Emergency department (ED) encounters for mental illness and substance use disorders [referred to here as behavioral health (BH) disorders] are increasing over time across the United States, as are BH comorbidities among ED patients.1,2 Patients with two or more BH diagnoses are twice as likely to use EDs compared with people without BH diagnoses.3 However, EDs may be overcrowded and poorly equipped to deliver care for BH conditions.4 To improve ED care, the American College of Emergency Physicians recommends that EDs integrate multiple BH processes, including routine screening for BH concerns such as alcohol use disorders and intimate partner violence, telehealth, the use of interdisciplinary team-based care, and referrals for ongoing care of BH conditions.5 These processes have been shown to efficiently identify previously undiagnosed BH problems,6,7 reduce wait times and time in the ED, and improve access to psychiatric consultation.8–10

The extent to which US hospital EDs have implemented recommended BH processes and whether specific hospital capabilities can support the integration of BH processes within EDs remains unclear. This study aims to understand the extent to which hospital capabilities are associated with greater adoption of BH care processes in ED. We examined the association of hospital capabilities and 4 BH care processes (screenings, a team care approach, telepsychiatry, and referrals to community-based BH clinicians) in a national sample of EDs.

First, because recent studies have found that hospitals using quality improvement methods were more likely to adopt care delivery innovations, we examined associations between BH care processes and hospital use of Lean, Six Sigma, or Robust Performance Improvement.11–13 Second, we examined whether hospital processes to disseminate best practices were associated with the use of BH care processes, as prior evidence observed that hospitals with processes to disseminate best-care practices were more likely to use clinical decision support tools to enhance care in other clinical domains.14 Third, based on evidence on the importance of innovation capacity in relation to other upstream care strategies,15 we examined whether barriers to innovation may impede the implementation of new processes like BH care in the ED. Last, we examined the hospital capacity to treat BH conditions based on the presence of designated beds for psychiatric or substance use disorders. Prior evidence indicates that hospitals with psychiatric treatment capacity were more likely to adopt telepsychiatry16 and it is possible that designated capacity is synergistic with other BH processes.

Overall, the hospital capabilities examined may indicate a hospital’s recognition of the value of new external information, tools, and processes to support hospital performance.17 They may also be an indication of resources to support organizational change and incorporate new BH processes into the workflows of busy EDs. Given that these BH processes can improve diagnosis and care quality for ED patients and BH integration in hospital EDs is strongly recommended by the World Health Organization (WHO), National Association of EMS Physicians, and US Preventive Services Task Force,7,18,19 our study can provide insights about how hospitals can improve the implementation of BH care processes.

METHODS Data and Sample

We linked the 2017/2018 National Survey of Healthcare Organizations and Systems (NSHOS) hospital survey with the 2017 American Hospital Association (AHA) Annual Survey. The NSHOS hospital survey (response rate=46.5%) collected information on organizational structures and processes of a national sample of critical access and general acute care hospitals between June 2017 and August 2018.20 The AHA Annual Survey provided information on hospital characteristics.21 We excluded 18 hospitals that were not in the NSHOS sample frame (eg, due to respondent organization or role changes) and 137 hospitals without responses to the AHA Annual Survey. The analytic sample included 602 hospitals across 50 states that completed the NSHOS and AHA surveys.

Measures Outcome Variables

NSHOS assessed 4 BH processes in hospital EDs: (1) routine screening and assessment for BH conditions, (2) team-based approaches to BH including mental health providers (ie, psychiatrist, psychiatric nurse practitioner), (3) the use of telehealth or telepsychiatry, and (4) referral to BH clinicians in the community. Routine BH screening was calculated as a composite index based on Yes versus No responses for each of the 6 conditions: (1) depression, (2) sexual, physical, or psychological abuse, (3) marijuana use, (4) opioid use, (5) alcohol abuse, and (6) other substance use [range=0–100; internal consistency reliability (α)=0.80]. The remaining 3 BH process measures were each dichotomous (Yes vs. No) (Appendix Table 1, Supplemental Digital Content 1, https://links.lww.com/MLR/C787).

