The Hospice Journal
ISSN: 0742-969X (Print) (Online) Journal homepage: http://www.tandfonline.com/loi/ippc19
Estimation of Survival Time in Terminal Cancer Patients: An Impedance to Hospice Admissions? Ronald S. Schonwetter, Thomas A. Teasdale, Porter Storey & Robert J. Luchi To cite this article: Ronald S. Schonwetter, Thomas A. Teasdale, Porter Storey & Robert J. Luchi (1990) Estimation of Survival Time in Terminal Cancer Patients: An Impedance to Hospice Admissions?, The Hospice Journal, 6:4, 65-79, DOI: 10.1080/0742-969X.1990.11882684 To link to this article: https://doi.org/10.1080/0742-969X.1990.11882684
Published online: 13 Sep 2017.
Submit your article to this journal
Article views: 1
View related articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=ippc19 Download by: [University of New England]
Date: 08 December 2017, At: 20:16
Downloaded by [University of New England] at 20:16 08 December 2017
Estimation of Survival Time in Terminal Cancer Patients: An Impedance to Hospice Admissions? Ronald S. Schonwetter Thomas A. Teasdale Porter Storey Robert J. Luchi
ABSTRACT. Accurate estimation of survival time in terminal can cer patients is difficult yet may provide useful information. A histor ical prospective study on 172 patients admitted to a home based hospice service was performed to determine which variables were best correlated with survival time. Mean and median survival were 48 and 22 days, respectively, representing a highly skewed distribu tion of life span in this sample. As age increased, survival time decreased. All Activities of Daily Living (ADLs) recorded (Bathing, Continence, Dressing and Transfer) as well as other measures of performance (mobility and pulse) and nutrition (appetite and nour ishment) were each strongly associated with survival. Multivariate Ronald S. Schonwetter, MD, is affiliated with the Division of Geriatrics, De partment of Internal Medicine, University of South Florida College of Medicine, Hospice of Hillsborough, Inc., Tampa, FL. He is also affiliated with the Huf fington Center on Aging, Baylor College of Medicine and the Veterans Affairs Medical Center, Houston, TX. Thomas A. Teasdale, MPH, is affiliated with the Huffington Center on Aging, Baylor College of Medicine and the Veterans Af fairs Medical Center, Houston, TX. Porter Storey, MD, is affiliated with the New Age Hospice, Houston, TX. Robert J. Luchi, MD, is affiliated with the Huf fington Center on Aging, Baylor College of Medicine and the Veterans Affairs Medical Center, Houston, TX. Requests for reprints should be sent to Ronald S. Schonwetter, MD, Univer sity of South Florida College of Medicine, 12901 Bruce B. Downs Boulevard, Box 19, Tampa, FL 33612. An earlier version of this paper was presented at the 14th International Con gress of Gerontology, Acapulco, Mexico, June, 1989. The Hospice Journal, Vol. 6(4) 1990 © 1991 by The Haworth Press, Inc. All rights reserved.
65
Downloaded by [University of New England] at 20:16 08 December 2017
66
THE HOSPICE JOURNAL
analysis limited significant variables to dressing ability, pulse rate, level of appetite and transferring ability. Outliers (survival > 180 days) were differentiated from the remainder of the sample by sig nificant differences in all ADLs recorded as well as the level of appetite. These findings establish the importance of assessing ADLs, a measure of functional status, and reinforce the importance of performance and nutrition measures when estimating length of survival in terminal cancer patients.
