Atlas›Domain I · Psychodynamic & depth psychology›Guide 083
Guide I-083

Sullivan's Sheppard Pratt ward

Harry Stack Sullivan1929Sheppard and Enoch Pratt Hospital
A 48 · L 40 · M 1214 interventions in the Trinity explorer1 described for recognition only
Client cardCompare
[source: …] where a claim comes from[uncertain] not confirmed in a primary source[synthesis] the writer's reconstruction

1. Introduction

Version followed: Harry Stack Sullivan's special reception ward at Sheppard and Enoch Pratt Hospital (organised 1929, run until he left in 1930), as described in his 1931 papers ("Socio-psychiatric research: its implications for the schizophrenia problem and for mental hygiene"; "The modified psychoanalytic treatment of schizophrenia"; both collected in Schizophrenia as a Human Process, 1962, introduced and annotated by Helen Swick Perry) and in Perry's biography Psychiatrist of America (1982). Sullivan's later interviewing concepts (1953, 1954) are used only where marked. Everything about running a present-day unit is reconstruction and is marked [synthesis].

What it is. A small admission ward (six beds, per Perry) for young men in an acute, early schizophrenic illness. It ran almost independently of the rest of the hospital and broke with its usual routines. Sullivan picked and trained the all-male staff himself, preferring people who had been through psychological difficulties like the patients'. Women were not allowed on the ward. He held long informal interviews with patients and long informal talks with staff, often in the evenings at his home [source: Encyclopedia.com entry (Conci); Blechner 2005, citing Chatelaine 1981; Perry 1982 for the bed count] [uncertain: the exact bed count and the use of "attendants" rather than "staff" come from Perry].

Plainly: this is not a manualised therapy. No manual, session protocol or fidelity measure exists. It is a historical ward model built on a theory, Sullivan's view of schizophrenia as a human, interpersonal process. Sullivan never analysed the ward's results systematically [source: Encyclopedia.com/Conci].

How change happens (Sullivan's idea). Schizophrenic states are human reactions to catastrophic failures of interpersonal living: panic, collapse of self-esteem, and overwhelming shame, very often around sexuality. Even the most disordered behaviour is still meaningful symbolic activity, so the person can be understood with enough contact. People recover when they are taken into a setting that accepts them: anxiety and shame come down, and everyday relationships with peers, not only the psychiatrist's interview, rebuild a workable self-regard. Psychotherapeutic inquiry into what went wrong comes after that [source: Encyclopedia.com/Conci; Blechner 2005; Sullivan 1931a, 1931b as collected 1962] [uncertain: wording reconstructed from secondary summaries, not from the 1931 texts directly].

2. Framework

Sullivan published no phase protocol, no session counts and no criteria for moving between stages. The stages below reconstruct the order his writings and his biographers emphasise: the milieu comes first, admission is decisive, peer acceptance comes before inquiry, and attention to the world the patient returns to comes last [synthesis]. The durations are suggestions for modern inpatient units (stays usually 1–4 weeks) and residential units (usually 2–6 months). They are not historical figures [synthesis] [uncertain: how long patients stayed on Sullivan's ward is not established].

Phase 0: Build the milieu (before any admission).

  • Goal: a small, protected unit with staff chosen and prepared for a non-shaming, peer-like stance. Sullivan selected and trained staff before and throughout [source: Encyclopedia.com/Conci].
  • Format: unit of about 6–12 beds, or a protected sub-unit of a larger ward. Keep the same core staff on most shifts. Write down which hospital routines the unit will modify, and get managers to agree [synthesis].
  • Move on when: every core staff member has had the induction (4.1–4.3), a daily staff-talk slot exists, and an admission script (4.4) is agreed [synthesis].

Phase 1: Reception (first hours to about 2 weeks).

  • Goal: get through the admission without adding humiliation or panic, and help the patient feel received by people rather than processed by an institution. Sullivan made this an admission ward because he thought the early contacts mattered most [source: Encyclopedia.com/Conci ("special admission ward")].
  • Format: one named staff member stays close for the first shift. Use as few interviewers and procedures as is safe. The psychiatrist makes brief, frequent, informal contacts, daily if possible. Staff record their observations each shift [synthesis].
  • Move on when: panic has visibly eased (sleep, eating, the patient tolerates others nearby), the patient accepts at least one staff member's company, and acute risk is managed under standard assessment [synthesis].

Phase 2: Peer acceptance and socialisation (roughly weeks 1–6, or longer in residential care).

  • Goal: the patient joins the ward's everyday social life, shares experience with peers and staff, and recovers some self-regard. Sullivan relied on these indirect, non-interview interactions as treatment in their own right [source: Encyclopedia.com/Conci].
  • Format: unstructured time with staff and patients (meals, games, walks, chores, talk). Longer individual conversations with the psychiatrist, 3–5 times a week. Daily staff discussion with the psychiatrist [synthesis, extrapolated from Sullivan's reported practice].
  • Move on when: the patient starts conversations, can talk about at least some of what happened without being overwhelmed by panic, and shows interest in other people [synthesis].

Phase 3: Inquiry (the "modified psychoanalytic" phase).

