Case Studies of Legal Issues End of Life

Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=wswe20

Journal of Social Work in End-of-Life & Palliative Care

ISSN: 1552-4256 (Print) 1552-4264 (Online) Journal homepage: https://www.tandfonline.com/loi/wswe20

Life-Sustaining Treatment Decisions: A Social Work Response to Meet Needs of Health Care Surrogates

Julia W. Buckey & Neil Abell

To cite this article: Julia W. Buckey & Neil Abell (2010) Life-Sustaining Treatment Decisions: A Social Work Response to Meet Needs of Health Care Surrogates, Journal of Social Work in End-of- Life & Palliative Care, 6:1-2, 27-50, DOI: 10.1080/15524256.2010.489221

To link to this article: https://doi.org/10.1080/15524256.2010.489221

Published online: 11 Jun 2010.

Submit your article to this journal

Article views: 678

View related articles

Citing articles: 11 View citing articles

PEER-REVIEWED ARTICLES

Life-Sustaining Treatment Decisions: A Social Work Response to Meet Needs of

Health Care Surrogates

JULIA W. BUCKEY University of Central Florida, Orlando, Florida, USA

NEIL ABELL Florida State University, Tallahassee, Florida, USA

Surrogates must possess essential patient information prior to legitimately exercising legal and moral obligations to act on patients’ medical care preferences. This descriptive, in vivo study examined factors influencing surrogate and proxy decisions (N ¼ 132) following life-sustaining treatment decisions. Patient communication and self-efficacy variables accounted for approxi- mately 38% of the variance in surrogates’ perceptions of benefits= barriers associated with decision making. Guided by patients’ advanced communication, respondents (97.8%) expressed high self-reliance and significant appreciation of benefits associated with their decisions. Clarifying surrogates’ and providers’ under- standing of patient care preferences during ICU=CCU admission may facilitate better adherence to patient wishes.

KEYWORDS end-of-life care, health care proxy, health care surrogate, Health Care Surrogate Preferences Scale, life-sustaining treatment decisions

In nearly 80% of the 2.5 million deaths that occur in U.S. hospitals each year, surrogates make 70–90% of the decisions to withhold or withdraw life-sustaining treatment from critically ill patients in the intensive care unit (ICU; Mcgee, 2000; Prendergast & Luce, 1997; Swigart, Lidz, Butterworth,

Received 17 January 2010; accepted 22 April 2010. Address correspondence to Julia W. Buckey, PhD, ACSW, University of Central Florida,

HPA1 Room 255, P.O. Box 163358, Orlando, FL 32816, USA. E-mail: [email protected]

Journal of Social Work in End-of-Life & Palliative Care, 6:27–50, 2010

Copyright # Taylor & Francis Group, LLC

ISSN: 1552-4256 print=1552-4264 online

DOI: 10.1080/15524256.2010.489221

27

& Arnold, 1996). Therefore, surrogates potentially influence nearly 1.5 million end-of-life treatment decisions per year. As the number of older people continues to grow, more and more Americans will likely experience this decision-maker role at some period during their lifetime.

An overall lack of advanced surrogate knowledge of patient care preferences complicates surrogate decision making. Surrogates better understand what is important to patients at the end of life when they have had detailed discussions with them. However, limited awareness about specific treatment choices and a lack of understanding of surrogate role requirements have remained largely unaddressed in prior examinations of surrogate and patient communications (Engelberg, Patrick, & Curtis, 2005; Gutheil & Heyman, 2005; Truog et al., 2001). Although differences often exist between what patients and their surrogates deem important end-of-life tasks, advanced knowledge of patient preferences has been found a necessary, but not sufficient, ingredient for surrogate action (Engelberg et al., 2005; Hines et al., 2001; Layde et al., 1995; Suhl, Simons, Reedy, & Garrick, 1994).

For many Americans, the recent uproar about so-called ‘‘death panels’’ (Lepore, 2009) evoked images of medical care driven by external mandates and little individualized choice. This underscored both public and political misunderstanding of the tremendous challenges faced by those making decisions in support of a loved one’s medical care at life’s end. In addition to highlighting the resistance and barriers associated with end-of-life discus- sions in American culture, this debate has emphasized Americans’ fears of ‘‘dying badly’’ (Court, 2009). At the same time, the debate countered the status quo by bringing to light the crucial importance of explicit com- munication by persons charged with making respectful health care choices on behalf of a loved one whose preferences would otherwise remain unspoken.

Social work professionals providing services to surrogates or extended family members involved in decision making in the ICU generally have relied on extensive professional training and experience in delivering short-term, crisis, or other brief treatment interventions to the surrogate population. Although the very public life and death of Terri Schiavo reminded us of the potentially negative and long-lasting consequences possible in relation- ships among patients’ survivors (Ditto, 2006), social work professionals have had few empirical findings available to guide them as they assist surrogates to discharge their duty of making treatment decisions. For that reason, this study sought to answer the question, ‘‘Did surrogates’ personal attributes, or their perceptions of communications, social support, or self-efficacy directly influence their perception of the benefits and barriers associated with their life-sustaining treatment decisions?’’ In other words, how well did these factors predict participants’ sense that they had done the right thing for their loved one?

28 J. W. Buckey and N. Abell

LITERATURE REVIEW

Development of Surrogate Roles and Responsibilities

During the late 20th century, the health care surrogate role emerged as an outgrowth of three interrelated developments: application of biomedical technology to critically ill patients, development of the field of bioethics, and judicial and legislative decisions directing surrogate and proxy actions (Health Care Advance Directives Act, 2009).

In recognizing the right of parents to speak and act on behalf of family members, federal and state courts acknowledged the health care surrogate role (Cruzan v. Director, Missouri Department of Health, 1990; Quinlan v. New Jersey, 1976). These extensions of individual patient rights from competent to incompetent patients paralleled passage of the Patient Self-Determination Act (PSDA, 1990), directing that all Medicare and Medi- caid patients be allowed to express their individual preferences for health care in advance of receiving medical treatment. This resulted in each state eventually moving to establish evidentiary standards that would determine specific conditions under which patients’ previously expressed treatment wishes would be honored. Consequently, individual state statutes have produced a diverse surrogate nomenclature consisting of a number of inter- changeable titles for patients’ representatives, such as health care surrogate, durable power of attorney for health care (DPAHC), health care proxy, or guardian. This language also reflected the varying surrogate duties and pro- cesses assigned within each state (Cruzan v. Director, Missouri Department of Health, 1990; Health Care Advance Directives Act, 2009).

Initially, research on surrogate activity focused research attention on medical patient decision making (Smeidra et al., 1990) or on patient- surrogate treatment decision concordance to gauge surrogate accuracy in predicting patient treatment preferences (Cogen et al., 1992; Hare, Pratt, & Nelson, 1992; Seckler, Meier, Mulvihill, & Cammer-Paris, 1991). Yet, these early studies produced inconsistent findings and offered little insight into the actual surrogate decision-making process (Engelberg et al., 2005). Instead, they produced an initial view of surrogate performance that persists today: surrogates were ‘‘unreliable’’ patient representatives, and their decisions ‘‘are no better than random chance’’ (Suhl et al., 1994, p. 90).

