Abstract
Background
In Taiwan, as the population ages, palliative care services (PCS) have expanded significantly to include comprehensive benefit plans for critically ill individuals, supported by reimbursements from the National Health Insurance program. However, incorporating palliative care into the medical management of these patients presents several challenges. We aim to evaluate the effects of palliative care interventions on medical resources in end-of-life scenarios, to promote earlier palliative care access and provide high-quality healthcare services for patients.
Methods
A total of 2202 patients were included in this study. Primary diagnosis and referral for PCS were assessed using ICD-10 and HNI code. All study subjects were divided into three groups: patients who did not receive PCS (no-PCS), patients who received PCS before their final hospital admission (PCS-before), and patients who received PCS after their final admission (PCS-after). We evaluated (i) the effects of PCS on eight medical resource utilization outcomes within the 30 days preceding death and (ii) the effects of early intervention on two major diseases.
Results
Initiating PCS before a patient’s last hospital admission was associated with less aggressive medical interventions in the 30 days before death, including reduced length of intensive care unit (ICU) [odds ratio (OR) = 0.25], and rates of endotracheal intubation (OR = 0.12), respiratory ventilator support (OR = 0.20), cardiopulmonary resuscitation (OR = 0.18), and blood transfusion (OR = 0.65). Among patients with cancer and lung diseases, those who received PCS prior to their final hospitalization of over 14 days experienced reduced hospitalization duration (OR = 0.52 and 0.24, respectively). Patients with lung disease also had significantly lower odds of ICU stays (OR = 0.44) and respiratory ventilation (OR = 0.33).
Conclusion
The timing of palliative care intervention critically impacts on duration of hospitalization and ICU stay and the need for intubation procedures or cardiopulmonary resuscitation. The findings can help the government and medical providers in developing comprehensive palliative care policies and programs to improve care quality and patient rights.
Keywords: palliative care, end-of-life, intensive care unit, endotracheal intubation, respiratory ventilator, transfusion
Introduction
Palliative care services (PCS) can enhance well-being for patients nearing the end of life, in addition to reducing healthcare burden and mitigating inefficiencies, particularly in regions with limited medical resources [1, 2]. In Asian countries such as Taiwan, Japan, and China, the demand for palliative care is growing [3]. However, according to a World Health Organization report, only 50% of countries have integrated palliative care into their national policies on noncommunicable diseases [4]. The key distinction between palliative care and general medical care lies in its approach to death, aiming to shift from aggressive treatment to active symptom management. Palliative care focuses on enhancing quality of life by controlling symptoms rather than curing the illness. Typically, patients are those with a terminal prognosis and a life expectancy of 6 months or less, based on medical assessments. The palliative care team strives to alleviate symptoms and maximize comfort for these patients, ensuring that even in their final months of life, they receive compassionate care. In Taiwan, where the palliative care program under national health insurance was only introduced within the last decade, just 62.3% of cancer patients received PCS before death in 2021. In contrast, the rate for noncancer patients was less than 30% [5].
PCS can lead to reductions in medical costs and aggressive treatment such as cardiopulmonary resuscitation (CPR) and respiratory ventilation [6–9]. Key metrics such as length of hospital stay, CPR, length of ICU (intensive care unit) stay, endotracheal intubation, respiratory ventilation, and blood transfusions are commonly used to assess the quality of end-of-life care [10, 11]. The timing of palliative care intervention may influence outcomes such as duration of hospitalization and ICU stay and the need for intubation procedures or CPR. Additionally, patient attitudes and clinical health services integration can affect PCS efficacy. Given that extended hospitalization and medical resource use can place a heavy burden on healthcare systems, the present study evaluated the effectiveness of PCS in end-of-life scenarios and the influence of intervention timing on medical resource use. Our findings provide valuable insights into how palliative care, especially early referral, can reduce medical futility and improve the quality of healthcare delivery.
Methods
Study design
We conducted a retrospective study at a teaching hospital in Taiwan that handles over 30 000 inpatients and nearly one million outpatient visits annually. Individuals aged 21 years and older who died between 2019 and 2021 were included. After applying exclusion criteria, a total of 2202 subjects were divided into three groups: patients who did not receive PCS (no-PCS), patients who received PCS before their final hospital admission (PCS-before), and patients who received PCS after their final admission (PCS-after). All electronic medical records related to patient care within the year preceding death were reviewed.
Data collection
The enrollment process is illustrated in Fig. 1. Among the 2202 patients included, a total of 1461 (66.4%) received PCS near the end of life. Within this group, 449 patients (20.4%) received PCS before their final hospital admission and 1012 patients (46.0%) after. National Health Insurance (NHI) codes used to identify PCS included P4401B, 02020B, 05601B, 05602B, 05603B, 05362C, and 05364C. Major diseases were defined using the following “International Classification of Diseases, 10th Revision (ICD-10)” codes: dementia (F01-F03), Alzheimer’s disease (G30-G32), stroke (I60-I69), heart failure (I50), chronic obstructive pulmonary disease (J40-J47), respiratory system diseases (J00-J99), liver diseases (K70-K77), acute kidney disease (N17), chronic kidney disease (N18 and N19), trauma or bone fracture (S00-S99), sepsis (A40 and A41), and cancer/neoplasms (C code).
