AIDS Care, 2015 Vol. 27, No. 3, 370–377, http://dx.doi.org/10.1080/09540121.2014.963019
Advance directives among people living with HIV: room for improvement Joshua A. Barocasa,b*, Kristine M. Erlandsonc, Blythe K. Belzera, Timothy Hessd and James Sosmand a Department of Medicine, University of Wisconsin School of Medicine and Public Health, Madison, WI, USA; bWilliam S. Middleton Memorial Veterans Hospital, Madison, WI, USA; cDivision of Infectious Diseases, Department of Medicine, University of ColoradoAnschutz Medical Campus, Aurora, CO, USA; dDivision of Infectious Diseases, University of Wisconsin School of Medicine and Public Health, Madison, WI, USA
(Received 6 July 2014; accepted 30 August 2014) While HIV has become a largely chronic disease, age-associated comorbidities are prevalent in people living with HIV (PLWH). Therefore, PLWH are appropriate for advance care planning (ACP) and advance directives (ADs) completion. We sought to characterize AD completion among outpatient PLWH. We conducted a retrospective chart review of PLWH who receive their routine care at the University of Wisconsin HIV clinic. Data were extracted from the electronic health record. Variables were entered into a stepwise multivariate logistic regression model to assess which factors were independently associated with AD completion. Five hundred and eighty eight charts were reviewed. Eighty-one percent of subjects were male and 72% were white; mean age was 46.8 years. ADs were completed by 134 subjects and 6.7% of those were completed at the HIV clinic. In the final multivariate model, those who had completed an AD were more likely to be older than age 45; ever been diagnosed with AIDS; have cardiovascular disease, neurologic disorder, chronic kidney disease, or malignancy. In this study, a small percentage of patients had documented ADs, with only a small proportion completed in the HIV clinic. The HIV clinic is an underutilized resource to offer ACP. Interventions are needed to provide the necessary ACP resources for PLWH. Keywords: HIV/AIDS; outpatient clinic; patient care; advance directives; advance care planning
Background By 2015, the majority of the people living with human immunodeficiency virus (PLWH) in the USA will be 50 years or older (CDC, 2007, 2013). The development and availability of antiretroviral therapy (ART) has transformed HIV care in the USA into a largely manageable, chronic disease. With effective ART, PLWH can expect a lifespan that is near the average of those living without HIV (Antiretroviral Therapy Cohort Collaboration, 2008; Collaboration of Observational HIV Epidemiological Research Europe (COHERE) in EuroCoord et al., 2012; May et al., 2014). Despite improvements in overall mortality, age-associated comorbidities such as cardiovascular disease, diabetes, cancer, functional impairment, and neurocognitive impairment are common and may be more prevalent in PLWH than in HIVuninfected persons (Boccara et al., 2013; Pathai, Bajillan, Landay, & High, 2013; Work Group for the HIV and Aging Consensus Project, 2012). Thus, with the increased disease burden associated with aging, decisions about preferences for care provided at the end-of life have become increasingly relevant. Advance care planning (ACP) is an organized process of communication to help individuals understand, reflect upon, and discuss goals for future healthcare decisions in the context of their values and beliefs *Corresponding author. Email: [email protected] © 2014 Taylor & Francis
(Wisconsin Medical Society, 2014). Advance directives (ADs), which allow patients to communicate health-care preferences in the event that he or she is no longer able to make decisions, are one component of ACP (Wisconsin Medical Society, 2014). Recent studies have shown that ACP discussions correlate with better health-care outcomes and satisfaction with care among various chronic disease populations such as chronic obstructive pulmonary disease (COPD) and diabetes. Patients with COPD report that ACP improves their health care and that the process of discussion regarding prognosis, treatment concerns, and preferences for care involved in ACP is valuable (Janssen, Engelberg, Wouters, & Curtis, 2012; MacPherson, Walshe, O’Donnell, & Vyas, 2012). Similarly, insulin-dependent diabetics have expressed a desire to discuss ADs (Toth, Gill, Godkin, & Lee, 1998). Findings by Martin et al. suggest that ACP can help HIV patients consider life goals they want to achieve, appoint a preferred proxy to whom they can express their preferences, strengthen relationships with loved ones, and prepare for death (Martin, Thiel, & Singer, 1999). They are also considered valuable by younger HIV patients when age-appropriate ACP documents are provided (Wiener et al., 2008). In addition, end-of-life discussions have not been shown to unduly distress patients with chronic diseases including PLWH
