Original Article
Knowledge About Hospice: Exploring Misconceptions, Attitudes, and Preferences for Care
American Journal of Hospice & Palliative Medicine® 2016, Vol. 33(1) 27-33 ª The Author(s) 2014 Reprints and permission: sagepub.com/journalsPermissions.nav DOI: 10.1177/1049909114546885 ajhpm.sagepub.com
John G. Cagle, PhD, MSW1, Daniel J. Van Dussen, PhD2, Krystal L. Culler, BA3, Iraida Carrion, PhD, LCSW4, Seokho Hong, MSW1, Jack Guralnik, MD5, and Sheryl Zimmerman, PhD6,7
Abstract Context: Poor knowledge and misperceptions about hospice are believed to be common, but there is scant evidence about the public’s understanding of hospice. Objectives: To examine hospice knowledge among geographically diverse adults; and describe linkages between knowledge, attitudes and beliefs about hospice care, and demographics. Methods: A small cross-sectional telephone survey of adults living in the contiguous US was conducted using randomly selected numbers (cell phone and landline) and over-sampling of minorities. Measures assessed knowledge (23-item test), attitudes (8-item scale), experiences and preferences related to hospice. Results: 123 participants completed surveys (response rate 46%). 106 (86%) had heard about hospice, 65 (54%) of whom had a personal experience with hospice. Participants had an average hospice knowledge test score of 18 (SD ¼ 3.4) indicating moderate knowledge of hospice. A majority of respondents (62%) did not know that hospice cannot provide concurrent cure-oriented care. Misperceptions about eligibility, coverage of hospice, the provision of hospice in nursing homes or to persons who live alone were not uncommon (missed by >20%). Greater knowledge of hospice was associated with more favorable attitudes about the hospice philosophy of care (r ¼ .22, p ¼ .023) and a greater preference for hospice (p ¼ .049). Respondents who were more educated, worked in the medical field, were non-Hispanic White, and had direct experience with hospice were also more likely to be more knowledgeable about hospice. Conclusion: Despite relatively high hospice awareness and favorability, myths and misperceptions about hospice still abound – and may drive ethnic disparities in end-of-life care. Educational interventions and future study are needed. Keywords hospice knowledge, end of life, preference for hospice, myths and misconceptions about hospice, attitudes and beliefs, Medicare Hospice Benefit
Background Approximately 1.65 million patients with life-threatening illness in the United States receive hospice care annually.1 Hospices provide interdisciplinary care and support to dying patients and their families by focusing on their physical, emotional, social, and spiritual needs. According to the Medicare Hospice Benefit, hospice patients must (1) have a life expectancy 6 months and (2) forego aggressive, cure-focused care.2 Studies have consistently demonstrated that hospice provides superior end-of-life care in terms of effective pain management, preservation of patient dignity, and family satisfaction when compared to alternate forms of care for terminally ill individuals.3,4 Health insurance plans, including Medicare and Medicaid, pay for hospice services. However, an estimated 40% of individuals who are eligible for hospice care do not receive it5—and minority
1
School of Social Work, University of Maryland, Baltimore, MD, USA Department of Gerontology, Youngstown State University, Youngstown, OH, USA 3 College of Health Solutions, Arizona State University, Phoenix, AZ, USA 4 School of Social Work, University of South Florida, Tampa, FL, USA 5 Division of Gerontology, Department of Epidemiology and Public Health, University of Maryland School of Medicine, Baltimore, MD, USA 6 School of Social Work, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA 7 Cecil G. Sheps Center for Health Services Research, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA 2
Corresponding Author: John G. Cagle, PhD, MSW, School of Social Work, University of Maryland, 525 West Redwood Street, 3W13, Baltimore, MD 21201, USA. Email: [email protected]
Downloaded from ajh.sagepub.com at University of York on February 17, 2016
American Journal of Hospice & Palliative Medicine® 33(1)
28 populations (Hispanic and African American, in particular) are underrepresented in hospice.1,6-10 Knowledge of hospice is an important and necessary component of end-of-life decision making.11,12 Yet evidence suggests a majority of patients and family members have no knowledge of hospice when they are referred, which further suggests that many families are being referred to hospice care without a good understanding of its scope of services, benefits, and limitations.13 Poor knowledge and misperceptions about hospice may contribute to poor attitudes about hospice, low rates of enrollment, and late referrals.12-16 Misperceptions about hospice care may also help explain existing racial and ethnic disparities in endof-life care. One study of minorities suggested that improved knowledge of hospice could increase access, timeliness of referrals, and better quality of care.17 Specifically, underutilization of hospice services and hospice referrals may be attributable to a lack of knowledge about hospice among minority groups.7,16,18-20 Understanding public perceptions about hospice may also help inform public education efforts to improve awareness of hospice as an option at the end of life. Although misconceptions about hospice are believed to be common, no studies have systematically investigated the prevalence of these beliefs in a broad geographically diverse sample. The goals of this study were to provide preliminary data on knowledge of hospice, including prevalence of misconceptions, and to describe the associations between knowledge of hospice, attitudes about hospice, preference for hospice and demographic characteristics.
