Substance use disorder treatment retention and completion: a prospective study of horse-assisted therapy (HAT) for young adults

Patients

The study sample comprised inpatients and day patients admitted between January 1, 2011 and June 30, 2012 to the Department of Addiction Treatment—Youth at Oslo University Hospital. The department treats men and women aged 16–26 years (but patients up to 35 years of age may be accepted) who have a primary diagnosis of mental and behavioral disorders due to psychoactive substance use (ICD 10). One hundred eleven patients entered treatment during the 18-month period. Three patients were discharged to other institutions for ongoing treatment, leaving 108 patients in the study.

Study design

At entry to treatment, all participating patients who had provided written informed consent were registered in the YATEP database. Recording usually started in the first week. The database comprises basic patient information, psychological tests, discharge status, and HAT participation data. Additional patient demographic, morbidity and treatment information, plus the dates of any temporary exit from treatment, can be drawn from the hospital’s electronic patient journal when required and matched anonymously to the YATEP database records. All individuals were followed from treatment entry to discharge.

Measures

The study outcomes were: (1) completion of treatment (primary outcome), (2) time in treatment (measured using number of treatment days), and (3) completion of 90 days of treatment or more (included because it is often identified as a critical period for effective treatment [2, 3, 11–13]).

Discharge status was categorized as (1) treatment completed, or (2) dropout. Treatment completed was defined as staying in treatment for the duration of the recommended treatment plan. This was determined by examining the YATEP discharge report and the clinician’s journal record. Those who left the program but returned within a 30 day period to continue with the remainder of the treatment course were considered to have completed treatment. Leaving the program but returning within a 30 day period was termed “temporary exit.” The number of days of temporary exit was excluded from the total treatment days at discharge. Dropout was defined as patient initiated treatment termination, or expulsion for rule violation prior to completing the agreed treatment period.

Psychological distress was measured using the Hopkins Symptom Check List 25 (HSCL-25), which is one of the assessment items in the YATEP. It consists of 25 questions that map respondents’ anxiety and depression [34]. It is scored on a scale from 1 (not bothered) to 4 (extremely bothered). The form is frequently used in Norwegian research projects with 1.75 as the risk cutoff in normal Norwegian populations [35]. In the analysis for this study, HSCL-25 was used as an indicator of psychological distress.

Severity of substances used was categorized as more severe [heroin, amphetamine, benzodiazepine, gamma hydroxybutyrate (or GHB) and cocaine] or less severe (cannabis and alcohol).

Treatment as usual (TAU)

The treatment site is part of the specialist health care system in Norway. Patients are referred by general practitioners and specialists or from other hospital departments. They must have a primary diagnosis of mental and behavioral disorders due to psychoactive substance use (ICD 10). The social services authority has oversight of this process. The treatment is a person-centered program which comprises individual and group therapy based on a biopsychosocial model with emphasis on mentalization-based theory and practice [36]. An individual treatment plan, which includes treatment goals, is prepared in cooperation with each patient. Medical treatment is offered, as well as assistance/counselling for accommodation, education, employment, living, adjustment and support. Psychological treatment is tailored to the individual’s specific problems and treatment goals. The likely duration of treatment is decided with the patient as part of the treatment plan, in accordance with their needs. It can include movement between units, such as from inpatient to day patient. In the day unit, as patients become more established in school, work or a domestic situation, the therapist gradually reduces contact until discharge.

The HAT intervention

HAT is an integral part of the department’s program of addiction treatment [37]. It comprises 12× 90-min sessions of body-orientated psychotherapy with horses. The animals have been selected and trained for this work to be strong, secure, responsive and interactive. Patients and staff are insured against injury by the hospital. Serious incidents and injuries must be recorded.

All patients are eligible to participate in HAT, but must be referred by their treating clinician. The referral can be requested by the patient or suggested by the clinician. A final decision on suitability and the treatment objectives of the individual’s HAT participation (for example, to strengthen boundary setting, or reduce anxiety, depression or aggression, etc.) are agreed at a preparatory meeting between the HAT therapist, the patient and the clinician. Patients have the opportunity to meet the horses and become involved in care activities (such as feeding) from their first day in treatment. They normally start the HAT program within 2–3 weeks. HAT therapists become part of the patient’s clinical team, with full access to the patient’s clinical record. Patients are encouraged to attend and participate fully but can choose not to undertake an activity, such as mounted work. Specific activities, level of participation and response at each HAT session are recorded by the HAT therapists in the patient’s electronic hospital journal.

