The authors are affiliated with the Department of Psychiatry and Behavioral Sciences, Duke University Medical Center, Box 3837, Durham, NC 27710 (e-mail: firstname.lastname@example.org).
Psychiatric advance directives allow competent persons to declare preferences and instructions for future mental health treatment in the event of an incapacitating crisis (1). They are transportable documents conveying information about patients' treatment preferences and history, including relevant medical conditions, emergency contact information, and medication side effects (1,2). Psychiatric advance directives aim to enhance patient autonomy when patients are most vulnerable and in need of quality care (3).
Studies show that most people with mental illness would complete a psychiatric advance directive if given assistance (4,5). Research suggests that psychiatric advance directives contain clinically useful information and rarely include medically inappropriate information (6,7). However, clinicians have voiced concerns (8,9). Providers have worried that psychiatric advance directives may be used to enforce treatment refusals (5,8).
Although patients can refuse treatment in their psychiatric advance directives, they can also explicitly state their reasons for these decisions. If patients cited reasons for refusal that providers felt were understandable, providers might be more willing to follow the directives. Conversely, if patients provided bizarre or incomprehensible reasons for refusal, clinicians might be less willing to comply with their requests. The purpose of this report is to examine whether a patient's reason for refusing treatment affects the likelihood of implementation of the psychiatric advance directive by clinicians.
Our study sampled 597 mental health professionals in North Carolina, including 167 psychiatrists and 237 clinical psychologists via a mailed questionnaire, with an additional 193 clinical social workers via an online survey. Data were collected between June 2004 and December 2004. We randomly selected psychiatrists and psychologists from state professional organization membership rosters and had a response rate of 32% for psychiatrists (167 of 522 psychiatrists) and 48% for psychologists (237 of 494 psychologists). Analyses showed no differences between responders and nonresponders. For logistical reasons, social workers were solicited by their professional organization through online newsletters linking to the survey. Because of this sampling method, we could not determine response rate for social workers.
Mean age of the respondents was 47.5±12.0 years (median 49 years, range 22—88 years), and 57% (330 of 581 who responded to question) were women. Eighty-six percent (N=515) of the sample were Caucasian, 3% (N=17) were African American, 2% (N=12) were Asian, 2% (N=12) were Hispanic, and 7% (N=41) did not identify their race or ethnicity. All participants were reimbursed $50 after completing the survey. The institutional review board of Duke University Medical Center found the study exempt from human subjects research review; therefore, explicit informed consent was not required.
Clinicians were presented with two scenarios, one in which a patient refused medications for a delusional reason and one in which conventional medical reasons were provided. Clinicians first read, "Mr. Smith is a 39-year-old white male with a long history of schizophrenia. He documents in a psychiatric advance directive he refuses psychotropic medications because he believes the FBI wishes to poison him. He currently presents as psychotic in the ER, has no history of violence, and shows the doctor the psychiatric advance directive, threatening to sue if it is not followed." Respondents were asked whether the psychiatric advance directive should be followed, using a 5-point scale (1, no; 2, probably not; 3, maybe; 4, probably; 5, definitely).
Clinicians then read, "Consider the same scenario as above, except Mr. Smith refuses medications because he has read some medications increase risk of diabetes and because he has a history of uncomfortable side effects including tardive dyskinesia, dry mouth, drowsiness, and confusion." Clinicians used the same scale to rate whether they thought this psychiatric advance directive should be honored.
To control for covariates, we measured clinicians' age, gender, race, profession, and work environment. Participants were asked about perceptions of the law, barriers to psychiatric advance directives, and knowledge of criteria for capacity assessment. Clinicians rated how much they valued family opinion and clients' autonomy. They also rated how much importance they placed on clients' insight, clients' cognition, and clients' history of unacceptable behaviors (such as violence, substance abuse, and suicide attempts) when deciding to follow clients' preferences.
