Bar6.qxd PSYCHIATRIC SERVICES ♦ June 1998 Vol. 49 No. 6 779977 The Effects of Animal-Assisted Therapy on Anxiety Ratings of Hospitalized Psychiatric Patients SSaannddrraa BB.. BBaarrkkeerr,, PPhh..DD.. KKaatthhrryynn SS.. DDaawwssoonn,, PPhh..DD.. Within the last decade, stud-ies supporting the healthbenefits of companion ani- mals have emerged (1–4). Cardiovas- cular effects are often the focus, due partly to findings from a 1980 study that reported longer survival rates fol- lowing myocardial infarction for pet owners compared with people with no pets (5). More recent evidence of cardiovascular benefit was document- ed in an Australian study involving 5,741 participants (6). The authors found that pet owners had significant- ly lower blood pressure and triglyc- eride levels compared with non-pet- owners, and the differences could not be explained by differences in ciga- rette smoking, diet, body mass index, or socioeconomic profile. Objective: Animal-assisted therapy involves interaction between pa- tients and a trained animal, along with its human owner or handler, with the aim of facilitating patients’ progress toward therapeutic goals. This study examined whether a session of animal-assisted therapy re- duced the anxiety levels of hospitalized psychiatric patients and whether any differences in reductions in anxiety were associated with patients’ diagnoses. Methods: Study subjects were 230 patients referred for therapeutic recreation sessions. A pre- and posttreatment crossover study design was used to compare the effects of a single animal-assist- ed therapy session with those of a single regularly scheduled thera- peutic recreation session. Before and after participating in the two types of sessions, subjects completed the state scale of the State-Trait Anxiety Inventory, a self-report measure of anxiety currently felt. A mixed-models repeated-measures analysis was used to test differences in scores from before and after the two types of sessions. Results: Sta- tistically significant reductions in anxiety scores were found after the animal-assisted therapy session for patients with psychotic disorders, mood disorders, and other disorders, and after the therapeutic recre- ation session for patients with mood disorders. No statistically signifi- cant differences in reduction of anxiety were found between the two types of sessions. Conclusions: Animal-assisted therapy was associated with reduced state anxiety levels for hospitalized patients with a vari- ety of psychiatric diagnoses, while a routine therapeutic recreation ses- sion was associated with reduced levels only for patients with mood dis- orders. (Psychiatric Services 49:797–801, 1998) Stress and anxiety are considered contributory factors to cardiovascular disease. Investigators have hypothe- sized that companion animals may serve to lower levels of stress and anx- iety (4,7,8). Several authors have re- ported lower blood pressure readings among adults and children when a previously unknown companion ani- mal is present during various stressful activities (5,9–14). Animals have been associated with positive effects on patients in a variety of health care settings (15). When an- imals were first introduced to these settings, they were generally brought for visits that were incidental to the treatment program. Currently, ani- mals are purposely included in treat- ment through various interventions broadly known as animal-assisted therapy. Animal-assisted therapy involves the use of trained animals in facilitat- ing patients’ progress toward thera- peutic goals (16). Interventions vary widely, from long-term arrangements in which patients adopt pets to short- term interactions between patients and a trained animal in structured ac- tivities. Although animals have typically been well received on psychiatric ser- vices, much of the data attesting to their benefits has been anecdotal (17–19). Several decades ago, Searles (20) and Levinson (21) addressed the therapeutic benefit of a companion dog for patients with schizophrenia, contending that the caring, human- canine relationship helped ground the patient in reality. Chronic mental- ly ill residents in supportive care Dr. Barker is associate professor of psychiatry, internal medicine, and anesthesiology and Dr. Dawson is affiliate assistant professor of biostatistics at the Medical College of Virginia, Virginia Commonwealth University, P.O. Box 980710, Richmond, Virginia 23298. Dr. Barker’s