AIDS Care (Evangelista et al., 2012; Lyon, Jacobs, Briggs, Cheng, & Wang, 2013; Lyon et al., 2010). Early, frequent, and thorough communication about ACP in PLWH may result in improved health outcomes and satisfaction. The overall prevalence of completed ADs in the USA is quite variable. AD completion rates in the general US population range from 26–44% (Matzo, Hijjazi, & Outwater, 2008; McAuley, McCutcheon, & Travis, 2008; Ramsaroop, Reid, & Adelman, 2007; Rao, Anderson, Lin, & Laux, 2014). Up to 72% of US adults aged 60 and older who were recently hospitalized and subsequently died had completed an AD (Silveira, Wiitala, & Piette, 2014). Similarly, 61–70% of older adults in long-term support settings such as nursing homes have an AD (Hirschman, Abbott, Hanlon, Prvu Bettger, & Naylor, 2012; Resnick, Schuur, Heineman, Stone, & Weissman, 2008). Among individuals with specific chronic diseases such as heart failure, diabetes, cystic fibrosis, and cancer, AD completion rates vary from 30–47% (Dunlay, Swetz, Mueller, & Roger, 2012; Evangelista et al., 2012; Hubert et al., 2013; Salmond & David, 2005; Sawicki, Dill, Asher, Sellers, & Robinson, 2008; Schellinger, Sidebottom, & Briggs, 2011; Toth et al., 1998; van Oorschot, Schuler, Simon, & Flentje, 2012). The 2012 Working Group for the HIV and Aging Consensus Project recommends that “older individuals with HIV, especially those with substantial illness burden, should be counseled in completing a durable power of attorney for health care and an advance directive, such as the Physician Order for Life Sustaining Treatment, or similar document” (Work Group for the HIV and Aging Consensus Project, 2012). Studies from the earlier AIDS era found that a majority of PLWH (62–84%) lacked an AD or ACP discussion (Gilligan & Jensen, 1995; Teno, Fleishman, Brock, & Mor, 1990; Wenger et al., 2001). A recent retrospective study of hospitalized PLWH by de Caprariis and colleagues reported only 8% of patients had an AD prior to hospital admission (de Caprariis, Carballo-Diéguez, Thompson, & Lyon, 2012). Erlandson et al. found that 47% of older PLWH on effective ART had completed some component of ACP (Erlandson et al., 2012). However, none of the reported studies assessed AD completion rates in the outpatient setting regardless of age or HIV-disease control. The aim of this study was to examine factors associated with AD completion among PLWH and identify the clinical setting in which they were completed.
371
and Clinics in Madison, WI. Patients were eligible for inclusion if they were (1) at least 18 years of age, (2) not imprisoned or institutionalized in mental health facilities, (3) did not have a legal guardian, and (4) were currently enrolled as patients of the UW HIV clinic as of 31 October 2013. The clinical and administrative database managed by the UW HIV clinic was used to identify patients who met these criteria. Trained study personnel (J.B. and B.B.) conducted a review of the electronic health record (EHR) for information that was not available through the clinical database. The study protocol was approved by the Institutional Review Board at the UW School of Medicine and Public Health. Variables Descriptive variables obtained through chart review and abstraction included: sociodemographic information (age, gender, race, ethnicity, relationship status, insurance status, country of origin), HIV risk factors (injection drug use, men who have sex with men [MSM], maternalfetal transmission, blood transfusion, unprotected heterosexual intercourse), characteristics of HIV-disease status (AIDS-defining illness, CD4 nadir, ever had a positive hepatitis C virus [HCV] serology), non-HIV comorbidities, ACP status (completed AD, setting in which AD completed), and health-care utilization (number of HIV clinic visits in previous five years). Insurance status was divided into private insurance and public or no insurance. Public insurance included Medicare, Medicaid, other state or Ryan White grant funding. Non-HIV comorbidities include prior diagnosis of any of the following: cardiovascular disease (myocardial infarction, coronary artery bypass graft, heart failure, arrhythmia, valvulopathy), cerebrovascular disease (stroke, transient ischemic attack [TIA]), diabetes mellitus (clinical diagnosis, use of hypoglycemic medications), chronic kidney disease (any stage), metabolic bone disease (osteopenia, osteoporosis), Central Nervous System (CNS) neurologic disorder (seizure disorder, cognitive impairment, dementia, encephalopathy not otherwise specified [NOS]), psychiatric illness (depression, anxiety, bipolar disorder, schizophrenia, psychotic illness), malignancy (excluding non-melanoma skin cancer), liver disease (hepatitis, hepatosteatosis, cirrhosis), and lung disease (COPD, asthma, obstructive sleep apnea, pulmonary embolism). Analysis
Methods Study population We conducted a retrospective chart review of HIVseropositive patients who received routine care in the HIV clinic at the University of Wisconsin (UW) Hospital
All sociodemographic and clinical variables were entered into a de-identified database for analysis. Descriptive statistics were used to characterize the study population. Study subjects who reported having a completed AD in the UW EHR were compared to those without an AD. Sociodemographic and clinical variables were
372
J.A. Barocas et al.
