Abstract
Context
Hospitalized patients with functional impairment have higher symptom burden and mortality. Little is known about how increased patient volume and acuity during the coronavirus disease 2019 (COVID-19) pandemic affected access to palliative care among patients with functional impairment.
Objectives
To examine changes in functional status and hospital outcomes among patients receiving inpatient palliative care consultation before, during and after the COVID-19 pandemic.
Methods
We conducted a retrospective, multisite cohort study of all adult patients (≥ 18 years) admitted to four hospitals in New York City, USA, who received inpatient palliative care consultation between March 1, 2019 and February 28, 2022 with documented functional status at the time of consultation measured by Karnofsky Performance Status scale.
Results
Among 13,180 eligible patients identified, patients’ functional status at the time of consultation decreased as palliative care consult volume increased with the onset of the pandemic. Compared to pre-pandemic, there was a statistically significant trend of lower functional status (P < 0.001) and higher in-hospital mortality (P < 0.001) among patients with noncancer and non-COVID-19 diagnoses two years after the pandemic. In contrast, patients with cancer had a statistically significant trend of higher functional status (P < 0.001) and no significant changes in in-hospital mortality over time.
Conclusion
As the healthcare system was stressed with high demand and limited resources, palliative care consultation prioritized highest acuity patients by shifting towards those with lower functional status and higher in-hospital mortality. This shift disproportionately affected noncancer patients. Innovative approaches to ensure upstream palliative care consultation during increased resource constraints are needed.
Key Words: Functional status, palliative care, COVID-19
Key Message
This article compares functional status at the time of palliative care consultation among a retrospective inpatient cohort before and during the COVID-19 pandemic. Increased consultation demands and limited resources shifted palliative care delivery to patients with lowest functional status and highest mortality. This shift disproportionately affected noncancer patients.
Introduction
Hospitalized patients with functional impairment have higher symptom burden and worse clinical outcomes,1, 2, 3, 4 and are more likely to be discharged to postacute care facilities instead of home.5 Accumulating evidence supports that early involvement of palliative care among patients with impaired function and serious illness can improve quality of life, decrease symptom burden, prevent nonbeneficial care, and reduce costs.6, 7, 8, 9, 10 Timely engagement of palliative care is recommended by several major medical societies which also represent illnesses that result in most common causes of death in the United States, including American Academy of Hospice and Palliative Medicine, American Society of Clinical Oncology,11 American Thoracic Society,12 American Heart Association and American Stroke Association.13
Despite a multitude of benefits associated with palliative care, the growing palliative care needs have been faced with significant workforce constraints, with an estimated shortage of more than 6000 hospice and palliative medicine physicians to meet the care demand in the United States even before the coronavirus disease 2019 (COVID-19) pandemic.14, 15, 16 The pandemic further exacerbated workforce shortage by adding significant morbidity and mortality burden to an already stressed healthcare system, amplifying existing disparities in access to healthcare resources.17 , 18 Two years after the first surge of COVID-19 pandemic, even as COVID-19 associated morbidity and mortality decreases, patients continued to avoid or postpone engagement with health care settings out of concerns for COVID-19 exposure,19 , 20 and those with pressing medical needs suffered from ongoing delays in multiple aspects of healthcare services.21, 22, 23, 24
The impact of increased patient volume and acuity due to the COVID-19 pandemic on the operational capacity of inpatient palliative care consult has not been well studied. We hypothesized that inpatient palliative care teams would have to triage the limited workforce towards patients with poorer function and higher risk of death, and in turn, minimize consultation volume capacity to initiate timely interventions for patients earlier in disease trajectory. In this study, we examined shifts in allocation of fixed specialty palliative care resource as the overall demand on the healthcare system increased and compared functional status and hospital outcomes among patients receiving inpatient palliative care consult before, during and after the COVID-19 pandemic.
Methods
Study Design
We conducted a retrospective, multi-site cohort study of all adult patients (aged ≥ 18) admitted to four hospitals within the Mount Sinai Health System (MSHS) who received inpatient specialty palliative care consultation between March 1, 2019 and February 28, 2022 with documented functional status at the time of initial consultation. Participating hospitals included one 1100-bed quaternary academic hospital (MSA) and three mid-sized community hospitals (MSB, MSC, MSD). All four hospitals have well-developed palliative care consultation teams with coverage available 24 hours a day, seven days a week. At each site, there are one-two general palliative care consult teams. Each team consisted of one attending physician, one fellow (MSA only) and one nurse practitioner, in collaboration with social workers and chaplains. Before the COVID-19 pandemic, each MSA team saw about 55–65 patients per month, and each team at the other sites saw about 30–35 patients per month. In addition to these teams, MSA had embedded palliative care clinicians within medical and surgical intensive care units (ICUs), hospital medicine, oncology, cardiology and the emergency department, which collectively saw another 60–70 patients per month. MSA also had a 14-bed inpatient palliative care unit caring for about 40 patients per month, staffed by an attending, a fellow, a nurse practitioner and a social worker. During the peak of the pandemic, the embedded palliative care clinicians models were expanded across MSHS to assist providers in emergency departments, hospital medicine and ICUs. A 24-hour telephonic palliative care consultation program was created to meet the demand that exceeded the in-person programs, and serious illness communication skills training were also implemented to expand the pool of primary palliative care clinicians.25 The study was approved by The Icahn School of Medicine at Mount Sinai Institutional Review Board.