More recent investigations have inquired into amounts and types of preliminary patient-surrogate communications, to move surrogate decision making from its initial presentation of surrogate duties limited to a simple, isolated decision, to a series of complex, multi-tasked responsibilities orche- strated by individuals who represent incapacitated patients (Buckey, 2007; Limerick, 2007; Swigart et al., 1996; Thelen, 2005; Vig, Taylor, Starks, Hopley, & Fryer-Edwards, 2006). Patient-surrogate communication preceding the patient’s need for medical treatment emerged as an instrumental component

Life-Sustaining Treatment Decisions 29

in surrogate decision-making studies. When surrogates knew patient wishes, they were committed to following them; when they did not, surrogates reported high levels of decision conflict and poorer post-mortem adjustments (Buckey, 2007; Carr & Moorman, 2008; Seckler et al., 1991; Wright et al., 2008).

Increasingly, examinations of surrogacy have revealed a stressfully complex assortment of tasks. When they knew little about patient care prefer- ences, surrogates displayed an ‘‘insatiable’’ need for information (Abbott, Sago, Breen, Abernethy, & Tulsky, 2001, p. 200), including human physiology, trajectory of patient illnesses, prognoses, and likely responses to treatment (Hiltunen, Medich, Chase, Peterson, & Forrow, 1999; Jacob, 1998; Swigart et al., 1996). Surrogates evaluated patients’ medical and emotional conditions together with prior illness histories, and factored what patients said into their decisions (Abbott et al., 2001; Hiltunen et al., 1999, Jacob, 1998; McNabney, Beers, & Siebens, 1994; Swigart et al., 1996). Less knowledge about patient preferences increased family conflict, surrogates’ personal distress, and number of treatment decisions required (Abbott et al., 2001; Hiltunen et al., 1999; Jacob, 1998; Lautrette et al., 2007).

Surrogates may feel both moral and ethical obligations to respect patient treatment wishes, regardless of their personal discomfort (Buckey, 2007; Swigart et al., 1996). Those who honored patients’ wishes reported greater suc- cess in post-mortem adjustment, and those who failed to do so reported more problems after the patient died (Jacob, 1998; Lautrette et al., 2007). Surrogates attempted to make decisions congruent with patient, surrogate, and family values, while avoiding decisions likely to increase patient pain or complicate treatment (Abbott et al., 2001; Hiltunen et al., 1999; Jacob, 1998; Vig et al., 2006). Decisions to withhold treatment were often based on patients’ poor prognoses, lack of mental acuity, and adverse reactions to treatment (Cogen et al., 1992; Hare et al., 1992; Layde et al., 1995; McNabney et al., 1994; Smedira et al., 1990).

Surrogates were routinely emotionally closer to patients than other family members, and often assumed a role of leadership in the family, expanding their responsibilities to care for other family members prior to, during, and following patient illness. These family leadership skills appeared to expedite decision making and included efforts to bring together family members divided in their preferences for patients’ medical care and treat- ment (Swigart et al., 1996).

Surrogates often view physicians as primary information sources about patient condition and prognosis, leading them to expect physicians to be available and involved during decision making (Engelberg et al., 2005; Lewis et al., 2006). Physician traits, such as availability, openness, and support, enhanced surrogate trust and produced greater adherence to physicians’ treat- ment recommendations (Abbott et al., 2001; Hiltunen et al., 1999; Jacob, 1998; Swigart et al., 1996). Conversely, and especially regarding decisions to forgo treatment, lower trust levels created greater decision-making difficulty

30 J. W. Buckey and N. Abell

(Hiltunen et al., 1999; Swigart et al., 1996). Collectively, surrogates’ subjective impressions of patient conditions, combined with information acquired from physicians, medical staff, and family members, aided surrogate decision making while mitigating their distress (Abbott et al., 2001; Jacob, 1998; Johnson et al., 2000; Hiltunen et al., 1999; McNabney et al., 1994; Swigart et al., 1996; Wright et al., 2008). In acquiring reliable information from multiple sources, surrogates made decisions after processing information obtained from patients and significant others; i.e., physicians, nurses, or family members (Hiltunen et al., 1999; Jacob, 1998; Johnson et al., 2000; Limerick, 2007; Swigart et al., 1996; Vig et al., 2006).

Incorporating Theory

To offset one major omission from prior studies, a conceptual model of surrogate decision making with variables based on relevant theory, the Health Belief Model was designed for this study (HBM; Rosenstock, Strecher, & Becker, 1988). The HBM has previously been applied to gauge individual preventive health, illness, and sick role behaviors and to emphasize readi- ness to act during periods of uncertainty. Although the HBM has not been applied to the health-related behaviors of persons acting on another’s behalf, the core of this value-expectancy model can be applied to surrogates who anticipate that particular actions taken on behalf of a critically ill patient will achieve a desired outcome. Their evaluation of potential patient benefits and barriers, or disadvantages, associated with treatment choices is central to the Surrogate Decision-Making Model (see Figure 1). For example, surrogates

FIGURE 1 Proposed Surrogate Decision-Making Model.

Life-Sustaining Treatment Decisions 31

may experience conflict when instituting a treatment possessing both advantages and disadvantages, such as stabilizing the patient’s medical condition while simultaneously risking greater medical complications resulting from the treatment choice.

We also adapted a key group of HBM social cognitive antecedents thought to precede individual surrogate health behaviors and modify factors influencing their action: demographic characteristics (age, gender, education, and ethnicity), sociopsychological attributes and beliefs (emotional closeness to patient and perceived social support), and structural influences (patient and physician communications). Thus, the model represents the interactive, psychosocial approach to surrogate decision making as it acknowledges the social and psychological realities of individual health decision behavior. The variable of self-efficacy was later introduced to enhance the model’s ability to account for greater variation in behavior (Rosenstock et al., 1988). Although self-efficacy had not been incorporated into subsequent HBM model tests as originally proposed by Rosenstock and colleagues, in other models examin- ing health behavior it had historically accounted for as much as 20% of the variance (Abraham & Sheeran, 2000). Thus, we included it here.

Physician assessment of other HBM variables, such as threat and suscep- tibility of severe (patient) illness, is a unique challenge in application of the HBM to surrogate decision making. These HBM variables are portrayed via physicians’ treatment recommendations, thus are not represented by surrogate decisions and were excluded from our model. Overall, the HBM represents an initial, interactive, psychosocial approach to surrogate decision making, reflecting the social-psychological reality of surrogate decision behaviors while incorporating the multiple influences potentially influencing health-related surrogate behaviors within health-related environments (Abraham & Sheenan, 2000). It offers a coherent approach to examine variables identified as being inconsistently measured in the surrogate literature, yet selectively found to influence surrogate decisions (Buckey, 2007; McNabney et al., 1994).

METHODOLOGY

While existing literature provides an initial profile of influences on surrogate decision making, methodological and conceptual limitations are often unadd- ressed, raising concerns about the validity of conclusions. George (2002) charges that researchers have failed to apply a uniform definition of surro- gacy. Others cite difficulties associated with prospective designs attempting to examine life-threatening treatment conditions before they have actually occurred, versus retrospective studies that risk bias when survivor recall may be clouded by personal bereavement and grief (Addington-Hall & McPherson, 2001). Finally, use and alteration of measurement scales with unknown or unreported psychometric properties compromised internal

32 J. W. Buckey and N. Abell

and statistical conclusion validity (Cogen et al., 1992; Hare et al., 1992; Smedira et al., 1990; Suhl et al., 1994). As a result, questions remain about the credibility of findings based on questionably measured constructs.