dichotomized and analyzed by Wald chi-squared tests or Fisher’s exact test where appropriate, with significance defined as p < 0.05. All variables were considered for entry into a stepwise multivariate logistic regression model to assess which factors were independently associated with AD. Variables were entered into the model at p < 0.10 and retained in the model at p < 0.05. Odds ratio (OR) and 95% confidence intervals (CI) were presented for the final model. Statistical analyses were conducted using STATA Version 11 (Cary, NC).
Results A total of 588 patient records that met inclusion criteria were reviewed for EHR data abstraction. Characteristics of the study subjects are shown in Table 1. The mean age of study subjects was 46.8 years (standard deviation (SD) ± 0.47) and mean CD4 count 634 cells/µL (SD ± 12.5 cells/µL). Most subjects were male and selfidentified as MSM. Nineteen percent of subjects were in a MSM relationship at the time of data extraction. Subjects were predominately white (72%) or AfricanAmerican (22%) whereas Asian/Pacific Islander and American Indian/Alaska Natives were smaller proportions of our study population, 1% and 4%, respectively. Subjects had been living with HIV for an average of 13.3 years (SD ± 4 months). Fifty-seven percent of subjects had at least 10 HIV-related outpatient visits in the previous five years and subjects were followed in the clinic for a mean of 8.7 years (SD ± 3 months). For 308 subjects (52%), their HIV provider also served as their primary care provider. Only one subject had no health coverage at the time of study eligibility, otherwise 41% of subjects held private health insurance. The majority of subjects had at least one non-HIV-related comorbidity with psychiatric illness being the most frequent (40%). Thirty-nine percent (228/588) had ever received a diagnosis of AIDS, 7% had ever had a positive HCV serology, and 5% reported previous or current injection drug use. ADs were completed by 134 subjects (23%). The average time since diagnosis of HIV was longer in those who completed ADs than those who did not (15 years versus 12.5 years). Table 2 shows the results of the stepwise multivariate logistic regression model measuring the association of subject characteristics with AD completion. In the final multivariate model, those who had completed an AD were more likely to be at least 45 years of age (OR: 3.4, 95% CI: 2.0–5.7); ever been diagnosed with AIDS (OR: 1.7, 95% CI: 1.1–2.7); have cardiovascular disease (OR: 2.3, 95% CI: 1.1–4.6), neurologic disorder (OR: 5.0, 95% CI: 2.2–12.1), chronic kidney disease (OR: 3.3, 95% CI: 1.3–8.3), or malignancy (OR: 2.8, 95% CI: 1.3–6.0).
