Volume 32, Issue 6 , Pages 519-531, December 2006
Pain, Demographics, and Clinical Characteristics in Persons Who Received Hospice Care in the United States
Article Outline
Abstract
Pain in persons who receive hospice care is not fully understood. The purpose of this study was to describe the demographics, clinical characteristics, and pain intensity of persons who received hospice care in the United States from 2000–2004. Data for this study were obtained from a provider of hospice pharmacy services and included information about the hospice organization, demographics and clinical characteristics, pain intensity, and opioid analgesic prescribing. Worst pain intensity during the previous 24 hours was assessed using a 0–10 numeric rating scale (0
=none, 10=worst) periodically during hospice care. During the study period, 347,555 persons received hospice services; 55.2% of these persons were female, 87.4% were Caucasian, and mean age was 75.3 years. At least one pain score was available for 156,887 (45.1%) individuals. Among persons with at least one pain score, pain was reported a mean of 2.9 times per person. Overall, mean pain intensity was mild, but severe pain was reported at least once by 20.3% of persons. Of individuals who reported severe pain at least once, mean age was 68.9 years, 53.7% were female, 78.0% were Caucasian, 7.4% were cared for in long-term care hospices, and 73.9% had a primary diagnosis of cancer. Severe pain was reported at least once by a substantial percentage of persons in this study. These data provide insight into pain reported by persons who received hospice care, and may be useful as process indicators of the quality of care.
Key Words: Pain, palliative care, hospice, database analysis, epidemiology, outcomes research
Introduction
Pain is an expected result of many diseases, medical care, surgical interventions, and trauma.1 Despite this expectation, evidence published during the past four decades indicates that pain is often undertreated in a variety of patient populations, including medical and surgical inpatients, persons in intensive care units, minority and elderly persons with cancer, and children.2, 3, 4, 5, 6, 7, 8, 9, 10 This phenomenon is important because undertreated pain contributes to unnecessary suffering and adverse clinical outcomes, and potentially adds to health care costs.11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24
Hospice care exists, in part, to provide outstanding supportive and symptomatic care for dying persons and their families.25 Despite this reason for being, currently available person-level information about pain relief during hospice care is limited. For instance, in the most recent National Home and Hospice Care Survey (NHHCS) conducted in 2000, information was collected about the demographics of persons who received hospice care, the length of time these individuals received services, and the types of resource use incurred, but not on the frequency and intensity of common symptoms, including pain.26 Similarly, in the 2002 National Hospice and Palliative Care Organization (NHPCO) data set, although patient and hospice demographic information is included, data about pain were also limited to the percentage of persons whose pain was at a comfortable level within 48 hours of admission.27 Other studies of pain in hospice patients generally emphasize the overall effectiveness of hospice care rather than describing patients' pain.28, 29, 30
The purpose of this study was to describe the overall demographics, clinical characteristics, and pain intensity of persons who received hospice care in the United States. These data are expected to be important for several reasons. First, by providing insight into pain in terminally ill individuals, data of this type can help reduce the unnecessary suffering at the end of life that is relatively common in the United States.31 Second, the Centers for Medicare and Medicaid Services have proposed revisions to conditions of participation that would require hospices to develop, implement, and maintain data-driven quality assurance programs. Third, a more detailed understanding of pain in hospice patients will help clinical decision makers to better understand how to use scarce resources.32
Methods
Data for this retrospective cohort study were derived from administrative information collected by a national pharmaceutical provider specializing in hospice care. Pharmacists and technicians collected these data prospectively from hospice nurses and staff during each patient's most recent episode of hospice care, and included information pertaining to the hospice organization, patient demographics, opioid analgesic use, pain intensity scores, and primary diagnoses. Telephone encounters between pharmacists and nurses or hospice staff occurred for a variety of reasons, including hospice admission or discharge, changes in medication orders or site of hospice care, request for a refill or renewal of an opioid analgesic prescription, or the patient's death. This administrative and clinical data set has existed since February 1, 2000, and the deidentified sample of the data used in these analyses was created on December 12, 2004. The study protocol was approved by the University of Washington Human Subjects Committee.
Construction of the analytic data set from the raw data is shown in Fig. 1. The sample for this analysis included all persons in the data set who were discharged from hospice care alive or who were deceased by December 12, 2004. We sought to describe persons whose hospice experience was complete; individuals who were still active hospice patients as of December 12, 2004 were excluded. Individuals were included in the analytic data set unless the data for those persons was not applicable to these analyses (i.e., still alive as of December 12, 2004, not a hospice patient) or if significant data were missing (i.e., no date of birth). Persons were also categorized by the setting of care received, either home-based hospice or long-term care (LTC) (nursing home) hospice. Age was calculated relative to the last day of hospice care (the date of death or of discharge from hospice) plus one day so that any persons who died the same day they were born would be considered one-day-old, and to be consistent with the methods used in the 2000 NHHCS.26 Similarly, the length of the hospice care episode was calculated as the number of days from the starting date of care until the date of death or discharge plus one day to insure that persons who entered and were discharged from hospice care on the same day were considered to have received hospice care for one day. The primary diagnosis refers to the International Classification of Disease, 9th revision, Clinical Modification (ICD-9-CM) code for the person's terminal illness.33, 34 Most persons had only one primary diagnosis. Individuals who had more than one primary diagnosis were categorized as having a primary diagnosis of cancer if any of those diagnostic codes corresponded to a malignant neoplasm. While an ICD-9-CM code is composed of three digits to the left of a decimal point and up to two digits to the right of the decimal point (i.e., XXX.XX), the analyses of primary diagnoses reported here were limited to the three digits to the left of the decimal point, unless otherwise noted. The patient's geographic region was defined using the U.S. Census Bureau map of regions and divisions.35
Fig. 1
Flow diagram of creation of analytic data set.
