Journal of Intensive Care Medicine

Integrating Palliative Care into Critical Care: A Quality Improvement Study Cynthia Hsu-Kim, Tara Friedman, Edward Gracely and James Gasperino J Intensive Care Med published online 5 March 2014 DOI: 10.1177/0885066614523923 The online version of this article can be found at:

Published by:

Additional services and information for Journal of Intensive Care Medicine can be found at: Email Alerts: Subscriptions: Reprints: Permissions:

>> OnlineFirst Version of Record - Mar 5, 2014 What is This?

Downloaded from at TEXAS SOUTHERN UNIVERSITY on November 10, 2014


Integrating Palliative Care into Critical Care: A Quality Improvement Study

Journal of Intensive Care Medicine 1-7 ª The Author(s) 2014 Reprints and permission: DOI: 10.1177/0885066614523923

Cynthia Hsu-Kim, MD1, Tara Friedman, MD2, Edward Gracely, PhD3, and James Gasperino, MD, PhD, MPH4,5

Abstract Background: Many terminally ill patients experience an increasing intensity of medical care, an escalation frequently not consistent with their preferences. In 2009, formal palliative care consultation (PCC) was integrated into our medical intensive care unit (ICU). We hypothesized that significant differences in clinical and economic outcomes exist between ICU patients who received PCC and those who did not. Methods: We reviewed ICU admissions between July and October 2010, identified 41 patients who received PCC, and randomly selected 80 patients who did not. We measured clinical outcomes and economic variables associated with patients’ ICU courses. Results: Patients in the PCC group were older (average 64 years, standard deviation [SD] 19.2 vs 55.6 years, SD 14.5; P ¼ .021) and sicker (median Acute Physiology and Chronic Health Evaluation IV score 85.5, interquartile range [IQR] 60.5-107.5 vs 60, IQR 39.2-74.75; P < .001) than the non-PCC controls. PCC patients received significantly more total days of ICU care on average (8 days, IQR 4-15 vs 4 days, IQR 2-7; P < .001), had more ICU admissions, and were more likely to die during their ICU stay (64.3% vs 12.5%, P < .001). Median total hospital charges per patient attributable to ICU care were higher in the PCC group than in the controls (US$315,493, IQR US$156,470-US$486,740 vs US$116,934, IQR US$54,750-US$288,660; P < .001). After we adjusted for ICU length of stay, we found that median ICU charges per day per patient did not differ significantly between the groups (US$37,463, IQR US$27,429-US$56,230 vs US$41,332, IQR US$30,149-US$63,288; P ¼ .884). Median time to PCC during the ICU stay was 7 days (IQR 2-14.5 days). Conclusions: Patients who received PCC had higher disease acuity, longer ICU lengths of stay, and higher ICU mortality than controls. ‘‘Trigger’’ programs in the ICU may improve utilization of PCC services, improve patient comfort, and reduce invasive, often futile end-of-life care. Keywords palliative care, intensive care unit, end-of-life care, hospice and palliative care medicine

Introduction In 1995, the Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments (SUPPORT) investigators found that physicians generally had a poor understanding of their patients’ wishes regarding their end-of-life care, often resulting in aggressive and painful care that was contrary to the patients’ preferences.1 A major finding of this trial indicated that improved communication between physicians and patients was not enough to change current practices concerning end-oflife care. The authors suspected that a greater societal change would be required to improve the experiences of seriously ill hospitalized patients at the end of life.1 More than a decade later, the Dartmouth Atlas of Health Care was published in response to burgeoning health care costs and demonstrated that while chronically ill patients still prefer to avoid hospitalization and intensive care during the final days of a terminal illness, the majority of health care expenditures occur in the last 2 years of life.2,3 In addition, evaluation of regional differences revealed that more costly care was

1 Department of Medicine, Division of Internal Medicine, Drexel University College of Medicine, Philadelphia, PA, USA 2 Vitas Palliative Care Solutions, Philadelphia, PA, USA 3 Department of Epidemiology and Biostatistics, School of Public Health, Drexel University, Philadelphia, PA, USA 4 Department of Medicine, Section of Critical Care Medicine, Drexel University College of Medicine, Philadelphia, PA, USA 5 Department of Environmental and Occupational Health, School of Public Health, Drexel University, Philadelphia, PA, USA

