Advance Directives State Requirements, Center Practices, and Participant Prevalence in Adult Day Services Centers: Findings From the 2016 National Study of Long-Term Care Providers

Abstract

Objectives

Adult day services centers (ADSCs) may serve as an entrée to advance care planning. This study examined state requirements for ADSCs to provide advance directives (ADs) information to ADSC participants, ADSCs’ awareness of requirements, ADSCs’ practice of providing AD information, and their associations with the percentage of participants with ADs.

Methods

Using the 2016 National Study of Long-Term Care Providers, analyses included 3,305 ADSCs that documented ADs in participants’ files. Bivariate and linear regression analyses were conducted.

Results

Nine states had a requirement to provide AD information. About 80.8% of ADSCs provided AD information and 41.3% of participants had documented ADs. There were significant associations between state requirements, awareness, and providing information with AD prevalence. State requirement was mediated by awareness.

Discussion

This study found many ADSCs provided AD information, and ADSCs that thought their state had a requirement and provided information was associated with AD prevalence, regardless of state requirements.

Advance directives (ADs), documents expressing end-of-life health care preferences, are an important component of care planning for individuals who require long-term care services. Having an AD can improve quality and satisfaction with care (Garrido, 2015) and reduce end-of-life health care spending (Nicholas, 2011). There is a paucity of research on ADs in adult day services centers (ADSCs). ADSCs provide services to community-dwelling older adults with disabilities and adults with mental illness or intellectual/developmental disabilities. In 2016, 4,600 ADSCs served 286,300 participants, many with high levels of need; for example, about one quarter had an activity of daily living limitation and one third had dementia (Lendon & Rome, 2018). ADSCs offer social and medical care to improve quality of life, reduce hospitalizations and nursing home admissions, and provide respite for caregivers (Fields et al., 2014). ADSCs may be an entrée to advance care planning by providing AD information and documentation to help facilitate communication and execution of participants’ care preferences across health care settings (Coleman, 2003; Sinuff et al. 2015).

Nursing homes, residential care communities, and home health agencies in the United States are required by the Patient Self Determination Act (1990) to provide information about ADs; however, ADSCs are state licensed/certified and not all states have such requirements. Awareness of requirements, providing AD information, and prevalence of ADs among participants likely vary considerably due to different state requirements (O’Keeffe et al., 2014). A recent report shows that about 38% of ADSC participants had an AD, which varied by region (Lendon et al., 2018). The extent of ADSC directors/staff awareness of their state’s requirement is unknown and may affect the provision of AD information. Conversely, ADSCs may enact this practice regardless of state requirements. Little is known about whether providing information affects the prevalence of ADs among participants. This report examines state requirements, awareness of state requirements, and the practice of providing AD information and their associations with AD prevalence among participants.

Method

Data Source

Data are from the ADSC survey of the 2016 National Study of Long-Term Care Providers (NSLTCP), conducted by the National Center for Health Statistics (NCHS). NCHS fielded a census using the National Adult Day Services Association’s list of ADSCs in the United States, regardless of membership. To be eligible for NSLTCP, ADSCs must (a) have been included in the database and in operation as of November 2015; (b) be licensed or certified by the state to provide adult day services, accredited by the Commission on Accreditation of Rehabilitation Facilities, authorized/set up to participate in Medicaid, or part of a Program of All-Inclusive Care for the Elderly; (c) have an average daily attendance of one or more participants in a typical week; and (d) have one or more participants enrolled at the time of the survey. NSLTCP was multimode, using mail, web, and telephone follow-up. The response rate was 61.8%. Target respondents were ADSC directors (or a knowledgeable staff member), who were instructed to consult participants’ records. There were two versions of the questionnaire with random assignment of ADSCs to each. Questions about ADs appeared in Version B. More details on NSLTCP are published elsewhere (NCHS, 2017).

