Appendix J
Commissioned Paper

Palliative Care/End-of-Life Measures

Sydney Dy and Joanne Lynn

INTRODUCTION

Recent advances in medical care and expansion of services offer tremendous potential for reducing suffering and improving quality of life for persons with life-threatening illnesses. However, study after study has demonstrated that these advances have not been translated well into clinical practice and that serious quality deficiencies persist for the care of this population (Teno, 2001). Few palliative care performance measures are included in population-based assessments of quality such as the National Healthcare Quality Report, or even in quality reports focused upon settings with high proportions of palliative care patients, such as nursing homes. Measuring quality for palliative care entails many challenges, including defining the denominator, adjusting for risk, accounting for patient preferences, assessing surrogate respondents, adjusting for differences in length of life arising from treatment choices, and evaluating patient-centered outcomes (Rosenfeld and Wenger, 2000). While measurable processes of care should be tightly linked to desirable outcomes, high-quality evidence of that linkage is quite uncommon in end-of-life care, and elements that reflect patient-centered care can be very difficult to measure.

On the other hand, assessing quality in care for the last years of life also has many opportunities for growth, including recent systematic reviews (Higginson et al., 2003; Lorenz et al., 2004), a national consensus project on clinical guidelines (National Consensus Project, 2004), and a large body of literature addressing the important domains and the development of measurement instruments. Palliative and end-of-life care measures must be



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 287
Performance Measurement: Accelerating Improvement Appendix J Commissioned Paper Palliative Care/End-of-Life Measures Sydney Dy and Joanne Lynn INTRODUCTION Recent advances in medical care and expansion of services offer tremendous potential for reducing suffering and improving quality of life for persons with life-threatening illnesses. However, study after study has demonstrated that these advances have not been translated well into clinical practice and that serious quality deficiencies persist for the care of this population (Teno, 2001). Few palliative care performance measures are included in population-based assessments of quality such as the National Healthcare Quality Report, or even in quality reports focused upon settings with high proportions of palliative care patients, such as nursing homes. Measuring quality for palliative care entails many challenges, including defining the denominator, adjusting for risk, accounting for patient preferences, assessing surrogate respondents, adjusting for differences in length of life arising from treatment choices, and evaluating patient-centered outcomes (Rosenfeld and Wenger, 2000). While measurable processes of care should be tightly linked to desirable outcomes, high-quality evidence of that linkage is quite uncommon in end-of-life care, and elements that reflect patient-centered care can be very difficult to measure. On the other hand, assessing quality in care for the last years of life also has many opportunities for growth, including recent systematic reviews (Higginson et al., 2003; Lorenz et al., 2004), a national consensus project on clinical guidelines (National Consensus Project, 2004), and a large body of literature addressing the important domains and the development of measurement instruments. Palliative and end-of-life care measures must be

OCR for page 287
Performance Measurement: Accelerating Improvement prominent in any national set of quality measures, since such a high proportion of care occurs in patients with life-threatening illness and since deficiencies in quality may cause particular harm in patients with little time or reserve remaining to recover from adverse effects. A national measurement set must consider the unique priorities and challenges of palliative care patients, as many measures associated with improved outcomes in a healthy population may be inappropriate or even harmful in patients with serious illness and limited prognoses. For the purposes of this paper, we will use the World Health Organization definition of palliative care as “an approach that improves the quality of life of patients and their families facing the problems associated with life-threatening illness, through the prevention and relief of suffering by means of early identification and impeccable assessment and treatment of pain and other problems, physical, psychosocial and spiritual” (World Health Organization, 2002). For our conceptual model, we will use the domains of the framework of the Toolkit of Instruments to Measure End of Life Care (Teno, 2000): Pain and other symptoms Emotional and cognitive symptoms Functional status Survival time and aggressiveness of care Advance care planning Continuity of care Spirituality Grief and bereavement Patient-centered reports and rankings (aka satisfaction) with the quality of care Caregiver well-being Quality of life For each domain, where appropriate, we have also organized measures into those applicable to assessment, management, and outcome. We have listed topics in this order in the text and Table J-1, and compared the results of our searches to these categories to determine where there are particular gaps in performance measurement for palliative care. METHODS AND SOURCES We limited our review to measurement sets particularly relevant to palliative care, as more general sets are under review in other parts of this project. We considered information from recent systematic reviews and consensus statements in palliative care, as well as previous reviews of quality indicators

