Abstract

Background: Quality of end-of-life care received by cancer patients has never been explored in an entire Asian country for all ages and cancer groups.

Patients and methods: Retrospective cohort study to examine trends in quality of end-of-life care among a cohort of 242 530 Taiwanese cancer patients who died in 2000–2006.

Results: In the last month of life, cancer care tended to become increasingly aggressive as shown by (i) intensive use of chemotherapy (15.45%–17.28%), (ii) frequent emergency room visits (15.69%–20.99%) and >14-day hospital stays (41.48%–46.20%), (iii) admissions to intensive care units (10.04%–12.41%), and (iv) hospital deaths (59.11%–65.40%). Use of cardiopulmonary resuscitation (13.09%–8.41%), intubation (26.01%–21.07%), and mechanical ventilation (27.46%–27.05%) decreased, whereas use of hospice services increased considerably (7.34%–16.83%). Among those receiving hospice services, rates of referrals to hospice services in the last 3 days of life decreased from 17.88% to 17.13% but remained steady after adjusting for selected covariates.

Conclusions: The quality of end-of-life care for Taiwanese cancer decedents was substantially inferior to that previously reported and to that recommended as benchmarks for not providing overly aggressive care near the end of life.

introduction

End-of-life care has become increasingly aggressive in the last decade [1, 2], resulting in disproportionately escalating health expenditures for decedents than for survivors [1, 3–5]. However, this increased treatment intensity does not correspond to improved health outcomes and patient or family satisfaction [6, 7], raising concerns about the appropriateness of providing aggressive care before patients die.

To systematically examine the quality of end-of-life cancer care, Earle et al. [8] developed major indicators of poor quality care currently available in administrative data. Those indicators were adopted by researchers from the United States [2, 9], Canada [10–12], and Portugal [13] to illustrate the quality of care received by cancer patients in the last months of life. However, such patterns of care have never been explored in an Asian country for all ages and cancer groups. The importance of evaluating such trends in Taiwan is highlighted by two facts: (i) cancer ranks fifth in Taiwan's National Health Insurance (NHI) outlays [14] and has been the leading cause of death since 1982, accounting for 28.1% of total deaths in 2006 [15] and (ii) due to universal insurance coverage by the NHI, trends in quality of cancer care in the last months of life can be analyzed in a population encompassing all age and cancer groups, facilitating comparison with international trends. Therefore, the purpose of this study was to evaluate trends in quality of care near the end of life in a population-based cohort of 242 530 cancer patients who died from 2000 to 2006 in Taiwan.

patients and methods

study design and sample

A retrospective cohort design was used to examine selected health care received for cancer patients near the end of life in Taiwan. Individual patient-level data were linked by a unique, scrambled personal identification number from computerized inpatient and outpatient NHI claims data, the Cancer Registration System (CRS) database, and the National Register of Deaths Database (NRDD). The completeness and accuracy of these databases were maintained by the Taiwan Department of Health (DOH). The unique characteristics of the NHI in Taiwan include universal insurance coverage, comprehensive benefits or services provided, and a single-payer system with the government as the sole insurer. Patients have free access to any health-care system and provider of their choice. Health-care systems are reimbursed for the services provided, and copayment is waived for patients with recognized major diseases, including malignancy. At the end of 2006, the NHI covered 99.0% of Taiwan's population [16]. The CRS included 97.34% of incident cancer cases with 86.80% completeness and 83.51%–90.06% accuracy [17]. The NRDD identified 242 530 cancer deaths from 2000 to 2006, with the cause of death specified as malignancy. The cancer decedents’ characteristics, broken down by year of death, are shown in Table 1 (categorical data) and Table 2 (continuous data).

Table 1.

Demographic characteristics (categorical) of Taiwanese cancer patients who died in 2000–2006

Characteristic Percent of patients with characteristic
 
P 
 Average, n = 242 530 2000, n = 31 554 2001, n = 32 992 2002, n = 34 342 2003, n = 35 096 2004, n = 35 808 2005, n = 37 065 2006, n = 35 673  
Gender         0.96 
    Male 64.23 64.51 64.45 64.12 63.49 64.21 64.37 64.47  
    Female 35.77 35.49 35.55 35.88 36.51 35.79 35.63 35.53  
Cancer diagnosis         <0.0001 
    Head and neck 8.37 8.31 8.22 7.98 8.38 8.51 8.57 8.56  
    Breast 3.75 3.67 3.79 3.53 3.94 3.67 3.86 3.78  
    Lung 19.75 19.84 19.87 19.93 19.61 19.65 19.60 19.79  
    Gastric 6.94 7.52 7.41 7.08 6.66 6.89 6.68 6.41  
    Colon–rectal 10.77 10.70 10.48 10.63 10.55 10.71 11.04 11.26  
    Hepatoma 19.55 19.02 19.44 20.22 19.95 19.49 19.14 19.55  
    Pancreatic 5.76 5.63 5.63 5.86 5.80 5.83 5.77 5.80  
    Hematological malignancies 5.56 5.70 5.47 5.41 5.54 5.54 5.54 5.72  
Time since diagnosis ≤1 year 58.07 62.09 60.93 59.66 57.36 57.02 55.09 55.18 <0.0001 
Characteristic Percent of patients with characteristic
 