Method We conducted a small cross-sectional telephone-based survey of community-dwelling adults aged 18þ living in the contiguous United States. A small population-based sample (N ¼ 123) was obtained using a blended sampling frame of randomly selected, validated cell phone and landline numbers. Oversampling was used to ensure sufficient representation of Hispanic and African American individuals. The survey was offered in English and Spanish. Surveys took 20 to 25 minutes to complete, and respondents were offered US$20 to US$30 compensation. Mobile phone users were offered slightly higher compensation to offset any charges related to use of cell phone minutes. The computerassisted telephone interviewing system was used to facilitate data collection.21 Data collection began in August 2012 and lasted for 6 weeks. Study methods were approved by institutional review boards at the University of Maryland, Baltimore, the University of North Carolina, and Youngstown State University.
Measures Survey topics included knowledge and attitudes about hospice, sources of information about hospice, previous experiences with hospice care, and demographics. Key variables are described in the following section.
Familiarity with hospice. Respondents’ self-described familiarity with hospice was assessed with 2 items. The first item asked: Have ever you heard of the term hospice? (Yes/No). Respondents who had heard of the term hospice were subsequently asked to describe their familiarity with hospice care, ranging from 1 ¼ not very familiar to 3 ¼ very familiar. Knowledge of Hospice. For individuals who reported that they had heard of hospice, we assessed their knowledge of hospice using a 23-item true/false test. The test was an expanded version of a 15-item knowledge of hospice test.22 With permission and input from the original authors, we added 8 additional items to the test. Face validity was established by an expert interdisciplinary panel of physicians, gerontologists, nurses, social workers, and psychologists and by literature review on common myths and misperceptions about hospice. The test is scored by totaling the number of correct responses (total possible score 0-23). Ordering of questions was randomized to prevent order bias. An item response analysis was conducted and all items were retained. The interitem reliability coefficient for the knowledge test was a ¼ .75. Attitudes about the hospice philosophy of care. Respondents’ attitudes toward hospice care were measured using an 8-item version of the Hospice Philosophy Scale 8 (HPS-8).23-25 The scale measures whether respondent attitudes are consistent with the hospice philosophy of care. The HPS-8 demonstrated good internal consistency (a ¼ .72 for this study), face, and factorial validity. The scale used branched 5-point response options, combined for a total score ranging from 8 to 40 with higher scores indicating more favorable attitudes toward the hospice philosophy of care. Items on the HPS-8 were also randomized to avoid order bias. Favorability. Respondents were asked whether they perceived hospice favorably using the following question: What is your general opinion of hospice? 1 ¼ not at all favorable to 3 ¼ very favorable. Preference for Hospice. Among those who had heard of hospice, respondents’ preference for hospice was measured using the following question: If you were dying, would you want the support of hospice services? (Yes/No). Direct Experience With Hospice. Respondents were asked whether they had personally experienced hospice services as a patient, close friend, or family member of a patient (Yes/ No). Among those who had a personal experience with hospice, additional questions were asked regarding primary diagnosis, whether the patient was still alive, and, if not, the quality of the death (good/bad). Furthermore, we assessed respondents’ direct level of experience (ie, direct knowledge of hospice) using a 3-point scale, ranging from: 1 ¼ Knew of someone who used hospice but was not familiar with the care they provided to 3 ¼ Knew of someone who used hospice and was very familiar with the care provided. If a respondent had multiple experiences with hospice, they were instructed to provide information on their most recent experience.
Downloaded from ajh.sagepub.com at University of York on February 17, 2016
Cagle et al
29
Analysis
Table 1. Sample Descriptives.a
Sample characteristics were summarized using univariate statistics appropriate to the measure. Multiple logistic regression was used to model demographics on whether respondents had heard of hospice (Yes/No). Linear multiple regression was employed to model demographic predictors of knowledge of hospice test scores. The variable ‘‘Direct Experience with Hospice’’ was added sequentially to this latter model to isolate the relative effects of experience on knowledge. For all models, demographics were simultaneously introduced, namely gender, age (in years), education level, prior experience working in the medical field (Yes/No), and race/ethnicity. Race/ethnicity was introduced using 2 indicator variables and compared non-Hispanic blacks and Hispanics to the reference group, non-Hispanic whites. Data were prescreened for missing values. The item receiving the most refusals (4%) inquired about household income. For the knowledge of hospice test, responses of ‘‘don’t know’’ were treated as incorrect. For the HPS-8, missing values, including ‘‘don’t know’’ responses and refusals, were imputed using multiple imputation. Casewise deletion was used for missing values on tests involving all other items. Pearson (continuous data) and Spearman (ordinal data) correlation coefficients were calculated for bivariate associations. Age was grouped into quartiles to depict its relationship with knowledge test scores. To further examine the relationship between knowledge of hospice and the HPS-8, we conducted a series of tests of association for each knowledge test item. We used independent t tests to identify mean differences on the HPS-8 between 2 groups: those who answered a given item correctly and those who did not. Mean differences in the HPS-8 were computed (mean incorrect mean correct) and reported. Using this calculation, a positive mean difference indicates that a correct score on a particular test item is associated with positive attitudes toward the hospice philosophy of care.
