Background

In 2009 and 2010, the Virginia General Assembly enacted major amendments (Virginia Code Annotated 2010), proposed by the Commission on Mental Health Law Reform (Commonwealth of Virginia Commission on Mental Health Law Reform 2008, 2009), to the Commonwealth’s Health Care Decisions Act (HCDA). These changes empowered persons with chronic, disabling health conditions including serious mental illness (SMI) to document, while competent, their treatment instructions and preferences in advance directives (Commonwealth of Virginia Commission on Mental Health Law Reform 2008, Bonnie et al. 2009). Through these advance directives (ADs), individuals could also authorize a healthcare agent to make treatment decisions for them during periods of incapacity. Virginia’s law is innovative in that it provides an integrated advance directive for mental health, medical, or end-of-life treatment decisions, without distinguishing legally between psychiatric and other causes of decisional incapacity. Efforts to implement the law statewide are underway, with the goal of optimizing the law’s potential to benefit persons with serious mental illness, enhance quality of care and outcomes in public behavioral healthcare systems, and ultimately promote population mental health in the community while protecting public safety. Information about key stakeholders’ support for the law, their perceptions of its benefits and drawbacks, and their own experience with ADs is needed to guide implementation of the HCDA in Virginia and to inform other states that may seek to follow Virginia’s policy lead in this area. This article reports the results of a survey of five key stakeholder groups—mental health program administrators, clinicians, service users, family members, and advocates—about their knowledge and perceptions of the HCDA and experience with ADs.

Advance directives for mental health treatment, often called psychiatric advance directives, represent an innovative and promising approach for promoting treatment engagement and continuity of care for people with serious mental health conditions and could play a significant role in improving long-term health and safety outcomes (Swanson et al. 2006a, 2008; Wilder et al. 2010). A major goal of these directives is to empower people with mental illness to control their own care and, in so doing, to reduce the need for coercive interventions during mental health crises (Commonwealth of Virginia Commission on Mental Health Law Reform 2008, Swanson et al. 2000, Scheyett et al. 2007). However, unlike psychiatric advance directive statutes in most states, the Virginia HCDA had a goal of integration of mental health and medical advance directives to create a comprehensive advance directive that would avoid mental health exceptionalism. Prior to the 2009/2010 amendments to the HCDA, Virginia law provided for ADs only in end-of-life care without any legally recognizable mechanism for individuals who did not have a terminal condition to execute an advance directive. The amendments, which allowed advance planning for any form of impaired decisional capacity, were explicit in their inclusion of all types of health care decisions, not just psychiatric care (Commonwealth of Virginia Commission on Mental Health Law Reform 2008). Virginia’s revised advance directive also includes an optional “Ulysses clause,” which allows a person to authorize future treatment even over their later objection during an incapacitating crisis (Henderson et al. 2008).

A previous survey of mental health consumers in five states showed that the majority of participants would want to complete a psychiatric advance directive, if given the necessary assistance, but very few had completed one (Swanson et al. 2006b). The reasons for lack of uptake are likely due to a combination of barriers at several levels, although the lack of direct, one-on-one assistance in the task of completing an advance directive may be the most significant barrier to widespread adoption of these promising instruments. The provision of trained facilitators to assist consumers in completing their ADs dramatically increased consumer rates of completion in a randomized study in North Carolina (Swanson et al. 2006a). Recognizing this, Virginia is currently undertaking an unprecedented effort to promote use of ADs by mental health consumers and to assist them in completing and executing these legal instruments. This effort was initially coordinated and overseen by the Commission on Mental Health Law Reform, and is now coordinated and overseen by a successor body, the Coordinating Committee for Promoting Use of Advance Directives by People with Mental Illness.

This study comprised part of the implementation effort in Virginia. For the implementation effort to be successful, a clear understanding of the current state of affairs regarding ADs in Virginia was needed. This study was designed to collect baseline data on knowledge, attitudes and experiences regarding the use of ADs in connection with mental health treatment. Five key stakeholder groups—administrators of hospitals and community mental health agencies called Community Service Boards (CSBs), clinicians, mental health service users, family members, and advocates—were surveyed to assess their knowledge of and attitudes about the HCDA and ADs. Because overcoming barriers to AD use is critical, this study also collected information on barriers to implementation. We expected that different stakeholder groups might identify very different barriers to implementation; therefore, inclusion of multiple stakeholders in the same study might offer a clearer picture of the challenges that must be addressed in order to effectively implement ADs in mental health care. Unlike previous surveys about psychiatric advance directives, which have tended to include a single stakeholder group (usually providers), this study is unique in its inclusion of core questions that were identical for all stakeholder groups, allowing direct comparison of attitudes, experiences, and beliefs about ADs between all interested parties.

