mental health help without going to the hospital: what exists, what the studies found, and when the hospital is the right call

the search that lands here is usually typed at a bad hour: mental health help, without the hospital. not because the hospital is wrong, but because of what it costs, what it takes, and what people have heard happens there. this page is what exists in between, what the studies found when they tested it, and the honest part: when the hospital is the right call.

here is what is on the page as of october 2026: the federal 988 faq on what the line does and when it involves 911, the one randomised trial and the one matched study of mobile crisis teams, a 2026 analysis of four states' medicaid data on who uses crisis stabilization units, the houston services by name from the county authority's own page, and samhsa's own list of when to call 911 instead. both directions, every number with its source.

someone to call: what 988 is, and what it is not

988 is a national line you can call, text, or chat with at 988lifeline.org, at any hour, for anyone, for suicide, mental health or substance use problems^[1],[6]. you do not have to be suicidal: samhsa's faq says there is no reason too big or too small to reach out, and that you can call about someone else^[1]. nobody asks for payment or insurance information^[1]. the lifeline's own page says the conversation is free and confidential^[6]. the faq says the wait after the greeting menu is typically under a minute, “at the time of publishing”, a phrase with no date attached^[1].

what happens when you call: a greeting, a menu (press 1 for the veterans crisis line, or choose spanish), or wait and you are routed to a local 988 centre by your approximate location or your phone's area code; dial zero to skip the menu^[1]. services run in english and spanish, with interpreters for calls in more than 240 languages^[1].

the question most people actually have is whether calling sends the police. samhsa's answer, in its own words: 988 is for emotional and behavioural health support “with minimal law enforcement involvement unless there is an immediate physical safety threat”; most crises are managed by counsellors “without involving law enforcement”; “a small percentage” of callers require activation of 911, “many of those” with the caller's consent and cooperation^[1]. the other direction, stated just as plainly on the same page: if you or the person you are calling about is in physical danger, the counsellor will contact 911, and in rare situations 911 may be asked to geolocate the phone through the carrier^[1]. “a small percentage” is not a number; samhsa does not publish one on that page.

what 988 is not: it is a phone line, not a service system. samhsa says the vision is a coordinated response that links callers to mobile crisis teams and stabilization centres, and that “currently, these crisis care services do not exist in all areas of the country”^[1]. whether the next two layers exist where you live is a county question, and the person on the 988 call is the fastest way to find out.

someone to come: mobile crisis teams

samhsa describes mobile crisis teams as mental health and substance use professionals, paraprofessionals and peer support workers who respond to people “in their community environments where people are comfortable to avoid unnecessary law enforcement involvement, emergency department use, and hospitalization”^[1]. what they typically do, from the same page: triage and screen, including explicitly for suicidality; assess; de-escalate; provide peer support from people in recovery; create a safety plan with you; coordinate follow-up; connect you to community support^[1]. in most places that have one, a two-person team is on call; some use telehealth as part of the response^[1].

two models exist, and the difference matters if police involvement is the thing you are trying to avoid: a mobile crisis team of behavioural health staff and peers, or a co-responder team in which law enforcement or emergency medical staff respond jointly with a mental health professional. which one shows up depends on the assessed risk, what the community has, and local law^[1]. availability “currently varies by the location”; samhsa says it has been expanding over the last several years, and that these services are “ideally” available around the clock, which is a way of saying they often are not^[1].

on the fear that a visit ends in a locked ward: samhsa's words are that the goal is assessment, de-escalation, safety planning and referral “in a manner that minimizes the need for hospitalization”, that “most people with suicidal thoughts will not require involuntary transport or hospitalization”, that when a team does judge hospitalisation necessary it works to meet the person's preference, and that “involuntary hospitalization may be needed in circumstances of imminent risk where the other efforts are not successful”^[1]. on cost: billing depends on the operator, payment codes exist for public and private payers, and samhsa says nobody should be denied or delayed access based on ability to pay, which it calls the standard practice of most teams^[1]. since the american rescue plan act of 2021, states can also claim an 85% federal match for qualifying community-based mobile crisis services under medicaid^[1].

what the studies found

the claim that mobile teams keep people out of hospital rests on surprisingly little, and the two best records point in different directions on different questions. the matched study first^[2]. in one public mental health system, 1,696 people who got a community mobile crisis response were matched after the fact to people who went through a hospital-based crisis service, on seven variables (sex, race, age, primary diagnosis, recent service use, substance use indication, and disability certification), leaving 1,100 pairs. the mobile crisis group was hospitalised 8 percentage points less often, and a person on the hospital route was 51% more likely to be admitted within 30 days (p<.001). treating more people in the community did not raise the risk of a later admission^[2].

hold the design next to the number. nobody was randomised: the two routes were chosen by whatever sent each person to one door or the other, and matching on seven measured variables cannot remove what was not measured. the people most likely to be hospitalised in that study were young, homeless, in acute trouble, referred by hospitals or the legal system, using substances, without income, and severely disabled^[2], which is a description of who the hospital route selects for, not only of what it does.

the randomised trial is the stronger design and the narrower question^[3]. at one urban emergency department, 120 people evaluated for suicidal thoughts, plans or behaviour and then discharged were randomised to follow-up either by a mobile crisis team in the community or at an outpatient clinic, with the same services on offer in both arms. first clinical contact happened for 69.6% of the mobile-team group against 29.6% of the clinic group (relative risk 2.35, 95% ci 1.55 to 3.56). and then the other direction: on symptoms and functioning, “no significant differences between groups” at two weeks or three months, and no difference between people who attended their first appointment and people who did not^[3]. the honest reading is that a team that comes to you gets you to the next appointment, and that getting to the next appointment did not, in that trial, make people measurably better by three months.