Characteristic
Results A total of 123 eligible individuals completed the survey (response rate of 46%). As Table 1 shows, nearly two-thirds (64%) of the respondents were female, majority (61%) white, with an average age of 48 years (standard deviation [SD] ¼ 17.4). Fifty-eight percent (58%; N ¼ 71) of the respondents completed the survey via cell phone, while the remaining 42% (N ¼ 52) were on a landline. A quarter (24%) of the participants identified themselves as black/African American. Thirteen percent (N ¼ 16) identified as Hispanic/Latino, 31% (N ¼ 5) of whom elected to complete the survey in Spanish. Based on regional census categories, 44% of respondents resided in the South, 28% in the Midwest, 18% in the West, and 11% were from the Northeast. A majority (86%; N ¼ 106) of respondents had heard of hospice, and most heard about it from a friend or family member (42%), while 28% heard about it through personal experience and 10% from television. The remaining respondents heard about it from a health care provider (7%), the
Age, M + SD Gender Female Male Race White Black Other Sampling strata Cell phone Landline US census region Midwest Northeast South West Ethnicity Hispanic Non-Hispanic Marital status Married/partnered Not married/partnered Employment status Employed full-time Employed part-time Student Retired Unemployed—seeking work Unemployed—not seeking work Education High school or less Some college or more Income Less than US$19 999 US$20 000-US$59 999 US$60 000-US$99 999 Over US$100 000 Insurance Medicare Medicaid Military coverage/VA benefits Private or employer coverage Some other type of insurance No coverage Self-rated health,b M + SD
N (%) 47.7 + 17.4 78 (63.4) 45 (36.6) 75 (61.0) 29 (23.6) 19 (15.4) 71 (58.0) 52 (42.0) 34 (27.6) 13 (10.7) 54 (43.9) 22 (17.9) 16 (13.0) 107 (87.0) 71 (58.2) 51 (41.8) 57 (46.3) 11 (8.9) 9 (7.3) 19 (15.4) 2 (9.8) 14 (11.4) 42 (34.1) 81 (65.9) 34 (29.3) 33 (28.4) 25 (21.6) 24 (20.7) 23 (18.7) 8 (6.5) 9 (7.3) 66 (53.7) 4 (3.3) 13 (10.6) 3.5 + 1.2
Abbreviations: SD, standard deviation; VA, veterans affairs. a N ¼ 123. b Self-rated health ranges from 1 (poor) to 5 (excellent).
Internet (2%), written material (2%), or another source (9%). As shown in Table 2, those who were unfamiliar with hospice were more likely to be younger (38 vs 58 years; P ¼ .016) and Hispanic. Forty-four percent of Hispanic respondents had not heard of hospice compared to 9% of nonHispanic individuals (P ¼ .038). Respondents who had heard of hospice were asked about their opinion of it, and responses were generally positive; 95% reported having a very favorable or somewhat favorable
Downloaded from ajh.sagepub.com at University of York on February 17, 2016
American Journal of Hospice & Palliative Medicine® 33(1)
30 Table 2. Predictors of Whether Respondents Had Heard of Hospice (Yes vs No).a Odds ratio (SE)
P
0.378 (1.32) 1.294 (0.641) 1.045 (0.018) 1.246 (0.215) 6.352 (1.087)
.461 .687 .016 .306 .089
Associations Between Knowledge, Attitudes, and Preferences for Hospice
Abbreviation: SE, standard error. a N ¼ 123. Predictors with P-values 5 years ago; and cancer was the leading patient diagnosis (52%), followed by heart disease/stroke (12%). Overall, 89% of respondents were either somewhat or very satisfied with the hospice care the patient received. In 61 of these cases, the hospice patient had died. The vast majority (92%; N ¼ 56) of deaths were expected and all but 1 were receiving hospice at the time of death. Most deaths (82%; N ¼ 50) were described as ‘‘good.’’
Knowledge of Hospice Among the 106 individuals given the knowledge test, the group answered on average 78% of questions correctly (mean [M] ¼ 18 items correct [SD ¼ 3.4] of a possible 23). Table 3 shows each test item listed in order from the highest proportion of incorrect responses to the least. Seven test items were incorrectly answered by >25% of respondents. First, only 38% of respondents correctly knew that hospice cannot provide curative care. Second, only 43% of respondents knew that hospices continue to provide support for families after the patient’s death. Third, 63% of respondents correctly answered that anyone can make a referral to hospice. Fourth, two-thirds (67%) were aware that hospice services are typically covered by insurance. Fifth, 70% were aware that hospice care can be provided in a nursing home. Sixth, 72% knew that hospice is available to any individual expected to live 6 months or less. Tied for the sixth most frequently missed question, 72% of those who took the test correctly knew that hospices do not necessarily have to stop services if an individual lives longer than 6 months.