P 
 Average, n = 242 530 2000, n = 31 554 2001, n = 32 992 2002, n = 34 342 2003, n = 35 096 2004, n = 35 808 2005, n = 37 065 2006, n = 35 673  
Gender         0.96 
    Male 64.23 64.51 64.45 64.12 63.49 64.21 64.37 64.47  
    Female 35.77 35.49 35.55 35.88 36.51 35.79 35.63 35.53  
Cancer diagnosis         <0.0001 
    Head and neck 8.37 8.31 8.22 7.98 8.38 8.51 8.57 8.56  
    Breast 3.75 3.67 3.79 3.53 3.94 3.67 3.86 3.78  
    Lung 19.75 19.84 19.87 19.93 19.61 19.65 19.60 19.79  
    Gastric 6.94 7.52 7.41 7.08 6.66 6.89 6.68 6.41  
    Colon–rectal 10.77 10.70 10.48 10.63 10.55 10.71 11.04 11.26  
    Hepatoma 19.55 19.02 19.44 20.22 19.95 19.49 19.14 19.55  
    Pancreatic 5.76 5.63 5.63 5.86 5.80 5.83 5.77 5.80  
    Hematological malignancies 5.56 5.70 5.47 5.41 5.54 5.54 5.54 5.72  
Time since diagnosis ≤1 year 58.07 62.09 60.93 59.66 57.36 57.02 55.09 55.18 <0.0001 
Table 2.

Demographic characteristics (continuous) of Taiwanese cancer patients who died in 2000–2006

Characteristic Mean ± standard deviation
 
P 
 Average, n = 242 530 2000, n = 31 554 2001, n = 32 992 2002, n = 34 342 2003, n = 35 096 2004, n = 35 808 2005, n = 37 065 2006, n = 35 673  
Age (years) 65.8 ± 14.9 64.9 ± 15.0 65.3 ± 14.9 65.5 ± 14.9 65.9 ± 14.8 66.0 ± 15.0 66.4 ± 14.9 66.6 ± 14.8 <0.0001 
Time since diagnosis (months) 23.1 ± 34.2 20.2 ± 30.7 20.6 ± 31.0 21.2 ± 31.9 22.2 ± 33.1 23.2 ± 34.2 24.9 ± 36.0 29.2 ± 39.8 <0.0001 
Characteristic Mean ± standard deviation
 
P 
 Average, n = 242 530 2000, n = 31 554 2001, n = 32 992 2002, n = 34 342 2003, n = 35 096 2004, n = 35 808 2005, n = 37 065 2006, n = 35 673  
Age (years) 65.8 ± 14.9 64.9 ± 15.0 65.3 ± 14.9 65.5 ± 14.9 65.9 ± 14.8 66.0 ± 15.0 66.4 ± 14.9 66.6 ± 14.8 <0.0001 
Time since diagnosis (months) 23.1 ± 34.2 20.2 ± 30.7 20.6 ± 31.0 21.2 ± 31.9 22.2 ± 33.1 23.2 ± 34.2 24.9 ± 36.0 29.2 ± 39.8 <0.0001 

indicators of poor quality of end-of-life care

The indicators used in this study for poor quality of end-of-life cancer care were adapted from Earle et al. [8] on the basis of the data available from the NHI claims database and the synthesized evidence of futility of life-sustaining treatments for terminally ill cancer patients, e.g. cardiopulmonary resuscitation (CPR) [18] and intubation for mechanical ventilation [19]. Specifically, indicators included (i) receiving chemotherapy in the last month of life, (ii) more than one emergency room (ER) visit, more than one inpatient hospital admission, or >14-day hospital stay in the last month of life, (iii) use of life-sustaining treatments such as admission to an intensive care unit (ICU), CPR, intubation, and mechanical ventilation in the last month of life, (iv) no use of hospice services in the last year of life or referral to hospice services in the last 3 days of life, and (v) dying in a hospital.

chemotherapy.

Use of chemotherapy was identified from both inpatient and outpatient claims, which specified the codes for all routes of chemotherapy administration but oral.

ER visits and inpatient hospital admissions.

ER visits were identified from inpatient and outpatient claims. Dates of hospitalization and discharge were identified from inpatient claims. Length of stay for each hospitalization was measured from the date of the hospital admission until discharge. Admission and discharge on the same day was counted as 1 day. If a cancer patient had multiple inpatient hospital admissions in the last month of life, lengths of stay were summed up across those admissions.

ICU admission was ascertained from inpatient hospital claims. Admissions to either coronary care or traditional intensive care units were considered ICU admissions.

CPR, intubation, and mechanical ventilation were identified from the inpatient claims by the codes specified for those procedures.

use of hospice services.

Regardless of whether a patient was enrolled in hospice home care or inpatient hospice services, each patient was counted only once. Duration of survival after hospice enrollment was calculated as the interval between the date of first enrollment in hospice care and the date of death. Duration of survival for cancer patients who enrolled in hospice care and were discharged or died on the same day was counted as 1 day.

referral to hospice services in the last 3 days of life.

Cancer patients were recognized as referred to hospice services in the last 3 days of life if their survival after hospice enrollment was ≤3 days.

death in a hospital.

