Sot. 5~. .L/ed. Vol. 32. No. I. pp. 83-87. 1991 Prmted I” Great Bnram. .&II nghtr reserved
NEED
0277-9536,91 S3.00 + 0.00 CopytIght S 1991 Pergamon Press plc
SATISFACTION
IN TERMINAL NORMA
CARE SETTINGS*
JEAN DAWS~N
School of Nursing of Widener University, Chester,
PA
19013, U.S.A.
Abstract-Research comparing hospice and conventional programs of care for the terminally ill has identified few measurable differences in the care provided to patients and their families. Nonetheless, hospice recipients frequently express a higher level of satisfaction with their program of care. This study compared the ability of hospice and conventional care settings to meet the basic emotional needs of families during a member’s dying and death from cancer. In addition, the relationship of basic needs satisfaction, perceptions about the nurse, and overall satisfaction with the program of care were explored. One hundred bereaved familial care givers completed a mail questionnaire concerning their perceptions of care at the site of a family member’s death. The sites were (1) the home, with care provided by a Medicare certified, community-based hospice program; (2) a hospital affiliated with a Medicare certified, community-based hospice program; (3) a hospital with its own hospice program; and (4) a conventional (non-hospice) hospital. Analyses of quantitative data supported two hypotheses about significant differences between hospice and conventional care. The conventional care group demonstrated the lowest levels of basic needs satisfaction, satisfaction with the psycho-social support of the nurse, and overall program satisfaction. As predicted, overall satisfaction with care was consistent across hospice groups. However, home hospice care provided the highest quality of basic needs satisfaction and the highest level of satisfaction with the nurse. Significant Pearson correlations supported the hypothesis that overall satisfaction is negatively related to unmet basic needs (r = -0.69) and positively related to the psycho-social support received from nurses (r = 0.73). The results of this study emphasize the importance of the nurse in the familial care giver’s feelings of satisfaction with a program of care. In addition, the results indicate that families experiencing a grief reaction are most likely to have basic needs met at home or when inpatient care is within a hospice rather than a conventional hospital. Ke_vwords-hospice,
terminal care, cancer, family satisfaction
few measurable differences in the care provided to patients and their families. Nonetheless, hospice recipients frequently express a higher level of satisfaction with their program of care. Why is hospice care more satisfying? This study attempts to clarify this question by evaluating how well programs of care meet the emotional needs of these familial care givers: those whose terminally ill family members died in the hospital while under the care of (1) a communitybased hospice program without an inpatient unit (Group 2); (2) a hospital-based hospice program without an inpatient unit (Group 3); and (3) a conventional care hospital (Group 4). The perceptions of the three hospital groups have been compared to each other and to those of a sample of families whose terminally ill members died at home under the care of a community-based hospice program (Group 1).
INTRODUCXION
After numerous years in which scholars documented the inadequacies of institutions in meeting the needs of the dying patient [l-4], hospice care has emerged as a humane alternative to the highly technical, depersonalized care charged to the modern hospital. Hospice care differs from conventional care in a number of ways, not the least of which is its emphasis on providing support to the total family. Patterned after a model of care developed in the 1960s at St Christopher’s Hospice in England, the goal of hospice is to meet the physical needs of the dying patient as well as the psycho-social needs of both the patient and his family [S]. Hospice accomplishes this, in part, by making available to families staff who have knowledge of grief and loss theory. Thus, the family is served by sensitive individuals who have the expertise to help. In the past decade, investigators have compared hospice and conventional care on variables such as emotional well-being [6-81, family morbidity [8-lo], activities of daily living and patient symptoms [I I], pain and comfort levels [I I-131, quality of life [12], and number of inpatient days and cost effectiveness [l 1, 12, 14-161. This research has identified
DEFINITIONS
In this study, a familial care giver is the adult identified on agency records as the person having the responsibility for decision making during the patient’s dying and death from cancer. The final hospitalization is the hospitalization during which the patient dies. Hospice programs are those that meet the standards of the National Hospice Organization (1979) [17].
*Presented at the International Conference on Community Nursing.
‘SHertogenbosch, March 1989.
The
Netherlands,
16-17
83
NORMA
84
JEANDAWSOON
A conventional care hospital has no identifiable hospice program or formal hospice affiliation. Inpatient care is directed by each patient’s private physician. HYPOTHESES
The following hypotheses are tested at the P < 0.05 level: 1. Familial care givers who are members of a hospice program during the patient’s final hospitalization will report a smaller degree of unmet need satisfaction and a higher level of satisfaction with hospital care and nursing care than will familial care givers who received conventional hospital care. 2. Regardless of the site of death, there will be no difference in perceptions on measures of satisfaction among the three groups of familial care givers served by hospice programs. 3. As measured by the appropriate scales, there will be a significant, negative correlation between the familial care givers’ overall satisfaction and their level of unmet needs and a positive correlation between overall satisfaction and satisfaction with the nurse. IMETHOD
The agencies providing the samples of bereaved familial care givers consisted of three non-profit agencies: an independent, Medicare certified, community-based hospice program for the terminally ill and two general, private, over 350 beds, teaching hospitals. The hospital with an in-house hospice program provided the services of a hospice team to inpatients. The conventional hospital had no special person available for terminal care support. Because identification of the samples required a search of the deceased patients’ medical records, procedures to address the issues of confidentiality and informed consent varied by institution. In general, the more complicated the institutional requirements, the lower were the return rates. Criteria for sample selection were as follows: 1. The terminally ill patient who died in the hospital had been admitted at least 24 hr prior to death. 2. The terminally ill patient who died at home was served by the hospice program of care for at least 1 week prior to death. 3. The patient’s death occurred at least 1 month prior, but not more than 12 months before, the researcher’s contact. 4. The patients, whose deaths were a result of a tissue cancer, and the familial care givers were 21 years of age or older. All individuals meeting the study’s criteria received a questionnaire by mail Non-respondents received a second packet by certified mail. This protocol achieved a total sample of 100 subjects, representing a return rate of 73% for Group 1 (33 out of 45 subjects); 71% for Group 2 (12 out of 17 subjects); 50% for Group 3 (25 out of 50 subjects); and 39% for Group 4 (30 out of 77 subjects).
