County of SAN Luis Obispo Board of Supervisors Agenda Item Transmittal (1) Department (2) Meeting Date (3)

R1

SLO County should commit to creating a single, integrated, and unified mental health services center that houses the PHF, the CSU, the MHET, outpatient coordination, juvenile mental health services, and that includes a medical health triage and screening facility where all Held persons, regardless of age, categorization, or insurance status, can be medically cleared prior to placement in an appropriate section of the mental health facility. This recommendation has not yet been implemented, but may be implemented in the future. The County Behavioral Health Department fully embraces the concept of a “one stop shop” and a full continuum of care for individuals with psychiatric or addiction care needs. The idea of a multi-service site has been planned as part of the anticipated Health Campus remodel, originally planned for 2024-25. This plan had to be set aside during the Covid-19 pandemic response. Once the Health Campus project is re-initiated, plans to develop a multiple service site will continue. The ability to fund such services to be agnostic of payor (that is, to be available for those with private insurance as well as those with Medi-Cal coverage) would need to be fully evaluated over time. The Mental Health Plan, which is the role of the County Behavioral Health system under contract with the Department of Health Care Services, in coordination with CenCal Health, could develop a unified service for individuals who are SLO County Medi-Cal beneficiaries. Integration, or at a minimum, co- location, of physical and behavioral health care is a goal of the California AIM (Advancing and Innovating Medi-Cal) project to improve Medi-Cal services across the State. Certain restrictions due to current State licensing and regulations limit the ability to have multiple services under one roof or provided by a single agency. One model of co-location, currently existing in Orange County, was only able to be created by having different providers for crisis stabilization, PHF, sobering, physical health screening, etc. At this time, the effectiveness of this project is still being evaluated; however, it provides one model to consider. The County hopes that State and Federal regulations may allow for more flexible implementation of projects such as this. R.2: SLO County should relieve the four private hospitals in our County of the responsibility for warehousing Held persons. The respondent notes that this item requires further analysis. The Health Agency has recently convened a new executive steering committee to develop collaborative approaches to address the impact of psychiatric crisis on the 4 local emergency departments. The Health Agency Interim Director will report on the progress of this committee in 6 months. The County fully recognizes the ongoing and statewide crisis in lack of appropriate resources for individuals experiencing psychiatric emergencies. This crisis is well documented; the attached report from the Rand Corporation provides detail. This crisis is not unique to SLO County. In fact, the length of stay for patients in local emergency departments is roughly on average with that of other hospitals in other parts of the State, even in counties where there are multiple privately operated psychiatric hospitals. The Health Agency complies with Federal civil rights laws and does not discriminate on the basis of race, color, national origin, age, disability, sex or any other protected class County of San Luis Obispo Health Agency 2180 Johnson Avenue | San Luis Obispo, CA 93401 | (P) 805-781-4719 | (F) 805-781-1273 6 of 52 slobehavioralhealth.org Two areas of concern have been the lack of psychiatric hospitals willing or able to accept individuals who test positive for Covid-19. These individuals have ended up remaining in emergency departments statewide for longer period of times than desired. The second area is the lack of pediatric psychiatric facilities, so youth have at times spent multiple days in emergency departments awaiting an appropriate placement. Neither of these situations are beneficial to the patients or their families, but are an unfortunate result of the lack of facilities either in SLO County or statewide to meet the specialized needs of these patients. Addressing this crisis effectively will require the full participation and engagement of multiple local entities, each with a distinct legal responsibility and area for potential impact. The County Health Agency has the responsibility to pay for inpatient treatment for County Medi-Cal beneficiaries. Beyond this legal responsibility, the County has also unilaterally increased services to provide care for individuals in psychiatric crisis. Hospitals with dedicated emergency departments have the responsibility to screen, stabilize, and transfer all individuals who present to their campuses. Responding to this crisis is a shared responsibility and the County is ready to continue working together toward a solution. R3. SLO County should seek the financial resources needed to hire and retain outpatient mental health services professional staff in sufficient number to allow for reasonable and customary management ratios. This recommendation requires further analysis. The Health Agency, in coordination with the County Human Resources Department, will continue to identify and pursue opportunities to recruit and retain qualified staff. This analysis will be followed up R4. SLO County should seek the financial resources needed to hire and retain mental health services professional staff in sufficient number to meet the needs of Held juveniles within our county. This recommendation requires further analysis. The responsibility of the County is to serve individuals who are Medi-Cal beneficiaries, or who are indigent. The responsibility of the Mental Health Plan, the contracted services between the County and the State Department of Health Care Services, is to pay for medically necessary inpatient care for all SLO County Medi-Cal beneficiaries. SLO County will continue to collaborate with other providers, both locally and statewide, to develop additional services for minors, including a continuum of crisis services. The Interim Health Agency Director has recently convened an Executive Steering Committee to seek collaborative partnerships to further develop resources for all community members requiring psychiatric care. R.5 The SLO County Sheriff’s Office, SLO County Behavioral Health Services, and the SLO County Board of Supervisors should jointly devise and implement a plan to ensure that properly trained and certified correctional officers are assigned in sufficient number to provide for the safety and security of all staff and held persons when such persons are in the County’s care and custody no matter which facility is responsible for the patient. This recommendation item requires further analysis. 2 7 of 52 At this time, the Health Agency contracts with a local security company that has been able to fully staff 2 FTEs of security 24/7 for the PHF and the Health Campus, as directed by the County CAO. Further discussion related to the Sheriff’s Office will be managed by the Interim Health Agency Director. 3 8 of 52 Evaluation Report CORPORATION RYAN K. MCBAIN, JONATHAN H. CANTOR, NICOLE K. EBERHART, SHREYA S. HUILGOL, INGRID ESTRADA-DARLEY Adult Psychiatric Bed Capacity, Need, and Shortage Estimates in California—2021 Summary Psychiatric beds are essential infrastructure for meeting the needs of individuals with mental health conditions. However, not all psychiatric beds are alike: They represent infrastructure within differ- ent types of facilities, ranging from acute psychiatric hospitals to community residential facilities. These facilities, in turn, serve clients with different needs: some who have high-acuity, short-term needs and others who have chronic, longer-term needs and KEY FINDINGS may return multiple times for care. ■ California faces an estimated 1.7-percent growth in its psychiatric bed California, like many parts need from 2021 to 2026. of the United States, is confront- ■ California faces shortages of psychiatric beds at all three major levels ing a shortage of psychiatric beds. of adult inpatient and residential care. This shortage manifests in high ■ Significant regional differences in the estimated shortfall of beds were bed occupancy rates and long wait noted at each level of care. lists for placements. However, determining the primary drivers ■ Growth in the need for psychiatric beds is projected to be largest in the Northern and Southern San Joaquin Valley. of this shortage—accounting for regional variation in psychiatric ■ Hard-to-place populations contribute disproportionately to bottle- bed capacity at different levels of necks in the existing system. care—is a challenging problem to ■ A majority of psychiatric facilities at all levels of care reported an tackle. Nevertheless, California inability to place individuals with comorbid dementia or traumatic brain is committed to expanding the injury, nonambulatory individuals, those requiring oxygen, and those mental health infrastructure, who tested positive for COVID-19. Individuals involved in the criminal justice system were reportedly difficult to place in community residen- including psychiatric bed capacity. tial settings. How, where, and to what extent 9 of 52 these investments should be made remains an open To estimate psychiatric bed need, we used several question. approaches for the purpose of triangulation. First, In this report, we estimated psychiatric bed we contacted psychiatric facilities throughout the capacity, need, and shortages for adults at each of state and spoke with administrative leaders at these three levels of care throughout California. These facilities to quantify bed occupancy rates, wait list three levels of care are acute, subacute, and commu- volume, average length of stay, and the number of nity residential services: individuals whom they would transfer to a higher or lower level of care if able to do so. Using the informa- • Acute care is directed toward those with the tion gathered, we were able to compute the number of highest acuity needs, is typically shorter term beds required—at each level of care in each region of (days to weeks), and is intended to stabilize the state—to reduce occupancy rates to 85 percent (a patients. standard ceiling) and accommodate wait list volume • Subacute care is directed toward those with and requested transfers. We calculated these estimates moderate- to high-acuity needs for a longer excluding state hospitals and, separately, including duration (multiple months). state hospitals, prioritizing the former approach. Our • Community residential services are intended rationale for this is that state hospital beds are gener- to address lower acuity and longer-term ally not considered part of the continuum of care at care (often multiple years) that is focused on a local level in terms of decisionmaking purposes. patient recovery. Second, we moderated this bottom-up estimate We computed these estimates with and without the by incorporating epidemiological information on inclusion of state hospitals, which often provide care regional variation in serious psychological distress for unique subpopulations who may be hard to place in (SPD) among adults, which serves as an indicator of other settings, including those with high acuity, long- psychiatric bed need. Third, as a top-down approach, term needs. Additionally, we projected growth in the we convened a Technical Expert Panel to deliberate need for psychiatric beds in the period of 2021 to 2026. and arrive at normative estimates of psychiatric bed need available from the research literature. Approach Lastly, we projected the need for psychiatric beds in the period from 2021 to 2026. To accomplish this, Our population of interest comprised adults (18 years we first quantified the prevalence of SPD according or older) throughout California. The corresponding to demographic categories (i.e., sex, race/ethnicity, sampling frame contained all psychiatric facilities and age group) among adults in California, using the with psychiatric beds serving adults throughout Cali- California Health Information Survey. From this, fornia’s 58 counties. Because individuals might access we were able to estimate the regional prevalence of psychiatric facilities (and beds) outside their county SPD in 2026, based on evolving demographic trends. of residence, we aggregated estimates at a regional Next, we cross-walked the estimated prevalence of level using the U.S. Census Bureau classification. SPD to the likelihood of requiring inpatient psychi- To estimate psychiatric bed capacity, we