Hospice patients periodically reported their worst pain intensity during the previous 24 hours to hospice nurses using a 0–10 numeric rating scale (NRS; 0=no intensity, 10=worst pain imaginable). These scales are valid and reliable tools to assess pain.36, 37 Pain intensity ratings on this scale were also categorized as no (NRS 0), mild (NRS 1–3), moderate (NRS 4–6), or severe (NRS 7–10). The first pain score for an individual was the one that was first reported to the pharmacy services provider after hospice care began. Similarly, the last pain score for a person was the one that was reported closest to the end of the hospice admission.
Opioids are the cornerstone of therapy for moderate to severe pain. Therefore, we focused our analyses on the drugs in this class. For instance, codeine-containing cough preparations were not included. Analgesic use was assessed by counting persons for whom an opioid was dispensed at least once during the study period and the number of times specific opioids were listed in the data set. Neither the specific reason a drug was dispensed nor doses used by the individual were available in this data set.
Results
Construction of Analytic Data Set (Fig. 1)
Of the 406,413 persons included in the raw data files, 23,068 individuals (5.7%) were removed from the data set because they were missing age or hospice information, were admitted to hospice prior to February 1, 2000 or discharged alive from hospice care after December 12, 2004, had an impossible primary diagnosis (i.e., males with uterine cancer or females with prostate cancer), or received nonhospice care from this provider. The data set used in these analyses included 347,555 persons who had died or been discharged from hospice care by December 12, 2004.
Demographics of Discharged Persons (Table 1)
Individuals in this data set received hospice care through 550 hospice organizations,approximately 15% of hospice programs in the United States in 2004.36 Most persons who received care were female, Caucasian, and had received home-based hospice care in the Northeast or Southern United States. At the time of discharge or death, the mean age of these individuals was 76 years, and 80% of persons were at least 60 years of age. Half of persons who received home-based hospice care and 65% of individuals who received LTC hospice care were female.
Table 1. Demographics of Discharged Persons (n=347,555)
| excelleRx, n (%) | 2000 National Home and Hospice Care Survey (%) | 2004 National Hospice and Palliative Care Organization (%) | ||
|---|---|---|---|---|
| Sex | ||||
| Female | 182,037 (52.4) | 50.2 | 56.7 | |
| Male | 151,709 (43.6) | 49.8 | 43.3 | |
| Unknown | 13,809 (4.0) | 0.0 | 0.0 | |
| Patient type | ||||
| Home-based hospice | 295,750 (85.1) | 53.4 (on admission) | ||
| LTC (nursing home) hospice | 51,805 (14.9) | 22.8 (on admission) | ||
| Patient status | ||||
| Deceased | 287,506 (82.7) | 85.5 | ||
| Discharged alive | 60,049 (17.3) | 14.5 | ||
| Age at discharge from hospice | ||||
| Mean age (years; median, SD, range) | 75.8 (78.4, 14.1, 0–105) | |||
| 2003 National Hospice & Palliative Care Organization | ||||
| Pediatrics (0–24 years of age) | 373 (0.1) | 0.36 | ||
| All persons | ||||
| 0–17 years of age | 1,155 (0.3) | 0.4 | ||
| 18–34 years of age | 2,140 (0.6) | 0.7 | ||
| 35–64 years of age | 66,376 (19.1) | 16.8a | ||
| 65–74 years of age | 69,460 (20.0) | 17.6a | ||
| 75–84 years of age | 111,982 (32.2) | 31.4a | ||
| 85+ years of age | 96,442 (27.8) | 33.3a | ||
| NHHCS categories | ||||
| Under 45 years of age (n, %) | 10,536 (3.0) | 3.9 | ||
| 45–64 years of age (n, %) | 59,135 (17.0) | 16.5 | ||
| Under 65 years of age (n, %) | 69,671 (20.0) | 20.4 | ||
| 65 years of age or older (n, %) | 277,884 (80.0) | 79.6 | ||
| 65–69 years of age (n, %) | 28,785 (8.3) | 10.1 | ||
| 70–74 years of age (n, %) | 40,675 (11.7) | 14.5 | ||
| 75–79 years of age (n, %) | 53,170 (15.3) | 12.5 | ||
| 80–84 years of age (n, %) | 58,812 (16.9) | 15.9 | ||
| 85+ years of age (n, %) | 96,442 (27.8) | 26.5 | ||
| Census region | ||||
| Northeast | 102,036 (29.5) | 20.1 | ||
| Midwest | 55,429 (16.0) | 25.7 | ||
| South | 131,108 (37.9) | 33.7 | ||
| West | 57,520 (16.6) | 20.6 | ||
| Unknown | 1,462 (0.4) | 0.0 | ||
| Race and ethnicity (347,555 total; 214,609 nonmissing) | Total | % without missing data | 2000 National Home and Hospice Care Survey (%) | 2004 National Hospice and Palliative Care Organization (%) |
| Asian or Pacific Islander | 1,911 (0.6) | 0.9 | 1.7 | |
| African-American | 17,701 (5.1) | 8.2 | 8.1 | 8.2 |
| Hispanic or Latino | 6,335 (1.8) | 3.0 | 3.8 | 6.3 |
| Native American | 246 (0.1) | 0.1 | 0.3 | |
| Caucasian | 187,386 (53.9) | 87.3 | 84.1 | 83.5 |
| Other | 1,030 (0.3) | 0.5 | 6.4 | |
| Missing or otherwise unknown | 132,946 (38.2) | — | 5.5 | — |
a2004 NHPCO estimates. |
Overall, the demographic profile of persons in the data set was similar to that reported in the 2000 NHHCS and the 2003–2004 NHPCO estimates.26, 38, 39, 40 The main differences observed between these data and other national estimates were related to the site and geographic location of hospice care and individuals. Specifically, 85% of study patients received home-based care compared to 53.4% of persons who were at home when they were admitted to hospice according to the 2004 NHPCO data.40 Additionally, compared to the NHHCS data, a greater percentage of persons in the current study were located in Northeastern and Southern states.26
Primary Diagnosis on Admission (Table 2)
Primary diagnosis information was available for 323,080 discharged or deceased persons. Cancer was the primary diagnosis for 55% of these individuals, and within this group, a malignant neoplasm of trachea, bronchus, or lung (ICD-9-CM 162 or 197.0) was the most common site of disease, followed by the large intestine or rectum (ICD-9-CM 153–154), and breast (ICD-9-CM 174–175). In comparison, 57.5% of persons in the 2000 NHHCS and 50% of persons in the 2002 NHPCO national data set had a primary diagnosis of cancer.26, 27 Other common primary diagnoses in the sample were circulatory diseases, respiratory diseases (mainly chronic obstructive pulmonary disease), and unspecified debility (ICD-9-CM 799.3).
