PREDICTORS OF SURROGATE INVOLVEMENT IN ADVANCE CARE PLANNING BEHAVIORS

Advance care planning (ACP) is a framework used to escribe the process of end-oflife care planning. ACP includes four behaviors: co mpleting a living will that documents end-of-life wishes, appointing a health c are proxy (HCP), discussing the use of interventions with loved ones, and discussin g quality versus quantity of life with loved ones. Studies have found that the number of individuals engaging in ACP is low. Further, even those who have completed a li ving will and HCP often have insufficient communication with loved ones regardin g their end-of-life wishes, often leading to discordant end-of-life care. The present tudy used constructs from the Transtheoretical Model (TTM) to examine predictors as ociated with surrogates’ ( N= 216) readiness (i.e., Stage of Change) to help thei r loved one complete ACP behaviors. Exploratory analyses including Analyses of Variance (ANOVAs) and correlations examined surrogate age, race, gender, experience wi th medical decision making relationship to the older patient, frequency of con tact, perceived positive consequences of ACP behaviors (Pros), perceived negative consequ ence of ACP behaviors (Cons), and ACP attitudes values and cultural beliefs (AVCB ) as potential predictors of surrogate readiness to assist their loved one in th e four ACP behaviors. A series of multiple regressions showed that across the four AC P behaviors, only Cons was a significant predictor of Stage of Change, suggestin g that as surrogates perceived more negative consequences of ACP, they were less ready to support the older patient in the process. This result is not consistent with other T TM research, in which Pros and Self Efficacy are generally stronger predictors of Stage of Change. This study also examines the level of agreement between older patie nts and their surrogates on readiness to engage in the four ACP behaviors. Diff erence scores between the surrogate Stages of Change and that of their older patient counterpart reflected discrepancies in readiness for ACP behaviors. ANOVA s and correlations were used to examine relationships between the level of agreemen t and the surrogate variables. Stage of Change distributions revealed surrogates w ere more ready to participate in ACP behaviors than the older patients, indicating t hat surrogates may play an important role in older patients’ readiness to comp lete ACP behaviors. Increased surrogate age was associated with less agreement on readi ess to engage in creating a living will. Female gender and higher surrogate rat ings on Pros was associated with more agreement in discussing interventions. Higher surrogate ratings of Cons was significantly associated with less agreement to dis cus interventions, and more agreement to discuss quality versus quantity of lif e. Higher ratings on AVCB was significantly associated with less agreement in rea diness to discuss interventions. Future longitudinal research may provide more infor mation regarding these

surrogate Stages of Change and that of their older patient counterpart reflected discrepancies in readiness for ACP behaviors. ANOVAs and correlations were used to examine relationships between the level of agreement and the surrogate variables.
Stage of Change distributions revealed surrogates were more ready to participate in ACP behaviors than the older patients, indicating that surrogates may play an important role in older patients' readiness to complete ACP behaviors. Increased surrogate age was associated with less agreement on readiness to engage in creating a living will. Female gender and higher surrogate ratings on Pros was associated with more agreement in discussing interventions. Higher surrogate ratings of Cons was significantly associated with less agreement to discuss interventions, and more agreement to discuss quality versus quantity of life. Higher ratings on AVCB was significantly associated with less agreement in readiness to discuss interventions.
Future longitudinal research may provide more information regarding these relationships.
iv Acknowledgements I would like to thank my major professor, Dr. Mark Robbins, for inviting me to join his research lab and mentoring me for the past three years. I would also like to thank my committee members Dr. Andrea Paiva, Dr. Colleen Redding, and Dr. Patricia Burbank. Their feedback and support throughout this process has been invaluable. I would also like to thank Dr. Terri Fried, who allowed me to use her data for this thesis. v