OCR for page 287
Performance Measurement: Accelerating Improvement for palliative care, relevant reports from the Institute of Medicine (Lunney et al., 2003; Teno et al., 2001), and other pertinent books and reports. We also reviewed articles and Web sites from recent RAND initiatives to define performance indicators. We performed Medline searches using the terms “quality indicator” and “performance measure” with the terms “palliative” and “end of life.” Finally, we reviewed Web sites for palliative care standards or indicator initiatives in other countries, including Canada, Australia, and the United Kingdom. MEASURE SETS Palliative Care Leading measurement sets in palliative care are described below, and pertinent measures are included in Table J-1. Dartmouth Atlas Wennberg and colleagues (1999, 2004) have used Medicare administrative data to evaluate a number of potential performance measures and to compare them across geographic regions defined by political division or hospital referral region. For the end-of-life measures, they have tabulated the services that Medicare recipients used in the last 6 months of life, showing wide variation by region and provider. Their measures include the number of days spent in the hospital; number of days spent in the intensive care unit; percentage of patients seeing 10 or more physicians; percentage ever enrolled in hospice; percentage of deaths occurring in the hospital; and percentage of deaths occurring in association with an intensive care unit. We describe several of these measures in more detail in Table J-1. Although the variation in these measures is striking, it is unclear whether those variations correlate with the quality of the end-of-life experience. Fisher et al. (2003a,b) did find that higher levels of resource utilization in the last 6 months of life were not associated with improved mortality or satisfaction for Medicare patients with serious illnesses, measuring regional satisfaction with the Medicare Current Beneficiary Survey. Drawbacks of retrospective analyses of patients who have died are discussed in the section below on challenges of measurement in end-of-life care. Brown Atlas of Dying The Brown Atlas (Teno, 2004) has extended the work of the Dartmouth group by using several additional data sources to examine regional variation in end-of-life care. The Atlas includes site of death information for

OCR for page 287
Performance Measurement: Accelerating Improvement TABLE J-1 Selected Potential Performance Measures for Palliative/End-of-Life Care Domain Category Name of Measure Description Pain Assessment   Pain measurement UHC Chart review Numerator: Patients who had any pain measurement within 48 hours of admission Denominator: Palliative care population hospital admissions Use of numeric pain scale UHC, Brown-QIO, VHA-QIO Chart review Numerator: Patients who had a numeric pain scale used Denominator: Palliative care or other population admissions with a pain score within 48 hours Pain as 5th vital sign VHA-QIO Across all settings Chart review Numerator: Patients who had pain assessed when other vital signs taken Denominator: All patients (unless lesser frequency indicated and documented in chart) Appropriate pain assessment Brown-QIO Assessment of pain intensity, 4 other elements Numerator: Patients with appropriate pain assessment Denominator: All NH residents with pain Treatment   Pain medication prescribed Brown-QIO Any pain medication Numerator: Any pain medication prescribed Denominator: All NH residents with pain Nonpharmacological treatment Brown-QIO Any nonpharmacological treatment in plan of care Numerator: Nonpharmacological treatment Denominator: All NH residents with pain Change in pain medication Brown-QIO Change in pain medication for uncontrolled pain Numerator: Change in pain medication Denominator: NH residents with daily pain and documented moderate-severe pain

OCR for page 287
Performance Measurement: Accelerating Improvement Psychometric Testing (Validity/Reliability) Prior Use References N Benchmarking Multiple settings   N Benchmarking Baier et al., 2004; Cleeland et al., 2003 N Improvement Cleeland et al., 2003 Y—e.g., Brief Pain Inventory Improvement Baier et al., 2004; Lorenz et al., 2004 N Improvement Baier et al., 2004 N Improvement Baier et al., 2004 N Improvement Baier et al., 2004