An attempt was made to address these and related limitations in our investigation of factors influencing surrogate and proxy decisions about life-sustaining treatments. As a result, this study differs from prior efforts in a number of ways. First, a uniform definition of surrogacy, encompassing multiple statutory definitions of adults functioning in the role of treatment decision maker on behalf of a critically ill medical patient was applied. In this study, surrogacy refers to ‘‘a competent adult who has not been expressly designated to make health care decisions for a particular incapacitated indi- vidual, but who, nevertheless, is authorized pursuant to s. 765.401 to make health care decisions for such individual’’ or ‘‘any competent adult expressly designated by a principal to make health care decisions on behalf of the prin- cipal upon the principal’s incapacity.’’ These definitions are compatible with the recent Florida Health Care Advance Directives Act (2009, Definitions sections, { 15 and { 16). Second, adhering to recommendations for conduct- ing research on end-of-life decision-making processes directly (Gessert, Curry, & Robinson, 2001), we conducted this study in real time, or while participants were actively engaged in health care surrogate decision making. People currently acting as surrogate and proxy for patients in the ICU=CCU (Critical Care Unit) were surveyed immediately after their decisions about a life-sustaining treatment and, where possible, before those decisions had been acted upon or their outcomes known; i.e., while patients continued to receive care. Thus, methodological concerns associated with the prospective– retrospective study debate were minimized as much as possible.

In addition to the conceptual framework described above, also employed were reliable, valid measurement tools; some, specifically developed and validated for use in this study (Buckey & Abell, 2004, 2009). A power analysis determined the required sample size needed to maximize the predictive power of the regression procedures employed (Hair, Black, Babin, Anderson, & Tatham, 2006). Again, the question addressed was ‘‘Did surrogates’ personal attributes, or their perceptions of communications, social support, or self-efficacy directly influence their life-sustaining treatment decisions?’’

Design

A real-time, descriptive, cross-sectional survey design was used to assess surrogate values, attitudes, and beliefs immediately after their decisions to initiate, withdraw, or withhold one of four life sustaining treatments: cardio- pulmonary resuscitation (CPR), mechanical ventilation, tube feedings (PRN or TPR), or dialysis. We surveyed a sample of surrogates from a diverse, heterogeneous population who were active in their decision-making roles.

Life-Sustaining Treatment Decisions 33

Sample

Respondents were adults making life-sustaining treatment decisions on behalf of incapacitated medical (ICU) patients during a current hospitaliza- tion, a prior hospitalization (such as with hospice patients), or for advanced Alzheimer’s patients being cared for at home. Surrogates qualified for the study if they had made at least one life-sustaining treatment decision on behalf of a patient and then continued in the surrogate role, and if their patient continued to receive some form of medical care.

Facilities were identified where surrogates might be found. One indicator of estimated surrogate activity was the number of hospital deaths, which also included patient age, gender, and race, reported quarterly by state-licensed hospitals to Florida’s Agency for Health Care Administration during the calen- dar years 2001–2003. Letters were sent to a broad range of hospital and other agency administrators seeking interest in study participation. Of 60 hospitals contacted, 8 administrators expressed study interest; of 18 hospices, 5 were interested, and no responses were obtained from the 55 long-term care facili- ties contacted. Final sample sites included three hospitals, two hospices, and two Alzheimer’s and Parkinson’s disease support agencies.

As the tertiary facilities engaged for this study (described below) maintain no formal records on surrogate action or on the utilization of life- sustaining treatment equipment, random sampling of study participants was not an option. Using a nonprobablity, purposive sampling approach based on the relevant characteristics of this highly specialized population (Henry, 1990), the sample blended respondents from rural with urban communities.

Measures

We used a combination of researcher-created items and standardized measurement scales to strengthen data reliability and statistical conclusion validity, and to increase the power of statistical tests employed (Tate, 1998). Of the 10 predictor variables represented in Figure 1, 4 were demographic— i.e., surrogate age, education level, gender, race-ethnicity—and 4 were single-item or summated rating scales developed for this study (physician and patient communications, religiosity-spirituality, and emotional closeness to the patient).

Existing standardized scales measured the remaining two predictor variables. Social support was measured by the 12-item Multidimensional Scale of Perceived Social Support (MSPSS; Zimet, Dahlem, Zimet, & Farley, 1988), and self-efficacy was measured by an adaptation of the 20-item Healthcare Surrogate Preferences Scale (HSPS; Buckey & Abell, 2004). While the original HSPS was designed to be predictive of self-efficacy (i.e., projecting what one would or could do), this view was modified for this study to ask participants to recall their judgments after the fact, following a surrogate decision. Thus,

34 J. W. Buckey and N. Abell

respondents had an opportunity to reflect on individual competencies required to conduct their duties while simultaneously evaluating personal beliefs about their role experience, including its tasks and responsibilities.

The Acceptance of Treatment Choice Inventory (ATCI), measuring surrogate perceptions of the advantages (benefits) and disadvantages (bar- riers) associated with evaluating life-sustaining treatment decisions, was used to measure the criterion variable, benefits, and barriers (Buckey & Abell, 2009). As established psychometrics are inherently sample dependent, reliability of all standardized measures were reassessed using Cronbach’s alpha (a) prior to hypothesis testing.

Although this study design was primarily quantitative, it also included a single qualitative item, ‘‘Is there anything else you feel we should know about your surrogate experience?’’ Located at the end of the survey, this item provided respondents an opportunity to share personal views on the subject.

Data Collection

Consent for use of human subjects was obtained from University and facility institutional review boards (IRBs), facility or agency administrators, and=or Boards of Directors. Potential study participants were approached in ICUs, hospice, or community support agency service locations to complete an anonymous, self-report survey about their feelings, attitudes, and beliefs following their decision to initiate, withhold, or withdraw patient treatment. Study participation was confidential and voluntary and the consent process specified a disclaimer that participation would in no way affect patient treatment.

Participants were requested not to place personal identifying information protected under HIPPA (1996); i.e., name, patient name, age, bed or room number, or any signature on the survey. Although participants were encoun- tered face-to-face when asked about their interest in the study, after data collection there was no way for the researcher to associate any response with any individual. Participants chose one of three methods of survey return: in person to the researcher or research associates, via facility drop boxes, or by mail. Other than survey completion, no other respondent activity was required. Counseling resources were listed in the packet to address potential respondent emotional reactivity.

Several facilities required the researcher to undergo HIPPA (2002) on-site training.The first author conducted on-site training to assist in standardizing data collection procedures for research associates who were recruited from facility staff members. Volunteer research associates were given study back- ground and data collection procedures, including paraphrasing of verbal scripts that were used during surrogate interactions and prior to survey administration. These associates used a verbal script to announce study avail- ability and assess study interest. HIPPA guidelines mandated that only facility

Life-Sustaining Treatment Decisions 35

staff could approach potential participants. After receiving an affirmative response, surrogates were personally presented a survey packet. Staff and potential respondents were familiarized with study procedures. After receiv- ing facility permission to post survey notices within facilities—i.e., in the ICU waiting room, on staff bulletin boards, or in newsletters—the researcher posted notices to publicize the study, ensuring researcher availability during day, evening, and night shifts.

Two hundred forty-six survey packets were distributed to potential respondents; 135 usable surveys were returned. The overall response rate of 55% was within the range of rates documented in the literature, from 37% (community mail survey, Hare et al., 1992) to 100% (medical inpatients, Suhl et al., 1994).

Data Management

Data analyses compared raw data runs using listwise deletion with runs using mean replacement values. As results of all model variables differed notably and yielded greater numbers of statistically significant correlations using mean replacement data, the more conservative, raw data values were used in data analyses (Tate, 1998). The influence of outliers, evident in four variables, resulted in deletion of three cases, yielding the total usable sample of 132. Sample sizes for specific analyses ranged from 105 to 128, each meeting criteria for adequate power in hypothesis testing. Data analyses were performed with SPSS 14.0.