Hospital Capabilities

We constructed 4 measures of hospital capabilities. First, we defined the use of quality improvement methods from the NSHOS as using any method: Lean, Six Sigma, or Robust Performance Improvement. Second, we measured dissemination processes for patient-care best practices from the NSHOS by constructing a count variable of 5 processes: regular staff meetings, regular listserv emails/newsletters, department representatives or champions, an electronic database of practice or system endorsed guidelines, and performance improvement events (Yes vs. No for each, composite range=0–5). Third, we assessed 6 barriers that impede hospital implementation of care delivery innovations reported in the NSHOS, including: “lack of a process for identifying beneficial innovations,” “lack of a process for disseminating information about innovations,” “not enough time to implement,” “insufficient financial resources to implement,” “lack the necessary knowledge/expertise to implement,” and “lack of incentives to implement.” We assigned each item 100 points if a “Major barrier,” 50 if a “Minor barrier,” and 0 if “Not a barrier” and then calculated the unweighted average of all items (range=0–100; α=0.80). Fourth, we created an indicator for the availability of inpatient beds for psychiatric or substance use care based on AHA survey data.

Control Variables

We controlled for hospital ownership, hospital participation in a network, accreditation by the Joint Commission, health system membership, teaching hospital status, and hospital size (number of beds). To account for characteristics of communities served by hospitals, we controlled for US Census region (Northeast, Midwest, South, and West); ZIP-code-level Rural-Urban Commuting Area codes categorized as urban, large-rural, and small-rural22; and the proportion of the population below the federal poverty levels in a hospital’s ZIP codes.

Statistical Analyses

Using NSHOS sampling and nonresponse weights, we first computed means and SD for continuous variables and frequencies and proportions for binary variables. Then, separate multivariable regression models were estimated for each of the 4 hospital ED BH processes to assess their association with hospital capabilities, controlling for hospital structural characteristics and area-level demographic characteristics (Table 1). We used linear regression for BH screening (index range: 0–100) and logistic regression for each of the 3 binary BH processes examined. Robust standard errors were used to account for the heteroscedasticity of the models. For ease of interpretation, we estimated marginal effects, which estimate the probability that an ED used the BH care process in response to a one-unit change in continuous hospital capability measures or a 0 to 1 change of binary hospital capability measures. Data analyses were conducted using Stata software, version 17.0.23

TABLE 1 - Integration of Behavioral Health Processes in Hospital EDs and Hospital Characteristics (N=602) N (%) or mean (SD) Behavioral health integration  Screening for mental and substance use conditions   Mean of 0 to 100 index (SD) 75.9 (29.8)  Use of a team approach to behavioral health, N (%) 277 (46.0)  Use of telepsychiatry, N (%) 240 (39.8)  Referrals to community-based behavioral health clinicians, N (%) 498 (82.7) Organizational capabilities  Any quality improvement method, N (%) 391 (65.0)  Dissemination of best patient-care practice approaches   Mean of number of approaches (SD) 3.9 (1.2)  Barriers to the adoption of evidence-based clinical treatments   Mean of 0 to 100 index (SD) 51.9 (26.0)  Having inpatient beds for psychiatric/substance use care, N (%) 177 (29.4) Structural characteristics  Participate in network, N (%) 302 (50.2)  Health system member, N (%) 391 (65.0)  Ownership, N (%)   Public 127 (21.1)   Private, nonprofit 409 (67.9)   Private, for profit 66 (11.0)  Joint Commission Accreditation, N (%) 378 (62.8)  Teaching hospital, N (%) 238 (39.5)  Hospital bed, mean (SD) 179 (213) Area-level characteristics  US Census region, N (%)   West 117 (19.4)   Northeast 88 (14.6)   Midwest 169 (28.1)   South 228 (37.9)  Rurality (by ZIP code), N (%)   Urban 311 (51.7)   Large-rural 115 (19.1)   Small-rural 176 (29.2)  % population below census poverty level, N (%)   <10% 147 (24.4)   10%–20% 324 (53.8)   20%–30% 86 (14.3)   >30% 35 (5.8)   Missing 10 (1.7)

ED indicates emergency department.