synthe- atric services, based on the proportional need for sized an array of data sets from state agencies that inpatient psychiatric services among individuals with are responsible for licensure of psychiatric beds. To versus without SPD, according to the National Survey supplement this information, we employed a strati- on Drug Use and Health (NSDUH). fied randomized sampling approach to administer a survey to collect data on the number of beds at facili- ties and the number of beds occupied. We provided Key Findings estimates to county points of contact at behavioral Psychiatric bed capacity. We estimated that Cali- health departments to review and revise them with fornia has a total of 5,975 beds at the acute level (19.5 an eye to improving accuracy. per 100,000 adults) and 4,724 at the subacute level (15.4 per 100,000 adults)—excluding state hospital 2 10 of 52 beds. If state hospital beds are included, these fig- ures increase to 7,679 (25.1 per 100,000 adults) and 9,168 beds (29.9 per 100,000 adults), respectively. We also observed large regional variation. For example, Growth in the need excluding state hospitals, acute bed capacity ranged for psychiatric beds is from 9.1 beds per 100,000 adults in the Northern San Joaquin Valley to 27.9 beds per 100,000 adults in the projected to be largest Superior region. For subacute bed capacity, regional estimates ranged from 7.4 to 31.8 beds per 100,000 in the Northern and adults. At the community residential level, we esti- mated that California has a total of 3,872 beds (12.7 Southern San Joaquin per 100,000 adults). Psychiatric bed need. Using observed occupancy Valley. rates, wait list volumes, and requested transfers, we estimated that California requires 50.5 inpatient psychiatric beds per 100,000 adults: 26.0 per 100,000 Psychiatric bed shortages. Synthesizing figures for at the acute level and 24.6 per 100,000 at the subacute bed capacity and bed need, we estimated that the state level, or 7,945 and 7,518 beds, respectively. At the has a shortfall of approximately 1,971 beds at the acute community residential level, we estimated a need of level (6.4 additional beds required per 100,000 adults) 22.3 beds per 100,000 adults. and a shortage of 2,796 beds at the subacute level (9.1 Estimated prevalence of SPD in California ranged additional beds required per 100,000 adults)—or 4,767 from 7.9 percent in the San Francisco Bay Area to subacute and acute beds combined, excluding state hos- 9.3 percent in the Southern San Joaquin Valley. When pital beds. If state hospitals were included in this esti- we incorporated this epidemiological information mate, the shortage of acute inpatient beds would shrink into our psychiatric bed need estimates, this intro- to 267, and there would be no observable shortage in duced regional variation in psychiatric bed need that beds at the subacute level. Separately, we estimated a ranged from 45.5 to 55.5 inpatient psychiatric beds per shortage of 2,963 community residential beds. 100,000 adults. Lastly, we collected secondary esti- The top-down estimates of psychiatric bed mates of psychiatric bed need from the academic lit- need—as drawn from the literature and our Techni- erature and our Technical Expert Panel. Using median cal Expert Panel—also indicated a bed shortage: 8.9 values, we generated a separate, top-down estimate beds per 100,000 adults at the acute level and 10.6 of psychiatric bed need: 27.5 beds per 100,000 adults beds per 100,000 adults at the subacute level. There- at the acute level and 25 per 100,000 at the subacute fore, our bottom-up and top-down estimates were level. We were unable to provide a comparable top- closely aligned. The remaining discrepancy likely down estimate of need for community residential beds pertains to differences in the configuration of health because of the significant heterogeneity within this systems throughout the United States and interna- classification and the paucity of academic literature. tionally, including availability of outpatient services We estimate that the magnitude of need for psychi- and alternatives to hospitalization, that drive need. atric beds is expected to grow modestly over the next When regional prevalence estimates for SPD were five years (2021 to 2026): by 1.7 percent. This is primar- incorporated, the gap in beds required reduced mod- ily due to shifting demographic trends, including adult estly: by 4.5 percent. We also documented significant population growth and increasing racial/ethnic diver- regional differences in the estimated shortfall of beds sity, because epidemiological data indicate that His- on the basis of the wide regional variation in psychiat- panic and Black adults experience SPD at higher rates ric bed capacity. For example, two regions of the state than do White adults. Growth in the need for psychi- appear to have sufficient acute inpatient psychiatric atric beds is projected to be largest in the Northern and bed capacity, whereas the remaining eight regions Southern San Joaquin Valley—by about 4.0 percent. 11 of 52 3 have a shortfall. At the subacute level, all regions such regions as the Central Coast, Inland (apart from the Northern San Joaquin Valley) appear Empire, and Southern San Joaquin Valley. to have a shortfall. However, the magnitude of this 2. Consider focusing on building or remodel- shortfall ranges from 5.1 additional beds required per ing infrastructure for the most hard-to-place 100,000 adults in the North Coast region of the state populations. Specific subpopulations appear to 17.2 additional beds required per 100,000 adults in to contribute disproportionately to bottle- the Southern San Joaquin Valley. necks in the current system, including an Lastly, we inquired about hard-to-place popula- inability to transfer patients with criminal tions. Here, we found that a majority of psychiatric justice involvement from the subacute level of facilities at all levels of care reported an inability to care to community residential settings. Given place individuals with comorbid dementia or trau- this, the state might need to consider alterna- matic brain injury, nonambulatory individuals, those tive arrangements for placing such popula- requiring oxygen, and those who tested positive for the tions, such as community-based and outpa- coronavirus disease 2019 (COVID-19). A majority of tient competency restoration programs. Here, respondents from community residential facilities also California could learn from other mental reported an inability to place individuals involved in health systems across the United States and the criminal justice system—particularly those with internationally. arson or sex offense convictions. 3. Set aside state funds for a system that reviews licensure data and periodically collects psychiatric facility–level information. Our Recommendations analysis and conclusions contain numerous Using these findings, we came up with three caveats, in large part because of poor data recommendations: quality. We wish to be transparent about this fact, with the hope that this serves as an impe- 1. Prioritize psychiatric bed infrastructure in tus for the state to consider investing in an the areas with the greatest need. In terms of adequate data review and monitoring system. an absolute shortfall of beds, the shortfall If the state were to allocate funds to routinely was greatest in terms of subacute beds, driven monitor and purge licensure data, policymak- partly by four regions (Los Angeles County, ers would be in a much stronger position to San Francisco Bay Area, Inland Empire, Supe- know what the existing capacity is at each rior region) that represented a shortfall of level of care—particularly at the community more than 2,000 beds—more than a quarter residential level. Likewise, the state should of all additional beds needed throughout the consider establishing a mechanism by which state. If policymakers examine the psychiatric psychiatric facilities report periodically on bed bed shortfall as a proportion of regional adult occupancy rates, wait list volume, number of population, this might lend greater weight to requested transfers to higher and lower levels regions with smaller or more rural popula- of care, and psychiatric patient boarding in tions: For example, the shortfall of subacute emergency departments. The state should also beds is 5.2 beds per 100,000 adults in Los consider collecting sociodemographic and Angeles County compared with 17.2 per clinical information on patients who use psy- 100,000 adults in the Southern San Joaquin chiatric beds. This would allow California to Valley. We also observed significant need for have a remarkably precise and sensitive system acute beds in such regions as the Northern for tracking the impact of investments that and Southern San Joaquin Valley and Central seek to address psychiatric bed shortages. Coast, while the shortfall at the community residential level was particularly notable in 4 12 of 52 Introduction Role of Psychiatric Beds Today, hundreds of Psychiatric beds are essential infrastructure for Californians in need of meeting the needs of individuals with serious mental health conditions.1 These beds serve several psychiatric beds are functions—including enabling safe, stable, and sup- held in hospital EDs or portive environments for individuals in acute mental health crises and for those with significant impair- county jails awaiting ment who require ongoing medical monitoring.2 Not all psychiatric beds are alike, because they openings in inpatient represent infrastructure within different types of facilities. For example, psychiatric beds in acute inpa- care settings. tient hospitals serve those in need of secure, 24-hour care and often include crisis stabilization units. The average length of stay in such states as California is resolution and triage services, acute inpatient ser- one to two weeks.3 By contrast, psychiatric beds in vices, subacute services, state hospitals, and com- such subacute facilities as mental health rehabilita- munity residential services. Clinical guidelines, such tion centers (MHRCs) or special treatment programs as the Level of Care Utilization System, enforced by at skilled nursing facilities (SNFs)—which provide the state of California through Senate Bill (SB) 855 longer-term recovery-oriented services, such as inde- (2020),8 provide a useful compass for matching an pendent life skills training—may remain occupied by individual’s needs with an appropriate level of care. the same individuals for many months.4 Ultimately, psychiatric beds represent an impor- tant component in a continuum of behavioral health Psychiatric Bed Shortfall care that includes integrated community services Psychiatric bed capacity is severely strained in Cali- ranging from prevention and screening to emergency fornia, as it is in much of the United States.9 The crisis response. Depending on the arrangement of present situation may be viewed, in part, as the long services, the need for psychiatric beds may look dif- tail of an effort to deinstitutionalize psychiatric ser- ferent. For example, the CrisisNow model of emer- vices throughout the United States during the latter gency care instituted in Arizona has significantly half of the 20th century: from a peak of 337 psy- reduced the state’s volume of psychiatric emergency chiatric beds per 100,000 individuals in the United department (ED) boarding.5 Assertive Community States in 1955 to a low of around 12 beds per 100,000 Treatment and similar models, such as Full Service in 2016.10 This transition to community-based Partnership programs, may also be key contribu- services—although well-intentioned—has resulted in tors. These models employ transdisciplinary teams a paucity of infrastructure to serve the needs of indi- to provide comprehensive services to patients who viduals who would otherwise benefit from a stable have needs that have not been adequately met by and supervised residence, particularly those with traditional approaches.6 According to the Substance serious mental illness (SMI).11 Abuse and Mental Health Services’ Behavioral Today, hundreds of Californians in need of psy- Health Treatment Services Locator, there are 133 chiatric beds are held in hospital EDs or county jails mental health treatment facilities in California that awaiting openings in inpatient care settings.12 In offer these services, though these services may be addition, a sizable percentage of chronically home- underreported.7 less individuals have an SMI.13 As county jail and With regard to psychiatric beds specifically, homeless populations continue to swell, these rising California embeds psychiatric beds within crisis numbers have created an increasing urgency to take 13 of 52 5 Estimating the need for psychiatric beds is