RESULTS

Respondents

The average surrogate age was 58 years (SD ¼ 12.5), slightly less than the aver- age reported age of patients (59 years). Patient ages over 89 (n ¼ 22) were not reportable under HIPPA (2002) privacy regulations. Consequently, actual mean age for patients is unknown, though certainly higher than the figure reported here. Respondents were three-fourths female (78.8%), predomi- nately white (82.5%), and reported 14.2 years of formal education (M ¼ 14.2, SD ¼ 2.89). Gauging themselves to be very religious or spiritual (M ¼ 7.49; SD ¼ 2.22) on a scale from 1 ¼ not at all important to 9 ¼ extremely important, most surrogates (n ¼ 111, or 84.1%) felt extremely to very emotion- ally close to the patient. Only a minority (n ¼ 5, or 3.8%) reported feeling neutral or distant toward the patient.

Spouses were the largest relationship subgroup (n ¼ 49, 37.1%), followed by daughters (n ¼ 47, 35.6%), sons (n ¼ 13, 9.8%), and others (n ¼ 23, 17.5%). Religious affiliation was primarily Protestant, 73 (55.3%), followed by Catholic, 44 (33.3%), and Jewish, 3 (2.3%). Nearly one half of

36 J. W. Buckey and N. Abell

respondents (n ¼ 64, or 48.5%) withheld or did not begin treatment. Forty (30.3%) initiated treatment, and 28 (21.2%) withdrew treatment. The greatest number of decisions (n ¼ 59, or 44.7%) were about CPR=do not resuscitate orders (DNRs); nearly one third of surrogates (n ¼ 36, or 27.3%) made deci- sions about mechanical ventilation; nearly one fifth (n ¼ 23, or 17.4%) made decisions about tube feedings. A minority (n ¼ 13, or 9.8%) made decisions about dialysis; only one (.8%) failed to specify a decision. Overall, 104 (78.8%) surrogates strongly agreed with the statement, ‘‘Looking back, I can accept that my decision was the best alternative for the patient.’’

Surrogates reported having dealt with patient illness an average of 36.5 months (SD ¼ 56.2 months); item responses ranged from 1 to 31 years, indi- cating the time surrogates had been actively involved with patient illness; 7 respondents failed to respond to this item. Cardiac=cardiopulmonary disease and cancer, the two most prevalent surrogate-identified patient diagnoses, paralleled the prevalence of each disease category in the general U.S. popu- lation (n ¼ 24, 18.4%, respectively), followed by cerebrovascular disease (n ¼ 17, 12.8%; Centers for Disease Control and Prevention, 2009). Interested readers are referred to Buckey (2007) for more information on surrogate and patient characteristics resulting from this study.

Existing and Newly Created Measures

Results of existing and new measures of study variables contained in the surrogate decision-making model are shown in Table 1. The MSPSS (Zimet et al., 1988) produced a Cronbach’s coefficient alpha representing scale reliability of a ¼ .943 for social support (M ¼ 5.96, SD ¼ 1.15); the adapted HSPS (Buckey & Abell, 2004), a ¼ .877, for surrogate self-efficacy (M ¼ 6.16; SD ¼ .76). Standard error of measurement (SEM) coefficients were computed as indicators of measurement error for the current sample; for

TABLE 1 Measurement of Key Constructs: Patient Communication, Physician Communi- cation, Social Support, Surrogate Self-Efficacy, and Benefits-Barriers

Measure and variable a SEM M SD

Existing measures MSPSS Social support .94 .27 5.96 1.15

Adapted HSPS Self-efficacy .88 .27 6.16 .76

New measures Physician communication .92 .56 7.64 2.02 Patient communication .91 .42 8.49 1.43 ATCI Benefits subscale .73 .39 6.32 .76 Barriers subscale .68 .54 1.95 .98 Benefits-Barriers composite .69 .38 .88 .68

Life-Sustaining Treatment Decisions 37

the MSPSS, SEM ¼ .27, and for the HSPS, SEM ¼ .27, results well below the threshold value of .3 (5% of 6.0 ¼ .30), thus both variables were retained for model testing (Springer, Abell, & Hudson, 2002).

Two new, researcher-created dimensions of surrogate communication variables measured the degree of surrogates’ perception of communication helpfulness, assessing first, whether communication occurred, and if so, the means by which it was shared (verbal or written) between (a) surrogates and physicians, and (b) surrogates and patients. Using a Likert scale where 1 ¼ not very helpful at all, to 9 ¼ extremely helpful, Physician communication (M ¼ 7.64, SD ¼ 2.02) resulted in a Cronbach’s coefficient of a ¼ .922 (SEM ¼ .56); and Patient communication (M ¼ 8.49, SD ¼ 1.43), a Cronbach’s coefficient of a ¼ .913 (SEM ¼ .42). The standard errors of measurement (SEMs) for physician communication were considerably above, and for patient communication only slightly above, the target threshold (�.40).

The initial validation of the 20-item Acceptance of Treatment Choice Inventory (ATCI), an evaluation of surrogate perceptions of the Benefits and Barriers associated with treatment decision choice, has been detailed elsewhere (Buckey & Abell, 2009). ATCI items combined (a) HBM constructs reflecting the individual costs-benefits analyses thought to occur during individual medical treatment decision making (Rosenstock et al., 1988), and (b) the empirical literature findings that reflected the potentially beneficial or harmful results surrogates associated with treatment decisions (Jacob, 1998; Swigart et al., 1996). Benefits scores (representing advantages surrogates associated with a treatment choice) were assigned positive values; Barriers scores (representing disadvantages surrogates associated with a treatment choice) were assigned negative values, on a 7-point Likert scale, where 7 ¼ very strongly agree and 1 ¼ very strongly disagree. The goal of the final ATCI composite score was to capture surrogate perceptions about whether their treatment decision would improve (produce greater Benefits) or worsen (produce greater Barriers) overall patient and surrogate situations. An overall higher score on the ATCI would indicate surrogates perceived greater advan- tages associated with a treatment decision; a lower score would indicate greater disadvantages associated with a treatment decision.

Final scale analysis mirrored the two originally intended subscales, with a slight modification. The Benefits subscale (M ¼ 6.32; SD ¼ .76) was inter- preted as decision support, as final items captured concerns about surrogate perceptions of the various sources of support for a treatment decision. The Barriers subscale (M ¼ 1.95; SD ¼ .98) was interpreted as an inability to act in the patient’s best interest, again reflecting content of items retained in the final analysis. The final Benefits-Barriers (decision support-inability to act) composite variable (n ¼ 117; M ¼ .88; SD ¼ .68; Range: �1.20 to þ2.40) was interpreted as higher scores reflecting surrogates’ perceptions of Benefits (i.e., greater support for surrogate decisions). Lower scores reflected greater perceptions of Barriers. As an example, the highest ATCI composite score,

38 J. W. Buckey and N. Abell

þ6, indicated that these surrogates (n ¼ 18) perceived greater benefits associated with their treatment decisions, while the lowest ATCI composite score, �1, indicated that this surrogate (n ¼ 1) perceived greater barriers associated with her treatment decision.

The subscale reliability for Benefits (decision support; a ¼ .726) was clearly acceptable for individual use, yet the subscale reliability for Barriers (inability to act; a ¼ .690) was marginal. Given the pilot nature of this study and the absence of stronger indicators of the Barriers-Benefits composite construct, we judged the ATCI as acceptable for use in these analyses. The SEM for the Benefits subscale (.39) slightly exceeded the threshold value of .30 (5% of 6.0 ¼ .30), and the Barriers subscale SEM (.54) exceeded the target value.