Table 3 also shows associations between individual knowledge test items and overall scores on the HPS-8. Specifically, 3 knowledge items were related to scores on the HPS-8. Respondents had more positive attitudes about the hospice philosophy of care when they correctly answered the following: (1) ‘‘Hospice staff are available by phone any time, day or night’’ [TRUE] (P ¼ .019); (2) ‘‘Hospice provides grief support to families’’ [TRUE] (P ¼ .006); and (3) ‘‘Hospice must stop services if an individual lives longer than six months’’ [FALSE] (P ¼.024).
Correlates of Hospice Knowledge and the Relative Effect of Direct Experience Knowledge test scores were negatively correlated with age (r ¼ .22, P ¼ .027). As Figure 1 shows, older respondents were less knowledgeable about hospice. However, older adults tended to have more favorable attitudes about the hospice philosophy of care (r ¼ .30, P ¼ .002). As seen in Table 4, when adjusting for demographics, 3 characteristics were associated with knowledge scores: education, experience working in the medical field, and race/ethnicity. Respondents who were more educated, or had prior work experience in the medical field, were more likely to have better knowledge of hospice (P ¼ .015 for both). Additionally, non-Hispanic black individuals tended to have lower knowledge of hospice (P < .001). On average, this group missed 4 more test questions than their caucasian counterparts (M ¼ 14.4, SD ¼ 6.8 vs M ¼ 18.4, SD ¼ 4.3, P ¼ .006). Age was not significantly associated with knowledge in this model as observed in bivariate tests. When direct experience with hospice was introduced (model 2 in Table 4), it was also associated with knowledge scores. More specifically, those who had more direct experience with hospice tended to have higher knowledge test scores (P ¼ .016). Additionally, when accounting for respondents’ direct
Downloaded from ajh.sagepub.com at University of York on February 17, 2016
Cagle et al
31
Table 3. Knowledge of Hospice Test Results by Item and Their Association with Attitudes Toward End-of-Life Scoresa Knowledge of Hospice test items [Answer] 1. While receiving hospice care, individuals can also receive treatments, such as chemotherapy, radiation, or surgery, to cure the person’s primary illness. [FALSE] 2. Hospice services end when the hospice patient dies. [FALSE] 3. Anyone can make a referral to hospice. [TRUE] 4. Hospice services are typically paid for by health insurance. [TRUE] 5. Hospice care can be provided in a nursing home. [TRUE] 6. Hospice must stop services if an individual lives longer than six months. [FALSE] 7. Hospice care is available to any individual expected to live six months or less. [TRUE] 8. People who live alone are able to receive hospice services. [TRUE] 9. A primary goal of hospice is to treat emotional needs of the dying individual and their family. [TRUE] 10. Individuals receiving hospice care cannot be taken to the hospital for treatment. [FALSE] 11. Hospice provides medications, treatments, medical equipment and supplies that are related to a patient’s primary illness. [TRUE] 12. Hospice care is only appropriate for people who have a few days to live. [FALSE] 13. A person with Alzheimer’s disease cannot have hospice services. [FALSE] 14. While in hospice, spouses, partners or children of the dying individual are not allowed to directly participate in his or her care. [FALSE] 15. Hospice provides grief support to families. [TRUE] 16. Only individuals over age 65 are eligible for hospice services. [FALSE] 17. Hospice care cannot be provided at home. [FALSE] 18. Hospice staff are available by phone any time, day or night. [TRUE] 19. Hospice care focuses on managing an individual’s pain and other symptoms. [TRUE] 20. Hospice care is available only to individuals with cancer. [FALSE] 21. Hospice care helps caregivers and family members as well as the dying individual. [TRUE] 22. Hospice care helps the dying individual by speeding up the dying process. [FALSE] 23. The hospice care team includes physicians, nurses, social workers, and chaplains. [TRUE]
% Incorrect
HPS-8b Mdif
P-valuec
62.3
-.11
.742
57.5 36.8 33.0 30.2 28.3 28.3 21.7 21.7 18.9 18.9
.23 .41 .002 .22 .82 .69 .41 .55 .05 -.32
.496 .234 .995 .541 .024 .059 .305 .168 .899 .452
15.1 15.1 13.2
.23 .31 .91
.622 .512 .061
12.3 11.3 9.4 8.5 8.5 6.6 5.7 5.7 2.8
1.36 .23 1.04 1.38 .41 .35 .22 .30 -.15
.006 .654 .065 .019 .491 .600 .758 .678 .880
Abbreviation: HPS-8, Hospice Philosophy Scale 8. a N ¼ 106. b The HPS-8 is an 8-item measure adapted from the Gerbino Hospice Philosophy Scale to assess general attitudes toward end-of-life care. Higher scores signify better attitudes toward end-of-life care. Possible score ranges from 8 to 40. The mean difference of HPS-8 between 2 groups (mean incorrect - mean correct) was calculated for each item. Thus, a positive mean difference indicates that a correct score on that particular item is associated with attitudes in favor of the hospice philosophy of care. c P-values pertain to the analysis of mean differences on the HPS-8 by knowledge test item; P-values
Knowledge About Hospice: Exploring Misconceptions, Attitudes, and Preferences for Care
American Journal of Hospice & Palliative Medicine® 2016, Vol. 33(1) 27-33 ª The Author(s) 2014 Reprints and permission: sagepub.com/journalsPermissions.nav DOI: 10.1177/1049909114546885 ajhpm.sagepub.com
