1158 ADVANCE DIRECTIVES Lahn et al. ADVANCE DIRECTIVES IN NURSING HOME TRANSFERS CLINICAL PRACTICE Advance Directives in Skilled Nursing Facility Residents Transferred to Emergency Departments MICHAEL LAHN, MD, BENJAMIN FRIEDMAN, MD, POLLY BIJUR, PHD, MARIANNE HAUGHEY, MD, E. J. GALLAGHER, MD Abstract. Objective: Ten years have passed since Congress enacted the Patient Self-Determination Act to promote the use of advance directives (ADs). This study was performed to determine the frequency, type, demographic distribution, and utility of ADs that accompany residents of skilled nursing facilities (SNFs) transferred to emergency departments (EDs). Methods: This was an observational, cross-sectional cohort of SNF residents, transferred to two urban, academic EDs. Chart review and physician interviews were conducted on consecutive patients arriving during 12-hour data collection shifts. Results: Among 715 patients entered, 315 [44%, 95% confidence interval (95% CI) = 40% to 48%] had an AD. Advance directives were significantly more prevalent among white (50%) than African American (34%) or Hispanic (39%) patients (p < 0.001), and varied from 0% to 94% among SNFs. Of the 315 patients with ADs, do-not-resuscitate (DNR) orders were the most prevalent (65%, 95% CI = 58% to 69%). Although 75% (95% CI = 69% to 81%) of the DNR orders addressed cardiopulmonary resuscitation (CPR), only 12% (95% CI = 8% to 16%) addressed intubation. Among 39 patients who required intubation or CPR, 44% had ADs, 82% (95% CI = 57% to 96%) of which were deemed useful. Conclusions: Despite a decade of legislation promoting their use, ADs are lacking in most SNF residents transferred to EDs for evaluation and in most settings in which a clinical indication exists for intubation or CPR. Variation in their prevalence appears to be associated with both ethnicity and SNF origin. Although about three-fourths of DNR ADs addressed CPR, only about one in ten offered guidance regarding intubation. When available, ADs are used in most instances to guide emergency care. Key words: advance directives; nursing homes; resuscitation orders; living wills; emergency treatment; cardiopulmonary resuscitation. ACADEMIC EMER- GENCY MEDICINE 2001; 8:1158 1162 TEN YEARS have passed since Congress enacted the Patient Self-Determination Act (PSDA) of 1991 1 in an effort to protect individuals rights concerning their medical care. A provision of the PSDA was designed to promote the use of advance directives (ADs), which are patient-directed guides or proxy designations intended to help coordinate medical care when individuals are unable to do so for themselves. Limited studies have addressed the effectiveness of the PSDA. 2,3 The issues of end-of-life care commonly arise in the emergency department (ED). These issues are particularly relevant to residents of skilled nursing facilities (SNFs) who are, in general, closer to the end of their lives than the general population. Decisions regarding end-of-life care are commonly made in the first minutes of ED treatment. Informative ADs that specifically address resuscitation options would be of great value to the emergency physician (EP). We know of no studies to date describing the characteristics of ADs in the specific cohort of SNF residents transferred to EDs for medical evaluation. The aim of this study was to estimate the frequency, types, and utility of ADs in this group of patients. From the Department of Emergency Medicine, Albert Einstein College of Medicine (ML, BF, PB, MH, EJG), Bronx, NY. Received April 3, 2001; revision received August 2, 2001; accepted August 14, 2001. Presented at the SAEM annual meeting, San Francisco, CA, May 2000. Address for correspondence and reprints: E. J. Gallagher, MD, Department of Emergency Medicine, Montefiore Medical Center, 111 East 210th Street, Bronx, NY 10467. Fax: 718-798- 6084; e-mail: jgallagh@montefiore.org METHODS Study Design. The study was a prospective, observational, cross-sectional survey of residents of SNFs, transferred to an ED for medical evaluation and treatment. Consecutive patients were enrolled during 96 twelve-hour shifts distributed among day, evening, and night tours in proportion to patient volume. Data were gathered using data col-