Place of death is coded by the NRDD as hospital, home, or other, and such codes were used by other researchers [20]. However, these codes were not used in our study. This choice was made due to the cultural meaning of dying at home [21]. Dying patients in Taiwan are commonly discharged ‘against medical advice’ and often with artificial respiratory support to allow family members to keep the patient alive to die at home. Therefore, if the date of discharge for the last admission was also the date of death [11], the patient was recognized as dying in the hospital.

demographic and disease-related characteristics

The patients’ gender and age at the time of death were obtained from the NRDD. Diagnoses and date of diagnosis were identified from the CRS.

data analysis

The rates of using selected indicators of quality of end-of-life care were examined by the Cochran-Armitage test [22] for statistically significant changes over the study period. The effect of year of death, i.e. the odds for each indicator in that year relative to the year 2000, on outcome variables was examined by multivariate logistic regression while controlling for age, gender, common diagnoses of cancer (lung, liver, colon–rectal, gastric, pancreatic, breast, and hematological malignancies), and survival time after cancer diagnosis. The effect of year of death on outcome variables was measured by adjusted odds ratio with 95% confidence interval (CI). For sensitivity analysis, all analyses were repeated on patients surviving within 1 year of cancer diagnosis. All analyses showed the same pattern as for the whole sample, but toward more aggressive care. Therefore, results are not reported for the cohort with ≤1 year survival after cancer diagnosis.

results

Averagely, 16.21% of cancer patients who died in 2000–2006 received chemotherapy in their last month of life (Table 3). Rates of chemotherapy in the last month of life increased from 15.45% in 2000 to 17.28% in 2006. When selected demographic and disease characteristics were controlled, the likelihood of receiving chemotherapy in the last month of life was significantly greater in 2004–2006 than in 2000 (Table 4).

Table 3.

Trends in indicators for quality of end-of-life care among Taiwanese cancer patients who died in 2000–2006

Quality indicator for end-of-life cancer care Percent of patients with indicator P 
Average, n = 242 530 2000, n = 31 554 2001, n = 32 992 2002, n = 34 342 2003, n = 35 096 2004, n = 35 808 2005, n = 37 065 2006, n = 35 673  
Chemotherapy in the last month of life 16.21 15.45 16.11 15.51 15.45 16.20 17.28 17.28 <0.0001 
More than one ER visit in the last month of life 18.34 15.69 17.51 17.40 17.31 18.98 20.01 20.99 <0.0001 
More than one hospital admission in the last month of life 14.21 13.86 15.09 14.97 14.25 13.85 13.89 13.62 <0.0001a 
>14-day hospital stay in the last month of life 44.42 41.48 43.08 43.91 43.85 45.89 45.97 46.20 <0.0001 
ICU admission in the last month of life 11.43 10.04 12.02 11.76 11.26 11.43 10.98 12.41 <0.0001 
CPR in the last month of life 10.59 13.09 12.74 11.59 10.37 9.93 8.59 8.41 <0.0001a 
Intubation in the last month of life 24.27 26.01 27.70 26.69 24.32 23.71 21.08 21.07 <0.0001a 
Mechanical ventilation in the last month of life 28.53 27.46 30.32 30.47 29.21 28.83 26.54 27.05 <0.0001a 
Use of hospice services in the last year of life 14.08 7.34 12.56 13.95 14.54 15.56 16.75 16.83 <0.0001 
Referral to hospice services in the last 3 days of lifeb 17.66 17.88 18.59 18.73 18.29 16.77 16.89 17.13 0.0036a 
Dying in an acute care hospital 63.23 59.11 61.84 63.21 63.23 65.15 64.06 65.40 <0.0001 
Quality indicator for end-of-life cancer care Percent of patients with indicator P 
Average, n = 242 530 2000, n = 31 554 2001, n = 32 992 2002, n = 34 342 2003, n = 35 096 2004, n = 35 808 2005, n = 37 065 2006, n = 35 673  
Chemotherapy in the last month of life 16.21 15.45 16.11 15.51 15.45 16.20 17.28 17.28 <0.0001 
More than one ER visit in the last month of life 18.34 15.69 17.51 17.40 17.31 18.98 20.01 20.99 <0.0001 
More than one hospital admission in the last month of life 14.21 13.86 15.09 14.97 14.25 13.85 13.89 13.62 <0.0001a 
>14-day hospital stay in the last month of life 44.42 41.48 43.08 43.91 43.85 45.89 45.97 46.20 <0.0001 
ICU admission in the last month of life 11.43 10.04 12.02 11.76 11.26 11.43 10.98 12.41 <0.0001 
CPR in the last month of life 10.59 13.09 12.74 11.59 10.37 9.93 8.59 8.41 <0.0001a 
Intubation in the last month of life 24.27 26.01 27.70 26.69 24.32 23.71 21.08 21.07 <0.0001a 
Mechanical ventilation in the last month of life 28.53 27.46 30.32 30.47 29.21 28.83 26.54 27.05 <0.0001a 
Use of hospice services in the last year of life 14.08 7.34 12.56 13.95 14.54 15.56 16.75 16.83 <0.0001 
Referral to hospice services in the last 3 days of lifeb 17.66 17.88 18.59 18.73 18.29 16.77 16.89 17.13 0.0036a 
Dying in an acute care hospital 63.23 59.11 61.84 63.21 63.23 65.15 64.06 65.40 <0.0001 
a

Indicates a decreasing trend by the Cochran-Armitage test; all other trends are increasing.

b

Among 34 139 patients admitted to hospice.