John G. Cagle, PhD, MSW1, Daniel J. Van Dussen, PhD2, Krystal L. Culler, BA3, Iraida Carrion, PhD, LCSW4, Seokho Hong, MSW1, Jack Guralnik, MD5, and Sheryl Zimmerman, PhD6,7
Abstract Context: Poor knowledge and misperceptions about hospice are believed to be common, but there is scant evidence about the public’s understanding of hospice. Objectives: To examine hospice knowledge among geographically diverse adults; and describe linkages between knowledge, attitudes and beliefs about hospice care, and demographics. Methods: A small cross-sectional telephone survey of adults living in the contiguous US was conducted using randomly selected numbers (cell phone and landline) and over-sampling of minorities. Measures assessed knowledge (23-item test), attitudes (8-item scale), experiences and preferences related to hospice. Results: 123 participants completed surveys (response rate 46%). 106 (86%) had heard about hospice, 65 (54%) of whom had a personal experience with hospice. Participants had an average hospice knowledge test score of 18 (SD ¼ 3.4) indicating moderate knowledge of hospice. A majority of respondents (62%) did not know that hospice cannot provide concurrent cure-oriented care. Misperceptions about eligibility, coverage of hospice, the provision of hospice in nursing homes or to persons who live alone were not uncommon (missed by >20%). Greater knowledge of hospice was associated with more favorable attitudes about the hospice philosophy of care (r ¼ .22, p ¼ .023) and a greater preference for hospice (p ¼ .049). Respondents who were more educated, worked in the medical field, were non-Hispanic White, and had direct experience with hospice were also more likely to be more knowledgeable about hospice. Conclusion: Despite relatively high hospice awareness and favorability, myths and misperceptions about hospice still abound – and may drive ethnic disparities in end-of-life care. Educational interventions and future study are needed. Keywords hospice knowledge, end of life, preference for hospice, myths and misconceptions about hospice, attitudes and beliefs, Medicare Hospice Benefit
Background Approximately 1.65 million patients with life-threatening illness in the United States receive hospice care annually.1 Hospices provide interdisciplinary care and support to dying patients and their families by focusing on their physical, emotional, social, and spiritual needs. According to the Medicare Hospice Benefit, hospice patients must (1) have a life expectancy 6 months and (2) forego aggressive, cure-focused care.2 Studies have consistently demonstrated that hospice provides superior end-of-life care in terms of effective pain management, preservation of patient dignity, and family satisfaction when compared to alternate forms of care for terminally ill individuals.3,4 Health insurance plans, including Medicare and Medicaid, pay for hospice services. However, an estimated 40% of individuals who are eligible for hospice care do not receive it5—and minority
1
School of Social Work, University of Maryland, Baltimore, MD, USA Department of Gerontology, Youngstown State University, Youngstown, OH, USA 3 College of Health Solutions, Arizona State University, Phoenix, AZ, USA 4 School of Social Work, University of South Florida, Tampa, FL, USA 5 Division of Gerontology, Department of Epidemiology and Public Health, University of Maryland School of Medicine, Baltimore, MD, USA 6 School of Social Work, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA 7 Cecil G. Sheps Center for Health Services Research, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA 2
Corresponding Author: John G. Cagle, PhD, MSW, School of Social Work, University of Maryland, 525 West Redwood Street, 3W13, Baltimore, MD 21201, USA. Email: [email protected]
Downloaded from ajh.sagepub.com at University of York on February 17, 2016
American Journal of Hospice & Palliative Medicine® 33(1)
28 populations (Hispanic and African American, in particular) are underrepresented in hospice.1,6-10 Knowledge of hospice is an important and necessary component of end-of-life decision making.11,12 Yet evidence suggests a majority of patients and family members have no knowledge of hospice when they are referred, which further suggests that many families are being referred to hospice care without a good understanding of its scope of services, benefits, and limitations.13 Poor knowledge and misperceptions about hospice may contribute to poor attitudes about hospice, low rates of enrollment, and late referrals.12-16 Misperceptions about hospice care may also help explain existing racial and ethnic disparities in endof-life care. One study of minorities suggested that improved knowledge of hospice could increase access, timeliness of referrals, and better quality of care.17 Specifically, underutilization of hospice services and hospice referrals may be attributable to a lack of knowledge about hospice among minority groups.7,16,18-20 Understanding public perceptions about hospice may also help inform public education efforts to improve awareness of hospice as an option at the end of life. Although misconceptions about hospice are believed to be common, no studies have systematically investigated the prevalence of these beliefs in a broad geographically diverse sample. The goals of this study were to provide preliminary data on knowledge of hospice, including prevalence of misconceptions, and to describe the associations between knowledge of hospice, attitudes about hospice, preference for hospice and demographic characteristics.