Table 2. Primary Diagnoses on Admission for Discharged Persons
| Primary Diagnosis (ICD-9-CM code) | exRx (%)a, (n=347,555) | 2000 National Home and Hospice Care Survey, (n=621,100, %)b | 2004 National Hospice and Palliative Care Organizationc, (%) |
|---|---|---|---|
| Infectious and parasitic diseases (001–139) | 3,211 (1.0) | 1.8† | |
| Human immunodeficiency virus (HIV) disease (042) | 1,743 (0.5) | 1.5† | 0.7 |
| Neoplasms (140–239) | 180,401 (55.8) | 58.4 | |
| Malignant neoplasms (140–208, 230–234) | 178,218 (55.2) | 57.5 | 45.9 |
| Large intestine & rectum (153–154) | 16,457 (5.1) | 8.3 | |
| Trachea, bronchus, & lung (162, 197.0) | 44,022 (13.6) | 19.4 | |
| Bone, connective tissue, & skin (170–173) | 5,264 (1.6) | 1.7† | |
| Breast (174–175) | 11,626 (3.6) | 2.6 | |
| Female genital organs (179–184) | 8,058 (2.5) | 2.5† | |
| Prostate (185) | 8,184 (2.5) | 3.3 | |
| Urinary organs (188–189) | 7,928 (2.5) | 2.5 | |
| Hematopoietic tissue (196, 200–208) | 11,156 (3.4) | 3.6 | |
| Other and unspecified sites | 73,581 (22.8) | ||
| All noncancer diagnoses | 144,862 (44.8) | ||
| Endocrine, nutritional, & metabolic disorders & immunity disorders (240–279) | 802 (0.2) | † | |
| Diabetes mellitus (ICD-9-CM 250) | 231 (0.1) | † | |
| Mental disorders (ICD-9-CM 290–319) | 14,680 (4.5) | 3.8 | |
| Diseases of the nervous system & sense organs (320–389) | 17,032 (5.3) | 5.2 | |
| Alzheimer's disease (331.0) | 8,172 (2.5) | 2.7† | |
| Dementia (all types) | 8.9 | ||
| Motorneuron disease (335.2) | 1,579 (0.5) | 1.9 (not including amyotrophic lateral sclerosis) | |
| ALS(335.20) | 1,467 (0.5) | 0.3 | |
| Diseases of the circulatory system (390–459) | 46,007 (14.2) | 11.7 | |
| Heart disease (391–392, 393–398, 402, 404, 415–416, 420–427, 428–429) | 30,845 (9.6) | 6.8 | 12.2 |
| Ischemic heart disease (410–414) | 1,960 (0.6) | † | |
| Heart failure (428) | 15,942 (4.9) | 3.8 (congestive heart failure ICD-9-CM 428.0) | |
| Cerebrovascular disease (430–438) | 11,142 (3.4) | 2.7 | Stroke/coma: 2.7 |
| Other diseases of the circulatory system (390, 403, 405, 417, 440–459) | 4,142 (1.3) | 4.8 | |
| Diseases of the respiratory system (460–519) | 19,966 (6.2) | 6.9 | 7.1 |
| Chronic obstructive pulmonary disease & allied conditions (490–496) | 726 (0.2) | 4.4 | |
| Diseases of the digestive system (520–579) | 6,253 (1.9) | 1.9† | Liver: 1.7 |
| Diseases of the genitourinary system (580–629) | 6,987 (2.2) | 1.2† | Kidney: 3.1 |
| Diseases of the musculoskeletal system and connective tissue (710–739) | 324 (0.1) | † | |
| Symptoms, signs, and ill-defined conditions (780–799) | 23,438 (7.2) | 5.6 | |
| Debility unspecified (799.3) | 15,928 (4.9) | 8.3 | |
| Unknown, other, or no primary diagnosis (% of all discharged persons) | 28,454 (8.8) | † | 7.2 |
aICD-9-CM codes were limited to numbers to the left of the decimal, except where noted otherwise. |
b2000 National Home and Hospice Care Survey. The † indicates that the figure does not meet standard of reliability or precision because sample size is less than 30 if shown without an estimate. If shown with an estimate, the sample size is between 30 and 59, or the sample size is greater than 59, but has a relative standard error of 30% or more. All digits of the ICD-9-CM codes are included. |
cDefinitions of NHPCO categories are as follows: Cancer: all cancers; Dementia: Alzheimer's disease, vascular dementia, and so on; Debility Unspecified: terminal debility, failure to thrive; Heart: all heart disease including congestive heart failure and primary sclerotic heart disease; HIV: all AIDS and HIV-related conditions; Kidney: end-stage renal disease; Liver: cirrhosis, advanced hepatitis, and other noncancer liver disease; Lung: chronic obstructive pulmonary disease and other noncancer lung diseases; Motorneuron: Parkinson's disease, Huntington's disease, multiple sclerosis, amyotrophic lateral sclerosis and stroke/coma categories were not further defined. |
Length of Hospice Service (Table 3)
The mean length of hospice service (LOS) among persons in this study was 60.8 days, with a median of 20 days. Nearly 60% of individuals had an LOS less than 30 days, and 23% died within a week of being admitted. Six percent of persons died 180 days or longer after starting hospice care. In comparison, in the 2000 NHHCS, the mean LOS was 47 days, with a median of 16 days, and 62.8% of persons had an LOS less than 30 days.26 In the 2003 NHPCO trend report, the mean LOS was 55.6 days (median 50.6), 32% of persons died within a week of being admitted, and 6% died 180 days or longer after beginning hospice care.39
Table 3. Length of Hospice Service for Discharged Persons (n, %)
| excelleRx, (n=347,555) | 2000 National Home and Hospice Care Survey, (n=621,100) | 2004 National Hospice and Palliative Care Organization | |
|---|---|---|---|
| Mean LOS (days; median, SD, range) | 60.8 (20, 109, 1–1,660) | 47 (16) | 56.8 (21.7) |
| Persons with LOS < 30 days | 203,828 (58.6) | 62.8% | |
| Persons who died within 7 days of admission (mean, median) | 79,609 (22.9) | 35.1% (30.7%) | |
| Persons with LOS ≥ 30 days | 143,727 (41.4) | 37.2% | |
| Persons with LOS ≥ 180 days | 31,039 (8.9) | ||
| Persons who died 180 days or longer after admission | 21,475 (6.2) | 9.2% (6.4%) |
Length of stay by primary diagnosis followed a similar pattern. Persons with cancer had a mean LOS of 52 days (median 21 days), compared to a mean of 43 days (median 18 days) in persons in the 2000 NHHCS and 41 days (median 40.9 days) in the 2003 NHPCO report.26, 39 The mean LOS for persons with all noncancer primary diagnoses was 70.6 days (median 19, SD 127.7). Average LOS also was 5 days longer in persons who received LTC hospice care compared to home-based hospice care, 8.5 days shorter for persons of Asian or Pacific Islander descent relative to persons of other races or ethnicity, and 11 days longer for females than males.