OCR for page 287
Performance Measurement: Accelerating Improvement Domain Category Name of Measure Description Adherence to guidelines Du Pen Adherence to “best practice” pain guidelines, defined as score of 2.5 on score of 0–3 Numerator: Adherence Denominator: Community oncology patients with pain of 3 or greater on 10-point scale Outcome   Rate of pain VHA-IHI % of patients with moderate-severe pain; various settings Patient perspective Numerator: % of patients with moderate or severe pain Denominator: All patients in setting Rate of pain in nursing homes Brown Atlas % of patients with moderate-severe pain; Collected from Minimum Data Set (MDS) Numerator: % of patients with moderate or severe pain over 7-day lookback period Denominator: All nursing home patients Persistent pain in nursing homes Brown Atlas % of nursing home patients with persistent pain Numerator: patients who still have moderate or excruciating pain on 2nd assessment 60–180 days after admission Denominator: Nursing home patients with pain on 1st assessment. Subgroups: persons with documented terminal illness; persons cognitively intact and able to report on their pain; patients with cancer Comfortable dying NDS % of patients whose pain was brought to a comfortable level within 48 hours of admission to hospice Numerator: patients answering that pain was brought to a comfortable level within 48 hours Denominator: patients uncomfortable due to pain on admission, able to self-report, and ≥18 years of age Pain relieved/reduced UHC Hospital Chart review Numerator: Pain relieved/reduced to <3/10 within 48 hours of admission Denominator: Palliative care population reporting pain on hospital admission Satisfaction Du Pen Satisfaction with current pain treatment; patients who would choose to have similar treatment again Numerator: Patients satisified with current pain treatment Denominator: Patients treated for pain

OCR for page 287
Performance Measurement: Accelerating Improvement Psychometric Testing (Validity/Reliability) Prior Use References N Improvement. Adherence was greater in intervention group and associated with reduced pain scores Du Pen et al., 1999 N Improvement Cleeland et al., 2003 MDS pain reporting has substantial validity issues. Currently undergoing further development as a CMS demonstration project Reporting, Improvement Teno et al., 2004; Baier et al., 2004 N Benchmarking Teno et al., 2002; Teno, 2004 N Benchmarking Connor et al., 2004 N   Y Improvement. Rates higher in intervention group Du Pen et al., 1999

OCR for page 287
Performance Measurement: Accelerating Improvement Domain Category Name of Measure Description Dyspnea Assessment   Dyspnea assessment UHC Dyspnea assessment within 24 hours of admission Hospital Chart review Numerator: Patients assessed for dyspnea within 24 hours of admission Denominator: Palliative care population admissions Outcome   Dyspnea relieved/reduced UHC Dyspnea relieved/reduced within 48 hours of admission Hospital Chart review Numerator: Patients with dyspnea reduced/relieved to ≤3 within 48 hours of admission Denominator: Patients with documented dyspnea Constipation Treatment   Bowel regimen UHC Bowel regimen within 24 hours of opioid administration Hospital Chart review Numerator: Patients with bowel regimen ordered within 24 hours or bowel regimen contraindicated Denominator: Palliative care population admissions started on opioids Emotional and cognitive symptoms Assessment   Depression and comorbid disease ACOVE Depression Screening for depression with new onset of serious comorbid conditions Community Numerator: Patient asked about or treated for depression or referred to mental health professional within 2 months of diagnosis of condition Denominator: Vulnerable elders who present with new onset of serious comorbid conditions, including malignancy Treatment   Recognizing depression ACOVE Depression Evaluation/treatment for depression if presents with depressive symptoms Community Numerator: Patient asked about or treated for depression or referred to mental health professional within 2 weeks of presentation Denominator: Vulnerable elders who present with new onset of symptoms of potential depression

OCR for page 287
Performance Measurement: Accelerating Improvement Psychometric Testing (Validity/Reliability) Prior Use References N   N N Tested in managed care organizations as part of ACOVE measurement set Benchmarking Nakajima and Wenger, 2003   Nakajima and Wenger, 2003

OCR for page 287
Performance Measurement: Accelerating Improvement Domain Category Name of Measure Description Care planning Process   Documentation of patient status UHC Documentation of all 4 aspects of patient status within 48 hours of admission: prognosis, functional status, psychosocial symptoms, symptom distress Numerator: Patients with all 4 aspects documented within 48 hours Denominator: Palliative care admissions Patient/family meeting UHC Patient/family meeting within 1 week of admission. Defined as documented discussion of patient preferences/plans for discharge disposition Hospital Chart review Numerator: Patients with patient/family meeting documented within 1 week of admission Denominator: Palliative care admissions Discharge planning UHC Plan for discharge disposition documented within 4 days of admission Hospital Chart review Numerator: Patients with discharge disposition documented within 4 days of admission Denominator: Palliative care population admissions Use of discharge planner UHC Discharge planner/social services arranged services required for discharge Hospital Chart review Numerator: Patients where discharge planner/social services arranged services required for discharge Denominator: Palliative care population admissions Advance directives and surrogates—outpatient ACOVE EOL Surrogate decision-maker should be documented in outpatient charts Chart review Community Numerator: Outpatient chart includes: (1) Advance directive indicating surrogate decision maker, (2) documentation of discussion of who would be surrogate or search for surrogate, or (3) indication that there is no identified surrogate Denominator: Vulnerable elders