ER, emergency room; ICU, intensive care unit; CPR, cardiopulmonary resuscitation.

Table 4.

Adjusted odds ratios for occurrences of indicators for quality of end-of-life cancer care in comparison with year 2000a

Quality indicator for end-of-life cancer care 2000 2001 2002 2003 2004 2005 2006 
AOR (95% CI) 
Chemotherapy in the last month of life Reference 1.05 (1.00, 1.10) 1.01 (0.97, 1.06) 1.05 (0.98, 1.07) 1.10*** (1.05, 1.15) 1.20*** (1.15, 1.25) 1.24*** (1.19, 1.30) 
More than one ER visit in the last month of life Reference 1.12*** (1.06, 1.16) 1.10*** (1.05, 1.15) 1.10*** (1.05, 1.15) 1.24*** (1.19, 1.29) 1.31*** (1.25, 1.36) 1.38*** (1.32, 1.44) 
More than one hospital admission in the last month of life Reference 1.09*** (1.05, 1.15) 1.08** (1.04, 1.14) 1.04 (1.00, 1.09) 1.02 (0.97, 1.06) 1.00 (0.96, 1.04) 0.95* (0.90, 0.99) 
>14-day hospital stay in the last month of life Reference 1.02 (0.99, 1.06) 1.06** (1.02, 1.09) 1.07*** (1.03, 1.10) 1.14*** (1.11, 1.18) 1.15*** (1.12, 1.19) 1.15*** (1.10, 1.19) 
ICU admission in the last month of life Reference 1.20*** (1.14, 1.26) 1.17*** (1.11, 1.23) 1.11*** (1.06, 1.17) 1.12*** (1.06, 1.18) 1.07* (1.01, 1.12) 1.15*** (1.09, 1.21) 
CPR in the last month of life Reference 0.93** (0.88, 0.97) 0.84*** (0.80, 0.88) 0.73*** (0.70, 0.77) 0.71*** (0.67, 0.74) 0.61*** (0.57, 0.64) 0.56*** (0.53, 0.59) 
Intubation in the last month of life Reference 1.04* (1.01, 1.08) 0.99 (0.95, 1.03) 0.87*** (0.84, 0.91) 0.84*** (0.81, 0.87) 0.73*** (0.70, 0.76) 0.70*** (0.67, 0.73) 
Mechanical ventilation in the last month of life Reference 1.10*** (1.06, 1.14) 1.11*** (1.07, 1.15) 1.04* (1.01, 1.07) 1.00 (0.97, 1.04) 0.90*** (0.87, 0.94) 0.90*** (0.86, 0.93) 
Use of hospice services in the last year of life Reference 1.80*** (1.70, 1.90) 2.02*** (1.91, 2.13) 2.15*** (2.04, 2.27) 2.32*** (2.20, 2.45) 2.49*** (2.36, 2.62) 2.59*** (2.45, 2.73) 
Referral to hospice services in the last 3 days of lifeb Reference 1.06 (0.93, 1.22) 1.07 (0.94, 1.22) 1.03 (0.90, 1.07) 0.93 (0.81, 1.06) 0.93 (0.82, 1.06) 0.90 (0.79, 1.04) 
Dying in an acute care hospital Reference 1.05** (1.01, 1.09) 1.11*** (1.07, 1.15) 1.12*** (1.08, 1.16) 1.20*** (1.17, 1.25) 1.15*** (1.12, 1.19) 1.21*** (1.17, 1.25) 
Quality indicator for end-of-life cancer care 2000 2001 2002 2003 2004 2005 2006 
AOR (95% CI) 
Chemotherapy in the last month of life Reference 1.05 (1.00, 1.10) 1.01 (0.97, 1.06) 1.05 (0.98, 1.07) 1.10*** (1.05, 1.15) 1.20*** (1.15, 1.25) 1.24*** (1.19, 1.30) 
More than one ER visit in the last month of life Reference 1.12*** (1.06, 1.16) 1.10*** (1.05, 1.15) 1.10*** (1.05, 1.15) 1.24*** (1.19, 1.29) 1.31*** (1.25, 1.36) 1.38*** (1.32, 1.44) 
More than one hospital admission in the last month of life Reference 1.09*** (1.05, 1.15) 1.08** (1.04, 1.14) 1.04 (1.00, 1.09) 1.02 (0.97, 1.06) 1.00 (0.96, 1.04) 0.95* (0.90, 0.99) 
>14-day hospital stay in the last month of life Reference 1.02 (0.99, 1.06) 1.06** (1.02, 1.09) 1.07*** (1.03, 1.10) 1.14*** (1.11, 1.18) 1.15*** (1.12, 1.19) 1.15*** (1.10, 1.19) 
ICU admission in the last month of life Reference 1.20*** (1.14, 1.26) 1.17*** (1.11, 1.23) 1.11*** (1.06, 1.17) 1.12*** (1.06, 1.18) 1.07* (1.01, 1.12) 1.15*** (1.09, 1.21) 
CPR in the last month of life Reference 0.93** (0.88, 0.97) 0.84*** (0.80, 0.88) 0.73*** (0.70, 0.77) 0.71*** (0.67, 0.74) 0.61*** (0.57, 0.64) 0.56*** (0.53, 0.59) 
Intubation in the last month of life Reference 1.04* (1.01, 1.08) 0.99 (0.95, 1.03) 0.87*** (0.84, 0.91) 0.84*** (0.81, 0.87) 0.73*** (0.70, 0.76) 0.70*** (0.67, 0.73) 
Mechanical ventilation in the last month of life Reference 1.10*** (1.06, 1.14) 1.11*** (1.07, 1.15) 1.04* (1.01, 1.07) 1.00 (0.97, 1.04) 0.90*** (0.87, 0.94) 0.90*** (0.86, 0.93) 
Use of hospice services in the last year of life Reference 1.80*** (1.70, 1.90) 2.02*** (1.91, 2.13) 2.15*** (2.04, 2.27) 2.32*** (2.20, 2.45) 2.49*** (2.36, 2.62) 2.59*** (2.45, 2.73) 
Referral to hospice services in the last 3 days of lifeb Reference 1.06 (0.93, 1.22) 1.07 (0.94, 1.22) 1.03 (0.90, 1.07) 0.93 (0.81, 1.06) 0.93 (0.82, 1.06) 0.90 (0.79, 1.04) 
Dying in an acute care hospital Reference 1.05** (1.01, 1.09) 1.11*** (1.07, 1.15) 1.12*** (1.08, 1.16) 1.20*** (1.17, 1.25) 1.15*** (1.12, 1.19) 1.21*** (1.17, 1.25) 
a