Pain Intensity (Table 4, Table 5)
At least one pain intensity score was available for 156,887 individuals (45.1% of all persons in this data set). Among persons with at least one reported pain intensity score, pain was reported an average of three times per person, or approximately once per week during a hospice admission of median length. Persons with at least one reported pain intensity score and who received LTC care had pain documented twice on average. Mean first and last reported pain scores were in the mild range. Overall, moderate to severe pain (NRS 4–10) accounted for 24.8% and 25.9% of first and last reported pain scores, respectively.
Table 4. Pain Intensity Scores for Discharged Persons (n, %)
| All Persons, (n=347,555) | Home Hospice, (n=295,750) | Long Term Care, (n=51,805) | |
|---|---|---|---|
| Mean number of times pain was reported per person (median, SD, range) | 1.3 (0, 2.4, 0–48) | 1.4 (0, 2.5, 0–48) | 0.8 (0, 1.5, 0–28) |
| Mean number of times pain was reprted per person for whom pain was reported at least once (n=156,887, median, SD, range) | 2.9 (2, 2.8, 1–48) | 3.0 (2, 2.9, 1–48) | 1.9 (1, 1.9, 1–28) |
| Number of persons who had pain reported (n, %) | |||
| 0 times | 190,668 (54.9) | 160,495 (54.3) | 30,173 (58.2) |
| 1 time | 63,440 (18.2) | 49,751 (16.8) | 13,689 (26.4) |
| 2 times | 33,777 (9.7) | 29,666 (10.0) | 4,111 (7.9) |
| 3 times | 20,161 (5.8) | 18,561 (6.3) | 1,600 (3.1) |
| 4 times | 12,490 (3.6) | 11,652 (3.9) | 838 (1.6) |
| 5 or more times | 27,019 (7.8) | 25,625 (8.7) | 1,394 (2.7) |
| Mean initial pain score for persons with only 1 pain score (median, SD, range) | n=63,440; 2.0 (0, 2.8, 0–10) | n=49,751; 2.0 (0, 2.9, 0–10) | n=13,689; 1.6 (0, 2.6, 0–10) |
| Initial pain intensity score rating by numeric rating scale range for persons with only 1 pain score (NRS; n, %) | n=63,440 | n=49,751 | n=13,689 |
| None (NRS 0) | 36,108 (56.9) | 27,528 (55.3) | 8,580 (62.7) |
| Mild (NRS 1–3) | 11,613 (18.3) | 9,307 (18.7) | 2,306 (16.9) |
| Moderate (NRS 4–6) | 9,228 (14.6) | 7,510 (15.1) | 1,718 (12.6) |
| Severe (NRS 7–10) | 6,491 (10.2) | 5,406 (10.9) | 1,085 (7.9) |
| Mean initial pain intensity score rating for persons with 2 or more pain scores (median, SD, range) | n=93,447; 2.1 (0, 2.9, 0–10) | n=85,504; 2.1 (0, 2.9, 0–10) | n=7,943; 1.8 (0, 2.8, 0–10) |
| Initial pain intensity score rating by NRS range (n, %) for persons with two or more pain scores | n=93,447 | n=85,504 | n=7,943 |
| None (NRS 0) | 51,569 (55.2) | 46,782 (54.7) | 4,787 (60.3) |
| Mild (NRS 1–3) | 17,649 (18.9) | 16,330 (19.1) | 1,319 (16.6) |
| Moderate (NRS 4–6) | 13,928 (14.9) | 12,914 (15.1) | 1,014 (12.8) |
| Severe (NRS 7–10) | 10,301 (11.0) | 9,478 (11.1) | 823 (10.4) |
| Mean last pain score for persons with two or more pain scores (median, SD; n=93,447, %) | 2.2 (0, 2.9, 0–10) | 2.2 (0, 2.9, 0–10) | 2.0 (0, 2.9, 0–10) |
| Last pain intensity score rating by numeric rating scale range for persons with two or more pain scores (NRS; n, %) | n=93,447 | n=85,504 | n=7,943 |
| None (NRS 0) | 50,271 (53.8) | 45,785 (53.6) | 4,486 (56.5) |
| Mild (NRS 1–3) | 17,459 (18.7) | 16,073 (18.8) | 1,386 (17.4) |
| Moderate (NRS 4–6) | 14,955 (16.0) | 13,740 (16.1) | 1,215 (15.3) |
| Severe (NRS 7–10) | 10,762 (11.5) | 9,906 (11.6) | 856 (10.8) |
| Number of persons with at least one pain score who reported severe pain at least once (n, %) | n=156,887 | n=135,255 | n=21,632 |
| Mean number of times pain was reported by primary diagnosis of cancer (median SD, range)? | 34,114 (21.7) | 31,070 (23.0) | 3,044 (14.1) |