OCR for page 287
Performance Measurement: Accelerating Improvement Psychometric Testing (Validity/Reliability) Prior Use References   Prognosis was least frequently documented, followed by functional status and psychosocial symptoms   N Benchmarking   N Benchmarking   Benchmarking The ACOVE indicators have been tested in managed care settings; further research is addressing quality improvement Research Wenger et al., 2003

OCR for page 287
Performance Measurement: Accelerating Improvement ening, summarizing into a small number of key dimensions, demonstration of broad applicability (region, type of illness, approach to care services, ethnic background), and demonstration that scores improve when processes of care improve (Teno et al., 2001; Teno, 2004). Many other potential measures of pain management and care planning are listed in Table J-1 and have some evidence to support their use. Areas with Measures That Need to Be Developed Many domains relevant to palliative care lack measures with sufficient supporting evidence for confidence even about whether further development of current approaches would yield useful measures. These include the treatment and prevention of most symptoms other than pain in patients who are very sick and nearing death. Measurement tools are available to address other physical and emotional symptoms, but insufficient work has yet been done to translate these into performance measures for this population. Measures for some symptoms have been developed for other populations, such as nausea in cancer treatment or depression in the elderly, but these do not have sufficient supporting evidence and have not been evaluated in the palliative care population. Many other areas, such as spirituality, life closure, and caregiver burden and bereavement, have measurement tools available, but generally research has not tested whether these vary with better care, whether they have ceiling effects, whether routine measurement is feasible, or most of the other attributes of useful measures of care system quality. Caregiving and caregiver concerns are areas with particular needs for further development. Caregivers are vital to many elements of the end-of-life experience, including psychosocial distress, life closure, and site of death. The quality and quantity of caregiving can affect many other measurement domains, including symptom management and advance care planning. In addition, the impact of caregiving on the caregivers can have consequences for their physical and emotional health. We identified no performance measures specific to caregiving. Although the After-Death Bereaved Family Interview is an interview of caregivers, it is oriented towards the caregiver’s perception of the patient’s experience rather than towards caregiver issues. In our systematic review of the end-of-life literature (Lorenz et al., 2004), we found that, although many measurement instruments have been developed to examine caregivers’ experiences, interventions for caregivers have had little consistent effect on these outcomes. Outcome measures also differed widely across studies; although caregiver burden was frequently studied, other outcomes included stress, depression, anxiety, satisfaction, caregiver morbidity and mortality, unmet needs, and institutionalization.

OCR for page 287
Performance Measurement: Accelerating Improvement The domain of grief and bereavement also has many available measurement instruments (Lorenz et al., 2004), but little is available to guide performance measurement. Bereavement may have significant impact on significant others’ health, including depression and suicidality, particularly for parents of children and widowed elders. However, recent systematic reviews have found that, despite a large number of interventions in the literature, there is no clear evidence that interventions are effective in improving the experience of a sizable population, except for the pharmacological treatment of depression (Forte et al., 2004). Much of the reason for the lack of demonstrated efficacy is the low quality and variability in measurement and interventions in the literature (Forte et al., 2004). As documented in our recent evidence report (Lorenz et al., 2004), many domains do not even have well-developed measurement tools for use in palliative care; in particular, continuity of care, dignity, and autonomy require further work on every stage—concepts, factors influencing the domain, reliability and validity, generalizability, and evidence that care system improvement affects the measures. Finally, few measurement tools have records of use across diverse populations, including pediatrics, and further research in performance measures will need to address differences among fatal diagnoses, ethnic groups, and age groups. Key Gaps in the Evidence Base Our recent systematic review of the end-of-life literature (Lorenz et al., 2004) summarized the major gaps in the palliative care evidence base, and many of these deficiencies affect the development of measures. The lack of research on the implications of alternative definitions of the end-of-life population hinders convergence on a routine denominator in palliative care research or improvement activities. The lack of palliative care measures (such as symptom levels) in most research on specific diseases also limits our ability to define populations with unmet palliative care needs. Although research has developed many measurement tools for different domains in palliative care, these measures have rarely been tested in different settings or populations, which limits their applicability for use in performance measurement. Performance measures in symptom management await studies on symptom prevalence in noncancer populations; on associations between processes and outcomes; and on how interventions can improve symptoms across populations. Some sustained research has developed better pain management, but research for other symptoms is mostly nonexistent. In advance care planning, the key issue is to understand how various interventions actually have impacts on achieving patients’ goals, an outcome that has mostly evaded assessment. Finally, little research is available to inform performance measures in continuity, spirituality, or caregiver issues.