To estimate the life expectancy for a patient with terminal cancer is a difficult yet important task for medical personnel. An accurate estimation allows appropriate admissions to a hospice, for services that might improve the patient's quality of dying.1 Reasonable as sessment of survival time provides patients and relatives a time frame to plan a strategy for emotionally and financially managing the terminal period. Additionally, it provides staff with an estimate of the time they have to develop their patient's insight into the dying process. There have been relatively few studies focusing on estimating survival in terminally ill cancer patients. One recent study2 com pared a terminal care service staff's estimate of survival in tenninal cancer patients, a functional rating produced from the Kamofsky Performance Scale (KPS) 3 and the patient's actual survival. This study reported that the KPS better correlated with actual survival than subjective estimates, which were overly optimistic. Reuben et al.,4 recently reported that the KPS, a measure of functional status, was the most important clinical factor in estimating survival time, but five other clinical symptoms had independent predictive value as well. Unfortunately, most patients with terminal cancer cluster around the lower end of the KPS, reducing the magnitude of varia tion in the scale and impairing discriminative potential. Three addi tional studies5• • reported physicians to be overly optimistic in terms of subjective predictions of survival time in hospice patients. These data suggest that physicians' estimates of length of survival in pa tients terminally ill from cancer should be interpreted very cau tiously and that other objective measures involving functional status evaluations are needed to better prognosticate. Medical staff, including physicians, may be uncomfortable and reluctant to discuss a patient's prognosis with the patient and fam ily, possibly due to the limited prognostic accuracy. Additionally, 6 7
Downloaded by [University of New England] at 20:16 08 December 2017
Research and Evaluation
67
doctors may believe the patient may blame the doctor personally for the poor outcome. It is not uncommon for staff to fear a poor emo tional reaction by the patient, but typically discover that the patient is relieved with the information. The staff or physician may also fear their personal reaction, which may be emotional when involv ing a close patient. Until the recent implementation of The Catastrophic Illness cov erage, Medicare reimbursed hospice services for a maximum of 210 days, and only for patients who were certified by a physician to have a terminal disease with a life expectancy of less than 6 months. As of January 1, 1989, there are no Medicare-specific limitations concerning length of survival for reimbursement. However, certifi cation of terminal illness is still required by a physician and the maximum amount reimbursed for hospice services is unchanged. Additionally, several private insurance policies place monetary lim itations on hospice service reimbursements. For these reasons, more precise estimates of survival time are needed in order for hos pice services to be cost-effective and to avoid utilizing the entire benefit prior to the patient's expiration. The purpose of the present study was to determine whether certain objective variables corre lated with survival time and whether they may be used to better predict length of survival in terminal cancer patients. METHODS
The New Age Hospice, founded in 1983, is a non-profit organi zation committed to providing physical, emotional and spiritual support to the terminally ill and their families. Their focus of care is in the patient's home, although there is an inpatient unit should that type of care be necessary. To qualify for admission to this hospice program, patients have expected survivals in terms of weeks or months as estimated by their referring physician. Three hundred thirty two medical charts were randomly selected from the charts of all patients admitted between March, 1986 and January, 1988 (n = 997). Data were collected only on patients admitted to the home-based community hospice program (n = 202). Patients were excluded if they had a terminal disease other than cancer (n = 18) or were discharged from the service for rea-
Downloaded by [University of New England] at 20:16 08 December 2017
68
THE HOSPICE JOURNAL
sons other than death (n = 12). One hundred seventy two patients were ultimately included in the present study. All study patients were initially seen and evaluated by The New Age Hospice physician and a registered nurse. Demographic data including age, sex, and marital status were recorded by study inves tigators. The following information was obtained from the physi cian's initial history and physical examination: patient acknowledg ment of terminal disease (yes or no); number of pre-admission medications; systolic and diastolic blood pressures; pulse rate; as signment of mobility (1 = normal to 5 = unable to turn in bed), appetite (1 = normal to 5 = refuses fluid), nourishment (1 = heavy to 5 = extreme cachexia) and pain (1 = none to 6 = over whelming). ADLs included transfer, bathing, dressing and conti nence only, and were evaluated by nurses on the initial visit and recorded as dependent or independent, according to the original definitions by Katz. 8 Length of survival was calculated from initial hospice visit to time of death. Statistical analysis was performed on the overall sample as well as on defined subgroups. Survival time was grouped in 30 day peri ods for chi square tests on categorical data. T Tests were used to determine significance levels of continuous independent variables. An analysis of variance was completed to eliminate possible over lap in the variables used in regression analysis. A forward stepwise regression analysis was completed to calculate correlation coeffi cients and to identify variables which would best predict survival time in these patients. An additional regression analysis was com pleted with outliers (survival > 180 days; n = 9) eliminated to determine whether this would increase the predictive value of the variables. Summary values for outliers and the remainder of the population were tested for potential differences. RESULTS
Characteristics of the study population are shown in Table 1. Fifty six percent of the sample were males and forty four percent were females. The mean age of the patients was 63 years, with 85% of the population 50 years of age and older. Slightly less than one half of the patients were married.