The instrument used in this study consisted of a demographic data form and these three 5-point. Likert-type scales: 1. The Need Satisfaction Scale consists of items necessary to the emotional health of the family of a terminally ill patient as identified by Hampe [18] and a review of the literature on loss and grief. The items are the need to be (1) kept informed of the patient’s condition, (2) able to do things to help the patient, (3) able to talk about fears and concerns with the people caring for the patient, (4) aware that the patient was free from pain and comfortable, (5) aware that the staff cared about them and were there to help them, (6) informed when the patient was dying, (7) present when the patient died, (8) able to have the time to say good-bye, and (9) able to talk with the staff after the patient’s death about events surrounding the terminal illness. A tenth need, the need for comfort and support from family members, was excluded from the instrument. Researchers have protested that it is not within the control of hospital care providers [ 191. Responses to the Need Satisfaction Scale were scored from 5 (strongly agree) to 1 (strongly disagree). Because subjects were asked to respond to each need both from the standpoint of how intensely the need was felt (Felt Need Subscale) and how well the program of care met the need (Met Need Subscale), the total Need Satisfaction Scale consisted of 18 items. With this method, it was possible to index an Unmet Need score. For example, if a subject rated the need to be with the patient at death with a 5 and the actual ability to be present at the time of death with a 2, the Unmet Need score was recorded as 3. The highest possible Unmet Need score for a program of care was 36, the minimum possible was 0. Support for content validity was achieved by submitting the scale to a group of 10 bereaved family members and 10 experienced oncology nurses. There was 100% agreement among the two groups that the nine items represented needs of grieving family members, that assistance from a program of care in meeting these needs was a reasonable expectation, and that-with the exception of the aforementioned item concerning support from family members-the list was comprehensive. In addition to content validity, support for criterion-related validity can be inferred by the strength of the significant relationship found in this study between unmet needs and overall satisfaction with the program of care (r = 0.69). Cronbach’s alpha reliabilities were 0.84 for the 18 items of the Need Satisfaction Scale, 0.74 for the Felt Need Subscale, and 0.84 for the Met Need Subscale. 2. The Nurse Satisfaction Scale is designed to measure clients’ satisfaction with the psycho-social support of the nurse. Responses are scored from 5 (strongly agree) to 1 (strongly disagree). Total satisfaction with the nurse is indicated by a score of 55, total dissatisfaction by a score of 1. Development of the scale has already been reported [20]. It includes support for content validity and coefficient alpha reliabilities of 0.89 and 0.81, respectively, in two trials. In this study, a Cronbach’s alpha of 0.96 was obtained.
Need Table
I. One-way
satisfaction
in terminal
Unmet needs Nurse satisfaction Overall satisfaction
G~OUD
2.4 50.7 4.8
1
Group 2
85
settings
ANOVA summary comparing four groups’ providers on mean satisfaction scores Mean
Scale
care
of familial
care
score Grotto
4.5 43.8 4.1
3
6.2 47. I 4.5
G~OUD 4
F value
10.7 38.0 3.6
I I .96** ll.83** 6.13’
Now. Unmet Needs: The higher the score. the less successful is the program of care in satisfying basic needs. Nurse Satisfaction: The higher the score, the higher the level of satisfaction with the nurse. Overall Satisfaction: The higher the overall satisfaction score, the higher is satisfaction with care at the site of death. “Group I =Community-based hospice care. 33 home deaths. Group 2 = Community-based hospice care, I2 hospital deaths. Group 3 = Hospitalbased hosaice care. 25 hosoital deaths. Group 4 = Hospital-based non-hospice care, 30 hospital deaths. lf = 0.001 l*P = 0.0001.
3. The Overall Satisfaction Scale consists of a single statement used to evaluate overall satisfaction with the agency providing care: “If you think about your needs or the needs of your family during the final hospitalization, how satisfied are you with the care and support you received from the hospital?” (The statement was amended appropriately for the home care group.) Responses were scored from 5 (very satisfied) to 1 (very dissatisfied).