Received July 2, 2013, and in revised form December 9, 2013. Accepted for publication December 11, 2013. Corresponding Author: James Gasperino, Department of Medicine, Section of Critical Care Medicine, Drexel University College of Medicine, 245N. 15th Street, MS#487, Philadelphia, PA 19102, USA. Email: [email protected]

Downloaded from at TEXAS SOUTHERN UNIVERSITY on November 10, 2014


Journal of Intensive Care Medicine

paradoxically associated with higher mortality, perhaps indicating that the care received by these patients at the end of life was excessive, resulting in both harm to the patient and waste of health care resources. This paradox is especially visible in the intensive care unit (ICU), where many chronically ill patients spend the end of their lives under an increased intensity of care with oftentimes poor outcomes.4 During the past decade, end-of-life care has evolved into a formal discipline that includes hospice and palliative care medicine (HPM). Subsequently, patients’ end-of-life needs are better served, whether through specialized palliative care units,5,6 home hospice services,7 inpatient consultation services,8-10 specialized cancer centers,11 or services incorporated within a multidisciplinary ICU team.12-14 The addition of palliative care services has decreased costs, increased patient satisfaction, and in some cases prolonged survival.15-19 The increase in programs has led to an associated increase in the number of certified palliative care physicians and fellowship training programs.10 Although many patients with terminal illnesses experience an increase in the intensity of medical care at the end of life, this escalation is frequently not consistent with their preferences.1,2 To better meet the needs of our patients and their families, in July 2010 Hahnemann University Hospital implemented a multidisciplinary hospice and palliative care service. The palliative care team included physicians certified in HPM, a nurse practitioner, and a case manager. During the first year of the program, this multidisciplinary team was integrated into our medical and neurologic intensive care services as a consultation service. As part of an ongoing quality improvement project to identify factors associated with quality and utilization of this service, we retrospectively evaluated the association between HPM, our palliative care consultation (PCC) service, and a number of clinical and economic outcomes of patients admitted to our medical and neurologic critical care services. We hypothesized that significant differences in clinical and economic outcomes exist between patients admitted to the ICU who received PCC and those who did not.

Methods Study Patients and Definitions We obtained approval for the study from the institutional review board of Drexel University College of Medicine. We then identified 42 patients on the medical and neurologic ICU services who had received PCC between July and October 2010. We excluded 1 patient who had a long and complicated hospital course that included multiple admissions to surgical, cardiac, and medical ICUs, the cost of which represented an outlier during the analysis of our data. Next, we identified a control ICU population of 80 patients, consisting of the first 20 consecutive admissions to the ICU for each month from July to October 2010, as an attempt to obtain a random, representative sample matched for month of admission. Patients were excluded from the study if they were an elective admission to the ICU or had received PCC during their hospital course

before admission to or after discharge from the ICU. During the study period, PCC was ordered at the discretion of the critical care team, which consisted of a critical care-trained attending physician and rotating resident house officers without any institution of a protocol or specific consultation criteria. The PCC consultants were not based in the ICU; rather they provided palliative care services throughout the hospital and consulted on patients in the ICU only after the initial consultation request was made by the critical care team.