Introduction
Current medical interventions, such as the use of life-sustaining machines, have the potential to prolong life far beyond the point at which an individual would live in the absence of these interventions. The use of these interventions often results in end-of-life care that is costly and exceeds the amount and duration of medical interventions the patient wished to receive (Zhang, 2009). Thus our intervention driven medical culture tends to put patients, their physicians, and their loved ones in a position of extending quantity of life at the expense of quality of life. Therefore, it is important for all involved in end-of-life care to consider the costs and benefits of life sustaining medical technologies. Research suggests that fewer life sustaining interventions could lead to better end-of-life outcomes. For example, Cohen et al. (2011) analyzed 701 retrospective reports from physicians of deceased patients in Brussels, Belgium and found that overall, patients who received palliative care services (i.e. services that aim to allay discomfort at end-of-life rather than attempting to prolong life) had greater feelings of well-being as measured by the Edmonton Symptom Assessment System (ESAS) as well as less shortness of breath, implying that they experienced less anxiety than those who did not receive palliative care. In addition, it was found that patients who received palliative care were more likely to die in the setting they preferred (i.e. mostly within the home) than those who did not receive palliative care (Cohen et al., 2011). These findings suggest that patients who received palliative care were more likely to have their preferences followed at end-oflife care. In order to receive the desired form of end-of-life care, the patient must explicitly make their wishes known to their loved ones and physician. The appropriate steps that are taken to accomplish this include the following: creating a living will (a document that includes end-of-life care wishes), and appointing a healthcare proxy (HCP), a patient advocate who ensures that physicians adhere to patient wishes. This process is described as Advance Care Planning (ACP), in which a patient can prepare for a time that they may be unable to communicate their healthcare preferences regarding the utilization of life-prolonging machines and procedures or palliative care.
A recent study that examined the need for ACP in medical settings found that out of 1083 hospitalized older patients, up to 57.2% required medical decision making by a surrogate and only 7.4% of the sample had a living will and 25% had an HCP document in their medical record (Torke et al., 2014). These findings highlight the current need for ACP completion in older patient populations.
ACP is best not thought of as a static process in which one's preferences remain constant throughout changing circumstances. It is important for all involved in the process to understand the older patient's preferences as they may shift and evolve, which requires adequate and continuous communication. Current research suggests that poor communication between a patient, their physician, and loved ones (i.e. surrogates) regarding end-of-life care typically results in care that is discordant with patient wishes (Swetz, Kuczewski, & Mueller, 2011). Therefore, in addition to completing documentation, the patient continuing to communicate their wishes to physicians and loved ones should improve the likelihood that they will receive the end-of-life care they desire (Hines, Toale, Heisel, & Baringer, 2000;Swetz et al., 2011). In addition to creating a living will and HCP, communication between the patient and surrogate, as well as communication between the patient and physician has been regarded as an important component of ACP (Fried et al., 2010).

Justification and Significance of the Study
The current literature on ACP and quality-of-life (QOL) ratings is mainly concerned with agreement between the patient and their HCP regarding end-of-life preferences (Capelle,Vlak, Algra,& Rinkel, 2010;Essen, 2004;Fried et al., 2011;Hawkins, Ditto, Danks, & Smucker, 2005;Hines et al., 2000;Jette et al., 2012;Parks et al., 2011;Rothman et al., 1991;Seckler, Meier, Mulvihill, & Paris, 1991;Shalowitz, Garrett-Mayer, & Wendler, 2006;Swetz et al., 2011;Winters & Parks, 2008). This research found that while surrogate ratings of patient QOL and physical ability (e.g. ambulatory abilities) are reasonably accurate, surrogate-patient agreement on advance care preferences is poor. It appears that surrogates have better understanding of their loved one's physical health rather than their preferences for treatment at end-of-life care. Further, while ACP improves end-of-life care outcomes for patients and their families, communication regarding this process between patients and their surrogates and physicians is lacking (Patel, Janssen,& Curtis, 2011). Thus, research is needed in order to understand factors associated with the level of surrogate involvement in ACP for a loved one.
Currently, no research specifically examines factors that potentially predict surrogates' readiness to assist a loved one in ACP. The present study attempts to address this question by using Multiple Regression modeling to assess whether demographics (i.e. age, race, gender, and relationship to older patient), frequency of interaction with the older patient, medical experiences, and Pros and Cons can discussions regarding end-of-life treatment, and physicians' personal beliefs on the subject do not significantly predict physician adherence to patient preferences (Torke et al., 2010). Given that physicians will typically act in favor of aggressive medical care (e.g. the use of life sustaining procedures) regardless of whether it adheres to patient wishes, it is important that the patient have a surrogate present to ensure that care is concordant with the patient's wishes. In summary, although surrogate decision making is also imperfect, physician decision making in end-of-life treatment is often more discordant with patient preferences (Shalowitz et al., 2006).
Even in the event that a surrogate is present, decisions made by surrogates are often inaccurate due to insufficient communication between the patient and the surrogate as well as overconfidence in both the patients and surrogates regarding surrogate decision-making accuracy (Hawkins et al., 2005;Hines et al., 2000;Shalowitz et al., 2006;Swetz et al. 2011). Although living will documents are an important component in ACP, designations are often rigid and can be difficult to interpret in ambiguous situations. For example, Mahon (2011) explains that a patient may consider a treatment if it will improve health, but may refuse the same treatment if it will prolong life at the expense of QOL. A surrogate who can accurately report the patient's wishes would be essential in this circumstance. Moreover, research suggests that patients' advance care preferences change significantly overtime, corresponding with changes in health and perspectives regarding quality versus quantity of life (Fried et al., 2007). Evidence suggesting that ACP is a dynamic process, further emphasizes the importance of effective and continuous surrogate-patient communication.
Although research on the effects of improved communication on ACP readiness is limited, the extant literature suggests that when patient-surrogate communication is sufficient, surrogate involvement in the process of ACP can improve understanding of patient end-of-life care preferences and improve adherence to the patient's wishes, thus allaying some of the stress and burden associated with difficult medical decision making that is experienced by all involved. The current literature on ACP supports the argument that older patient-surrogate communication is as important in ACP as completing a living will and HCP (Hines, et al., 2000;Shalowitz et al., 2006;Swetz et al., 2011;Torke, et al., 2010). In the present study, communication regarding the use of life sustaining interventions and communication regarding quality versus quantity of life are included as two behaviors that are relevant to ACP. In summary, communication plays an important role in ACP and the present study attempts to examine this factor by assessing a frequency of contact measure as a predictor of surrogate readiness to assist their loved one in ACP and agreement in Stage of Change of the four ACP behaviors (i.e. creating a living will, creating a HCP, communicating preferences regarding the use of life sustaining treatments, and communication regarding quality versus quantity of life).
Programs aimed at raising awareness of the importance of ACP in patients nearing end-of-life can improve quality of death by ensuring that patients' medical wishes are followed (Fried et al., 2012;Levy, Morris, & Kramer, 2008). However, older patients often approach end-of-life without ensuring that their advance care wishes are known through a living will or advance directive, a HCP, or through adequate communication with surrogates and physicians. Although patients often view ACP in a positive light, rates of completion of living wills and HCP are low (Salmond & Estrella, 2005;Silveira, Kim, & Langa, 2010). More specifically, Silveira et al. (2010) reviewed medical records of 3746 subjects (≥ 60-years-old) and found that individuals who completed a living will were more likely to request limited care and those who completed a HCP were more likely to die outside of the hospital. Further, patients who did not complete a living will or HCP were more likely to receive aggressive care at end-of-life in a hospital. In summary, older patients who do not make their wishes known by completing ACP and communicating their wishes to physicians and loved ones could potentially receive end-of-life care that is more aggressive than they might prefer. This information highlights the importance of understanding factors associated with successful completion of ACP. More importantly, with completion rates of ACP being unacceptably low (Silveira et al., 2010, Torke et al., 2014, surrogate aid in the process of completing a living will and HCP could improve end-of-life care outcomes. Therefore, it is important that the present study examine factors that could serve as predictors of surrogate readiness to assist an older patient in completing a living will and HCP.