e-mail address is sbbarker@hsc.vcu.edu. PSYCHIATRIC SERVICES ♦ June 1998 Vol. 49 No. 6779988 homes who were visited by puppies had decreased depression after the visits, compared with a matched con- trol group (22). More recently, Arnold (23) de- scribed the use of therapy dogs with patients with dissociative disorders. Benefits included the dog’s calming in- fluence, ability to alert the therapist early to clients’ distress, and facilita- tion of communication and interac- tion. Others have proposed that an an- imal can serve as a clinical bridge in psychotherapy, providing an entree to more sensitive issues (16,24,25). On an inpatient psychiatric unit, animal-assisted therapy was found to attract the greatest number of pa- tients among those who selected groups to attend voluntarily and was found to be the most effective in at- tracting isolated patients (26). Other researchers found that a group meet- ing for psychiatric inpatients held in a room where caged finches were lo- cated had higher attendance and higher levels of patient participation, and was associated with more im- provement in scores on the Brief Psychiatric Rating Scale, compared with a matched group held in a room without birds (27). Anecdotally, psy- chiatric patients who are withdrawn and nonresponsive have been de- scribed as responding positively to a therapy dog with smiles, hugs, and talking (16). For elderly patients with dementia, lower heart rates and noise levels were associated with the presence of a therapy dog (28), and patients with Alzheimer’s disease sig- nificantly increased socialization be- haviors when a therapy dog was nearby (29). Based on the evidence in the litera- ture associating companion animals with anxiety reduction and with posi- tive responses from clinical popula- tions, this study investigated the ef- fect of an animal-assisted therapy group session on the anxiety levels of psychiatric inpatients. Also of re- search interest was whether any anxi- olytic effect found varied by diagnos- tic group. Methods A pre- and posttreatment crossover design was used for this study. Changes in anxiety ratings were com- pared for the same patients under two conditions: a single animal-assist- ed therapy group session and a single therapeutic recreation group session that served as a comparison condi- tion. The setting for this study was the inpatient psychiatry service of an ur- ban academic medical center. The service treats adult patients with a full range of acute psychiatric disorders. The average length of stay is seven to eight days. The animal-assisted therapy session consisted of approximately 30 min- utes of group interaction with a ther- apy dog and the dog’s owner. During the semistructured session, which was held once a week, the owner talked generally about the dog and encouraged discussion about pa- tients’ pets as the dog moved freely about the room interacting with pa- tients or carrying out basic obedience commands. The comparison condition was a therapeutic recreation group session held on the unit on the day following the animal-assisted therapy session. Therapeutic recreation sessions were held daily on the unit. They varied in content, including education about how to spend leisure time, presenta- tions to increase awareness of leisure resources in the community, and mu- sic and art activities. Coordination of both the animal-assisted therapy ses- sions and the therapeutic recreation sessions was shared by three recre- ational therapists. The study used the state scale of the State-Trait Anxiety Inventory to measure patients’ levels of anxiety be- fore and after the animal-assisted therapy session and the therapeutic recreation session (30). The State- Trait Anxiety Inventory is a brief, easy-to-administer self-report mea- sure that is widely used in research and clinical practice. The state scale, which measures the level of anxiety felt at the present time, has been found to be sensitive to changes in transitory anxiety expe- rienced by patients in mental health treatment. The inventory consists of 20 items related to feelings of appre- hension, nervousness, tension, and worry. For each item, subjects circle one of four numbers corresponding to ratings of not at all, somewhat, moderately so, or very much so. In- struments are scored by calculating the total of the weighted item re- sponses. Scores can range from 20 to 80, with greater scores reflecting higher levels of anxiety. The