Method We conducted a small cross-sectional telephone-based survey of community-dwelling adults aged 18þ living in the contiguous United States. A small population-based sample (N ¼ 123) was obtained using a blended sampling frame of randomly selected, validated cell phone and landline numbers. Oversampling was used to ensure sufficient representation of Hispanic and African American individuals. The survey was offered in English and Spanish. Surveys took 20 to 25 minutes to complete, and respondents were offered US$20 to US$30 compensation. Mobile phone users were offered slightly higher compensation to offset any charges related to use of cell phone minutes. The computerassisted telephone interviewing system was used to facilitate data collection.21 Data collection began in August 2012 and lasted for 6 weeks. Study methods were approved by institutional review boards at the University of Maryland, Baltimore, the University of North Carolina, and Youngstown State University.
Measures Survey topics included knowledge and attitudes about hospice, sources of information about hospice, previous experiences with hospice care, and demographics. Key variables are described in the following section.
Familiarity with hospice. Respondents’ self-described familiarity with hospice was assessed with 2 items. The first item asked: Have ever you heard of the term hospice? (Yes/No). Respondents who had heard of the term hospice were subsequently asked to describe their familiarity with hospice care, ranging from 1 ¼ not very familiar to 3 ¼ very familiar. Knowledge of Hospice. For individuals who reported that they had heard of hospice, we assessed their knowledge of hospice using a 23-item true/false test. The test was an expanded version of a 15-item knowledge of hospice test.22 With permission and input from the original authors, we added 8 additional items to the test. Face validity was established by an expert interdisciplinary panel of physicians, gerontologists, nurses, social workers, and psychologists and by literature review on common myths and misperceptions about hospice. The test is scored by totaling the number of correct responses (total possible score 0-23). Ordering of questions was randomized to prevent order bias. An item response analysis was conducted and all items were retained. The interitem reliability coefficient for the knowledge test was a ¼ .75. Attitudes about the hospice philosophy of care. Respondents’ attitudes toward hospice care were measured using an 8-item version of the Hospice Philosophy Scale 8 (HPS-8).23-25 The scale measures whether respondent attitudes are consistent with the hospice philosophy of care. The HPS-8 demonstrated good internal consistency (a ¼ .72 for this study), face, and factorial validity. The scale used branched 5-point response options, combined for a total score ranging from 8 to 40 with higher scores indicating more favorable attitudes toward the hospice philosophy of care. Items on the HPS-8 were also randomized to avoid order bias. Favorability. Respondents were asked whether they perceived hospice favorably using the following question: What is your general opinion of hospice? 1 ¼ not at all favorable to 3 ¼ very favorable. Preference for Hospice. Among those who had heard of hospice, respondents’ preference for hospice was measured using the following question: If you were dying, would you want the support of hospice services? (Yes/No). Direct Experience With Hospice. Respondents were asked whether they had personally experienced hospice services as a patient, close friend, or family member of a patient (Yes/ No). Among those who had a personal experience with hospice, additional questions were asked regarding primary diagnosis, whether the patient was still alive, and, if not, the quality of the death (good/bad). Furthermore, we assessed respondents’ direct level of experience (ie, direct knowledge of hospice) using a 3-point scale, ranging from: 1 ¼ Knew of someone who used hospice but was not familiar with the care they provided to 3 ¼ Knew of someone who used hospice and was very familiar with the care provided. If a respondent had multiple experiences with hospice, they were instructed to provide information on their most recent experience.
Downloaded from ajh.sagepub.com at University of York on February 17, 2016
Cagle et al
29
Analysis
Table 1. Sample Descriptives.a
Sample characteristics were summarized using univariate statistics appropriate to the measure. Multiple logistic regression was used to model demographics on whether respondents had heard of hospice (Yes/No). Linear multiple regression was employed to model demographic predictors of knowledge of hospice test scores. The variable ‘‘Direct Experience with Hospice’’ was added sequentially to this latter model to isolate the relative effects of experience on knowledge. For all models, demographics were simultaneously introduced, namely gender, age (in years), education level, prior experience working in the medical field (Yes/No), and race/ethnicity. Race/ethnicity was introduced using 2 indicator variables and compared non-Hispanic blacks and Hispanics to the reference group, non-Hispanic whites. Data were prescreened for missing values. The item receiving the most refusals (4%) inquired about household income. For the knowledge of hospice test, responses of ‘‘don’t know’’ were treated as incorrect. For the HPS-8, missing values, including ‘‘don’t know’’ responses and refusals, were imputed using multiple imputation. Casewise deletion was used for missing values on tests involving all other items. Pearson (continuous data) and Spearman (ordinal data) correlation coefficients were calculated for bivariate associations. Age was grouped into quartiles to depict its relationship with knowledge test scores. To further examine the relationship between knowledge of hospice and the HPS-8, we conducted a series of tests of association for each knowledge test item. We used independent t tests to identify mean differences on the HPS-8 between 2 groups: those who answered a given item correctly and those who did not. Mean differences in the HPS-8 were computed (mean incorrect mean correct) and reported. Using this calculation, a positive mean difference indicates that a correct score on a particular test item is associated with positive attitudes toward the hospice philosophy of care.
