Coercion and the Inpatient Treatment Alliance

The benefits of a strong treatment alliance (also called therapeutic alliance or therapeutic relationship) between patients and their psychiatrists are widely recognized and noncontroversial. Among psychiatric inpatients, a strong treatment alliance is associated with improvement at discharge (1), increased satisfaction with treatment (2), and decreased risk for violence (3). A meta-analysis of inpatient and outpatient samples reported a correlation between scores on alliance scales and positive clinical outcomes (4).

The treatment alliance is potentially harmed when patients experience coercion and are forced to comply with treatment plans they have not chosen and may not want. Three studies in Europe found that inpatients who reported higher levels of perceived coercion gave lower ratings of the treatment alliance (2, 5, 6).

Despite the importance of this topic, empirical research on the relationship between coercion and the treatment alliance specifically among inpatients is limited. Only a few studies exist, their sample sizes are small for quantitative analyses (N=79–164), and they are thinly spread across just a few countries (Australia, Ireland, England, and Switzerland) (5–8). Additionally, quantitative methods that reduce complex experiences to statistical associations reveal little about how coercive experiences affect the treatment alliance. Qualitative research is needed; however, only one such study has addressed coercion and the treatment alliance among inpatients. That study was small (N=19) and recruited some participants by advertising in a mental health newsletter, increasing the risk of selection bias (9).

This study addressed this gap in the literature. Psychiatric inpatients were interviewed during their hospitalization, and themes related to coercion and the treatment alliance were qualitatively analyzed. These data are intended to complement existing quantitative studies, adding deeper understanding of how experiences with coercion may affect the inpatient treatment alliance.

Methods

The study was approved by the institutional review board at the New York State Psychiatric Institute. A checklist of the consolidated criteria for reporting qualitative studies with comments is included in the online supplement (10).

The interview questionnaire was developed by two authors (R.E.L. and J.L.B.) after reviewing relevant literature. Additional authors (L.B.D. and R.B.D.) provided critical review, and feedback was incorporated into the questionnaire. The questionnaire remained consistent throughout all interviews to facilitate comparisons of participants’ responses. Interviewers were advised to ask follow-up questions to encourage participants to elaborate on their experiences and to explore emerging themes.

Participants were enrolled from two psychiatric inpatient units at two different campuses of a large, multisite, academic medical center in New York City. One unit has 24 beds and special staffing for electroconvulsive therapy, the other unit has 30 beds and special staffing for treatment of patients with co-occurring substance use disorders. Both units treat adults (ages 18 and older), are locked, and have an average length of stay of approximately 2 weeks.

From the outset, this study was designed to enroll 50 participants (25 from each unit). This target was selected in an effort to include enough participants so that a range of experiences was represented, while also limiting data to a manageable size so that in-depth analysis of each participant’s experiences remained possible. This sample represents approximately 15%−20% of the inpatients treated on these units during the study period. The current sample is larger than other qualitative studies on this topic among inpatients (N=19) (9), and outpatients (N=6–15) (11, 12).

Participants were recruited during a 3-month period in 2017. The goal was to include the perspectives of all psychiatric inpatients, so inclusion criteria were broad and exclusion criteria were minimal. Persons were eligible to participate if they were psychiatric inpatients, age 18 or older, and English speaking and if their inpatient psychiatrist considered them clinically appropriate for the interview. Clinically appropriate meant that the psychiatrist felt the patient was able to tolerate the interview and able to provide meaningful data and that participation would not interfere with the patient’s treatment or discharge processes. Patients were ineligible if they were unwilling or unable to participate meaningfully in an interview.

Attending psychiatrists identified eligible inpatients under their care and notified the interviewers (R.E.L. or J.L.B.). An interviewer then approached the patient, introduced the study, offered the chance to participate, and, if the patient was willing, arranged a time to conduct the interview. Interviews were conducted on the inpatient units. The interviewer was not a member of the patient’s treatment team.

Written informed consent was obtained at the time of the interview and documented by the interviewer. Interviewers read an introduction to the study and asked the patient questions about his or her demographic characteristics. Participants then completed the Inpatient Treatment Alliance Scale, a 10-item self-report instrument designed for use on inpatient units. Participants assigned a score (0–6) indicating their level of agreement with statements about three major domains of the treatment alliance (bond, collaboration, and goals) (13). Interviewers then asked the participants open-ended questions about the treatment alliance and coercive practices they had experienced during the current hospitalization and prior hospitalizations.

The full range of coercive psychiatric treatment practices is broad and may include interactions with staff, imbalances of power, seclusion and restraint, forced medication, and lack of patient involvement in treatment planning. This study focused on a subset of coercive experiences: court-mandated treatment, involuntary hospitalization, and locked doors. These experiences were chosen because they were easily operationalized, emotionally salient, and prevalent among participants.