  • Goal: an individual, collaborative review of the relationships, events and shame-laden experiences that led into the illness, done without lowering self-esteem. Sullivan modified psychoanalytic method for schizophrenia by putting the patient's security first and inquiry after [source: Sullivan 1931b title and 1932 epitome; content from secondary summaries] [uncertain: details of the 1931b procedure].
  • Format: 2–3 individual sessions a week, roughly 45–60 minutes. The milieu continues alongside [synthesis].
  • Move on when: the patient can give a coherent account of the lead-up to the illness that includes their own part and other people's, and can name early signs of renewed panic [synthesis].

Phase 4: Re-entry.

  • Goal: prepare the return to the interpersonal world (family, work, peers) that the illness came out of, and line up continuing relationships [synthesis, consistent with Sullivan's environmental emphasis] [uncertain: Sullivan's 1931 paper on environmental factors].
  • Format: family or significant-other meetings, graded leave, an outpatient therapist and peer contacts arranged before discharge [synthesis].
  • Move on (discharge) when: there is a named person outside, a relapse-signature plan exists, and the patient has handled at least one graded leave [synthesis].

Running throughout: psychiatrist–staff conversation as the engine of the unit (4.2) and detailed shift records (4.12) [source: Encyclopedia.com/Conci for both practices].

3. Guiding principles

  • People are human first. In the later formulation: "much more simply human than otherwise" [source: Sullivan 1953, The Interpersonal Theory of Psychiatry; page not checked] [uncertain: this phrasing comes after the ward period]. Practical meaning: no staff behaviour, joke or label treats patients as a different kind of being.
  • Psychotic behaviour is meaningful. Sullivan concluded from detailed records of several hundred patients that even the most disturbed behaviour is symbolic activity within the human realm [source: Encyclopedia.com/Conci]. Practical meaning: staff ask "what is this about for him?" before "how do we stop it?", except where safety demands.
  • Anxiety and lost self-esteem drive the illness; the milieu must lower them. [source: Encyclopedia.com/Conci; Blechner 2005] Practical meaning: test every procedure, rule and remark against "does this add shame or fear?"
  • First contacts are decisive. [source: Encyclopedia.com/Conci] Practical meaning: put the best staff and the most care into admission day.
  • The ward is the treatment. Human interactions outside the psychiatrist's interview are therapeutic [source: Encyclopedia.com/Conci]. Practical meaning: staff conversations are treatment, not supervision duty.
  • Peers heal. Patients with similar backgrounds benefit from sharing their experiences [source: Encyclopedia.com/Conci]. Secondary accounts frame staff as the peer relationships patients had missed earlier in life [source: Wikipedia mirror, unsourced claim] [uncertain]. Practical meaning: staff relate as fellow people, not custodians, and the unit encourages patients to talk with each other.
  • Choose staff for who they are. Sullivan preferred staff who had lived through difficulties like the patients' [source: Encyclopedia.com/Conci]. Practical meaning: hire for warmth, low reactivity to bizarre or sexual material, and comfort with one's own history. Today, include peer-support workers with lived experience [synthesis].
  • Regulate shame, especially sexual shame. [source: Blechner 2005, citing Chatelaine 1981] Practical meaning: sexual fears, wishes and identity can be spoken about without shock, ridicule or moralising.
  • Protect the unit from institutional routine. [source: Encyclopedia.com/Conci] Practical meaning: negotiate exemptions from routines that degrade people (mass processing, unexplained searches, rotating strangers) where safety allows.
  • The psychiatrist is a participant observer and runs the human environment. [source: Sullivan's general method, later formalised 1954; his practice of informal talks with staff per Encyclopedia.com/Conci] Practical meaning: the psychiatrist's main job is shaping the staff's attitudes as much as individual sessions.
  • Keep a research attitude. Sullivan was director of clinical research and worked from detailed records [source: Encyclopedia.com/Conci]. Practical meaning: record observations precisely and review them.
  • Watch the course. Sullivan's papers distinguished recoverable courses from ones that harden into chronic paranoid or withdrawn adjustments [source: Sullivan 1928, "Tentative criteria of malignancy in schizophrenia," in the 1962 collection] [uncertain]. Practical meaning: track whether the patient is reconnecting or settling into fixed grievance or withdrawal, and escalate if it is the latter.
  • Modern limits [synthesis]. Admit all genders and do not exclude staff by gender or select them by sexual orientation. Respect boundaries and patient rights. Work alongside standard psychiatric care (medication by the prescriber, risk assessment, physical health). Sullivan's design choices that conflict with these are described historically in 4.14.

4. Intervention inventory

The sources record the practices listed below (staff selection, staff talks, a protected admission ward, peer sharing, non-shaming talk about sexuality, long informal interviews, detailed records, and later inquiry). They do not record step-by-step procedures. Every step sequence and all example wording is therefore [synthesis], and example language is illustrative throughout. Items 4.1–4.13 are for modern use. Item 4.14 is historical only.

4.1 Selecting unit staff for attitude

  • Purpose / when: In Phase 0, and with every new hire. Sullivan chose staff carefully and preferred people who had been through psychological difficulties like the patients' [source: Encyclopedia.com/Conci].
  • Steps [synthesis]:
    1. Write a short profile: warm, curious, slow to react, not easily shocked by bizarre, sexual or hostile material, able to be with someone without fixing them, comfortable with their own history.
    2. At interview, describe a scenario: a frightened new patient says staff are poisoning him, or reveals a sexual fear. Ask what the candidate would feel, say and do.
    3. Screen out people who ridicule, moralise, or speak about patients in us/them terms ("the crazies," "manipulative").
    4. Include peer-support workers with lived experience of psychosis or psychiatric admission, the closest modern counterpart to Sullivan's preference [synthesis].
    5. Select on attitude, not on gender or sexual orientation (see 4.14).
  • Example language (illustrative): "Tell me about a time someone said something that shocked you. What did you do with your face in the first few seconds?"
  • Watch for: Candidates whose answers stress control and compliance over contact. It is working when patients name particular staff as people they trust within the first week.