ACADEMIC EMERGENCY MEDICINE December 2001, Volume 8, Number 12 1159 lectors trained in the study protocol. The data collectors were also trained in how to audit charts and records that accompanied patients from the SNFs as well as what the various ADs were and looked like. The study was approved by the Institutional Review Board (IRB) of the Montefiore Medical Center and the Committee of Clinical Investigations (CCI) of the Albert Einstein College of Medicine. Since this was an observational study that did not affect patient care in any way, informed consent was waived by the IRB and the CCI. Study Setting and Population. The patients were enrolled from the EDs of two large, urban academic centers with a combined adult ED census of 130,000 patients. Both centers care for a high percentage of elders, and each center routinely receives residents from two different groups of SNFs. Data were gathered by review of transfer documents, direct interview of patients, and the EP primarily responsible for the care of that individual. TABLE 1. Distribution of Patient Characteristics with and without Advance Directives (ADs) Patients with ADs % Patients without ADs % Total %(n)* All patients 44 56 100 (715) Sex Male 38 62 100 (250) Female 48 52 100 (459) Race/ethnicity White 50 50 100 (403) African American 34 66 100 (185) Hispanic 39 61 100 (110) Age <65 years 34 66 100 (87) 65 years 46 54 100 (619) *Numbers do not add to 715 due to missing data: 6 patients were missing data on sex, 17 on race/ethnicity, and 9 on age. p = 0.01; 2 = 6.66. p = 0.001; 2 = 14.33. p = 0.04, 2 = 4.04. Study Protocol. A data collection instrument was completed for all eligible subjects. All patients transferred from a SNF, inpatient hospice, or other inpatient chronic care facility to one of the two EDs were entered. Patients transferred from geriatric day-care centers, home hospice, or other nonresidential facilities were excluded. Information on the presence, type, and content of ADs was obtained by examining documents accompanying the patient. Family members and/or SNF personnel were contacted if clarification of an AD was needed. Skilled nursing facilities were not routinely called looking for ADs unless a family member or other visitor advised us that documentation was possibly not transferred with the patient. Three types of ADs were identified: the living will (LW), the designation of a health care proxy (HCP), and the do-not-resuscitate (DNR) order. LWs are written documents that express in advance your specific instructions and choices about various types of medical treatments and certain medical conditions. They are recognized as evidence of your wishes if you are seriously ill and unable to communicate. 4 A HCP is a patient-appointed surrogate who decides medical care when patients lose the ability to decide for themselves. 4 A DNR order tells medical professionals not to perform cardiopulmonary resuscitation (CPR) in the event that a patient stops breathing or his or her heart stops beating. We chose to incorporate DNR orders into the general category of ADs. Clearly if a patient fills out a DNR order while still having the capacity to do so, this meets the previously stated definition of an AD. If a family member or close friend consents to a DNR order on an incapacitated patient, this is still a prospective way of influencing future medical decisions. In New York State, the decision by the surrogate can be based on the patient s previous discussion with the surrogate or what the surrogate believes is in the patient s best interest. 5 In New York State, a DNR order from a SNF or other hospital remains effective until an attending physician cancels it or issues an order continuing it. 6 We believed that the DNR order could be interpreted as a form of AD and hence decided to report its prevalence with the two more traditional ADs: the LW and the HCP. Utility of the AD was assessed by determining the proportion of each type of AD that addressed critical medical decisions regarding CPR and intubation. The DNR orders and LWs were examined to see whether CPR and intubation were addressed. For ADs that designated a HCP, the proxy was called if the doctor indicated a need to discuss the patient s care. The proxy was then asked whether he or she had discussed CPR and/or intubation with the patient. A second approach to assessing usefulness of the AD was to ask physicians whose patients had an indication for intubation or CPR whether the AD was helpful to them in caring for the patient. Intubations and performance of CPR were recorded. Data Analysis. Data were entered into Epi-Info (Ver. 6, USD, Inc., Snellville, GA) and exported into SPSS (Ver. 9, SPSS Inc., Chicago, IL) for statistical analysis. Ninety-five percent confidence intervals (95% CIs) were used to describe the precision of estimates. For comparison of proportions, the re-