Research and Evaluation
69
Table l Characteristics of study population (n=l72)
Downloaded by [University of New England] at 20:16 08 December 2017
AGE
GENDER
< 25 .. • ...... • .. • • • • ... • 2% 25 - 49
13%
50 - 74
57%
> 74
28%
Male ................... 56% Female ................. 44%
MARITAL STATUS
Married ................ 48% Not married ............ 52%
Overall mean survival was 48 days while the median survival was 22 days. Only 35% of the patients lived beyond seven weeks (approximate mean survival). The distribution of life span is highly skewed (Figure 1) with the majority of these terminal cancer pa tients surviving only three weeks. Table 2 shows survival in days by significant variables collected as categorical data. As age increased, survival time decreased sig nificantly. Survival time decreased in patients with extremely high or low pulse rates. Performance (mobility) and nutrition (appetite and nourishment) were associated with shorter survival in patients who were less mobile, ate less, and were physically thinner. Inde pendence in each of the ADLs recorded (bathing, continence, dress ing and transfer) was strongly associated with increased survival in terminal cancer patients. Table 3 shows the correlation coefficients and significance levels of each variable with survival time. All ADLs recorded were noted to have p values of < 0.001. Those variables not significantly associated with survival were gender,
THE HOSPICE JOURNAL
70
Figure 1 PROPORTIONAL SURVIVAL CURVE N=172
Downloaded by [University of New England] at 20:16 08 December 2017
100
@
40
20: I I
o,______._____________ IJ 100 1:x:J
200
SURVIAL IN DAYS
marital status, number of pre-admission medications, systolic and diastolic blood pressures and level of pain. A stepwise regression analysis (n = 172) limited significant vari ables to dressing (p < 0.001), pulse (p = 0.01), appetite (p = 0.01) and transfer (p = 0.02). These four variables accounted for
Downloaded by [University of New England] at 20:16 08 December 2017
Research and Evaluation
71
30% (r2 = 0.30) of the variation in survival time. A second step wise regression analysis included 163 subjects and excluded the outliers (n = 9). No decrease in variation of survival time (r2 = 0.29) was shown. Figure 2 represents the predicted and observed survival time in days using the initial regression analysis. It appears that many of the patients who died under 7 weeks had overestima tions in their predicted length of survival while many patients who died after 7 weeks had underestimations of their predicted life span. Comparison of the outliers and the remainder of the population resulted in differences in each AOL included ((bathing (p < 0.001); dressing (p < 0.001 ); continence (p = 0.008); and transfer (p = 0.024)) as well as the appetite level (p = 0.038). No additional variables differentiated the outliers from the remainder of the study sample. DISCUSSION
Median survival was just over 3 weeks for patients referred to this outpatient hospice service and thus, physicians may have re ferred patients too late to achieve full benefits of these services. Median survival in an in-patient hospice, typically for patients that are more ill and closer to death, was 24.5 days5 and previous data from The National Hospice Survey revealed a median survival of 35 days.• There are several possible reasons for the pattern of late re ferral in this sample. The most probable is that physicians were overestimating survival time in their terminal cancer patients, as this has been well documented.2• • • A second possible reason is that hospice services may not be well known to community physicians. However, the majority of referrals are from institutions within The Texas Medical Center, where the existence of hospice services is most widely known. A third possible explanation for late referral is that it may be difficult for physicians to refer patients to a hospice as this may contribute to ideas of professional failure in not controlling the disease process. A correlate of this may be that physicians pre fer to personally continue primary care of their patient until death and erroneously believe that a hospice referral may prevent this occurrence. Thus, although the exact cause for late referrals to hos pice could not be ascertained, it is possible that a greater benefit 5 67