Benefits and Barriers

Simple regression was used to test hypotheses between predictor variables and the Benefits-Barriers criterion (see Table 2). After deleting three cases with excessive influence to improve model performance (Tate, 1998), patient communication accounted for 25% of the variance in Benefits-Barriers [R2 ¼ .254, p < .001, F(1,110) ¼ 37.48]. The more conservative adjusted R2, appropriate for comparisons containing different numbers of variables, sample sizes, or both, was essentially equivalent ðR2adj ¼ :247Þ. These find- ings were consistent with other studies addressing surrogates’ need for pati- ent information, suggesting that surrogates may seek greater quantity and quality of information than patients are willing to provide (Abbot et al., 2001; Covinsky et al., 2000; Hines et al., 2001; Jacob, 1998, Swigart et al., 1996).

Self-efficacy accounted for approximately 22% of variance in perceived Benefits-Barriers [R

2 ¼ .223, p < .001, F(1,115) ¼ 32.751; R2adj ¼ :216]. This find- This finding mirrored the conceptual literature on medical decision making but has not been previously reported in surrogate studies. Self-efficacy repeatedly accounts for approximately 20% of model variability when more broadly examining individual health behaviors (Abraham & Sheeran, 2000).

Nearly 5% of surrogates’ self-efficacy could be accounted for by patient communication [R

2 ¼ .048, p < .013, F(1,126) ¼ 6.407 R2adj ¼ :041].

TABLE 2 Bivariate Relationships: Predicting the Impact of Patient Communication and Surrogate Self-Efficacy on Perceived Benefits-Barriers of Life-Sustaining Treatment Decisions

Variable B SE B B

Patient communication (n ¼ 112) .379�� .062 .504 Self-efficacy (n ¼ 116) .790�� .138 .472 Self-efficacy and patient communication (n ¼ 128) .100� .039 .220

Note. R 2 ¼ .254 and ðR2adj ¼ :247Þ for Patient communication; R

2 adj ¼ .223 and R2adj ¼ :216 for Self-efficacy;

and R2 ¼ .048 and R2adj ¼ :041 for Patient communication and Self-efficacy. �p < .05. ��p < .001.

Life-Sustaining Treatment Decisions 39

Thus, surrogates aware of patient wishes in advance may have greater confidence in performing their surrogate duties, knowing that patients’ wishes were followed. When viewed in conjunction with the statistically sig- nificant but relatively small correlation between these variables (r ¼ .220), these findings suggest measurement of separate, but related, constructs (Hair et al., 2006).

Using hierarchical multiple regression (HMR) as the procedure of choice for testing conceptual models (Cohen & Cohen, 1983), the less con- ceptually important variables—in this case, demographic variables of age, education, gender, ethnicity, and religious-spiritual beliefs—enter the model first (Group 1), followed by physician and patient communications (Group 2), and finally, by social support and self-efficacy (Group 3; see Figure 2). Cases assessed in the HMR analysis (n ¼ 105) produced a ratio of cases to predictor variables (9) of approximately 12:1, greater than the minimum 5:1 ratio, but slightly less than the ideal, approximately 15 or 20:1 (Hair et al., 2006).

As shown in Table 3, regression of the demographic variables in Step 1 did not significantly predict Benefits-Barriers (R

2 ¼ .031). Introduction of physician and patient communication variables in Step 2 was significant (R2 ¼ .153, F ¼ 2.50, F[.05; 7, 97] ¼ 6.36, p ¼ .001) reinforcing the overall strength of the relationship between patient communication and Benefits- Barriers reported in the previous simple regression model. The resulting change in R2 (DR2 ¼ .122) indicated that communication with the patient and physician accounted for approximately 12% of variability in perceived Benefits-Barriers. Lastly, after controlling for the influence of variables previously introduced, social support and self-efficacy were entered into

FIGURE 2 Applied Hypothesized Model of Surrogate Decision Making.

40 J. W. Buckey and N. Abell

the model in Step 3. This resulted in a statistically significant R2 ¼ .252 (F ¼ 3.56, F[.05; 9, 95] ¼ .718, p ¼ .001). The DR2 ¼ .099 (9.9%) in the final step indicates that social support and self-efficacy increased the explanatory power of the model by an additional 10%.

Post hoc diagnostics improved the fit of the regression model and resulted in statistically significant coefficients showing no multicollinearity concerns (Hair et al., 2006). Final model results, improved by deletion of one influential case (N ¼ 104), are shown in Table 3. The final model chan- ged as follows: R

2 (from .252 to .381), R2adj (from .181 to .322), and DR 2 (from

.099 to .140). Of particular note were changes in the final model in Step 2, in which patient communication and self-efficacy explained 12.5% and 12.4% of surrogate decision making, respectively. Thus, surrogate communication with patients, and surrogate knowledge of and functionality in their roles, accounted for approximately one quarter of the variability in surrogates’ per- ceptions of Benefits and Barriers associated with their treatment decisions. Our preliminary power analysis supported testing simple and multiple regression R

2 coefficients of medium effect size (ES ¼ .15) considering all variables entered into the hierarchical model. But as indicated by the final coefficient of determination, R2 ¼ .381 ðR2adj ¼ :322Þ, the overall study effect size is relatively large, given typical effect sizes in the social sciences litera- ture that range from .25 to .50 (Cohen & Cohen, 1983).

TABLE 3 Summary of Hierarchical Regression Initial (n ¼ 105) and Final (n ¼ 104) Surrogate Decision Models: Predicting Perceived Benefits-Barriers

Variable B SE B b

Initial model Step 1 Age �.004 .011 �.041 Education �.047 .046 �.104 Gender �.238 .332 �.073 Ethnicity .501 .356 .148 Religiosity-spirituality �.008 .066 �.012

Step 2 Physician communication .025 .061 .038 Patient communication .305� .082 .350

Step 3 Social support �.031 .111 �.027 Self-efficacy .628� .185 .336

Final model Patient communication .291�� .067 .362 Self-efficacy .673�� .155 .391

Note. Step 1: R2 ¼ .031; R2adj ¼ �:018. DR 2 ¼ .031.

Step 2: R 2 ¼ .153; R2adj ¼ :092. DR

2 ¼ .122. Step 3: R2 ¼ .252; R2adj ¼ :181. DR

2 ¼ .099. Final regression model: R2 ¼ .381. R2adj ¼ :322. DR

2 ¼ .140. �p < .05. ��p < .001.

Life-Sustaining Treatment Decisions 41

DISCUSSION

Study findings underscore the importance of early communications between patients and their surrogates in enhancing surrogates’ understanding of patient treatment preferences (Gutheil & Heyman, 2005; Karel, Powell, & Cantor, 2004). Consistent with previous research, surrogates in this study were aware of patient preferences, and they attempted to comply with them when making ‘‘tough’’ decisions (Buckey, 2007, p. 157; Jacob, 1998). Accord- ingly, this advanced understanding of patient wishes enabled them to carry out excruciating, emotionally painful life-and-death decisions reflecting their knowledge of ‘‘what Mom wanted’’ (Buckey, 2007, p. 140). Following patient treatment directives not only provide surrogates a degree of personal gratifi- cation and emotional relief, but may bring closure to a highly emotional, stressful life experience, a closure enhanced by acceptance of the decision made, regardless of patient outcomes. Reflecting on their personal accept- ance of decisions made under real time conditions, the majority of surrogates in this study (98%) felt they made the best decision possible, a decision that they and others can ‘‘live with’’ (Buckey, 2007, p. 158).