a thorny undertaking. There are no standardized Given the diversity and approaches or best practices, and health systems geographic distribution are constructed differently at the regional and state levels. However, prior literature outlines at least of both adults and four methods for calculating the number of psy- chiatric beds that are needed to address population psychiatric facilities needs. These methods include (1) expert consensus, (2)a normative approach, (3) a population health in California, needs approach, and (4) an observed outcomes approach.16 We briefly survey these approaches, including their throughout the state are strengths and limitations. Expert consensus. Calculating psychiatric bed heterogeneous. need by expert consensus relies on open discussion among content and methods experts—who opera- tionally define a set of relevant principles, deliberate action. For example, in early 2021, an estimated 1,600 evidence, and then achieve mutual agreement on adults in need of psychiatric beds were residing in standards. A key example of this approach is a 2008 county jails because they had been deemed incompe- report by the Treatment Advocacy Center (TAC),17 tent to stand trial and were unable to be placed by the which interviewed 15 experts to deliberate and arrive Department of State Hospitals (DSH).14 Legislators, at a suggested benchmark for measuring psychiatric meanwhile, have called for an overhaul of state psy- bed need: 40 to 60 beds per 100,000 in population.18 chiatric services—particularly against the backdrop of Although expert consensus is a helpful method the COVID-19 pandemic.15 for arriving at an estimate of psychiatric bed need, this approach has limitations. The report by TAC did Measuring Need for Psychiatric Beds not outline the precise deliberations that led experts to arrive at their conclusion, making it challenging Estimating the need for psychiatric beds in Califor- to scrutinize the estimate or replicate the process. nia is essential for at least three reasons. First, when Furthermore, without a set of operational definitions estimates on the need for psychiatric beds are paired for such terms as psychiatric bed, it remains unclear with estimates on the existing bed capacity, evalu- whether a specific benchmark would translate in other ators can determine the magnitude of the shortfall settings or how one would go about allocating 40 to 60 in beds throughout the state, allowing policymak- beds per 100,000 in population across different levels ers to discuss investments accordingly. Second, the of care. The number and types of beds required are estimation process can serve a diagnostic function: liable to depend on local context and resources. Given the diversity and geographic distribution Normative approach. The normative approach of both adults and psychiatric facilities in Califor- to calculating psychiatric bed need is predicated nia, needs throughout the state are heterogeneous. on an assumption that jurisdictions (or countries) Lastly, if the need for psychiatric beds is estimated with similar health systems and demographic char- at multiple intervals, these estimates can provide an acteristics are likely to require a similar number ongoing feedback mechanism for fine-tuning invest- of psychiatric beds. In this scenario, a jurisdiction ments. Because needs are dynamic, investments in with more-robust psychiatric facility infrastruc- infrastructure at any given time may only partially ture has the potential to serve as a comparator for address problems. This factor should create an impe- others. For example, the Organisation for Economic tus to determine how effective initial investments Co-operation and Development has employed this were and to plan for new investments for new needs approach to compare mental health infrastructure have arisen. across its 38 member countries, providing annual- 6 14 of 52 ized estimates of psychiatric beds per 100,000 in for example, the Department of Health used popu- population—with a low of 3 beds per 100,000 in lation and demographic estimates to calculate the population in Mexico to 259 per 100,000 in popula- total psychiatric bed need as 30 beds per 100,000 in tion in Japan (the United States ranks seventh from population.23 Other states (Mississippi, Oklahoma, the bottom at 25 per 100,000).19 The World Health Missouri) have performed similar calculations, using Organization has followed a similar model with its a fixed population ratio to arrive at estimates of need Mental Health Atlas project.20 Coupled with other ranging from 20 to 117 psychiatric beds per 100,000 indicators (e.g., hospital readmission rates), compara- in population.24 This wide range in values is indica- tive analyses can be used to determine whether and tive of disconcerting variation in the parameters used to what extent more-robust infrastructure translates to come up with these estimates. to improved population health outcomes. Calculating the number of psychiatric beds The main challenge with the normative approach needed from this population health approach assumes is that there are large differences in the number of certain targets for specific mental health conditions, psychiatric beds per capita even within countries which may be more or less accurate. For example, that have similar health and economic systems. Fur- there is a paucity of epidemiological data on the likeli- thermore, understanding the difference in reported hood of an individual requiring specific services based numbers of psychiatric beds is difficult because there on mental health diagnosis and illness severity, and on is no standard definition for psychiatric bed. Coun- the optimal duration and intensity of service provi- tries like Italy do not consider residential treatment sion. Relying on limited information risks generating facilities to constitute inpatient care, producing a inaccurate predictions for the number of psychiatric much lower estimate per 100,000 in population com- beds needed and may underpin the wide variation pared with countries that include residential treat- observed in estimates across states—which are often ment facilities.21 Additionally, what works for some using differing sets of assumptions. countries and systems in terms of psychiatric bed Observed outcomes approach. The observed needs and mental health services may not work in outcomes approach is based on the observation that others. Individual countries may first need to define heterogeneous psychiatric bed capacity across coun- their core mental health services and set data-driven ties and states is liable to have observable effects on targets to meet those needs. One recent study found health systems and populations.25 By looking at the that, out of 32 mental health plans developed across relationship between psychiatric bed capacity and key five countries, only four plans included specific tar- performance indicators—such as wait times, occu- gets for their core services—including psychiatric pancy rates, length of stay, emergency room board- bed needs.22 These sorts of comparative metrics are ing, and population health outcomes—researchers also generally lacking within the United States across have the potential to calculate minimum and optimal regions and counties. psychiatric bed capacity requirements from an induc- Population health approach. A population tive perspective.26 Along these lines, in 2020, the San health approach identifies the prevalence of mental Francisco Department of Public Health conducted a health conditions within a geographic area and simulation to assess psychiatric bed need, which ana- then applies a set of standards to meet population lyzed more than 25,000 mental health–related admis- health needs in accordance with these prevalence sions and accounted for variable bed occupancy rates estimates—including quantifying psychiatric services across different levels of care.27 The authors of the and corresponding infrastructure, such as psychi- report concluded that observed bed occupancy rates atric beds. One advantage of this approach is that greater than 85 percent have the potential to contrib- prevalence estimates for particular mental health ute to bottlenecks and flow issues over the long run conditions (and, by extension, needs that derive and suggested an additional 97 beds across four types from these conditions) are sensitive to the underly- of facilities to achieve zero wait time. ing demographic characteristics of the region—such The observed outcomes approach has also been as age, sex, and income distributions. In Tennessee, criticized.28 Specifically, researchers have argued that 15 of 52 7 key performance indicators tend to concentrate on bill’s provisions allow local governments to purchase process measures pertaining to hospital administra- psychiatric facilities to prevent closures and require that tion, which may have little correspondence to patient facility owners give residents greater advanced notice outcomes. Furthermore, hospital functioning may be prior to closure. Several complementary bills signed into dependent on a wide array of factors, such as employ- law earlier this year, including AB 27, AB 362, AB 816, ment rates or social determinants of health within AB 977, AB 1220, AB 1443, and SB 400, represent part the local community. Therefore, these extraneous of a $22 billion investment to address homelessness and factors may function as confounders during analysis the need for behavioral health services. The investments unless they are incorporated as covariates. include $3 billion dedicated to housing for those with Ultimately, as noted earlier in this report, there are acute behavioral and physical health issues—with an no consensus best practices for determining psychiat- expectation of creating approximately 22,000 new beds ric bed needs. The most robust approach may therefore and treatment slots.33 be to inspect the problem from multiple vantage points Additional Assembly legislation, AB 2265, has and methodologies to converge on a triangulated set of focused on increasing access to treatment by allow- estimates for psychiatric bed needs. This includes—to ing California counties to use mental health services the extent possible—pressure testing the approaches funds to address not only mental health conditions outlined above by performing sensitivity analyses. but also substance use disorders, with the goals of We have therefore elected to assume this triangulated enhancing care coordination and creating an inte- approach in our analysis, as further described in the grated behavioral health care system.34 In the context “Methods” section. of psychiatric bed infrastructure, this legislation is particularly relevant for populations in need of longer-term rehabilitative services who are coping California’s Investment in Infrastructure with comorbid substance use and mental health con- California has stated a commitment to expand mental ditions. It was signed into law on September 24, 2020. health infrastructure, including psychiatric bed capacity.29 In early 2020, the California Mental Health Purpose of This Report Services Authority (CalMHSA) announced the forma- tion of a Behavioral Health Task Force, appointed to In this report, we provide an estimate of current advise Governor Newsom on efforts to reform and psychiatric bed capacity throughout the ten census advance behavioral health services throughout the regions of California, according to three overarching state.30 Since the onset of the COVID-19 pandemic levels of care: acute inpatient care, subacute inpa- in 2020, Newsom has signed numerous bills into law tient care, and community residential treatment. We that aim to increase Californians’ access to mental then compare the measure of current psychiatric health services. As noted earlier in this report, these bed capacity with estimates of psychiatric bed need. bills include SB 855, which requires commercial health Lastly, we project bed capacity needs over the next insurance plans outside Medi-Cal to provide medically five years, based on evolving demographic trends necessary treatments for all mental health conditions throughout the state. and substance use disorders.31 The bill also requires We note that the primary bed estimates provided that health plans provide services that comply with in this report do not include state hospitals, although level of care determinations as outlined in the Level of we provide secondary estimates for which state hos- Care Utilization System, American Society of Addic- pitals are included. This decision was based on three tion Medicine criteria, and other clinical guidelines for factors. First, more than 90 percent of psychiatric pediatric populations. beds at state hospitals are occupied by individuals Separately, Assembly Bill (AB) 2377—which was involved in the criminal justice system.35 Thus, state also ratified in 2020—sought to mitigate the impact of hospitals serve a set of clients with unique constraints adult residential facility closures and help residents at that do not apply to local behavioral health con- risk of homelessness in California.32 Specifically, the tinuums that we examine in detail throughout this 8 16 of 52 report. In addition, the forensic population census ization of these facilities and portfolio of services may has continued an upward trend that portends a con- vary. Given this factor, we assigned facilities to three tinued reduction in beds available for non-forensic levels of care—matching closely to a conceptual model individuals. Second, the exclusion of state hospitals established by the County Behavioral Health Directors has the secondary benefit of modeling need for psy- Association of California.37 The levels were defined by chiatric beds if state hospitals were transitioned to two axes: first, the acuity of need being attended to, local, community-based alternatives. Third, unlike ranging from emergent crises to nonemergent, ongo- other types of facilities, state hospitals are not broadly ing supports; second, typical length of stay, ranging distributed throughout the state. This distribution from short term (days to weeks) to long term (months results in computational challenges for determining to years). With the exception of community residen- psychiatric bed shortages in regions that contain state tial facilities, length of stay is usually time-delimited hospitals, especially because not all patients within according to the particular type of facility. state hospitals are residents from the region in which Operationally, we defined the three levels of care that state hospital is located. We discuss further as follows: details in the “Methods” section. 