reliable internal consistency and validity.²⁶ Self-reported pain was measured with the Wong-Baker FACES Pain Rating Scale (WBS).²⁷ At visit’s end, participants were asked to complete the Next Visit Expectations questionnaire (Appendix Table 5). Details on questionnaires are in Supplemental Methods (Appendix).

Statistical analyses

Descriptive statistics (e.g., median and mean [SD]) were used for initial analysis. Anxiety (MCDAS), fear (CFSS-DS), and pain (WBS) scales, Next Visit Expectations questionnaire, and HR data were evaluated using the Mann-Whitney U test. Cortisol and α-amylase data were log₁₀-transformed before analysis and a Mann-Whitney U test was used to test for group differences. A significance threshold of p<0.05 was used. Given the small sample size and non-normally distributed data, non-parametric correlations (Spearman rank correlation) and group comparisons (Wilcoxon test) were performed. Analyses were performed using JMP Pro (Version-17.2.0). Details on figure creation and statistics are in Supplemental Methods (Appendix).

Results

Study participants

Participants (N=39) were well-matched across groups for sex (AAT: 61% female; control: 57% female) and age distributions (AAT: mean, 9.17 years; control: mean, 9.86 years) (Figure 1A, Table 1). Those who identified as Latin American represented a majority of both groups (AAT: 72%; control: 62%). Restorations were performed in 61% of all participants (AAT: 67%; control: 57%) while extractions were performed in 36% of patients (AAT: 33%; control: 38%).

Table 1:

Sample demographic characteristics

	Control (n=21)	Intervention (AAT) (n=18)
Age in years
Mean ± SD	9.86 ± 1.98	9.17 ± 2.09
Median	9	9
Range	7 to 14	7 to 14
Sex
Female	57% (n=12)	61% (n=11)
Male	43% (n=9)	39% (n=7)
Race
African American	14% (n=3)	17% (n=3)
Asian/Pacific Islander	5% (n=1)	0% (n=0)
Caucasian	19% (n=4)	11% (n=2)
Latin American	62% (n=13)	72% (n=13)
Ethnicity
Hispanic	57% (n=12)	78% (n=14)
Non-Hispanic	43% (n=9)	22% (n=4)
Received an injection
Yes	86% (n=18)	89% (n=16)
No	14% (n=3)	11% (n=2)
Received nitrous
Yes	71% (n=15)	89% (n=16)
No	29% (n=6)	11% (n=2)
Type of procedure received
Restoration	57% (n=12)	67% (n=12)
Extraction	38% (n=8)	33% (n=6)
Only injection*	5% (n=1)	0% (n=0)
Frankl Score
Mean ± SD	3.67 ± 0.58	3.78 ± 0.43
Median	4.00	4.00
Frankl 4	71% (n=15)	78% (n=14)
Frankl 3	24% (n=5)	22% (n=4)
Frankl 2	5% (n=1)	0% (n=0)
Pets at home?
Yes	62% (n=13)	67% (n=12)
No	38% (n=8)	33% (n=6)
Anxiety disorder?
Yes	14% (n=3)	0% (n=0)
No	86% (n=18)	100% (n=18)
How afraid to go to the dentist?
Very afraid	19% (n=4)	0% (n=0)
Somewhat afraid	24% (n=5)	6% (n=1)
Only a little afraid	29% (n=6)	39% (n=7)
Not afraid at all	29 % (n=6)	56 % (n=10)
HR device used
Shimmer	67% (n=10)	83% (n=10)
Polar	33% (n=5)	17% (n=2)

^*One control participant received a local anesthetic injection and no subsequent procedure due to uncooperative behavior.

Of the 21 participants assigned to the control group, 15 colored the picture of a dog and 6 had 2 minutes of quiet time. Intraorally injected local anesthetic (2% lidocaine, 1:100,000 epinephrine after topical) was planned for all participants and administered to all except 5 participants (n=34 received an injection, 87%, 18 in the control group, 16 in the AAT group), due to the provider’s decision to change treatments during the visit (n=5 did not receive an injection, 13%, 3 in the control group, 2 in the AAT group). Frankl scores were similar for control and AAT participants, indicating children had a positive to definitely positive attitude towards the dentist (Table 1) (Control: mean, 3.7; AAT: mean, 3.8).²⁸

Effect of AAT on physiological measurements

No differences were observed between groups at any event point for salivary cortisol or α-amylase (Figure 2A–B, Appendix Tables 6–7).

Figure 2. Physiologic objective measures.