Data Collection
Using health system palliative care clinical database (MSHS-PC), which extracts data from electronic health records consultation template, and health system administrative billing database (MSX), we collected information on patient demographics (i.e., age, sex, and race), serious illness diagnoses (using a modified method of identifying serious illness groups based on ICD-10 codes),26 ICU admission, ICU length of stay, hospital length of stay, discharge disposition and in-hospital death. We defined primary palliative care diagnoses by first using diagnoses (noncancer vs. cancer) documented by palliative care clinicians. If cancer was documented as one of the two palliative care diagnoses, the patient was categorized in the cancer group; otherwise, the patient was categorized as noncancer. Among noncancer diagnoses, we defined patients with COVID-19 using ICD-10 codes. Using MSHS-PC data from the initial palliative care consultation note, we also collected date of consultation and patients’ functional status at the time of consultation. Since March 2019, patients’ functional status at consultation was routinely assessed and documented by palliative care specialists across the participating hospital sites using the Karnofsky Performance Status (KPS) scale.27 KPS is a validated scale to assess functional status in a variety of contexts of palliative care, and correlates with both resource utilization and prognosis at the end of life.27 The KPS scale ranges from 0% to 100%, with 100% indicating normal activity and 0% marking death. In this study, we further categorized patients into mild (KPS 80%–100%), moderate (KPS 50%–70%), severe (KPS 30%–40%) and very severe (KPS 10%–20%) functional impairment.
Statistical Analysis
We performed descriptive analysis of patients’ demographic and clinical characteristics, KPS at time of consultation, and patient outcomes. Continuous variables were reported as means with standard deviations (SD) and medians with interquartile ranges (IQR). Categorical variables were reported as frequencies and percentages. We used Chi-square test for comparisons of categorical variables, and Student's t-test and one-way analysis of variance (ANOVA) for comparisons of continuous variables. We used Jonckheere-Terpstra test, a rank-based nonparametric test, to assess if there was a statistically significant trend between the ordinal independent variable KPS at consultation and the ordinal dependent variable designating time periods of consultation. For sensitivity analysis, we conducted bivariate comparisons between patients excluded due to missing KPS at the time of consultation and the patients included for final analysis. We also examined the trends in number of days between admission to consultation and between consultation to discharge, and hospital and ICU length of stay after excluding patients with hospital length of stay > 30 days. Missing data were handled by listwise deletion. All analyses were conducted using STATA/SE 15.1 (StataCorp LP). Two-sided P values < 0.05 were considered statistically significant.
Results
We identified a total of 13,913 patients who received inpatient palliative care consultation during the study period, among which we excluded 16 (0.1%) patients with missing discharge date or disposition, and 717 (5.2%) patients with missing KPS at the time of consultation. A final sample of 13,180 (94.7%) patients was included in the final analysis (Table 1 ). Patients had a mean age of 70.3 years (SD 16.4) and 50.6% were female. There were 4068 palliative care consults during the year pre-pandemic (March 2019–Feb 2020), 4560 consults during pandemic year one (March 2020–February 2021) and 4552 consults during pandemic year two (March 2021–February 2022). The median KPS assessed at consultation was 40% [IQR 20%–50%], 30% [IQR 20%–50%] and 30% [IQR 20%–50%], respectively. About half of the patients (50.5%) had noncancer diagnoses prompting palliative care consult, as compared to 9% patients with COVID-19 and 40.5% patients with cancer. Patients excluded from the study due to missing KPS were disproportionately older, received palliative care consultation pre-pandemic, had noncancer diagnoses, and were admitted to one community hospital MSD (Supplementary Table 1 ); however, this was only a small percentage (5.2%) of a large sample population.
Table 1.
Demographic and Clinical Characteristics of Patients Receiving Inpatient Palliative Care Consult Across Four Hospitals
| Total (N = 13,180) |
Pre pandemic March 19–February 20 (n = 4068) |
Pandemic Year 1 March 20–February 21 (n = 4560) |
Pandemic Year 2 March 21– ebruary 22 (n = 4552) |
P- Value | |||||
|---|---|---|---|---|---|---|---|---|---|
| No. | % | No. | % | No. | % | No. | % | ||
| Hospital | |||||||||
| MSA | 7306 | 55.4 | 2361 | 58.0 | 2590 | 56.8 | 2355 | 51.7 | <0.001 |