Adjusted for gender, age, common diagnoses, and postdiagnosis survival.

b

Among 34 139 patients admitted to hospice.

*P < 0.05; **P < 0.001; ***P ≤ 0.0001.

AOR, adjusted odds ratio; CI, confidence interval; ER, emergency room; ICU, intensive care unit; CPR, cardiopulmonary resuscitation.

Averagely, 18.34% of cancer patients in their last month of life visited the ER more than once (Table 3). Compared with 2000, the likelihood of frequent (more than one) ER visits was stable from 2001 to 2003; thereafter, it increased at a rate of 7% per year (Table 4). During the last month of life, averagely, 14.21% of cancer patients were frequently (more than one) admitted to a hospital, and 44.42% spent >14 days in a hospital (Table 3). Compared with 2000, the adjusted odds of frequent hospitalizations in the last month of life increased significantly in 2001–2002, regressed to a comparable level in 2003–2005, and decreased significantly in 2006 (Table 4). Over the 7-year period studied, except for 2001, cancer patients in their last month of life showed a significantly increasing trend to stay >14 days in a hospital (Table 4).

Different patterns of change were found during the study period between ICU care and other life-sustaining treatments such as CPR, intubation, and mechanical ventilation in the last month of life. Averagely, only slightly more than one-tenth (11.43%) of cancer patients received ICU care in the last month of their life, but a significant trend was seen toward more use of such care over the study period (from 10.04% in 2000 to 12.41% in 2006) (Tables 3 and 4). However, the opposite trends were found for CPR (from 13.09% in 2000 to 8.41% in 2006) or intubation (from 26.01% in 2000 to 21.07% in 2006) in the last month of life. The adjusted odds of use of mechanical ventilation increased significantly from 2000 to 2003, regressed in 2004 to a level comparable to that of 2000, and decreased significantly in 2005–2006 (Table 4). Over time, use of mechanical ventilation in the last month of life showed a decreasing trend by the Cochran-Armitage test (Table 3).

The rate of use of hospice services was small for each study year, but more than doubled over time from 7.34% to 16.83% (Table 3). After selected demographic and disease characteristics were controlled for the small proportion of patients who received hospice care, rates of referrals to hospice services in the last 3 days of life remained steady over the study period (Table 4). The majority of cancer patients (on an average, 63.23%) died in an acute care hospital. Indeed, cancer patients became significantly more likely over time to die in an acute care hospital (Table 4).

discussion

In this study, we found evidence that Taiwanese cancer patients who died in 2000–2006 were increasingly more likely to receive aggressive care in the last month of life, including (i) intensive chemotherapy, (ii) more than one ER visit and >14-day hospital stays, and (iii) ICU admissions. Similar patterns of aggressive care were observed for US Medicare patients who died between 1993 and 1996 [2], except that Earle et al. [2] reported a significantly increased trend in frequent hospitalizations coupled with an insignificant change in >14-day hospital stays. However, even though the sample in that study [2] comprised patients who died within 1 year of cancer diagnosis, the extent of aggressive care for every indicator was less than in our study. Furthermore, the extent of aggressiveness found in our study for each indicator was greater than that reported for intensive chemotherapy [2, 10, 12, 23], frequent ER visits [2, 12], frequent hospitalizations [2, 12], 14-day hospital stays [2, 12], and ICU admissions [2, 10, 12] in the last month of life (all rates were compared for the whole study cohort).

Our study results indicate that, from 2000 to 2006, use of hospice services increased and use of CPR, intubation, and mechanical ventilation decreased in the last month of life. These trends reflect actions of the Taiwan government in 2000. The hospice movement has been slowly adopted in Taiwan since 1983, primarily due to the initiative of professional societies and charities [24]. On 7 June2000, the government passed the Hospice-Palliative Care Act to allow advanced medical directives to forgo CPR and to legalize palliative care for terminally ill patients [25]. Soon after this legislation was enacted, the DOH facilitated utilization of hospice services by issuing new NHI reimbursement regulations effective from 1 July 2000 to provide inpatient hospice care to terminally ill cancer patients. Correspondingly, from 2000 to 2006, use of hospice services in Taiwan increased 2.59-fold and use of CPR, intubation, and mechanical ventilation decreased by factors of 0.56, 0.70, and 0.90, respectively.