Transtheoretical Model
The Transtheoretical model (TTM) is a comprehensive and integrative model of behavior change that has been applied as an organizing framework to understand behavior change and decision-making in dozens of contexts (Prochaska et al., 1994).
The following key constructs are included in the TTM: Stages of Change, Decisional Balance, Self-efficacy and Processes of Change. The Stages of Change construct is based on a series of stages that individuals tend to move through in efforts to achieve behavior change (Prochaska et al., 1994). The Decisional Balance construct was adapted from Janis and Mann's (1977) decision making model. This construct is essentially a decisional balance checklist of comparative gains (Pros) and losses (Cons) (Prochaska et al., 1994). The Self Efficacy construct was derived from Bandura (1977) and this term refers to an individual's confidence in their ability to complete a behavior. The Processes of Change construct is aimed to examine behaviors that represent activities that people use to progress through the Stages of Change; these include experiential and cognitive processes (e.g. consciousness raising and self-reevaluation), which are used in the earlier stages, and behavioral processes (e.g. helping relationships and self-liberation), which are used in later stages (Fried et al. 2010). The TTM has shown consistent results when applied to 48 different health behaviors in 10 different countries (Hall and Rossi, 2008), indicating that the TTM is generalizable as a model for behavior change across many different groups. Moreover, the TTM Stages of Change have been used as a conceptual framework for social workers discussing advance care planning with patients that had advanced illnesses (Rizzo et al., 2010). Fried and colleagues (2012) were the first to develop and validate scales that measure the TTM constructs as applied to ACP behaviors in a sample of older patients. In the present study, two key constructs of the TTM will be analyzed: Decisional Balance and Stage of Change. The Stages of Change range from a stage that indicates that a participant is not considering changing a behavior to a stage that indicates the participant has performed the behavior more than 6 months prior.
Precontemplation defines the stage in which an individual is not planning on changing a behavior in the foreseeable future (i.e. within the next 6 months). Contemplation is the stage that indicates an individual is planning to take steps toward behavior change in the next 6 months. Preparation is characterized by an individual planning to change a behavior in the more immediate future (i.e. within the next 30 days). Action is the stage that indicates an individual has made behavior changes within the preceding 6 months. Maintenance is a stage that indicates an individual is continuing the desired behavior; in the context of the present study, Maintenance describes an individual who completed ACP documentation and communicated their preferences to surrogates prior to 6 months (Fried et al., 2010). The Decisional Balance construct (Pros and Cons) is an intermediate indicator of change, meaning that there is a consistent relationship between the level of Pros and Cons and Stage of Change: as individuals progress from Precontemplation to action, the maximum increase in Pros is typically twice the maximum decrease in Cons (Hall and Rossi, 2008).
Given that TTM constructs have shown consistent results across many different behaviors, including the context of ACP (Fried et al., 2012), this model of behavior change would be an appropriate framework for understanding readiness to complete ACP behaviors. In the present study, the (TTM) Stages of Change were used in the surrogate sample to assess readiness to help an older patient in ACP (Fried et al., 2010). Based on questions regarding thoughts on ACP completion behaviors, participants were placed into their respective Stage of Change (i.e. Precontemplation, Contemplation, Preparation, Action/Maintenance). Surrogates were also asked a series of questions regarding the Pros and Cons of helping a loved one complete ACP.