Of the 134 subjects with completed ADs, 6.7% were completed at the UW HIV clinic, while the majority were completed in the inpatient or pre-surgical setting (44%), another non-UWHIV outpatient clinic or community-based organization (CBO; 34%), or with an attorney/notary (13%), as shown in Figure 1. More subjects in our study completed their AD per year in the last three years compared to prior time periods. From 1 January 2010 to 31 October 2013, 47 subjects completed an AD whereas from 1 January 1990 to 31 December 1999, a total of 10 subjects completed an AD and from 1 January 2000 to 31 December 2009, a total of 77 subjects completed an AD. Discussion and conclusions In this retrospective review of PLWH who were receiving care at the UW HIV clinic, we found ADs were completed by less than one-quarter of subjects. In our study, PLWH who had completed ADs were more likely to be older; have a history of AIDS-defining illness; and have comorbidities including cardiovascular disease, neurologic disease, chronic kidney disease, or malignancy. Furthermore, the vast majority of PLWH did not complete their AD at the UW HIV clinic, but instead, were more likely to complete an AD during an inpatient or pre-surgical hospitalization or at another non-UWHIV outpatient clinic or CBO. AD completion percentage in this study was similar to those reported in other studies of PLWH. Our finding that 23% had completed ADs falls in between the two most recently reported completion percentages for PLWH (8–47%), but is lower than those of other chronic diseases (Barakat et al., 2013; de Caprariis et al., 2012; Erlandson et al., 2012; Rao et al., 2014; Salmond & David, 2005; Schellinger et al., 2011). Several previously identified barriers to ACP for PLWH may contribute to low AD completion. Since there are many competing priorities, HIV providers may not offer the opportunity to complete an AD in clinic. Additionally, HIV providers have expressed discomfort with end-oflife discussions particularly in situations in which the patients are younger or there are cultural differences between provider and patient (Curtis & Patrick, 1997). Providers may have concerns about their patients’ perceptions of the ACP discussion and possibly unmasking patient anxiety or depression about end of life issues (Curtis & Patrick, 1997; Gott et al., 2009; Knauft, Nielsen, Engelberg, Patrick, & Curtis, 2005). These barriers are similar to those expressed by providers treating other chronic diseases (Gott et al., 2009; MacPherson et al., 2012). We found certain subsets of PLWH are more likely to complete ADs than others. Older patients were more likely to have completed an AD, similar to studies of the
AIDS Care
373
Table 1. Characteristics of PLWH, by completion of advance directive (AD). Characteristics Total (%) Age
Advance directives among people living with HIV: room for improvement Joshua A. Barocasa,b*, Kristine M. Erlandsonc, Blythe K. Belzera, Timothy Hessd and James Sosmand a Department of Medicine, University of Wisconsin School of Medicine and Public Health, Madison, WI, USA; bWilliam S. Middleton Memorial Veterans Hospital, Madison, WI, USA; cDivision of Infectious Diseases, Department of Medicine, University of ColoradoAnschutz Medical Campus, Aurora, CO, USA; dDivision of Infectious Diseases, University of Wisconsin School of Medicine and Public Health, Madison, WI, USA
(Received 6 July 2014; accepted 30 August 2014) While HIV has become a largely chronic disease, age-associated comorbidities are prevalent in people living with HIV (PLWH). Therefore, PLWH are appropriate for advance care planning (ACP) and advance directives (ADs) completion. We sought to characterize AD completion among outpatient PLWH. We conducted a retrospective chart review of PLWH who receive their routine care at the University of Wisconsin HIV clinic. Data were extracted from the electronic health record. Variables were entered into a stepwise multivariate logistic regression model to assess which factors were independently associated with AD completion. Five hundred and eighty eight charts were reviewed. Eighty-one percent of subjects were male and 72% were white; mean age was 46.8 years. ADs were completed by 134 subjects and 6.7% of those were completed at the HIV clinic. In the final multivariate model, those who had completed an AD were more likely to be older than age 45; ever been diagnosed with AIDS; have cardiovascular disease, neurologic disorder, chronic kidney disease, or malignancy. In this study, a small percentage of patients had documented ADs, with only a small proportion completed in the HIV clinic. The HIV clinic is an underutilized resource to offer ACP. Interventions are needed to provide the necessary ACP resources for PLWH. Keywords: HIV/AIDS; outpatient clinic; patient care; advance directives; advance care planning