Characteristic
Results A total of 123 eligible individuals completed the survey (response rate of 46%). As Table 1 shows, nearly two-thirds (64%) of the respondents were female, majority (61%) white, with an average age of 48 years (standard deviation [SD] ¼ 17.4). Fifty-eight percent (58%; N ¼ 71) of the respondents completed the survey via cell phone, while the remaining 42% (N ¼ 52) were on a landline. A quarter (24%) of the participants identified themselves as black/African American. Thirteen percent (N ¼ 16) identified as Hispanic/Latino, 31% (N ¼ 5) of whom elected to complete the survey in Spanish. Based on regional census categories, 44% of respondents resided in the South, 28% in the Midwest, 18% in the West, and 11% were from the Northeast. A majority (86%; N ¼ 106) of respondents had heard of hospice, and most heard about it from a friend or family member (42%), while 28% heard about it through personal experience and 10% from television. The remaining respondents heard about it from a health care provider (7%), the
Age, M + SD Gender Female Male Race White Black Other Sampling strata Cell phone Landline US census region Midwest Northeast South West Ethnicity Hispanic Non-Hispanic Marital status Married/partnered Not married/partnered Employment status Employed full-time Employed part-time Student Retired Unemployed—seeking work Unemployed—not seeking work Education High school or less Some college or more Income Less than US$19 999 US$20 000-US$59 999 US$60 000-US$99 999 Over US$100 000 Insurance Medicare Medicaid Military coverage/VA benefits Private or employer coverage Some other type of insurance No coverage Self-rated health,b M + SD
N (%) 47.7 + 17.4 78 (63.4) 45 (36.6) 75 (61.0) 29 (23.6) 19 (15.4) 71 (58.0) 52 (42.0) 34 (27.6) 13 (10.7) 54 (43.9) 22 (17.9) 16 (13.0) 107 (87.0) 71 (58.2) 51 (41.8) 57 (46.3) 11 (8.9) 9 (7.3) 19 (15.4) 2 (9.8) 14 (11.4) 42 (34.1) 81 (65.9) 34 (29.3) 33 (28.4) 25 (21.6) 24 (20.7) 23 (18.7) 8 (6.5) 9 (7.3) 66 (53.7) 4 (3.3) 13 (10.6) 3.5 + 1.2
Abbreviations: SD, standard deviation; VA, veterans affairs. a N ¼ 123. b Self-rated health ranges from 1 (poor) to 5 (excellent).
Internet (2%), written material (2%), or another source (9%). As shown in Table 2, those who were unfamiliar with hospice were more likely to be younger (38 vs 58 years; P ¼ .016) and Hispanic. Forty-four percent of Hispanic respondents had not heard of hospice compared to 9% of nonHispanic individuals (P ¼ .038). Respondents who had heard of hospice were asked about their opinion of it, and responses were generally positive; 95% reported having a very favorable or somewhat favorable
Downloaded from ajh.sagepub.com at University of York on February 17, 2016
American Journal of Hospice & Palliative Medicine® 33(1)
30 Table 2. Predictors of Whether Respondents Had Heard of Hospice (Yes vs No).a Odds ratio (SE)
P
0.378 (1.32) 1.294 (0.641) 1.045 (0.018) 1.246 (0.215) 6.352 (1.087)
.461 .687 .016 .306 .089
Associations Between Knowledge, Attitudes, and Preferences for Hospice
Abbreviation: SE, standard error. a N ¼ 123. Predictors with P-values 5 years ago; and cancer was the leading patient diagnosis (52%), followed by heart disease/stroke (12%). Overall, 89% of respondents were either somewhat or very satisfied with the hospice care the patient received. In 61 of these cases, the hospice patient had died. The vast majority (92%; N ¼ 56) of deaths were expected and all but 1 were receiving hospice at the time of death. Most deaths (82%; N ¼ 50) were described as ‘‘good.’’
Knowledge of Hospice Among the 106 individuals given the knowledge test, the group answered on average 78% of questions correctly (mean [M] ¼ 18 items correct [SD ¼ 3.4] of a possible 23). Table 3 shows each test item listed in order from the highest proportion of incorrect responses to the least. Seven test items were incorrectly answered by >25% of respondents. First, only 38% of respondents correctly knew that hospice cannot provide curative care. Second, only 43% of respondents knew that hospices continue to provide support for families after the patient’s death. Third, 63% of respondents correctly answered that anyone can make a referral to hospice. Fourth, two-thirds (67%) were aware that hospice services are typically covered by insurance. Fifth, 70% were aware that hospice care can be provided in a nursing home. Sixth, 72% knew that hospice is available to any individual expected to live 6 months or less. Tied for the sixth most frequently missed question, 72% of those who took the test correctly knew that hospices do not necessarily have to stop services if an individual lives longer than 6 months.