| MSB | 2096 | 15.9 | 671 | 16.5 | 689 | 15.1 | 736 | 16.2 | |
| MSC | 1982 | 15.0 | 550 | 13.5 | 705 | 15.5 | 727 | 16.0 | |
| MSD | 1796 | 13.6 | 486 | 11.9 | 576 | 12.6 | 734 | 16.1 | |
| Age, years | |||||||||
| Mean (SD) | 70.3 (16.4) | 70.0 (16.5) | 69.9 (16.0) | 70.8 (16.7) | 0.019 | ||||
| Median [IQR] | 72 [60–83] | 72 [60–83] | 71 [60–82] | 72 [61–84] | |||||
| <50 | 1441 | 10.9 | 475 | 11.7 | 481 | 10.5 | 485 | 10.7 | 0.004 |
| 50–64 | 3106 | 23.6 | 926 | 22.8 | 1135 | 24.9 | 1045 | 23.0 | |
| 65–79 | 4351 | 33.0 | 1,366 | 33.6 | 1530 | 33.6 | 1455 | 32.0 | |
| ≥80 | 4282 | 32.5 | 1,301 | 32.0 | 1414 | 31.0 | 1567 | 34.4 | |
| Gender | |||||||||
| Female | 6663 | 50.6 | 2,127 | 52.3 | 2193 | 48.1 | 2343 | 51.5 | <0.001 |
| Race | |||||||||
| White | 4351 | 33.0 | 1,391 | 34.2 | 1451 | 31.8 | 1509 | 33.2 | 0.004 |
| Non-Hispanic Black | 3242 | 24.6 | 982 | 24.1 | 1091 | 23.9 | 1169 | 25.7 | |
| Hispanic | 2798 | 21.2 | 893 | 22.0 | 989 | 21.7 | 916 | 20.1 | |
| Othera | 2233 | 16.9 | 650 | 16.0 | 806 | 17.7 | 777 | 17.1 | |
| Unknown/missing | 556 | 4.2 | 152 | 3.7 | 223 | 4.9 | 181 | 4.0 | |
| KPS at palliative care consultation | |||||||||
| Mean (SD) | 33.5 (18.7) | 35.6 (16.8) | 32.4 (19.3) | 32.6 (19.4) | <0.001 | ||||
| Median [IQR] | 30 [20-50] | 40 [20-50] | 30 [20-50] | 30 [20-50] | |||||
| 10%–20% | 5497 | 41.7 | 1200 | 29.5 | 2134 | 46.8 | 2163 | 47.5 | <0.001 |
| 30%–40% | 4309 | 32.7 | 1827 | 44.9 | 1284 | 28.2 | 1198 | 26.3 | |
| 50%–70% | 2989 | 22.7 | 949 | 23.3 | 992 | 21.8 | 1048 | 23.0 | |
| 80%–100% | 385 | 2.9 | 92 | 2.3 | 150 | 3.3 | 143 | 3.1 | |
| Number of serious illness diagnoses | |||||||||
| 0 | 2579 | 19.6 | 766 | 18.8 | 911 | 20.0 | 902 | 19.8 | 0.034 |
| 1 | 6556 | 49.7 | 2031 | 49.9 | 2203 | 48.3 | 2322 | 51.0 | |
| ≥2 | 4045 | 30.7 | 1271 | 31.2 | 1446 | 31.7 | 1328 | 29.2 | |
| Primary palliative care diagnosis | |||||||||
| Noncancer | 6658 | 50.5 | 2187 | 53.8 | 2067 | 45.3 | 2404 | 52.8 | <0.001 |
| COVID-19 | 1180 | 9.0 | 1 | 0.0 | 741 | 16.2 | 438 | 9.6 | |
| Cancer | 5342 | 40.5 | 1880 | 46.2 | 1752 | 38.4 | 1710 | 37.6 | |
| Time between admission to palliative care consultation, days | |||||||||
| Mean (SD) | 7.3 (11.1) | 6.6 (10.1) | 7.1 (11.4) | 7.9 (11.8) | <0.001 | ||||
| Median [IQR] | 4 [2-9] | 4 [1-8] | 4 [1-9] | 4 [2-10] | |||||
| ≤2 | 4839 | 36.7 | 1583 | 38.9 | 1752 | 38.4 | 1504 | 33.0 | <0.001 |
| 3–7 | 4481 | 34.0 | 1438 | 35.3 | 1469 | 32.2 | 1574 | 34.6 | |
| 8–14 | 2169 | 16.5 | 596 | 14.7 | 753 | 16.5 | 820 | 18.0 | |
| ≥15 | 1691 | 12.8 | 451 | 11.1 | 586 | 12.9 | 654 | 14.4 | |
| Time between palliative care consultation and discharge, days | |||||||||
| Mean (SD) | 11.4 (17.8) | 10.9 (17.7) | 11.4 (18.8) | 11.7 (16.7) | 0.13 | ||||
| Median [IQR] | 6 [2-13] | 6 [2-12] | 6 [2-13] | 7 [3-14] | |||||
| ≤2 | 3467 | 26.3 | 1173 | 28.8 | 1,197 | 26.2 | 1,097 | 24.1 | <0.001 |
| 3–7 | 4151 | 31.5 | 1274 | 31.3 | 1,456 | 31.9 | 1,421 | 31.2 | |
| 8–14 | 2642 | 20.0 | 779 | 19.1 | 912 | 20.0 | 951 | 20.9 | |
| ≥15 | 2920 | 22.2 | 842 | 20.7 | 995 | 21.8 | 1,083 | 23.8 | |
| Hospital length of stay, days | |||||||||
| Mean (SD) | 18.6 (23.0) | 17.6 (22.4) | 18.5 (24.4) | 19.7 (22.1) | <0.001 | ||||
| Median [IQR] | 12 [6-22] | 11 [6-21] | 12 [7-22] | 13 [7-24] | |||||
| 0–7 | 3962 | 30.1 | 1398 | 34.4 | 1347 | 29.5 | 1,217 | 26.7 | <0.001 |
| 8–14 | 3684 | 28.0 | 1129 | 27.8 | 1336 | 29.3 | 1,219 | 26.8 | |
| 15–30 | 3435 | 26.1 | 949 | 23.3 | 1174 | 25.7 | 1,312 | 28.8 | |
| >30 | 2099 | 15.9 | 592 | 14.6 | 703 | 15.4 | 804 | 17.7 | |
| ICU admission | |||||||||
| Yes | 4706 | 35.7 | 1,366 | 33.6 | 1689 | 37.0 | 1,651 | 36.3 | 0.002 |
| ICU length of stay, days | |||||||||
| Mean (SD) | 12.2 (19.1) | 10.4 (15.4) | 13.5 (24.2) | 12.2 (15.4) | <0.001 | ||||
| Median [IQR] | 7 [3-15] | 6 [3-12] | 8 [3-16] | 7 [3-15] | |||||
| Discharge disposition | |||||||||
| Home | 3679 | 27.9 | 1,258 | 30.9 | 1260 | 27.6 | 1,161 | 25.5 | <0.001 |
| Postacute care | 2498 | 19.0 | 798 | 19.6 | 768 | 16.8 | 932 | 20.5 | |
| Home hospice | 1254 | 9.5 | 411 | 10.1 | 406 | 8.9 | 437 | 9.6 | |
| Inpatient hospice | 1212 | 9.2 | 477 | 11.7 | 366 | 8.0 | 369 | 8.1 | |
| Deceased | 4537 | 34.4 | 1,124 | 27.6 | 1760 | 38.6 | 1653 | 36.3 | |
Race categorized as “Other” include races documented as Asian, American Indian or Alaska Native, Native Hawaiian or Pacific Islander, or other.
Definition of abbreviations: ICU= Intensive Care Unit, KPS = Karnofsky Performance Status
Continuous variables are presented as mean (standard deviation) and median [interquartile range].
Categorical variables are presented as frequency and percentage by column.
P-values are obtained by Chi-square tests and ANOVA for comparisons of categorical and continuous variables, respectively.
Supplementary Table 1.