Figure 1.
Flowchart of the patient enrollment process in the study.
Indicators of medical resource use
We examined eight indicators of medical resource use within 30 days before death [2, 12]: (I) emergency room (ER) visits, (II) hospitalization lasting over 14 days, (III) hospitalization lasting over 30 days, (iv) ICU stay, (v) endotracheal intubation, (vi) respiratory ventilation, (VII) CPR, and (viii) blood transfusion. We assessed the effectiveness of different forms of PCS, including palliative referral, family palliative care consultations, and in-hospital palliative care. Additionally, we analyzed the length of hospitalization before death and the timing of PCS initiation.
Statistical analysis
We used multiple logistic regression to explore the effects of PCS initiated before or after final hospitalization on medical resource use at the end of life. The strengths of the relationships were expressed as odds ratios (ORs) with corresponding 95% confidence intervals (CIs). One-way ANOVA and chi-squared tests were performed to assess homogeneity across the three groups. Within-group comparisons were made using Bonferroni correction t-tests. The Mann–Whitney U test compared the medians of multiple groups. All analyses were conducted using SAS version 9.3 (SAS Institute Inc, Cary, NC). Statistical significance was set at P < .05.
Abbreviations
PCSs: Palliative care services; NHI: National health insurance; ER: Emergency room; ICU: Intensive care unit; CPR: Cardiopulmonary resuscitation; LOS: Length of hospital stay; OR: Odds ratio; CI: Confidence interval.
Results
Comparison of the baseline characteristics of the study subject
Table 1 summarizes the demographic characteristics of the 2202 study participants. Among whom 1264 (57.4%) were male and 938 (42.6%) were female. The three subgroups significantly (P < .001) differed in age. The average age of the no-PCS group was older (76.5 ± 13.4 years) than that of the PCS-before (66.8 ± 13.7 years) and PCS-after (71.4 ± 14.7 years) groups. The PCS-before group had a high proportion of patients aged 41–60 years (32.3%) and 61–80 years (47.2%), and the PCS-after group had a high proportion of patients aged 61–80 years (44.8%) and >80 years (31.7%).
Table 1.
Demographic characteristic of the 2202 study participants
| Characteristic | Total, n = 2202 |
No-PCS, n = 741 (33.6%) |
PCS-before their final admission, n = 449 (20.4%) |
PCS-after their final admission, n = 1012 (46.0%) |
P |
|---|---|---|---|---|---|
| Gender, n (%) | |||||
| Male | 1264 (57.4%) | 439 (59.2%) | 254 (56.6%) | 571 (56.4%) | .460 |
| Female | 938 (42.6%) | 302 (40.8%) | 195 (43.4%) | 441 (43.6%) | |
| Age, mean ± SD | 72.2 ± 14.8 | 76.5 ± 13.4 a | 66.8 ± 13.7 b | 71.4 ± 14.7 C | <.001 |
| Age Group, n (%) | |||||
| 21–40 | 44 (2.0%) | 10 (1.4%) | 10 (2.2%) | 24 (2.4%) | <.001 |
| 41–60 | 466 (20.3%) | 87 (11.7%) | 145 (32.3%) | 214 (21.2%) | |
| 61–80 | 969 (44.0%) | 304 (41.0%) | 212 (47.2%) | 453 (44.8%) | |
| >80 | 743 (33.7%) | 340 (45.9%) | 82 (18.3%) | 321 (31.7%) | |
| Palliative support, n (%) | |||||
| Palliative consultant period before death | |||||
| 1–3 months before death | 1350 (92.4%) | 340 (75.7%) | 1010 (99.8%) | <.001 | |
| 4–6 months before death | 51 (3.5%) | 50 (11.1%) | 1 (0.1%) | ||
| 7–12 months before death | 60 (4.1%) | – | 59 (13.1%) | 1 (0.1%) | |
| Comorbidities, n (%) | |||||
| Lung disease | 1350 (61.3%) | 553 (74.6%) | 201 (44.8%) | 596 (58.9%) | <.001 |
| Cancer | 1179 (53.4%) | 148 (20.0%) | 392 (87.3%) | 639 (63.1%) | <.001 |
| Sepsis | 603 (27.4%) | 289 (39.0%) | 75 (16.7%) | 239 (23.6%) | <.001 |
| Chronic kidney disease | 581 (26.4%) | 249 (33.6%) | 69 (15.4%) | 263 (26.0%) | <.001 |
| Bone fracture | 490 (22.3%) | 212 (28.6%) | 76 (16.9%) | 202 (20.0%) | <.001 |
| COPD | 473 (21.5%) | 188 (25.4%) | 76 (16.9%) | 209 (20.7%) | .002 |
| Acute kidney disease | 403 (18.3%) | 172 (23.2%) | 64 (14.3%) | 167 (16.5%) | <.001 |
| Dementia | 380 (17.3%) | 180 (24.3%) | 48 (10.7%) | 152 (15.0%) | <.001 |
