A Rapid Review of “Low-Threshold” Psychiatric Medication Prescribing: Considerations for Street Medicine and Beyond

For psychiatric providers, perhaps the most challenging aspect of prescribing in low-threshold settings is making a DSM-based psychiatric diagnosis to justify a prescribed medication. Individuals treated in low-threshold settings often do not have access to their psychiatric records. They may not have social support from friends, family, or others who may be able to provide meaningful collateral history. They may not tolerate the lengthy probing assessments required to establish a clear DSM-5 diagnosis (7). They may not recall or may not feel comfortable disclosing previous manic episodes or past traumas. Duration of symptoms may be difficult to elicit, and substance use may complicate the diagnostic picture. For some individuals, the clinician may be able to obtain past clinical information from public databases for Medicaid recipients, the electronic medical record used in the clinical setting (e.g., the Care Everywhere feature in Epic), summaries of previous clinic or hospital discharges, collateral history from a case manager or social support, or pharmacy records. In practice, however, no widely applicable survey or standard exists for solving such diagnostic dilemmas. Prescribers instead need to rely on their acumen in the moment and use sound clinical judgment to reasonably diagnose and safely treat these individuals.

For individuals with prominent diagnostic uncertainty, a symptom-based (i.e., psychosis, mood, anxiety, trauma, and substance use) diagnostic and treatment approach may therefore be the most reasonable way to weigh the risks and benefits of medication use (Table 1). Medications can be collaboratively chosen to maximize symptom relief while minimizing risks for harm. This approach is particularly suitable when the medication recipient may not be available for a follow-up visit and when obtaining blood work may be impractical. In effect, when initially prescribing to such recipients in low-threshold settings, a provider’s best assumption might be that no in-person or laboratory monitoring may be possible.

Across all symptom clusters encountered in low-threshold settings, a few factors should be considered when choosing medications. First, given the significant care barriers experienced by many individuals treated in low-threshold settings, these individuals are at high risk for missing doses and running out of medication. Thus, prescribing medications with withdrawal or discontinuation syndromes, in particular medications with short half-lives, might lead to distress and later reluctance to undergo further medication trials (22). Simple medication regimens (e.g., one medication with a moderate-to-long half-life and dosed once per day) might aid in adherence. Second, gastrointestinal adverse effects can be particularly troubling for persons without access to restrooms. Slower titration or avoidance of medications with potent gastrointestinal effects should be considered (22). Withdrawal symptoms, distressing gastrointestinal effects, or any other adverse medication-related effect could influence an individual’s willingness to follow up for further treatment. Third, medications requiring refrigeration or secure storage should not be prescribed to persons who do not have access to these amenities (28). Finally, factors such as limited financial means and lack of access to personal transportation can be barriers to acquiring medications from pharmacies. Ideally, low-threshold practitioners might develop partnerships with pharmacies that deliver medications to the clinic or to nonresidential settings. Treatment team members, including peers, outreach workers, community health workers, or navigators, could help by accompanying clients to the pharmacy or even by picking up medications from the pharmacy and delivering them to clients. If nothing else, providers should attempt to use nearby pharmacies that clients could reasonably access by foot or public transportation.

A first symptom cluster to consider is psychosis, which may include symptoms such as hallucinations, delusions, and disorganized thinking and behavior. Second-generation antipsychotics (SGAs) are a reasonable medication class to use in treating individuals with psychosis in low-threshold settings. SGAs do not require routine or extensive laboratory monitoring when they are prescribed for short courses, although longer-term SGA use is associated with weight gain and risk for developing metabolic syndrome and type 2 diabetes mellitus (29). SGAs are preferred over first-generation antipsychotics (FGAs) because SGAs have a lower risk for inducing debilitating extrapyramidal symptoms that can impair movement (2). Because SGAs do not differ significantly from one another in efficacy for managing psychotic symptoms, patient choice as well as past responses, adverse effect profile, and cost should be paramount drivers for selecting an SGA (30). One adverse effect that warrants special consideration for individuals who sleep in unsafe spaces (such as shelters, outdoors, and others) is sedation. Individuals who are sedated may be unable to defend themselves from unpredictable interpersonal violence (7, 22, 28).

For individuals who have a known treatment history, have been seen several times at a given clinical site, and are amenable to injection medications, long-acting injectable antipsychotics (LAIs) may be appropriate for psychosis management. Expert consensus recommends a brief oral trial (between 4 and 14 days) of the antipsychotic before administering an LAI (31). In a large, prospective trial using a national database, LAIs reduced rehospitalization rates by 20%–30% relative to oral antipsychotic medications (32). Numerous other studies have shown benefits of SGA LAIs relative to oral agents in relapse prevention and rehospitalization rate reduction (33–35). Several studies have reported that LAIs can be particularly helpful for medication adherence for persons who are not housed (36–38).

A second symptom cluster to consider is mood-related symptoms, which include mania, hypomania, depression, and mixed mood symptoms.