Methods

Our study was approved by the University of Virginia IRB for Social and Behavioral Sciences and by the Duke University Medical Center IRB. We developed an accessible internet-based survey to assess stakeholders’ knowledge, experience, and opinions concerning Virginia’s HCDA and the use of advance directives in mental health care. In developing survey items, we created a set of core questions relevant across stakeholder groups, as well as more tailored questions for each of the primary constituency groups: treatment program administrators, clinicians, consumers, family, and advocates. The core questions allowed comparison of responses across groups while group-specific questions provided additional detail on each group’s interests, knowledge, and needs. Survey questions were based on pre-existing surveys used in psychiatric advance directive research in North Carolina (Elbogen et al. 2006). The initial survey items were reviewed by the Coordinating Committee, which included members from each stakeholder group. We used feedback from committee members to refine questions, and then piloted the survey on a group of 10 individuals, including clinician, consumer, and administrator volunteers identified by the committee. Final changes were made based on the pilot test results prior to opening the survey to respondents.

To qualify for survey participation, individuals had to report employment as a facility administrator or clinician (e.g., psychiatrist, psychologist, mental health social worker, or mental health counselor) at a Virginia community service board, psychiatric hospital, or residential mental health treatment facility, or self-identify as a mental health service recipient, family member, or a mental health advocate. To recruit survey participants, we submitted survey information and URL links to professional or advocacy organizations with these stakeholder members.Footnote 1 These organizations, in turn, directly emailed the link to members, included the link within a regularly published newsletter, or posted the link on their website. The survey was available for completion for 30 days (throughout June 2010); potential participants received 2 reminder e-mails via their professional or advocacy organizations, one at day 15, and a second at day 25.

A total of 460 individuals responded to the survey. There were 67 administrator responses, 268 clinician responses, 40 consumer responses, 60 family responses, and 25 advocate responses. Response rate estimates varied by stakeholder group. We were uncertain how many individuals in each group had an opportunity to view the URL link, with the exception of the administrator surveys. Administrator surveys were sent to a single individual at each of 16 state facilities in DBHDS, and 90–100 members of VHHA, resulting in a 58–63% administrator response rate. Extrapolating from the number of individuals believed to have received the clinician and consumer/family/advocate survey links,Footnote 2 we estimate response rates of 10 and 7%, respectively.

Of those participants who began the survey, 55 (82%) of the administrator, 238 (88%) of the clinician, 39 (98%) of the consumer, 52 (87%) of the family, and 21 (84%) of the advocate group completed it. Among all groups, the majority of respondents were female (55–70%), Caucasian (56–80%), and middle-aged (mean ages ranged from 46 to 58). Greater than 95% of administrators and clinicians held at least a bachelor’s degree; about 60% of consumers, family, and advocates did as well. Administrators and clinicians tended to be experienced (23 and 15 average years of experience, respectively). More clinicians than administrators worked in an outpatient setting (85% vs. 58%). Both groups reported that the majority of their consumers had either no insurance or only Medicaid, and about half of each group reported that they worked primarily with consumers with psychotic illnesses.

Results

Below, we report survey results across all stakeholder groups, organized by survey theme: opinions, current practice, knowledge, and perceived barriers. Sample sizes vary considerably across items given variation in item completion within and across stakeholder groups. Because Virginia law specifies an inclusive advance directive for medical and mental health care, our survey used the term “advance directive” rather than “psychiatric advance directive” throughout. We attempted to clarify survey instructions and questions through the frequent inclusion of the phrases “advance directives for individuals with serious mental illness,” and “advance directives that include mental health care.”