| Stroke/Alzheimer | 347 (15.8%) | 191 (25.8%) | 26 (5.8%) | 130 (12.9%) | <.001 |
| Heart failure | 290 (13.2%) | 156 (21.1%) | 17 (3.8%) | 117 (11.6%) | <.001 |
| Liver disease/cirrhosis | 272 (12.4%) | 67 (9.0%) | 52 (11.6%) | 153 (15.1%) | .001 |
| Medical resource utilization, n (%) | |||||
| ER visits in their last month | 1577 (71.6%) | 558 (75.3%) | 339 (75.5%) | 680 (67.2%) | <.001 |
| Hospitalization lasting 14–30 days | 1117 (50.7%) | 308 (41.6%) | 191 (42.5%) | 618 (61.1%) | <.001 |
| Hospitalization lasting over 30 days | 399 (18.1%) | 112 (15.1%) | 53 (11.8%) | 234 (23.1%) | <.001 |
| ICU stays | 991 (45.0%) | 518 (69.9%) | 87 (19.4%) | 386 (38.1%) | <.001 |
| Endotracheal intubation | 219 (10.0%) | 143 (19.3%) | 14 (3.1%) | 62 (6.1%) | <.001 |
| Respiratory ventilator | 760 (34.5%) | 413 (55.7%) | 57 (12.7%) | 290 (28.7%) | <.001 |
| CPR | 111 (5.0%) | 76 (10.3%) | 10 (2.2%) | 25 (2.5%) | <.001 |
| Blood transfusion | 1289 (58.5%) | 457 (61.7%) | 220 (49.0%) | 612 (60.5%) | <.001 |
COPD: Chronic obstructive pulmonary disease.
Continuing variables were analyzed using the ANOVA method, while categorical variables were assessed using the chi-squared test. Statistical significance was set at P < .05. Significant comparisons are indicated as follows: PCS-before group vs. no-PCS group; PCS-after group vs. no-PCS group; PCS-before group vs. PCS-after group.
In the PCS-before group, palliative consultation or care was initiated 1–3 months before death for 75.7% of patients, 4–6 months before death for 11.1% of patients, and 7–12 months before death for 13.1% of patients. Across all groups, the two most common comorbidities were lung disease (61.3%) and cancer (53.4%). Among patients with cancer, 87.3% were in the PCS-before group, whereas patients with other chronic diseases, such as heart failure, stroke, or Alzheimer’s, were less likely to receive PCS either before or after final hospitalization. In contrast, initiating PCS during final hospitalization was more common for patients with lung disease (58.9%) and cancer (63.1%). For other diseases, the prevalence of PCS initiation after final hospitalization ranged from 11.6% to 26.0%.
Aggressive care and medical resource use among PCS groups
Table 1 shows the levels of medical resource use of each group. ER visits differed significantly among the three groups (P < .001). The PCS-before group used significantly fewer medical resources relative to both other groups (P < .001), including lower rates of hospital stays longer than 30 days (11.8%), ICU admissions (19.4%), endotracheal intubation (3.1%), respiratory ventilation (12.7%), CPR (2.2%), and blood transfusion (49.0%). Patients in the PCS after group also used fewer medical resources overall, including fewer ER visits (67.2%), shorter ICU stays (38.1%), and a reduced need for endotracheal intubations (6.1%), respiratory ventilation (28.7%), and CPR (2.5%). Table 2 shows that no-PCS group received more blood transfusions (median of 2 units) than both the PCS-before and PCS-after groups (P < .001).
Table 2.
Effect of PCS on medical resource utilization
| No-PCS, n = 741 (33.6%) | PCS-before their final hospital admission, n = 449 (20.4%) | PCS-after their final hospital admission, n = 1012 (46.0%) | |||||
|---|---|---|---|---|---|---|---|
| Indicators | Median | (q1, q3) | Median | (q1, q3) | Median | (q1, q3) | P |
| Number of ER visits in last month | 1.0 | (1.0, 1.0) | 1.0 | (1.0, 1.0) | 1.0 | (1.0, 1.0) | <.001 |
| Days of hospitalization before death | 11.0 | (5.0, 22.0) | 11.0 | (6.0, 20.0) | 17.0 | (10.0, 28.0) | <.001 |
| ICU stay days | 4.0 | (0.0, 10.0) | 0.0 | (0.0, 0.0) | 0.0 | (0.0, 7.0) | <.001 |
| Blood transfusion | 2.0 | (0.0, 4.0) | 0.0 | (0.0, 4.0) | 1.0 | (0.0, 4.0) | <.001 |
| Days of their final hospitalization (mean ± SD) | 16.6 ± 17.3 | 15.0 ± 12.7 | 23.1 ± 22.2 | ||||
| PCS intervention after admission (mean ± SD) | 10.2 ± 15.4 | ||||||
Data presented as the median (25th and 75th percentiles). Statistical significance (P < .05) was determined using the independent-sample Kruskal–Wallis test.