Background By 2015, the majority of the people living with human immunodeficiency virus (PLWH) in the USA will be 50 years or older (CDC, 2007, 2013). The development and availability of antiretroviral therapy (ART) has transformed HIV care in the USA into a largely manageable, chronic disease. With effective ART, PLWH can expect a lifespan that is near the average of those living without HIV (Antiretroviral Therapy Cohort Collaboration, 2008; Collaboration of Observational HIV Epidemiological Research Europe (COHERE) in EuroCoord et al., 2012; May et al., 2014). Despite improvements in overall mortality, age-associated comorbidities such as cardiovascular disease, diabetes, cancer, functional impairment, and neurocognitive impairment are common and may be more prevalent in PLWH than in HIVuninfected persons (Boccara et al., 2013; Pathai, Bajillan, Landay, & High, 2013; Work Group for the HIV and Aging Consensus Project, 2012). Thus, with the increased disease burden associated with aging, decisions about preferences for care provided at the end-of life have become increasingly relevant. Advance care planning (ACP) is an organized process of communication to help individuals understand, reflect upon, and discuss goals for future healthcare decisions in the context of their values and beliefs *Corresponding author. Email: [email protected] © 2014 Taylor & Francis
(Wisconsin Medical Society, 2014). Advance directives (ADs), which allow patients to communicate health-care preferences in the event that he or she is no longer able to make decisions, are one component of ACP (Wisconsin Medical Society, 2014). Recent studies have shown that ACP discussions correlate with better health-care outcomes and satisfaction with care among various chronic disease populations such as chronic obstructive pulmonary disease (COPD) and diabetes. Patients with COPD report that ACP improves their health care and that the process of discussion regarding prognosis, treatment concerns, and preferences for care involved in ACP is valuable (Janssen, Engelberg, Wouters, & Curtis, 2012; MacPherson, Walshe, O’Donnell, & Vyas, 2012). Similarly, insulin-dependent diabetics have expressed a desire to discuss ADs (Toth, Gill, Godkin, & Lee, 1998). Findings by Martin et al. suggest that ACP can help HIV patients consider life goals they want to achieve, appoint a preferred proxy to whom they can express their preferences, strengthen relationships with loved ones, and prepare for death (Martin, Thiel, & Singer, 1999). They are also considered valuable by younger HIV patients when age-appropriate ACP documents are provided (Wiener et al., 2008). In addition, end-of-life discussions have not been shown to unduly distress patients with chronic diseases including PLWH
AIDS Care (Evangelista et al., 2012; Lyon, Jacobs, Briggs, Cheng, & Wang, 2013; Lyon et al., 2010). Early, frequent, and thorough communication about ACP in PLWH may result in improved health outcomes and satisfaction. The overall prevalence of completed ADs in the USA is quite variable. AD completion rates in the general US population range from 26–44% (Matzo, Hijjazi, & Outwater, 2008; McAuley, McCutcheon, & Travis, 2008; Ramsaroop, Reid, & Adelman, 2007; Rao, Anderson, Lin, & Laux, 2014). Up to 72% of US adults aged 60 and older who were recently hospitalized and subsequently died had completed an AD (Silveira, Wiitala, & Piette, 2014). Similarly, 61–70% of older adults in long-term support settings such as nursing homes have an AD (Hirschman, Abbott, Hanlon, Prvu Bettger, & Naylor, 2012; Resnick, Schuur, Heineman, Stone, & Weissman, 2008). Among individuals with specific chronic diseases such as heart failure, diabetes, cystic fibrosis, and cancer, AD completion rates vary from 30–47% (Dunlay, Swetz, Mueller, & Roger, 2012; Evangelista et al., 2012; Hubert et al., 2013; Salmond & David, 2005; Sawicki, Dill, Asher, Sellers, & Robinson, 2008; Schellinger, Sidebottom, & Briggs, 2011; Toth et al., 1998; van Oorschot, Schuler, Simon, & Flentje, 2012). The 2012 Working Group for the HIV and Aging Consensus Project recommends that “older individuals with HIV, especially those with substantial illness burden, should be counseled in completing a durable power of attorney for health care and an advance directive, such as the Physician Order for Life Sustaining Treatment, or similar document” (Work Group for the HIV and Aging Consensus Project, 2012). Studies from the earlier AIDS era found that a majority of PLWH (62–84%) lacked an AD or ACP discussion (Gilligan & Jensen, 1995; Teno, Fleishman, Brock, & Mor, 1990; Wenger et al., 2001). A recent retrospective study of hospitalized PLWH by de Caprariis and colleagues reported only 8% of patients had an AD prior to hospital admission (de Caprariis, Carballo-Diéguez, Thompson, & Lyon, 2012). Erlandson et al. found that 47% of older PLWH on effective ART had completed some component of ACP (Erlandson et al., 2012). However, none of the reported studies assessed AD completion rates in the outpatient setting regardless of age or HIV-disease control. The aim of this study was to examine factors associated with AD completion among PLWH and identify the clinical setting in which they were completed.