and 3 (P < 0.01). In each instance, home hospice care is viewed more favorably than inpatient hospice care. Pearson correlations revealed a significant, negative correlation between overall satisfaction with one’s program of care and unmet need scores (r = -0.69, P = O.OOOl),as well as a positive correlation between overall satisfaction and the level of satisfaction with the psycho-social support of the nurse (r = 0.73, P = 0.0001). DlSCUSSlON
RESULTS
Of the
total sample of familial care givers (N = IOO), most were Caucasian (95%), female (66%), and either the patient’s husband (28%) or wife (38%). The mean age of the familial care giver was 59.5 years. Chi square analyses revealed no significant differences among the four groups in sex, income, education, or relationship to the patient. The results of an analysis of variance (ANOVA) of the four groups’ scores on measures of satisfaction are shown in Table 1. As may be seen, hospice groups score more favorably on all measures compared to the conventional care group. The conventional care group has a greater need satisfaction deficit (F = 11.96, P = O.OOOl), demonstrates the lowest degree of satisfaction with the psycho-social support received from nurses (F = 11.83, P = O.OOOl), and reports the lowest level of overall satisfaction with their program of care (F = 6.13, P = 0.001). The differences in scores are greatest between the home hospice group (Group 1) and the conventional care group (Group 4). Although the ANOVAs reported in Table 1 reveal significant differences among the four groups, ANOVA does not evaluate significant differences between any two groups. In order to test the hypothesis of no differences in the scores between pairs of hospice groups, a series of f-tests were carried out and are presented in summary form in Table 2. As may be seen, while overall satisfaction with their programs of care is comparable, significant differences do exist between hospice groups on some measures. The mean differences (see Table 1) in nurse satisfaction scores between Groups 1 and 2 and between Groups 1 and 3 are significant at the P -z 0.05 level. Also, significant differences are apparent in the extent of unmet needs between Groups 1
Analyses of differences between groups supported the study’s hypothesis that basic needs are more likely to be satisfied within hospitals to a significantly greater extent when inpatient care is supported by a hospice program. In comparison to conventional care, a hospital with an in-house hospice provided a higher quality of psycho-social support from nurses, and its group of familial care givers experienced less incongruity between what they felt they needed from the staff and what was actually provided. Not surprisingly, their overall satisfaction with the hospital care was higher. A question implicit in this study is, “What happens to the community-based hospice’s concept of care when a member is admitted to, and dies, in a hospital without its own in-house hospice team?’ Based on the results of the investigation, the answer is not clear. It was hypothesized that levels of satisfaction in such a group of familial care givers would be higher than satisfaction levels in a group served by a conventional hospital. The prediction was only Table 2. r-Test summary table comparing scores by group ’ Satisfaction Unmet
Group I I I 2 2 3
and and and and and and
2 3 4 3 4 4
needs
satisfaction
scores
Norse
Overall
* * .*
l
l l
. .
l
**
.*
.
l
l
‘Group I = Community-based hospice care, 33 home deaths. Group 2 = Community-based hospice care, I2 hospital deaths. Group 3 = Hospital-based hospice care, 25 hospital deaths. Group 4 = Hospitalbased non-hospice care, 30 hospital deaths. lf < 0.05 l*P < 0.005 l**p < 0.001.
86
NORMA JEA.VDAWOF;
partially supported. Significant differences were found in the higher extent of unmet needs identified by the conventional care group. Although mean scores for the hospice-affiliated group appeared sub-
stantially higher, no significant differences were found between the groups in their levels of satisfaction with the nurses or their overall satisfaction with care. Any conclusions about the care provided to the community-based hospice group who experienced a final hospitalization must be made with caution. To what extent the lack of significant differences was a function of the small sample size must be considered. Since P values depend on sample size, identical mean scores could yield lower significance levels in a larger sample. In the same vein, conclusions based on an analyses of multiple r-tests also must be made with caution [Zl]. One authority suggests using a significance level of 0.01 when multiple t-tests are employed, emphasizing that even this may not control a “galloping” alpha leading to a Type 1 error [22]. If a 0.01 level of probability were employed for this study’s data analyses, any differences observed in the care provided to Group 2 would lack statistical validation of significance. Despite the fact that no significant differences were predicted among hospice groups, some were revealed. The community-based, hospice program’s home care group reported greater satisfaction with nurses than did their group experiencing a final hospitalization in a hospital lacking an in-house hospice team. Similarly, the home care group was more satisfied with nurses and had fewer unmet needs than the group cared for by a hospital with its own hospice program. This difference in unmet needs would remain significant even at an alpha of 0.01. The greatest differences in scores occurred between the home care group and the conventional care group. The home care group revealed the lowest unmet need scores, the greatest satisfaction with the psycho-social support from nurses, and the greatest overall satisfaction with their program of care. These findings are consistent with those of other researchers. Buckingham [6] found greater satisfaction expressed by home hospice patients for some aspects of hospice care compared to a conventional home care group, and Kane and his colleagues [I l] found that two of five groups of familial care givers reported greater satisfaction with their involvement in the patients’ care than did the conventional care group. The findings are also consistent with those of the National Hospice Study whose researchers reported that hospice families receiving home care were more satisfied with the site of death of the patient than were conventional care groups [ 121. A hypothesis of this research stated that overall satisfaction with care would be related negatively to the familial care givers’ level of unmet need satisfaction and positively to their level of satisfaction with the nurse. This hypothesis was supported. A substantial, significant, negative relationship was found between overall satisfaction and unmet needs, and a substantial, significant, positive relationship was found between overall satisfaction and the level of satisfaction with the nurse. Not surprisingly, the
results indicate that the greater the gap between what familial care givers feel they need and what the program and the program’s nurses offer, the lower is the familial care givers’ satisfaction with the program of care. At the outset of this research, when the study was as yet an ill-defined idea, the fact that hospice care often was evaluated by recipients as more satisfying than conventional care was intriguing. What made it so was the fact that there was little data to explain why hospice care should be viewed more favorably. This study approached the evaluation of programs of care from the perspective that families coping with a member’s terminal illness have basic emotional needs, the satisfaction of which may influence feelings of satisfaction with the program of care. This thesis has been supported by the study’s results. It would not be inappropriate to wonder if satisfaction in the home was higher because the home setting itself was more satisfying. However, nothing in the analyses of the study’s data supported this possibility. The data evaluation indicated that a program of care’s intervention, particularly psycho-social support by the nurse, played a strong role in overall satisfaction. The home hospice’s staff simply provided a consistently higher quality of basic needs support to familial care givers. The findings of this study suggest that public policy should encourage the availability of hospice personnel in terminal care settings. It should be noted, though, that the inpatient settings of this study did not provide as high a level of basic needs satisfaction as the home, even when the hospital had its own hospice program. The extent to which this was a result of a policy of scattered bed admissions throughout the hospitals (in contrast to admission to a single unit) needs further investigation. Although this study did not evaluate care in a hospital with its own hospice inpatient unit, it is possible that a hospice unit would provide the ideal milieu for the family of the patient dying within an institution. Parkes [23], in fact, has suggested that the ideal program may be one that offers a combination of home care and inpatient care in a hospice unit. However, the presence of a hospice unit is not without cost. It is desirable because it may be able to protect families from added burdens to their grief. It is undesirable because of the added costs incurred through increased utilization of inpatient beds where such a unit exists [23].