somewhere to go that is not an emergency room

the third layer is a building: a crisis stabilization unit, a psychiatric emergency service, a 23-hour observation unit, names vary by state. the first multi-state look at who actually uses them was published in jama health forum in may 2026, from 2022 medicaid claims in arizona, massachusetts, oklahoma and washington: 144,193 people and 347,920 acute behavioural-health visits to either a crisis stabilization unit or an emergency department^[4]. the share handled in the units rather than the emergency department ran from 55.9% in arizona (60,371 of 108,048 visits) to 13.2% in oklahoma (3,811 of 28,810)^[4]. same federal programme, four-fold difference in whether the door exists.

the finding that should change how you read every “alternative to the hospital” claim is the diagnosis mix. schizophrenia and related disorders were the most common reason for a unit visit in three of the four states (27.7% in arizona, 30.5% in oklahoma, 27.8% in washington). suicide-related diagnoses were rare in the units, from 0.02% of visits in massachusetts to 0.8% in washington, and common in emergency departments, from 17.1% in massachusetts to 30.4% in oklahoma^[4]. people in a suicidal crisis were, in 2022, still overwhelmingly going through the emergency department. the authors counted doors, not outcomes; the study cannot say which door was better for anyone^[4]. alcohol-related disorders were a large share of visits in both settings^[4].

houston, by name

this page exists partly because the search “mental health help without hospital houston” was landing on our houston sliding-scale therapy page, which does not answer it. the answer is the harris center for mental health and idd, the local mental health authority for harris county, read from its own service page on 16 october 2026^[5]:

• 24/7 crisis and access line: 713-970-7000. the site says it is always open with counsellors available if you need mental health support^[5].
• mobile crisis outreach team (mcot): mobile crisis response for people of all ages “who are experiencing a mental health crisis and are unable/unwilling to access services independently”, using the least restrictive approach until the crisis resolves or the person is placed in an appropriate setting. staffing listed on the page includes a psychiatrist, nurses, licensed clinicians and two peer support specialists. the page also says mcot may collaborate with the houston police department and the harris county sheriff's office when indicated^[5].
• psychiatric emergency services (pes), at the neuropsychiatric center: psychiatric evaluation and initial treatment for all ages in psychiatric crisis, staffed around the clock by psychiatrists, nurses, social workers, counsellors, technicians and peer specialists. it “accepts voluntary and involuntary admissions”; depending on the evaluation, a person may be referred to an outpatient provider, admitted to the extended observation unit, transferred to an inpatient facility, or sent to another programme^[5]. that is the both-directions fact of this page: the same door that can send you home with an outpatient referral can also admit you.
• crisis stabilization unit (csu), at the same centre: inpatient stabilisation for harris county residents aged 18 and over in an acute crisis. the page says it is voluntary, requires a referral by a doctor, runs 24 hours a day and 365 days a year, and that the average stay is approximately 3 to 5 days^[5].
• crisis call diversion: crisis phone counsellors sit inside houston's 911 dispatch and assess whether a behavioural-health call needs police or medical response, or can be diverted to phone referral or the mobile team. you cannot call it directly; it takes referrals only from 911 call takers and dispatchers^[5].

the page also describes a crisis intervention response team that pairs a licensed clinician with a police officer, and a homeless outreach team of officers and care coordinators^[5]. if avoiding police contact is your constraint, say so on the access line; the page lists both police-paired and clinician-only responses, and which one comes is a decision made on the call. outside harris county, the same structure usually exists under your county's local mental health authority, and 988 can route you to it^[1].

when the hospital is the right call

this is samhsa's list, not ours, of when 911 is appropriate rather than 988: “a suicide attempt in progress. a person has a specific suicide or self-harm plan that they are or intend to immediately carry out and they have the means to carry it out. a person is suspected of being in the midst of an overdose. a person is experiencing physical symptoms such as chest pain or shortness of breath that could be either a panic attack or an acute medical emergency.” the next sentence on the page is an invitation to reach out earlier, “before you get to this point”^[1].

and the limit of this page: it can tell you what each layer is and what the studies found. it cannot tell you which side of that list you are on tonight. the 988 counsellor's job, and the mobile team's, is exactly that judgment, made with you. for the other question people bring to this search, whether medication is what keeps people out of hospital or what puts them there, the register studies are on their own page.

what to ask on the call

questions that the records above make worth asking, in whatever order helps: does my county have a mobile crisis team, and is it clinician-only or paired with police^[1]; is there a crisis stabilization unit or psychiatric emergency service near me, does it take walk-ins or does it need a referral, and is it voluntary^[4],[5]; what would make you contact 911 during this call^[1]; what happens after tonight, and who calls whom^[3]. the trial evidence is that the follow-up appointment is the thing most people never get to^[3]; asking who is responsible for it is not a small question.

where talking to people fits

below the crisis layers there is a quieter one: warmlines, which are peer-staffed lines for the nights that are bad but not emergencies, with hours that vary by state (what they are and how to find one), and peer support, which is people with lived experience rather than clinicians. veisund is the second kind: free, under a handle, post any hour, and people who have been through something similar can answer. it is not a crisis service and it is not care. one automated exception we name: if a post suggests immediate risk, the app responds with crisis resources. if any part of this page described tonight, the number is 988, not an app.