1. acute, representing highly structured, around- We supplement these quantitative analyses with the-clock medically monitored inpatient care input from our panel of technical experts. Using the for individuals at heightened risk of harm to combined results, we outline a series of recommenda- themselves or others, or those who are other- tions. These recommendations pertain to the expan- wise unable to care for themselves sion of psychiatric bed capacity to address existing 2. subacute, representing around-the-clock gaps, and they are situated in California’s context of inpatient care that includes specialized pro- ongoing legislative efforts to establish a holistic con- gramming in a controlled environment with a tinuum of behavioral health care services. significant degree of supervision but with less intensive medical monitoring and interven- tion than acute care Methods 3. residential, representing nonhospital pro- Population and Scope grams in which individuals live on the prem- ises of a facility and are provided with consis- Our population of interest comprised all adults (18 tent programming to promote interpersonal years or older) in California, across all 58 counties. and independent living skills, with staff pres- Because individuals may access psychiatric facili- ent 24 hours a day, seven days a week. ties (and beds) outside their county of residence, we Table 1 details the types of facilities contained aggregated estimates of population, capacity, and within each of these levels of care, based on facility need at a regional level using the U.S. Census Bureau licensure information. classification: Superior California, North Coast, San State hospitals represent a unique type of institu- Francisco Bay Area, Northern San Joaquin Valley, tion. In other settings, acute care is short term and Central Coast, Southern San Joaquin Valley, Inland focused on stabilizing patients, whereas residential Empire, Los Angeles County, Orange County, and San Diego-Imperial.36 Although our denominator for care is lower intensity and geared toward long-term medical and nonmedical supports.38 In this sense, the calculations in this report included all adults in the two axes described above (acuity and length of California—because all adults have the potential to stay) are aligned. The expectation is that patients may use inpatient psychiatric services—a disproportion- transition up and down the care continuum based ate number of those using psychiatric beds are adults on their needs at a given time point. However, state with SMIs, such as schizophrenia, bipolar disorder, hospitals are less dynamic and often provide care for and major depressive disorder. subcategories of patients with long-term, high acuity There are many types of psychiatric facilities in needs or based on medical necessity. For example, California. Depending on the county, the character- 17 of 52 9 TABLE 1 individuals who are incompetent to stand trial and are Levels of Care and Corresponding Adult awaiting transfer to an alternative care setting such Psychiatric Bed Infrastructure as a state hospital.42 Our rationale for this was that, although jail units are not suitable to serve as psychiat- Level of Care Types of Facilities Included ric beds (and therefore should not factor into capacity), Acute Acute psychiatric hospitals; psychiatric they are nevertheless housing individuals in need of (Level 3) health facilities; general acute care psychiatric beds. This was not possible for ED board- hospitals with psychiatric wards; acute beds at state hospitals ing. Although California generates an annual hospital utilization report, these reports do not contain ED Subacute General or specialized subacute facilities; (Level 2) MHRCs; SNFs with specialized treatment boarding rates among patients with mental health con- programs; institutions for mental disease; ditions.43 Likewise, there are no formal tallies of the subacute beds at state hospitals number of individuals with mental health conditions Residential Adult residential treatment facilities; who are receiving permanent supportive housing and (Level 1) enhanced or augmented board-and-care facilities; social rehabilitation facilities would otherwise benefit from alternative placement in a setting with psychiatric beds. NOTE: For a definition of psychiatric health facilities, see California De- partment of Health Care Services, 2021b. For a definition of institutions for mental disease, see California Department of Health Care Services, 2021c. Procedures state hospitals often house forensic patients (i.e., Focus group discussions. As a preliminary step, those involved with the criminal justice system), and, we conducted focus group discussions with county in California, many state hospital beds are reserved leaders at behavioral health departments throughout for such patients.39 This raises two questions: Should the state and members at CalMHSA and the County state hospitals be placed within the care continuum; Behavioral Health Directors Association. These and, if so, where? For the purposes of this report, we discussions focused on conceptual issues that these have remained agnostic to the first question, com- individuals were confronting with regard to sup- puting capacity and bed shortages with and without porting psychiatric bed needs. We focused on four inclusion of state hospitals. For the latter, we allo- topics: (1) perceived structural drivers of psychiatric cated acute state hospital beds to the acute level and bed shortages, (2) populations who were challenging subacute beds to the subacute level, both of which are to place in psychiatric facilities, (3) defining the care documented in licensure data sets.40 continuum, including which types of facilities cor- Specific populations and bed categories were respond to which levels of care, and (4) other areas deemed to be outside the scope of our analysis. of note that would be important for RAND research- Regarding populations, we excluded children and ado- ers to consider. We took detailed notes from each lescents, for whom there are subtle but important dif- discussion and deliberated the feedback provided to ferences in the care continuum and who were therefore us to develop our methodological approach, which is deemed to merit a separate analysis.41 Regarding bed detailed below. categories, we omitted beds corresponding to perma- Estimation of capacity. We downloaded the nent supportive housing, those in county jails, and most current licensure data available for each type those in EDs used for boarding patients with mental of psychiatric facility, using public data sets from health conditions. These categories are seldom quanti- the California Department of Public Health,44 Cali- fied as psychiatric beds because they are not exclu- fornia Department of State Hospitals,45 California sively reserved for populations with mental health Department of Health Care Services,46 and Cali- conditions, though it may be the case that individuals fornia Department of Social Services.47 These data with mental health conditions occupy one of these bed sets were then merged into a master file of facilities types. with psychiatric beds in California, and each facility For quantification of psychiatric bed need, we was geocoded at the address and county levels using incorporated the number of jail units occupied by ArcGIS Desktop 10.8.48 10 18 of 52 To validate facility licensure data, we executed need. In total, the Technical Expert Panel consisted two additional steps. First, we contacted behavioral of four participants whose names and titles can be health directors in all 58 counties in October 2021, found in Appendix B, alongside all major prompts providing them with an inventory of facilities within used during the panel’s discussion. their county and soliciting revisions. Where discrep- Estimation of need, Approach 3: population ancies arose, we prioritized the revisions detailed health assessment. Lastly, we drew from epide- by the county point of contact. Second, we reviewed miological data reported in the California Health 2,500 online entries for community residential Information Survey (CHIS) and NSDUH.50 Both the facilities to estimate the percentage of these facilities CHIS and NSDUH employ the Kessler 6, which is a within each county that provided services to indi- measure of psychological distress.51 A score of 13 or viduals with mental health conditions rather than greater indicates SPD, which is a marker of probable those with other types of disabilities or the elderly. SMI. This allowed us to rebalance regional need for For example, if we queried 60 community residential psychiatric beds based on (1) estimated regional prev- facilities in Orange County and found that 5 of 60 alence of SPD among adults, according to the CHIS, were psychiatric facilities with psychiatric beds, we and (2) expected utilization of psychiatric inpatient inferred that this percentage (5/60 = 8.3%) applied services among adults with versus without SPD in at the population level—that is, that 8.3 percent of the United States, based on NSDUH (which contains Orange County’s community residential facilities and estimates of self-reported receipt of inpatient mental beds were psychiatric facilities that contained psychi- health services in the past year).52 A substantial body atric beds. of literature has shown that SPD is a correlate of SMI Estimation of need, Approach 1: survey of and that those with SPD have a greater need for both psychiatric facilities. Using the observed outcomes outpatient and inpatient services.53 approach described earlier in this report, we drew a Projection of need. In addition to estimat- random sample of psychiatric facilities throughout ing current need for psychiatric beds, we projected California using Stata’s v.17 runiform command49 change in need over the next five years (2021 to and attempted to contact facility directors to inquire 2026). These estimates drew from expected trends about bed occupancy, average length of stay, wait in population growth and demographic shifts in the list volume, and the number of patients that facility age, sex, and racial/ethnic composition of the state directors recommended for transfer to a higher or according to U.S. Census Bureau information. The lower level of care. In the event that we were unable “Analysis” section provides a fuller description of the to reach an administrative leader at a particular facil- computations involved in this. ity, we made up to four additional attempted con- tacts per facility. For facilities that we were unable to Measures reach, we imputed estimates based on median values within the facility type and county, adjusting for total Psychiatric beds. We defined psychiatric beds as number of beds. Calls were made in October and beds within psychiatric facilities that have the pri- November 2021.