Predictors of Surrogate Stage of Change
Based on previous findings, several independent variables can be tested as predictors of Stage of Change for each of the four ACP behaviors in the present study.
Surrogate Age and Gender. Evidence suggesting that age and gender play a role in surrogate ACP involvement is scarce. However, Hines and colleagues (2000) found that of the patients who chose a parent as surrogate, mothers were chosen significantly more often (78%) than fathers (22%). Additionally, although the gender differences in other categories (i.e. spouse, child, sibling, and friend) were not significant, the overall frequency of females in the Hines et al. (2000) study was higher than males. The greater frequency of female surrogates might suggest that females in the present study will be more likely to be at a later stage of readiness for ACP than males. In addition, Alano and colleagues (2010) examined predictors associated with completion of advance directives in a sample of older adults (> 65years-old) and found that significantly more females completed advance directives than males. Although this finding is not derived from a surrogate sample, it suggests the likelihood that females will participate in ACP is higher than males. It was also found that increased age was a predictor of completion of an advance directive (Alano et al., 2010). For example, participants that were above age 85 were 6.3 times more likely to complete an advance directive than participants 59-75 years old. These results suggest that increased age in the surrogate sample may be associated with greater likelihood of being at a later Stage of Change to assist a loved one in ACP.
Surrogate Race. It is expected that race will be a significant predictor of Stage of Change in each of the four ACP behaviors. More specifically, Whites will have greater likelihood of being in a later Stage of Change than Nonwhites. Hopp & Duffy (2000) used logistic regression to compare differences in ACP and end-of-life care decision making amongst Whites and Blacks. Results from this study showed that Whites were significantly more likely to discuss ACP with loved ones, complete a living will, and appoint a HCP. In addition, Fried and colleagues (2010) found that, in a sample of old patients (≥ 65-years-old), the majority of Whites were in the Action/Maintenance Stage of Change (54%), while the majority of Nonwhites were in the Precontemplation Stage of Change (59%). These findings suggests that the likelihood of White surrogates discussing ACP and assisting older patients in the planning process will be higher than Nonwhites in this study.

Surrogate-Older Patient Relationship.
Research comparing the accuracy of surrogates with different relationships to the older patient, specifically between children and spouses, found that spouses are more likely to be chosen as a HCP and are significantly more accurate than children in following patient preferences at endof-life (Parks et al., 2011). Although there is currently no research specifically comparing spouse and child readiness to assist in the ACP process, evidence was found that the frequency of spouses serving as surrogates was significantly greater (55%) than adult children (23%) (Hines et al., 2000). These findings suggest that in the present study, it is likely that spouses will be at a later Stage of Change for ACP behaviors than children of older participants.

Surrogate-Older Patient Communication. The literature on ACP suggests
that communication is a key component of surrogate accuracy and comfort with the topic (Fried & O'Leary, 2008;Hines et al., 2000;Lang & Quill, 2004). This research implies that surrogates who have nuanced understanding of the older patient's end-oflife wishes will be more willing to assist that person in ACP. It is expected that surrogates who communicate more often with their loved one will have greater likelihood of being involved with assisting the older patient in each of the four ACP behaviors.
Surrogate Life Experience with Medical Decisions. Medical experiences might serve as predictors of readiness to assist a loved one in ACP. Previous research has identified that experience of major surgery makes old adults' completion of advance directives six times more likely to occur (Alano et al., 2010), suggesting that surrogates with more medical experiences in the present study will have greater likelihood of being at a later Stage of Change in readiness to assist in ACP. Further, in a study assessing factors that help or hamper surrogates' willingness to assist an older loved one in ACP found that previous experience with medical decision making is a factor that helps surrogates in this process (Vig et al., 2007).