Sensitivity Analysis of Population Excluded due to Missing KPS
| KPS missing (n=717) |
KPS documented (n=13,180) |
||||
|---|---|---|---|---|---|
| No. | % | No. | % | p-value | |
| Time period of palliative care consultation | |||||
| Prepandemic: March 19 – February 20 | 394 | 55.0 | 4068 | 30.9 | <0.001 |
| Pandemic Year 1: March 20 – February 21 | 235 | 32.8 | 4560 | 34.6 | |
| Pandemic Year 2: March 21 – February 22 | 88 | 12.3 | 4552 | 34.5 | |
| Hospital | |||||
| MSA | 108 | 15.1 | 7306 | 55.4 | <0.001 |
| MSB | 26 | 3.6 | 2096 | 15.9 | |
| MSC | 134 | 18.7 | 1982 | 15.0 | |
| MSD | 449 | 62.6 | 1796 | 13.6 | |
| Age, days | |||||
| Mean (SD) | 77.2 (14.0) | 70.3 (16.4) | <0.001 | ||
| Median [IQR] | 79 [68-87] | 72 [60-83] | |||
| <50 | 28 | 3.9 | 1441 | 10.9 | <0.001 |
| 50–64 | 104 | 14.5 | 3106 | 23.6 | |
| 65–79 | 230 | 32.1 | 4351 | 33.0 | |
| ≥80 | 355 | 49.5 | 4282 | 32.5 | |
| Gender | |||||
| Female | 382 | 53.3 | 6663 | 50.6 | 0.16 |
| Race | |||||
| White | 329 | 45.9 | 4351 | 33.0 | <0.001 |
| Non-Hispanic Black | 100 | 13.9 | 3242 | 24.6 | |
| Hispanic | 163 | 22.7 | 2798 | 21.2 | |
| Othera | 109 | 15.2 | 2233 | 16.9 | |
| Unknown/missing | 16 | 2.2 | 556 | 4.2 | |
| Number of serious illness diagnosis | |||||
| 0 | 134 | 18.7 | 2579 | 19.6 | 0.37 |
| 1 | 345 | 48.1 | 6556 | 49.7 | |
| ≥2 | 238 | 33.2 | 4045 | 30.7 | |
| Primary palliative care diagnosis | |||||
| Noncancer | 590 | 82.3 | 6658 | 50.5 | <0.001 |
| COVID | 85 | 11.9 | 1180 | 9.0 | |
| Cancer | 42 | 5.9 | 5342 | 40.5 | |
| Time between admission to palliative care consultation, days | |||||
| Mean (SD) | 5.6 (7.0) | 7.3 (11.1) | <0.001 | ||
| Median [IQR] | 3 [1-7] | 4 [2-9] | |||
| ≤2 | 300 | 41.8 | 4839 | 36.7 | 0.001 |
| 3–7 | 257 | 35.8 | 4481 | 34.0 | |
| 8–14 | 101 | 14.1 | 2169 | 16.5 | |
| ≥15 | 59 | 8.2 | 1691 | 12.8 | |
| Time between palliative care consultation and discharge, days | |||||
| Mean (SD) | 7.1 (9.2) | 11.4 (17.8) | <0.001 | ||
| Median [IQR] | 4 [2-9] | 6 [2-13] | |||
| ≤2 | 266 | 37.1 | 3467 | 26.3 | <0.001 |
| 3–7 | 234 | 32.6 | 4151 | 31.5 | |
| 8–14 | 126 | 17.6 | 2642 | 20.0 | |
| ≥15 | 91 | 12.7 | 2920 | 22.2 | |
| Length of stay, days | |||||
| Mean (SD) | 12.6 (12.8) | 18.6 (23.0) | <0.001 | ||
| Median [IQR] | 8 [5-16] | 12 [6-22] | |||
| 0–7 | 321 | 44.8 | 3962 | 30.1 | <0.001 |
| 8–14 | 190 | 26.5 | 3684 | 28.0 | |
| 15–30 | 151 | 21.1 | 3435 | 26.1 | |
| >30 | 55 | 7.7 | 2099 | 15.9 | |
| ICU admission | |||||
| Yes | 175 | 24.4 | 4706 | 35.7 | <0.001 |
| ICU length of stay, days | |||||
| Mean (SD) | 8.2 (10.5) | 12.2 (19.1) | 0.006 | ||
| Median [IQR] | 4 [2-11] | 7 [3-15] | |||
| Discharge disposition | |||||
| Home | 149 | 20.8 | 3679 | 27.9 | <0.001 |
| Postacute care | 179 | 25.0 | 2498 | 19.0 | |
| Outpatient hospice | 104 | 14.5 | 1254 | 9.5 | |
| Inpatient hospice | 106 | 14.8 | 1212 | 9.2 | |
| Deceased | 179 | 25.0 | 4537 | 34.4 | |
Race categorized as “Other” include races documented as Asian, American Indian or Alaska Native, Native Hawaiian or Pacific Islander, or other.
Definition of abbreviations: ICU= Intensive Care Unit, KPS = Karnofsky Performance Status.
Continuous variables are presented as mean (standard deviation) and median [interquartile range].
Categorical variables are presented as frequency and percentage by column.
P values are obtained by Chi-square test and Student's t-test for comparisons of categorical and continuous variables, respectively.
Fig. 1 showed that mean KPS at consultation was inversely affected by consult volume. Increasing number of consults was observed during the pandemic and was correlated with decreased functional status and increased in-hospital mortality among consult patients as compared to pre-pandemic levels. Mortality of consult patients also trended closely with overall hospital mortality at the time and was usually several folds higher than in-hospital mortality (Fig. 1b).
Fig. 1.
Trends of functional status, consult volume, and mortality before, during and after the COVID-19 pandemic. (A) Trends of patients’ functional status at the time of initial inpatient palliative care consultation in relationship to consult volume. (B) Trends of patients’ functional status in relationship to consult mortality and in-hospital mortality. KPS = Karnofsky Performance Status. For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.