Source: 2017/2018 National Survey of Healthcare Organizations and Systems (NSHOS) hospital survey with the 2017 American Hospital Association (AHA) Annual Survey. All descriptive statistics were weighted using National Survey of Healthcare Organizations and systems (NHSOS) survey sampling weights.

RESULTS Descriptive Statistics Behavioral Health Processes

Routine screening for BH conditions was common in hospital EDs (mean=75.9, SD=29.8). Nearly half (46%) of the hospitals used team-based approaches to BH, 40% used telehealth or telepsychiatry, and 83% referred patients to BH clinicians in the community (Table 1).

Hospital Capabilities

Most hospitals (N=391, 65%) reported using at least one quality improvement method. Hospitals used multiple approaches to disseminate the best patient-care practices, averaging 4 out of 6 approaches (SD=1.2). The average index of barriers to the use of care delivery innovations was 51.9 out of 100 (SD=26.0). A minority (29%) of the hospitals had inpatient care capacity for psychiatric or substance use disorders (Table 1).

Multivariable Regression Analyses

In adjusted analyses, hospitals using more methods for disseminating best practices conducted routine screening and assessment for more BH conditions in the ED. For one additional process used to disseminate best patient-care practices, the screening index (0–100 scale) increased by 4.07 points (P<0.01) (Table 2). Further, hospitals with more barriers to using care delivery innovations implemented BH screening for fewer conditions. The BH screening index was 0.15 points lower per 1-point increase in the 0–100 barrier index (P<0.01), highlighting that EDs with more methods to disseminate best practices and fewer barriers to the use of care innovations were more likely to screen for BH conditions.

TABLE 2 - Associations Between Hospital Characteristics and Behavioral Health Processes in Emergency Departments Behavioral health processes in hospital emergency departments Screening for mental health/substance use disorders Team approach to behavioral health Telehealth/telepsychiatry Referrals to community-based behavioral health clinicians Hospital characteristics Coefficients [95% CI] Marginal effects [95% CI] Marginal effects [95% CI] Marginal effects [95% CI] Organizational capabilities  Barriers to the adoption of care delivery innovations −0.15 [−0.24; −0.05]** −0.28 [−0.41; −0.14]*** −0.06 [−0.21; 0.10] 0.09 [−0.21; 0.03]  Any quality improvement method −0.51 [−6.65; 5.63] 1.64 [

−

7.00; 10.28] 6.13 [−3.66; 15.92] 2.40 [−4.88; 9.68]  Dissemination of best patient- care practices 4.07 [1.69; 6.46]** 4.41 [1.22; 7.59]** 3.29 [−0.50; 7.07] 0.66 [−2.02; 3.35]  Having inpatient beds for psychiatric/substance use care 2.03 (−3.84; 7.90]) 27.7 [19.77; 35.68]*** 1.65 [−7.77; 11.06] 0.72 [−7.09; 8.52] Structural characteristics  Participate in network −0.83 [−5.85; 4.19] −4.15 [−11.61; 3.30] −7.69 [−15.93; 0.55] −6.36 [−12.95; 0.24]  Health system member −2.95 [−9.10; 3.19] −3.91 [−13.21; 5.38] 8.92 [−1.12; 18.97] 1.68 [−6.01; 9.36]  Ownership (reference=public)   Private, nonprofit 3.58 [−4.21; 11.37] 9.72 [−1.55; 21.00] 3.85 [−8.70; 16.40] 5.61 [−3.67; 14.90]   Private, for profit 0.71 [−12.83; 14.25] 3.11 [−18.83; 25.05] 14.02 [−8.67; 36.71] 4.69 [−11.35; 20.72]   Joint Commission Accreditation −2.66 [−8.31; 2.99] −8.87 [−17.87; 0.12] −5.83 [−15.28; 3.62] −9.76 [−17.22; −2.31]*    Teaching hospital 4.58 [−1.42; 10.58] 10.83 [1.66; 20.01]* −7.65 [−17.94; 2.64] 0.51 [−7.17; 8.20]    Hospital beds −1.88 [−4.86; 1.11] 3.05 [−3.27; 9.36] 4.09 [−1.40; 9.57] 4.64 [−1.08; 10.36] Area-level characteristics  US Census region (ref=West)   Northeast 2.07 [−5.79; 9.94] −14.85 [−27.68; −2.02]* 9.53 [−4.94; 24.01] 3.07 [−8.97; 15.10]   Midwest −1.91 [−9.10; 5.29] −12.00 [−22.36; −1.64]* 7.06 [−4.60; 18.72] −2.97 [−11.71; 5.77]   South −3.57 [−10.51; 3.36] −22.57 [−32.99; −12.16]*** 7.77 [−3.98; 19.52] 4.96 [−4.75; 14.68]  Rurality (by ZIP code, ref=urban)   Large-rural −4.57 [−12.38; 3.25] −5.44 [−17.45; 6.56] 2.37 [−10.62; 15.37] 2.81 [−7.74; 13.37]   Small-rural −1.70 [−9.75; 6.36] −2.43 [−14.35; 9.49] 11.30 [−1.56; 24.15] −8.38 [−17.64; 0.89]  % population below census poverty level (ref=