A. Salivary cortisol levels at visit events. Each dot represents a study participant B. Salivary amylase (ɑ-amylase) levels at visit events. Each dot represents a study participant. C. Heart rate data with corresponding visit events. D. Behavioral coding results for relaxed lower body between groups. Control (gray) and AAT (blue) groups are displayed on the x-axis. The y-axis shows the mean duration (in minutes) of relaxed lower body during dental procedure. Significance threshold p<0.05. H1: After participant’s calming period, but before the intervention, not concurrent with saliva collection swab placement. H2: During intervention (AAT, Control). H3: Immediately after intervention (AAT, Control). H4: Immediately after movement to the clinical treatment room. H5: During first injection. H6: Immediately after first injection. H7: Placement of first isolation or mouth prop. H8: Beginning procedure with handpiece or forceps into mouth. H9: During procedure (10 minutes into the procedure). H10: After the procedure (removal of the mouth prop, patient is seated upright). S1: Baseline at start of study visit. S2: Eight minutes after intervention (AAT or control). S3: Eight minutes into the dental procedure. S4: Eight minutes after completion of the procedure at the end of the study visit.

For every event, there was a smaller range in HR variance among the AAT group compared with the control group (Figure 2C, Appendix Table 8). At events immediately after movement to the clinical treatment room and between events immediately after the procedure started and 10 minutes into the procedure (Figure 2C), there was a decrease in HR observed in the AAT group that was not present in the control group.

Effect of AAT on behavioral measures of the Paediatric Dental Pain, Anxiety, and Fear Coding Approach

Participants who received AAT had a significantly longer mean duration of relaxed lower body than control participants (Figure 2D) (p=0.0204). Post-procedure CFSS-DS and MCDAS scores had significant positive correlations with the number of verbal indicators, such as crying and moaning (CFSS-DS correlation, 0.48; MCDAS correlation, 0.46; p=0.004) (Appendix Table 9). Both baseline and post-procedure CFSS-DS scores were also positively correlated with MCDAS survey scores collected pre- and post-procedure. Finally, post-injection pain was significantly higher (correlation, 0.33; p=0.048) among those who had more hand movements during the procedure.

Effect of AAT on self-reported assessments

Post-procedure, members of the control group reported significantly more pain on WBS than the AAT group (p=0.001) (Figure 3A, Appendix Table 6). In the control group, 53% reported increased pain post-procedure, and 21% of AAT participants reported increased pain.

Figure 3. Survey responses.

A. Wong-Baker FACES Pain Rating Scale responses with corresponding events. Box plot visualization of perceptual pain scores compared between the animal-assisted therapy (AAT) (blue) and control (gray) groups for each visit event. Significant differences (p<0.05) are indicated with an asterisk (*). Each dot in the box plot represents a study participant. B. Next Visit Expectation questionnaire responses. A higher score indicates a more positive outlook for future visits. C. Anxiety Scale Responses. Scatter plot for both anxiety surveys (Children’s Fear Survey Schedule-Dental Subscale, Modified Child Dental Anxiety Scale), each shown on a separate axis. Each dot presents the change in anxiety (post-study visit score minus start of visit baseline score) for each participant. A higher, positive score indicates an increase in anxiety from the start of the study visit to the end.

Participants were asked to self-assess fear and anxiety at the start of the visit and after the dental procedure to evaluate changes across the appointment; trends were observed, although no change or group comparison was significant. When evaluating fear using CFSS-DS, 35% of the control group reported higher scores (e.g., higher fear) while 56% of the AAT group reported higher scores after the procedure (not significant) (Figure 3C). Higher scores for MCDAS (e.g., higher anxiety) were reported in 45% of the control group and 37% of the AAT group. The control group had a mean (SD) change in fear of −2.05 (13.29) points and a mean (SD) change in anxiety of −0.65 (6.34), indicating decreased fear and anxiety over the visit. The AAT group reported a mean (SD) increase of 1.577 (5.90) on the fear scale and a mean (SD) decrease of 0.195 (4.71) on the anxiety scale. There was a higher, more positive overall score on the Next Visit Expectations questionnaire among AAT participants, although this was not significant, as our sample size was likely underpowered for this measure (Figure 3B).

Discussion

This pilot prospective controlled clinical trial was designed to explore the use of AAT in a pediatric dental setting. If effective, it may improve pediatric dental experiences and mitigate long-term DA and dental avoidance.^2–4 However, there is a dearth of comprehensive data on how to study AAT’s effects in dentistry and whether AAT is efficacious at managing dental fear, anxiety, pain, or behavior. This pilot was designed to begin evaluating AAT in dental settings and identify feasible outcome measures for future studies.

Stress hormones were studied in pediatric dental patients as markers of sympathetic nervous system (SNS) activation via saliva, a readily available biofluid.^29,30 Cortisol is a marker of the hypothalamic-pituitary-adrenal axis and levels peak approximately 20 through 30 minutes after onset of stress.^29,31 Salivary cortisol was investigated in children aged 2 through 5 years with behavior management challenges undergoing dental prophylaxis.^32,33 Salivary α-amylase was used in pediatric dental settings as a screening tool for fear.³⁴ In our study, no differences were found in these hormones, likely because collecting saliva was a new experience for patients, causing SNS arousal, or because saliva was collected only 8 minutes after each visit event; longer wait times were not feasible in this setting. The inter-individual variability of cortisol and α-amylase levels in our dataset may also be attributed to their diurnal rhythms; although all appointments took place between 9 AM and 5 PM, we could not control or limit the scheduled times of dental visits.