Opinions About Advance Directives

All respondents held generally favorable opinions of ADs, although clinicians were slightly, but not significantly, less positive than administrators or consumers. Greater than 90% (354/380) of respondents agreed with the statement, “Advance directives that include mental health care will give people with serious mental health problems more control over their lives.” Greater than 90% also agreed that ADs will lead to a better understanding by providers of what consumers want for treatment in both crisis (341/368) and outpatient settings (357/379), and to an improved quality of life for consumers (323/354). Over 80% agreed that ADs would improve the relationship between mental health providers and consumers (283/332), as well increase providers’ likelihood of listening to consumers (327/376) and being helpful to them (285/351). Only 21% agreed that ADs would have “very little impact one way or the other” (73/341) and 18% believed that ADs would lead to increased medical malpractice lawsuits (51/279).

Advance Directive Policy and Practice

Current policy about ADs and use of ADs within clinical settings were assessed in the administrator and clinician group surveys. 67% of clinicians (166/249) reported that they had discussed ADs with at least one of their clients, but only 19% (47/249) reported that they had personally assisted at least one client with completing an AD. Our results also indicated differences in reported current interventions associated with ADs between administrator and clinician stakeholders in both inpatient and outpatient facilities (Table 1). Administrators reported higher use of ADs compared to clinicians, particularly in the inpatient setting.

Table 1 Reported current practice of the use of advance directives in facilities providing treatment to individuals with severe mental illness
Full size table

Knowledge of Advance Directives

Knowledge of the legal underpinnings of ADs in Virginia, as defined in the HCDA, varied substantially across and between stakeholder groups. 100% (17/17) of inpatient and 88% (29/33) of outpatient administrators had heard of the HCDA, while 67% (12/18) of inpatient and 62% (130/209) of outpatient clinicians had heard of the law. 82% (14/17) of inpatient and 24% (8/33) of outpatient administrators had attended a training specifically on recent amendments to the HCDA (which included changes to AD law), while 17% (3/18) of inpatient and 16% (34/209) of outpatient clinicians had been to such a training. 100% of both inpatient and outpatient administrators reported having received general training on ADs, while only about 30% of inpatient (5/18) and outpatient (65/209) providers reported having received AD training.

The majority of respondents to the consumer, family, and advocate surveys reported familiarity with the HCDA (72% or 18/25 of advocates, 58% or 22/38 of consumers, and 53% or 30/57 of family). 49% (19/39) of consumers reported having personally completed an AD; of these all but 1 had appointed a health care agent (18/39). Among consumers who did not have an AD, 75% (15/20) reported that they would like to have one.

Several questions in the administrator and clinician forms of the survey asked participants to choose whether a factual statement was true or false (e.g., “The person [who has completed an AD that includes mental health care] will never be involuntarily hospitalized”). The scores for these questions were compiled to create an aggregate measure that reflected the participant’s level of accurate knowledge about ADs. Scores could range from 0 to 4. The average score for clinicians was slightly but significantly lower than the score for administrators (2.9 vs. 3.1, P < 0.02, two-sided t test with unequal variances). There was a positive association between participation in a training on ADs and knowledge score (P < 0.03, two-sided t test with unequal variances) but no significant association with having discussed ADs with consumers (P = 0.11) or assisted them in completing one (P = 0.9).

Perceived Barriers to Advance Directive Implementation and Use

Consumers, clinicians, and administrators were asked about barriers that might prevent a consumer from completing an AD. Clinicians and administrators additionally were asked about barriers that might prevent clinicians from assisting consumers with ADs and about barriers that might prevent ADs being used in a crisis.

Consumer Completion of Advance Directives

Table 2 shows the percent of each group who agreed with statements describing potential barriers to consumers trying to complete an AD. Overall, relatively more clinicians agreed with statements identifying barriers than did administrators, consumers, family, or advocates. Identified barriers also differed between groups. For example, 29% administrators agreed with the statement, “Consumers think no one will pay attention to the AD,” but 77% of advocates agreed with this. On the other hand, over 75% of clinicians agreed with statements indicating that consumers would be unable to identify a health care proxy they trusted, that consumers did not trust legal documents, and that consumers did not think ADs were relevant to them, while less than 25% of consumers agreed with any of these statements. In a multivariable regression model in which the total number of identified barriers was regressed onto respondent type, sex, age, and race, clinicians identified significantly more barriers (beta = 0.54, 95% CI 0.22–0.86, P < 0.001) than other groups (administrators, consumers, family, or advocates), while consumers identified significantly fewer barriers (beta = −0.37, 95% CI = −0.73–−0.011, P < 0.05) than other groups. There were no significant differences by sex, age, race, or other respondent type.