Interviewers asked about sign-out letters (whether participants had ever submitted one) as a marker for patient disagreement with the psychiatrist’s treatment plan or timeline (a situation in which the treatment alliance is at risk and a patient may feel pressure to accept a plan he or she does not want). In New York, a sign-out letter is the mechanism by which voluntary or involuntary patients can formally request discharge; the letter triggers either a discharge or a trip to court, where a judge may order discharge or retention for up to 60 days.

Diagnostic information of participants was reported separately by the treating psychiatrist. Researchers did not review patient records. The only unique patient identifier collected was the sound of the patient’s voice. Interviews were audio-recorded and transcribed by R.E.L. or J.L.B.

We used a grounded theory approach to analyze the data, which allowed themes to emerge from the interviews. Midway through the data collection, we (R.E.L. and J.L.B.) reviewed the interviews and developed a preliminary codebook of themes. After 50 interviews, there were no new ideas or concepts emerging, suggesting that theme saturation had been achieved. We (R.E.L. and M.P.-C., both attending psychiatrists on the inpatient units) reviewed the transcripts, created a final version of the codebook, and coded each interview. Two of us (R.E.L. and M.P.-C.) coded the interviews independently and discussed discrepancies until consensus was reached. Interview questions and content codes are available in the online supplement.

We managed the data by using NVivo 12 (version 12.1.1.256, QSR International) and Excel 2013 (Microsoft).

Results

Sample

Sixty patients were invited to participate, and 50 completed the interview. Participants’ demographic and clinical characteristics are shown in Table 1. Most of the participants reported a strong treatment alliance; however, a wide range of experiences was represented (Table 2). Interviewers offered a definition of a treatment alliance as “the doctor and patient working together.” Participants’ descriptions of strong treatment alliances, as well as more difficult treatment relationships, are included in the online supplement.

Court-Mandated Treatment

Five patients had ever been to court for treatment-related reasons. Two said the treatment alliance was harmed because the court experience led to involuntary medication, a state hospitalization, and an adversarial relationship.

Two reported that the treatment alliance with the psychiatrist had remained strong despite going to court, because “he was right there with me” or because the psychiatrist explained the plan in a way the patient appreciated.

One patient said the experience had no impact on the treatment alliance because the psychiatrist and patient already knew what the other wanted. “They knew I wanted to leave and it just wasn’t, it had no impact on [the relationship]. They had a very firm idea: they wanted to keep me.”

Submitting a Sign-out Letter

Thirteen patients had ever submitted a sign-out letter (either during this hospitalization or prior hospitalizations), formally requesting discharge. None of these participants thought submitting a sign-out letter had harmed the treatment alliance.

For some (N=3), the sign-out letter was purely motivated by a desire to leave the hospital, and the inpatient treatment alliance played no role in the decision. One gave no consideration to the inpatient alliance because “I already have [a psychiatrist] on the outside.”

Importantly, sometimes the sign-out letter was more about forcing a conversation, and less about forcing a discharge (N=3). One patient who submitted a sign-out letter and then retracted it said, “The conversation [about discharge] needed to happen and it wasn’t happening, and I needed to feel more in control of what was going on.”

Two patients did not think the sign-out letter had changed the treatment relationship, because the psychiatrist already knew the patient wanted to leave.

While there was not an expectation that psychiatrists would agree to the discharge (“psychiatrists can be very adamant”), two patients said they did not think the psychiatrists would treat them differently because of the sign-out letter (“I think they do have our best interests at heart”).

One patient submitted and then retracted a sign-out letter out of concern it would make it harder for him to work with his psychiatrist.

Involuntary Hospitalizations

Nineteen patients reported having been involuntarily hospitalized at some point in their lives, and one additional patient had been involuntarily held in a psychiatric emergency room. Six patients said the alliance was harmed by being involuntarily hospitalized. Experiences ranged from feeling “a little bit frustrated” to “I lost trust in psychiatry completely” and “It was horrible. I had no idea I would end up here.” Patients said the involuntary experiences were difficult because they felt as though they had no control over decisions, they did not feel like full participants in treatment planning, and they were not able to manage their lives outside the hospital (e.g., not able to make money). Three patients said it was difficult to engage in treatment because they did not want to be there.

Many said the inpatient treatment alliance was unaffected (N=7) or remained strong (N=6) despite the involuntary status. One patient explained that this was because the emergency room doctor who hospitalized him against his will was different from the inpatient psychiatrist who treated him. Another participant said she knew she needed the hospitalization. One suggested that the overall treatment experience was more important than voluntary or involuntary status.