The sessions are planned and provided by two qualified therapists who are also Norwegian Level 1 Riding Instructors. The program design is structured for small groups (maximum four participants per session), but includes provision for individual work on specific needs if required. It involves a three-way interactive (positive triangulated) process in which the patient works in emotional safety with the horse on activities selected with his/her therapist to address agreed goals. During sessions, the horse will respond naturally to environmental factors (for example, the proximity of other horses, or a sudden loud noise). Similarly, it will react to the physical and emotional state of the patient (for example, a request lacking focus or clarity is unlikely to produce the desired movement from the horse, and an aggressive request may be met with resistance). The therapist, in leading the process, can both read and influence the horse, and provide reflective feedback to the patient on the relationship, reactions and responses between the horse and patient.

Activities can involve any combination of herd behavior observation, stable duties, and ground, mounted and/or driving work with the horses. Observation of the herd can promote discussion of social interaction and relationships and stable duties promote responsibility, routine and reliability. Groundwork is used to address issues relating to boundaries/contact, anxiety/trust, communication/connection, mastery (of new skills, the horse and self), body awareness and focus. Mounted work addresses posture, balance/centering, coordination, rhythm/regulation, mastering of anxiety and focus. Carriage driving can be used to promote forward thinking and outlook, and, with other passengers, it can engender a sense of empowerment, group responsibility and care. These activities involve good healthy exercise, having fun, and learning new skills. However, while physical exercise, fun and skill acquisition are important, the prime purpose of this program is therapy and contribution to successful treatment.

The focus of the first four sessions is on getting to know about horses, herd behavior, basic handling and safety. The following eight sessions are tailored to meet the individual’s therapy objectives using a range of group and individual ground-based, mounted or driving exercises as outlined in the stable manual (unpublished).

The HAT program has been developed at Oslo University Hospital over time, largely by Lysell [22], a qualified and experienced body-oriented psychotherapist. It also draws on theoretical and practice material from a number of relevant equine-assisted therapy schools [16, 20, 38–42]. It uses many of the usual equine-assisted/facilitated therapy exercises but places stronger emphasis on those relevant to substance use disorders, such as boundary setting, development of trust and control of emotional affect. It differs from most other horse therapy programs in two aspects. First, the patients have responsibility for the horses after hours, giving greater emphasis on care, routine, reliability and responsibility (all relevant to substance use recovery). Second, after the four introductory sessions, the HAT program does not follow a sequenced routine. Rather, specific activities and therapeutic processing are targeted at individual patient needs and are sequenced at appropriate points throughout the patient’s HAT program.

HAT treatment outcome is not assessed per se. It is included as part of the patient’s overall treatment outcome assessment, as measured by change in the YATEP psychological instruments and, in particular, by whether individuals complete their agreed substance use treatment program.

Statistical analysis

In this naturalistic, intention-to-treat study, univariate and multivariate analyses were used to assess the relationship between treatment completion and HAT plus a range of patient factors (gender, age, education, number and severity of substances used, psychological distress and number of temporary exits). We included time in treatment as an additional outcome measure. It is assessed using univariate analysis of both the mean (113 days) and the reported critical minimum period for effective treatment (90 days). However, we excluded time in treatment from the logistic regression because of the obvious relationship between longer time in treatment and treatment completion.

Pearson Chi squared and independent-samples t test were used to test the relationship between discharge status and HAT. Odds ratio (OR) was used to test the strength of the relationship. Potential confounding variables relating to the patient (age, sex, education, number and severity of substances used, and psychological distress), time in treatment (mean time and temporary exits) and HAT participation were controlled for using logistic regression analysis. Linear and other interactive associations were checked. SPSS Version 21 (IBM Corp., Armonk, NY, USA) was used.

Study participants’ characteristics

In the 6 months prior to intake, 32.4 % of the 108 patients had used a single drug, 28.7 % had used two drugs and 39.9 % had used three or more drugs. Cannabis was the most commonly used primary drug (38.9 %), followed by alcohol (18.5 %), amphetamine + cocaine (15.8 %), heroin (15.7 %) and other drugs (11.1 %). The average age of first use of the primary drug was 17 years, (results not shown) with 52.8 % of all patients being 15 years or younger when they first used their primary drug (Table 1).

At referral, all patients had a primary diagnosis of mental and behavioral disorders due to psychoactive substance use (ICD 10). Fifty-eight (53.7 %) of the patients were reported to have no psychiatric comorbid condition, 34 (31.5 %) had one, and 16 (14.8 %) had two or more. The most common of the comorbid conditions (results not shown) were behavioral disorders (36.2 %), followed by neurotic/stress (including post-traumatic stress disorder) disorders (34.8 %), mood disorders (20.3 %) and other disorders (8.7 %). Approximately 1 week after entry, 89 (82.4 %) patients’ first Hopkins Symptom Check List 25 (HSCL-25) scores were equal to or above the psychosocial risk cutoff. Sixteen patients (14.8 %) were prescribed substitution medicine (15 buprenorphine, 1 methadone) during all or part of their treatment (Table 1).

