Providing Behavioral Health Services in Nursing Homes is Difficult: Findings from a National Survey

Abstract

Objectives:

Behavioral health (BH) disorders affect 65–90% of nursing home (NH) residents. Access to BH services in NHs has been generally considered inadequate, but the empirical evidence is sparse. We examined the availability of BH services and identified facility-level factors associated with the difficulty of providing BH services in NHs.

Design:

A national random sample of 3,996 NHs was identified. Two structured surveys with questions about BH service availability, quality, satisfaction, staffing, staff education, turnover, and service barriers, were mailed to administrators and directors of nursing in each NH between July-December 2017.

Setting/Participants:

Completed surveys were obtained from 1,079 NHs (27% response rate). Descriptive statistics and multivariable logistic regressions were employed.

Measurements:

Four outcome measures: 1) Adequacy of BH staff education; 2) Ability to meet resident BH service needs; 3) Adequacy of coordination/collaboration between NH/community providers; and 4) Availability of necessary facility infrastructure, were based on 5-point Likert scales.

Results:

BH service needs were unmet in one-third of NHs; almost half lacked appropriate staff BH education. Over 30% reported having inadequate coordination of care between NH and community providers, and 26.2% had inadequate infrastructure for residents’ referrals/transport. Staff BH education was less problematic in NHs with Alzheimer’s units (OR:0.6;p-value<0.05), lower RN turnover (OR:0.7;p-value<0.05), and more psychiatrically trained RNs (OR:0.5;p-value<0.001) and social workers (OR:0.6;p-value<0.05). Lower RN turnover (OR:0.7;p-value<0.05) and more psychiatrically trained RNs (OR:0.6;p-value<0.05) were associated with fewer NHs reporting being unable to meet BH service needs. Having more psychiatrically trained RNs (OR:0.6;p-value<0.05) was associated with fewer NHs reporting inadequate coordination with community providers.

Conclusion:

Inadequate BH education and psychiatric training among NH staff were associated with sub-par provision of BH services in this care setting. New initiatives that increase access to BH providers and services, and improve staff education are urgently needed in NHs.

INTRODUCTION

Behavioral health (BH) disorders affect 65–90% of nursing home (NH) residents.^1–3 The proportion of NH admissions among residents with BH disorders is increasing.^4–6 These residents often require access to appropriate psychiatric expertise for treatments/symptom management.⁷ While the need for BH services in NHs is recognized, access to such services is generally considered inadequate, and may be further exacerbated by disparities in admissions to high quality NHs among residents with BH disorders.⁵

Following the 1987 Institute of Medicine report on inadequate and inappropriate treatment of NH residents with BH disorders,⁸ Congress enacted the Nursing Home Reform Act, as part of the Omnibus Budget Reconciliation Act (OBRA), which included regulations to address unmet BH needs in NHs. In the early 1990s, the level of BH services provided to NH residents was still reported as well below actual need.⁹ National-level surveys conducted between 1995–2004 demonstrated that the overall availability of BH services in US NHs did not improve nearly 20 years following OBRA’87. By 2004, only 25% of US NHs provided BH services on a regular basis, 24% on as needed basis, 28% both on a regular and as needed basis, and 22% provided no such services.¹⁰

Two recent surveys of BH services were conducted in NHs in Florida¹¹ and Hawaii.¹² Administrators in Florida NHs reported being highly satisfied with BH service availability/quality. Findings from Hawaii were more negative and mixed.¹² Two thirds of the respondents agreed that access to BH services was problematic and 50% reported BH-related staff education and training as insufficient.

Currently, we know little about US NHs’ ability to provide quality BH services. Motivated by the increasing population of NH residents with BH conditions, limitations of the few prior small-scale surveys, and lacking national-level data, we conducted a large, random, national survey of US NHs. The study objectives were to: 1) examine NHs’ perceptions of BH services availability and quality, and 2) identify facility-level factors associated with difficulty in providing BH services in NHs.

METHODS

Primary and secondary data

A national random sample of 3,996 NHs was identified from the Nursing Home Compare (NHC) website, maintained by the Centers for Medicare and Medicaid Services (CMS). Two versions of the same survey were mailed, one to NH administrators and another to directors of nursing (DONs), between July-December 2017. Survey packets included an information letter, survey, and pre-addressed stamped return envelope. An 800-phone number was made available to call with questions, two follow-up mailings were sent to non-responding NHs 1 and 2 months after the initial mailing, and the survey was made electronically available to initial non-respondents and in subsequent mailings. Surveys contained 10 questions (based on extensive literature review) regarding availability and quality of, and satisfaction with, BH services and providers, staffing, staff education, and turnover.

Information on key NH characteristics was obtained from the NHC file,¹³ LTCfocus dataset,¹⁴ and Rural-Urban Commuting Area Codes (RUCA) dataset.¹⁵

The study protocol was reviewed and approved by the Institutional Review Board.