internal consistency for the state scale of the State-Trait Anxiety Inventory is high; median alpha coef- ficient is .93 (30). The construct valid- ity is supported by studies showing that state scale scores are higher un- der stressful conditions. Procedures A total of 313 adult psychiatric pa- tients consecutively referred for ther- apeutic recreation over an eight- month period in 1996 were eligible for the study. Patients are referred for therapeutic recreation as soon as they are stable enough to participate in group activities, generally within 24 to 72 hours of admission. When patients were initially re- ferred for therapeutic recreation, they were asked to sign a consent form to participate in a group session involving a therapy dog. Patients were not eligible to participate if they had any known canine allergies, were fearful of dogs, or did not sign a con- sent form. Study subjects attended both an animal-assisted therapy group session and a therapeutic recreation group session. The two types of sessions were held once a The animal-assisted therapy session consisted of about 30 minutes of group interaction with a therapy dog and the dog’s owner. PSYCHIATRIC SERVICES ♦ June 1998 Vol. 49 No. 6 779999 week on consecutive days at the same time on each day. The three recreational therapists providing services to the inpatient psychiatry unit volunteered to assist with the study. Because the thera- pists were not blind to the treatment condition, steps were taken to mini- mize bias by training the therapists in standard data collection proce- dures. At the beginning and end of each animal-assisted therapy group session and the comparison thera- peutic recreation group session the following day, the recreational thera- pist administered the State-Trait Anxiety Inventory. The therapists read the instrument verbatim to any patient who had difficulties reading. For the animal-assisted therapy group, the pretreatment instrument was completed before the dog en- tered the room. Two female owners of therapy dogs volunteered to provide the animal-as- sisted therapy sessions. The first vol- unteer provided the therapy for the initial four months of the study; then she became ill and could not contin- ue. The second volunteer agreed to continue the study following the same format used by the first volunteer. Her participation required reversing the days that the animal-assisted ther- apy session and the therapeutic recre- ation session were offered. The dogs and owners met hospital policy for participating in animal-as- sisted therapy, including documenta- tion of the dog’s current vaccinations, controllability, and temperament. The volunteers were advised of the animal-assisted therapy group session and given direction on how to lead the therapy group. Analysis Instruments were scored twice for ac- curacy by one of the authors using the scoring keys for the State-Trait Anxi- ety Inventory. A mixed-models re- peated-measures analysis was used to compare pre- and posttreatment dif- ferences in anxiety scores between and within the animal-assisted thera- py condition and the therapeutic recreation condition by diagnostic category. Results Because this study was conducted in a clinical setting, pre- and posttreat- ment measures on all subjects under both conditions were difficult to ob- tain. Six patients refused to partici- pate because of canine allergies or fear of dogs. Of the 313 patients who were eligible for the study, 73 percent (N=230) participated in at least one animal-assisted therapy group session or one recreation group session and completed a pre- and a posttreatment measure for the session. Fifty patients completed a pre- and a posttreatment measure for both types of sessions. Failure to complete all four measures was primarily due to time conflicts with medical treatments and patient discharges. Patient characteristics The mean±SD age of the 313 pa- tients referred for therapeutic recre- ation was 37±12 years, and their mean length of stay was 10.98±8.88 days. A total of 174 patients were women, and 139 were men. The ma- jority were black (169 subjects, or 54 percent) and single (195 subjects, or 63 percent). They had completed an average of 11.3±2.6 years of educa- tion. For analysis, patients were catego- rized by primary discharge diagnosis. The diagnoses were collapsed into four categories: mood disorders, in- cluding all depressive, bipolar, and other mood disorders, for 154 