THE HOSPICE JOURNAL
72
Table 2 Survival Time for Significant Variables mean s urvival in days (n)
Downloaded by [University of New England] at 20:16 08 December 2017
overall mean survival a 48 days
< 25
AGE:
25 - 49
44 (98)
74
42 (49)
< 75
22 (10)
75 - 99
67 (65)
100 - 124
39 (86)
124
13 (10)
0.023
PULSE:
>
p
0.011 Normal
MOBILITY:
p
76 (22)
50 - 74
>
p
64 (3)
35 ( 1 )
Walks in home
72 (60)
Bears weight
47 (44)
Sits up
37 (30)
Unable to turn
18 (37)
< 0.0001 Norma 1
APPETITE:
66 (9)
Decreased i ntake
74 (38)
Soft food
52 (75)
Occassional fluid
19 (42)
Refuses fluid p < 0.0001
12 (8)
73
Research and Evaluation
Downloaded by [University of New England] at 20:16 08 December 2017
NOURISHMENT:
Heavy
Normal build
63 (34)
Thin
46 (79)
Very thin
42 (43)
Extreme cachexia P
65 (8)
10 (8)
0.043
BATHING:
Dependent
39 (148)
Independent
110 (22)
P = 0.00015 CONTINENCE:
Dependent
30 ( 92)
Indepencient
69 (78)
P < 0.0001 DRESSING:
Dependent
37 (142)
Independent
1 06 (28)
P
ISSN: 0742-969X (Print) (Online) Journal homepage: http://www.tandfonline.com/loi/ippc19
Estimation of Survival Time in Terminal Cancer Patients: An Impedance to Hospice Admissions? Ronald S. Schonwetter, Thomas A. Teasdale, Porter Storey & Robert J. Luchi To cite this article: Ronald S. Schonwetter, Thomas A. Teasdale, Porter Storey & Robert J. Luchi (1990) Estimation of Survival Time in Terminal Cancer Patients: An Impedance to Hospice Admissions?, The Hospice Journal, 6:4, 65-79, DOI: 10.1080/0742-969X.1990.11882684 To link to this article: https://doi.org/10.1080/0742-969X.1990.11882684
Published online: 13 Sep 2017.
Submit your article to this journal
Article views: 1
View related articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=ippc19 Download by: [University of New England]
Date: 08 December 2017, At: 20:16
Downloaded by [University of New England] at 20:16 08 December 2017
Estimation of Survival Time in Terminal Cancer Patients: An Impedance to Hospice Admissions? Ronald S. Schonwetter Thomas A. Teasdale Porter Storey Robert J. Luchi
ABSTRACT. Accurate estimation of survival time in terminal can cer patients is difficult yet may provide useful information. A histor ical prospective study on 172 patients admitted to a home based hospice service was performed to determine which variables were best correlated with survival time. Mean and median survival were 48 and 22 days, respectively, representing a highly skewed distribu tion of life span in this sample. As age increased, survival time decreased. All Activities of Daily Living (ADLs) recorded (Bathing, Continence, Dressing and Transfer) as well as other measures of performance (mobility and pulse) and nutrition (appetite and nour ishment) were each strongly associated with survival. Multivariate Ronald S. Schonwetter, MD, is affiliated with the Division of Geriatrics, De partment of Internal Medicine, University of South Florida College of Medicine, Hospice of Hillsborough, Inc., Tampa, FL. He is also affiliated with the Huf fington Center on Aging, Baylor College of Medicine and the Veterans Affairs Medical Center, Houston, TX. Thomas A. Teasdale, MPH, is affiliated with the Huffington Center on Aging, Baylor College of Medicine and the Veterans Af fairs Medical Center, Houston, TX. Porter Storey, MD, is affiliated with the New Age Hospice, Houston, TX. Robert J. Luchi, MD, is affiliated with the Huf fington Center on Aging, Baylor College of Medicine and the Veterans Affairs Medical Center, Houston, TX. Requests for reprints should be sent to Ronald S. Schonwetter, MD, Univer sity of South Florida College of Medicine, 12901 Bruce B. Downs Boulevard, Box 19, Tampa, FL 33612. An earlier version of this paper was presented at the 14th International Con gress of Gerontology, Acapulco, Mexico, June, 1989. The Hospice Journal, Vol. 6(4) 1990 © 1991 by The Haworth Press, Inc. All rights reserved.