We used chi square analyses to examine whether honoring psychiatric advance directives—defined as probably or definitely following Mr. Smith's request—differed by clinical profession or between the two scenarios. We employed multivariate logistic regression using the SPlus statistical software program to control for covariates in the relationship between the reason for refusal and clinicians' willingness to honor psychiatric advance directives.
Twenty-two percent of respondents (129 of 585 who responded to this question) reported that they would follow the psychiatric advance directive in which a delusional reason for refusal was given. For the conventional medical reasoning scenario, 72% (420 of 584 who responded to this question) reported that they would follow the psychiatric advance directive. Responses to the two scenarios were significantly different across the entire sample (χ2=47, df=1, p<.001).
Psychiatrists, psychologists, and social workers were all more likely to honor the psychiatric advance directive refusing treatment for conventional medical reasons (Table 1). There were also no significant differences in this pattern based on age, gender, or having worked in an emergency setting.
The decision to follow a psychiatric advance directive was regressed on reason for refusal in order to control for covariates. None of the covariates was significant, but the reason for refusal (delusional versus conventional medical) remained significant (adjusted odds ratio=2.49, 95% confidence interval=2.06—2.91, p<.001).
Our data suggest that clinicians who read psychiatric advance directives are implicitly concerned with the reasons for refusal of standard treatment as well as with the refusal itself. Specifically, psychiatric advance directives with conventional medical explanations for refusal of medications were much more likely to be respected than were psychiatric advance directives with refusals based on a delusional reason. In a previous study in which 164 psychiatrists read a hypothetical scenario of treatment refusal in a psychiatric advance directive with no reason given, 53% responded that they would honor the psychiatric advance directive (10). This value falls approximately halfway between our results showing 22% would honor psychiatric advance directives with delusional reasoning and 72% would honor psychiatric advance directives with conventional medical reasoning, indicating that reasons for refusal appear to shift clinicians' support of psychiatric advance directives in the expected direction.
Clinicians may be interpreting reasons given in the hypothetical psychiatric advance directive as evidence for Mr. Smith's competence at the time of writing it. In the absence of other evidence, delusional reasons for medication refusal may be sufficient for some clinicians to decide Mr. Smith was not competent when he wrote his psychiatric advance directive, and the directive is therefore not valid. Clinicians may also reason that it would be countertherapeutic to honor the delusional psychiatric advance directive but therapeutic to honor the conventional medical refusal, especially as it appeals to concerns about side effects and adverse outcomes. Another factor underlying clinicians' reasoning could be inherent ethical conflict between patient autonomy and welfare.
There are several study limitations. First, because clinicians' decisions were based on short vignettes that were limited in detail, we do not know whether clinicians would decide similarly in actual clinical cases involving treatment refusal. The short length of the vignettes may also have led to difficulty in interpretation. For example, a clinician who interpreted the second Mr. Smith as refusing only medications with specific side effects, rather than all medications, might have been more likely to honor the second directive. Second, we could not control for every covariate and inadvertently may have omitted important variables from the survey. Third, the results reflect North Carolina clinicians' opinions and are based on the statutes of that state; future research should study clinicians in other jurisdictions.
Although many people have worried that treatment refusals in psychiatric advance directives will be unacceptable to clinicians, the data presented here suggest that clinicians are more sophisticated in their decision making. Our results reveal that rather than automatically overriding treatment refusals, clinicians examine the underlying reasons for refusals before deciding how to proceed. If a psychiatric advance directive presents evidence of conventional medical reasoning behind treatment refusal, the survey suggests that most clinicians will try to honor their patient's wishes.
This study was supported by research grant R01-MH063949 from the National Institute of Mental Health and by Independent Research Scientist Career Award K02-MH67864. The study was also supported by the John D. and Catherine T. MacArthur Foundation Research Network on Mandated Community Treatment.
Dr. Swanson has received research support from Eli Lilly. Dr. Swartz has received research support from Bristol-Myers Squibb, Pfizer, and Eli Lilly; speaking honoraria and travel funds from AstraZeneca; and educational fees from GlaxoSmithKline. Dr. Van Dorn has received research support from Eli Lilly.