| Yes | 1.6 (1, 2.6, 0–48, n=178,218) | 1.6 (1, 2.6, 0–48, n=165,933) | 1.2 (1, 2.1, 0–27, n=12,265) |
| No | 1.1 (0, 2.1, 0–38, n=144,862) | 1.3 (0, 2.3, 0–38, n=107,027) | 0.7 (0, 1.3, 0–28, n=37,835) |
| Missing | 0.4 (0, 1.4, 0–33, n=24,475) | 0.4 (0, 1.5, 0–33, n=22,270) | 0.2 (0, 1.0, 0–13, n=1,705) |
Table 5. Mean Number of Times Severe Pain Was Reported (mean, median, SD, range)
| All Persons | Persons with at least 1 Pain Score | Persons with at least 1 Severe Pain Score (NRS 7–10) | |
|---|---|---|---|
| Was primary diagnosis cancer? | |||
| Yes | 0.2 (0, 0.7, 0–21, n=178,218) | 0.4 (0, 0.9, 0–21, n=89,422) | 1.5 (1, 1.1, 1–21, n=24,808) |
| No | 0.1 (0, 0.4, 0–25, n=144,862) | 0.2 (0, 0.6, 0–25, n=63,228) | 1.4 (1, 0.9, 1–25, n=8,507) |
| Missing | 0.0 (0, 0.3, 0–15, n=24,475) | 0.2 (0, 0.7, 0–15, n=4,237) | 1.3 (1, 1.0, 1–15, n=799) |
| Race | |||
| African American | 0.2 (0, 0.7, 0–11, n=17,701) | 0.4 (0, 0.8, 0–11, n=11,797) | 1.5 (1, 1.0, 1–11, n=2,833) |
| Asian or Pacific Islander | 0.2 (0, 0.6, 0–8, n=1,911) | 0.3 (0, 0.7, 0–8, n=1,243) | 1.4 (1, 0.9, 1–8, n=262) |
| Caucasian | 0.2 (0, 0.6, 0–25, n=187,386) | 0.3 (0, 0.8, 0–25, n=121,527) | 1.5 (1, 1.1, 1–25, n=26,260) |
| Hispanic | 0.2 (0, 0.7, 0–11, n=6,335) | 0.4 (0, 0.8, 0–11, n=4,244) | 1.5 (1, 1.0, 1–11, n=1,077) |
| Native American | 0.2 (0, 0.9, 0–10, n=246) | 0.4 (0, 1.0, 0–10, n=163) | 1.7 (1, 1.7, 1–10, n=34) |
| Other | 0.2 (0, 0.7, 0–8, n=1,030) | 0.4 (0, 0.9, 0–8, n=620) | 1.5 (1, 1.2, 1–8, n=152) |
| Missing or unknown | 0.0 (0, 0.3, 0–15, n=132,946) | 0.3 (0, 0.7, 0–15, n=17,293) | 1.4 (1, 0.9, 1–15, n=3,496) |
| Patient type | |||
| Hospice | 0.2 (0, 0.6, 0–25, n=295,750) | 0.3 (0, 0.8, 0–25, n=135,255) | 1.5 (1, 1.1, 1–25, n=31,070) |
| LTC | 0.1 (0, 0.4, 0–13, n=51,805) | 0.2 (0, 0.5, 0–13, n=21,632) | 1.3 (1, 0.8, 1–13, n=3,044) |
| Sex | |||
| Female | 0.1 (0, 0.6, 0–25, n=182,037) | 0.3 (0, 0.8, 0–25, n=84,215) | 1.5 (1, 1.0, 1–25, n=18,031) |
| Male | 0.2 (0, 0.6, 0–21, n=151,709) | 0.3 (0, 0.8, 0–21, n=72,662) | 1.5 (1, 1.0, 1–21, n=16,082) |
| Unknown | 0.0 (0, 0.0, 0–2, n=13,809) | 0.2 (0, 0.6, 0–2, n=10) | 2.0 (2, NA, 2–2, n=1) |
| Age | |||
| Under 65 years of age | 0.3 (0, 0.8, 0–25, n=69,671) | 0.6 (0, 1.1, 0–25, n=34,239) | 1.7 (1, 1.3, 1–25, n=11,606) |
| 65 years of age or older | 0.1 (0, 0.4, 0–15, n=277,884) | 0.2 (0, 0.6, 0–15, n=122,648) | 1.4 (1, 0.9, 1–15, n=22,508) |
Among persons with at least one reported pain score, 23% of persons who received home-based care and 14% of individuals who received LTC care reported having severe pain at least once. Persons with a primary diagnosis of cancer had pain scores reported more often than persons without cancer. In addition, severe pain was more common among persons with cancer than among persons with other primary diagnoses and in individuals under 65 years of age compared to older persons.
Opioid Analgesic Use (Table 6, Table 7)
Opioid analgesics were dispensed at least once to 28.4% of persons in the sample. These drugs were dispensed a total of 338,604 times, indicating that some persons were prescribed more than one opioid, used more than one strength of the same drug, or had a drug prescribed more than once. The drugs dispensed most often were morphine, oxycodone, and fentanyl. Methadone was dispensed for 8% of persons, and codeine, hydrocodone, or propoxyphene were dispensed to 0.9%, 2.0%, and 3.2% of individuals, respectively (Table 6).
