Assessing Mechanisms of Anxiety Reduction in Animal-assisted Interventions

The specific aims of this project are to (1) test the mechanisms by which AAIs reduce anxiety, and (2) determine if the anxiolytic effect of social and physical interaction is moderated by level of pre-existing social anxiety. To achieve these aims, 75 adolescents (age 13-17) will undergo a well-validated laboratory-based social evaluative stressor, the Trier Social Stress Task for Children, and be randomly assigned to one of three conditions: 1) no interaction with a dog (control condition), 2) social interaction only (no physical interaction) with a therapy dog; or 3) social interaction plus physical interaction with a therapy dog. Using a multivariate approach, three levels of outcome data will be collected: a) self-reported experience (anxiety), b) autonomic physiology (heart rate), and c) behavioral performance (error rates on mental math task). In addition, the interactions will be videotaped and behavioral coding will be used to explore the specific social behaviors between the participant and the dog that may predict anxiety reduction (such as frequency or type of social referencing or physical contact).

• animal-assisted intervention (Other): Interaction with a therapy dog
• active control (Other): Interaction with a stuffed dog

Primary outcomes

• Self-reported Affective Experience [Self-reported anxiety was measured at six time points, during: (Time 1; 0 min) baseline, (Time 2; 30 min) anticipation, (Time 3; 35 min) preparation, (Time 4; 45 min) speech, (Time 5; 60 min) recovery period 1, and (Time 6; 75 min) recovery period 2.]: The state scale of the State-Trait Anxiety Inventory (STAI) was used to measure state-level anxiety. We used the six-item short form of the STAI, which asks participants to rate how each of the six words reflects their feelings (calm, upset, relaxed, worried, tense, content). The short form was originally administered as a four-point scale, which we further modified to a three-point scale for feasibility in administering repeatedly over a short time period (response options for each item followed the format: very calm, calm, not calm). Responses to the six items were used to create a sum score at each time point with a possible range of 3 to 18 (higher scores indicate higher levels of anxiety).

Secondary outcomes

• Autonomic Physiological Reactivity: Electrodermal Activity [Continuous through the 2 hour experiment.]: Electrodermal activity (EDA) was recorded at 4 Hz by an Empatica E4 wristband sensor. EDA is reported in μSiemens with higher values representing higher levels of autonomic reactivity. EDA was recorded continuously, and for analysis was assessed over six time windows during the experiment: Time 1 = beginning of study; Time 2 = baseline period; Time 3 = anticipation phase; Time 4 = last 5 min of the stressor; Time 5 = recovery 1; Time 6 = recovery 2. Each time point was 5 minutes in duration.
• Autonomic Reactivity: Heart Rate [continuous through the 2 hour experiment]: Heart rate (beats per minute) Heart rate was measured via photoplethysmography from the Empatica E4 wristband using Empatica's proprietary algorithm, which automatically imputes missing data from the photoplethysmograph signal and corrects for motion artifacts. The heart rate is computed as the average heart rate values that spans 10 seconds. This average HR is computed at 1 Hz. HRwas recorded continuously, and for analysis was assessed over six time windows during the experiment: Time 1 = beginning of study; Time 2 = baseline period; Time 3 = anticipation phase; Time 4 = last 5 min of the stressor; Time 5 = recovery 1; Time 6 = recovery 2. Each time point was 5 minutes in duration.
• Cognitive Performance - Number of Errors [1 hour into 2 hour experiment]: Number of errors (incorrect answer to subtraction task) during mental math task; better performance was characterized by fewer errors. Number of errors ranged from 0 to 8.
• Cognitive Performance - Lowest Number Reached/Highest Number of Correct Responses [1 hour into 2 hour experiment]: Highest number of correct responses in serial subtraction task. To adjust for the different levels of subtraction based on age level, lowest number reached was operationalized by calculating the number of correct responses (a higher score indicating better performance). Number of correct answers ranged from 1 to 41.

Sex: All, ages: 13 Years to 17 Years, accepts healthy volunteers

Inclusion Criteria:

\- Low, mid-range, and high levels of social anxiety

Exclusion Criteria:

* Fear of dogs
* Allergy to dogs

Participant flow

Outcome results

• Self-reported Affective Experience (Primary), Mean, units on a scale: Control - Stuffed Dog 10.19 +/- 1.81; Therapy Dog - Social 10.00 +/- 2.07; Therapy Dog - Social + Physical 10.28 +/- 1.60 (p .723)
• Autonomic Physiological Reactivity: Electrodermal Activity (Secondary), Mean, μSiemens: Control - Stuffed Dog 7.32 +/- 3.03; Therapy Dog - Social 6.56 +/- 2.52; Therapy Dog - Social + Physical 8.01 +/- 3.41 (p .271)
• Autonomic Reactivity: Heart Rate (Secondary), Mean, Beats per minute: Control - Stuffed Dog 79.48 +/- 7.92; Therapy Dog - Social 78.73 +/- 10.55; Therapy Dog - Social + Physical 80.44 +/- 11.36 (p .638)
• Cognitive Performance - Number of Errors (Secondary), Mean, number of errors: Control - Stuffed Dog 3.29 +/- 1.95; Therapy Dog - Social 3.36 +/- 2.24; Therapy Dog - Social + Physical 3.20 +/- 2.20 (p .868)
• Cognitive Performance - Lowest Number Reached/Highest Number of Correct Responses (Secondary), Mean, highest number of correct responses: Control - Stuffed Dog 12.95 +/- 8.85; Therapy Dog - Social 13.41 +/- 9.48; Therapy Dog - Social + Physical 14.92 +/- 12.59 (p .809)

Adverse events

Time frame: 75 minutes

Limitations and caveats: Heart rate data was of poor quality due to movement artifacts and therefore unable to be reliably interpreted.

• Results publication: Mueller MK, Anderson EC, King EK, Urry HL. Null effects of therapy dog interaction on adolescent anxiety during a laboratory-based social evaluative stressor. Anxiety Stress Coping. 2021 Jul;34(4):365-380. doi: 10.1080/10615806.2021.1892084. Epub 2021 Mar 2. PMID 33650444

• Study Protocol (PDF, 2019-01-31)
• Informed Consent Form (PDF, 2019-01-31)
• Statistical Analysis Plan (PDF, 2020-04-22)

• Megan K Mueller, PhD, Principal investigator, Tufts University