Surrogate Pros and Cons. It is expected that endorsements in the Decisional
Balance construct of the TTM will serve as predictors of Stage of Change. More specifically, previous research examining the validity and reliability of the TTM in ACP behavior change on a sample of older adults (≥ 65-years-old), found that an increase in Pros and a decrease in Cons of ACP was associated with increased readiness in ACP participation (Fried et al., 2012). It is expected that results from the present study will be consistent with results from the Fried and colleagues (2012) study in that increased Pros and decreased Cons will be associated with later Stages of Change (i.e. greater readiness to help a loved one in ACP behaviors).

Surrogate Attitudes Values and Cultural Beliefs (AVCB).
Medical misconceptions as well as religious beliefs that regard a higher power as the entity that determines health outcomes have been shown to influence medical decision making, particularly in the context of organ donation and ACP. In a review examining the cultural influences on African American's willingness to become an organ donor, Morgan (2006) highlights the common medical misconceptions and religious beliefs that influence the decision to become an organ donor. The view that physicians will not save the lives of those who declare themselves as organ donors is a common medical belief that discourages organ donation in African Americans. Further, religiosity can influence the decision to donate organs in two ways: religious leaders might not emphasize the importance of certain health behaviors, and religious myths might discourage certain health behaviors (Morgan, 2006). Similar misconceptions have also been found to deter blood donation (Burditt et al., 2009).Given that medical mistrust and misconceptions seem to be a barrier to health care planning, Fried and colleagues (2012) developed the ACP Values and Beliefs scale (AVCB) that will be analyzed in the present study. This scale incudes items that inquire about the misconceptions and attitudes that might deter individuals from completing ACP. Fried and colleagues (2012) found that in a sample of 304 older patients, ACP values and beliefs are significantly associated with Stages of Change, which measure readiness to complete all ACP behaviors. Given these previous findings, it is reasonable to predict that this scale will show similar associations to Stages of Change in the surrogate sample.

Surrogate-Older Patient Discrepancy on Stage of Change
In addition to examining predictors of surrogate Stage of Change in each ACP behavior, the present research aims to examine factors associated with surrogate-older patient discrepancy. This set of analyses will describe the nature of the relationships between variables including surrogate age, gender, race, relationship with the older patient (Relationship), frequency of contact with the loved one (Communication), experiences with medical decision making (Life Experience), Pros, Cons, AVCB, and surrogate-older patient discrepancy on Stage of Change in each of the four ACP behaviors.

Hypotheses
The present study attempts to further understand the factors associated with surrogate involvement in ACP by examining potential predictors of surrogate readiness to assist a loved one in ACP behaviors as well as better understanding the relationships between surrogate-older patient discrepancies on the level of readiness for completion of ACP behaviors and the predictor variables. ANOVAs and Chisquare tests will be used as exploratory analyses to test group differences between the predictor variables and Stage of Change as well as the surrogate-older patient discrepancy on Stage of Change in each of the four ACP behaviors. Multiple Regression analyses will be used to assess whether variables such as age, race, gender, Relationship, Communication, Life Experience, Pros and Cons, and AVCB, can predict surrogate readiness to assist a loved one in ACP behaviors. The following hypotheses will be tested: 1. Predictors of surrogate readiness to assist a loved one (i.e. Stage of Change) in each of the four ACP behaviors will include the following: a spousal relationship to the older patient, interacting with the older patient daily, White race, female gender, increased age, greater experience with medical decision making, increased Pros, and less religious and medical misconceptions related to end-of-life care.
2. Exploratory analyses will be run in order to elucidate the relationship between the surrogate variables, which include age, gender, race, Relationship, Communication, Life Experience, Pros, Cons, and AVCB, and the surrogateolder patient Discrepancy variable.

Participants and Recruitment
Data for the present study were collected from 218 older patients and their surrogate, who was identified by the older patient, from two primary care facilities and one senior center. Older patients in this study were screened by physicians as not having a diagnosis of dementia. Trained interviewers conducted in-person interviews with the older patients, as well as phone interviews with the surrogates. Of note, while a total of 304 older patients participated in this study, roughly two thirds of this sample identified a surrogate who subsequently agreed to participation. In addition, the majority (i.e. approximately two thirds) of the older patients who identified surrogates were in the Action/Maintenance Stage of Change across the four ACP behaviors.