REFERENCES I.Charmaz K. The Social Reaky of Death: Dealh in Contemporary America. Addison-Wesley, Reading, 1980. 2. Kubler-Ross E. On Death and Dying. Macmillan. New York, 1969. 3. Sudnow D. Passing On. Prentice Hall, Englewood. NJ, 1967. 4. Vachon M., Freedman K., Formo A.. Rogers J.. Lyall W. and Freeman S. The final illness is cancer: The widow’s perspective. Can. Med. Assoc. J. 117, 1151-1154, 1977.
Need satisfaction in terminal care settings 5. Stoddard S. The Hospice Mocemenl: A Better Way of Dying. Stein and Day, Briarcliff Manor. NY. 1978.
6. Buckingham R. Objectives and methodology of the study. In First American Hospice: Three Years of Home Care (Edited by Lack S. and Buckingham R.), pp. 211-264. Hospice Inc., New Haven, 1978. 7. Godkin A., Krant M. and Doster N. The impact of hospice care on families. Inr. 1. Psychiar. Med. 13, 153-165, 1983-84. 8. Parkes C. and Parkes J. “Hospice” versus “hospital” care: Reevaluation after 10 years as seen by surviving spouses. Postgrad. Med. J. 60, 120-124, 1984. 9. Cameron J. and Parkes C. Terminal care: Evaluation of effects on surviving family of care before and after bereavement. Postgrad. med. J. 59, 73-78, 1983. IO. Parkes C. Terminal care: Evaluation of inpatient service at St Christopher’s Hospice. Part 2. Self-assessments of effects of the service on surviving spouses. Postgrad. med. J. 55, 517-522,
1979.
II. Kane R., Bernstein L., Wales J., Leibowitz A. and Kaplan S. A randomised controlled trial of hospice care. LANCA 1, 890-894, 1982. 12. Greer D., Mor V., Morris J., Sherwood S., Kidder D. and Bimbaum H. An alternative in terminal care: Results of the National Hospice Study. J. chron. Dis. 39, 9-26.
1986.
13. Parkes C. Terminal care: Evaluation of inpatient service at St Christopher’s Hospice. Part 1. Views of surviving spouses on effects of the service on the patient. Postgrad. med. J. 55, 517-522,
1979.
87
14. Amado A., Cronk B. and Milio R. Cost of terminal care: Home hospice vs hospital. Nurs. Ourlook 27, 522-526. 1979. 15. Van Buren L. Hospice home care: A cost analysis of a sample of patients seen during 1980. Home Hlrh Rec. 2, 10-13. 1980.
16. Haid M., Fowler M., Nicklin 0. and LeToumeau E. People and dollars: The experience of one hospice. Sot. Med. J. 77, 47@472.
1984.
17. National Hospice Organization. Standards of a Hospice Program of Care (6th rev.). National Hospice Organization, McLean, VA, 1979. 18. Hampe S. Needs of the grieving spouse in a hospital setting. Nurs. Res. 24, 113-120, 1975. 19. Dracup K. and Breu C. Using nursing research findings to meet the needs of grieving spouses. Nurs. Res. 27, 212-216, 1978. 20. Risser H. Development of an instrument to measure patient satisfaction with nurses and nursing care in primary care settings. Nurs. Res. 24, 45-52, 1975. 21. Kennedy J. and Bush A. An Inrroducrion IO the Design and Analysis of Experiments in Educarion and Psychology. University Press of America, New York, 1985.
22. Reid B. Potential sources of type I error and possible solutions to avoid a “galloping” alpha rate. Nurs. Res. 32, 190-191, 1983. 23. Parkes C. Not always! J. Palliat. Care 4, 50-52. 1988. 24. Wilson B., Blosse R., Tucker J. and Spector K. Hospice care: Perspectives on a Blue Cross plan’s community pilot program. Inquiry 20, 322-327, 1983.