pa- tients (49.2 percent); psychotic disor- ders, including schizophrenia, schizo- affective disorder, and other psychot- ic disorders, for 80 patients (25.6 per- cent); substance use disorders, for 52 patients (16.6 percent); and all other disorders, including anxiety, cogni- tive, personality, and somatization disorders, for 27 patients (8.6 per- cent). TTaabbllee 11 Mean pretreatment, posttreatment, and change scores on the State-Trait Anxiety Inventory for hospitalized psychiatric pa- tients with various diagnoses who participated in an animal-assisted therapy session or therapeutic recreation Animal-assisted therapy (N=93) Therapeutic recreation (N=137) Diagnosis and measure N Mean SD F1 p< N Mean SD F1 p< Mood disorders Pretreatment 53 47.58 12.73 83 47.58 12.73 Posttreatment 45 42.35 12.74 80 43.68 10.78 Change 44 4.05 9.08 6.71 .01 80 4.08 9.08 16.8 .001 Psychotic disorders Pretreatment 34 48.47 15.26 45 46.31 12.55 Posttreatment 26 42.65 12.53 39 43.41 14.56 Change 26 5.77 13.72 7.62 .006 39 2.08 9.06 1.90 ns Substance use disorders Pretreatment 16 48.93 13.17 15 45.33 11.30 Posttreatment 13 44.07 14.00 14 43.28 8.19 Change 12 5.17 10.00 2.66 ns 13 .69 7.79 .09 ns Other disorders Pretreatment 11 52.81 11.68 5 50.40 19.56 Posttreatment 11 45.54 13.74 5 50.20 18.74 Change 11 7.27 9.67 5.06 .026 5 .20 6.87 .003 ns 1 df=1, 194 PSYCHIATRIC SERVICES ♦ June 1998 Vol. 49 No. 6880000 Comparison of therapy groups Table 1 shows the mean scores of the 230 study participants on the State- Trait Anxiety Inventory before and af- ter attending an animal-assisted ther- apy group session and a therapeutic recreation group session as well as the mean change scores. Change scores were calculated using data from pa- tients with measures at both pre- and posttreatment time points. The F test and p values show the significance of the change across time. No statistical- ly significant differences in anxiety change scores were found between animal-assisted therapy and thera- peutic recreation. Although no signif- icant between-group differences were found, within-group differences were statistically significant for both animal-assisted therapy and thera- peutic recreation (F=6.71, df= 1, 194, p=.01, and F=16.81, df=1, 194, p<.001, respectively). Among patients who participated in therapeutic recreation, only patients with mood disorders had a significant mean decrease in anxiety. Among pa- tients who participated in animal-as- sisted therapy, patients with mood disorders, psychotic disorders, and other disorders had a significant mean decrease in anxiety. This find- ing suggests that animal-assisted ther- apy reduces anxiety for a wider range of patients than the comparison con- dition of therapeutic recreation. Discussion and conclusions Spielberger (30) provided normative State-Trait Anxiety Inventory scores for neuropsychiatric patients based on data from male veterans. Com- pared with the normative patients with depressive reaction, the patients with mood disorders in the study re- ported here had somewhat lower mean pretreatment scores (47.58± 12.73, compared with 54.43±13.02). The pretreatment scores of the pa- tients with psychotic disorders in this study were slightly higher than the scores for the normative patients with schizophrenia (48.47±15.26, com- pared with 45.70±13.44). In this study, no significant differ- ence was found between the anxiety change scores after patients partici- pated in animal-assisted therapy and after patients participated in thera- peutic recreation. However, this lack of difference could be due to the small number of patients (N=50) who completed all four study measures. A power analysis of the magnitude of differences between the change scores for animal-assisted therapy and therapeutic recreation indicated that larger samples would be needed to achieve an 80 percent power level at an alpha of .05: a sample of 300 pa- tients with psychotic disorders, 125 patients with substance use disorders, and 61 patients with other disorders. For patients with mood disorders, the difference in anxiety change scores was too small for any reasonably sized study to detect a significant differ- ence. For within-group differences, a sig- nificant reduction in anxiety after therapeutic recreation was found on- ly for patients with mood disorders, whereas a significant reduction after animal-assisted therapy was found for patients with mood disorders, psy- chotic disorders, and other disorders. The