R2

SLO County should relieve the four private hospitals in our County of the responsibility for warehousing Held persons. The respondent notes that this item requires further analysis. The Health Agency has recently convened a new executive steering committee to develop collaborative approaches to address the impact of psychiatric crisis on the 4 local emergency departments. The Health Agency Interim Director will report on the progress of this committee in 6 months. The County fully recognizes the ongoing and statewide crisis in lack of appropriate resources for individuals experiencing psychiatric emergencies. This crisis is well documented; the attached report from the Rand Corporation provides detail. This crisis is not unique to SLO County. In fact, the length of stay for patients in local emergency departments is roughly on average with that of other hospitals in other parts of the State, even in counties where there are multiple privately operated psychiatric hospitals. The Health Agency complies with Federal civil rights laws and does not discriminate on the basis of race, color, national origin, age, disability, sex or any other protected class County of San Luis Obispo Health Agency 2180 Johnson Avenue | San Luis Obispo, CA 93401 | (P) 805-781-4719 | (F) 805-781-1273 6 of 52 slobehavioralhealth.org Two areas of concern have been the lack of psychiatric hospitals willing or able to accept individuals who test positive for Covid-19. These individuals have ended up remaining in emergency departments statewide for longer period of times than desired. The second area is the lack of pediatric psychiatric facilities, so youth have at times spent multiple days in emergency departments awaiting an appropriate placement. Neither of these situations are beneficial to the patients or their families, but are an unfortunate result of the lack of facilities either in SLO County or statewide to meet the specialized needs of these patients. Addressing this crisis effectively will require the full participation and engagement of multiple local entities, each with a distinct legal responsibility and area for potential impact. The County Health Agency has the responsibility to pay for inpatient treatment for County Medi-Cal beneficiaries. Beyond this legal responsibility, the County has also unilaterally increased services to provide care for individuals in psychiatric crisis. Hospitals with dedicated emergency departments have the responsibility to screen, stabilize, and transfer all individuals who present to their campuses. Responding to this crisis is a shared responsibility and the County is ready to continue working together toward a solution. R3. SLO County should seek the financial resources needed to hire and retain outpatient mental health services professional staff in sufficient number to allow for reasonable and customary management ratios. This recommendation requires further analysis. The Health Agency, in coordination with the County Human Resources Department, will continue to identify and pursue opportunities to recruit and retain qualified staff. This analysis will be followed up R4. SLO County should seek the financial resources needed to hire and retain mental health services professional staff in sufficient number to meet the needs of Held juveniles within our county. This recommendation requires further analysis. The responsibility of the County is to serve individuals who are Medi-Cal beneficiaries, or who are indigent. The responsibility of the Mental Health Plan, the contracted services between the County and the State Department of Health Care Services, is to pay for medically necessary inpatient care for all SLO County Medi-Cal beneficiaries. SLO County will continue to collaborate with other providers, both locally and statewide, to develop additional services for minors, including a continuum of crisis services. The Interim Health Agency Director has recently convened an Executive Steering Committee to seek collaborative partnerships to further develop resources for all community members requiring psychiatric care.

R3

SLO County should seek the financial resources needed to hire and retain outpatient mental health services professional staff in sufficient number to allow for reasonable and customary management ratios. This recommendation requires further analysis. The Health Agency, in coordination with the County Human Resources Department, will continue to identify and pursue opportunities to recruit and retain qualified staff. This analysis will be followed up

R4

SLO County should seek the financial resources needed to hire and retain mental health services professional staff in sufficient number to meet the needs of Held juveniles within our county. This recommendation requires further analysis. The responsibility of the County is to serve individuals who are Medi-Cal beneficiaries, or who are indigent. The responsibility of the Mental Health Plan, the contracted services between the County and the State Department of Health Care Services, is to pay for medically necessary inpatient care for all SLO County Medi-Cal beneficiaries. SLO County will continue to collaborate with other providers, both locally and statewide, to develop additional services for minors, including a continuum of crisis services. The Interim Health Agency Director has recently convened an Executive Steering Committee to seek collaborative partnerships to further develop resources for all community members requiring psychiatric care. R.5 The SLO County Sheriff’s Office, SLO County Behavioral Health Services, and the SLO County Board of Supervisors should jointly devise and implement a plan to ensure that properly trained and certified correctional officers are assigned in sufficient number to provide for the safety and security of all staff and held persons when such persons are in the County’s care and custody no matter which facility is responsible for the patient. This recommendation item requires further analysis. 2 7 of 52 At this time, the Health Agency contracts with a local security company that has been able to fully staff 2 FTEs of security 24/7 for the PHF and the Health Campus, as directed by the County CAO. Further discussion related to the Sheriff’s Office will be managed by the Interim Health Agency Director. 3 8 of 52 Evaluation Report CORPORATION RYAN K. MCBAIN, JONATHAN H. CANTOR, NICOLE K. EBERHART, SHREYA S. HUILGOL, INGRID ESTRADA-DARLEY Adult Psychiatric Bed Capacity, Need, and Shortage Estimates in California—2021

R5

The SLO County Sheriff’s Office, SLO County Behavioral Health Services, and the SLO County Board of Supervisors should jointly devise and implement a plan to ensure that properly trained and certified correctional officers are assigned in sufficient number to provide for the safety and security of all staff and held persons when such persons are in the County’s care and custody no matter which facility is responsible for the patient. This recommendation item requires further analysis. 2 7 of 52 At this time, the Health Agency contracts with a local security company that has been able to fully staff 2 FTEs of security 24/7 for the PHF and the Health Campus, as directed by the County CAO. Further discussion related to the Sheriff’s Office will be managed by the Interim Health Agency Director. 3 8 of 52 Evaluation Report CORPORATION RYAN K. MCBAIN, JONATHAN H. CANTOR, NICOLE K. EBERHART, SHREYA S. HUILGOL, INGRID ESTRADA-DARLEY Adult Psychiatric Bed Capacity, Need, and Shortage Estimates in California—2021 Summary Psychiatric beds are essential infrastructure for meeting the needs of individuals with mental health conditions. However, not all psychiatric beds are alike: They represent infrastructure within differ- ent types of facilities, ranging from acute psychiatric hospitals to community residential facilities. These facilities, in turn, serve clients with different needs: some who have high-acuity, short-term needs and others who have chronic, longer-term needs and KEY FINDINGS may return multiple times for care. ■ California faces an estimated 1.7-percent growth in its psychiatric bed California, like many parts need from 2021 to 2026. of the United States, is confront- ■ California faces shortages of psychiatric beds at all three major levels ing a shortage of psychiatric beds. of adult inpatient and residential care. This shortage manifests in high ■ Significant regional differences in the estimated shortfall of beds were bed occupancy rates and long wait noted at each level of care. lists for placements. However, determining the primary drivers ■ Growth in the need for psychiatric beds is projected to be largest in the Northern and Southern San Joaquin Valley. of this shortage—accounting for regional variation in psychiatric ■ Hard-to-place populations contribute disproportionately to bottle- bed capacity at different levels of necks in the existing system. care—is a challenging problem to ■ A majority of psychiatric facilities at all levels of care reported an tackle. Nevertheless, California inability to place individuals with comorbid dementia or traumatic brain is committed to expanding the injury, nonambulatory individuals, those requiring oxygen, and those mental health infrastructure, who tested positive for COVID-19. Individuals involved in the criminal justice system were reportedly difficult to place in community residen- including psychiatric bed capacity. tial settings. How, where, and to what extent 9 of 52 these investments should be made remains an open To estimate psychiatric bed need, we used several question. approaches for the purpose of triangulation. First, In this report, we estimated psychiatric bed we contacted psychiatric facilities throughout the capacity, need, and shortages for adults at each of state and spoke with administrative leaders at these three levels of care throughout California. These facilities to quantify bed occupancy rates, wait list three levels of care are acute, subacute, and commu- volume, average length of stay, and the number of nity residential services: individuals whom they would transfer to a higher or lower level of care if able to do so. Using the informa- • Acute care is directed toward those with the tion gathered, we were able to compute the number of highest acuity needs, is typically shorter term beds required—at each level of care in each region of (days to weeks), and is intended to stabilize the state—to reduce occupancy rates to 85 percent (a patients. standard ceiling) and accommodate wait list volume • Subacute care is directed toward those with and requested transfers. We calculated these estimates moderate- to high-acuity needs for a longer excluding state hospitals and, separately, including duration (multiple months). state hospitals, prioritizing the former approach. Our • Community residential services are intended rationale for this is that state hospital beds are gener- to address lower acuity and longer-term ally not considered part of the continuum of care at care (often multiple years) that is focused on a local level in terms of decisionmaking purposes. patient recovery. Second, we moderated this bottom-up estimate We computed these estimates with and without the by incorporating epidemiological information on inclusion of state hospitals, which often provide care regional variation in serious psychological distress for unique subpopulations who may be hard to place in (SPD) among adults, which serves as an indicator of other settings, including those with high acuity, long- psychiatric bed need. Third, as a top-down approach, term needs. Additionally, we projected growth in the we convened a Technical Expert Panel to deliberate need for psychiatric beds in the period of 2021 to 2026. and arrive at normative estimates of psychiatric bed need available from the research literature. Approach Lastly, we projected the need for psychiatric beds in the period from 2021 to 2026. To accomplish this, Our population of interest comprised adults (18 years we first quantified the prevalence of SPD according or older) throughout California. The corresponding to demographic categories (i.e., sex, race/ethnicity, sampling frame contained all psychiatric facilities and age group) among adults in California, using the with psychiatric beds serving adults throughout Cali- California Health Information Survey. From this, fornia’s 58 counties. Because individuals might access we were able to estimate the regional prevalence of psychiatric facilities (and beds) outside their county SPD in 2026, based on evolving demographic trends. of residence, we aggregated estimates at a regional Next, we cross-walked the estimated prevalence of level using the U.S. Census Bureau classification. SPD to the likelihood of requiring inpatient psychi- To