NEED
0277-9536,91 S3.00 + 0.00 CopytIght S 1991 Pergamon Press plc
SATISFACTION
IN TERMINAL NORMA
CARE SETTINGS*
JEAN DAWS~N
School of Nursing of Widener University, Chester,
PA
19013, U.S.A.
Abstract-Research comparing hospice and conventional programs of care for the terminally ill has identified few measurable differences in the care provided to patients and their families. Nonetheless, hospice recipients frequently express a higher level of satisfaction with their program of care. This study compared the ability of hospice and conventional care settings to meet the basic emotional needs of families during a member’s dying and death from cancer. In addition, the relationship of basic needs satisfaction, perceptions about the nurse, and overall satisfaction with the program of care were explored. One hundred bereaved familial care givers completed a mail questionnaire concerning their perceptions of care at the site of a family member’s death. The sites were (1) the home, with care provided by a Medicare certified, community-based hospice program; (2) a hospital affiliated with a Medicare certified, community-based hospice program; (3) a hospital with its own hospice program; and (4) a conventional (non-hospice) hospital. Analyses of quantitative data supported two hypotheses about significant differences between hospice and conventional care. The conventional care group demonstrated the lowest levels of basic needs satisfaction, satisfaction with the psycho-social support of the nurse, and overall program satisfaction. As predicted, overall satisfaction with care was consistent across hospice groups. However, home hospice care provided the highest quality of basic needs satisfaction and the highest level of satisfaction with the nurse. Significant Pearson correlations supported the hypothesis that overall satisfaction is negatively related to unmet basic needs (r = -0.69) and positively related to the psycho-social support received from nurses (r = 0.73). The results of this study emphasize the importance of the nurse in the familial care giver’s feelings of satisfaction with a program of care. In addition, the results indicate that families experiencing a grief reaction are most likely to have basic needs met at home or when inpatient care is within a hospice rather than a conventional hospital. Ke_vwords-hospice,
terminal care, cancer, family satisfaction
few measurable differences in the care provided to patients and their families. Nonetheless, hospice recipients frequently express a higher level of satisfaction with their program of care. Why is hospice care more satisfying? This study attempts to clarify this question by evaluating how well programs of care meet the emotional needs of these familial care givers: those whose terminally ill family members died in the hospital while under the care of (1) a communitybased hospice program without an inpatient unit (Group 2); (2) a hospital-based hospice program without an inpatient unit (Group 3); and (3) a conventional care hospital (Group 4). The perceptions of the three hospital groups have been compared to each other and to those of a sample of families whose terminally ill members died at home under the care of a community-based hospice program (Group 1).
INTRODUCXION
After numerous years in which scholars documented the inadequacies of institutions in meeting the needs of the dying patient [l-4], hospice care has emerged as a humane alternative to the highly technical, depersonalized care charged to the modern hospital. Hospice care differs from conventional care in a number of ways, not the least of which is its emphasis on providing support to the total family. Patterned after a model of care developed in the 1960s at St Christopher’s Hospice in England, the goal of hospice is to meet the physical needs of the dying patient as well as the psycho-social needs of both the patient and his family [S]. Hospice accomplishes this, in part, by making available to families staff who have knowledge of grief and loss theory. Thus, the family is served by sensitive individuals who have the expertise to help. In the past decade, investigators have compared hospice and conventional care on variables such as emotional well-being [6-81, family morbidity [8-lo], activities of daily living and patient symptoms [I I], pain and comfort levels [I I-131, quality of life [12], and number of inpatient days and cost effectiveness [l 1, 12, 14-161. This research has identified
DEFINITIONS
In this study, a familial care giver is the adult identified on agency records as the person having the responsibility for decision making during the patient’s dying and death from cancer. The final hospitalization is the hospitalization during which the patient dies. Hospice programs are those that meet the standards of the National Hospice Organization (1979) [17].
*Presented at the International Conference on Community Nursing.
‘SHertogenbosch, March 1989.
The
Netherlands,
16-17
83
NORMA
84
JEANDAWSOON
A conventional care hospital has no identifiable hospice program or formal hospice affiliation. Inpatient care is directed by each patient’s private physician. HYPOTHESES
The following hypotheses are tested at the P < 0.05 level: 1. Familial care givers who are members of a hospice program during the patient’s final hospitalization will report a smaller degree of unmet need satisfaction and a higher level of satisfaction with hospital care and nursing care than will familial care givers who received conventional hospital care. 2. Regardless of the site of death, there will be no difference in perceptions on measures of satisfaction among the three groups of familial care givers served by hospice programs. 3. As measured by the appropriate scales, there will be a significant, negative correlation between the familial care givers’ overall satisfaction and their level of unmet needs and a positive correlation between overall satisfaction and satisfaction with the nurse. IMETHOD
The agencies providing the samples of bereaved familial care givers consisted of three non-profit agencies: an independent, Medicare certified, community-based hospice program for the terminally ill and two general, private, over 350 beds, teaching hospitals. The hospital with an in-house hospice program provided the services of a hospice team to inpatients. The conventional hospital had no special person available for terminal care support. Because identification of the samples required a search of the deceased patients’ medical records, procedures to address the issues of confidentiality and informed consent varied by institution. In general, the more complicated the institutional requirements, the lower were the return rates. Criteria for sample selection were as follows: 1. The terminally ill patient who died in the hospital had been admitted at least 24 hr prior to death. 2. The terminally ill patient who died at home was served by the hospice program of care for at least 1 week prior to death. 3. The patient’s death occurred at least 1 month prior, but not more than 12 months before, the researcher’s contact. 4. The patients, whose deaths were a result of a tissue cancer, and the familial care givers were 21 years of age or older. All individuals meeting the study’s criteria received a questionnaire by mail Non-respondents received a second packet by certified mail. This protocol achieved a total sample of 100 subjects, representing a return rate of 73% for Group 1 (33 out of 45 subjects); 71% for Group 2 (12 out of 17 subjects); 50% for Group 3 (25 out of 50 subjects); and 39% for Group 4 (30 out of 77 subjects).