Conversely, when knowledge of patient preferences for treatment is unknown, unclear, or inadequate, surrogates had to rely on additional fac- tors, including theirs and others’ values and preferences for patient treatment (Lewis et al, 2006; Limerick, 2007; Vig et al., 2006). The impact of this lack of knowledge was articulated by one respondent, age 54, acting on behalf of her husband, age 57, who openly admitted, ‘‘We’d never talked, so I didn’t know what to do.’’ (Buckey, 2007, p. 141).

Concerns associated with fulfilling the surrogate role included its pragmatic-introspective nature, which we termed the head-heart dilemma, one characterized by internal logical versus emotional struggles. Head- related activities, captured by HBM value expectancy theory, included logical thought processing and its follow-through, acting on patient wishes in order to attain the desired patient treatment decision (Limerick, 2007). One respon- dent separated her actual treatment decision from its personal repercussions, identifying the most difficult part of her surrogacy as ‘‘watching them go, even though I knew it was what she wanted and she was ready’’ (Buckey, 2007, p. 134). One respondent (who happened to be a health care professional) indicated she did ‘‘a little magical thinking’’ in deciding her mother’s treatment, acknowledging, ‘‘Being a health care professional doesn’t make these decisions any easier’’ (Buckey, 2007, p. 134).

As the limited availability of end-of-life quantitative measures unique to surrogate needs have inhibited progress in end-of-life investigations (George, 2002; National Institutes of Health [NIH], 2004), the attempt in this study to capture two variables—self-efficacy and benefits and barriers associated with treatment decisions—offer end-of-life researchers a possible

42 J. W. Buckey and N. Abell

beginning for reliable, valid measurement of surrogate decision activity. As a new variable in the surrogate decision literature, surrogate self-efficacy may expand surrogate investigations through its representation of inherent com- plexities in the surrogate role. The HSPS (Buckey & Abell, 2004), developed early in this research agenda, provides a reliable, valid measurement tool specifically to assess self-efficacy in the context of surrogate decisions. Preliminary scale items associated with treatment decisions developed for the ATCI (Buckey & Abell, 2009), such as rallying family support, obtaining information from others, and inward reflection on considerations leading to a final treatment decision, mirrored conceptual themes in other studies (Limerick, 2007; Swigart et al., 1996). Lastly, we have piloted evaluation of benefits and barriers perceived by surrogates. This promotes greater reliance on conceptual literature regarding surrogates’ evaluative process, previously unaddressed in quantitative investigations.

Even with attention to selection of a rigorous measure such as the MSPSS subscales identifying sources of support: family, friends, and signifi- cant others (Zimet et al., 1988), the social support variable did not emerge as statistically significant, accounting for less than .8% (sr

2 ¼ .008) of the DR2 ¼ .099 reported in the HMR, Step 3. Prior studies identified family and others as sources of guidance during decision making (Abbott et al., 2001; Cogen et al., 1992; Hiltunen et al., 1999; Jacob, 1998; Smedira et al., 1990; Swigart et al., 1996). Considering the strong showing of ATCI items identify- ing surrogates’ influence from and connection with others in treatment decisions, this finding was surprising. Although a positive relationship did emerge between social support and benefits (r ¼ .245, p < .001), it may be that surrogates view support or guidance obtained during decision-making deliberations as a benefit; and beyond family and friends, investigators have not acknowledged ICU health care professionals (i.e., physicians, nurses, social workers, pastoral care) as sources of such support (D. Heyland, per- sonal communication, November 3, 2006). Future research should attempt to define and clarify the potential role of social support by conceptualizing help received from formal care providers.

Limitations

Access to only three tertiary care facilities—one, a major University teaching hospital, along with the Hospice and service agencies (Alzheimer’s and Parkinson’s Association)—was a limitation (small sample size). Two hospitals declined participation primarily due to concerns about researcher contact with surrogate representatives at such an emotionally volatile time, a concern echoed one facility ethics committee. Administrators felt that engaging with researchers at such a critical time in the patient’s and surrogate’s experiences would adversely affect family responses on patient satisfaction surveys; thus

Life-Sustaining Treatment Decisions 43

they declined participation. It should be noted that researcher restriction by facility administrators was identified as a barrier to conducting end-of-life research in the first National Institutes of Health End-of-Life Consensus Report (NIH, 2004). Social work researchers pursuing surrogate decision research should be aware of, and prepared to deal with, potential facility resistance.

However, our experiences in this in vivo study, conducted while surro- gates exercised the duties of their role, indicated that such concerns may be unwarranted. It was only after speaking with one emotionally distraught surrogate who declined participation that we discovered that a research associate had, in violation of study protocol, inappropriately approached the surrogate prior to his treatment decision, rather than afterward. In contrast, one mother-daughter couple expressed appreciation to us for com- piling a survey, thus allowing them to reflect and affirm their decision to withdraw their loved one from the ventilator per his wishes.

Another limitation of this study was the occurrence of Terri Schiavo’s death (March 2005), midway through data collection. This unforeseen event was accompanied by excessive media coverage of which participants were almost certainly aware, and posed a major threat to the internal validity of our study (Pedhauzur & Schmelkin, 1991). A visible spike in survey returns nearly tripled submissions in the 2 weeks prior through the 2 weeks following Ms. Schiavo’s death, as her life-and-death struggle precipitated international attention and heightened awareness of moral and ethical dilem- mas associated with surrogate decisions.

This study was conducted in Florida. Consequently, all methodological and applied issues drawn from the data are circumscribed by Florida statute (Health Care Advance Directives Act, 2009). Generalizations can only cautiously be made to other settings, though it is our belief that the broader issues regarding surrogate decisions; specifically, the evaluation of the presenting benefits and barriers that are the emphasis of this study, are more readily generalizable than particulars of law or title.

Despite these limitations, this study produced meaningful results. In addition to the new measures noted above, the findings also provide support for more intensive study of the underlying surrogate decision-making pro- cess, specifically investigating factors relevant to making, and making peace with, surrogate decisions. Our findings also support expanding measurement of the types and nature of patient-surrogate communication, highlighted as a need in the NIH (2004) report.

Implications for Social Work Practice

With previous research findings indicating a lack of surrogate awareness of and preparation for assuming their roles (Bern-Klug, Gessert, & Forbes, 2001; Moorman & Carr, 2008), the finding that increased surrogate

44 J. W. Buckey and N. Abell

self-efficacy enhanced perceived benefits associated with surrogate decisions can serve to guide social workers as they help surrogates prepare to carry out their dual responsibilities. These include communicating with patients to learn and understand patients’ treatment preferences and implementing those preferences when patients become unable to do so.

As surrogates in this study affirmed, with little or no advance knowledge of patient preferences, surrogates cannot represent patients accurately. Therefore, social workers might begin by assessing surrogates’ knowledge of patient treatment preferences early, at minimum, upon patient entry into the critical or intensive care unit, especially if patients are in imminent danger of losing consciousness=capacity for decision making. Initially, as in any crisis situation, social workers can assess and corroborate surrogates’ knowl- edge about patient care preferences during initial patient admission, or when patient transfer to ICU=CCU is anticipated or occurs.

Secondly, social workers can increase or enhance patient-surrogate communications, if indicated. Interventions should target enhancing or clarifying previous and current communications, especially for surrogates entering the ICU=CCU situation and who have little knowledge of patient care preferences and for patients who remain communicative and cognizant of their surroundings. Doing so may establish a more secure base of infor- mation for surrogates’ future use. As patients’ treatment needs intensify over time, having clarified surrogate understanding of patient preferences in the beginning may prove useful.

