Class Format
The DFV classes were standardized, and each week followed a consistent structure. Dance for Veterans is a 1-hour class that begins with a greeting and an expression of gratitude as represented by movements developed by individual class members. After listening to an introduction, the seated participants perform yogalike stretches that promote relaxation and improve flexibility. The stretches are followed by rhythm games. Participants repeat and change rhythms to the sounds of upbeat songs, thereby enhancing their observation and listening skills, creativity, and sense of mastery. Then, in Brain Dance, the middle part of the class, memory and coordination are challenged as participants learn a 7-part movement progression.23 Last is a group creative exploration call-and-response activity, usually a game in which the group coordinates participants’ names with their specific individual movements. Each participant says his or her name and creates an individual movement to represent himself or herself; the group then echoes that participant’s name and movement. This activity fosters group cohesion and creativity while improving attention, memory, and a sense of self-worth.
Instructor Training
The 12-week course of DFV classes was led by Dr. Steinberg-Oren and Dr. Krasnova as part of the LAACC general mental health program. The instructors received intensive training in DFV implementation from Sarah Wilbur, a doctoral student in the UCLA department of World Arts and Cultures/Dance and one of the founders of DFV. Training involved written materials and a half-day retreat. Ms. Wilbur modeled the class for 8 weeks. Then she observed the teachers and provided corrective feedback for another 4 weeks. After the 12-week training period, Ms. Wilbur attended class periodically to monitor how closely the instructors were following the prescribed class format and to provide helpful suggestions and new exercises.
Participants
Veterans receiving outpatient psychiatric care for PTSD at LAACC were recruited for the DFV class. They had undergone a thorough psychiatric interview and been found to meet the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM) criteria for PTSD. All underwent physical screening by a primary care provider to rule out preexisting medical conditions that would contraindicate taking the class. Participation was voluntary. Data analysis was approved by the institutional review board of VAGLAHCS.
Data Collection
Sixty-one veterans entered the class on a rolling basis from August 2012 to November 2014. At the participants first class, they completed a demographic questionnaire. For each of the first 12 sessions attended, they were asked to complete the State-Trait Anxiety Inventory (STAI) form Y before and immediately after class.24 The STAI is a self-report questionnaire that measures short-term state anxiety and long-term trait anxiety as characterized by tension, apprehension, nervousness, and worry. It lists the same 20 items twice, first for state anxiety and then for trait anxiety. This valid and reliable measure of generalized anxiety, which has been used in hundreds of research studies, has test–retest intervals ranging from 1 hour to more than 3 months.24,25 Veterans in the study were also asked to provide qualitative feedback on any mood or sense-of-self changes experienced from the time they entered class to once it was completed.
For data analysis, a final sample of 20 veterans was selected. These veterans had completed at least 12 preclass and postclass STAI ratings within the 4-month period. The other 41 veterans in the study were not included in the data analysis because of inconsistent attendance, tardiness, or leaving class without completing a questionnaire. Further, because a large amount of STAI trait data was missing, only state items were analyzed. The data of the veterans who completed their ratings were double-entered to minimize recording errors.
Of the 20 completers (all men), 7 (35%) were African American, 7 (35%) were Hispanic, 5 (25%) were white, and 1 (5%) declined to report race. Completers’ ages ranged from the 40s to the late 70s; 40% (the largest grouping) were between ages 60 and 69 years. Noncompleters’ demographic data were comparable. Of the 41 noncompleters, 14 (34%) were African American, 15 (37%) were Hispanic, 8 (19%) were white, and 4 (10%) were Asian or Pacific Islander. Noncompleters’ ages also ranged from the 40s to the late 70s, with the largest grouping (58%) between ages 60 and 69 years. Thus, the authors did not find any significant differences between completers and noncompleters.
Results
A mixed-effects linear model was used to assess whether participation length (in weeks), testing time (preclass vs postclass), or the interaction of these two variables were significant predictors of state anxiety as measured by STAI. This model included a random intercept by participant to account for differences in baseline stress levels. Analyses revealed a significant main effect of testing time on STAI state scores, t(458) = 7.48, P < .0001, such that class participation appeared to be associated with a mean decrease of 11 points on the state scale (Figure). However, participation length was not a significant predictor of STAI state scores, t(458) = 1.20, P = .233, and there was no interaction effect, t(458) = –0.57, P = .567.