The instrument used in this study consisted of a demographic data form and these three 5-point. Likert-type scales: 1. The Need Satisfaction Scale consists of items necessary to the emotional health of the family of a terminally ill patient as identified by Hampe [18] and a review of the literature on loss and grief. The items are the need to be (1) kept informed of the patient’s condition, (2) able to do things to help the patient, (3) able to talk about fears and concerns with the people caring for the patient, (4) aware that the patient was free from pain and comfortable, (5) aware that the staff cared about them and were there to help them, (6) informed when the patient was dying, (7) present when the patient died, (8) able to have the time to say good-bye, and (9) able to talk with the staff after the patient’s death about events surrounding the terminal illness. A tenth need, the need for comfort and support from family members, was excluded from the instrument. Researchers have protested that it is not within the control of hospital care providers [ 191. Responses to the Need Satisfaction Scale were scored from 5 (strongly agree) to 1 (strongly disagree). Because subjects were asked to respond to each need both from the standpoint of how intensely the need was felt (Felt Need Subscale) and how well the program of care met the need (Met Need Subscale), the total Need Satisfaction Scale consisted of 18 items. With this method, it was possible to index an Unmet Need score. For example, if a subject rated the need to be with the patient at death with a 5 and the actual ability to be present at the time of death with a 2, the Unmet Need score was recorded as 3. The highest possible Unmet Need score for a program of care was 36, the minimum possible was 0. Support for content validity was achieved by submitting the scale to a group of 10 bereaved family members and 10 experienced oncology nurses. There was 100% agreement among the two groups that the nine items represented needs of grieving family members, that assistance from a program of care in meeting these needs was a reasonable expectation, and that-with the exception of the aforementioned item concerning support from family members-the list was comprehensive. In addition to content validity, support for criterion-related validity can be inferred by the strength of the significant relationship found in this study between unmet needs and overall satisfaction with the program of care (r = 0.69). Cronbach’s alpha reliabilities were 0.84 for the 18 items of the Need Satisfaction Scale, 0.74 for the Felt Need Subscale, and 0.84 for the Met Need Subscale. 2. The Nurse Satisfaction Scale is designed to measure clients’ satisfaction with the psycho-social support of the nurse. Responses are scored from 5 (strongly agree) to 1 (strongly disagree). Total satisfaction with the nurse is indicated by a score of 55, total dissatisfaction by a score of 1. Development of the scale has already been reported [20]. It includes support for content validity and coefficient alpha reliabilities of 0.89 and 0.81, respectively, in two trials. In this study, a Cronbach’s alpha of 0.96 was obtained.
Need Table
I. One-way
satisfaction
in terminal
Unmet needs Nurse satisfaction Overall satisfaction
G~OUD
2.4 50.7 4.8
1
Group 2
85
settings
ANOVA summary comparing four groups’ providers on mean satisfaction scores Mean
Scale
care
of familial
care
score Grotto
4.5 43.8 4.1
3
6.2 47. I 4.5
G~OUD 4
F value
10.7 38.0 3.6
I I .96** ll.83** 6.13’
Now. Unmet Needs: The higher the score. the less successful is the program of care in satisfying basic needs. Nurse Satisfaction: The higher the score, the higher the level of satisfaction with the nurse. Overall Satisfaction: The higher the overall satisfaction score, the higher is satisfaction with care at the site of death. “Group I =Community-based hospice care. 33 home deaths. Group 2 = Community-based hospice care, I2 hospital deaths. Group 3 = Hospitalbased hosaice care. 25 hosoital deaths. Group 4 = Hospital-based non-hospice care, 30 hospital deaths. lf = 0.001 l*P = 0.0001.
3. The Overall Satisfaction Scale consists of a single statement used to evaluate overall satisfaction with the agency providing care: “If you think about your needs or the needs of your family during the final hospitalization, how satisfied are you with the care and support you received from the hospital?” (The statement was amended appropriately for the home care group.) Responses were scored from 5 (very satisfied) to 1 (very dissatisfied).