The observed overall decrease in functional status at the time of consultation was driven by the relative increase in the percentage of patients with noncancer and COVID-19 who had the lowest KPS group (KPS 10%–20%) (Fig. 2 b). Jonckheere-Terpstra test showed that there was a statistically significant trend of lower KPS among patients with noncancer diagnoses over the years from pre-pandemic to pandemic year two (P < 0.001). In contrast, there was a statistically significant trend of higher KPS among cancer patients over time (P < 0.001). Of note, changes in KPS among COVID-19 patients from pandemic year one to pandemic year 2 were not statistically significant as assessed by Chi-square test (P = 0.26).
Fig. 2.
Changes in distribution of functional status groups by diagnoses before, during and after the COVID-19 pandemic. (A) Number of consults for each functional status group by diagnoses before, during and after the pandemic. (B) Percentage of each functional status group by diagnoses before, during and after the pandemic. KPS = Karnofsky Performance Status. For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.
Among patients within each functional status group, there was an overall increase in in-hospital mortality among patients who received a consultation in the pandemic years as compared to the prepandemic period (Fig. 3 a). This was statistically significant among patients with KPS 10%–20% (P = 0.018), KPS 30%–40% (P < 0.001) and KPS 50%–70% (P = 0.002), but not among those with mild functional impairment (KPS 80%–100%). When stratified by noncancer, COVID-19 and cancer diagnoses, patients with noncancer diagnoses had a significantly higher in-hospital mortality of 40.3% in pandemic year one and 37.7% in pandemic year two as compared to 30.4% prepandemic (P < 0.001), whereas patients with COVID-19 (pandemic year one 67.5%; pandemic year two 66.7%; P = 0.78) or cancer (pre-pandemic 24.4%; pandemic year one 24.4%; pandemic year two 26.6%; P = 0.25) had no significant changes in mortality over time (Supplementary Fig. 1 ).
Fig. 3.
Changes in functional status group in-hospital mortality over time and by diagnoses. (A) In-hospital mortality of each functional status group before, during and after the COVID-19 pandemic. (B) In-hospital mortality of each functional status group by diagnoses. KPS = Karnofsky Performance Status. For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.
Supplementary Fig. 1.
In-hospital deaths and mortality by year and diagnoses.
Across all diagnoses, those with very severe functional impairment (KPS 10%–20%) at time of consultation had more than 50% in-hospital mortality (52.9% among noncancer patients; 72.7% among COVID-19 patients; 57.1% among cancer patients) (Fig. 3b). Patients with COVID-19 had significantly higher in-hospital mortality as compared to those with a noncancer or cancer diagnoses across all degrees of functional impairment at time of consultation (P < 0.001 for KPS 10%–20%, KPS 30%–40% and KPS 50%–70%; P = 0.039 for KPS 80%–100%). There was a graded decrease in mortality with better functional status across all diagnoses (Fig. 3b).
There was a statistically significant increase in the mean number of days between hospital admission and consultation over time (6.6 days (SD 10.1) prepandemic; 7.1 days (SD 11.4) in pandemic year one; 7.9 days (SD 11.8) in pandemic year two; P < 0.001) (Table 1). Similarly, hospital and ICU length of stay also increased significantly over time (Table 1). The findings remained consistent when sensitivity analysis was performed among patients with length of stay ≤ 30 days (Supplementary Table 2 ). When stratified by diagnoses, the mean number of admission-to-consult days (P < 0.001), hospital length of stay (P = 0.003), and ICU days (P = 0.028) all increased significantly over time among patients with noncancer diagnoses (Supplementary Table 3 ).
Supplementary Table 2.
Sensitivity Analysis of Patients With Length of Stay ≤ 30 Days
| Total (N = 11,081) |
March 19 – February 20 (n = 4068) |
March 20 – February 21 (n = 4560) |
March 21 – February 22 (n = 3748) |
P-Value | |
|---|---|---|---|---|---|
| Number of days between admission to palliative care consultation | |||||
| Mean (SD) | 5.0 (5.0) | 4.6 (4.7) | 5.0 (5.1) | 5.3 (5.0) | <0.001 |
| Median [IQR] | 3[1–7] | 3 [1–6] | 3 [1–7] | 4 [2–7] | |
| Number of days between palliative care consultation and discharge | |||||
| Mean (SD) | 6.6 (5.9) | 6.2 (5.8) | 6.5 (5.9) | 6.9 (6.1) | <0.001 |
| Median[IQR] | 5 [2–9] | 5 [2–9] | 5 [2–9] | 5 [2–10] | |
| Length of stay | |||||
| Mean (SD) | 11.5 (7.4) | 10.8 (7.2) | 11.5 (7.3) | 12.2 (7.5) | <0.001 |
| Median[IQR] | 10 [6–16] | 9 [5–15] | 10 [6–16] | 11 [6–17] | |
| ICU days | |||||
| Mean (SD) | 7.3 (6.0) | 6.6 (5.5) | 7.8 (6.3) | 7.5 (6.0) | <0.001 |
| Median [IQR] | 6 [3–10] | 5 [2–9] | 6 [3–11] | 6 [3–10] | |
Definition of abbreviations: ICU= intensive care unit.
Continuous variables are presented as mean (standard deviation) and median [interquartile range].
P-values are obtained by ANOVA.
Supplementary Table 3.