Measures of BH service availability and key outcomes

NHs’ perceptions of BH service availability were assessed in two survey questions: one asking about the degree of difficulty the NH has in providing BH services, and another regarding the extent to which NHs consider various aspects of the admission process when admitting patients with a recent history or current diagnosis of psychiatric/behavioral problems.¹²

Four outcome measures summarizing aspects of service availability were constructed based on responses to the following statements: 1) the level of BH education among our staff is an issue in our ability to provide good BH services; 2) we are unable to fully meet resident needs for BH services; 3) there is inadequate coordination/collaboration between facility staff and community providers; and 4) we lack adequate facility infrastructure (e.g. to make referrals, transport residents to services). Each statement was assessed on a 5-point Likert scale with a higher number denoting poorer outcome.

Other covariates

We included other covariates that may be important in understanding availability of BH services in NHs.^5,10–12 NH-level characteristics from the NHC file¹³ (state, ownership, nursing hours/resident day, 5-star ratings for quality and staffing, NH market bed competition) and LTCfocus data¹⁴ (NH bed size, occupancy rate, presence of Alzheimer’s unit, percent Medicare/Medicaid residents) were included. Each NH was categorized as rural/urban based on zip codes from the RUCA database.¹⁶ Staff turnover for RNs and CNAs, and proportions of psychiatrically trained physicians, nurse practitioners/physician assistants, and social workers were obtained from the survey.

Analyses

We first compared responding NHs to all facilities in our national sample (Table 1). Chi-square and t-tests were used for statistical inference as appropriate. We then fit separate multivariable logistic regression models to assess the relationship between NH factors and outcome measures. All analyses were conducted at the NH level using R version 3.5.1.

Table 1.

Sample Characteristics: A comparison of NHs responding to the survey and the national random sample of NHs.

Characteristic	Survey Returned (N=1,079)	National Random Sample (N=3,996)	p-value (χ2 or t-test)
For profit	62.5%	72.6%	<0.001
Chain	53.9%	58.8%	0.005
Quality star rating: 1 Star	10.8%	10.5%	0.95
Quality star rating: 5 Stars	33.6%	33.5%
Staffing star rating: 1 Star	8.7%	11.5%	0.001
Staffing star rating: 5 Stars	14.2%	10.9%
Total beds	Mean: 108.2, SD: 64.0	Mean: 110.0, SD: 59.6	0.42
Alzheimer’s unit	19.4%	15.4%	0.002
Percent Medicaid	Mean: 58.3, SD: 22.6	Mean: 59.2, SD: 22.4	0.21
Percent Medicare	Mean: 13.4, SD: 12.2	Mean: 14.7, SD: 12.9	0.002
Occupancy rate	Mean: 82.2, SD: 14.1	Mean: 82.1, SD: 14.1	0.89
Urban	62.5%	66.9%	0.009
Market bed competition (1-HHI)	Mean: 0.3, SD: 0.3	Mean: 0.6, SD: 0.3	<0.001

RESULTS

Overall, 1,264 surveys were obtained from 1,079 NHs (27% NH response rate). If a NH returned multiple surveys (185 NHs returned more than 1), the survey from the individual with the longest tenure in their position at the NH was included in the final sample.

Responding NH characteristics

Responding NHs were less likely to be for-profit or chain-affiliated compared to the national sample, and more likely to report having an Alzheimer’s unit (Table 1). Responding NHs were less likely to be located in urban areas or counties with more competition for NH beds and more likely to have a 5-star rating for staffing quality. However, there were no statistically significant differences in 5-star overall quality rating, occupancy rate, bed size or the percent of Medicaid/Medicare residents between the responding NHs and national sample.

BH service availability and influence on admission process

NHs rated the degree of difficulty in providing BH services (selecting ‘NA’ if the service was not available) (Figure 1). NHs indicated that BH medication management (NA for 1.1% of NHs), assessment for emotional/behavioral problems/disorders (NA for 2.3% of NHs), and behavior management (NA for 2.4% of NHs) were generally provided, although 19.8%, 26.5%, and 30.6% of NHs reported the provision of these services as being difficult-to-very difficult, respectively. However, BH services for family support, group counseling/therapy, and substance abuse counseling were largely unavailable (35.5%, 49.3%, and 56.3%, respectively); with 1/4 to 1/3 of NHs reporting the delivery of these services to be difficult-to-very difficult (24.3%, 29.6%, and 28.7%, respectively) when available. While referral to specialized programs/services for emotional/behavioral problems/disorders and BH crisis intervention were reported as generally available (12% and 15.5%, respectively, were not available), 40.2% and 45.1% of NHs reported significant difficulty providing these services, respectively.

Figure 1.

Nursing homes’ rating of the degree of difficulty in providing and availability of behavioral health services. Difficult to Provide=4–5 on a Likert scale. Service not Available=‘NA’ response.

NHs identified the extent to which aspects of caring for residents with BH disorders were considered when admitting patients with a recent history or current diagnosis of psychiatric/behavioral problems. Among the important-to-very-important considerations at admission were: the necessity of providing greater staff attention to residents with behavioral symptoms (88.8%); the concern for resident and staff safety (86.6%); and the difficulty in hospitalizing BH patients when the need arose (71.9%) (Figure 2). Other factors reported as important-to-very-important at admission included: difficulty accessing psychiatric support and follow-up (60.8%); inadequate staff training (58.2%); delays in admission caused by Preadmission Screening and Resident Review (PASRR) (39.0%); and a primary facility focus on short-term rehabilitation and skilled nursing residents (37.8%).