4.2 Psychiatrist–staff conversations (the unit's engine)

  • Purpose / when: Daily, all phases. Sullivan talked with staff often, informally and at length, often in the evenings at his home, and treated those talks as central to the ward [source: Encyclopedia.com/Conci].
  • Steps [synthesis]:
    1. Hold a 30–45 minute informal staff talk each day, and protect the time. A weekly longer reflective meeting can replace home visits; do not hold meetings at home today (boundaries, confidentiality).
    2. Ask each staff member for one concrete exchange with a patient: what was said, what the patient did next, and what the staff member felt.
    3. With the group, work out what the patient may have been communicating and what anxiety or shame was in play, on both sides.
    4. Agree one adjustment for the next shift, such as "sit with him at meals without asking questions."
    5. Notice staff anxiety or contempt and talk about it without blame. Staff attitudes are treated as the treatment variable.
  • Example language (illustrative): "When he told you to get out, what happened in you? … And what do you think he was protecting himself from right then?"
  • Watch for: Talks drifting into rule enforcement or gossip. It is working when staff begin to predict patients' reactions and change their own approach without being told.

4.3 Staff induction in the non-shaming stance

  • Purpose / when: Phase 0, then refreshed. Sullivan trained his staff himself [source: Encyclopedia.com/Conci]. The content below is reconstructed from his principles [synthesis].
  • Steps [synthesis]:
    1. Teach the core ideas in plain language: psychosis as frightened, meaningful human behaviour; panic and shame as drivers; first contacts matter; ordinary companionship as treatment.
    2. Role-play the first 30 minutes of an admission, a patient disclosing a sexual fear, and a patient hearing voices that insult him. Debrief what added or reduced shame.
    3. Agree the unit's "never" list: no mocking, no talking about patients in their hearing as if they were absent, no threats framed as treatment, no pathologising sexuality.
    4. Set disclosure boundaries: staff may share ordinary human experience briefly to normalise. They never share sexual details or seek closeness for their own needs (see 4.8, 4.14).
  • Example language (illustrative, trainer): "Your job tonight is not to get him to make sense. It's to be someone who doesn't make him feel crazier."
  • Watch for: Staff who agree in training but revert under stress. Pick this up in 4.2.

4.4 Reception of the newly admitted patient

  • Purpose / when: First hours and days (Phase 1). Sullivan set up the ward as an admission service because he put the greatest weight on early contacts [source: Encyclopedia.com/Conci].
  • Steps [synthesis]:
    1. Assign one named staff member to greet the patient and stay with them through the first shift.
    2. Explain each necessary procedure before it happens, in one or two sentences, and give choices wherever possible (order, timing, who is present).
    3. Limit the number of new faces and repeated history-taking on day one. Combine assessments where safe and do the standard risk assessment once, with care.
    4. Introduce the patient to one or two other patients by name. Offer food, a seat and quiet.
    5. The psychiatrist makes a brief, unhurried first contact that is mainly about the patient's comfort and fear, not a full mental-state interrogation (the full assessment can follow over the next day).
    6. Record what frightened or calmed the patient (4.12).
  • Example language (illustrative): "I'm Sam. I'll be around you most of today. Nobody here is going to rush you. If something we do doesn't make sense, ask me and I'll explain."
  • Watch for: Rising panic (pacing, hypervigilance, mutism, aggression) after procedures, which signals the reception added threat. It is working when the patient seeks out the named staff member, eats, and sleeps.

4.5 Protecting the unit from institutional routine

  • Purpose / when: Phase 0 and ongoing. Sullivan's ward ran almost independently of the hospital and broke with its usual practices [source: Encyclopedia.com/Conci]. One secondary source says medical staff and other authority figures were largely kept off the ward [source: Wikipedia mirror, unsourced] [uncertain].
  • Steps [synthesis]:
    1. List the routines on the parent ward that add shame or depersonalisation: mass rounds, rotating agency staff, routine searches, public medication queues, uniforms, large ward rounds where the patient faces many strangers.
    2. For each, propose a safe alternative, such as small reviews with two or three known clinicians or individual medication times, and get governance sign-off.
    3. Ask visiting professionals to come in through a known staff member and keep their contact brief.
    4. Keep mandatory safety, legal and physical-health procedures. Change how they are done, not whether they are done.
  • Example language (illustrative, to management): "We're not asking to skip safety checks. We're asking that they be done by staff he knows, with an explanation, instead of by whoever is on."
  • Watch for: Drifting back to routine under bed pressure or staff shortage. It is working when audits show fewer staff strangers per patient per day.