Subgroup Analysis by Different Diagnoses
| Total (N = 13,180) | April 19 – February 20 (n = 3733) | March 20 – February 21 (n = 4560) | March 21 – February 22 (n = 4552) | P Value | ||
|---|---|---|---|---|---|---|
| Number of days between admission to palliative care consultation | ||||||
| Noncancer | Mean (SD) | 7.9 (12.0) | 7.2 (10.1) | 7.7 (13.6) | 8.7 (12.0) | <0.001 |
| Median[IQR] | 4 [2–9] | 4 [2–9] | 4 [2–9] | 5 [2–10] | ||
| COVID | Mean (SD) | 10.3 (10.8) | – | 10.1 (11.0) | 10.6 (10.3) | 0.46 |
| Median[IQR] | 8 [3–14] | – | 8 [3–14] | 8 [4–14] | ||
| Cancer | Mean (SD) | 5.8 (9.9) | 5.9 (10.0) | 5.2 (7.7) | 6.2 (11.6) | 0.009 |
| Median[IQR] | 3 [1–7] | 3 [1–7] | 3 [1–6] | 3 [1–7] | ||
| Number of days between palliative care consultation and discharge | ||||||
| Noncancer | Mean (SD) | 11.4 (19.4) | 10.7 (17.8) | 11.9 (22.8) | 11.7 (17.5) | 0.068 |
| Median[IQR] | 6 [2–13] | 5 [2–12] | 6 [2–13] | 6 [2–14] | ||
| COVID | Mean (SD) | 12.2 (16.4) | – | 12.2 (16.4) | 11.9 (15.5) | 0.72 |
| Median[IQR] | 6 [2–15] | – | 6 [2–15] | 6 [3–15] | ||
| Cancer | Mean (SD) | 11.1 (15.8) | 11.2 (17.4) | 10.4 (13.9) | 11.6 (15.8) | 0.060 |
| Median[IQR] | 7 [3–13] | 6 [2–13] | 6 [3–13] | 7 [3–14] | ||
| Length of stay, days | ||||||
| Noncancer | Mean (SD) | 19.4 (25.3) | 17.9 (22.6) | 19.7 (30.1) | 20.4 (22.9) | 0.003 |
| Median[IQR] | 12 [6–23] | 11 [6–21] | 11 [6–23] | 13 [7–25] | ||
| COVID | Mean (SD) | 22.5 (20.7) | – | 22.3 (21.0) | 22.5 (19.2) | 0.92 |
| Median[IQR] | 17 [10–28] | – | 16 [9–27] | 17 [10–29] | ||
| Cancer | Mean (SD) | 16.9 (20.3) | 17.1 (22.0) | 15.6 (16.7) | 17.8 (21.6) | 0.004 |
| Median[IQR] | 11 [6–20] | 11 [6–20] | 10 [6–19] | 12 [7–22] | ||
| ICU days | ||||||
| Noncancer | Mean (SD) | 12.3 (20.7) | 11.1 (14.5) | 13.6 (29.0) | 12.2 (16.1) | 0.028 |
| Median[IQR] | 7 [3–14] | 7 [3–13] | 8 [3–15] | 7 [3–15] | ||
| COVID | Mean (SD) | 17.8 (17.4) | – | 18.2 (18.3) | 17.1 (15.9) | 0.46 |
| Median[IQR] | 13 [7–23] | – | 13 [7–23] | 13 [7–21] | ||
| Cancer | Mean (SD) | 8.0 (13.0) | 8.7 (17.3) | 7.0 (7.7) | 8.2 (10.0) | 0.23 |
| Median[IQR] | 4 [2–9] | 5 [2–9] | 4 [2–9] | 5 [2–9] | ||
Definition of abbreviation: ICU = intensive care unit.
Continuous variables are presented as mean (standard deviation) and median [interquartile range].
P-values are obtained by ANOVA for Noncancer and Cancer subgroups, and Student's t-test for COVID subgroup.
Discussion
In our study, we found that as palliative care consultation volume increased with the onset of the COVID-19 pandemic, the functional status of patients at the time of consultation decreased. When the consult volume was highest, patients who received consultation had the lowest functional status at the time of consultation, which was also associated with increased in-hospital mortality. Interestingly, the functional status of patients with cancer at the time of consultation was higher during the pandemic, and there was no significant difference in the mortality of patients with cancer over time. Our findings highlighted that the increased consult volume during the pandemic limited the inpatient palliative care team's availability to initiate early, upstream interventions, especially among patients with noncancer diagnoses.
During the peak of the pandemic, healthcare systems incorporated multiple strategies to meet the increasing inpatient palliative care needs, including redesigning consult team work flow, rapid scaling-up of telehealth and electronic consults, and increased training and educational tools for nonpalliative care providers.28, 29, 30 However, despite these efforts, during times of increased volume, we continued to observe a lower functional status in patients at the time of consultation. As we allocated specialty palliative care resource to patients with the lowest functional status, those with less yet still significant functional impairment were unable to receive palliative care, leaving them with potentially unmet needs. This delay in palliative care initiation may have led to less timely interventions for symptom management, emotional support, and serious illness communication. Innovative consultation models are required to meet the increasing volume of inpatients with palliative care needs while at the same time improve access to palliative care for those with less functional impairment who are earlier in their illness trajectory. The KPS may serve as an easy-to-measure marker to target palliative care services towards those with the most severe functional impairment (KPS 10%–20%) and highest mortality, as well as identify earlier opportunities to deliver palliative care to those with severe functional impairment (KPS 30%–40%), and even those with moderate impairment (KPS 50%–70%).
We also observed a widening gap in access to inpatient palliative care between patients with noncancer and cancer diagnoses in times of health system stress. While noncancer and cancer patients have similar symptom burden and would equally benefit from palliative care,31 , 32 disparities among noncancer patients in receiving palliative care have been well described even before the pandemic.33, 34, 35, 36 With increasing consult volume during the pandemic, our findings suggested that noncancer patients disproportionally suffered the consequences of resource shortages, as suggested by lower KPS at time of consultation, increased in-hospital mortality and greater lengthening of admission-to-consult time, when compared to patients with cancer. There are several potential explanations: providers may preferentially recognize palliative care needs among cancer patients due to existing referral culture, while palliative care needs among patients with noncancer diagnoses are more challenging to identify given their less predictable illness trajectory and prognosis. Additionally, an established oncology palliative care consultation program in our health system with specified trigger criteria to initiate consultation for patients with advanced cancer might have “protected” cancer patients from resource constraints. Further research is required to understand the mechanisms behind these disparities to promote equity in palliative care delivery.