The data were nationally representative of about 4,600 ADSCs. Of the 1,426 respondents (4,594, weighted) to Version B, the analytic sample was further restricted to 1,045 (3,305, weighted) ADSCs that maintained documentation of ADs in participant files (78.1% of ADSCs). Exclusion was necessary to maintain the same denominator throughout all analyses because of an embedded skip pattern where NSLTCP collected the number of participants with ADs only from ADSCs that maintained documentation.

Measures

State requirement and ADSC awareness

Determining a “state requirement” for ADSCs to provide AD information entailed reviewing several sources, beginning with the Assistant Secretary for Planning and Evaluation’s Regulatory Review of Adult Day Services: Final Report, 2014 Edition (O’Keefe et al, 2014). The report was confirmed or supplemented by information from state ADSC associations and aging/public health agency websites, using the following search terms: advance directive, proxy, physician orders for life-sustaining treatments, medical orders for life-sustaining treatments, living will, and advance care planning. A state was considered to have a requirement if a source stipulated ADSCs must provide AD information to their participants and/or caregivers and was in effect before 2016 (Supplementary Appendix A). Based on independent reviews by two authors (J. P. Lendon and C. Caffrey), there was 100% interrater agreement.

State requirement awareness was measured by “Does your state require your center to provide information about ADs to participants or their families?” (yes/no/do not know). Awareness was collapsed to two categories for the multivariate analyses (yes/no and do not know).

Providing AD information

Providing AD information was measured by “Advance directives are written documentation and may include health care proxies, durable power of attorney, living wills, do not resuscitate (DNR) orders, or physician or medical orders for life-sustaining treatments (POLST or MOLST). Does this ADSC provide any information about advance directives to participants or their families?” (yes/no).

Participants with ADs

The percentage of participants with ADs was based on two questions with a skip. The first question: “Does this adult day services center typically maintain documentation of participants’ advance directives or have documentation that an advance directive exists in participant files?” If yes, “Of the current participants, how many have documentation of an advance directive in their file?” The number of participants with an AD was converted to a percentage by dividing it by the total number of participants in the respective ADSC.

Covariates

Covariates included Census region, chain and ownership status, Medicaid licensure, electronic health records use, and model type. Chain status referred to ownership by a person, group, or organization with two or more ADSCs. Types of ownership included private nonprofit, private for-profit or publicly traded or limited liability, and government. Medicaid licensure referred to ADSCs authorized or set up to participate in a Medicaid plan. Electronic health records use referred to the use of computerized participant health/personal information. ADSCs were categorized as a medical model (designed to meet “only health/medical needs,” “primarily health/medical needs and some social/recreational needs,” or “equally social/recreational and health/medical needs”) or social model (designed to meet “only social/recreational needs” or “primarily social/recreational needs and some health/medical needs”). ADSC size was the total number of participants.

Analyses

Univariate and bivariate analyses were performed to describe the overall sample, using chi-square and t tests and pair-wise tests, where applicable, for statistical significance. Multivariate linear regression examined the association between a state requirement and the percentage of ADSC participants with ADs (Model A). State requirement awareness was then entered into the model (Model B), to examine possible mediating effects, followed by providing AD information (Model C). Mediation of state requirement awareness was tested by bootstrapping the indirect effect and reporting bias-corrected 95% confidence interval (Hayes, 2009). Each model included all covariates.

Statistical significance tests were two-sided at p less than .05 significance level. Data analyses were performed using Stata/SE version 14 and complex survey weights to estimate robust standard errors. Univariate and bivariate analyses excluded missing cases on a case-by-case basis. Multivariate models excluded cases with missing data on any of the variables. Missing data ranged from 0.1% for medical model to 2.8% for the number of participants with ADs. Diagnostics showed no multicollinearity in the multivariate models.

Results

State Requirement, Awareness, and Providing Information

Nine states had requirements for ADSCs to provide AD information, with 22.3% of ADSCs located in these states (Table 1). About 39.2% of ADSCs thought their state had a requirement, 36.6% thought their state did not, and 24.3% did not know. In states with a requirement, 66.1% thought there was a requirement, compared to 31.4% in states without requirement. Overall, 80.8% of ADSCs provided AD information. In states with a requirement, 90.3% of ADSCs provided information, compared to 78.1% in states without a requirement.