1160 ADVANCE DIRECTIVES Lahn et al. ADVANCE DIRECTIVES IN NURSING HOME TRANSFERS Figure 1. Distribution of advance directives (ADs) in 18 skilled nursing facilities. sults of chi-square testing with a p-value were reported. RESULTS Seven hundred fifteen patients were enrolled in the course of 96 twelve-hour shifts. The mean age of the study subjects was 79 years (SD 13). Eighty-eight percent of the patients were 65 years old or older. Females composed 65% of the subjects entered. The majority of the subjects (57%) were white; 26% were African American, and 16% were Hispanic. Eighty-five percent of the study subjects came from one of 17 SNFs, each of which transferred at least ten residents during the study period. Fifty-two percent of the subjects were transferred from one of five SNFs. Forty-four percent (95% CI = 40% to 48%) of the subjects had an AD present (n = 315). Advance directives were significantly more common among females, whites, and patients aged 65 years or older (Table 1). The proportion of patients presenting with ADs varied widely among transferring facilities (Fig. 1). Among SNFs with ten or more transfers, the proportion of patients with ADs ranged from 0% to 94%. Of the 315 subjects with ADs, 64% (95% CI = 59% to 69%) had a DNR order present, 60% (95% CI = 55% to 65%) had a HCP documented, and 12% (95% CI = 8% to 16%) had LW documentation. Thirty-six percent of these patients had more than one type of AD present. The most common combination of different types of ADs was DNR order and HCP. Information about whether the AD addressed CPR and intubation was gathered directly from documents accompanying the patient for those with DNR orders and LWs. For patients with a HCP, this information was obtained only from proxies whom the treating physician contacted. Physicians thought that 39 of the 189 (21%) proxies should be contacted, and were able to do so in 90% of the cases (35/39). All of the proxies were aware of their status as the HCP. Table 2 shows the proportion of patients with each specific type of AD whose AD addressed CPR and intubation. Approximately three-fourths of all three types of ADs addressed CPR. In contrast, only 12% of DNR orders addressed intubation, while about 70% of the HCPs and LWs addressed this issue. Cases where a DNR order did not address either aspect of resuscitation occurred in instances where the facility s transfer sheet stated DNR without confirmatory documentation. There was a clinical indication for intubation in 39 patients, ten of whom also had an indication for CPR. Of these 39 patients, 17 (44%, 95% CI = 28% to 60%) had an AD present. Among these 17 patients, the treating physician found the AD to be useful in guiding care in 14 patients, or 82% (95% CI = 57% to 96%). Only one of the ten patients with indications for CPR had an AD present, and it was found to be useful in the decision not to resuscitate the patient. DISCUSSION Residents of SNFs are commonly transferred to EDs for medical evaluation and treatment. Some of these patients are critically ill and in need of immediate resuscitation. Emergency physicians typically lack a prior relationship with their patients, and are consequently unaware of the individuals wishes. Among patients unable to communicate treatment preferences, ADs are extremely helpful. In their absence, physicians ordinarily err on the side of resuscitation, since the alternative may result in a wrongful death. Commonly, however, well-intentioned efforts to reverse a terminal event result only in its temporary postponement rather than a meaningful extension of life. Prolongation of dying is too often accompanied by needless suffering of patients and families, while drawing heavily on scarce health care resources. We found that a minority (44%) of SNF resi-
ACADEMIC EMERGENCY MEDICINE December 2001, Volume 8, Number 12 1161 dents seen in our EDs had an AD available. Because other ED-based studies were not restricted to elder or institutionalized patients, the prevalence of ADs in our study is nearly double that reported by previous investigators (22%, 27%, and 23%). 7 9 This higher prevalence is still disturbingly low when one considers that the Congressional Patient Self-Determination Act of 1991 clearly requires all hospitals and other institutions receiving Medicare or Medicaid funding to promote the use of ADs. We found variability among the different types of ADs used in our patients. Do-not-resuscitate orders and HCPs were over four times more prevalent than LWs. There was little variability in how well the various ADs addressed CPR: DNR, 75%; LW, 72%; and HCP, 72%. However, DNR orders addressed intubation in only 12% of the cases compared with 69% of the HCPs contacted and 71% of the LWs. We were able to show some differences in possible usefulness among the three types of ADs found. Do-not-resuscitate orders are common but are limited to resuscitation and only rarely address intubation. They are in fact misused when their presence is used to guide management in nonresuscitation issues. 