Social workers can also develop strategies to address surrogate com- munication difficulties. If the patient is noncommunicative, the surrogate’s or family’s recounting of the most recently expressed patient treatment prefer- ences may also aid in making future treatment decisions, especially should the most recent preferences change in the case of a marked physical decline, such as an increase in patient pain or discomfort. Social workers listening to surro- gates’ narratives about patients’ life histories and lifestyles may offer yet another method to prepare surrogates to examine an eventual, but currently unstated question, ‘‘What would Mom want?’’ in a supportive, reassuring way.

Finally, social workers can enhance surrogate knowledge of patient preferences through tapping into the knowledge of others. For surrogates with extended family support systems, social workers can encourage surro- gates to marshal their resources, notifying significant others in the patient’s life who are available and willing to support the surrogate during patient care and in making difficult future decisions.

As an estimated 90% of patients die after decisions to withhold or with- draw treatment (Truog et al., 2001), the surrogate role is often complicated by a debilitating level of emotional and physical stress. Surrogates in this study reported being overwhelmed with the multiple tasks expected of them. Social workers can serve as liaisons between health care providers and surrogates—providing concrete, updated patient information; supporting

Life-Sustaining Treatment Decisions 45

surrogate emotional, physical, and psychological well-being through direct services; and advocating for provision of appropriate collateral services, specific interventions, or referrals for follow-up care.

To address these potential voids in surrogate self-efficacy during face-to-face interviews, social workers can topically explore surrogates’ understanding of their duties. In our highly mobile society, surrogate repre- sentatives may arrive at the patient’s bedside from across the country or around the world; thus, an assessment and orientation of the surrogate’s understanding of what decisions may be expected and when they would most likely occur can inform both surrogate and social worker of any knowledge or information gaps to be addressed.

Filling these gaps might include sharing printed materials or online resources on topics including surrogate duties, state statutory requirements, and facility rules, regulations, or procedures that will influence surrogates’ abilities to carry out their duties. Other examples to increase knowledge of surrogate role functioning may include provision for independent education, such as online resources, available through individual or facility resources, or overnight accommodations near the patient, or ICU=CCU quiet rooms designed to support family members.

As relationships are established, social workers may explore prepara- tory, ‘‘What if?’’ scenarios in surrogate discussions, or offer to help rebuild connections with distant family members. As members of interdisciplinary health care teams, social workers excel at facilitating communications (Black, 2005). Helping surrogates develop relationships with medical or health care facility staff, supporting contact among them, and providing surrogates clear, concrete information needed for making treatment decisions can be crucial in the environments of most fast-paced medical treatment facilities.

Responding to surrogate needs can extend traditional social work services into a realm that benefits surrogate and patient partnerships through critically important communication and education. By assessing and expand- ing what surrogates know about patient preferences early in the patient’s critical care regimen, social workers can assist surrogates to clarify and articu- late their understanding of patient care preferences prior to the time a surro- gate decision is needed. During this process, social work efforts can focus on preparing surrogates for subsequent complex decisions, thereby averting future unnecessary or unwanted medical care and its associated costs. Thus, social work interventions can simultaneously clarify and educate surrogates on the expectations associated with the surrogate role, while continuing to assist and support surrogates during decision making in the critical care environment.

This study reinforces that advanced knowledge of patients’ explicit treatment preferences guides surrogate treatment decisions, even in the face of personal surrogate distress and hardship. The degree of importance surro- gates attached to honoring patient wishes was reflected in the significance of

46 J. W. Buckey and N. Abell

the patient communication and self-efficacy variables. The emerging benefits-barriers variable pointed to a greater reliance on theory to capture surrogate evaluations of treatment options. Focusing future investigations on how surrogates can be better prepared to know patient preferences in advance and to execute their roles within supportive, responsive medical systems may improve understanding of surrogates’ decision-making experiences.

Potential respondents’ emotional distress is a paramount concern in end-of-life research. This study showed that approaching surrogates with sensitivity, respect, and support during an emotionally volatile period is a viable research approach, particularly when potential respondents under- stand that the research may benefit others facing similar challenges in the future. Reported benefits associated with their treatment decisions suggested they derived a degree of personal satisfaction in bringing closure to a highly emotional, stressful life experience.

REFERENCES

Abbott, K. H., Sago, J. G., Breen, C. M., Abernethy, A. P., & Tulsky, J. A. (2001). Families looking back: One year after discussion of withdrawal or withholding of life-support. Critical Care Medicine, 29, 197–201.

Abraham, C., & Sheeran, P. (2000). Understanding and changing health behaviour: From health beliefs to self-regulation. In C. Abraham, & P. Sheeran (Eds.), Understanding and changing health behaviour: From health beliefs to self regulation, (pp. 3–24). Amsterdam: Harwood Academic Publishers.

Addington-Hall, J., & McPherson, C. (2001). After-death interviews with surrogates= bereaved family members: Some issues of validity. Journal of Pain and Symptom Management, 22, 784–790.

Bern-Klug, M., Gessert, C., & Forbes, S. (2001). The need to revise assumptions about the end-of-life: Implication for social work practice. Health & Social Work, 26, 38–48.

Black, K. (2005). Advance directive communication practices: Social workers’ contri- butions to the interdisciplinary health care team. Social Work in Health Care, 40(3), 39–55.

Buckey, J. W. (2007). Factors affecting life-sustaining treatment decisions by health care surrogates and proxies. Dissertation Abstracts International, Section A. Humanities and Social Sciences, 68(05), 2169.

Buckey, J. W., & Abell, N. (2004). Validation of the Health Care Surrogate Preferences Scale. Social Work, 49, 432–439.

Buckey, J. W., & Abell, N. (2009). Factors affecting life-sustaining treatment decisions by health care surrogates and proxies: Assessing benefits and barriers. Social Work in Health Care, 48(4), 386–401.

Carr, D., & Moorman, S. M. (2008). Spouses’ effectiveness as end-of-life health care surrogates: Accuracy, uncertainty, and errors of overtreatment or undertreat- ment. The Gerontologist, 48(6), 811–819.

Life-Sustaining Treatment Decisions 47

Centers for Disease Control and Prevention. (2009). Deaths and mortality. Retrieved January 11, 2010, from http://www.cdc.gov/nchs/FASTATS/deaths.htm

Cogen, R., Patterson, B., Chavin, S., Cogen, J., Landsberg, L., & Posner, J. (1992). Surrogate decision-maker preferences for medical care of severely demented nursing home patients. Archives of Internal Medicine, 152, 1885–1888.

Cohen, J., & Cohen, P. (1983). Applied multiple regression=correlation analysis for the behavioral sciences. Hillsdale, NJ: L. Erlbaum Associates.

Court, A. (Producer). (2009, November 22). Sixty Minutes: The co$t of dying [Television broadcast]. New York: CBS News. Retrieved December 9, 2009, from http://www.cbsnews.com/video/watch/?id=5737437n&tag=cbsnewsMain ColumnArea.2

Covinsky, K. E., Fuller, J. D., Yaffe, K., Johnston, C. B., Hamel, M. B., Lynn, J., et al. (2000). Communication and decision making in seriously ill patients: Findings of the SUPPORT project. Journal of the American Geriatrics Society, 48, S187–S193.

Cruzan, v. Director, Missouri Department of Health, 497 U.S. 261 (1990). Ditto, P. H. (2006). What would Terri want? On the psychological challenges of

surrogate decision making. Death Studies, 30, 135–148. Engelberg, R. A., Patrick, D. L., & Curtis, J. R. (2005). Correspondence between

patients’ preferences and surrogates’ understandings for dying and death. Journal of Pain and Symptom Management, 30(6), 498–509.