Table 6. Opioid Analgesic Use Among Persons Who Used an Opioid at least Once (n=98,864)
| Drug | Persons Who Used a Specific Opioid (n=122,023, %) | Total Mentions of Drug (n=338,604, %) |
|---|---|---|
| Butorphanol | 2 (0.002) | 2 (0.001) |
| Codeine | 1,083 (0.9) | 2,345 (0.7) |
| Fentanyl | 34,047 (27.9) | 93,342 (27.6) |
| Hydrocodone | 2,435 (2.0) | 5,081 (1.5) |
| Hydromorphone | 801 (0.7) | 1,726 (0.5) |
| Levorphanol | 2 (0.002) | 4 (0.001) |
| Meperidine | 8 (0.007) | 28 (0.01) |
| Methadone | 1,633 (1.3) | 4,753 (1.4) |
| Morphine | 38,692 (31.7) | 96,817 (28.6) |
| Opium | 569 (0.5) | 1,244 (0.4) |
| Oxycodone | 38,307 (31.4) | 123,948 (36.6) |
| Pentazocine | 5 (0.004) | 10 (0.003) |
| Propoxyphene | 3,925 (3.2) | 8,061 (2.4) |
| Tramadol | 514 (0.4) | 1,243 (0.4) |
Table 7. Opioid Analgesic Use by Demographics (n, %)
| Yes (n=98,864) | |
|---|---|
| Sex | |
| Female (n=182,037) | 51,958 (28.5) |
| Male (n=151,709) | 46,563 (30.7) |
| Unknown (n=13,809) | 340 (2.5) |
| Patient type | |
| Hospice (295,750) | 93,419 (31.6) |
| Long-term care (51,805) | 5,445 (10.5) |
| Age | |
| Under 65 years of age (n=69,671) | 27,283 (39.2) |
| Over 65 years of age (n=277,884) | 71,581 (25.8) |
| Race | |
| Asian or Pacific Islander (n=1,911) | 690 (36.1) |
| African American (n=17,701) | 6,703 (37.9) |
| Caucasian (n=187,386) | 63,831 (34.1) |
| Hispanic (n=6,335) | 2,440 (38.5) |
| Native American (n=246) | 94 (38.2) |
| Other (n=1,030) | 360 (35.0) |
| Unknown (n=132,946) | 24,746 (18.6) |
| Geographic location | |
| Northeast (n=102,036) | 29,958 (29.4) |
| Midwest (n=55,429) | 16,119 (29.1) |
| South (n=131,108) | 38,059 (29.0) |
| West (n=57,520) | 14,680 (25.5) |
| Unknown (n=1,462) | 48 (3.3) |
| Primary diagnosis is cancer | |
| Yes (n=178,218) | 68,559 (38.5) |
| No (n=144,862) | 25,989 (17.9) |
| Unknown (n=24,475) | 4,316 (17.6) |
Opioids were dispensed more frequently for males, persons who received home-based hospice care, or persons who were under 65 years of age, Hispanic individuals, persons in Northeastern states, and persons with a primary diagnosis of cancer (Table 7). Among persons who received LTC hospice care, opioids were dispensed for 10.5%, compared to 31.6% of persons who received home-based hospice care.
Discussion
In this study of a large sample of persons who received hospice care in the United States from 2000–2004, we present the first estimates from a new, population-based data set. To our knowledge, these estimates represent the first detailed estimates of pain in persons who received hospice care. Because these data represent the first use of this data set, it is important to describe the data and estimate its validity compared to other national data sources. In addition, there are some differences between the current data set and the NHHCS and NHPCO data sets. Specifically, although demographic information is included as part of each data set, the current data include more recent demographic information about this population, as well as medication use and longitudinal estimates of pain intensity. Primary diagnosis data in the data set is also categorized by ICD-9-CM code, rather than by descriptive terms, as used in the NHPCO data set. As a result, we believe the current data represent a unique source that is suitable for health services research.
The demographic and clinical profile of persons in the current analyses was similar to the estimates derived from the 2000 NHHCS and the NHPCO National Data Set.26, 27, 38, 39, 40 Most persons who received hospice care were white, and the percentage of African American, Asian, and Hispanic persons who received hospice care in this study was lower than the proportion of these minority groups in the 2000 United States population.41 There are many possible reasons for these differences, including that attitudes toward hospice and other care at the end of life may differ between racial and ethnic groups and that hospice care may be less readily available to persons in minority groups.
The length of hospice service experienced by persons in the current study was approximately 25% longer than that reported in the 2000 NHHCS.26 Among the approximately 97% of persons in this study whose LOS was a year or less, the mean LOS of 47.6 days and median of 19 days was nearly identical to the NHHCS estimate.
At least one NRS score of a person's worst pain intensity in the past 24 hours was available for fewer than half of persons in this study. Among persons with at least one pain score, pain was reported a mean of three times (or approximately once a week assuming an LOS of median duration) in persons who received home-based hospice care, and twice in persons who received LTC hospice care. In addition, pain was reported less frequently for persons without cancer than for persons with cancer. While balancing caregiver and respondent burden is important and the optimal frequency of pain assessment in this population is not known, it is impossible to know how severe a person's pain is without asking. Furthermore, these estimates suggest that although some decrease in monitoring of symptoms may be expected as the end of life approaches, pain may be monitored less closely among persons in LTC hospices and in persons who do not have a primary diagnosis of cancer. Since pain and suffering at the end of life are common in the United States, and given the relatively short period of time that most persons receive hospice care, it makes sense that more frequent monitoring of pain would contribute to earlier identification of persons with pain that is difficult to control.25, 26 Furthermore, because predicting pain in ill persons is difficult at best, the estimates from the current study suggest that pain may be underestimated in hospice patients with few or no reported pain intensity scores, such as persons who did not have cancer or who received care in LTC hospices.42
Pain intensity was relatively mild for most persons at the first and last assessment. However, approximately one-fifth of persons reported severe pain in the previous 24 hours at least once during their hospice experience, particularly among persons with a primary diagnosis of cancer. The inference we can make from occasional ratings of the intensity of pain or other symptoms is relatively limited because pain intensity is just one dimension of a complex physical and emotional experience.43 In addition, persons were asked to rate their worst pain in the previous 24 hours. This is useful as a proxy for more detailed information, but we cannot tell from these data alone whether a person had a single episode of severe breakthrough pain or if their pain was generally severe during that day. Pain intensity ratings in the severe range (NRS 7–10) were more common from persons with cancer as a primary diagnosis and younger individuals. This observation may be due, in part, to how pain is viewed by persons of different ages. Older persons, for instance, may be more likely to believe that pain is a normal part of the aging and dying process, and, therefore, less likely to report it.