371
and Clinics in Madison, WI. Patients were eligible for inclusion if they were (1) at least 18 years of age, (2) not imprisoned or institutionalized in mental health facilities, (3) did not have a legal guardian, and (4) were currently enrolled as patients of the UW HIV clinic as of 31 October 2013. The clinical and administrative database managed by the UW HIV clinic was used to identify patients who met these criteria. Trained study personnel (J.B. and B.B.) conducted a review of the electronic health record (EHR) for information that was not available through the clinical database. The study protocol was approved by the Institutional Review Board at the UW School of Medicine and Public Health. Variables Descriptive variables obtained through chart review and abstraction included: sociodemographic information (age, gender, race, ethnicity, relationship status, insurance status, country of origin), HIV risk factors (injection drug use, men who have sex with men [MSM], maternalfetal transmission, blood transfusion, unprotected heterosexual intercourse), characteristics of HIV-disease status (AIDS-defining illness, CD4 nadir, ever had a positive hepatitis C virus [HCV] serology), non-HIV comorbidities, ACP status (completed AD, setting in which AD completed), and health-care utilization (number of HIV clinic visits in previous five years). Insurance status was divided into private insurance and public or no insurance. Public insurance included Medicare, Medicaid, other state or Ryan White grant funding. Non-HIV comorbidities include prior diagnosis of any of the following: cardiovascular disease (myocardial infarction, coronary artery bypass graft, heart failure, arrhythmia, valvulopathy), cerebrovascular disease (stroke, transient ischemic attack [TIA]), diabetes mellitus (clinical diagnosis, use of hypoglycemic medications), chronic kidney disease (any stage), metabolic bone disease (osteopenia, osteoporosis), Central Nervous System (CNS) neurologic disorder (seizure disorder, cognitive impairment, dementia, encephalopathy not otherwise specified [NOS]), psychiatric illness (depression, anxiety, bipolar disorder, schizophrenia, psychotic illness), malignancy (excluding non-melanoma skin cancer), liver disease (hepatitis, hepatosteatosis, cirrhosis), and lung disease (COPD, asthma, obstructive sleep apnea, pulmonary embolism). Analysis
Methods Study population We conducted a retrospective chart review of HIVseropositive patients who received routine care in the HIV clinic at the University of Wisconsin (UW) Hospital
All sociodemographic and clinical variables were entered into a de-identified database for analysis. Descriptive statistics were used to characterize the study population. Study subjects who reported having a completed AD in the UW EHR were compared to those without an AD. Sociodemographic and clinical variables were
372
J.A. Barocas et al.
dichotomized and analyzed by Wald chi-squared tests or Fisher’s exact test where appropriate, with significance defined as p < 0.05. All variables were considered for entry into a stepwise multivariate logistic regression model to assess which factors were independently associated with AD. Variables were entered into the model at p < 0.10 and retained in the model at p < 0.05. Odds ratio (OR) and 95% confidence intervals (CI) were presented for the final model. Statistical analyses were conducted using STATA Version 11 (Cary, NC).