Table 2 Stakeholder perception of barriers that might prevent consumers with severe mental illness from completing an advance directive
Full size table

Clinician Facilitation of Advance Directives

Table 3 shows the percent of administrators and clinicians who agreed with statements describing potential barriers for clinician facilitation of ADs. A majority of both groups perceived significant barriers to facilitation by clinicians; these concerns were more widely expressed by administrators in most cases, although clinicians were particularly likely to express concern that they lacked training to carry out facilitation. On the other hand, it is noteworthy that a majority of both groups disagreed with the statement that consumers would make inappropriate requests in their ADs. A multivariable regression model in which the total number of identified facilitation barriers was regressed onto respondent type (clinician versus administrator), clinical setting (outpatient vs. inpatient), sex, age, and race, revealed no significant predictors of response.

Table 3 Administrator and clinician reported barriers to facilitating completion and use of advance directives for consumers with severe mental illness
Full size table

Accessing and Using Advance Directives During a Crisis

Table 3 also shows the percent of administrators and clinicians who agreed with statements describing potential barriers to using ADs in a crisis. Over 90% of respondents agreed with the concern that the consumer would be too psychotic during a crisis to remember to notify providers that they have an AD, and 95% of administrators agreed that consumers would want to change their minds about their ADs during crises. No statements were endorsed by less than 50% of the respondents, indicating overall high levels of concern about crisis barriers. A multivariable regression model in which the total number of identified facilitation barriers was regressed onto respondent type (clinician versus administrator), clinical setting (outpatient vs. inpatient), sex, age, and race, revealed no significant predictors of response.

Overall Endorsement of Barriers

Aggregate barrier scores for clinicians and administrators were created by totaling the number of barriers endorsed in all categories, corrected for the number of questions that were not answered. Barrier scores could range from 0 to 22. The mean barrier score was 15.4 (SD 3.3) for administrators and 15.8 (SD 2.7) for clinicians. Barrier scores were inversely correlated with the knowledge scores previously described (Spearman rank-order correlation = −0.15, P < 0.01); that is, lower knowledge of ADs was correlated with endorsing a higher number of barriers. Additionally, barrier scores were significantly lower in clinicians who reported having personally discussed ADs with consumers (15.5 vs. 16.4, P < 0.02, two-sided t test with unequal variances), and among clinicians and administrators reporting a policy of discussing ADs with at least 50% of their consumers (14.2 vs. 16.0, P < 0.0001, two-sided t test with unequal variances). Clinicians who reported they had assisted at least one consumer in completing an AD also reported lower barrier scores, but this was not statistically significant (15.0 vs. 15.9, P = 0.1, two-sided t test with unequal variances).

Conclusions

Our survey indicates that all relevant stakeholder groups broadly support the implementation of ADs under Virginia’s Health Care Decision Act, but the level of baseline knowledge, use, and perception of barriers of ADs varies by group. Responses to opinion-based questions about the theoretical benefits of ADs for mental health were overwhelmingly positive. Less than a fifth of respondents endorsed negative attitudes about ADs (i.e., that they will not be helpful, that they will lead to increased malpractice lawsuits, etc.). Despite enthusiasm for ADs, substantial barriers—a lack of accurate knowledge, incompatible attitudes, and system barriers—threaten to derail implementation; these barriers need to be identified, confronted, and creatively addressed over time.

Specific training in legal and clinical aspects of ADs varied widely between stakeholder groups, with administrators being more knowledgeable and better trained than providers and inpatient staff being more knowledgeable and better trained than outpatient staff. The lack of knowledge and training about ADs by providers is particularly important in light of the finding that there was an inverse association between perceived barriers to implementation of ADs and knowledge about them. Receipt of training about ADs was associated with increased knowledge, suggesting that expanded training is needed to improve clinical knowledge of ADs, and thus decrease barriers to their use.