Study Design and Measures We retrospectively reviewed the patients’ charts for information regarding sex, age, race, Acute Physiology and Chronic Health Evaluation IV (APACHE IV) score, total days of hospital stay, total days of ICU stay, the number of ICU readmissions within a given hospital stay, number of ventilator days, number of diagnoses and procedures billed, all-cause mortality, hospital charges for the ICU admission, and days before PCC was ordered. To obtain the average cost per day per patient, the cost for each patient’s ICU stay was divided by that patient’s length of stay. This number was then averaged for all the individual patients to obtain the average cost per day. Comorbidities were obtained from the patients’ medical histories as documented in either their initial history or their initial ICU admission note. Comorbidities counted were the following: localized and metastatic malignancies, leukemias, and lymphomas; psychiatric disorders including depression, anxiety, bipolar disorder, schizophrenia, and alcohol or substance abuse; vascular conditions such as abdominal aortic aneurysm, peripheral vascular disease, vascular malformations, decubitus ulcers, amputations, and thromboembolic disease; chronic infections including HIV infection, recurrent urinary tract infections, aspergillomas, and herpes simplex virus infection; pulmonary conditions including sarcoidosis, obstructive sleep apnea, interstitial lung disease, pulmonary hypertension, chronic obstructive lung disease, and asthma; chronic kidney disease and end-stage renal disease on dialysis; gastrointestinal disorders including peptic ulcer disease, inflammatory bowel disease, diverticulitis, hepatitis C infection, hepatocellular carcinoma, end-stage liver disease, and hepatorenal syndrome; neurologic disorders including Parkinson disease, peripheral neuropathies, seizure disorder, dementia, history of stroke, paraplegias, and amyotrophic lateral sclerosis; cardiac diseases including congestive heart failure, hypertension, coronary artery disease, valvulopathies, and hypertrophic cardiomyopathies; dysrhythmias including atrial fibrillation; hematologic disorders such as anemia of chronic disease, hemolytic anemia, polycythemia vera, and heparin-induced thrombocytopenia; endocrinologic conditions such as adrenal insufficiency, Cushing syndrome, diabetes mellitus, hyperlipidemia, hypothyroidism, hyperparathyroidism, and obesity; and rheumatologic conditions including gout, dermatomyositis, lupus, osteoporosis, and rheumatoid arthritis. The patients’ medical records were also reviewed for the presence of a code status and the number of times that the

Downloaded from at TEXAS SOUTHERN UNIVERSITY on November 10, 2014

Hsu-Kim et al


code status changed within the index ICU admission. The code status levels (ie, level of care) at our institution are as follows in decreasing levels of intensity: ‘‘full code,’’ that is, do not resuscitate (DNR)/do not intubate (DNI); maintenance care; comfort care; and withdrawal of care (ie, withdrawal of life-sustaining therapies). The level-designated DNR/DNI involves the option to choose either one or both the options and the ability to choose or deny other elements of care, including dialysis or antibiotics. Maintenance care is a continuation of the therapeutic efforts that are in place at the time, without the option to escalate medical care. For instance, if vasopressors and antibiotics are already being administered, they may be continued; however, newly diagnosed anemia may not be treated with transfusion if such an intervention was not initially part of the treatment plan. Comfort care is the replacement of active therapies with medications designed to provide comfort and with the caveat that therapeutic measures such as intravenous inotropes or dialysis may be continued as individual measures to relieve, for example, shortness of breath. Withdrawal of care (ie, withdrawal of life-sustaining measures) allows for cessation of all active treatment and frequently includes the initiation of an intravenous narcotic. Clinically, withdrawal of care is reserved for patients who would not survive long enough without therapy to benefit from the sole provision of comfort therapies for any meaningful period. Within our institution, code status is designated by a levelof-care form located in a designated section of the patient’s medical record. Per our hospital’s policy, this form must be signed by a house staff physician immediately after discussion with the family and cosigned by an attending physician within 24 hours. Forms with an attending physician’s signature are valid indefinitely until replaced by a second form bearing an attending physician’s signature. Because all forms with a signature indicate that a discussion has taken place with the patient’s family regarding a shift in the level of intensity of care, for the purposes of this study, we counted all forms dated within the ICU stay and including a signature by either a member of the house staff or an attending physician. If a level-of-care form was missing from the patient’s medical record, we considered the code status to be ‘‘full code’’ for the entire ICU stay.

Methods of Measurement The APACHE IV score was calculated using an online calculator on the Web site of the Middle East Critical Care Assembly ( To calculate the average charges per day, the charge for each patient’s ICU stay was divided by that patient’s ICU length of stay to obtain ICU charges per day for that patient. Once this number was calculated per patient, the average was calculated for all patients. Because hospital charge data were obtained only for the initial ICU stay, any readmissions to a surgical or medical ICU were not included in the cost calculations.