65
Downloaded by [University of New England] at 20:16 08 December 2017
66
THE HOSPICE JOURNAL
analysis limited significant variables to dressing ability, pulse rate, level of appetite and transferring ability. Outliers (survival > 180 days) were differentiated from the remainder of the sample by sig nificant differences in all ADLs recorded as well as the level of appetite. These findings establish the importance of assessing ADLs, a measure of functional status, and reinforce the importance of performance and nutrition measures when estimating length of survival in terminal cancer patients.
To estimate the life expectancy for a patient with terminal cancer is a difficult yet important task for medical personnel. An accurate estimation allows appropriate admissions to a hospice, for services that might improve the patient's quality of dying.1 Reasonable as sessment of survival time provides patients and relatives a time frame to plan a strategy for emotionally and financially managing the terminal period. Additionally, it provides staff with an estimate of the time they have to develop their patient's insight into the dying process. There have been relatively few studies focusing on estimating survival in terminally ill cancer patients. One recent study2 com pared a terminal care service staff's estimate of survival in tenninal cancer patients, a functional rating produced from the Kamofsky Performance Scale (KPS) 3 and the patient's actual survival. This study reported that the KPS better correlated with actual survival than subjective estimates, which were overly optimistic. Reuben et al.,4 recently reported that the KPS, a measure of functional status, was the most important clinical factor in estimating survival time, but five other clinical symptoms had independent predictive value as well. Unfortunately, most patients with terminal cancer cluster around the lower end of the KPS, reducing the magnitude of varia tion in the scale and impairing discriminative potential. Three addi tional studies5• • reported physicians to be overly optimistic in terms of subjective predictions of survival time in hospice patients. These data suggest that physicians' estimates of length of survival in pa tients terminally ill from cancer should be interpreted very cau tiously and that other objective measures involving functional status evaluations are needed to better prognosticate. Medical staff, including physicians, may be uncomfortable and reluctant to discuss a patient's prognosis with the patient and fam ily, possibly due to the limited prognostic accuracy. Additionally, 6 7
Downloaded by [University of New England] at 20:16 08 December 2017
Research and Evaluation
67
doctors may believe the patient may blame the doctor personally for the poor outcome. It is not uncommon for staff to fear a poor emo tional reaction by the patient, but typically discover that the patient is relieved with the information. The staff or physician may also fear their personal reaction, which may be emotional when involv ing a close patient. Until the recent implementation of The Catastrophic Illness cov erage, Medicare reimbursed hospice services for a maximum of 210 days, and only for patients who were certified by a physician to have a terminal disease with a life expectancy of less than 6 months. As of January 1, 1989, there are no Medicare-specific limitations concerning length of survival for reimbursement. However, certifi cation of terminal illness is still required by a physician and the maximum amount reimbursed for hospice services is unchanged. Additionally, several private insurance policies place monetary lim itations on hospice service reimbursements. For these reasons, more precise estimates of survival time are needed in order for hos pice services to be cost-effective and to avoid utilizing the entire benefit prior to the patient's expiration. The purpose of the present study was to determine whether certain objective variables corre lated with survival time and whether they may be used to better predict length of survival in terminal cancer patients. METHODS