Locked Doors

All patients were interviewed on locked inpatient units. Four thought the locked doors helped the alliance by reducing distractions and helping them to focus on their treatment. Most (N=37) said the locked doors had no effect on the treatment alliance, often because it was what they expected.

Six reported the locked doors negatively affected the alliance.

Relationship Between Hospital Experiences and Future Adherence

Interviewers asked participants whether there were any experiences during this hospitalization, good or bad, that affected their willingness to adhere to their outpatient treatment plan. Among the 40 codable responses, a majority (N=27) said their inpatient experiences had no effect on their willingness to adhere to the outpatient plan, sometimes making a sharp distinction between inpatient and outpatient experiences.

Eight participants reported positive experiences that increased their willingness to be adherent, including working together with the psychiatrist to create the outpatient plan and having helpful therapeutic experiences that provided a model for what aftercare could be like.

Five participants reported negative experiences that decreased their willingness to be adherent. These included feeling intimidated by hospital staff, believing medications or dosages were inappropriately prescribed, witnessing other patients have medication side effects, having difficulty relating to other patients and their problems, and feeling they had served their time (“They can’t force me to go to no treatment plan. Like, this [expletive] don’t make no sense. I’ve been here enough time. This is enough torture for me.”).

Discussion

In 50 qualitative interviews on two inpatient psychiatry units, participants reported a wide range of experiences with specific coercive practices and their effects on the inpatient treatment alliance. Some participants described an imbalance of power, a lack of control, and insufficient participation in treatment planning as damaging to the inpatient alliance. Other participants, who felt the treatment alliance was maintained despite coercive experiences, spoke of good communication with the doctor, understanding the rationale behind various interventions, and feeling the doctor was trying to keep the patients’ best interests in mind. These interviews provide a glimpse of what some inpatients experience.

These results do not call into question previous findings that coercive experiences tend to weaken the treatment alliance (5–7, 9). Outpatient studies have consistently found that fear of coercion is an obstacle to seeking psychiatric care (14, 15). In a large qualitative study, outpatients described choice and autonomy as central to recovery and coercion as a major hindrance (16). Similarly, a quantitative study among outpatients reported that higher treatment satisfaction was associated with a better working alliance and less perceived coercion (17). Our results add to this literature, however, by showing that, for some patients, coercion does not necessarily spell the end of a treatment alliance.

This study expands possible explanations for the apparent paradox in the literature that although coercion is a barrier to engagement in care (5–7, 9, 14, 15), measures of coercion among inpatients have not been associated with poor engagement at follow-up (18, 19). Prima facie explanations include the possibility that patients who find coercion adversely affecting the alliance avoid further contact and do not provide follow-up data as well as the reality that patient experiences are complex, individual, and constantly evolving in ways that cannot be easily reduced to quantitative scales. In this study, we found that some patients and clinicians in the sample were able to establish a strong alliance despite the patient’s experience of coercion in the hospital. Additionally, some patients experienced sharp distinctions between engagement with an emergency room psychiatrist, an inpatient psychiatrist, and an outpatient team, such that an alliance (or conflict) that happened in one setting did not translate to other settings. Patients were able to have nuanced and individualized views of different clinicians.

Use of the Inpatient Treatment Alliance Scale is a methodological strength of the study design (13). Future studies may benefit from using mixed methods to stratify participants according to their treatment alliance scores and then asking more specifically targeted questions.

Limitations of the study included recruitment only in one city and within one hospital system. Not all participants had experienced all the forms of coercion assessed in the interview. The study did not assess length of time since illness onset. Data were not available on study nonparticipants, so it is not known whether their experiences differed from study participants. Many coercive experiences were not specifically assessed in this study, including forced medication, seclusion, and restraint, which are important categories of experiences that can affect the treatment alliance. The interview questions tended to stratify treatment relationships into two categories (good alliance or difficult relationship), and future studies could take a more nuanced approach by exploring varieties of treatment relationships. It would be especially important to understand the views of individuals whose experience of coercion was so aversive that they avoided treatment entirely.

Conclusions

In this qualitative study of 50 inpatients on two psychiatric units, participants did not report a straightforward relationship between selected coercive experiences and a disrupted treatment alliance. While some participants felt that coercive experiences made it difficult or impossible to work with the inpatient psychiatrist, others experienced the inpatient psychiatrist as well intentioned, even when coercive measures had been used. While coercion is never desirable in clinical settings, patients, clinicians, and policy makers should consider that, for some patients, positive treatment experiences and a good treatment alliance remain possible even when coercive measures have been used. In addition to minimizing the use of coercive procedures, clinicians should focus on how all aspects of treatment can promote collaboration and partnership.