Opioid use information was available for 28.4% of persons in this data set, not including use of a comfort pack, which contains a small amount (∼15ml) of morphine solution and other medications. The most frequently prescribed drugs were morphine, oxycodone, and fentanyl. Methadone or hydromorphone was dispensed for a small percentage of persons. Additionally, codeine, meperidine, or propoxyphene was dispensed at least once for approximately 5% of persons who were prescribed an opioid at least once. Although dispensing of these three analgesics was not widespread, it was still somewhat unexpected, given that use of these drugs is widely discouraged due to relatively low potency, the potential for genetic polymorphisms in metabolic enzymes to interfere with codeine activation, and toxic metabolites. We also observed notable differences in opioid dispensing between persons who received home-based hospice care and individuals who received LTC hospice care, and between persons with a primary diagnosis of cancer and individuals with noncancer primary diagnoses.
Lack of information about opioid dispensing for an individual may have occurred for several reasons. The person may not have received any medications in this class, they may have received medications from other pharmacies, or the hospice may have received pharmacy services from some other provider. Evidence in the published biomedical literature suggests that elderly and minority patients with cancer are less likely to be treated with opioid analgesics, although hospice patients are also more likely to receive regular pain medication than nonhospice patients.4, 28, 29, 30 In addition, most of the persons in this data set were elderly, and advancing age may affect how disease presents. For example, as other authors have noted, pain due to some types of cancer may be less intense in older adults compared to younger individuals.44, 45 Out-of-network resource use is a possibility, although there are financial incentives for most hospice patients to get their medications from the hospice. In 2000, Medicare was the primary payer for hospice care, and the Medicare Hospice Benefit includes medications as part of the per-diem rate paid to hospices.26, 46 As a result, persons who chose to get medications from sources other than the hospice or its pharmacy provider would have been responsible for paying for those drugs, similar to the situation for persons with prescription-drug insurance who use out-of-network services. Hospices may have used a different pharmacy provider for some medications, although all but two of the hospices in this data set had contracted with this provider to be the regular source for pharmacy services.
There are some limitations to this work. First, data were collected on an irregular schedule. As a result, data for some hospices and patients are more complete than others. In these analyses, we found no clear data collection pattern. Second, some data are missing. Although all data fields are technically required, data may be missing or incorrect because variables often have an “unknown” or “not specified” response category, the nurse may not have asked the patient to rate their symptoms, the person may not be able to rate their symptoms, or the nurse or hospice staff member may not be willing to answer questions about the patient's demographic characteristics. In addition, technicians or pharmacists may not collect these data when it is available. As a result, some variables, such as race, have a relatively high degree of missingness. Additionally, each telephone encounter between a hospice nurse and pharmacy personnel generally results in additional data collection; however, if the nurse does not call about a patient, no data are collected for that person. Thus, it is impossible to distinguish between missing data that were not collected at all, and data that may have been collected but not shared with the pharmacy provider. Third, although it is reasonable to assume that ICD-9-CM codes were chosen to the best of the coder's ability, the precision of the primary diagnosis is not known. Fourth, drug data must be added to the database by hand. As a result, there is potential for selection and information biases.47 We have attempted to account for these limitations by excluding persons with substantially inconsistent demographic data, comparing these estimates to other national data, and by limiting our diagnosis analyses.
We believe these results have several potentially valuable uses. First, to our knowledge, this analysis is the first describing the pain experience of a cohort of persons who received hospice care. As a result, these data contribute to the understanding of pain in persons who receive hospice care, and are useful for generating hypotheses about ways to improve care for persons at the end of life. For example, using demographics and clinical characteristics may be useful as proxies to estimate possible pharmacogenomic differences in response to opioid analgesics. Second, using the processes of medical care is a common way to estimate the quality of medical care, as in the Health Plan Employer Data and Information Set.48 Furthermore, the pain management standards promulgated by the Joint Commission on Accreditation of Healthcare Organizations include the assessment of the intensity of pain in all patients.49 Thus, although there are presently no indicators to estimate the quality of pain care, these results may contribute to such an effort. Third, given the current emphasis on payment for performance, if health care organizations are able to prove that they can provide improved pain management, they may also be able to negotiate increased reimbursement for doing so.
Conclusions
These data contribute to the body of knowledge about hospice care, and, specifically, about pain management in persons at the end of life. Information of this type is expected to be useful in understanding how to improve care for this population.
Acknowledgments
Scott Strassels was supported by a predoctoral fellowship from excelleRx, Incorporated. The contributions of Calvin J. Alt, Jr., RPh, J. Nile Barnes, PharmD, Kevin Bain, PharmD, BCPS, Michael Groh, PharmD, Terri Maxwell, RN, PhDc, Ewan McNicol, BSPharm, MS, Suzanne Novak, MD, PhD, JoAnne Reifsnyder, PhD, APRN, BC-PCM, and Doug Weschules, PharmD, BCPS, are gratefully acknowledged.