Our study has several limitations. We relied on the functional status assessment of patients at the time of consultation, which might not reflect the functional status of patients at the time of admission, and we were not able to compare changes in functional status between consult patients and other hospitalized patients who did not receive a palliative care consultation. During the COVID-19 pandemic, healthcare systems experienced rapid changes in demand which affected workflow and shifted some characteristics of patients who sought medical care or were admitted to the hospital. As a result, we were not able to measure and control for these shifting patient and hospital factors due to the pandemic, which might have contributed to changes in patients’ functional status and in-hospital mortality. We used billing codes to facilitate data collection, which may lead to confounding by under-counting and misclassification of COVID-19 and serious illness diagnoses. Although we included a large patient population across four tertiary and community hospitals within urban healthcare systems, our findings may not be generalizable to other settings or geographic areas.
Conclusion
Our study provided new insights about the shifts in inpatient specialty palliative care utilization towards patients with lower functional status and higher in-hospital mortality when faced with increasing workforce demands during the COVID-19 pandemic, which resulted in the loss of the previously gained momentum of providing more upstream consultation to patients earlier in illness trajectory with less functional impairment. Importantly, patients with noncancer diagnoses were more vulnerable to this shift in resource allocation as compared to those with cancer. Future studies should investigate innovative approaches to address the unmet needs among patients with noncancer diagnoses and ensure timely palliative care delivery to patients with less functional impairment, especially in times of resource constraints.
Disclosures and Acknowledgments
The authors declare no competing financial interests. Dr. Laura P. Gelfman received support from the National Institute on Aging [K23AG049930] and the Cambia Health Foundation Sojourns Scholar Leadership Program. Dr. R. Sean Morrison received support from the Mount Sinai Older Adults Independence Center [P30AG028741] and the National Institute on Aging [P01AG066605].
All authors contributed substantially to the study design, data interpretation, and drafting/revision of the manuscript. Dr. Luyi Xu and Ms. Li Zeng completed data acquisition. Dr. Luyi Xu and Dr. Laura P. Gelfman completed statistical analysis. Dr. Luyi Xu and Dr. Laura P. Gelfman had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
References
- 1.Lage DE, El-Jawahri A, Fuh CX, et al. Functional impairment, symptom burden, and clinical outcomes among hospitalized patients with advanced cancer. J Natl Compr Canc Netw. Jun 2020;18:747–754. doi: 10.6004/jnccn.2019.7385. [DOI] [PubMed] [Google Scholar]
- 2.Portz JD, Kutner JS, Blatchford PJ, Ritchie CS. High symptom burden and low functional status in the setting of multimorbidity. J Am Geriatr Soc. Oct 2017;65:2285–2289. doi: 10.1111/jgs.15045. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Zampieri FG, Bozza FA, Moralez GM, et al. The effects of performance status one week before hospital admission on the outcomes of critically ill patients. Intensive Care Med. 2017;43:39–47. doi: 10.1007/s00134-016-4563-5. [DOI] [PubMed] [Google Scholar]
- 4.Turcotte LA, Zalucky AA, Stall NM, et al. Baseline frailty as a predictor of survival after critical care: a retrospective cohort study of older adults receiving home care in Ontario, Canada. Chest. 2021;160:2101–2111. doi: 10.1016/j.chest.2021.06.009. [DOI] [PubMed] [Google Scholar]
- 5.Mehta A, Chai E, Berglund K, Rizzo E, Moreno J, Gelfman LP. Using admission karnofsky performance status as a guide for palliative care discharge needs. J Palliat Med. 2021;24:910–913. doi: 10.1089/jpm.2020.0543. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Bajwah S, Oluyase AO, Yi D, et al. The effectiveness and cost-effectiveness of hospital-based specialist palliative care for adults with advanced illness and their caregivers. Cochrane Database Syst Rev. 2020;9 doi: 10.1002/14651858.CD012780.pub2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Quinn KL, Shurrab M, Gitau K, et al. Association of receipt of palliative care interventions with health care use, quality of life, and symptom burden among adults with chronic noncancer illness: a systematic review and meta-analysis. JAMA. 2020;324:1439–1450. doi: 10.1001/jama.2020.14205. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.May P, Normand C, Cassel JB, et al. Economics of palliative care for hospitalized adults with serious illness: a meta-analysis. JAMA Intern Med. 2018;178:820–829. doi: 10.1001/jamainternmed.2018.0750. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Gaertner J, Siemens W, Meerpohl JJ, et al. Effect of specialist palliative care services on quality of life in adults with advanced incurable illness in hospital, hospice, or community settings: systematic review and meta-analysis. BMJ. 2017;357:j2925. doi: 10.1136/bmj.j2925. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Qureshi D, Tanuseputro P, Perez R, Pond GR, Seow HY. Early initiation of palliative care is associated with reduced late-life acute-hospital use: a population-based retrospective cohort study. Palliat Med. 2019;33:150–159. doi: 10.1177/0269216318815794. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ferrell BR, Temel JS, Temin S, et al. Integration of palliative care into standard oncology care: American Society of Clinical Oncology Clinical Practice Guideline update. J Clin Oncol. 2017;35:96–112. doi: 10.1200/JCO.2016.70.1474. [DOI] [PubMed] [Google Scholar]
- 12.Lanken PN, Terry PB, Delisser HM, et al. An official American Thoracic Society clinical policy statement: palliative care for patients with respiratory diseases and critical illnesses. Am J Respir Crit Care Med. 2008;177:912–927. doi: 10.1164/rccm.200605-587ST. [DOI] [PubMed] [Google Scholar]