The New Age Hospice, founded in 1983, is a non-profit organi zation committed to providing physical, emotional and spiritual support to the terminally ill and their families. Their focus of care is in the patient's home, although there is an inpatient unit should that type of care be necessary. To qualify for admission to this hospice program, patients have expected survivals in terms of weeks or months as estimated by their referring physician. Three hundred thirty two medical charts were randomly selected from the charts of all patients admitted between March, 1986 and January, 1988 (n = 997). Data were collected only on patients admitted to the home-based community hospice program (n = 202). Patients were excluded if they had a terminal disease other than cancer (n = 18) or were discharged from the service for rea-
Downloaded by [University of New England] at 20:16 08 December 2017
68
THE HOSPICE JOURNAL
sons other than death (n = 12). One hundred seventy two patients were ultimately included in the present study. All study patients were initially seen and evaluated by The New Age Hospice physician and a registered nurse. Demographic data including age, sex, and marital status were recorded by study inves tigators. The following information was obtained from the physi cian's initial history and physical examination: patient acknowledg ment of terminal disease (yes or no); number of pre-admission medications; systolic and diastolic blood pressures; pulse rate; as signment of mobility (1 = normal to 5 = unable to turn in bed), appetite (1 = normal to 5 = refuses fluid), nourishment (1 = heavy to 5 = extreme cachexia) and pain (1 = none to 6 = over whelming). ADLs included transfer, bathing, dressing and conti nence only, and were evaluated by nurses on the initial visit and recorded as dependent or independent, according to the original definitions by Katz. 8 Length of survival was calculated from initial hospice visit to time of death. Statistical analysis was performed on the overall sample as well as on defined subgroups. Survival time was grouped in 30 day peri ods for chi square tests on categorical data. T Tests were used to determine significance levels of continuous independent variables. An analysis of variance was completed to eliminate possible over lap in the variables used in regression analysis. A forward stepwise regression analysis was completed to calculate correlation coeffi cients and to identify variables which would best predict survival time in these patients. An additional regression analysis was com pleted with outliers (survival > 180 days; n = 9) eliminated to determine whether this would increase the predictive value of the variables. Summary values for outliers and the remainder of the population were tested for potential differences. RESULTS
Characteristics of the study population are shown in Table 1. Fifty six percent of the sample were males and forty four percent were females. The mean age of the patients was 63 years, with 85% of the population 50 years of age and older. Slightly less than one half of the patients were married.
Research and Evaluation
69
Table l Characteristics of study population (n=l72)
Downloaded by [University of New England] at 20:16 08 December 2017
AGE
GENDER
< 25 .. • ...... • .. • • • • ... • 2% 25 - 49
13%
50 - 74
57%
> 74
28%
Male ................... 56% Female ................. 44%
MARITAL STATUS
Married ................ 48% Not married ............ 52%
Overall mean survival was 48 days while the median survival was 22 days. Only 35% of the patients lived beyond seven weeks (approximate mean survival). The distribution of life span is highly skewed (Figure 1) with the majority of these terminal cancer pa tients surviving only three weeks. Table 2 shows survival in days by significant variables collected as categorical data. As age increased, survival time decreased sig nificantly. Survival time decreased in patients with extremely high or low pulse rates. Performance (mobility) and nutrition (appetite and nourishment) were associated with shorter survival in patients who were less mobile, ate less, and were physically thinner. Inde pendence in each of the ADLs recorded (bathing, continence, dress ing and transfer) was strongly associated with increased survival in terminal cancer patients. Table 3 shows the correlation coefficients and significance levels of each variable with survival time. All ADLs recorded were noted to have p values of < 0.001. Those variables not significantly associated with survival were gender,