Anxiety is a third symptom cluster to consider among clients in low-threshold settings. According to professional treatment guidelines, the first-line treatment for generalized anxiety disorder is an SRI or SNRI (25). Even if a formal anxiety disorder diagnosis cannot be made, given the relative safety of SRI and SNRI medications, individuals who present with primary anxiety symptoms in the absence of acute substance use or evidence of a bipolar diathesis could benefit from these medications.

Additionally, given their rapid action and U.S. Food and Drug Administration approval for a variety of anxiety presentations, benzodiazepines could be considered in the treatment of unspecified anxiety. However, as we discuss below, the risks of prescribing benzodiazepines in low-threshold settings might outweigh potential benefits.

A final class of medications that might be considered for managing anxiety is antihistamines, such as hydroxyzine. These are non–dependence-forming agents that can be effective for treating individuals with acute anxiety (53). Again, providers should use caution when prescribing these sedating medications to individuals who sleep in potentially unsafe places.

A fourth, and common, symptom cluster seen in low-threshold settings is trauma-related symptoms. Recent evidence raises questions about whether psychotherapy, the previous treatment standard, is more effective than medication in the treatment of individuals diagnosed as having posttraumatic stress disorder (PTSD) (54). In low-threshold settings, upfront medication prescription may be appropriate during phases of engagement and shared decision making regarding possible short- or long-term therapy initiation. According to professional treatment guidelines, first-line medication classes for PTSD are SRIs or venlafaxine (an SNRI) (26). As discussed above, SRIs and SNRIs have low risk profiles and do not require routine laboratory monitoring, making them suitable for use in low-threshold settings. Some evidence suggests that prazosin can effectively manage PTSD-associated nightmares. However, in a recent trial of combat veterans, prazosin did not alleviate distressing dreams or improve sleep quality (55). Current evidence recommends avoiding benzodiazepines in acute-trauma settings, given their lack of efficacy (56).

A final symptom cluster that is common in low-threshold settings involves illicit substance use.

It is important to consider whether a given prescription dosage could increase the risk of death or serious injury in the setting of overdose. Various strategies, such as bubble-packing the medication, prescribing 7 days of the medication with refills, or, if legal and feasible, holding the prescription at the low-threshold site and dispensing a week’s worth of medication at a time, could reduce this risk for self-harm.

When considering how many pills or refills of a medication to dispense, prescribers must balance a range of factors, including promoting longitudinal treatment engagement, managing acute or subacute symptoms, or continuing medications an individual is currently taking. Each goal might dictate a unique time line for prescribing initial and subsequent medications. Limiting medication fills to 30 days with no refills could encourage treatment reengagement. Follow-up visits could be essential for monitoring medication responses and offering additional resources to patients. During these visits, providers could also address adverse effects that could interfere with treatment. Further considerations that may influence duration of prescriptions include, but are not limited to, availability of prescription drug–monitoring program records for cross-referencing, availability of collateral for verification of previous medication regimens, duration of use and tolerability of the current medication, medical risk of a given prescription to a specific recipient, risk for medication misuse or diversion, and cost of the medication. Ultimately, if a patient is well known to a low-threshold service provider, the prescriber may determine that refills can be reasonably and safely dispensed to the patient. Besides these timing considerations, it is important to concretely consider where patients might receive their next medication prescription, including but not limited to the current low-threshold setting, another low-threshold setting, or an established outpatient clinic to which they have been newly connected.

Although it may be tempting to assume that the prescriptions provided in low-threshold settings will serve as a bridge to more enduring, focused, and stable outpatient psychiatric care, significant risks arise if the medication recipient does not present for follow-up care and if medications become a “bridge to nowhere.” The greatest benefit that prescribers at low-threshold settings may provide in such instances is to be welcoming and empathic, establish a therapeutic and collaborative relationship, prescribe safe and tolerable medications for symptom relief, encourage repeat visits, and work to connect the individual to additional resources. The practitioner’s goal might remain connecting the patient to long-term outpatient care even when larger social factors might nullify this possibility in practice.

Deciding whether to prescribe controlled substances in low-threshold settings can be a challenging dilemma for treatment providers. Benzodiazepines are helpful in the acute treatment of individuals with anxiety, and stimulants can manage functionally impairing attentional conditions. However, with the exception of buprenorphine, which has an evidence base for use in low-threshold settings (59, 60, 62, 71), we feel that prescription of a controlled medication requires an established treatment relationship that is grounded in mutual trust. This relationship helps ensure that these medications, for which the risks for misuse are severe, are being taken safely. The low-threshold setting often does not allow for implementation of such safeguards. Therefore, we caution against routinely prescribing controlled substances other than buprenorphine. However, many individuals served in low-threshold settings could benefit immensely from appropriately and safely prescribed controlled substances. Perhaps the most effective role prescribers could play in such situations is to help these individuals engage with longitudinal care resources in their communities, where these medications can be safely prescribed. It is also worth noting that the COVID-19 pandemic has made access to some of these additional resources more challenging. This reduced access may justify more liberal prescription of take-home supplies of controlled substances for management of substance use disorders, while the crisis persists (72).