Medical resource use in two major diseases among patients who received PCS
In Taiwan, patients with cancer have been most likely to receive PCS, although PCS for patients with other chronic diseases has begun to surge. Table 3 focuses specifically on the 1461 patients who received palliative care for two main diseases—cancer and lung disease. Among patients with cancer, the proportion of hospital stays exceeding 14 days (41.0% vs. 56.5%, P < .001) and 30 days (9.6% vs. 16.4%, P = .021) significantly differed between those who received PCS before or after final hospital admission. Patients with lung disease in the PCS-before group experienced significant reductions (P < .005) in-hospital stays longer than 14 days (39.5% vs. 69.2%) and 30 days (10.5% vs. 33.2%), ICU stays (47.4% vs. 67.1%), and endotracheal intubation (31.6% vs. 52.8%).
Table 3.
Medical resource utilization among patients who received PCS by disease group
| Cancer | Lung Disease | Cancer & Lung Disease | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Characteristic | Receiving PCS, n = 1461 | PCS-before final admission, 229 (41.0%) | PCS-after final admission, 329 (59.0%) | P | PCS-before final admission, 38 (11.7%) | PCS-after final admission, 286 (88.3%) | P | PCS-before final admission, 163 (34.5%) | PCS-after final admission, 310 (65.5%) | P |
| Gender, n (%) | ||||||||||
| Male | 825 (56.5%) | 122 (53.3%) | 173 (52.6%) | .872 | 18 (47.4%) | 164 (57.3%) | .244 | 107 (65.6%) | 195 (62.9%) | .555 |
| Female | 636 (43.5%) | 107 (46.7%) | 156 (47.4%) | 20 (52.6%) | 122 (42.7%) | 56 (34.4%) | 115 (37.1%) | |||
| Age Group, n (%) | ||||||||||
| 21–40 | 34 (2.3%) | 6 (2.6%) | 7 (2.1%) | .200 | 0 (0.0%) | 5 (1.7%) | .776 | 4 (2.5%) | 12 (3.9%) | .153 |
| 41–60 | 359 (24.6%) | 84 (36.7%) | 104 (31.6%) | 4 (10.5%) | 29 (10.1%) | 50 (30.7%) | 71 (22.9%) | |||
| 61–80 | 665 (45.5%) | 114 (49.8%) | 162 (49.2%) | 11 (28.9%) | 96 (33.6%) | 83 (50.9%) | 158 (51.0%) | |||
| >80 | 403 (27.6%) | 25 (10.9%) | 56 (17.0%) | 23 (60.5%) | 156 (54.5%) | 26 (16.0%) | 69 (22.3%) | |||
| Medical resource utilization, n (%) | ||||||||||
| ER visits in their last month | 1019 (69.7%) | 170 (74.2%) | 222 (67.5%) | .086 | 34 (89.5%) | 188 (65.7%) | .003 | 120 (73.6%) | 197 (63.5%) | .027 |
| Hospitalization lasting 14-30 days | 809 (55.4%) | 94 (41.0%) | 186 (56.5%) | <.001 | 15 (39.5%) | 198 (69.2%) | <.001 | 75 (46.0%) | 189 (61.0%) | .002 |
| Hospitalization lasting over 30 days | 287 (19.6%) | 22 (9.6%) | 54 (16.4%) | .021 | 4 (10.5%) | 95 (33.2%) | .004 | 26 (16.0%) | 76 (24.5%) | .031 |
| ICU stays | 473 (32.4%) | 18 (7.9%) | 28 (8.5%) | .784 | 18 (47.4%) | 192 (67.1%) | .017 | 49 (30.1%) | 122 (39.4%) | .046 |
| Endotracheal intubation | 76 (5.2%) | 4 (1.7%) | 5 (1.5%) | .834 | 1 (2.6%) | 22 (7.7%) | .496 | 9 (5.5%) | 29 (9.4%) | .145 |
| Respiratory ventilator | 347 (23.8%) | 7 (3.1%) | 17 (5.2%) | .227 | 12 (31.6%) | 151 (52.8%) | .014 | 38 (23.3%) | 100 (32.3%) | .042 |
| CPR | 35 (2.4%) | 4 (1.7%) | 1 (0.3%) | .165 | 1 (2.6%) | 11 (3.8%) | 1.000 | 5 (3.1%) | 10 (3.2%) | .926 |
| Blood transfusion | 832 (56.9%) | 100 (43.7%) | 172 (52.3%) | .045 | 20 (52.6%) | 187 (65.4%) | .124 | 97 (59.5%) | 195 (62.9%) | .470 |
| PCS intervention after admission (days, mean ± SD) | 6.8 ± 15.0 | 15.8 ± 16.4 | 8.8 ± 13.6 | |||||||
Data are presented as case numbers (frequency %). Statistical significance (P < .05) was determined using the chi-squared test. If the case number was less than 5, Fisher’s exact test was used for analysis.
Length of hospitalization by disease subgroup
Table 4 shows that length of hospitalization for the PCS-before group ranged from 11.0 to 13.0 days across the three disease subgroups, and patients in the PCS-after group had longer hospital stays (15.0, 20.0, and 17.5 days for each subgroup, respectively). In clinical care, patients are more likely to receive aggressive therapies before being referred for PCS, which can extend hospital stays. Patients with lung disease in the PCS-after group had longer ICU stays (median, 8.0 days).
Table 4.