and 3 (P < 0.01). In each instance, home hospice care is viewed more favorably than inpatient hospice care. Pearson correlations revealed a significant, negative correlation between overall satisfaction with one’s program of care and unmet need scores (r = -0.69, P = O.OOOl),as well as a positive correlation between overall satisfaction and the level of satisfaction with the psycho-social support of the nurse (r = 0.73, P = 0.0001). DlSCUSSlON
RESULTS
Of the
total sample of familial care givers (N = IOO), most were Caucasian (95%), female (66%), and either the patient’s husband (28%) or wife (38%). The mean age of the familial care giver was 59.5 years. Chi square analyses revealed no significant differences among the four groups in sex, income, education, or relationship to the patient. The results of an analysis of variance (ANOVA) of the four groups’ scores on measures of satisfaction are shown in Table 1. As may be seen, hospice groups score more favorably on all measures compared to the conventional care group. The conventional care group has a greater need satisfaction deficit (F = 11.96, P = O.OOOl), demonstrates the lowest degree of satisfaction with the psycho-social support received from nurses (F = 11.83, P = O.OOOl), and reports the lowest level of overall satisfaction with their program of care (F = 6.13, P = 0.001). The differences in scores are greatest between the home hospice group (Group 1) and the conventional care group (Group 4). Although the ANOVAs reported in Table 1 reveal significant differences among the four groups, ANOVA does not evaluate significant differences between any two groups. In order to test the hypothesis of no differences in the scores between pairs of hospice groups, a series of f-tests were carried out and are presented in summary form in Table 2. As may be seen, while overall satisfaction with their programs of care is comparable, significant differences do exist between hospice groups on some measures. The mean differences (see Table 1) in nurse satisfaction scores between Groups 1 and 2 and between Groups 1 and 3 are significant at the P -z 0.05 level. Also, significant differences are apparent in the extent of unmet needs between Groups 1
Analyses of differences between groups supported the study’s hypothesis that basic needs are more likely to be satisfied within hospitals to a significantly greater extent when inpatient care is supported by a hospice program. In comparison to conventional care, a hospital with an in-house hospice provided a higher quality of psycho-social support from nurses, and its group of familial care givers experienced less incongruity between what they felt they needed from the staff and what was actually provided. Not surprisingly, their overall satisfaction with the hospital care was higher. A question implicit in this study is, “What happens to the community-based hospice’s concept of care when a member is admitted to, and dies, in a hospital without its own in-house hospice team?’ Based on the results of the investigation, the answer is not clear. It was hypothesized that levels of satisfaction in such a group of familial care givers would be higher than satisfaction levels in a group served by a conventional hospital. The prediction was only Table 2. r-Test summary table comparing scores by group ’ Satisfaction Unmet
Group I I I 2 2 3
and and and and and and
2 3 4 3 4 4
needs
satisfaction
scores
Norse
Overall
* * .*
l
l l
. .
l
**
.*
.
l
l
‘Group I = Community-based hospice care, 33 home deaths. Group 2 = Community-based hospice care, I2 hospital deaths. Group 3 = Hospital-based hospice care, 25 hospital deaths. Group 4 = Hospitalbased non-hospice care, 30 hospital deaths. lf < 0.05 l*P < 0.005 l**p < 0.001.
86
NORMA JEA.VDAWOF;
partially supported. Significant differences were found in the higher extent of unmet needs identified by the conventional care group. Although mean scores for the hospice-affiliated group appeared sub-
stantially higher, no significant differences were found between the groups in their levels of satisfaction with the nurses or their overall satisfaction with care. Any conclusions about the care provided to the community-based hospice group who experienced a final hospitalization must be made with caution. To what extent the lack of significant differences was a function of the small sample size must be considered. Since P values depend on sample size, identical mean scores could yield lower significance levels in a larger sample. In the same vein, conclusions based on an analyses of multiple r-tests also must be made with caution [Zl]. One authority suggests using a significance level of 0.01 when multiple t-tests are employed, emphasizing that even this may not control a “galloping” alpha leading to a Type 1 error [22]. If a 0.01 level of probability were employed for this study’s data analyses, any differences observed in the care provided to Group 2 would lack statistical validation of significance. Despite the fact that no significant differences were predicted among hospice groups, some were revealed. The community-based, hospice program’s home care group reported greater satisfaction with nurses than did their group experiencing a final hospitalization in a hospital lacking an in-house hospice team. Similarly, the home care group was more satisfied with nurses and had fewer unmet needs than the group cared for by a hospital with its own hospice program. This difference in unmet needs would remain significant even at an alpha of 0.01. The greatest differences in scores occurred between the home care group and the conventional care group. The home care group revealed the lowest unmet need scores, the greatest satisfaction with the psycho-social support from nurses, and the greatest overall satisfaction with their program of care. These findings are consistent with those of other researchers. Buckingham [6] found greater satisfaction expressed by home hospice patients for some aspects of hospice care compared to a conventional home care group, and Kane and his colleagues [I l] found that two of five groups of familial care givers reported greater satisfaction with their involvement in the patients’ care than did the conventional care group. The findings are also consistent with those of the National Hospice Study whose researchers reported that hospice families receiving home care were more satisfied with the site of death of the patient than were conventional care groups [ 121. A hypothesis of this research stated that overall satisfaction with care would be related negatively to the familial care givers’ level of unmet need satisfaction and positively to their level of satisfaction with the nurse. This hypothesis was supported. A substantial, significant, negative relationship was found between overall satisfaction and unmet needs, and a substantial, significant, positive relationship was found between overall satisfaction and the level of satisfaction with the nurse. Not surprisingly, the
results indicate that the greater the gap between what familial care givers feel they need and what the program and the program’s nurses offer, the lower is the familial care givers’ satisfaction with the program of care. At the outset of this research, when the study was as yet an ill-defined idea, the fact that hospice care often was evaluated by recipients as more satisfying than conventional care was intriguing. What made it so was the fact that there was little data to explain why hospice care should be viewed more favorably. This study approached the evaluation of programs of care from the perspective that families coping with a member’s terminal illness have basic emotional needs, the satisfaction of which may influence feelings of satisfaction with the program of care. This thesis has been supported by the study’s results. It would not be inappropriate to wonder if satisfaction in the home was higher because the home setting itself was more satisfying. However, nothing in the analyses of the study’s data supported this possibility. The data evaluation indicated that a program of care’s intervention, particularly psycho-social support by the nurse, played a strong role in overall satisfaction. The home hospice’s staff simply provided a consistently higher quality of basic needs support to familial care givers. The findings of this study suggest that public policy should encourage the availability of hospice personnel in terminal care settings. It should be noted, though, that the inpatient settings of this study did not provide as high a level of basic needs satisfaction as the home, even when the hospital had its own hospice program. The extent to which this was a result of a policy of scattered bed admissions throughout the hospitals (in contrast to admission to a single unit) needs further investigation. Although this study did not evaluate care in a hospital with its own hospice inpatient unit, it is possible that a hospice unit would provide the ideal milieu for the family of the patient dying within an institution. Parkes [23], in fact, has suggested that the ideal program may be one that offers a combination of home care and inpatient care in a hospice unit. However, the presence of a hospice unit is not without cost. It is desirable because it may be able to protect families from added burdens to their grief. It is undesirable because of the added costs incurred through increased utilization of inpatient beds where such a unit exists [23].