4.6 Staff as companions: everyday peer contact

  • Purpose / when: Phases 1–3. This is the ward's core: therapeutic value of ordinary human interaction beyond the psychiatrist's interview [source: Encyclopedia.com/Conci].
  • Steps [synthesis]:
    1. Staff spend most of their shift with patients (meals, cards, walks, chores, music), not in the office.
    2. They follow the patient's lead on topics and do not push for disclosure or symptom reports.
    3. They talk about ordinary things (sport, work, family, the weather) as one person to another. This is how self-regard comes back: being treated as worth talking to.
    4. When the patient says something strange, they show interest and calm, and do not argue or endorse (see 4.10).
    5. They bring notable exchanges to 4.2.
  • Example language (illustrative): "I'm going to walk round the garden. Come if you like. No need to talk."
  • Watch for: Over-familiarity or staff using patients for company or attention. It is working when the patient starts conversations and asks about others.

4.7 Fostering patient-to-patient sharing

  • Purpose / when: Phase 2 onward. Sullivan held that patients with similar backgrounds benefit from sharing their experiences [source: Encyclopedia.com/Conci].
  • Steps [synthesis]:
    1. Arrange shared activities in small groups of 2–4 that need cooperation (cooking, a game, a garden job).
    2. Pair a newer patient with one further along, with both agreeing.
    3. Hold an optional, loosely structured, staff-present conversation group about "what it was like coming in" and "what helped." Its purpose is recognition, not therapy-by-confrontation.
    4. Step in quickly on ridicule, scapegoating or sexual pressure between patients (see Section 5).
  • Example language (illustrative): "A couple of people here felt the same panic the first week. Would it help to hear how it went for them?"
  • Watch for: One patient's psychotic material frightening another, and pairings that turn into dependence or exploitation. It is working when patients spontaneously reassure or include each other.

4.8 Normalising talk about shame-laden experience, including sexuality

  • Purpose / when: Phases 1–3, whenever sexual fear, identity conflict or other shame drives panic. On Sullivan's ward, staff were encouraged to talk casually about homosexual experience so that patients would not feel shame or fear about it [source: Blechner 2005, citing Chatelaine 1981]. Sullivan's papers link schizophrenic panic to sexual and other social shame [source: Sullivan 1929, "Archaic sexual culture and schizophrenia," 1962 collection, cited by Blechner] [uncertain]. The modern form keeps the function (de-shaming) and drops the historical form (see 4.14) [synthesis].
  • Steps [synthesis]:
    1. Make it explicit in the unit's culture that sexual orientation, gender identity, sexual fears and sexual history can be spoken of without shock or moralising.
    2. When a patient hints at sexual shame, respond calmly and matter-of-factly, without pursuing detail. Say it is common and speakable.
    3. Staff may use brief, general normalising statements ("lots of people worry about that"). They do not disclose their own sexual experiences.
    4. Offer the individual sessions (4.9, 4.11) as the place for fuller exploration.
    5. Where the shame comes from family or community condemnation, say so. Blechner stresses that a milieu must address social condemnation of a core part of identity [source: Blechner 2005].
  • Example language (illustrative): "You looked frightened when that came up. Nothing about who you're attracted to is going to shock anyone here, and it doesn't make you less of a person. If you want to talk about it more, Dr … is a good person to bring it to."
  • Watch for: Staff curiosity becoming intrusive or sexualised, and patients disclosing faster than they can tolerate (a panic spike afterwards). It is working when the topic stops triggering panic or withdrawal.

4.9 The psychiatrist's long informal interviews

  • Purpose / when: Phases 1–2 especially, continuing into 3. Sullivan held frequent, informal, lengthy interviews with patients [source: Encyclopedia.com/Conci] and made strong connections with very withdrawn patients [source: Blechner 2005]. Structure drawn from his later interview method is marked [source: Sullivan 1954, The Psychiatric Interview] [uncertain].
  • Steps [synthesis]:
    1. Meet often (daily in Phase 1 if possible) and informally: on the ward, walking, sitting together. Do not open with symptom checklists.
    2. Take a participant-observer stance. Notice your own anxiety and how it shapes what you ask.
    3. Track the patient's anxiety minute by minute. When it rises, slow down, change topic smoothly, or name the difficulty plainly.
    4. Ask about concrete events and people ("what happened then?", "who was there?") rather than abstractions.
    5. Avoid interpretations that expose or humiliate. Self-esteem comes before insight.
    6. Long silences are allowed. Staying present with a withdrawn patient is itself contact.
  • Example language (illustrative): "I'm not here to test you. I'd like to understand what the last few weeks were like from where you were standing."
  • Watch for: The patient becoming more agitated after sessions, which means the pace or topic is too threatening. It is working when the patient waits for or seeks out the sessions and gives more concrete detail over time.

4.10 Responding to psychotic communication as meaningful

  • Purpose / when: Any phase, whenever delusional, hallucinatory or bizarre communication appears. This rests on Sullivan's conclusion that even the most disturbed behaviour is symbolic activity within the human realm, so understanding is possible with enough contact [source: Encyclopedia.com/Conci].
  • Steps [synthesis]:
    1. Neither argue with nor endorse the content. Show you take the experience seriously.
    2. Listen for the interpersonal theme: humiliation, threat, being found out, longing for closeness, fear of one's own impulses.
    3. Reflect the feeling and the theme in ordinary words.
    4. Respond to the need under it: safety, company, dignity, distance.
    5. Bring the hypothesised meaning to 4.2 for the team to test against further observations.
  • Example language (illustrative): "The voices keep calling you disgusting. That sounds exhausting, and lonely. I don't see you that way."
  • Watch for: Staff "decoding" out loud in ways that feel intrusive. Keep hypotheses for the team. It is working when the patient's communications become less guarded and more connected to real events.