If a psychiatric medication has a narrow therapeutic index, has high toxicity risk, or requires frequent laboratory monitoring or other regular and ongoing evaluations, it may not be appropriate to prescribe the medication in a low-threshold setting. Such medications include lithium, tricyclic antidepressants, and clozapine. Safely prescribing these medications, even for short courses, is inherently difficult for individuals with an uncertain ability to follow up with treatment, unknown medication adherence, and unknown medical histories. The exception may be for individuals who received one of these medications in a previous setting (e.g., during a recent inpatient admission) and who need a short course of the medication to bridge them to an outpatient visit with a known provider. Still, risks and benefits should be weighed carefully, and collateral history would likely be needed to ensure safe and appropriate medication prescription. For example, a provider in a low-threshold setting might prescribe sufficient clozapine to prevent a patient from missing two consecutive days of the medication, which would necessitate medication retitration, a high-risk scenario. To prescribe clozapine, the provider would need to confirm the medication dose and timing of the last dose, meet federal Clozapine Risk Evaluation and Mitigation Strategy program requirements (including providing an absolute neutrophil count, in accordance with the patient’s monitoring frequency), locate a pharmacy to fill the medication, and ensure that the patient is reconnected with their long-term outpatient psychiatric clinic for subsequent clozapine refills.

To gain a more complete picture of a patient’s social situation, with the hope that this broader understanding will lead to safer and more targeted psychiatric medication prescribing, prescribers could consider using the Structural Vulnerability Assessment Tool (73). Derived from social-scientific studies on social determinants of health, this structured interview guide assesses an individual’s access to the social, economic, and health resources that may affect their experience with medical and mental health services. In a low-threshold setting, the provider may modify assessment questions to specifically evaluate access to social services, proximity to mental health emergency centers, and presence of a social support system that could assist with medication management.

History of previous medication trials can also be difficult to ascertain in low-threshold settings. Individuals may be distrustful of medications if they have had adverse effects from the medications in the past. Tools such as the Psychiatric Medication History (74) might aid prescribers in gathering information about past medication experiences in a structured and prompt manner. If an individual cannot recall past medications, the provider might also contact one or more pharmacies at which the individual has previously filled prescriptions or reference shared electronic medical records (e.g., Care Everywhere) to obtain this information.

Finally, and importantly, providing value-concordant care—or care aligned with recipients’ treatment goals and preferences—is a major challenge for psychiatric medication prescribing in low-threshold settings. This issue is especially poignant when medication preferences differ between medication recipients and providers. In such situations, to achieve value-concordant care, prescriber attention to the subjective experiences of medication recipients is crucial (75). Individuals may be hesitant or unwilling to take psychiatric medications or to participate in research conducted in these settings. The reasons for this resistance are complex and range from personal experiences to broader cultural influences. Special consideration should be made to the disempowering experiences that many individuals have had with the health care system. Indeed, these experiences have been documented in clinical research, where individuals treated in low-threshold settings have received inappropriate psychotropic medication prescriptions at high rates (23). Such experiences can foster understandable distrust toward medical institutions and their practices (76). For these reasons, it is crucial that informed‐consent discussions with patients regarding the risks, benefits, and alternatives to each recommended medication be open and honest. The decision to prescribe medications must not be made lightly. Shared decision making can help providers explore care recipients’ values and work toward a consensus for agreeable goals of care (77, 78). These goals could ultimately include more intensive and frequent psychiatric care, including regular engagement in behavioral or other therapies. Notably, prioritization of medication recipients’ preferences has been associated with longer and more stable treatment relationships with providers (79).

Our analysis had several limitations. First, given the rapid, unsystematic nature of the literature review, it is possible that we missed studies specific to prescribing in low-threshold outpatient psychiatric settings. Still, we identified only two immediately applicable sources after a review of >2,000 abstracts obtained with an extensive list of search terms, indicating an evident dearth of literature on this topic. Second, because little evidence-based work has been published on the prescription of most medications in low-threshold settings, the data on these prescribing practices have, by default, largely been extrapolated from research conducted in more clinical and controlled settings. We used the term “considerations” rather than “recommendations” throughout this review to avoid overstating our conclusions, given the lack of published and applicable research in this area. Finally, although these medication-specific considerations were literature based, a comprehensive review of all clinical psychopharmacologic literature for each medication was beyond the scope of this study. For practice standards, we referenced well-established and comprehensive psychopharmacology texts and up-to-date professional society and government prescribing guidelines. We assume that readers of this review who are prescribers are knowledgeable about the indications, uses, benefits, and common and serious adverse effects of the medications they prescribe. We encourage prescribers to consult comprehensive prescribing resources for complete details on all medications.