Our results also show a pattern of increasingly aggressive care in the last month of life in conjunction with a significant increase in hospice use, as previously observed [2]. These findings are consistent with the suggestion that hospice utilization adds to the aggressive care received by cancer patients near the end of life rather than substituting for intensive management [2, 26, 27]. In support of this proposition, we found that cancer patients who received hospice care were 1.56 times more likely to have been referred to hospice in the last 3 days of life (95% CI 1.49–1.63, P < 0.0001) if they had received at least one of the following treatments: intensive chemotherapy, frequent ER visits and hospitalizations, ICU care, CPR, intubation, or mechanical ventilation in the last month of life.

Despite evidence of an increasing trend in 2000–2006 toward use of hospice services in the last year of life among Taiwanese cancer patients, hospice services are by no means widely used in Taiwan. Averagely, one in seven Taiwanese cancer patients received hospice care before they died, which is far below the proportion in other countries. In the United States, in 2002, two-thirds of Medicare cancer patients [28] had used a hospice service before they died. In England, the rate of using inpatient hospices and hospice home care in 2003 was estimated to be 18.0% and 70% [29], respectively. The proportion of terminally ill cancer patients receiving hospice services was 66%, 67%, and 68.2% in Singapore [30], Hong Kong [31], and South Australia [32], respectively. Furthermore, the rates of referrals to hospice care in the last 3 days of life (16.77%–18.73%) in our study were higher than those previously reported (13.8%–17.0% and 4.2%) [2, 12].

The majority of cancer patients in our study (59.11%–65.40%) died in a hospital, despite the cultural significance of dying at home and the preference of most terminally ill cancer patients in Taiwan to die at home [21]. Our data for the proportion of cancer patients dying in a hospital are higher than those recently documented worldwide. Percentages of hospital deaths for cancer patients have been reported as 34.6% in Italy [33], 37% in the United States [34], 38.8% in Korea [35], 40% in Canada [36], 47% in South East England [37], and 56.4% in South Australia [38]. Furthermore, in contrast to the decreasing trend observed by Earle et al. [2], our study found an increasing propensity to die in a hospital. Such a high incidence of hospital deaths may be associated with the substantial demands for providing end-of-life care at home [39] and limited use of hospice services, especially for hospice home care services [less than one-third of those who used hospice services received it at home (data not shown)].

The strength of our study lies in its population-based sample, which encompassed all age and cancer groups of Taiwan. Similar issues have been explored only in elderly patients (>65 years) [2], selected cancer groups [2, 12], or limited geographic areas [2, 10, 12]. However, our study has several limitations. First, the preferences for end-of-life care and place of death of terminally ill cancer patients were not addressed. Secondly, the influence of available health-care resources and hospital characteristics on the aggressiveness of care provided at the end of life was not explored. Thirdly, our study used administrative data to assess quality of care, which is not always a precise approach because it depends on the quality of coding and important clinical features are frequently not collected in administrative data. People who die cannot always be considered terminally ill. The appropriateness of aggressive care or adequate proportion of patients receiving hospice services will depend on an accurately defined population of patients dying from complications of curative cancer treatment or dying of cancer, respectively [40]. The appropriateness of aggressive care identified in our study can only serve as a red flag for areas where quality of care needs to be further investigated. Furthermore, the associations between aggressive treatment and other health outcomes, such as quality of life and satisfaction with care, were not explored in our study.

Despite these limitations, results from this study are informative. By identifying areas in which hospitals seem to inadequately manage end-of-life care, these results can facilitate national policy makers’ and health-care professionals’ efforts to improve care for cancer patients near the end of life. The quality of care for cancer decedents identified in our study was substantially inferior to that previously reported [2, 10, 12] and to that recommended by Earle et al. [9] as benchmarks for not providing overly aggressive care near the end of life. The greater extent of aggressive cancer care observed for Taiwanese cancer decedents than for cancer patients who died in other countries [2, 10, 12, 23] may result from the characteristics of universal insurance coverage, almost unlimited comprehensive services provided, and the reimbursement rules of pay-for-service inherent in Taiwan's NHI. Alternatively, such patterns of end-of-life cancer care may be due to limited dissemination of hospice/palliative care philosophy and the reluctance of health-care professionals to disclose prognosis and discuss a transition from curative to palliative care. Indeed, terminally ill cancer patients and their family caregivers in Taiwan have frequently reported that health-care professionals inadequately communicate about prognosis and the goals of future care [41]. However, when considering the goals of future end-of-life care, comfort care was most preferred by terminally ill cancer patients in Taiwan [42]. Exploring and obtaining knowledge of specific end-of-life care preferences of a cancer patient will help health-care professionals tailor their care to meet the patient's actual, rather than his/her inferred needs and to limit the use of futile aggressive treatment at the end of life.