HAT participation

Forty-three patients (39.8 %) received treatment as usual (non-HAT group) and 65 (60.2 %) had treatment as usual plus HAT (HAT intervention group). The intervention was well tolerated with no reported adverse treatment effect or injury arising from HAT. None of the patients were withdrawn from HAT for clinical reasons.

Although this was a naturalistic study, there were no major differences in baseline characteristics between the HAT group and the non-HAT group. However, significantly more HAT participants were aged 27 years or more (24.6 vs 9.3 %, p = 0.03) and their average age was slightly older than non-HAT participants (mean 23.7 years, SD = 3.4 vs mean 22.1 years, SD = 3.2, p < 0.02) (results not shown). They were also more likely than non-HAT participants to have one or more temporary exits from treatment (49.2 vs 25.6 %, p = 0.05) (Table 1).

Thirty-seven (56.9 %) of HAT participants completed their treatment compared with only six (14.0 %) of non-HAT participants, p < 0.001 (Table 1).

Time in treatment

HAT participants remained in treatment for a significantly longer period (p < 0.001) than those who did not participate in HAT (mean 141 days, SD = 93.6, vs mean 70 days, SD = 73.8). They were almost four times more likely to remain in treatment for 90 days or more (OR 3.9 CI 1.7–8.8, p = 0.001) (results not shown).

Treatment discharge status

At discharge, 43 (39.8 %) of the 108 patients in the study had completed treatment and 65 (60.3 %) had dropped out. Treatment completion was significantly associated with only HAT participation [χ² (1,108) = 19.9, p < 0.001] and length of time in treatment {for both mean days in treatment [χ² (1,108) = 30.2, p < 0.001] and the critical 90 days period [χ² (1,108) = 32.8, p < 0.001] (Table 2)}.

Prediction of treatment completion

Strengths and limitations

The size and relative homogeneity of the population studied is a strength when compared with other studies of therapy involving horses. However, there are limitations to both internal and external validity. Nonrandom assignment, including possible self-selection bias, is the most obvious limitation. The possible novelty effect of HAT is another. These shortcomings need to be addressed in large controlled studies that seek to clarify HAT’s role, including causal inference, in substance use disorder therapy, in treatment outcomes and treatment effect.

HAT participation was voluntary. Because there were no data on why patients joined or did not join the HAT program, it was not possible to form an opinion on whether the HAT participants stayed in treatment longer because they participated in the HAT program or because of some other positive factor, which also led them to participate in HAT. Data on patients’ motivation and program satisfaction (such as those used by Decker et al. [43]) might have been useful but were not available.

Dropout, the most frequently used measure of addiction treatment outcome, is fraught with definitional and other problems [3, 4, 6, 10]. There is no consistent definition of the term. Many studies fail to define how they have used it and/or omit discussion of the impact of temporary exits. We did not find any studies that included dropout as a measure of treatment outcome for HAT, and therefore comparisons were not possible.

Our findings are consistent with claims that HAT has a positive effect on psychosocial illness. However, we acknowledge this could be due, in whole or in part, to a variety of possible confounding factors that were not examined (such as, for example the “honeymoon” effect of a new program, the outdoor environment or the ambience around the stables). These factors lend themselves initially to qualitative investigation.

It is possible that a similar effect might be obtained more economically with another animal, such as a dog. Inclusion of dogs in substance misuse therapy has been found to improve therapeutic relationships [56]. However, we have found no studies which compare the use of horses with other animals as an adjunct therapy for substance misuse but note that Nurenberg [32] and colleagues in their comparison did find a significantly positive effect of therapy involving horses, but not dogs, when incorporated into psychotherapy for aggressive behavior.

Relevance

Ongoing care was not addressed in the study. However, HAT may have considerable potential as a community-based activity for outpatient treatment. If available from accredited public and private providers in community settings, it may enable engagement in healthy, pleasurable outdoor activities with a nonsubstance use peer group.

There is a continuous struggle to find what motivates substance use disorder patients to remain healthy. Rigorous evidence of safety and efficacy is required before HAT can become a conventionally accepted treatment with the associated provision of health insurance cover [57–59]. The positive findings from this treatment facility-based study, and HAT’s broader community potential, prompt further investigation of both the underlying therapeutic processes and the longer-term impact of HAT on substance use disorder treatment.