4 Health care proxies are easy to fill out and require a low level of decision-making ability (a patient chooses a person who will make decisions for the patient when he or she cannot). Health care proxies offer flexibility because they can be applied to nonresuscitation issues and, when used correctly, allow the patient s views to be heard even after the loss of the ability to communicate. A limitation to their use in our study was that EPs were unable to contact proxies in 10% of the cases. This contact requirement inherent to the use of HCPs may not be a significant problem in a nonemergency decision, where time is not of the essence. Another weakness in the use of the HCP was demonstrated by the lack of previous discussion of intubation (31% never discussed) and CPR (28% never discussed). A proxy s decisions should be based on previous discussion with the patient. 4 Living wills have the potential for having the most utility for EPs since they frequently address common emergency scenarios that need immediate decisions. Our data showed that LWs were in fact more likely to address both intubation and CPR than DNR orders. A limitation to their use is that they do not address all the possible treatments or conditions that may arise. Living wills also require a high level of capacity to decide on these very complex decisions. This limits their use in elders, who may not have such decision-making abilities. The prevalence of LWs in our patient population was low. TABLE 2. Percentage of Advance Directives that Addressed Cardiopulmonary Resuscitation (CPR) and Intubation, by Type of Advance Directive Type n Addressed CPR % (95% CI) Addressed Intubation % (95% CI) Do-not-resuscitate order 201 75 (69, 81) 12 (8, 16) Health care proxy 35* 72 (55, 83) 69 (55, 83) Living will 38 72 (58, 86) 71 (57, 85) *Thirty-nine warranted telephone contact; 35 were successfully contacted. LIMITATIONS AND FUTURE QUESTIONS Our study was undertaken in only two urban EDs within the same county. This may diminish the ability to generalize these results to other geographic regions. A striking disparity in the prevalence of ADs was found among the various SNFs. Methods that one facility used to generate a high prevalence should be investigated and propagated. Indeed, a clinical trial of such an undertaking would help to identify realistic strategies for dealing with this important public health problem. There is a possibility that SNF residents had ADs, but were transferred without their ADs accompanying them. This would have increased the percentage of SNF residents having an AD, but would not have any clinical relevance since an AD unavailable to the treating physician cannot alter patient management. Another area for future investigation is the persistent ethnic gap that we and other investigators 7 have found in the use of ADs. We do not have sufficient data to assess the extent to which this might be associated with race rather than ethnicity, nor can we determine whether this association might be confounded by socioeconomic status. One of the study s predetermined goals was to examine the utility of various ADs with respect to intubation and CPR. Our low frequency of intubation and CPR among patients with ADs made it unlikely that we would detect a difference in ED physician perceptions of utility if one were in fact present. CONCLUSIONS Despite legislation promoting their use, ADs are found in less than half of the residents of SNFs transferred to our ED for evaluation. There is large variability in the prevalence of ADs among various SNFs. Ethnicity maybe a factor in the utilization of ADs. The most prevalent AD is the DNR order. Despite its popularity, the DNR order infrequently addresses intubation. When available, ADs are used by physicians to guide emergency care.
1162 ADVANCE DIRECTIVES Lahn et al. ADVANCE DIRECTIVES IN NURSING HOME TRANSFERS References 1. Omnibus Budget Reconciliation Act of 1990. Pub. No. 101-508, Sections 4206 and 4751. Washington, DC: Government Printing Office, 1990. 2. Mezey M, Mitty E, Rappaport M, Ramsey G. Implementation of the Patient Self-Determination Act (PSDA) in nursing homes in New York City. J Am Geriatr Soc. 1997; 45:43 9. 3. Park D, Eaton TA, Larson EJ, Palmer H. Implementation of the Patient Self-Determination Act: a study of its implementation. South Med J. 1994; 87:971 7. 4. Department of Health, State of New York. Your Rights as a Hospital Patient in New York State. Albany, NY, Jan 1998, pp 7 8, 21 27. 5. New York State Public Health Law: Chapter 45, Article 29-B Orders Not To Resuscitate. 2965. Albany, NY, 2001. 6. New York State Public Health Law: Chapter 45, Article 29-B Orders Not To Resuscitate. 2971. Albany, NY, 2001. 7. Ishihara KK, Wrenn K, Wright SW, Socha CM, Cross M. Advance directives in the emergency department: too few, too late. Acad Emerg Med. 1996; 3:50 3. 8. Llovera I, Mandel FS, Ryan JG, Ward MF, Sama A. Are emergency department patients thinking about advanced directives? Acad Emerg Med. 1997; 4:976 80. 9. Llovera I, Ward MF, Ryan JG, et al. Why don t emergency department patients have advance directives? Acad Emerg Med. 1999; 6:1054 60.