References
- Acute pain management: Operative or medical procedures and trauma. Clinical Practice Guideline No. 1. AHCPR Pub. No. 92-0032. Rockville, MD: Agency for Healthcare Policy and Research, Public Health Service, U.S. Department of Health and Human Services; February 1992;
- . Incidence and characteristics of pain in a sample of medical-surgical inpatients. Pain. 1987;30:69–78
- Pain outcomes after intestinal surgery. Outcomes Manag Nurs Pract. 2001;5:41–46
- Pain and its treatment in outpatients with metastatic cancer. N Engl J Med. 1994;330:592–596
- . Undertreatment of medical inpatients with narcotic analgesics. Ann Intern Med. 1973;78:173–181
- Analysis of narcotic analgesic usage in the treatment of postoperative pain. JAMA. 1983;250:926–929
- . Acute pain management: programs in US hospitals and experiences and attitudes among US adults. Anesthesiology. 1995;83:1090–1094
- . Assessment of patient satisfaction utilizing the American Pain Society's Quality Assurance Standards on acute and cancer-related pain. J Pain Symptom Manage. 1994;9:5–11
- . A controlled trial to improve care for seriously ill hospitalized patients. The study to understand prognoses and preferences for outcomes and risks of treatments (SUPPORT). JAMA. 1995;274:1591–1598
- Symptoms and suffering at the end of life in children with cancer. N Engl J Med. 2000;342:326–333
- Management of cancer pain. Clinical Practice Guideline No. 9. AHCPR Publication 94-0592. Rockville, MD: Agency for Health Care Policy and Research, US Department of Health and Human Services, Public Health Service; March 1994;
- Guideline for the management of acute and chronic pain in sickle-cell disease. American Pain Society Clinical Practice Guidelines Series, No. 1. Glenview, IL: American Pain Society; 1999;
- Guideline for the management of arthritis pain in osteoarthritis, rheumatoid arthritis, and juvenile chronic arthritis. American Pain Society Clinical Practice Guidelines Series, No. 2. Glenview, IL: American Pain Society; 2002;
- Guideline for the management of cancer pain in adults and children. APS Clinical Practice Guidelines Series, No. 3. Glenview, IL: American Pain Society; 2005;
- . Surgical stress: the role of pain and analgesia. Br J Anaesth. 1989;63:189–195
- Preemptive epidural analgesia and recovery from radical prostatectomy. A randomized controlled trial. JAMA. 1998;279:1076–1082
- . Psychological influences on surgical recovery. Perspectives from psychoneuroimmunology. Am Psychol. 1998;53:1209–1218
- . Retrospective evaluation of unanticipated admissions and readmissions after same day surgery and associated costs. J Clin Anesth. 2002;14:349–353
- . What decline in pain intensity is meaningful to patients with acute pain?. Pain. 2003;105(1–2):151–157
- . Importance of effective pain control. Acta Anaesthesiol Scand. 1987;31(Suppl 85):1–16
- . Acute pain. Lancet. 1999;353:2051–2058
- . Effect of postoperative analgesia on surgical outcome. Br J Anaesth. 2001;87:62–72
- . Postoperative analgesia: economics, resource use, and patient satisfaction in an urban teaching hospital. Anesth Analg. 2002;94:130–137
- . Chronic pain as an outcome of surgery—a review of predictive factors. Anesthesiology. 2000;93:1123–1133
- . Access to hospice care: expanding boundaries, overcoming barriers. Hasting Center Report Special Supplement. 2003;33(2):S3–S59Available from http://www.thehastingscenter.org/pdf/access_hospice_care.pdfAccessed June 22, 2005
- . Characteristics of hospice care discharges and their length of service: United States, 2000. National Center for Health Statistics. Vital Health Stat 13. 2003;(154):1–36
- . Measuring hospice care: the National Hospice and Palliative Care Organization National Hospice Data Set. J Pain Symptom Manage. 2004;28:316–328
- . A randomized controlled trial of hospice care. Lancet. 1984;1(8382):890–894
- . Does receipt of hospice care in nursing homes improve the management of pain at the end of life?. J Am Geriatr Soc. 2002;50:507–515
- . Hospice enrollment and pain assessment and management in nursing homes. J Pain Symptom Manage. 2003;26:791–799
- In: Field MJ, Cassel CK editor. Approaching death. Improving care at the end of life. Washington, DC: National Academy Press; 1997;
- Medicare announces new proposed rule to improve hospice care. Available from http://new.cms.hhs.gov/apps/media/press/release.asp?Counter=1469Accessed February 21, 2006
- International classification of diseases, 9th revision, clinical modification (ICD-9-CM). Available from http://www.cdc.gov/nchs/about/otheract/icd9/abticd9.htmAccessed June 21, 2005
- ICD-9-CM. International Classification of Diseases, 9th Revision. Clinical Modification. 5th edition. Los Angeles, CA: Practice Management Information Corporation; 1997;
- Census regions and divisions of the United States. Available from http://www.census.gov/geo/www/us_regdiv.pdfAccessed January 20, 2006
- . The measurement of clinical pain intensity a comparison of six methods. Pain. 1986;27:117–126
- . The subjective experience of acute pain. An assessment of the utility of 10 indices. Clin J Pain. 1989;5:153–159
- NHPCO's 2004 Facts and Figures. Available from http://www.nhpco.org/Accessed January 20, 2006
- National Hospice and Palliative Care Organization. National Trend Summary 2000–2003.
- National Hospice and Palliative Care Organization. 2004 National Summary Statistics and Trends.
- Population by race and Hispanic or Latino Origin for the United States: 1990 and 2000 (PHC-T-1). Available from http://www.census.gov/population/www/cen2000/phc-t1.htmlAccessed February 24, 2006
- . Pain and satisfaction with pain control in hospitalized medical patients: no such thing as low risk. Arch Intern Med. 2004;164:175–180
- . Definition of pain. Available from http://www.iasp-pain.org/terms-p.html#PainAccessed June 21, 2005
- . Visceral pain. Pain: Clinical Updates. 2005;XII(6):1–635
- . An international survey of cancer pain characteristics and syndromes. IASP Task Force on Cancer Pain. International Association for the Study of Pain. Pain. 1999;82(3):263–274
- . Social Security Online. Hospice care; hospice program. Available from www.ssa.govAccessed February 13, 2004
- In: Rothman KJ, Greenland S editor. Modern epidemiology. 2nd ed.. Philadelphia, PA: Lippincott-Raven Publishers; 1998;
- The Health Plan Employer Data and Information Set (HEDIS). Available from http://www.ncqa.org/Programs/HEDIS/index.htmAccessed June 21, 2005
- . Joint commission focuses on pain management. Available from http://www.jcaho.org/news+room/health+care+issues/jcaho+focuses+on+pain+management.htmAccessed June 21, 2005
This study was funded in part by excelleRx, Inc.
PII: S0885-3924(06)00519-7
doi:10.1016/j.jpainsymman.2006.06.005
© 2006 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved.
Volume 32, Issue 6 , Pages 519-531, December 2006