<

10%)   10%–20% −4.74 [−10.62; 1.14] −2.56 [−11.94; 6.82] −2.82 [−12.77; 7.14] −5.82 [−13.88; 2.23]   20%–30% −0.42 [−7.81; 6.98] −1.95 [−14.84; 10.94] −10.29 [−23.16; 2.58] −3.66 [−14.24; 6.93]   >30% −4.23 [−5.19; 13.66] 5.68 [−10.48; 21.84] −7.60 [−25.57; 10.37] −6.88 [−21.77; 8.00]

Number of hospital beds were standardized and its coefficients are the changes in the outcomes associated with a standard deviation change in the number of beds.

**P<0.01.

***P<0.001.

Source: Variables from 2017/2018 National Survey of Healthcare Organizations and Systems (NSHOS) hospital survey: all organizational capabilities except for “Having inpatient beds for psychiatric/substance use care,” rurality, and % population below census poverty level. Variables from the 2017 American Hospital Association (AHA) Annual Survey: all structural characteristics, and US Census region. Estimates based on linear regression (for screening for mental/substance disorders) or logistic regressions of the remaining dependent variables as a function of the hospital characteristics shown in the table. 95% CIs are in brackets.

Similarly, hospitals were more likely to use team-based approaches for BH in the ED when more methods were used to disseminate best practices (4.41 percentage point increase in use of team-based for an additional method of disseminating best practices, P<0.01) and when they experienced fewer barriers to innovation (0.28 percentage point increase in the use of team-based approaches for 1-point reduction in the barrier index, P<0.001). Hospitals were also more likely to report using team-based approaches when the hospital reported capacity for inpatient care for psychiatric/substance use disorders compared with hospitals without inpatient care capacity (P<0.001).

Neither the use of telepsychiatry in EDs nor the presence of a referral system to community-based BH clinicians was associated with hospital capabilities. Teaching hospitals were more likely to adopt team-based approaches compared to nonteaching hospitals (P<0.05).

DISCUSSION

To our knowledge, our study is the first national assessment of the integration of BH processes within hospital EDs. Most hospitals were found to use screening for BH conditions and had referral processes to community-based BH clinicians in place. Approximately half of the hospitals used a team approach to BH. A minority had implemented telepsychiatry.