THE HOSPICE JOURNAL
70
Figure 1 PROPORTIONAL SURVIVAL CURVE N=172
Downloaded by [University of New England] at 20:16 08 December 2017
100
@
40
20: I I
o,______._____________ IJ 100 1:x:J
200
SURVIAL IN DAYS
marital status, number of pre-admission medications, systolic and diastolic blood pressures and level of pain. A stepwise regression analysis (n = 172) limited significant vari ables to dressing (p < 0.001), pulse (p = 0.01), appetite (p = 0.01) and transfer (p = 0.02). These four variables accounted for
Downloaded by [University of New England] at 20:16 08 December 2017
Research and Evaluation
71
30% (r2 = 0.30) of the variation in survival time. A second step wise regression analysis included 163 subjects and excluded the outliers (n = 9). No decrease in variation of survival time (r2 = 0.29) was shown. Figure 2 represents the predicted and observed survival time in days using the initial regression analysis. It appears that many of the patients who died under 7 weeks had overestima tions in their predicted length of survival while many patients who died after 7 weeks had underestimations of their predicted life span. Comparison of the outliers and the remainder of the population resulted in differences in each AOL included ((bathing (p < 0.001); dressing (p < 0.001 ); continence (p = 0.008); and transfer (p = 0.024)) as well as the appetite level (p = 0.038). No additional variables differentiated the outliers from the remainder of the study sample. DISCUSSION
Median survival was just over 3 weeks for patients referred to this outpatient hospice service and thus, physicians may have re ferred patients too late to achieve full benefits of these services. Median survival in an in-patient hospice, typically for patients that are more ill and closer to death, was 24.5 days5 and previous data from The National Hospice Survey revealed a median survival of 35 days.• There are several possible reasons for the pattern of late re ferral in this sample. The most probable is that physicians were overestimating survival time in their terminal cancer patients, as this has been well documented.2• • • A second possible reason is that hospice services may not be well known to community physicians. However, the majority of referrals are from institutions within The Texas Medical Center, where the existence of hospice services is most widely known. A third possible explanation for late referral is that it may be difficult for physicians to refer patients to a hospice as this may contribute to ideas of professional failure in not controlling the disease process. A correlate of this may be that physicians pre fer to personally continue primary care of their patient until death and erroneously believe that a hospice referral may prevent this occurrence. Thus, although the exact cause for late referrals to hos pice could not be ascertained, it is possible that a greater benefit 5 67
THE HOSPICE JOURNAL
72
Table 2 Survival Time for Significant Variables mean s urvival in days (n)
Downloaded by [University of New England] at 20:16 08 December 2017
overall mean survival a 48 days
< 25
AGE:
25 - 49
44 (98)
74
42 (49)
< 75
22 (10)
75 - 99
67 (65)
100 - 124
39 (86)
124
13 (10)
0.023
PULSE:
>
p
0.011 Normal
MOBILITY:
p
76 (22)
50 - 74
>
p
64 (3)
35 ( 1 )
Walks in home
72 (60)
Bears weight
47 (44)
Sits up
37 (30)
Unable to turn
18 (37)
< 0.0001 Norma 1
APPETITE:
66 (9)
Decreased i ntake
74 (38)
Soft food
52 (75)
Occassional fluid
19 (42)
Refuses fluid p < 0.0001
12 (8)
73
Research and Evaluation
Downloaded by [University of New England] at 20:16 08 December 2017
NOURISHMENT:
Heavy
Normal build
63 (34)
Thin
46 (79)
Very thin
42 (43)
Extreme cachexia P
65 (8)
10 (8)
0.043
BATHING:
Dependent
39 (148)
Independent
110 (22)
P = 0.00015 CONTINENCE:
Dependent
30 ( 92)
Indepencient
69 (78)
P < 0.0001 DRESSING:
Dependent
37 (142)
Independent
1 06 (28)
P