Comparison of PCS impact on hospitalization and medical resource utilization in patients with cancer and lung disease
| Cancer, n = 558 | Lung disease, n = 324 | Cancer and Lung disease, n = 473 | |||||||
|---|---|---|---|---|---|---|---|---|---|
| PCS-before final admission, 229 (41.0%) | PCS-after final admission, 329 (59.0%) | PCS-before final admission, 38 (11.7%) | PCS-after final admission, 286 (88.3%) | PCS-before final admission, 163 (34.5%) | PCS-after final admission, 310 (65.5%) | ||||
| Indicators | Median (q1, q3) | Median (q1, q3) | P | Median (q1, q3) | Median (q1, q3) | P | Median (q1, q3) | Median (q1, q3) | P |
| ER visits in last month (number) | 1.0 (0.0, 2.0) |
1.0 (0.0, 1.0) |
<0.001 | 1.0 (1.0, 1.0) |
1.0 (0.0, 1.0) |
0.022 | 1.0 (0.0, 2.0) |
1.0 (0.0, 1.0) |
<0.001 |
| Hospitalization before death (days) | 11.0 (6.0, 19.0) |
15.0 (9.0, 23.0) |
0.001 | 11.5 (5.0, 19.0) |
20.0 (12.0, 35.0) |
<0.001 | 13.0 (7.0, 22.0) |
17.5 (10.0, 29.0) |
<0.001 |
| ICU stay (days) | 0.0 (0.0, 0.0) |
0.0 (0.0, 0.0) |
0.141 | 0.0 (0.0, 6.0) |
8.0 (0.0, 16.0) |
<0.001 | 0.0 (0.0, 3.0) |
0.0 (0.0, 5.0) |
0.036 |
| Blood transfusion (units) | 0.0 (0.0, 2.0) |
0.0 (0.0, 2.0) |
0.098 | 0.0 (0.0, 3.0) |
1.0 (0.0, 4.0) |
0.178 | 2.0 (0.0, 4.0) |
2.0 (0.0, 4.0) |
0.897 |
Data are presented as the median (quantile 25% and 75%). Statistical significance (P < .05) was determined using the Mann–Whitney U test.
Effects of PCS on medical resources use
Figure 2A shows medical resource use among patients in the PCS-before and PCS-after groups. Logistic regression found significantly lower rates of ICU admission (OR = 0.25, P < .001), endotracheal intubation (OR = 0.12, P < .001), respiratory ventilation (OR = 0.20, P < .001), CPR (OR = 0.18, P < .001), and blood transfusion (OR = 0.65, P = .003) for the PCS-before group. In contrast, patients in the PCS-after group had higher incidence of longer hospital stays over 14 days (OR = 2.39, P < .001) and 30 days (OR = 1.97, P < .001) relative to those in the no-PCS group. Figure 3A and B display outcomes for patients with cancer and lung disease who received PCS. Patients with cancer in the PCS-before group had significantly fewer hospital stays over 14 days (OR = 0.52, P < .001) and 30 days (OR = 0.49, P = .010). Patients with lung disease in the PCS-before group also had reduced rates of hospital stays over 14 and 30 days (OR = 0.24, P < .001 and OR = 0.19, P = .003, respectively), ICU admission (OR = 0.44, P = .030), and respiratory ventilation (OR = 0.33, P = .006). However, no significant differences were observed in endotracheal intubation, CPR, or blood transfusion.
Figure 2.
Forest plot showing the ORs for healthcare resource utilization: (a) No-PCS vs. PCS started before their final admission, and (b) No-PCS vs. PCS started after the last admission. ORs were calculated following a univariate logistic regression analysis for eight independent variables related to healthcare resource utilization. Data are presented as ORs with 95% CIs.
Figure 3.
Forest plot of OR for healthcare resource utilization. (a) Cancer patients had PCS intervention before their final hospitalization compared with PCS intervention after their final hospitalization. (b) Lung disease patients had PCS intervention before their final hospitalization compared with PCS intervention after their final hospitalization. (c) PCS interventions before vs. after last hospitalization in patients with cancer and lung disease. ORs after a univariate logistic regression analysis for eight independent variables related to healthcare resource utilization. Data presented as OR with 95% CI.
Discussion
Statement of principal findings
Palliative care, once primarily focused on patients with cancer, is now increasingly used to manage chronic disorders [13, 14]. Measures such as the need for respiratory ventilation, endotracheal intubation, CPR, blood transfusion, and length of stay in the hospital and ICU are commonly used to assess the quality of end-of-life medical care. We found that patients who received PCS were significantly less likely to undergo CPR relative to those who did not receive PCS. Notably, early initiation of PCS significantly reduced the use of aggressive interventions such as endotracheal intubation (OR = 0.12) and respiratory ventilation (OR = 0.20). These results may also reflect the growing influence of Taiwan’s “Patient Right to Autonomy Act and Hospice Palliative Care Act,” which allows terminally ill patients to forgo CPR through do-not-resuscitate orders and advance care plans. Palliative care has been shown to result in improved pain documentation, fewer ER visits, and reduced hospitalizations [15, 16]. In Taiwan, 10%–30% of cancer patients seek emergency and intensive care services within 1 month before death. Patients’ preferences and the affordability of healthcare under the NHI system may increase the rate of ER visits [17].