Hospitals with more extensive methods for disseminating patient-care best practices and fewer barriers to implementing care delivery innovations were more likely to conduct routine screening and assessment for BH conditions and to use team-based approaches to BH in the ED. Our results offer context for past research suggesting the ED was an important setting to address BH problems given the high rates of BH conditions among ED patients.24 Implementing universal screening may require hospital investment in personnel or increased workloads for existing staff. Hospitals would require support or reallocation of resources to improve BH care processes and balance the benefits and costs of integrating these BH care processes for patients and staff. Similarly, team-based care for emergency psychiatric services that involves nurses, physicians, and psychiatrists can improve wait times and reduce the use of restraint.8 Hospital policies that focus on information sharing and communication across disciplines might support the use of team care models in EDs. Programs or policies that enable hospitals to develop formal approaches to share patient-care best practices that are enabled by structured communication opportunities can strengthen BH assessment and treatment capabilities in EDs. One example is the Agency for Healthcare Research and Quality’s Learning Health Systems initiative that assists organizations in improving the uptake of effective clinical interventions into routine practice through evidence appraisal, information sharing, and dissemination of best practices.25

None of the hospital capabilities assessed in the study were associated with telepsychiatry or referrals to community-based BH clinicians. Prior research observed that the receipt of telepsychiatry in the ED was associated with improved outcomes including reduced time in the ED, less crowding, reduced travel time for physicians, and improved access to on-site psychiatric consultation.9,10 Patients with BH conditions need coordinated and accessible care across community and hospital settings. Our null results suggest that hospitals need more resources to integrate BH processes in the ED, including licensed psychiatrists or other BH providers, reimbursement mechanisms for telepsychiatry, and increased access to community-based BH care providers.16,26,27 More evidence is needed to better understand how hospitals can foster the use of telepsychiatry and referrals to community-based BH clinicians and improve the effectiveness of these systems.

However, limitations should be considered when interpreting our results. First, we examined whether hospital EDs adopted BH care processes, but we did not assess the types of screening tools for mental health and substance abuse disorders, the structure of team-based care models, telepsychiatry modalities, referral processes, and follow-up, or how consistently BH care processes were used. Second, we were not able to control for several patients (eg, payer mix, illness severity) and organizational characteristics (eg, availability of mental health professionals). The inclusion of additional covariates could alter the results and the impact of these covariates on BH processes in the ED should be examined when relevant data are available. Third, the lack of statistically significant associations observed for some outcomes could be due to the modest sample size. Finally, we were not able to examine whether the use of BH services in hospital EDs was associated with patient outcomes. However, given scant evidence on hospital implementation of BH processes within EDs at a national scale, this study provides a foundation for future research.

CONCLUSIONS

Our study is the first to explore hospital adoption of recommended BH care processes in EDs. We found that methods for disseminating patient-care best practices and strategies to encourage the use of care delivery innovations were associated with a greater use of BH screening and the use of team-based approaches to care. More evidence, however, is needed about how hospitals can integrate telepsychiatry services and strengthen ED referral systems for community-based BH clinicians.