4.11 Inquiry phase: reviewing the lead-up to the illness

  • Purpose / when: Phase 3, once panic has eased and the patient is relating to others. The term "modified psychoanalytic treatment" is Sullivan's (1931b). The sequence below is reconstructed [synthesis] [uncertain: 1931b procedure].
  • Steps [synthesis]:
    1. Agree on the task together: "to understand what happened, so you can see it coming next time."
    2. Build a timeline of the months before admission: relationships, losses, humiliations, sexual or identity crises, sleep and withdrawal.
    3. For each turning point, ask what the patient felt, what others did, and what the patient did to protect themselves.
    4. Name the pattern (for example, isolation following humiliation, then suspicion, then panic) in the patient's own words.
    5. Link it to present ward relationships when there are parallels, gently and without accusation.
    6. Turn the pattern into early-warning signs and a plan (carried into 4.13).
  • Example language (illustrative): "So after the trouble at work, you stopped seeing anyone, and that's when the feeling of being watched started. Does that order sound right to you?"
  • Watch for: Re-emerging panic or psychotic symptoms when certain topics come up. If so, step back to Phase 2 contact (Section 5). It is working when the patient can tell the story coherently and recognises their own part without collapsing into self-blame.

4.12 Detailed observational records and case review

  • Purpose / when: All phases. Sullivan developed his views by studying detailed records of several hundred patients and was director of clinical research [source: Encyclopedia.com/Conci].
  • Steps [synthesis]:
    1. Each shift, record verbatim fragments of meaningful exchanges, what happened just before panic or calm, and who the patient sought out or avoided.
    2. Keep descriptions behavioural and human. Avoid labels like "attention-seeking."
    3. Review the records weekly in 4.2 to refine hypotheses about each patient's anxieties and what helps.
    4. Track outcomes simply (contact with others, sleep, panic episodes, readmissions) so the unit can learn.
  • Example language (illustrative, note): "21:10: refused meds after the new agency nurse came in; calmer after Sam sat with him; said 'they all know.'"
  • Watch for: Notes becoming defensive or custodial. It is working when patterns emerge that change staff behaviour.

4.13 Re-entry: working with the patient's interpersonal world

  • Purpose / when: Phase 4. The rationale is Sullivan's view of illness arising in interpersonal and social circumstances. The procedure is [synthesis] [uncertain: Sullivan's 1931 environmental-factors paper].
  • Steps [synthesis]:
    1. With consent, meet family or significant others. Explain the illness as a human reaction to overwhelming anxiety and shame, not as moral failure.
    2. Identify specific pressures to change (ridicule, condemnation of identity, isolation) and supports to strengthen.
    3. Plan graded leave and debrief each one on the unit.
    4. Connect the patient with at least one ongoing relationship: an outpatient therapist, a peer group, a trusted friend.
    5. Write a relapse-signature plan from 4.11.
  • Example language (illustrative, to family): "What helped most here was people treating him as one of them. What would make home feel like that?"
  • Watch for: A home setting that repeats the shaming (a common cause of relapse; consider supported housing). It is working when leaves go without a panic spike.