Data Analysis Data were collected prospectively and analyzed retrospectively. Univariate analysis included comparison of patient characteristics between the PCC and the control groups using independent sample t-test for continuous variables. Sex, race, and mortality were compared using chi-square (w2) analysis. We used the Mann-Whitney U test for significance to measure differences between economic variables such as hospital charges. Hospital mortality includes all patients who died on an inpatient unit or floor during that same hospitalization. Patients who may have died in hospice were not included in this calculation. All-cause mortality was calculated as a percentage of patients dying for any reason during their hospital stay (admission to discharge date). All statistical tests used a 2-tailed a of .05 for statistical significance. Statistical analyses were performed using IBM SPSS Statistics 19 software. Because of the wide, unequal distribution of the APACHE scores and other clinical parameters, these values are reported as median values with interquartile ranges (IQRs). The ages of the patients, however, were normally distributed and therefore are reported as means and standard deviations (SDs).

Results Patient Characteristics Of the 121 patients included in the study, 41 patients (the PCC group) received consultation from our palliative care service and 80 patients (the control group) did not. The demographic data and clinical characteristics are summarized in Table 1. The 2 groups had a similar racial distribution; however, the PCC group had a higher percentage of women than the control group (58% vs 36%; P ¼ .019). Patients in the PCC group were older (average 64 years, SD 19.2 vs 55.6 years, SD 14.5; P ¼ .021) and sicker (median APACHE IV score 85.5, IQR 60.5-107.5 vs 60, IQR 39.2-74.75; P < .001) than those in the control group. The palliative care group had significantly longer ICU stays (8 days, IQR 4-15 vs 4 days, IQR 2-7; P < .001) as well as a higher mortality (64.3% vs 12.5%; P < .001). The PCC patients had more procedures performed than patients in the control group (5 procedures, IQR 3-9 vs 4 procedures, IQR 2-9; P ¼ .037) and a trend toward more diagnoses per patient (17 diagnoses, IQR 14-23 vs 14 diagnoses, IQR 10-22.7; P ¼ .059). Despite the longer ICU stays, the 2 groups had no significant difference in the length of total hospital stays (13 days, IQR 7-24 vs 7.5 days, IQR 3-24.5; P ¼ .145). Ventilator days were also not significantly different between the 2 groups; the PCC group spent a median of 3 days ventilated as compared with 0 days in the control group (P ¼ .097). Interestingly, patients in the PCC had slightly though significantly fewer comorbid conditions per patient than their control counterparts (3 comorbid conditions, IQR 2-5 vs 4 comorbid conditions, IQR 3-5.75; P ¼ .041).

Palliative Care-Related Outcomes The median length of ICU stay for the PCC group after PCC was 3 days (IQR 1-5 days). On admission to the ICU, 85.4%

Downloaded from at TEXAS SOUTHERN UNIVERSITY on November 10, 2014


Journal of Intensive Care Medicine

Table 1. Baseline Characteristics of 121 Study Participants Receiving Palliative Care Consultation in the ICU Versus Controls (July to October 2010).a Characteristics Age Sex Male Female Race White African American Other APACHE IV score Total hospital days Total ICU days Number of diagnoses Number of procedures Ventilator days Comorbidities Hospital stay mortality

Palliative Care (n ¼ 41) 64 y (SD 19.2)

Control Group (n ¼ 80) 55.6 y (SD 14.5)

41.5% (17) 58.5% (24)

63.8% (51) 36.3% (29)

48.7% 48.7% 2.6% 85 13 8 17 5 3 3 64.3%

40% (32) 55% (44) 5% (4) 53.5 (IQR 39.2-74.75) 7.5 (IQR 3-24.5) 4 (IQR-2-7) 14 (IQR 10-22.7) 4 (IQR 2-9) 0 (IQR 0-4) 4 (IQR 3-5.75) 12.5% (10)

P Value .021 .019

.573 (20) (20) (1) (IQR 60.5-107.5) (IQR 7-24) (IQR 4-15) (IQR 14-23) (IQR 3-9) (IQR 0-8.5) (IQR 2-5) (26)

Integrating Palliative Care into Critical Care: A Quality Improvement Study.

Many terminally ill patients experience an increasing intensity of medical care, an escalation frequently not consistent with their preferences. In 20...
197KB Sizes 2 Downloads 3 Views