Measures
The following measures were developed by Fried and colleagues (2010).
Although these measures were tailored for administration to the respective samples, they were only validated in the older patient sample; not the surrogate sample (Fried et al. 2010(Fried et al. , 2012. Descriptive Information. Demographic information was collected, which includes age, race, gender, and the relationship with the older patient (i.e. Child, Spouse, Significant other, or Other relationship). In addition, surrogates reported the frequency of contact with the older patient as a proxy measure for the amount of older patient-surrogate communication.
Surrogate Age and Gender. The older patients had a mean age of 75 (SD= ± 7) and the majority of the older patient sample were female (73%). The surrogate sample had a mean age of 60 (SD=± 12.7) and were also majority female (68%).  (19%), and Less than weekly (4%). Given that those surrogates who contacted the older patient less than weekly only made up 4% of the sample, the frequency of contact categories were dichotomized into Daily and Weekly or less. Surrogates who had daily contact with the older patient constituted 76% of the sample (N= 167) and 24% of the sample (N=51) contacted the older patient weekly or less.
The goal for collapsing the above described independent variable categories is to address limitations in sample size in some participant subgroups and to improve the statistical power of the analyses. Table 1 presents the demographic information of the older patient and surrogate samples.

Surrogate Life Experience with Medical Decisions.
A series of six questions was used to assess surrogates' prior experiences with medical experiences and decision making. The surrogates were asked if they have experienced medical problems themselves or if they have witnessed others experience medical problems or had to make difficult medical decisions for others. The Life Experience index includes items such as, "Have you ever faced a life-threatening illness?" and "Have you ever had to make a medical decision for someone who was dying?" Participants were asked to respond 'Yes' or 'No' to each item. See Appendix A for a complete list of the Life Experience items. These items were not formally developed as a scale and as such internal consistency was modest (α= .48). The items were summed and treated as an index and used as a predictor variable for the present analyses.

Surrogate Decisional Balance (Pros and Cons).
The Decisional Balance measure included 6 Pros (i.e., the perceived benefits) and 6 Cons (i.e. the perceived limitations) of assisting a loved one in ACP. These items were rated on a 5 point scale from 1="Strongly Agree" to 5="Strongly Disagree". Items are listed in Appendix B.
Surrogate AVCB. The AVCB measure includes 7 items that assess religious beliefs and medical myths associated with ACP. Surrogates endorsed these items on a 5-point scale assessing the level of agreement with each statement with 1 representing 'Strongly Agree' and 5 'Strongly Disagree.' Chronbach's alpha indicates good internal consistency (α= .86).The AVCB items include "My loved one does not need to do advance care planning because once you reach a certain age, the doctors aren't going to use machines," and "My loved one does not need to do advance care planning because once it becomes clear that you are dying, the doctors aren't going to use machines." See Appendix C for the AVCB items.  Correlations compared relationships with discrepancy and continuous variables, which include age, Life Experience, Pros, Cons, and AVCB and Discrepancy (i.e. levels of agreement) on Stages of Change in each of the four algorithms. In addition, Stage of Change distribution in the surrogate and older patients was evaluated in order to examine patterns in overall readiness to complete each of the four ACP behaviors. Figure 2, across the four ACP behaviors, older patients reported being less ready to complete ACP behaviors than the surrogate sample. For example, comparison of Stage of Change distribution across the two samples in the HCP behavior shows that more older patients were in Precontemplation (31.3%) and Contemplation (7.6%), whereas more surrogates were in Preparation (15.4%) and

As illustrated in
Action/Maintenance (23.7%). In addition to notable differences in the level of readiness in the two groups, the overall distribution of Stage of Change shows that the majority of participants are in Action/Maintenance, with the exception of older patients in the HCP behavior.
Living Will. A significant positive correlation was found between Discrepancy in the Living Will behavior and age, r(215) = .14, p = .040. No significant correlations were found between Pros, Cons, and AVCB and Discrepancy. In addition, no significant differences on Discrepancy were observed amongst, race, gender, Relationship, and Communication in the Living Will behavior.