Despite efforts by the government, professional societies, and charities, many cancer patients in Taiwan who could potentially benefit from hospice care do not receive it in time to ameliorate the dying process. Research is warranted to investigate factors influencing hospice utilization (especially for hospice home care) and the timing of hospice referrals. Knowledge derived from such research may provide insights into developing effective interventions to overcome identified barriers to hospice utilization. National policy makers and health-care professionals should increase support for families to provide end-of-life care at home, thereby fulfilling most patients’ wishes to die at home and benefiting both patients and society.

Our study results indicate that Taiwanese cancer patients increasingly received aggressive care in their last month of life, with a small proportion of decedents enrolled in hospice care in 2000–2006. With the enactment of global budgeting to limit inpatient hospital care in 2003, a trend developed toward fewer frequent hospitalizations coupled with an accelerated increase of frequent ER visits in the last month of life for cancer patients in 2004–2006. Cancer patients’ use of hospice services in Taiwan may be expanded by a 2005 DOH policy to integrate hospice services into acute care. The impact of these policies should be evaluated by continually monitoring the quality indicators identified in our study for the impact on patterns of care for cancer patients near the end of life in Taiwan.

funding

Bureau of Health Promotion, Department of Health, Taiwan, R.O.C. (DOH96-HP-1510); National Health Research Institute (NHRI-EX97-9406PC).

References

1.
Barnato
AE
McClellan
MB
Kagay
CR
Garber
AM
Trends in inpatient treatment intensity among Medicare beneficiaries at the end of life
Health Serv Res
 , 
2004
, vol. 
39
 (pg. 
363
-
375
)
2.
Earle
CC
Neville
BA
Landrum
MB
, et al.  . 
Trends in the aggressiveness of cancer care near the end of life
J Clin Oncol
 , 
2004
, vol. 
22
 (pg. 
315
-
321
)
3.
Hogan
C
Lunney
J
Gabel
J
Lynn
J
Medicare beneficiaries’ costs of care in the last year of life
Health Aff
 , 
2001
, vol. 
20
 (pg. 
188
-
195
)
4.
Lubitz
JD
Riley
G
Trends in Medicare payments in the last year of life
N Engl J Med
 , 
1993
, vol. 
328
 (pg. 
1092
-
1096
)
5.
Scitovsky
AA
“The high cost of dying” revisited
Milbank Q
 , 
1994
, vol. 
72
 (pg. 
561
-
591
)
6.
Fisher
ES
Wennberg
DE
Stukel
TA
, et al.  . 
The implications of regional variations in Medicare spending. Part 2: health outcomes and satisfaction with care
Ann Intern Med
 , 
2003
, vol. 
138
 (pg. 
288
-
298
)
7.
Teno
JM
Mor
V
Ward
N
, et al.  . 
Bereaved family member perceptions of quality of end-of-life care in U.S. regions with high and low usage of intensive care unit care
J Am Geriatr Soc
 , 
2005
, vol. 
53
 (pg. 
1905
-
1911
)
8.
Earle
CC
Park
ER
Lai
B
, et al.  . 
Identifying potential indicators of the quality of end-of-life cancer care from administrative data
J Clin Oncol
 , 
2003
, vol. 
21
 (pg. 
1133
-
1138
)
9.
Earle
CC
Neville
BA
Landrum
MB
, et al.  . 
Evaluation claims-based indicators of the intensity of end-of-life cancer care
Int J Qual Health Care
 , 
2005
, vol. 
17
 (pg. 
505
-
509
)
10.
Barbera
L
Paszat
L
Chartier
C
Indicators of poor quality end-of-life cancer care in Ontario
J Palliat Care
 , 
2006
, vol. 
22
 (pg. 
12
-
17
)
11.
Barbera
L
Paszat
L
Qiu
F
End-of-life care in lung cancer patients in Ontario: aggressiveness of care in the population and a description of hospital admissions
J Pain Symptom Manage
 , 
2008
, vol. 
35
 (pg. 
267
-
274
)
12.
Grunfeld
E
Lethbridge
L
Dewar
R
, et al.  . 
Towards using administrative databases to measure population-based indicators of quality of end-of-life care: testing the methodology
Palliat Med
 , 
2006
, vol. 
20
 (pg. 
769
-
777
)
13.
Braga
S
Miranda
A
Fonseca
R
, et al.  . 
The aggressiveness of cancer care in the last three months of life: a retrospective single centre analysis
Psycho-oncology
 , 
2007
, vol. 
16
 (pg. 
863
-
868
)
14.
Department of Health
Health Expenditures by Disease Category
  
http://www.doh.gov.tw/statistic/data/graphic/2005/graphic/graphic.doc (10 February 2008, date last accessed)
15.
Department of Health
Statistics of Causes of Death, Volume I
  