Blood transfusions are an essential supportive treatment in both aggressive and palliative medical care [18]. The patient groups in our study significantly differed in transfusion practices. Transfusion prevalence was lower in the PCS-before group (49.0%) relative to the no-PCS group (61.7%), with the latter also receiving a higher average number of blood units. Given that blood is a finite resource, its availability and allocation are critical in the treatment of severe illness [19]. However, the evaluation criteria for blood transfusions in clinical settings are not always applicable to patients receiving palliative care [20]. Anemia is highly prevalent in the ICU, with rates reaching up to 98%, and 27.0%–62.1% of patients with pneumonia experience moderate to mild anemia during their hospital stay [21, 22]. Although some studies have suggested that blood transfusions should not be withheld from patients with terminal cancer receiving palliative care, such decisions are heavily influenced by the treatment goals of both the patient and the medical team [23, 24]. Transfusions may not substantially influence disease recovery or survival in patients with terminally illnesses, which makes addressing the appropriateness of their administration during palliative care essential for avoiding potentially futile use of a limited resource.
On average, patients received PCS approximately 10.2 ± 15.4 days after their hospital admission (Table 2), indicating that the timing of palliative care initiation was likely influenced by patient condition and medical practice rather than standardized protocols. Currently, there is no universally accepted optimal time for initiating palliative care. In our study, palliative care was typically started around 10 days after admission, which may be too late for patients with terminal illnesses who could benefit from earlier end-of-life intervention. Aggressive care practices and medical resource use differ considerably depending on underlying diseases and comorbidities. A JAMA study reported that 40% of patients who died from sepsis while hospitalized were never referred for PCS before admission [25]. Older adults with septicemia tend to require more intensive care, including an increased number of hospital visits and lengthened hospital and ICU stays [26, 27]. Patients with lung disease often receive more aggressive treatments, such as endotracheal intubation and respiratory ventilation, because clinicians may overestimate their survival. Our results indicate that patients with lung disease or severe infections were referred for palliative care less frequently than those with cancer [11.7% (n = 38) vs. 41.0% (n = 229), Table 3]. PCS was initiated later for patients with lung disease (15.8 ± 16.4 days after hospitalization) than those with cancer (6.8 ± 15.0 days), suggesting a potential delay in PCS referral, which could result in missed opportunities for timely intervention.
Strengths and limitations
While this study provides valuable insights into how palliative care, especially early referral, can reduce medical futility and improve the quality of healthcare delivery, there were limitations. Medical records were categorized into only 11 diagnostic codes, which may have introduced bias due to inaccurate coding or overlapping diagnoses for patients with multiple conditions. Additionally, our retrospective approach started from the patient’s death and looked backward. Analysis of medical resource use was confined to the month preceding death. In contrast, early palliative care interventions were only considered within the year before death, although some patients may have received palliative care even earlier. This analysis focused on the length of hospital stay, ventilator and medication use, and emergency interventions, which can vary widely, complicating efforts to retrospectively standardize or accurately assess costs. Cost analysis was not a primary objective of this study and was therefore not included in our findings. Furthermore, socio-demographics characteristics—such as income, education level, and ethnicity—were inaccessible through the hospital’s medical information system. Due to privacy restrictions, data on patients who died in other hospitals or at home were also unavailable for analysis.
Interpretation within the context of the wider literature
Multidisciplinary approaches have shown that palliative care improves end-of-life outcomes [28]. Taiwan’s “Patient Right to Autonomy Act” enables patients to make informed decisions about their healthcare, ensuring that their rights are respected at the end of life and preventing unnecessary medical treatment. The growing recognition and acceptance of PCS by both medical professionals and patients underscore its importance [29, 30].
Implications for policy, practice, and research
Although palliative care in Taiwan has traditionally been focused on patients with cancer, new policies intend to extend insurance coverage for PCS to terminal chronic illnesses. Patients in our study who received PCS experienced fewer aggressive medical interventions, including respiratory assistance and endotracheal intubation, demonstrating the benefits of early palliative care initiation. Future efforts should focus on further integrating PCS into standard medical practice, encouraging earlier referrals, and providing individualized care plans for symptom management and spiritual support for patients and their families.
Conclusion
Unnecessary and aggressive medical interventions can burden families and strain healthcare resources. The timing of palliative care initiation plays a critical role in reducing the use of aggressive medical treatment. Early palliative care intervention is essential to address the holistic needs of patients and their families—encompassing medical, physical, social, and spiritual dimensions—and to ensure that dignity is maintained throughout the duration of terminal illness.
Acknowledgements
The authors would like to thank the palliative care team members for their understanding and continuous support for this project.
Contributor Information
Chia-Chia Lin, Department of Family Medicine, Taichung Tzu-Chi Hospital, Buddhist Tzu-Chi Medical Foundation, No. 88 Fong-Shing Road, Taichung City 42743, Taiwan.
Tsing-Fen Ho, Department of Medical Laboratory Science and Biotechnology, Central Taiwan University of Science and Technology, No. 666 Buzih Road, Taichung City 40601, Taiwan.