George, L. K. (2002). Research design in end-of-life research: State of science. The Gerontologist, 42(Suppl. 3), 86–98.

Gessert, C. E., Curry, N. M., & Robinson, A. (2001). Ethnicity and end-of-life care: The use of feeding tubes. Ethnicity & Disease, 11, 97–106.

Gutheil, I. A., & Heyman, J. C. (2005). Communication between older people and their health care agents: Results of an intervention. Health & Social Work, 30(3), 107–116.

Hair, J. F., Jr., Black, B., Babin, B., Anderson, R. E., & Tatham, R. L. (2006). Multivari- ate data analysis (6th ed.). Upper Saddle River, NJ: Prentice Hall.

Hare, J., Pratt, C., & Nelson, C. (1992). Agreement between patients and their self-selected surrogates on difficult medical decisions. Archives of Internal Medicine, 152, 1049–1054.

Health Care Advance Directives Act, XLIV FL Stat xx 765.101, 205–401 (2009). Retrieved May 2, 2010, from http://www.flsenate.gov/Statutes/index.cfm?App_ mode=Display_Statute&URL=Ch0765/ch0765.htm

Health Insurance and Portability and Accountability Act of 1996 (HIPAA), 45 CFR x 160.103-164. (2002).

Henry, G. (1990). Practical sampling. Newburg Park, CA: Sage. Hiltunen, E. F., Medich, C., Chase, S., Peterson, L., & Forrow, L. (1999). Family

decision making for end-of-life treatment: The SUPPORT nurse narratives. The Journal of Clinical Ethics, 10, 126–134.

Hines, S. C., Glover, J. J., Babrow, A. S., Holley, J. L., Badzek, L. A., & Moss, A. H. (2001). Improving advance care planning by accommodating family prefer- ences. Journal of Palliative Medicine, 4, 481–489.

Jacob, D. A. (1998). Family members’ experiences with decision making for incom- petent patients in the ICU: A qualitative study. American Journal of Critical Care, 7, 30–36.

48 J. W. Buckey and N. Abell

Johnson, N., Cook, D., Giacomini, M., & Willms, D. (2000). Towards a ‘‘good’’ death: End-of-life narratives constructed in an intensive care unit. Culture, Medicine and Psychiatry, 24(3), 275–295.

Karel, M. J., Powell, J., & Cantor, M. D. (2004). Using a Value Discussion Guide to facilitate communication in advance care planning. Patient Education and Counseling, 55, 22–31.

Lautrette, A., Darmon, M., Megarbane, B., Joly, L. M., Chevret, S., Adrie, C., et al. (2007). A communication strategy and brochure for relatives of patients dying in the ICU. New England Journal of Medicine, 356, 469–478.

Layde, P. M., Beam, C. A., Broste, S. K., Connors, A. F., Jr., Desbiens, N., Lynn, J., et al. (1995). Surrogates’ predictions of seriously ill patients’ resuscitation preferences. Archives of Internal Medicine, 4, 518–523.

Lepore, J. (2009, November 30). The politics of death. The New Yorker, p. 60. Retrieved May 2, 2010, from http://www.newyorker.com/reporting/2009/11/ 30/091130fa_fact_lepore

Lewis, C. L., Hanson, L. C., Golin, C., Garrett, J. M., Cox, C. E., Jackman, A., et al. (2006). Surrogates’ perceptions about feeding tube placement decisions. Patient Education and Counseling, 61, 246–252.

Limerick, M. H. (2007). The process used by surrogate decision makers to withhold and withdraw life-sustaining measures in an intensive care environment. Oncology Nursing Forum, 34(2), 331–340.

Mcgee, D. C. (2000, November). Withdrawing life support from the critically ill. Chest, 118, 1238–1239.

McNabney, M. K., Beers, M. H., & Siebens, H. (1994). Surrogate decision-makers’ satisfaction with the placement of feeding tubes in elderly patients. The Journal of the American Geriatrics Society, 42, 161–168.

Moorman, S. M., & Carr, D. (2008). Spouses’ effectiveness as end-of-life health care surrogates: Accuracy, uncertainty, and errors of overtreatment or undertreat- ment. The Gerontologist, 48(6), 811–819.

National Institutes of Health. (2004). National Institutes of Health state-of-the-science conference statement on improving end-of-life care. Retrieved May 3, 2010, from http://consensus.nih.gov/2004/2004EndOfLifeCareSOS024html.htm

Patient Self-Determination Act, Pub. L. No. 101-508, xx 4206 & 4751, 104 Stat. 1388 (1990). Pedhazur, E. J., & Schmelkin, L. P. (1991). Measurement, design, and analysis: An

integrated approach. Hillsdale, NJ: Lawrence Erlbaum Associates. Prendergast, T. J., & Luce, J. M. (1997). Increasing incidence of withholding and

withdrawal of life support from the critically ill. American Journal of Respiratory Care Medicine, 155, 15–20.

Quinlan v. New Jersey, 70 N.J. 10, 355 A.2d 647 (1976). Rosenstock, I. M., Strecher, V. J., & Becker, M. H. (1988). Social learning theory and

the Health Belief Model. Health Education Quarterly, 15, 175–183. Seckler, A. B., Meier, D. E., Mulvihill, M., & Cammer-Paris, B. E. (1991). Substituted

judgment: How accurate are proxy predictions? Annals of Internal Medicine, 115, 92–98.

Smedira, N. G., Evans, B. H., Grais, L. S., Cohen, N. H., Lo, B., Cooke, M., et al. (1990). Withholding and withdrawal of life support from the critically ill. The New England Journal of Medicine, 322, 309–315.

Life-Sustaining Treatment Decisions 49

Springer, D. W., Abell, N., & Hudson, W. W. (2002). Creating and validating rapid assessment instruments for practice and research: Part 1. Research on Social Work Practice, 12(3), 408–439.

Suhl, J., Simons, P., Reedy, T., & Garrick, T. (1994). Myth of substituted judgment. Archives of Internal Medicine, 154, 90–96.

Swigart, V., Lidz, C., Butterworth, V., & Arnold, R. (1996). Letting go: Family willingness to forgo life support. Heart and Lung, 25(6), 483–494.

Tate, R. (1998). An introduction to modeling outcomes in the behavioral and social sciences, (2nd ed.). Edina, MN: Burgess Publishing.

Thelen, M. (2005). End-of-life decision making in intensive care. Critical Care Nurse, 25, 28–37.

Truog, R. D., Cist, F. M., Bracket, S. E., Burns, J. P., Curley, M. A. Q., Danis, M., et al. (2001). Recommendations for end-of-life care in the intensive care unit: The Ethics Committee of the Society of Critical Care Medicine. Critical Care Medicine, 20(12), 2332–2348.

Vig, E. K., Taylor, J. S., Starks, H., Hopley, E. K., & Fryer-Edwards, K. (2006). Beyond substituted judgment: How surrogates navigate end-of-life decision-making. Journal of The American Geriatrics Society, 54, 1688–1693.

Wright, A. A., Zhang, B., Ray, A., Mack, J. W., Trice, E., Balboni, T., et al. (2008). Association between end-of-life discussions, patient mental health, medical care near death, and caregiver adjustment. Journal of the American Medical Association, 300(14), 1665–1673.

Zimet, G. D., Dahlem, N. W., Zimet S. G., & Farley, G. K. (1988). The Multidimen- sional Scale of Perceived Social Support. Journal of Personality Assessment, 52, 30–41.

50 J. W. Buckey and N. Abell