16.
Bureau of National Health Insurance
NHI Profile
  
17.
Bureau of Health Promotion
Cancer Registry Annual Report, 2003 Taiwan
  
http://crs.cph.ntu.edu.tw/crs_c/annual.html (10 February 2008, date last accessed)
18.
Reisfield
GM
Wallace
SK
Munsell
MF
, et al.  . 
Survival in cancer patients undergoing in-hospital cardiopulmonary resuscitation: a meta-analysis
Resuscitation
 , 
2006
, vol. 
71
 (pg. 
152
-
160
)
19.
Groeger
JS
White
P
Nierman
DM
, et al.  . 
Outcome for cancer patients requiring mechanical ventilation
J Clin Oncol
 , 
1999
, vol. 
17
 (pg. 
991
-
997
)
20.
Lin
HC
Lin
CC
A population-based study on the specific locations of cancer deaths in Taiwan, 1997–2003
Support Care Cancer
 , 
2007
, vol. 
15
 (pg. 
1333
-
1339
)
21.
Tang
ST
Liu
TW
Lai
MS
McCorkle
R
Discrepancy in the preferences of place of death between terminally ill cancer patients and their primary family caregivers in Taiwan
Soc Sci Med
 , 
2005
, vol. 
61
 (pg. 
1560
-
1566
)
22.
Armitage
P
Tests for linear trends in proportions and frequencies
Biometrics
 , 
1955
, vol. 
11
 (pg. 
375
-
386
)
23.
Emanuel
EJ
Young-Xu
Y
Levinsky
NG
, et al.  . 
Chemotherapy use among Medicare beneficiaries at the end of life
Ann Intern Med
 , 
2003
, vol. 
138
 (pg. 
639
-
644
)
24.
Lai
YL
Su
WH
Palliative medicine and the hospice movement in Taiwan
Support Care Cancer
 , 
1997
, vol. 
5
 (pg. 
348
-
350
)
25.
Department of Health Act of Palliative Care
 
26.
Campbell
DE
Lynn
J
Louis
TA
Dholakia
R
Medicare program expenditures associated with hospice use
Ann Intern Med
 , 
2004
, vol. 
140
 (pg. 
269
-
277
)
27.
Emanuel
EJ
Cost savings at the end of life. What do the data show?
JAMA
 , 
1996
, vol. 
275
 (pg. 
1907
-
1914
)
28.
Connor
SR
Elwert
F
Spence
C
, et al.  . 
Geographic variation in hospice use in the United States in 2002
J Pain Symptom Manage
 , 
2007
, vol. 
34
 (pg. 
277
-
285
)
29.
Hospice information
Hospice and Palliative Care Facts and Figures 2005
  
30.
Goh
CR
The Asia Pacific Hospice Palliative Care Network: a network for individuals and organizations
J Pain Symptom Manage
 , 
2002
, vol. 
24
 (pg. 
128
-
133
)
31.
Tse
DM
Chan
KS
Lam
WM
, et al.  . 
The impact of palliative care on cancer deaths in Hong Kong: a retrospective study of 494 cancer deaths
Palliat Med
 , 
2007
, vol. 
21
 (pg. 
425
-
433
)
32.
Hunt
RW
Fazekas
BS
Luke
CG
, et al.  . 
The coverage of cancer patients by designated palliative services: a population-based study, South Australia, 1999
Palliat Med
 , 
2002
, vol. 
16
 (pg. 
403
-
409
)
33.
Beccaro
M
Costantini
M
Rossi
PG
, et al.  . 
Actual and preferred place of death of cancer patients. Results from the Italian survey of the dying of cancer (ISDOC)
J Epidemiol Community Health
 , 
2006
, vol. 
60
 (pg. 
412
-
416
)
34.
Flory
J
Young-Xu
Y
Gurol
I
, et al.  . 
Place of death: U.S. trends since 1980
Health Aff
 , 
2004
, vol. 
23
 (pg. 
194
-
200
)
35.
Yun
YH
Lim
MK
Choi
KS
, et al.  . 
Predictors associated with the place of death in a country with increasing hospital deaths
Palliat Med
 , 
2006
, vol. 
20
 (pg. 
455
-
461
)
36.
Neutel
CI
Bishop
ML
Harper
SD
Gaudette
LA
Proportion of cancer deaths occurring in hospital, Canada, 1994-2000
Can J Public Health
 , 
2005
, vol. 
96
 (pg. 
264
-
268
)
37.
Davies
E
Linklater
KM
Jack
RH
, et al.  . 
How is place of death from cancer changing and what affects it? Analysis of cancer registration and service data
Br J Cancer
 , 
2006
, vol. 
95
 (pg. 
593
-
600
)
38.
Foreman
LM
Hunt
RW
Luke
CG
Roder
DM
Factors predictive of preferred place of death in the general population of South Australia
Palliat Med
 , 
2006
, vol. 
20
 (pg. 
447
-
453
)
39.
Tang
ST
Li
CY
Liao
YC
Factors associated with depressive distress among Taiwanese family caregivers of cancer patients at the end of life
Palliat Med
 , 
2007
, vol. 
21
 (pg. 
249
-
257
)
40.
Gagnon
B
Mayo
NE
Laurin
C
, et al.  . 
Identification in administrative databases of women dying of breast cancer
J Clin Oncol
 , 
2006
, vol. 
24
 (pg. 
856
-
862
)
41.
Tang
ST
Liu
TW
Lai
MS
, et al.  . 
Congruence of knowledge, experiences and preferences for disclosure of diagnosis and prognosis between terminally-ill cancer patients and their family caregivers in Taiwan
Cancer Invest
 , 
2006
, vol. 
24
 (pg. 
360
-
366
)
42.
Tang
ST
Liu
TW
Lai
MS
, et al.  . 
Concordance of preferences for end-of-life care between terminally ill cancer patients and their family caregivers in Taiwan
J Pain Symptom Manage
 , 
2005
, vol. 
30
 (pg. 
510
-
518
)