estimate psychiatric bed capacity, we synthe- atric services, based on the proportional need for sized an array of data sets from state agencies that inpatient psychiatric services among individuals with are responsible for licensure of psychiatric beds. To versus without SPD, according to the National Survey supplement this information, we employed a strati- on Drug Use and Health (NSDUH). fied randomized sampling approach to administer a survey to collect data on the number of beds at facili- ties and the number of beds occupied. We provided Key Findings estimates to county points of contact at behavioral Psychiatric bed capacity. We estimated that Cali- health departments to review and revise them with fornia has a total of 5,975 beds at the acute level (19.5 an eye to improving accuracy. per 100,000 adults) and 4,724 at the subacute level (15.4 per 100,000 adults)—excluding state hospital 2 10 of 52 beds. If state hospital beds are included, these fig- ures increase to 7,679 (25.1 per 100,000 adults) and 9,168 beds (29.9 per 100,000 adults), respectively. We also observed large regional variation. For example, Growth in the need excluding state hospitals, acute bed capacity ranged for psychiatric beds is from 9.1 beds per 100,000 adults in the Northern San Joaquin Valley to 27.9 beds per 100,000 adults in the projected to be largest Superior region. For subacute bed capacity, regional estimates ranged from 7.4 to 31.8 beds per 100,000 in the Northern and adults. At the community residential level, we esti- mated that California has a total of 3,872 beds (12.7 Southern San Joaquin per 100,000 adults). Psychiatric bed need. Using observed occupancy Valley. rates, wait list volumes, and requested transfers, we estimated that California requires 50.5 inpatient psychiatric beds per 100,000 adults: 26.0 per 100,000 Psychiatric bed shortages. Synthesizing figures for at the acute level and 24.6 per 100,000 at the subacute bed capacity and bed need, we estimated that the state level, or 7,945 and 7,518 beds, respectively. At the has a shortfall of approximately 1,971 beds at the acute community residential level, we estimated a need of level (6.4 additional beds required per 100,000 adults) 22.3 beds per 100,000 adults. and a shortage of 2,796 beds at the subacute level (9.1 Estimated prevalence of SPD in California ranged additional beds required per 100,000 adults)—or 4,767 from 7.9 percent in the San Francisco Bay Area to subacute and acute beds combined, excluding state hos- 9.3 percent in the Southern San Joaquin Valley. When pital beds. If state hospitals were included in this esti- we incorporated this epidemiological information mate, the shortage of acute inpatient beds would shrink into our psychiatric bed need estimates, this intro- to 267, and there would be no observable shortage in duced regional variation in psychiatric bed need that beds at the subacute level. Separately, we estimated a ranged from 45.5 to 55.5 inpatient psychiatric beds per shortage of 2,963 community residential beds. 100,000 adults. Lastly, we collected secondary esti- The top-down estimates of psychiatric bed mates of psychiatric bed need from the academic lit- need—as drawn from the literature and our Techni- erature and our Technical Expert Panel. Using median cal Expert Panel—also indicated a bed shortage: 8.9 values, we generated a separate, top-down estimate beds per 100,000 adults at the acute level and 10.6 of psychiatric bed need: 27.5 beds per 100,000 adults beds per 100,000 adults at the subacute level. There- at the acute level and 25 per 100,000 at the subacute fore, our bottom-up and top-down estimates were level. We were unable to provide a comparable top- closely aligned. The remaining discrepancy likely down estimate of need for community residential beds pertains to differences in the configuration of health because of the significant heterogeneity within this systems throughout the United States and interna- classification and the paucity of academic literature. tionally, including availability of outpatient services We estimate that the magnitude of need for psychi- and alternatives to hospitalization, that drive need. atric beds is expected to grow modestly over the next When regional prevalence estimates for SPD were five years (2021 to 2026): by 1.7 percent. This is primar- incorporated, the gap in beds required reduced mod- ily due to shifting demographic trends, including adult estly: by 4.5 percent. We also documented significant population growth and increasing racial/ethnic diver- regional differences in the estimated shortfall of beds sity, because epidemiological data indicate that His- on the basis of the wide regional variation in psychiat- panic and Black adults experience SPD at higher rates ric bed capacity. For example, two regions of the state than do White adults. Growth in the need for psychi- appear to have sufficient acute inpatient psychiatric atric beds is projected to be largest in the Northern and bed capacity, whereas the remaining eight regions Southern San Joaquin Valley—by about 4.0 percent. 11 of 52 3 have a shortfall. At the subacute level, all regions such regions as the Central Coast, Inland (apart from the Northern San Joaquin Valley) appear Empire, and Southern San Joaquin Valley. to have a shortfall. However, the magnitude of this 2. Consider focusing on building or remodel- shortfall ranges from 5.1 additional beds required per ing infrastructure for the most hard-to-place 100,000 adults in the North Coast region of the state populations. Specific subpopulations appear to 17.2 additional beds required per 100,000 adults in to contribute disproportionately to bottle- the Southern San Joaquin Valley. necks in the current system, including an Lastly, we inquired about hard-to-place popula- inability to transfer patients with criminal tions. Here, we found that a majority of psychiatric justice involvement from the subacute level of facilities at all levels of care reported an inability to care to community residential settings. Given place individuals with comorbid dementia or trau- this, the state might need to consider alterna- matic brain injury, nonambulatory individuals, those tive arrangements for placing such popula- requiring oxygen, and those who tested positive for the tions, such as community-based and outpa- coronavirus disease 2019 (COVID-19). A majority of tient competency restoration programs. Here, respondents from community residential facilities also California could learn from other mental reported an inability to place individuals involved in health systems across the United States and the criminal justice system—particularly those with internationally. arson or sex offense convictions. 3. Set aside state funds for a system that reviews licensure data and periodically collects psychiatric facility–level information. Our Recommendations analysis and conclusions contain numerous Using these findings, we came up with three caveats, in large part because of poor data recommendations: quality. We wish to be transparent about this fact, with the hope that this serves as an impe- 1. Prioritize psychiatric bed infrastructure in tus for the state to consider investing in an the areas with the greatest need. In terms of adequate data review and monitoring system. an absolute shortfall of beds, the shortfall If the state were to allocate funds to routinely was greatest in terms of subacute beds, driven monitor and purge licensure data, policymak- partly by four regions (Los Angeles County, ers would be in a much stronger position to San Francisco Bay Area, Inland Empire, Supe- know what the existing capacity is at each rior region) that represented a shortfall of level of care—particularly at the community more than 2,000 beds—more than a quarter residential level. Likewise, the state should of all additional beds needed throughout the consider establishing a mechanism by which state. If policymakers examine the psychiatric psychiatric facilities report periodically on bed bed shortfall as a proportion of regional adult occupancy rates, wait list volume, number of population, this might lend greater weight to requested transfers to higher and lower levels regions with smaller or more rural popula- of care, and psychiatric patient boarding in tions: For example, the shortfall of subacute emergency departments. The state should also beds is 5.2 beds per 100,000 adults in Los consider collecting sociodemographic and Angeles County compared with 17.2 per clinical information on patients who use psy- 100,000 adults in the Southern San Joaquin chiatric beds. This would allow California to Valley. We also observed significant need for have a remarkably precise and sensitive system acute beds in such regions as the Northern for tracking the impact of investments that and Southern San Joaquin Valley and Central seek to address psychiatric bed shortages. Coast, while the shortfall at the community residential level was particularly notable in 4 12 of 52 Introduction Role of Psychiatric Beds Today, hundreds of Psychiatric beds are essential infrastructure for Californians in need of meeting the needs of individuals with serious mental health conditions.1 These beds serve several psychiatric beds are functions—including enabling safe, stable, and sup- held in hospital EDs or portive environments for individuals in acute mental health crises and for those with significant impair- county jails awaiting ment who require ongoing medical monitoring.2 Not all psychiatric beds are alike, because they openings in inpatient represent infrastructure within different types of facilities. For example, psychiatric beds in acute inpa- care settings. tient hospitals serve those in need of secure, 24-hour care and often include crisis stabilization units. The average length of stay in such states as California is resolution and triage services, acute inpatient ser- one to two weeks.3 By contrast, psychiatric beds in vices, subacute services, state hospitals, and com- such subacute facilities as mental health rehabilita- munity residential services. Clinical guidelines, such tion centers (MHRCs) or special treatment programs as the Level of Care Utilization System, enforced by at skilled nursing facilities (SNFs)—which provide the state of California through Senate Bill (SB) 855 longer-term recovery-oriented services, such as inde- (2020),8 provide a useful compass for matching an pendent life skills training—may remain occupied by individual’s needs with an appropriate level of care. the same individuals for many months.4 Ultimately, psychiatric beds represent an impor- tant component in a continuum of behavioral health Psychiatric Bed Shortfall care that includes integrated community services Psychiatric bed capacity is severely strained in Cali- ranging from prevention and screening to emergency fornia, as it is in much of the United States.9 The crisis response. Depending on the arrangement of present situation may be viewed, in part, as the long services, the need for psychiatric beds may look dif- tail of an effort to deinstitutionalize psychiatric ser- ferent. For example, the CrisisNow model of emer- vices throughout the United States during the latter gency care instituted in Arizona has significantly half of the 20th century: from a peak of 337 psy- reduced the state’s volume of psychiatric emergency chiatric beds per 100,000 individuals in the United department (ED) boarding.5 Assertive Community States in 1955 to a low of around 12 beds per 100,000 Treatment and similar models, such as Full Service in 2016.10 This transition to community-based Partnership programs, may also be key contribu- services—although well-intentioned—has resulted in tors. These models employ transdisciplinary teams a paucity of infrastructure to serve the needs of indi- to provide comprehensive services to patients who viduals who would otherwise benefit from a stable have needs that have not been adequately met by and supervised residence, particularly those with traditional approaches.6 According to the Substance serious mental illness (SMI).11 Abuse and Mental Health Services’ Behavioral Today, hundreds of Californians in need of psy- Health Treatment Services Locator, there are 133 chiatric beds are held in hospital EDs or county jails mental health treatment facilities in California that awaiting openings in inpatient care settings.12 In offer these services, though these services may be addition, a sizable percentage of chronically home- underreported.7 less individuals have an SMI.13 As county jail and With regard to psychiatric beds specifically, homeless populations continue to swell, these rising California embeds psychiatric beds within crisis numbers have created an increasing urgency to take 13 of 52 5 Estimating the