Chang-Hung Lin, Center of Quality Management, Taichung Tzu-Chi Hospital, Buddhist Tzu-Chi Medical Foundation, No. 88 Fong-Shing Road, Taichung City 42743, Taiwan.
Nu-Man Tsai, Department of Medical Laboratory and Biotechnology, Chung Shan Medical University, No. 110, Section 1, Jianguo North Road, Taichung 40201, Taiwan; Department of Life-and-Death Studies, Nanhua University, 55 Section 1, Nanhua Road, Dalin Township, Chiayi 622301, Taiwan.
Yu-Hung Kuo, Department of Research, Taichung Tzu-Chi Hospital, Buddhist Tzu-Chi Medical Foundation, No. 88 Fong-Shing Road, Taichung City 42743, Taiwan.
Ju-Huei Chien, Department of Medical Laboratory Science and Biotechnology, Central Taiwan University of Science and Technology, No. 666 Buzih Road, Taichung City 40601, Taiwan; Department of Research, Taichung Tzu-Chi Hospital, Buddhist Tzu-Chi Medical Foundation, No. 88 Fong-Shing Road, Taichung City 42743, Taiwan.
Author contributions
Chia-Chia Lin (Resources, Methodology, Data curation), Chang-Hung Lin (Resources, Methodology, Data curation), Tsing-Fen Ho (Conceptualization, Formal Analysis, Writing—review and editing), Nu-Man Tsai (Review), Yu-Hung Kuo (Investigation, Statistical Analysis), Ju-Huei Chien (Conceptualization, Data Curation, and Writing—original draft preparation). All the authors have read and agreed to the published version of the manuscript.
Conflict of interests
None declared.
Funding
This research was supported by the Ministry of Education of Taiwan (PMN1122369), and the Taichung Tzu- Chi Hospital, Taichung, Taiwan (TTCRD112-27). The funders had no role in study design, data collection and analysis, publication decisions, or manuscript preparation.
Data availability
The authors confirm that all data generated and analyzed during this study are included in this published article. The data sets analyzed during the current study contain patient medical information; they are not publicly available due to the patient protection policy of Tzu Chi Hospital. Inquiries about the study data and protocol will be shared on reasonable request to the corresponding author.
Ethics and other permissions
This is a retrospective observational study. All participants’ data were gathered via the hospital information system, and informed consent could not be obtained individually. For this study, the informed consent was waived by the research ethics committee of Taichung Tzu-Chi Hospital. The data were anonymized and unidentifiable. Ethical approval was obtained from the Research Ethics Committee of Taichung Tzu-Chi Hospital (REC 113–27).
References
- 1. Chang TS, Su YC, Lee CC. Determinants for aggressive end-of-life care for oral cancer patients: a population-based study in an Asian country. Medicine 2015;94:e460. doi: 10.1097/MD.0000000000000460 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Earle CC, Landrum MB, Souza JM et al. Aggressiveness of cancer care near the end of life: is it a quality-of-care issue?. J Clin Oncol 2008;26:3860–6. doi: 10.1200/JCO.2007.15.8253 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Lu Y, Gu Y, Yu W. Hospice and palliative care in China: development and challenges. Asia PacJ Oncol Nurs 2018;5:26–32. doi: 10.4103/apjon.apjon_72_17 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Connor S. Global Atlas of Palliative Care. 2nd edn. London: World Health Organization, 2023. [Google Scholar]
- 5. Ministry of Health and Welfare Annual Report 2021. Health Promotion Administration Ministry of Health and Welfare ROC. [Google Scholar]
- 6. Barbera L, Seow H, Sutradhar R et al. Quality indicators of end-of-life care in patients with cancer: what rate is right?. J Oncol Pract 2015;11:e279–87. doi: 10.1200/JOP.2015.004416 [DOI] [PubMed] [Google Scholar]
- 7. Patrick DL. Measuring and improving the quality of dying and death. Ann Internal Med 2003;139:410–5. doi: 10.7326/0003-4819-139-5_Part_2-200309021-00006 [DOI] [PubMed] [Google Scholar]
- 8. Thongprayoon C, Kaewput W, Petnak T et al. Impact of palliative care services on treatment and resource utilization for hepatorenal syndrome in the United States. Medicines (Basel, Switzerland) 2021;8:21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. 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;21:1133–8. doi: 10.1200/JCO.2003.03.059 [DOI] [PubMed] [Google Scholar]
- 10. Davis MP, Van Enkevort EA, Elder A et al. The influence of palliative care in hospital length of stay and the timing of consultation. Am J Hosp Palliat Care 2022;39:1403–9. doi: 10.1177/10499091211073328 [DOI] [PubMed] [Google Scholar]
- 11. Kyeremanteng K, Gagnon LP, Thavorn K et al. The Impact of palliative care consultation in the ICU on length of stay: a systematic review and cost evaluation. J Intensive Care Med 2018;33:346–53. doi: 10.1177/0885066616664329 [DOI] [PubMed] [Google Scholar]