Results A total of 588 patient records that met inclusion criteria were reviewed for EHR data abstraction. Characteristics of the study subjects are shown in Table 1. The mean age of study subjects was 46.8 years (standard deviation (SD) ± 0.47) and mean CD4 count 634 cells/µL (SD ± 12.5 cells/µL). Most subjects were male and selfidentified as MSM. Nineteen percent of subjects were in a MSM relationship at the time of data extraction. Subjects were predominately white (72%) or AfricanAmerican (22%) whereas Asian/Pacific Islander and American Indian/Alaska Natives were smaller proportions of our study population, 1% and 4%, respectively. Subjects had been living with HIV for an average of 13.3 years (SD ± 4 months). Fifty-seven percent of subjects had at least 10 HIV-related outpatient visits in the previous five years and subjects were followed in the clinic for a mean of 8.7 years (SD ± 3 months). For 308 subjects (52%), their HIV provider also served as their primary care provider. Only one subject had no health coverage at the time of study eligibility, otherwise 41% of subjects held private health insurance. The majority of subjects had at least one non-HIV-related comorbidity with psychiatric illness being the most frequent (40%). Thirty-nine percent (228/588) had ever received a diagnosis of AIDS, 7% had ever had a positive HCV serology, and 5% reported previous or current injection drug use. ADs were completed by 134 subjects (23%). The average time since diagnosis of HIV was longer in those who completed ADs than those who did not (15 years versus 12.5 years). Table 2 shows the results of the stepwise multivariate logistic regression model measuring the association of subject characteristics with AD completion. In the final multivariate model, those who had completed an AD were more likely to be at least 45 years of age (OR: 3.4, 95% CI: 2.0–5.7); ever been diagnosed with AIDS (OR: 1.7, 95% CI: 1.1–2.7); have cardiovascular disease (OR: 2.3, 95% CI: 1.1–4.6), neurologic disorder (OR: 5.0, 95% CI: 2.2–12.1), chronic kidney disease (OR: 3.3, 95% CI: 1.3–8.3), or malignancy (OR: 2.8, 95% CI: 1.3–6.0).
Of the 134 subjects with completed ADs, 6.7% were completed at the UW HIV clinic, while the majority were completed in the inpatient or pre-surgical setting (44%), another non-UWHIV outpatient clinic or community-based organization (CBO; 34%), or with an attorney/notary (13%), as shown in Figure 1. More subjects in our study completed their AD per year in the last three years compared to prior time periods. From 1 January 2010 to 31 October 2013, 47 subjects completed an AD whereas from 1 January 1990 to 31 December 1999, a total of 10 subjects completed an AD and from 1 January 2000 to 31 December 2009, a total of 77 subjects completed an AD. Discussion and conclusions In this retrospective review of PLWH who were receiving care at the UW HIV clinic, we found ADs were completed by less than one-quarter of subjects. In our study, PLWH who had completed ADs were more likely to be older; have a history of AIDS-defining illness; and have comorbidities including cardiovascular disease, neurologic disease, chronic kidney disease, or malignancy. Furthermore, the vast majority of PLWH did not complete their AD at the UW HIV clinic, but instead, were more likely to complete an AD during an inpatient or pre-surgical hospitalization or at another non-UWHIV outpatient clinic or CBO. AD completion percentage in this study was similar to those reported in other studies of PLWH. Our finding that 23% had completed ADs falls in between the two most recently reported completion percentages for PLWH (8–47%), but is lower than those of other chronic diseases (Barakat et al., 2013; de Caprariis et al., 2012; Erlandson et al., 2012; Rao et al., 2014; Salmond & David, 2005; Schellinger et al., 2011). Several previously identified barriers to ACP for PLWH may contribute to low AD completion. Since there are many competing priorities, HIV providers may not offer the opportunity to complete an AD in clinic. Additionally, HIV providers have expressed discomfort with end-oflife discussions particularly in situations in which the patients are younger or there are cultural differences between provider and patient (Curtis & Patrick, 1997). Providers may have concerns about their patients’ perceptions of the ACP discussion and possibly unmasking patient anxiety or depression about end of life issues (Curtis & Patrick, 1997; Gott et al., 2009; Knauft, Nielsen, Engelberg, Patrick, & Curtis, 2005). These barriers are similar to those expressed by providers treating other chronic diseases (Gott et al., 2009; MacPherson et al., 2012). We found certain subsets of PLWH are more likely to complete ADs than others. Older patients were more likely to have completed an AD, similar to studies of the
AIDS Care
373
Table 1. Characteristics of PLWH, by completion of advance directive (AD). Characteristics Total (%) Age