- 13.Braun LT, Grady KL, Kutner JS, et al. Palliative care and cardiovascular disease and stroke: a policy statement from the American Heart Association/American stroke association. Circulation. 2016;134:e198–e225. doi: 10.1161/CIR.0000000000000438. [DOI] [PubMed] [Google Scholar]
- 14.Dumanovsky T, Augustin R, Rogers M, Lettang K, Meier DE, Morrison RS. The growth of palliative care in U.S. Hospitals: a status report. J Palliat Med. 2016;19:8–15. doi: 10.1089/jpm.2015.0351. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Kamal AH, Bull JH, Swetz KM, Wolf SP, Shanafelt TD, Myers ER. Future of the palliative care workforce: preview to an impending crisis. Am J Med. 2017;130:113–114. doi: 10.1016/j.amjmed.2016.08.046. [DOI] [PubMed] [Google Scholar]
- 16.Lupu D, American Academy of Hospice, Palliative Medicine Workforce Task F Estimate of current hospice and palliative medicine physician workforce shortage. J Pain Symptom Manage. 2010;40:899–911. doi: 10.1016/j.jpainsymman.2010.07.004. [DOI] [PubMed] [Google Scholar]
- 17.Sleeman KE, Cripps RL, Murtagh FEM, et al. Change in activity of palliative care services during the Covid-19 pandemic: a multinational survey (CovPall) J Palliat Med. 2022;25:465–471. doi: 10.1089/jpm.2021.0315. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Chalk D, Robbins S, Kandasamy R, et al. Modelling palliative and end-of-life resource requirements during COVID-19: implications for quality care. BMJ Open. 2021;11 doi: 10.1136/bmjopen-2020-043795. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Gonzalez D, Karpman M, Kenney GM, Zuckerman S. Urban Institute; Washington, DC: 2021. Delayed and forgone health care for nonelderly adults during the COVID-19 pandemic. [Google Scholar]
- 20.Mogharab V, Ostovar M, Ruszkowski J, et al. Global burden of the COVID-19 associated patient-related delay in emergency healthcare: a panel of systematic review and meta-analyses. Global Health. 2022;18:58. doi: 10.1186/s12992-022-00836-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Czeisler ME, Marynak K, Clarke KEN, et al. Delay or avoidance of medical care because of COVID-19-related concerns - United States, June 2020. MMWR Morb Mortal Wkly Rep. 2020;69:1250–1257. doi: 10.15585/mmwr.mm6936a4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.De Luca G, Verdoia M, Cercek M, et al. Impact of COVID-19 pandemic on mechanical reperfusion for patients with STEMI. J Am Coll Cardiol. 2020;76:2321–2330. doi: 10.1016/j.jacc.2020.09.546. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Maringe C, Spicer J, Morris M, et al. The impact of the COVID-19 pandemic on cancer deaths due to delays in diagnosis in England, UK: a national, population-based, modelling study. Lancet Oncol. 2020;21:1023–1034. doi: 10.1016/S1470-2045(20)30388-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Moynihan R, Sanders S, Michaleff ZA, et al. Impact of COVID-19 pandemic on utilisation of healthcare services: a systematic review. BMJ Open. 2021;11 doi: 10.1136/bmjopen-2020-045343. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Gelfman LP, Morrison RS, Moreno J, Chai E. Palliative care as essential to a hospital system's pandemic preparedness planning: how to get ready for the next wave. J Palliat Med. 2021;24:656–658. doi: 10.1089/jpm.2020.0670. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Kelley AS, Ferreira KB, Bollens-Lund E, Mather H, Hanson LC, Ritchie CS. Identifying older adults with serious illness: Transitioning from ICD-9 to ICD-10. J Pain Symptom Manage. 2019;57:1137–1142. doi: 10.1016/j.jpainsymman.2019.03.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Abernethy AP, Shelby-James T, Fazekas BS, Woods D, Currow DC. The Australia-modified Karnofsky Performance Status (AKPS) scale: a revised scale for contemporary palliative care clinical practice [ISRCTN81117481] BMC Palliat Care. 2005;4(7) doi: 10.1186/1472-684X-4-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Blinderman CD, Adelman R, Kumaraiah D, et al. A Comprehensive approach to palliative care during the coronavirus pandemic. J Palliat Med. 2021;24:1017–1022. doi: 10.1089/jpm.2020.0481. [DOI] [PubMed] [Google Scholar]
- 29.Ankuda CK, Woodrell CD, Meier DE, Morrison RS, Chai E. A beacon for dark times: palliative care support during the coronavirus pandemic. Catalyst Non-Issue Content. 2020;1(3) doi: 10.1056/CAT.20.0204. [DOI] [Google Scholar]
- 30.Ganeshan S, Humphreys J, Judson T. Reimagining the inpatient palliative care consult: lessons from COVID-19. Am J Manag Care. 2021;27:e215–e217. doi: 10.37765/ajmc.2021.88704. [DOI] [PubMed] [Google Scholar]
- 31.Mounsey L, Ferres M, Eastman P. Palliative care for the patient without cancer. Aust J Gen Pract. 2018;47:765–769. doi: 10.31128/AJGP-07-18-4625. [DOI] [PubMed] [Google Scholar]
- 32.Moens K, Higginson IJ, Harding R, Euro I. Are there differences in the prevalence of palliative care-related problems in people living with advanced cancer and eight non-cancer conditions? A systematic review. J Pain Symptom Manage. 2014;48:660–677. doi: 10.1016/j.jpainsymman.2013.11.009. [DOI] [PubMed] [Google Scholar]
- 33.Tobin J, Rogers A, Winterburn I, et al. Hospice care access inequalities: a systematic review and narrative synthesis. BMJ Support Palliat Care. 2022;12:142–151. doi: 10.1136/bmjspcare-2020-002719. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Rosenwax L, Spilsbury K, McNamara BA, Semmens JB. A retrospective population based cohort study of access to specialist palliative care in the last year of life: who is still missing out a decade on? BMC Palliat Care. 2016;15:46. doi: 10.1186/s12904-016-0119-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Quinn KL, Wegier P, Stukel TA, Huang A, Bell CM, Tanuseputro P. Comparison of palliative care delivery in the last year of life between adults with terminal noncancer illness or cancer. JAMA Netw Open. 2021;4 doi: 10.1001/jamanetworkopen.2021.0677. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Lau C, Meaney C, Morgan M, Cook R, Zimmermann C, Wentlandt K. Disparities in access to palliative care facilities for patients with and without cancer: a retrospective review. Palliat Med. 2021;35:1191–1201. doi: 10.1177/02692163211007387. [DOI] [PMC free article] [PubMed] [Google Scholar]