need for psychiatric beds is a thorny undertaking. There are no standardized Given the diversity and approaches or best practices, and health systems geographic distribution are constructed differently at the regional and state levels. However, prior literature outlines at least of both adults and four methods for calculating the number of psy- chiatric beds that are needed to address population psychiatric facilities needs. These methods include (1) expert consensus, (2)a normative approach, (3) a population health in California, needs approach, and (4) an observed outcomes approach.16 We briefly survey these approaches, including their throughout the state are strengths and limitations. Expert consensus. Calculating psychiatric bed heterogeneous. need by expert consensus relies on open discussion among content and methods experts—who opera- tionally define a set of relevant principles, deliberate action. For example, in early 2021, an estimated 1,600 evidence, and then achieve mutual agreement on adults in need of psychiatric beds were residing in standards. A key example of this approach is a 2008 county jails because they had been deemed incompe- report by the Treatment Advocacy Center (TAC),17 tent to stand trial and were unable to be placed by the which interviewed 15 experts to deliberate and arrive Department of State Hospitals (DSH).14 Legislators, at a suggested benchmark for measuring psychiatric meanwhile, have called for an overhaul of state psy- bed need: 40 to 60 beds per 100,000 in population.18 chiatric services—particularly against the backdrop of Although expert consensus is a helpful method the COVID-19 pandemic.15 for arriving at an estimate of psychiatric bed need, this approach has limitations. The report by TAC did Measuring Need for Psychiatric Beds not outline the precise deliberations that led experts to arrive at their conclusion, making it challenging Estimating the need for psychiatric beds in Califor- to scrutinize the estimate or replicate the process. nia is essential for at least three reasons. First, when Furthermore, without a set of operational definitions estimates on the need for psychiatric beds are paired for such terms as psychiatric bed, it remains unclear with estimates on the existing bed capacity, evalu- whether a specific benchmark would translate in other ators can determine the magnitude of the shortfall settings or how one would go about allocating 40 to 60 in beds throughout the state, allowing policymak- beds per 100,000 in population across different levels ers to discuss investments accordingly. Second, the of care. The number and types of beds required are estimation process can serve a diagnostic function: liable to depend on local context and resources. Given the diversity and geographic distribution Normative approach. The normative approach of both adults and psychiatric facilities in Califor- to calculating psychiatric bed need is predicated nia, needs throughout the state are heterogeneous. on an assumption that jurisdictions (or countries) Lastly, if the need for psychiatric beds is estimated with similar health systems and demographic char- at multiple intervals, these estimates can provide an acteristics are likely to require a similar number ongoing feedback mechanism for fine-tuning invest- of psychiatric beds. In this scenario, a jurisdiction ments. Because needs are dynamic, investments in with more-robust psychiatric facility infrastruc- infrastructure at any given time may only partially ture has the potential to serve as a comparator for address problems. This factor should create an impe- others. For example, the Organisation for Economic tus to determine how effective initial investments Co-operation and Development has employed this were and to plan for new investments for new needs approach to compare mental health infrastructure have arisen. across its 38 member countries, providing annual- 6 14 of 52 ized estimates of psychiatric beds per 100,000 in for example, the Department of Health used popu- population—with a low of 3 beds per 100,000 in lation and demographic estimates to calculate the population in Mexico to 259 per 100,000 in popula- total psychiatric bed need as 30 beds per 100,000 in tion in Japan (the United States ranks seventh from population.23 Other states (Mississippi, Oklahoma, the bottom at 25 per 100,000).19 The World Health Missouri) have performed similar calculations, using Organization has followed a similar model with its a fixed population ratio to arrive at estimates of need Mental Health Atlas project.20 Coupled with other ranging from 20 to 117 psychiatric beds per 100,000 indicators (e.g., hospital readmission rates), compara- in population.24 This wide range in values is indica- tive analyses can be used to determine whether and tive of disconcerting variation in the parameters used to what extent more-robust infrastructure translates to come up with these estimates. to improved population health outcomes. Calculating the number of psychiatric beds The main challenge with the normative approach needed from this population health approach assumes is that there are large differences in the number of certain targets for specific mental health conditions, psychiatric beds per capita even within countries which may be more or less accurate. For example, that have similar health and economic systems. Fur- there is a paucity of epidemiological data on the likeli- thermore, understanding the difference in reported hood of an individual requiring specific services based numbers of psychiatric beds is difficult because there on mental health diagnosis and illness severity, and on is no standard definition for psychiatric bed. Coun- the optimal duration and intensity of service provi- tries like Italy do not consider residential treatment sion. Relying on limited information risks generating facilities to constitute inpatient care, producing a inaccurate predictions for the number of psychiatric much lower estimate per 100,000 in population com- beds needed and may underpin the wide variation pared with countries that include residential treat- observed in estimates across states—which are often ment facilities.21 Additionally, what works for some using differing sets of assumptions. countries and systems in terms of psychiatric bed Observed outcomes approach. The observed needs and mental health services may not work in outcomes approach is based on the observation that others. Individual countries may first need to define heterogeneous psychiatric bed capacity across coun- their core mental health services and set data-driven ties and states is liable to have observable effects on targets to meet those needs. One recent study found health systems and populations.25 By looking at the that, out of 32 mental health plans developed across relationship between psychiatric bed capacity and key five countries, only four plans included specific tar- performance indicators—such as wait times, occu- gets for their core services—including psychiatric pancy rates, length of stay, emergency room board- bed needs.22 These sorts of comparative metrics are ing, and population health outcomes—researchers also generally lacking within the United States across have the potential to calculate minimum and optimal regions and counties. psychiatric bed capacity requirements from an induc- Population health approach. A population tive perspective.26 Along these lines, in 2020, the San health approach identifies the prevalence of mental Francisco Department of Public Health conducted a health conditions within a geographic area and simulation to assess psychiatric bed need, which ana- then applies a set of standards to meet population lyzed more than 25,000 mental health–related admis- health needs in accordance with these prevalence sions and accounted for variable bed occupancy rates estimates—including quantifying psychiatric services across different levels of care.27 The authors of the and corresponding infrastructure, such as psychi- report concluded that observed bed occupancy rates atric beds. One advantage of this approach is that greater than 85 percent have the potential to contrib- prevalence estimates for particular mental health ute to bottlenecks and flow issues over the long run conditions (and, by extension, needs that derive and suggested an additional 97 beds across four types from these conditions) are sensitive to the underly- of facilities to achieve zero wait time. ing demographic characteristics of the region—such The observed outcomes approach has also been as age, sex, and income distributions. In Tennessee, criticized.28 Specifically, researchers have argued that 15 of 52 7 key performance indicators tend to concentrate on bill’s provisions allow local governments to purchase process measures pertaining to hospital administra- psychiatric facilities to prevent closures and require that tion, which may have little correspondence to patient facility owners give residents greater advanced notice outcomes. Furthermore, hospital functioning may be prior to closure. Several complementary bills signed into dependent on a wide array of factors, such as employ- law earlier this year, including AB 27, AB 362, AB 816, ment rates or social determinants of health within AB 977, AB 1220, AB 1443, and SB 400, represent part the local community. Therefore, these extraneous of a $22 billion investment to address homelessness and factors may function as confounders during analysis the need for behavioral health services. The investments unless they are incorporated as covariates. include $3 billion dedicated to housing for those with Ultimately, as noted earlier in this report, there are acute behavioral and physical health issues—with an no consensus best practices for determining psychiat- expectation of creating approximately 22,000 new beds ric bed needs. The most robust approach may therefore and treatment slots.33 be to inspect the problem from multiple vantage points Additional Assembly legislation, AB 2265, has and methodologies to converge on a triangulated set of focused on increasing access to treatment by allow- estimates for psychiatric bed needs. This includes—to ing California counties to use mental health services the extent possible—pressure testing the approaches funds to address not only mental health conditions outlined above by performing sensitivity analyses. but also substance use disorders, with the goals of We have therefore elected to assume this triangulated enhancing care coordination and creating an inte- approach in our analysis, as further described in the grated behavioral health care system.34 In the context “Methods” section. of psychiatric bed infrastructure, this legislation is particularly relevant for populations in need of longer-term rehabilitative services who are coping California’s Investment in Infrastructure with comorbid substance use and mental health con- California has stated a commitment to expand mental ditions. It was signed into law on September 24, 2020. health infrastructure, including psychiatric bed capacity.29 In early 2020, the California Mental Health Purpose of This Report Services Authority (CalMHSA) announced the forma- tion of a Behavioral Health Task Force, appointed to In this report, we provide an estimate of current advise Governor Newsom on efforts to reform and psychiatric bed capacity throughout the ten census advance behavioral health services throughout the regions of California, according to three overarching state.30 Since the onset of the COVID-19 pandemic levels of care: acute inpatient care, subacute inpa- in 2020, Newsom has signed numerous bills into law tient care, and community residential treatment. We that aim to increase Californians’ access to mental then compare the measure of current psychiatric health services. As noted earlier in this report, these bed capacity with estimates of psychiatric bed need. bills include SB 855, which requires commercial health Lastly, we project bed capacity needs over the next insurance plans outside Medi-Cal to provide medically five years, based on evolving demographic trends necessary treatments for all mental health conditions throughout the state. and substance use disorders.31 The bill also requires We note that the primary bed estimates provided that health plans provide services that comply with in this report do not include state hospitals, although level of care determinations as outlined in the Level of we provide secondary estimates for which state hos- Care Utilization System, American Society of Addic- pitals are included. This decision was based on three tion Medicine criteria, and other clinical guidelines for factors. First, more than 90 percent of psychiatric pediatric populations. beds at state hospitals are occupied by