size of these reductions was sim- ilar to differences reported by Wilson (13) for college students whose anxi- ety scores were measured under vary- ing levels of stress. No significant reduction was found in anxiety scores for patients with substance use disorders after either animal-assisted therapy or therapeu- tic recreation. This lack of difference may be due to the small sample size or due to a relationship between state anxiety and physiological withdrawal that is less amenable to change within one session of animal-assisted therapy or therapeutic recreation. The reduction in anxiety scores for patients with psychotic disorders was twice as great after animal-assisted therapy as after therapeutic recre- ation. This finding suggests that ani- mal-assisted therapy may offer pa- tients with psychotic disorders an in- teraction that involves fewer de- mands compared with traditional therapies. As Arnold (23) contends, perhaps the therapy dog provides some sense of safety and comfort not found in more traditional inpatient therapies. Alternatively, the dog may provide a nonthreatening diversion from anxiety-producing situations (31). Or perhaps it is the physical touching of the dog that reduces pa- tients’ anxiety, as has been reported for other populations (12). In this study setting, animal-assist- ed therapy was offered only one day each week. It would be interesting to study the effect of more frequent ex- posure to determine if the reduced anxiety is partly due to novelty or if increased exposure results in further anxiety reductions. Although some patients in the study remained hospi- talized long enough to participate in more than one animal-assisted thera- py session, there were not enough such patients to permit investigation of the effect of repeated exposure. Therefore, data from their initial ani- mal-assisted therapy and therapeutic recreation sessions were used for analyses. It is not possible to determine how much the dog or the owner con- tributed independently to the reduc- tions in anxiety found in this study. Al- though the study’s purpose was to ex- amine the effect of animal-assisted therapy, further examination of the effect of its components is needed. Because many owners of therapy dogs volunteer their time to come to psychiatric units, animal-assisted therapy appears to be a cost-effective intervention. However, volunteers may not participate consistently. In this The reduction in anxiety scores for patients with psychotic disorders was twice as great after animal-assisted therapy as after therapeutic recreation. PSYCHIATRIC SERVICES ♦ June 1998 Vol. 49 No. 6 880011 study, a second therapy dog and own- er, a potential confounding variable, were introduced after the first owner became ill. Use of nonvolunteers could strengthen future studies by providing more consistent treatment conditions. Finally, although the results pro- vide evidence of the immediate effect on state anxiety of a single session of animal-assisted therapy, further study is needed to determine if patients’ overall level of anxiety is affected. Further studies of the effect of ani- mal-assisted therapy on psychiatry services are needed to replicate the findings from this study and to ad- vance our understanding of the ther- apeutic benefits of the human-animal interaction. ♦ Acknowledgments The authors thank Al Best, Ph.D., for his assistance with statistical analysis and Pat Conley, Helen Brown, and Claudette Mc- Daniel for their assistance with data col- lection. References 1. Akiyama A, Holtzman JM, Britz VVE: Pet ownership and health status during be- reavement. Omega 17:187–193, 1986 2. Rowan AN: Do companion animals provide a health benefit? (edtl). Anthrozoös 4:212, 1991 3. Serpell J: Beneficial effects of pet owner- ship on some aspects of human health and behavior. Journal of the Royal Society of Medicine 84:717–720, 1991 4. Siegel JM: Stressful life events and use of physician services among the elderly: the moderating role of pet ownership. Journal of Personality and Social Psychology 58: 1081–1086, 1990 5. Friedman E, Katcher AH, Lynch JJ, et al: Animal companions and one-year survival of patients after discharge from a coronary care unit. Public Health Reports 95:307– 312, 1980 6. Anderson WP, Reid CM, Jennings GL: Pet ownership and risk factors for cardiovascu- lar disease. Medical Journal of Australia 157: 298–301, 1992 7. Davis JH: Animal-facilitated therapy in stress