Table 3 also shows associations between individual knowledge test items and overall scores on the HPS-8. Specifically, 3 knowledge items were related to scores on the HPS-8. Respondents had more positive attitudes about the hospice philosophy of care when they correctly answered the following: (1) ‘‘Hospice staff are available by phone any time, day or night’’ [TRUE] (P ¼ .019); (2) ‘‘Hospice provides grief support to families’’ [TRUE] (P ¼ .006); and (3) ‘‘Hospice must stop services if an individual lives longer than six months’’ [FALSE] (P ¼.024).
Correlates of Hospice Knowledge and the Relative Effect of Direct Experience Knowledge test scores were negatively correlated with age (r ¼ .22, P ¼ .027). As Figure 1 shows, older respondents were less knowledgeable about hospice. However, older adults tended to have more favorable attitudes about the hospice philosophy of care (r ¼ .30, P ¼ .002). As seen in Table 4, when adjusting for demographics, 3 characteristics were associated with knowledge scores: education, experience working in the medical field, and race/ethnicity. Respondents who were more educated, or had prior work experience in the medical field, were more likely to have better knowledge of hospice (P ¼ .015 for both). Additionally, non-Hispanic black individuals tended to have lower knowledge of hospice (P < .001). On average, this group missed 4 more test questions than their caucasian counterparts (M ¼ 14.4, SD ¼ 6.8 vs M ¼ 18.4, SD ¼ 4.3, P ¼ .006). Age was not significantly associated with knowledge in this model as observed in bivariate tests. When direct experience with hospice was introduced (model 2 in Table 4), it was also associated with knowledge scores. More specifically, those who had more direct experience with hospice tended to have higher knowledge test scores (P ¼ .016). Additionally, when accounting for respondents’ direct
Downloaded from ajh.sagepub.com at University of York on February 17, 2016
Cagle et al
31
Table 3. Knowledge of Hospice Test Results by Item and Their Association with Attitudes Toward End-of-Life Scoresa Knowledge of Hospice test items [Answer] 1. While receiving hospice care, individuals can also receive treatments, such as chemotherapy, radiation, or surgery, to cure the person’s primary illness. [FALSE] 2. Hospice services end when the hospice patient dies. [FALSE] 3. Anyone can make a referral to hospice. [TRUE] 4. Hospice services are typically paid for by health insurance. [TRUE] 5. Hospice care can be provided in a nursing home. [TRUE] 6. Hospice must stop services if an individual lives longer than six months. [FALSE] 7. Hospice care is available to any individual expected to live six months or less. [TRUE] 8. People who live alone are able to receive hospice services. [TRUE] 9. A primary goal of hospice is to treat emotional needs of the dying individual and their family. [TRUE] 10. Individuals receiving hospice care cannot be taken to the hospital for treatment. [FALSE] 11. Hospice provides medications, treatments, medical equipment and supplies that are related to a patient’s primary illness. [TRUE] 12. Hospice care is only appropriate for people who have a few days to live. [FALSE] 13. A person with Alzheimer’s disease cannot have hospice services. [FALSE] 14. While in hospice, spouses, partners or children of the dying individual are not allowed to directly participate in his or her care. [FALSE] 15. Hospice provides grief support to families. [TRUE] 16. Only individuals over age 65 are eligible for hospice services. [FALSE] 17. Hospice care cannot be provided at home. [FALSE] 18. Hospice staff are available by phone any time, day or night. [TRUE] 19. Hospice care focuses on managing an individual’s pain and other symptoms. [TRUE] 20. Hospice care is available only to individuals with cancer. [FALSE] 21. Hospice care helps caregivers and family members as well as the dying individual. [TRUE] 22. Hospice care helps the dying individual by speeding up the dying process. [FALSE] 23. The hospice care team includes physicians, nurses, social workers, and chaplains. [TRUE]
% Incorrect
HPS-8b Mdif
P-valuec
62.3
-.11
.742
57.5 36.8 33.0 30.2 28.3 28.3 21.7 21.7 18.9 18.9
.23 .41 .002 .22 .82 .69 .41 .55 .05 -.32
.496 .234 .995 .541 .024 .059 .305 .168 .899 .452
15.1 15.1 13.2
.23 .31 .91
.622 .512 .061
12.3 11.3 9.4 8.5 8.5 6.6 5.7 5.7 2.8
1.36 .23 1.04 1.38 .41 .35 .22 .30 -.15
.006 .654 .065 .019 .491 .600 .758 .678 .880
Abbreviation: HPS-8, Hospice Philosophy Scale 8. a N ¼ 106. b The HPS-8 is an 8-item measure adapted from the Gerbino Hospice Philosophy Scale to assess general attitudes toward end-of-life care. Higher scores signify better attitudes toward end-of-life care. Possible score ranges from 8 to 40. The mean difference of HPS-8 between 2 groups (mean incorrect - mean correct) was calculated for each item. Thus, a positive mean difference indicates that a correct score on that particular item is associated with attitudes in favor of the hospice philosophy of care. c P-values pertain to the analysis of mean differences on the HPS-8 by knowledge test item; P-values