ADSC Participants With ADs

Overall, 41.3% of participants had an AD (Table 1), which did not differ by state requirement. Among ADSCs that thought their state had a requirement, 52.3% of participants had ADs, followed by 38.3% in ADSCs that did not know, and 31.1% in ADSCs that thought the state did not have a requirement (Table 2). Among ADSCs that provided information, 46.8% of participants had ADs, compared to 18.6% in ADSCs that did not.

Linear regression models of the percentage of participants with ADs are presented in Table 3. In Model A, state requirement had a statistically significant positive association with the percentage of participants with ADs (b = 7.7, SE = 2.8, p = .006). Model B shows that state requirement awareness mediated the association between having a state requirement and the percentage with ADs. ADSCs that thought their state had a requirement (b = 20.1, SE = 2.3, p < .001) had greater percentages of participants with ADs, compared to ADSCs that thought their state had no requirement or did not know. The direct effect of state requirement on AD prevalence was no longer statistically significant (b = 2.4, SE = 2.9, p = .413), but the indirect effect (5.4, bias-corrected 95% CI = 3.3–8.1), by way of awareness, was statistically significant.

Model C found that ADSCs that provided information (b = 24.3, SE = 2.4, p < .001) had a greater percentage of participants with ADs, compared to ADSCs that did not provide information. Awareness (b = 14.2, SE = 2.4, p < .001) was also still associated with percentage of participants with ADs.

Discussion

This study examined the associations between AD prevalence and having a state requirement for ADSCs to provide AD information to participants; ADSC’s awareness of state requirements; and providing AD information. Despite the absence of a federal mandate and less than a quarter of ADSCs located in nine states with a requirement, about 80% of ADSCs that documented ADs provided AD information and 41% of their participants had ADs.

A state requirement was associated with the prevalence of ADs; however, this association was mediated by awareness. ADSC directors/staff who think there is a state requirement (despite incorrect perceptions of requirements) and providing AD information were important independent drivers of the prevalence of ADs, regardless of actual state requirements. These findings suggest ADSCs may believe providing AD information is a useful practice for many reasons, it is encouraged by other entities, or have other AD-related state requirements. For example, this practice may be influenced by Medicaid participation requirements, provider association recommendations, attitudes of ADSC owners/operators, or participants’ health needs. ADSCs serve a unique population with complex health care needs who may benefit from advance care planning. This study suggests that educating participants about ADs may be positively associated with the percentage of participants with ADs.

This study has several limitations. The 2016 NSLTCP collected aggregated participant data; therefore, this study focused on regulatory and organizational factors and not individual-level sociodemographic/cultural/health characteristics that might also be related to having an AD. Analyses were restricted to only ADSCs that maintained AD documentation, which differ from ADSCs that do not document (Lendon et al., 2018), and might undercount participants with ADs. ADSCs that documented ADs were less likely to be in the West and more likely to be Medicaid licensed and to use electronic health records. State requirements were based on independent reviews of secondary resources. Despite 100% agreement between the authors, regulations in some states might have been missed. The study could not determine if the survey respondent was unaware of a state requirement, whereas another employee might have been aware.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Acknowledgments

The findings and conclusions in this article are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention, National Center for Health Statistics (NCHS). The authors thank Amy Branum, PhD, Lauren Harris-Kojetin, PhD, and Alexander Strashny, PhD (all from NCHS) for their valuable feedback on this article. We are grateful to the journal’s peer reviewers for their suggestions to improve this article for publication.

Conflict of Interest

None declared.

Author Contributions

All authors planned the study. J. P. Lendon and C. Caffrey conducted and verified statistical analyses. J. P. Lendon wrote the first draft of the article, and C. Caffrey and D. T. Lau provided substantial revisions.

References