An additional biometric measure was HR; a drop in HR immediately after movement to the clinical treatment room and after the procedure started (Figure 2C) was observed and may reflect a decrease in DA. Because anxiety is anticipatory in nature, once an anxiety-inducing event occurs (e.g., going into the treatment room or starting a procedure), the HR decreases. Contrary to this, we observed a less pronounced, steady decline in HR among control participants.

DA, fear, and pain were also objectively measured using the Paediatric Dental Pain, Anxiety and Fear Coding Approach²². The significantly longer duration of relaxed lower body found in the AAT group (Figure 2D) may suggest a reduced response to fear and anxiety. Both the CFSS-DS and MCDAS positively correlated with verbal indicators (e.g., crying or moaning and verbal responses to pain), meaning those who reported greater fear and anxiety also vocalized more frequently. The positive correlation between hand movements and post-injection pain scores indicated that participants were more likely to put their hands over their mouth or chest when experiencing pain (Appendix Table 9). Dental providers can visually track these behaviors as a reflection of patient’s DA, fear and pain.

Anxiety and fear are related but distinct states that exist on a continuum to alert people to situations perceived as threatening; anxiety prevents or prepares for future threatening situations, and fear is an in-the-moment response to manage immediate threats with an organized nervous system arousal response.^35–37 Separating fear and anxiety in clinical dental settings is operationally challenging and was addressed here using a fear and a separate anxiety questionnaire. As expected, results of these surveys showed positive but imperfect correlations, verifying that they correlate but measure distinct underlying constructs. When evaluating fear using the CFSS-DS scale, a larger percentage of the AAT group (56%) reported more fear after the procedure than the control group (35%) (Figure 3C). Higher scores were reported on the MCDAS in a larger proportion of the control group (45%) than the AAT group (37%), suggesting a larger percentage of those in the control group reported higher anxiety after the procedure. With an underpowered sample, neither trend reached significance.

Self-report scales using the WBS are validated for measuring pain in pediatric populations.²⁷ Dental pain has also been linked to anxiety.³⁸ Our findings showed that the AAT group reported significantly less pain post-procedure, suggesting that an AAT intervention may reduce post-procedure pain perceptions and DA. Although not significant, an overall higher, more positive score in the Next Visit Expectations questionnaire among AAT participants suggests a reduction in DA when anticipating future appointments. Altogether, these data suggest that AAT may reduce DA for improved oral health outcomes, although future studies are needed for definitive evidence, guided by our lessons learned (described below).

Limitations:

Due to its pilot nature, our study was insufficiently powered to detect small and medium effect sizes. Block randomization was not possible due to limited availability of the therapy dog handler. There were variations in dental provider, procedure type, length, use of nitrous oxide and local anesthetic, and presence of parents or guardians in the treatment room that we could not control, as the study was designed for applicability to current pediatric practice. Variation in procedure type and use of anesthetic administration was a limitation, as a sealant is likely a less anxiety-inducing for most participants than a restoration or crown involving an injection; however, all participants were scheduled to receive an invasive dental procedure requiring an injection, such that the anticipatory anxiety was likely more uniform than the ultimate procedures chosen by the dental care provider. Future studies should have more uniform procedures or large enough cohorts to allow for stratification.

Due to variable appointment times and a small sample, data could not be effectively stratified according to sex, age, or time of day; therefore, we cannot rule out impacts of diurnal hormonal variation and other confounding variables.²⁹ Similarly, our study was too underpowered to compare patients whose caregivers were present versus absent from the clinic, as their presence may influence the patients’ experience; to limit the effects of the caregiver’s presence, parents and guardians were asked to minimize interaction with their child during the visit. Our naturalistic study design reflects realistic pediatric settings with real-time treatment decisions; therefore, the study could not influence provider treatment decisions or caregivers’ presence. We used a brief therapy dog intervention; however, longer interactions with the dog chairside throughout visits should be evaluated for efficacy. Potential effects on DA and pain perceptions were seen, even with a brief, 2-minute therapy dog intervention.

Conclusions

There was less variation in HR range among AAT participants than the control group, with decreases in HR after key events, such as procedure start, which may suggest a decrease in DA. No differences were noted in fear and anxiety scales, and salivary cortisol and α-amylase levels. The AAT group had longer durations of relaxed lower body and reported significantly less pain at the end of the procedure than the control group. Although a larger randomized controlled trial is needed to reach sufficient power, our pilot study contributes to the growing literature supporting use of AAT in dental settings and guides methodology for future investigations. AAT may be an effective adjunctive non-pharmacological anxiety and behavioral management tool for improving the experience of pediatric dental patients in the dental office and reducing future avoidance behavior.

Data Availability Statement

The data supporting the findings of this study are available within the article and its supplementary materials.