mediation. Holistic Nursing Practice 2:75–83, 1988 8. Patronek GJ, Glickman LT: Pet ownership protects against the risks and consequences of coronary heart disease. Medical Hy- potheses 40:245–249, 1993 9. Baun M, Bergstrom N, Langston N, et al: Physiological effects of petting dogs: influ- ence of attachment, in The Pet Connection. Edited by Anderson RK. Minneapolis, Uni- versity of Minnesota Press, 1984 10. Friedman E, Katcher AH, Thomas SA, et al: Social interaction and blood pressure: influence of companion animals. Journal of Nervous and Mental Disease 171:461–465, 1983 11. Katcher AH, Friedman E, Beck AM, et al: Looking, talking, and blood pressure: the physiological consequences of interaction with the living environment, in Our Lives With Companion Animals. Edited by Katcher AH, Beck AM. Philadelphia, Uni- versity of Pennsylvania Press, 1983 12. Vormbrock JK, Grossberg JM: Cardiovas- cular effects of human–pet dog interac- tions. Journal of Behavioral Medicine 11: 509–517, 1988 13. Wilson CC: The pet as an anxiolytic inter- vention. Journal of Nervous and Mental Disease 179:482–489, 1991 14. Nagengast SL, Baun MM, Leibowitz MJ, et al: The effects of the presence of a com- panion animal on physiological and behav- ioral distress in children during a physical examination. Abstracts of the Delta Society Twelfth Annual Conference. Renton, Wash, Delta Society, 1993 15. Barba BE: The positive influence of ani- mals: animal-assisted therapy in acute care. Clinical Nurse Specialist 9:199–202, 1995 16. Voelker R: Puppy love can be therapeutic, too. JAMA 274:1897–1899, 1995 17. Beck A: The therapeutic uses of animals. Veterinary Clinics of North America, Small Animal Practice 15:2, 1985 18. Beck A, Katcher A: A new look at animal- assisted therapy. Journal of the American Veterinary Medical Association 184:414– 421, 1984 19. Draper RJ, Gerber GJ, Layng EM: Defin- ing the role of pet animals in psychothera- py. Psychiatric Journal of the University of Ottawa 15:169–172, 1990 20. Searles H: The Non-Human Environment. New York, International Universities Press, 1960 21. Levinson BM: The dog as co-therapist. Mental Hygiene 46:59–65, 1962 22. Francis G, Turner J, Johnson S: Domestic animal visitation as therapy with adult home residents. International Journal of Nursing Studies 22:201–206, 1985 23. Arnold JC: Therapy dogs and the dissocia- tive patient: preliminary observations. Dis- sociation 8:247–252, 1995 24. Barker SB, Barker RT, Dawson KS, et al: The use of the family life space diagram in establishing interconnectedness: a prelimi- nary study of sexual abuse survivors, their significant others, and pets. Individual Psy- chology, in press 25. Mallon GP: Utilization of animals as thera- peutic adjuncts with children and youth: a review of the literature. Child and Youth Care Forum 21:53–67, 1992 26. Holcomb R, Meacham M: Effectiveness of an animal-assisted therapy program in an inpatient psychiatric unit. Anthrozoös 2:259–264, 1989 27. Beck A, Seraydarian L, Hunter G: The use of animals in the rehabilitation of psychi- atric inpatients. Psychological Reports 58: 63–66, 1986 28. Walsh PG, Mertin PG, Verlander DF, et al: The effects of a “pets as therapy” dog on persons with dementia in a psychiatric ward. Australian Occupational Therapy Journal 42:161–166, 1995 29. Batson K, McCabe BW, Baun MM, et al: The effect of a therapy dog on socialization and physiologic indicators of stress in per- sons diagnosed with Alzheimer’s disease, in Animals, Health, and Quality of Life: Ab- stract Book. Paris, France, AFIRAC, 1995 30. Spielberger CD: State-Trait Anxiety Inven- tory Manual. Palo Alto, Calif, Mind Gar- den, 1977 31. Arkow P: How to Start a Pet Therapy Pro- gram. Alameda, Calif, Latham Foundation, 1982 FFiirrsstt--PPeerrssoonn AAccccoouunnttss IInnvviitteedd ffoorr CCoolluummnn Patients, former patients, family members, and mental health professionals are invited to submit first-person accounts of experiences with mental illness and treat- ment for the Personal Accounts column of Psychiatric Services. Maximum length is 1,600 words. The column appears every other month. Material to be considered for publication should be sent to the column editor, Jeffrey L. Geller, M.D., M.P.H., at the Department of Psychiatry, University of Massachusetts Medical School, 55 Lake Avenue North, Worcester, Massachusetts 01655. Authors may publish under a pseudonym if they wish. PSYCHIATRIC SERVICES ♦ June 1998 Vol. 49 No. 6880022