Surrogate Readiness to Help a Loved One in ACP Behaviors
Medical interventions have the potential to prolong life far beyond the point at which an individual would live in the absence of these interventions. Given that medical interventions can be costly and often exceed the amount of care a patient would wish to receive at end-of-life (Zhang, 2009) In order to further explore this issue, the present study examined predictors associated with surrogate readiness to help a loved one in doing the four ACP behaviors. The overall results did not support the hypothesis that age, race, gender, Life Experience, Relationship, Communication, Pros, and AVCB would be significant predictors of Stage of Change in all of the ACP behaviors. However, exploratory analyses revealed that Pros, Cons, and AVCB were significantly associated with Stage of Change in the Living Will behavior. Higher ratings on Pros was associated with more surrogate readiness to help a loved one complete a living will document, whereas higher ratings on Cons and AVCB was associated with less surrogate readiness to help a loved one complete a Living Will. Additionally, in the ACP behaviors, HCP, Interventions, and Quality versus Quantity of Life, significant differences were found among Stages of Change in Pros and Cons. In each of these behaviors, higher ratings on Pros was associated with more surrogate readiness to help a loved one complete HCP, discuss the use of end of life interventions as well as focus on quality versus quantity of life with their loved one. Conversely, higher ratings on Cons was associated with less readiness to help a loved one complete HCP, and discuss the use of interventions as well as quality versus quantity of life with their loved one. The association between increased Pros and greater surrogate readiness, as well as the association between increased Cons and less surrogate readiness is consistent with the relationships between Pros and Cons, and Stage of Change in the older patient sample (Fried et al., 2012). A significant predictor in each surrogate ACP behavior was Cons, or perceived negative consequences of a given ACP behavior. These findings indicate that lower ratings of Cons is associated with helping a loved one in each of the four ACP behaviors. In other words, surrogates who perceived fewer negative consequences to performing ACP were more ready to help an older patient (i.e. loved one) complete and discuss ACP behaviors. This finding is consistent with previous research that suggesting that ACP non-completers were more concerned with negative beliefs associated with ACP (e.g. "advanced directives are too binding") than ACP completers (Beck, Brown, Boles, & Barret, 2002). It may be helpful for health care providers to focus on addressing the negative aspects or Cons of ACP in their consultations with patients and their loved ones, thus promoting more surrogate involvement in the ACP process. This is consistent with implications of the patient perceived barriers to ACP examined by Schickedanz and colleagues (2009). This finding is not consistent with previous TTM research across many behavioral domains, which suggests that Pros are more strongly associated with readiness for behavior change than are Cons (Hall & Rossi, 2008). Results from Schickedanz and colleagues (2009) suggested six barrier themes that impeded patients from completing ACP behaviors, which include the following: "perceiving ACP as irrelevant (84%); personal barriers (53%); relationship concerns (46%); information needs (36%); health encounter time constraints (29%), and problems with advance directives (29%). Some barriers were endorsed at all steps (e.g., perceiving ACP as irrelevant). Others were endorsed at individual steps (e.g., relationship concerns for family/friend discussions, time constraints for doctor discussions, and problems with advance directives for documentation)" (p. 5). Schickedanz and colleagues (2009) drew conclusions similar to the present study regarding the importance for health care providers to address barriers or perceived negative consequences of completing ACP behaviors. This finding may have significant implications for the development of future intervention programs in medical care settings.

Agreement on Readiness to Complete ACP Behaviors
Research suggests that insufficient communication between the older patient and their surrogate often leads to end-of-life care that is discordant with patient wishes (Shalowitz et al., 2006;Torke, Moloney, Siegler, Abalos, & Alexander, 2010). Given that little is known regarding the factors that may contribute to this discordance, this set of analyses aimed to examine surrogate variables that may be associated with the level of discrepancy among older patient-surrogate Stages of Change in each of the four ACP behaviors. The difference in overall Stage of Change distribution between the surrogates and older patients suggests that surrogates may be more willing to participate in ACP behaviors than older patients. This implies that involving loved ones in the ACP process may encourage older patients to complete ACP behaviors. In comparison to other ACP behaviors, Stage of Change distributions suggest that older patients may experience more difficulty in appointing an HCP, which further implicates the importance of older patient communication with loved ones regarding end-of-life care wishes. Reasons why these patterns of readiness are observed may be elucidated in future research. These findings may also have implications for clinical discussions.
Results indicated that age was significantly associated with Discrepancy in surrogate-older patient Stage of Change in the Living Will behavior. More specifically, older surrogates were more likely to have discordance with their loved one in readiness to complete a living will. Given that the strength of this association was relatively week, this result should be interpreted with caution. No surrogate variables were found to be significantly associated with Discrepancy in readiness to complete the HCP behavior.
In terms of Discrepancy on Stage of Change in the Interventions behavior, results indicated that higher ratings on Pros was associated with more surrogate-older patient agreement, whereas higher ratings of Cons and AVCB was associated with less surrogate-older patient agreement. In addition, results indicate that female surrogates were more likely to be on the same level of readiness as their loved ones to discuss the use of medical interventions at end-of-life care than male surrogates. Pros was significantly associated with Discrepancy between surrogate and older patient Stage of Change in Interventions, which suggests that surrogates who perceive more positive consequences related to discussing medical interventions at end-of-life care were more likely to be at the same level of readiness as their loved one to complete this behavior.

Surrogate ratings on Cons was significantly correlated with Discrepancy on
Stages of Change in the Quality versus Quantity of Life behavior, indicating that surrogates who endorsed more perceived negative consequences of completing the Quality versus Quantity of Life behavior were more likely to be on the same level of readiness as their loved ones. Overall, results suggest that it may be helpful to address perceived positive and negative aspects of ACP, as well as cultural beliefs that may play a role in older persons' readiness to communicate medical wishes with their loved ones. However, given that these results are mixed, they should be interpreted with caution.