Figure 2.

Nursing homes’ rating of the extent to which aspects of caring for residents with behavioral health disorders are considered when admitting patients with a recent history or current diagnosis of psychiatric/behavioral problems.¹²

NH satisfaction with the provision of BH services

NHs also rated primary challenges in their ability to deliver BH services to residents. Almost half (47.3%) reported the level of BH education among staff affected their ability to provide BH services (Supplementary Table S1). At least 1/3 (33.3%) of NHs reported being unable to adequately meet residents’ BH service needs. Inadequate coordination between facility staff and community providers was an issue for 31.8% of NHs, and 26.2% reported lacking adequate infrastructure to make referrals/transport residents to services outside their facility.

The results of multivariable logistic regression models identifying NH factors associated with BH services availability/satisfaction are shown in Supplementary Table S1. NHs were less likely to find staff BH education problematic if they had an Alzheimer’s unit (OR 0.6;p-value<0.05), lower RN turnover (OR 0.7;p-value<0.05), and more psychiatrically trained RNs (OR 0.5;p-value<0.001) and social workers (OR 0.6;p-value<0.05). NHs with lower RN turnover (OR 0.7;p-value<0.05) and more psychiatrically trained RNs (OR 0.6;p-value<0.05) were less likely to report being unable to meet resident needs for BH services. NHs with more psychiatrically trained RNs (OR 0.6;p-value<0.05) were less likely to report having inadequate coordination with community providers. Urban NHs (OR 0.6;p-value<0.05), and those with lower RN turnover (OR 0.7;p-value<0.05), higher RN staffing (OR 0.7;p-value<0.05), and more psychiatrically trained physicians (OR 0.5;p-value<0.05) were less likely to report inadequate facility infrastructure, while those with higher CNA turnover (OR 1.8;p-value<0.01) were more likely to report inadequate infrastructure to make referrals/transport residents to services.

DISCUSSION

While most NHs reported being able to provide basic BH services (e.g., problem assessment, on-site management or referral to specialty care), between 20–40% reported the provision of these services as difficult-to-very-difficult. Almost half reported that lacking appropriate staff education was an obstacle in providing BH services, and at least 1/3 were not able to adequately meet residents’ BH needs. The ability to provide BH services also influenced NHs’ perceived impact on admissions. For example, PASRR requirements were instituted because of concerns that individuals with serious mental illness might reside in NHs lacking adequate resources, but today almost 40% of surveyed NHs viewed PASRR as hindering admission screening and causing delays, and more than 60% were concerned because of the perceived difficulty in accessing psychiatric support after NH admission. If NHs routinely provided BH services PASRR should not be viewed as hindering admission nor should service provision after admission be viewed as an impediment in so many facilities.

More than 30 years after OBRA’87, we found relatively little progress in NHs’ ability to adequately provide BH services to residents. This is troubling as the proportion of NH population with BH conditions is large and increasing, driven largely by NH admission of residents diagnosed with depression.^1,10 In most states Medicaid covers only very basic BH services. Since Medicaid reimbursement rates are generally very low, there is little incentive for NHs to enhance these services to long-term care residents (who are largely funded by Medicaid). Thus, it is not surprising that almost half of NHs reported access to specialized BH services or BH crisis interventions, respectively, were difficult-to-very-difficult to provide, if available at all.

We did not find that NH quality, as measured by the CMS 5-star system, was statistically significantly associated with any of our outcome measures. Prior research suggests that residents with BH conditions face disparities in accessing high quality homes.⁵ While it is possible that this access disparity negatively impacts overall outcomes, our findings suggest that inadequate access to BH services seems to be problematic for both lower and higher star-rated NHs. Similarly, while for-profit NHs are usually found to have lower care quality, we found no difference in this respect Vis-a-Vis our BH outcome measures. However, as in other studies demonstrating that NHs with higher staffing provide better care quality, our findings show that consistent (i.e. less turnover) and adequate nurse staffing, and psychiatrically trained RNs and physicians, were predictive of a NH’s ability to provide BH services to residents in need.

Several limitations should be noted. As in any voluntary survey, selection bias is possible. The responding NHs were somewhat different from the random sample, with the responding NHs showing higher quality of care (higher 5-star ratings). This suggests that our findings may somewhat underestimate the inadequate state of BH services in NHs. Respondents’ bias may be a limitation, but is not a significant concern because of the large sample and relatively high response rate. As in other research, potential omitted variable bias may be present.

In conclusion, this study found that consistent and adequate nurse staffing, and staff BH education and psychiatric training were associated with higher availability of BH services in NHs. Further research is needed to understand the association between BH service availability and residents’ outcomes.

Supplementary Material

Supp Table

Supplementary Table S1. Results from multivariable logistic regression models assessing relationships between nursing home characteristics and outcome measures.