4.14 Historical practices not to be reproduced (description only)

  • Exclusion of women and gender-segregated staffing. No female nurses were allowed on the ward, not even as visitors [source: Blechner 2005; Encyclopedia.com/Conci]. Stated rationale (inferred by commentators): to lower sexual anxiety in these young men [uncertain]. Current limit: excluding staff or patients by gender breaches employment and equality law in most jurisdictions. Patient-preference single-gender spaces are a separate, policy-governed matter.
  • Harm account: Exclusion of women and gender-segregated staffing What was actually done. In 1929 Sullivan organised a special admission ward for young male patients diagnosed with schizophrenia at Sheppard and Enoch Pratt Hospital, near Baltimore. Patients came straight onto it at admission rather than being moved there later. Sullivan hand-picked every staff member, and all of them were men. "No women were allowed on the ward" [source: Encyclopedia.com, Stanton entry]. In the account of a surviving attendant, "no female nurses were allowed even to come into the ward" [source: Blechner 2005, p. 4, from Chatelaine's 1981 interviews]. The ward ran almost independently of the rest of the hospital, so the rule held around the clock: nursing, meals, activities and evenings all took place in an all-male setting. The arrangement continued under Sullivan's successor, William Silverberg, after Sullivan left in 1930 [source: Encyclopedia.com, Stanton entry]. No source describes patients or families being told about the design or asked to agree to it. Psychiatric admission at the time involved no consent process of the modern kind [uncertain]. Whether mothers, sisters or other female visitors were also kept off the ward is not documented. "No women were allowed" suggests they were [uncertain]. The rationale given by later commentators is that keeping out women, and men hostile to homosexuality, lowered the sexual anxiety and shame that Sullivan thought drove these young men's panic [source: Blechner 2005, pp. 4–5, 7]. How it caused harm. No source documents an injury, death or complaint linked to this rule. Sullivan never analysed the ward's results systematically [source: Encyclopedia.com, Stanton entry]. Blechner's "86% cure rate" comes without a cited source or method, so it cannot be used either to clear the rule or to condemn it. The risks of the design are therefore [synthesis]:
    • Avoidance became the treatment. Taking a feared group out of the environment removes the trigger without building tolerance of it. Discharge then brings the patient back into a mixed world he has not practised meeting.
    • Family contact was cut. If female relatives were excluded, patients lost contact with family members during the most frightening phase of illness, and families were shut out of care [uncertain].
    • The same theory was applied to everyone. Every patient's distress was assumed to be about sex and gender. The ward's design applied that assumption to all of them, whatever their individual formulation.
    • Staff were excluded by category. Qualified women were barred from the work because of what they were, not because of anything they did.
    Recognising it today.
    • Blanket bans on a sex, gender or other group of staff or visitors, justified by a theory about the patients' sexuality or "triggers" rather than by individual preference or safety assessment.
    • Programmes that limit family contact as a form of therapy ("we need to separate him from his mother's influence").
    • Any service that treats keeping a feared stimulus away as the treatment itself, with no plan to rebuild tolerance.
    • How legitimate practice differs: same-sex sleeping and bathroom areas, or a chosen gender of key worker or chaperone, are matters of patient choice and safety, set by policy, decided case by case and reviewed. They do not remove a whole group from the clinical team [synthesis].
    What to do instead / helping those harmed. Offer individual preferences (gender of key worker, single-sex areas) on request. Where fear of a group is part of the problem, work on it gradually, with the patient's agreement. Involve family unless the patient declines or there is a safety reason [synthesis]. When assessing someone who spent time in a closed or segregated programme, ask how contact with family and the outside world was restricted, and what they were told about why. Shame or anger about a past "treatment" design is a legitimate reaction, not a symptom [synthesis]. Staffing rules that exclude by sex or gender should be raised with the service's governance lead and, where relevant, its employment or equality officer [synthesis].
  • Staff selection by sexual orientation and encouragement of same-sex affection. Secondary accounts say Sullivan chose staff who were gay or very easy-going about homosexuality [source: Blechner 2005, citing Chatelaine 1981]. One historian's archival study describes his encouraging same-sex affection on the ward [source: Wikipedia summary of Wake, Private Practices, 2011] [uncertain: the extent and nature of this are not documented, and no specific harm was documented there]. Current limit: hiring cannot select by orientation. Any sexual or romantic contact, or sexualised intimacy, between staff and patients is a serious boundary violation and is prohibited. De-shaming is done through attitude (4.8), never through staff intimacy.
  • Harm account: Staff selection by sexual orientation and encouragement of same-sex affection What was actually done.
    • Selection. Sullivan chose the attendants himself. According to staff interviewed decades later, they were "either openly homosexual or extremely easygoing about it," and he removed "homophobic persons from the staff as much as possible" [source: Blechner 2005, pp. 4, 7, from Chatelaine 1981; the main witness was William Elliott, who joined the ward in 1929]. He also preferred people who had been through psychological disturbances like the patients' [source: Encyclopedia.com, Stanton entry]. Selection therefore rested on the staff member's own sexuality and history, not on any described training standard.
    • Talk on the ward. Staff were encouraged to talk casually with each other about their homosexual experiences, within patients' hearing, so that the patients would not feel shame or fear about such feelings [source: Blechner 2005, p. 4]. These patients were young men admitted in acute psychosis, many of them recorded as having "homosexual conflicts" [source: Wikipedia summary of Wake 2011].
    • Affection between people on the ward. Wake's study of the ward's patient records describes Sullivan's approach as "encouraging same-sex affection" [source: Wikipedia summary of Wake 2011]. It is not known between whom (patients with each other, or patients with staff), in what form, or how far it went [uncertain]. Blechner's article describes no sexual contact between staff and patients [source: Blechner 2005].
    • Staff culture and theory. Sullivan met staff informally and at length, often in the evenings at his home [source: Encyclopedia.com, Stanton entry]. Staff later described him as becoming more open about his own sexuality over the life of the ward [source: Blechner 2005, pp. 4–5]. In a 1929 paper he wrote that "any actual material arising from voluntary or unwilling concrete sexual activities has a beneficial effect" on such patients. Blechner notes that this has never been tested [source: Blechner 2005, p. 14, quoting Sullivan 1929].