individuals Separately, Assembly Bill (AB) 2377—which was involved in the criminal justice system.35 Thus, state also ratified in 2020—sought to mitigate the impact of hospitals serve a set of clients with unique constraints adult residential facility closures and help residents at that do not apply to local behavioral health con- risk of homelessness in California.32 Specifically, the tinuums that we examine in detail throughout this 8 16 of 52 report. In addition, the forensic population census ization of these facilities and portfolio of services may has continued an upward trend that portends a con- vary. Given this factor, we assigned facilities to three tinued reduction in beds available for non-forensic levels of care—matching closely to a conceptual model individuals. Second, the exclusion of state hospitals established by the County Behavioral Health Directors has the secondary benefit of modeling need for psy- Association of California.37 The levels were defined by chiatric beds if state hospitals were transitioned to two axes: first, the acuity of need being attended to, local, community-based alternatives. Third, unlike ranging from emergent crises to nonemergent, ongo- other types of facilities, state hospitals are not broadly ing supports; second, typical length of stay, ranging distributed throughout the state. This distribution from short term (days to weeks) to long term (months results in computational challenges for determining to years). With the exception of community residen- psychiatric bed shortages in regions that contain state tial facilities, length of stay is usually time-delimited hospitals, especially because not all patients within according to the particular type of facility. state hospitals are residents from the region in which Operationally, we defined the three levels of care that state hospital is located. We discuss further as follows: details in the “Methods” section. 1. acute, representing highly structured, around- We supplement these quantitative analyses with the-clock medically monitored inpatient care input from our panel of technical experts. Using the for individuals at heightened risk of harm to combined results, we outline a series of recommenda- themselves or others, or those who are other- tions. These recommendations pertain to the expan- wise unable to care for themselves sion of psychiatric bed capacity to address existing 2. subacute, representing around-the-clock gaps, and they are situated in California’s context of inpatient care that includes specialized pro- ongoing legislative efforts to establish a holistic con- gramming in a controlled environment with a tinuum of behavioral health care services. significant degree of supervision but with less intensive medical monitoring and interven- tion than acute care Methods 3. residential, representing nonhospital pro- Population and Scope grams in which individuals live on the prem- ises of a facility and are provided with consis- Our population of interest comprised all adults (18 tent programming to promote interpersonal years or older) in California, across all 58 counties. and independent living skills, with staff pres- Because individuals may access psychiatric facili- ent 24 hours a day, seven days a week. ties (and beds) outside their county of residence, we Table 1 details the types of facilities contained aggregated estimates of population, capacity, and within each of these levels of care, based on facility need at a regional level using the U.S. Census Bureau licensure information. classification: Superior California, North Coast, San State hospitals represent a unique type of institu- Francisco Bay Area, Northern San Joaquin Valley, tion. In other settings, acute care is short term and Central Coast, Southern San Joaquin Valley, Inland focused on stabilizing patients, whereas residential Empire, Los Angeles County, Orange County, and San Diego-Imperial.36 Although our denominator for care is lower intensity and geared toward long-term medical and nonmedical supports.38 In this sense, the calculations in this report included all adults in the two axes described above (acuity and length of California—because all adults have the potential to stay) are aligned. The expectation is that patients may use inpatient psychiatric services—a disproportion- transition up and down the care continuum based ate number of those using psychiatric beds are adults on their needs at a given time point. However, state with SMIs, such as schizophrenia, bipolar disorder, hospitals are less dynamic and often provide care for and major depressive disorder. subcategories of patients with long-term, high acuity There are many types of psychiatric facilities in needs or based on medical necessity. For example, California. Depending on the county, the character- 17 of 52 9 TABLE 1 individuals who are incompetent to stand trial and are Levels of Care and Corresponding Adult awaiting transfer to an alternative care setting such Psychiatric Bed Infrastructure as a state hospital.42 Our rationale for this was that, although jail units are not suitable to serve as psychiat- Level of Care Types of Facilities Included ric beds (and therefore should not factor into capacity), Acute Acute psychiatric hospitals; psychiatric they are nevertheless housing individuals in need of (Level 3) health facilities; general acute care psychiatric beds. This was not possible for ED board- hospitals with psychiatric wards; acute beds at state hospitals ing. Although California generates an annual hospital utilization report, these reports do not contain ED Subacute General or specialized subacute facilities; (Level 2) MHRCs; SNFs with specialized treatment boarding rates among patients with mental health con- programs; institutions for mental disease; ditions.43 Likewise, there are no formal tallies of the subacute beds at state hospitals number of individuals with mental health conditions Residential Adult residential treatment facilities; who are receiving permanent supportive housing and (Level 1) enhanced or augmented board-and-care facilities; social rehabilitation facilities would otherwise benefit from alternative placement in a setting with psychiatric beds. NOTE: For a definition of psychiatric health facilities, see California De- partment of Health Care Services, 2021b. For a definition of institutions for mental disease, see California Department of Health Care Services, 2021c. Procedures state hospitals often house forensic patients (i.e., Focus group discussions. As a preliminary step, those involved with the criminal justice system), and, we conducted focus group discussions with county in California, many state hospital beds are reserved leaders at behavioral health departments throughout for such patients.39 This raises two questions: Should the state and members at CalMHSA and the County state hospitals be placed within the care continuum; Behavioral Health Directors Association. These and, if so, where? For the purposes of this report, we discussions focused on conceptual issues that these have remained agnostic to the first question, com- individuals were confronting with regard to sup- puting capacity and bed shortages with and without porting psychiatric bed needs. We focused on four inclusion of state hospitals. For the latter, we allo- topics: (1) perceived structural drivers of psychiatric cated acute state hospital beds to the acute level and bed shortages, (2) populations who were challenging subacute beds to the subacute level, both of which are to place in psychiatric facilities, (3) defining the care documented in licensure data sets.40 continuum, including which types of facilities cor- Specific populations and bed categories were respond to which levels of care, and (4) other areas deemed to be outside the scope of our analysis. of note that would be important for RAND research- Regarding populations, we excluded children and ado- ers to consider. We took detailed notes from each lescents, for whom there are subtle but important dif- discussion and deliberated the feedback provided to ferences in the care continuum and who were therefore us to develop our methodological approach, which is deemed to merit a separate analysis.41 Regarding bed detailed below. categories, we omitted beds corresponding to perma- Estimation of capacity. We downloaded the nent supportive housing, those in county jails, and most current licensure data available for each type those in EDs used for boarding patients with mental of psychiatric facility, using public data sets from health conditions. These categories are seldom quanti- the California Department of Public Health,44 Cali- fied as psychiatric beds because they are not exclu- fornia Department of State Hospitals,45 California sively reserved for populations with mental health Department of Health Care Services,46 and Cali- conditions, though it may be the case that individuals fornia Department of Social Services.47 These data with mental health conditions occupy one of these bed sets were then merged into a master file of facilities types. with psychiatric beds in California, and each facility For quantification of psychiatric bed need, we was geocoded at the address and county levels using incorporated the number of jail units occupied by ArcGIS Desktop 10.8.48 10 18 of 52 To validate facility licensure data, we executed need. In total, the Technical Expert Panel consisted two additional steps. First, we contacted behavioral of four participants whose names and titles can be health directors in all 58 counties in October 2021, found in Appendix B, alongside all major prompts providing them with an inventory of facilities within used during the panel’s discussion. their county and soliciting revisions. Where discrep- Estimation of need, Approach 3: population ancies arose, we prioritized the revisions detailed health assessment. Lastly, we drew from epide- by the county point of contact. Second, we reviewed miological data reported in the California Health 2,500 online entries for community residential Information Survey (CHIS) and NSDUH.50 Both the facilities to estimate the percentage of these facilities CHIS and NSDUH employ the Kessler 6, which is a within each county that provided services to indi- measure of psychological distress.51 A score of 13 or viduals with mental health conditions rather than greater indicates SPD, which is a marker of probable those with other types of disabilities or the elderly. SMI. This allowed us to rebalance regional need for For example, if we queried 60 community residential psychiatric beds based on (1) estimated regional prev- facilities in Orange County and found that 5 of 60 alence of SPD among adults, according to the CHIS, were psychiatric facilities with psychiatric beds, we and (2) expected utilization of psychiatric inpatient inferred that this percentage (5/60 = 8.3%) applied services among adults with versus without SPD in at the population level—that is, that 8.3 percent of the United States, based on NSDUH (which contains Orange County’s community residential facilities and estimates of self-reported receipt of inpatient mental beds were psychiatric facilities that contained psychi- health services in the past year).52 A substantial body atric beds. of literature has shown that SPD is a correlate of SMI Estimation of need, Approach 1: survey of and that those with SPD have a greater need for both psychiatric facilities. Using the observed outcomes outpatient and inpatient services.53 approach described earlier in this report, we drew a Projection of need. In addition to estimat- random sample of psychiatric facilities throughout ing current need for psychiatric beds, we projected California using Stata’s v.17 runiform command49 change in need over the next five years (2021 to and attempted to contact facility directors to inquire 2026). These estimates drew from expected trends about bed occupancy, average length of stay, wait in population growth and demographic shifts in the list volume, and the number of patients that facility age, sex, and racial/ethnic composition of the state directors recommended for transfer to a higher or according to U.S. Census Bureau information. The lower level of care. In the event that we were unable “Analysis” section provides a fuller description of the to reach an administrative leader at a particular facil- computations involved in this. ity, we made up to four additional attempted con- tacts per facility. For facilities that we were unable to Measures reach, we imputed estimates based on median values within the facility type and county, adjusting for total Psychiatric beds. We defined psychiatric beds as number of beds. Calls were made in October and beds within psychiatric facilities that have the pri- November 2021.