- 12. Kang DW, Shim YB, Lee EK et al. Healthcare resource utilization and medical costs in patients with terminal cancer during best supportive care. PLoS One 2022;17:e0269565. doi: 10.1371/journal.pone.0269565 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Kim S, Lee K, Kim S. Knowledge, attitude, confidence, and educational needs of palliative care in nurses caring for non-cancer patients: a cross-sectional, descriptive study. BMC Palliat Care 2020;19:105. doi: 10.1186/s12904-020-00581-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Mounsey L, Ferres M, Eastman P. Palliative care for the patient without cancer. Austr J Gen Pract 2018;47:765–9. doi: 10.31128/AJGP-07-18-4625 [DOI] [PubMed] [Google Scholar]
- 15. Hjermstad MJ, Kolflaath J, Løkken AO et al. Are emergency admissions in palliative cancer care always necessary? Results from a descriptive study. BMJ Open 2013;3:e002515.doi: 10.1136/bmjopen-2012-002515 (7 January 2025, date last accessed). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Lin YC, Hsu HK, Lai TS et al. Emergency department utilization and resuscitation rate among patients receiving maintenance hemodialysis. J Formos Med Assoc 2019;118:1652–60. doi: 10.1016/j.jfma.2019.01.007 [DOI] [PubMed] [Google Scholar]
- 17. Lee YH, Chu D, Yang NP et al. Emergency visits among end-of-life cancer patients in Taiwan: a nationwide population-based study. BMC Palliat Care 2015;14:25. doi: 10.1186/s12904-015-0016-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Garraud O. Transfusion at the border of the “intention-to-treat”, in the very aged person and in palliative care: a debate. Transfus Clin Biol 2021;28:367–9. doi: 10.1016/j.tracli.2021.08.350 [DOI] [PubMed] [Google Scholar]
- 19. Thomas J, Jensen L, Nahirniak S et al. Anemia and blood transfusion practices in the critically ill: a prospective cohort review. Heart Lung 2010;39:217–25. doi: 10.1016/j.hrtlng.2009.07.002 [DOI] [PubMed] [Google Scholar]
- 20. Raval JS, Cooling LL. Red blood cell transfusion in palliative care: what are we doing and why are we doing it? Transfusion 2018;58:3–4. doi: 10.1111/trf.14437 [DOI] [PubMed] [Google Scholar]
- 21. Czempik PF, Wilczek D, Herzyk J et al. Hospital-acquired anemia in patients hospitalized in the intensive care unit: a retrospective cohort study. J Clin Med 2022;11:3939. doi: 10.3390/jcm11143939 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Reade MC, Weissfeld L, Angus DC et al. The prevalence of anemia and its association with 90-day mortality in hospitalized community-acquired pneumonia. BMC Pulm Med 2010;10:15. doi: 10.1186/1471-2466-10-15 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Marote S, Marinho J, Silva MC et al. Transfusion practices in patients with advanced cancer: a retrospective study in a palliative care service. Porto Biomed J 2022;7:e195. doi: 10.1097/j.pbj.0000000000000195 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Goksu SS, Gunduz S, Unal D et al. Use of blood transfusion at the end of life: does it have any effects on survival of cancer patients?. Asian Pac J Cancer Prev 2014;15:4251–4. doi: 10.7314/APJCP.2014.15.10.4251 [DOI] [PubMed] [Google Scholar]
- 25. Rhee C, Jones TM, Hamad Y et al. Prevalence, underlying causes, and preventability of sepsis-associated mortality in US acute care hospitals. JAMA Netw Open 2019;2:e187571. doi: 10.1001/jamanetworkopen.2018.7571 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Guillon A, Hermetet C, Barker KA et al. Long-term survival of elderly patients after intensive care unit admission for acute respiratory infection: a population-based, propensity score-matched cohort study. Crit Care 2020;24:384. doi: 10.1186/s13054-020-03100-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Anderson ME, Glasheen JJ, Anoff D et al. Understanding predictors of prolonged hospitalizations among general medicine patients: a guide and preliminary analysis. J Hosp Med 2015;10:623–6. doi: 10.1002/jhm.2414 [DOI] [PubMed] [Google Scholar]
- 28. Tsai YS, Tsai WC, Chiu LT et al. Effectiveness of palliative care before death in reducing emergency care utilization for patients with terminal cancer and trends in the utilization of palliative care from 2005–2018. Healthcare (Basel) 2023;11:2907. doi: 10.3390/healthcare11212907 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Chen IH, Kuo SF, Lin YK et al. Knowledge of and barriers to palliative care perceived by healthcare providers before and after promotion of the patient autonomy act: a cross-sectional study. Int J Environ Res Public Health 2022;19:3884. doi: 10.3390/ijerph19073884 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Wu YL, Yang CY, Lin TW et al. Factors impacting advance decision making and health care agent appointment among Taiwanese urban residents after the passage of patient right to autonomy act. Healthcare (Basel) 2023;11:1478. doi: 10.3390/healthcare11101478 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The authors confirm that all data generated and analyzed during this study are included in this published article. The data sets analyzed during the current study contain patient medical information; they are not publicly available due to the patient protection policy of Tzu Chi Hospital. Inquiries about the study data and protocol will be shared on reasonable request to the corresponding author.