    How it caused harm. No source records a specific patient injured by these practices. The evidence is thin: retrospective recollections gathered about fifty years later, and summaries of a book. Blechner, a sympathetic commentator, presents the aim of removing patients from homophobia as a real insight [source: Blechner 2005, p. 5]. The aim of reducing shame is legitimate. The harm lies in the means [synthesis]:
    • Sexualised environment for people who could not consent to it. Staff discussing their sex lives in front of acutely psychotic young men exposed patients to sexual material they had not chosen, at a time when their judgement and boundaries were at their most fragile.
    • Affection as treatment. Encouraging affection as part of treatment, in a closed unit run by staff with power over patients, blurs the line between care and intimacy. That is the setting in which staff sexual exploitation of patients develops.
    • A theory that excuses abuse. A theory holding that even unwilling sexual activity can benefit patients removes the very safeguard (consent) that protects them.
    • Labels imposed by the theory. Patients were placed in a programme built on an assumption about their sexuality, recorded as "homosexual conflict." That may not have been their own account of themselves.
    • Selection by identity and loyalty. Choosing staff for personal identity and closeness to the leader, and meeting them informally at his home, builds an in-group in which boundary problems are less likely to be named or reported.
    Recognising it today.
    • Staff or therapists sharing their sexual history or current sex life with clients, framed as "normalising" or "modelling openness."
    • Touch, cuddling, "holding", "reparenting" or romantic closeness offered as corrective emotional experience. Talk of "special" relationships with particular patients.
    • Theories that sexual experience, or a relationship with the therapist, will heal.
    • Teams recruited for loyalty to a charismatic leader, socialising privately with him or her, where questioning the method is treated as a personal failing.
    • How legitimate practice differs: affirming care for LGBTQ+ clients, and paid peer-support workers with lived experience. Peer workers have a defined role, training, supervision and the same boundary rules as other staff. Therapist self-disclosure is brief, deliberate, for the client's benefit and discussed in supervision, never sexual [synthesis].
    What to do instead / helping those harmed. Reduce shame through attitude: neutral, non-judgemental questions about sexuality and identity, and an affirming stance (see 4.8). Apply the ward's sexual-safety policy, and keep a clear rule of no sexual or romantic contact between staff and patients, in any direction and with no exceptions for "therapeutic" reasons. When assessing someone who was in a programme like this, ask directly and calmly whether any helper touched them, talked to them about sex, or had a romantic or sexual relationship with them. Expect shame, self-blame and confusion about whether it "counted" as abuse. Do not assume their sexual orientation from what the programme recorded. Sexual contact or sexualised behaviour by a clinician must be reported through safeguarding and to the clinician's professional regulator, following local law and the client's wishes as far as the law allows. In supervision, a colleague's sexualised self-disclosure or "special" patient relationships are a reason to act early [synthesis].
  • Alcohol as an aid to contact or sedation. Some accounts of Sullivan's era say he sometimes used alcohol with acutely disturbed patients [uncertain]. This is mentioned only so that no reader adopts it. It is not acceptable practice. Any sedation is a medical decision under standard prescribing.
  • Harm account: Alcohol as an aid to contact or sedation What was actually done. This could not be established. None of the sources for this guide mention alcohol on Sullivan's ward: Blechner's account of the ward, the Encyclopedia.com entries (Stanton; Conci) and the Wikipedia summaries [source: Blechner 2005; Encyclopedia.com]. Sullivan's 1931 papers, the 1962 collection and Perry's biography, where any such detail would be recorded, are not drawn on here [uncertain: the claim that Sullivan gave alcohol to acutely disturbed patients remains unverified; no setting, frequency, method or consent procedure can be described]. Psychiatrists of the period did experiment with alcohol in schizophrenia, for example in stuporous patients, alongside barbiturate "narcosis" and drug-assisted interviews [source: titles only: "Psychopharmacological study of schizophrenia … cardiazol, sodium amytal and alcohol in schizophrenic stupor," Journal of Mental Science; "Narcotherapy of schizophrenia by intravenous alcohol drip method," CIA Reading Room document] [uncertain]. How it caused harm. No documented case from Sullivan's ward was found. In general, alcohol given to a person in acute psychosis carries these risks [synthesis]:
    • Physical danger. Oversedation, vomiting and choking, falls and injuries, and dangerous interactions with other sedative drugs.
    • Worse mental state. Disinhibition and agitation, worsening confusion, and alcohol-related worsening of psychosis.
    • Unreliable disclosure. Material produced while intoxicated is open to suggestion and unreliable, as with other drug-assisted interviews.
    • Lost capacity to consent. The patient loses the capacity to agree to, or refuse, what is happening in the session.
    • A model of drinking to cope. For someone at risk of a substance-use problem, a helper who uses alcohol to ease contact teaches them to drink to cope.
    Recognising it today.
    • A clinician or support worker drinking with a client, holding sessions in bars, offering "one drink to help you relax", or tolerating intoxication in sessions as useful.
    • Any substance, including cannabis or psychedelics outside approved, ethics-reviewed trials, used by a helper to make a client "open up."
    • Sedation given by non-prescribers, or "as needed" medication used to manage a ward rather than to treat the patient.
    • How legitimate practice differs: medication is prescribed by an authorised prescriber against a documented indication, with consent or a lawful framework, and monitored. Research uses of psychoactive drugs run under ethics approval and regulatory oversight [synthesis].
    What to do instead / helping those harmed. For acute agitation or panic, use the ward's de-escalation approach first (see the security-before-inquiry principle above). Any sedation is a prescriber's decision under standard protocols. Build contact through presence and patience, not through altered states [synthesis]. When assessing clients, ask whether any helper ever gave them alcohol or drugs, and screen for substance-use problems that may have been encouraged that way. A clinician supplying alcohol or drugs to a patient is a fitness-to-practise matter to report to their regulator and, where the client is at risk, through safeguarding [synthesis].