REFERENCES 1. Capp R, Hardy R, Lindrooth R, et al. National trends in emergency department visits by adults with mental health disorders. J Emerg Med. 2016;51:131–135.e1. 2. Weiss AJ, Barrett ML, Heslin KC, et al. Trends in emergency department visits involving mental and substance use disorders, 2006-2013. 2016. HCUP (Healthcare Cost and Utilization Project) Statistical Brief #216 Accessed 31 October, 2021. https://www.hcup-us.ahrq.gov/reports/statbriefs/sb216-Mental-Substance-Use-Disorder-ED-Visit-Trends.pdf 3. Weilburg JB, Wong HJ, Sistrom CL, et al. Behavioral health factors as predictors of emergency department use in the high-risk, high-cost Medicare population. Psychiatr Serv. 2018;69:1230–1237. 4. Stefan S. Emergency Department Treatment of the Psychiatric Patient: Policy Issues and Legal Requirements. Oxford University Press; 2006. 5. American College of Emergency Physicians. Policy Compendium. Vol. November. 2022:95-96, 203, 211, 486. Accessed 15 February, 2023. https://www.acep.org/globalassets/new-pdfs/policy-statements/policy-compendium.pdf 6. Thompson RG Jr, Mullinax S, De Monte R, et al. Effectiveness of a self-administered computerized mental health screening tool in the emergency department. Psychiatr Serv. 2023;74:1180–1184. 7. Chun TH, Duffy SJ, Linakis JG. Emergency department screening for adolescent mental health disorders: the who, what, when, where, why and how it could and should be done. Clin Pediatr Emerg Med. 2013;14:3–11. 8. Okafor M, Wrenn G, Ede V, et al. Improving quality of emergency care through integration of mental health. Community Ment Health J. 2016;52:332–342. 9. Zhong C, Freeman RE, Boggs KM, et al. Receipt of telepsychiatry and emergency department visit outcomes in New York state. PsychiatrQuart. 2021;92:1109–1127. 10. Reliford A, Adebanjo B. Use of telepsychiatry in pediatric emergency room to decrease length of stay for psychiatric patients, improve resident on-call burden, and reduce factors related to physician burnout. Telemed e-Health. 2019;25:828–832. 11. Carman K, Paez K, Stephens J, et al. Improving Care Delivery Through Lean: Implementation Case Studies. Agency for Healthcare Research and Quality; 2014. 12. Steinfeld B, Scott J, Vilander G, et al. The role of lean process improvement in implementation of evidence-based practices in behavioral health care. J Behav Health Serv Res. 2015;42:504–518. 13. Idemoto L, Williams B, Blackmore C. Using lean methodology to improve efficiency of electronic order set maintenance in the hospital. BMJ Open Qual. 2016;5:u211725; w4724. 14. Bui LN, Marshall C, Miller-Rosales C, et al. Hospital adoption of electronic decision support tools for preeclampsia management. Qual Manage Health Care. 2022;31:59–67. 15. Brewster AL, Fraze TK, Gottlieb LM, et al. The role of value‐based payment in promoting innovation to address social risks: a cross‐sectional study of social risk screening by us physicians. Milbank Q. 2020;98:1114–1133. 16. Li Z, Harrison SE, Li X, et al. Telepsychiatry adoption across hospitals in the United States: a cross-sectional study. BMC Psychiatry. 2021;21:1–12. 17. Cohen WM, Levinthal DA. Absorptive capacity: a new perspective on learning and innovation. Adm Sci Q. 1990;35:128–152. 18. Krist AH, Davidson KW, Mangione CM, et al. Screening for unhealthy drug use: US Preventive Services Task Force Recommendation Statement. Jama. 2020;323:2301–2309. 19. Mulvaney-Day N, Marshall T, Downey Piscopo K, et al. Screening for behavioral health conditions in primary care settings: a systematic review of the literature. J Gen Intern Med. 2018;33:335–346. 20. AHRQ-Funded Center of Excellence: Dartmouth–Berkeley–Harvard–Mayo Clinic. About NSHOS. Updated. 2018. Accessed 4 November, 2021. https://sites.dartmouth.edu/coe/nshos/ 21. American Hospital Association. AHA Annual Survey Database (2017 Edition). 2023. Accessed 4 November 4, 2021. https://www.ahadata.com/aha-annual-survey-database-asdb/ 22. University of Washington. Rural-urban community area codes (RUCAs). Accessed 31 October, 2020. http://depts.washington.edu/uwruca/ 23. StataCorp LLC. Stata: Release 17 Statistical Software. StataCorp LP; 2021. 24. Kene M, Miller Rosales C, Wood S, et al. Feasibility of expanded emergency department screening for behavioral health problems. Am J Manag Care. 2018;24:585–591. 25. Greene SM, Holmes KL. Learn to fly: training and competencies to support the multidisciplinary workforce needs of learning health systems. Learn Health Syst. 2022;6:e10347. 26. Zachrison KS, Boggs KM, Hayden EM, et al. Understanding barriers to telemedicine implementation in rural emergency departments. Ann Emerg Med. 2020;75:392–399. 27. Nesper AC, Morris BA, Scher LM, et al. Effect of decreasing county mental health services on the emergency department. Ann Emerg Med. 2016;67:525–530.