Limitations
An important limitation of the present study is that the later Stages of Change are over-represented in the surrogate sample in each of the ACP behaviors, with the majority (i.e. ≥ 58%) of participants in the Action/Maintenance Stage of Change and a small portion (i.e. < 8%) in the Precontemplation Stage of Change. This overrepresentation of later Stages of Change suggests that the present sample may represent a subset of the general population of older persons and their loved ones, thus the present findings may be limited in generalizability. Further, cross-sectional data is limiting with regard to understanding behavior change; therefore, results from the present data can only indicate which factors may be relevant in longitudinal studies.
Although the scales utilized in the surrogate sample were validated in the older patient sample (Fried et al. 2012), confirmatory factor analysis has not been performed on these scales in the surrogate sample. Thus, the present results involving the Decisional Balance and AVCB measures should be interpreted with caution. In addition, the present study only examined predictor variables derived from the surrogate sample. Examining variables from the older patient sample may provide more information regarding predictors of surrogate Stage of Change as well as Discrepancy on Stages of Change.
Another limitation in the present study is that the Discrepancy measure merely reveals discordance between the surrogate and older patient Stage of Change in each ACP behavior. This measure does not describe the nature of the discordance; therefore, conclusions derived from this measure are limited. For example, although a surrogate-older patient pair may have Optimal Agreement in the Precontemplation Stage of Change, this level of agreement holds different implications than a pair in the Action/Maintenance Stage of Change of a given ACP behavior.

Future Directions
As the data in the present study are cross-sectional, longitudinal data would add more information regarding the longevity of a TTM intervention in ACP behaviors and surrogate readiness to assist their loved ones in the process. TTM data on ACP support between patients and surrogates assessed at multiple time points is a future goal for this research.
This study indicates that Cons, or perceived negative aspects of helping a loved one in the ACP process, can influence a surrogate's readiness to participate in the process. Therefore, it may be important for future studies to work toward better understanding effective ways to address these perceived negative consequences or barriers in order to promote surrogate's helping loved ones in the ACP process. Schickedanz and colleagues (2009)  Given that previous experiences in medical decision making has been identified as a factor that may increase a surrogate's willingness to help a loved one in ACP (Vig et al., 2007) it may be beneficial to develop a valid and reliable measure of medical decision-making experience in a surrogate sample. This may provide more information regarding the importance of this factor in surrogate involvement in ACP. The present study only examined independent variables derived from the surrogate sample; therefore, future studies may want to examine independent variables from the older patient sample as well. Future research could examine relationships between older patient independent variables and surrogate Stages of Change. In addition, it may be important to examine independent variables from the older patient sample as predictors of older patient and surrogate Discrepancy on Stage of Change in each behavior. This research may provide more valuable information aimed to guide health care provider discussions with patients and their loved ones regarding ACP.
Given that the Discrepancy measure provides limited information regarding surrogate-older patient discordance on readiness to complete ACP behaviors, it may be beneficial to develop a measure that more accurately describes the nature of the surrogate-older patient discordance and elucidate more reasons why a discrepancy may exist on readiness to complete ACP behaviors.

B. ADVANCE CARE PLANNING -PROS AND CONS
Let us take a moment to review what advanced care planning Consists of: 1) Living Will, 2) Health Care Proxy, 3) Talking to your loved ones and doctor about medical care you do or do not want to receive at the end-of-life.
Here are some possible advantages and disadvantages of planning for your future medical care in the ways that I just asked you about. Please tell me how much you agree or disagree with these statements. Please choose your answer from a scale of 1 to 5, where 1 means strongly agree, 2 means agree, 3 means neither agree nor disagree, 4 means disagree and 5 means strongly disagree. Now, some of these questions may not be clear to you, if so please respond with "I do not understand the question" as your response.

C. ADVANCE CARE PLANNING-ATTITUDES VALUES AND BELIEFS
Here are some additional beliefs about issues related to advance care planning. Please tell me how strongly you agree with the following statements, continuing to use the same answer categories. Now, some of these questions may not be clear to you so please respond with "I do not understand the question" as your response.
How strongly do I agree that: Strongly Agree Agree Neither Agree nor Disagree Disagree Strongly Disagree Doesn't Understand Question

D. STEPS FOR ADVANCED CARE PLANNING
Now we will talk about the specific steps involved in advanced care planning and what steps your loved one has taken. 1. A living will is not the same as a regular will. This is a piece of paper that generally includes a statement saying that if a person's condition is thought to be terminal or if the person is permanently unconscious, then the person should not be kept alive through life support systems.
Has your loved one made out a living will? GO TO Q. 2←Yes -1 GO TO Q. 1b←No -2 Don't know -8 1a.
Are you ready to ask your loved one if he/she has a living will?