Perceived barriers differed somewhat among stakeholders. Clinicians and administrators appeared to underestimate consumers’ desire for ADs. They were also more concerned about consumers’ ability to complete ADs than were the consumers themselves. Consumers and family members, on the other hand, expressed skepticism about clinician and administrator willingness to respect or implement their ADs. The combination of clinicians who think consumers do not want ADs and consumers who think clinicians will not respect their ADs logically results in very few ADs being completed. Educating one group of stakeholders about the opinions of other stakeholder groups might improve this situation. Because the survey examined perceptions and beliefs, rather than actual behavior, it is impossible to know if stakeholder report would coincide with behavior in the clinic. It is possible that clinicians accurately gauge consumers’ inability to complete ADs, and that consumers hold well-founded concerns that clinicians will override or ignore their ADs in a mental health crisis.

In this survey, administrators tended to report higher use of ADs than did clinicians. Administrators may be more likely to reflect the policy of an institution when answering questions about AD use, while clinicians may be more likely to report the reality of clinical services as they observe it on a daily basis. If administrators are not aware that ADs are being used far less than policy dictates, they may be equally unaware of the need for top-down efforts to increase use through additional training and provision of clinical time for AD completion. However, terminology used in the survey may partly explain this apparent discrepancy. Although we focused our survey on advance directives which included mental health treatment, the use of the generic term “advance directive” may have introduced a lack of precision in meaning, making it somewhat difficult to interpret the results relating to current practice.

It is likely that most individuals had at least some knowledge of the pre-existing laws governing medical advance directives, and may have based their reports on their conceptualization of the older form of advance directive. For example, hospital administrators are probably very familiar with Joint Commission requirements that all individuals be asked whether they have an advance directive when they are admitted to a hospital and likely already have practices in place to ensure this happens. Their report of higher use of advance directives may reflect this focus on medical advance directives, while mental health clinicians and consumers may have been more focused on the psychiatric use of advance directives.

Like all surveys, our study is limited by multiple potential sources of error. Foremost is the problem of low response rates, particularly in the consumer, family, and advocate groups, as well as the difficulty of determining accurate response rates for internet surveys where the denominator is unknown. Selection bias is worsened with lower response rates, and is therefore most likely in our consumer, family, and advocate groups. Bias among consumer respondents is suggested by the very high rate of AD use reported by our consumer respondents. The completion rate of ADs is much higher than previous surveys of psychiatric advance directive use, which have found very low rates of AD completion among consumers. For example, a study of psychiatric AD use in 5 cities reported a prevalence of 3.9–12.9% completion of ADs (Swanson et al. 2006b), far lower than the 49% completion rate found in the current survey. The elevated rate is probably due to the nature of our consumer respondents, who are all active members of consumer advocacy groups and therefore likely better informed about and more interested in consumer-directed tools to increase treatment autonomy. However, several attitudinal questions on this survey were identical to a prior study completed in North Carolina in which consumers with severe mental illness were randomly selected from community clinics (Swanson et al. 2006a; Van Dorn et al. 2006), and responses to those questions were not significantly different from the responses of consumers in the current survey. This is somewhat reassuring regarding the generalizability of the current survey findings.

The prospects for successful implementation of ADs in Virginia seem reasonably promising. The level of enthusiasm and support among all stakeholder groups is very high. It is significantly higher than in previous surveys in North Carolina (Elbogen et al. 2006; Swanson et al. 2003), perhaps because enactment of the law was preceded by a three-year period of consensus-building among the stakeholders by the Commission. In addition, a coordinated implementation effort was undertaken by the stakeholder leadership immediately after the law went into force, taking a multi-pronged approach. The necessity of “bottom-up” as well as “top-down” adoption is evident in the difference between clinician and administrator responses to this survey. Although it does not appear necessary to convince people of the theoretical advantages of ADs, each stakeholder group has areas where additional education is needed. Consumers and families need to feel confident that, if they complete an AD, it will be respected and adhered to by clinicians. Targeted, ongoing training sessions are needed for clinicians to increase their comfort level with administering and using ADs. Additionally, potential barriers to use in crisis settings need to be addressed, for example through AD registries that would allow clinicians to locate ADs even if consumers were psychotic at the time. If clinicians and administrators remain skeptical that ADs will be used in a crisis, they are unlikely to invest the significant time and effort needed to implement their facilitation and use.