5. Handling problems

The sources say almost nothing about handling specific difficulties. Unless tagged otherwise, everything in this section is [synthesis], derived from Sullivan's principles (anxiety and shame reduction, peer acceptance, protected milieu).

  • Persistent panic or overwhelm. Cut demands to a minimum. Keep one known staff member nearby in shifts. Lower stimulation (quiet room with company, not seclusion as a first resort). Stop inquiry-phase work and go back to Phase 1 contact. Ask the prescriber to review; do not assume the milieu alone is enough. In 4.2, look at what on the unit may be adding threat.
  • Withdrawal and mutism (the "stuck" patient). Keep up regular, undemanding presence: sit nearby, talk a little about ordinary things, invite without pressure. Sullivan reportedly made strong connections with very withdrawn patients [source: Blechner 2005]. Over weeks, review for a hardening withdrawn course (Section 3, "Watch the course") and get a second opinion and medical review if there is no movement.
  • Suspicion and hostility toward staff. Do not argue with the accusation. Acknowledge the fear and give the patient control over distance ("I'll sit over here"). Make sure staff behaviour is open and predictable: explain everything and avoid whispering in sight. In 4.2, check whether a real slight set it off.
  • "Resistance" to inquiry (Phase 3). Treat it as a sign the topic threatens self-esteem, not as defiance. Go back to safer ground, say the topic can wait, and come back to it later in smaller steps. Never force interpretations.
  • Non-adherence to medication or the programme. Find out the patient's reasons (side effects, shame, the meaning of the medication to him) in the same non-shaming way. Arrange prescriber review. Separate the relationship from compliance: staff stay warm whatever the patient decides, within legal and safety limits.
  • Ruptures (a patient feels betrayed or humiliated by a staff member). Name it promptly. The staff member acknowledges their part simply, without over-apologising or defending. Bring it to 4.2. If the relationship cannot be repaired, change the key worker without shaming the staff member, and tell the patient why.
  • Peer ridicule, scapegoating or sexual pressure between patients. Step in immediately and calmly. Protect the target. Talk with the instigator privately about what drove it, often their own fear. Change pairings and activities. Unwanted sexual contact goes through the safeguarding protocol.
  • Staff shaming or custodial drift. Raise it in 4.2 with specific examples. Retrain, and if it persists, move the person off the unit. Patients cannot benefit from peer acceptance if staff withhold it.
  • Over-involvement and boundary drift (staff becoming confidants, special friendships, gifts, contact outside the unit). This is the characteristic risk of a peer-style milieu. Set explicit rules: no contact outside the unit, no sexual or romantic content, no personal social media. Monitor in supervision and act on early signs. See 4.14.
  • Sexual panic or identity crisis driving psychosis. Use 4.8 in particular: calm, non-moralising acknowledgement, a protected place to talk in individual sessions, and no pressure to disclose or label. Check for family condemnation and plan for it at re-entry (4.13).
  • Institutional pushback (bed pressure, short stays, staffing rotas). Protect the core minimum: one named key worker per shift, a daily staff talk, and a non-shaming admission. Collect simple outcome data (4.12) to argue the case.
  • Short admissions (days, not weeks). Compress the model: do Phase 1 fully, start Phase 2 contact, and hand the inquiry (4.11) and re-entry work (4.13) to outpatient or residential follow-up through a warm, in-person introduction.
  • Mixed-gender and diverse units today. Sullivan's ward was all-male and young. Apply the same principles to any patient group. Match peer pairings to patient preference, and keep shame regulation sensitive to gender, culture and sexuality.
  • Suicide risk, aggression or acute medical problems. The milieu does not replace standard risk assessment and management. Use the unit's suicide and violence risk protocols, observation levels and medical review. Where restrictive measures are unavoidable, carry them out with explanation and dignity, and debrief afterwards with the patient and in 4.2.
  • Staff burnout or vicarious distress. Long contact with frightened, psychotic patients is demanding. Make sure supervision, rotation and staff support exist. Sullivan's emphasis on talking with staff extends naturally to looking after them.

Sources

Sources:

Further reading (not drawn on unless listed above):

  • Sullivan, H. S. (1931a). Socio-psychiatric research: its implications for the schizophrenia problem and for mental hygiene. American Journal of Psychiatry, 87.
  • Sullivan, H. S. (1931b). The modified psychoanalytic treatment of schizophrenia. American Journal of Psychiatry, 88.
  • Sullivan, H. S. (1962). Schizophrenia as a Human Process (introduction and commentaries by H. S. Perry). New York: Norton. Includes the 1928 "Tentative criteria of malignancy" and 1929 "Archaic sexual culture and schizophrenia" papers.
  • Sullivan, H. S. (1953). The Interpersonal Theory of Psychiatry. New York: Norton.
  • Sullivan, H. S. (1954). The Psychiatric Interview. New York: Norton.
  • Perry, H. S. (1982). Psychiatrist of America: The Life of Harry Stack Sullivan. Cambridge, MA: Belknap/Harvard.
  • Chatelaine, K. L. (1981). Harry Stack Sullivan: The Formative Years. Washington, DC: University Press of America.
  • Wake, N. (2006). "The full story by no means all told": Harry Stack Sullivan at Sheppard-Pratt, 1922–1930. History of Psychology, 9.
  • Wake, N. (2011). Private Practices: Harry Stack Sullivan, the Science of Homosexuality, and American Liberalism. New Brunswick: Rutgers University Press.

Limits of the sources: Sullivan's 1931 papers, the 1962 collection and Perry's biography are not drawn on. Ward details therefore rest on reliable secondary summaries and are tagged [uncertain] where they go beyond them.

Added for the harm accounts in 4.14:

← Previous · I-082Intensive psychoanalytic psychotherapy of schizophreniaNext · I-084 →Need-adapted treatment