REFERENCES I.Charmaz K. The Social Reaky of Death: Dealh in Contemporary America. Addison-Wesley, Reading, 1980. 2. Kubler-Ross E. On Death and Dying. Macmillan. New York, 1969. 3. Sudnow D. Passing On. Prentice Hall, Englewood. NJ, 1967. 4. Vachon M., Freedman K., Formo A.. Rogers J.. Lyall W. and Freeman S. The final illness is cancer: The widow’s perspective. Can. Med. Assoc. J. 117, 1151-1154, 1977.
Need satisfaction in terminal care settings 5. Stoddard S. The Hospice Mocemenl: A Better Way of Dying. Stein and Day, Briarcliff Manor. NY. 1978.
6. Buckingham R. Objectives and methodology of the study. In First American Hospice: Three Years of Home Care (Edited by Lack S. and Buckingham R.), pp. 211-264. Hospice Inc., New Haven, 1978. 7. Godkin A., Krant M. and Doster N. The impact of hospice care on families. Inr. 1. Psychiar. Med. 13, 153-165, 1983-84. 8. Parkes C. and Parkes J. “Hospice” versus “hospital” care: Reevaluation after 10 years as seen by surviving spouses. Postgrad. Med. J. 60, 120-124, 1984. 9. Cameron J. and Parkes C. Terminal care: Evaluation of effects on surviving family of care before and after bereavement. Postgrad. med. J. 59, 73-78, 1983. IO. Parkes C. Terminal care: Evaluation of inpatient service at St Christopher’s Hospice. Part 2. Self-assessments of effects of the service on surviving spouses. Postgrad. med. J. 55, 517-522,
1979.
II. Kane R., Bernstein L., Wales J., Leibowitz A. and Kaplan S. A randomised controlled trial of hospice care. LANCA 1, 890-894, 1982. 12. Greer D., Mor V., Morris J., Sherwood S., Kidder D. and Bimbaum H. An alternative in terminal care: Results of the National Hospice Study. J. chron. Dis. 39, 9-26.
1986.
13. Parkes C. Terminal care: Evaluation of inpatient service at St Christopher’s Hospice. Part 1. Views of surviving spouses on effects of the service on the patient. Postgrad. med. J. 55, 517-522,
1979.
87
14. Amado A., Cronk B. and Milio R. Cost of terminal care: Home hospice vs hospital. Nurs. Ourlook 27, 522-526. 1979. 15. Van Buren L. Hospice home care: A cost analysis of a sample of patients seen during 1980. Home Hlrh Rec. 2, 10-13. 1980.
16. Haid M., Fowler M., Nicklin 0. and LeToumeau E. People and dollars: The experience of one hospice. Sot. Med. J. 77, 47@472.
1984.
17. National Hospice Organization. Standards of a Hospice Program of Care (6th rev.). National Hospice Organization, McLean, VA, 1979. 18. Hampe S. Needs of the grieving spouse in a hospital setting. Nurs. Res. 24, 113-120, 1975. 19. Dracup K. and Breu C. Using nursing research findings to meet the needs of grieving spouses. Nurs. Res. 27, 212-216, 1978. 20. Risser H. Development of an instrument to measure patient satisfaction with nurses and nursing care in primary care settings. Nurs. Res. 24, 45-52, 1975. 21. Kennedy J. and Bush A. An Inrroducrion IO the Design and Analysis of Experiments in Educarion and Psychology. University Press of America, New York, 1985.
22. Reid B. Potential sources of type I error and possible solutions to avoid a “galloping” alpha rate. Nurs. Res. 32, 190-191, 1983. 23. Parkes C. Not always! J. Palliat. Care 4, 50-52. 1988. 24. Wilson B., Blosse R., Tucker J. and Spector K. Hospice care: Perspectives on a Blue Cross plan’s community pilot program. Inquiry 20, 322-327, 1983.