OCR for page 287
Performance Measurement: Accelerating Improvement Gaps in Understanding How Population Measures Need to Be Altered for the Palliative Care/End-of-Life Population Existing measures may apply to an elderly population or one defined by a particular diagnosis, but these need testing and adaptation to be sure that they will apply well to the palliative care population. For example, in ACOVE, a panel of geriatric experts found that only 130/203 of the indicators intended for vulnerable elders were still appropriate for patients with a prognosis of 6 months of less, and many of the general measures could be more useful if specifically adapted to the palliative care population (Solomon et al., 2003). Walter et al. (2004) found that not accounting for the seriousness of underlying illness, patient preferences, or clinician judgment can seriously compromise the performance of a quality measure. In populations with high proportions of patients who are ill or do not want aggressive care, high rates of screening may reflect badgering and imprudent decisions rather than quality, and low rates may be perfectly appropriate. Measure sets addressing populations with high proportions of palliative care patients need to include measures relevant to palliative care issues. For example, Mitchell et al. (2004) found that the 6-month mortality among newly admitted nursing home residents with advanced dementia was over 30 percent. However, measure sets in these settings often do not include appropriate elements of palliative care. For example, elements such as documentation of proxy decision makers, decisions to forgo resuscitation or hospitalization, or prognosis and symptoms might greatly improve the appropriateness of MDS for the high proportion of nursing home patients who need palliative care (American Academy of Hospice and Palliative Medicine, 2004). CHALLENGES TO APPLYING THESE MEASURES FOR THE PURPOSES OF QUALITY IMPROVEMENT, PAY FOR PERFORMANCE, AND PUBLIC REPORTING Challenges of Outcomes in Palliative Care Two major challenges to using outcome measures in palliative and end-of-life care are validity and adjustment for patient characteristics and preferences. Although many potential measures are objective (such as site of death) or have undergone careful development and extensive psychometric testing (such as the After-Death Bereaved Family Interview), the validity of these measures as indicators of the overall quality of palliative care has not been well established. Site of death is a good example of concerns about validity. Increasing the numbers of patients who die at home appears, at first glance, to be a

OCR for page 287
Performance Measurement: Accelerating Improvement laudable objective. Site-of-death information can generally be reliably obtained from death certificate or Medicare data. However, measuring whether dying at home is an important outcome may depend on how the question is asked. One national survey found that more than 60 percent of the elderly and more than 80 percent of seriously ill patients would prefer to die at home. However, in another national survey of seriously ill patients, in a list of nine attributes of what was important at the end of life, dying at home was ranked last (Steinhauser et al., 2000). Only 35 percent of patients and 30 percent of bereaved family members agreed that dying at home was important (Steinhauser et al., 2000). Whether a patient dies at home may depend on patient and caregiver preferences, and the patient’s perceptions of caregiver burden. For example, Fried et al. (1999) found that the primary concern of patients who preferred to be at home was the desire to be with their family members, while those who chose other settings were more concerned about their families’ ability to care for them and burden on their families. One would expect that the element that would be more important than the location at the time of death would be the patient’s preference as to where to live when near to death, but that question has not yet been asked in a research context. Dying at home may also be strongly dependent on whether supportive resources are available in that locality. Pritchard et al. (1998) found that in-hospital death increased with greater hospital bed availability and use and decreased with greater nursing home and hospice availability and use. Hospital bed availability was the most powerful predictor, far outstripping patient preference. However, Pritchard et al. also pointed out that the arrangements for care in a locality enmeshed a broad array of social patterns and expectations, including the behavior of the police and the neighbors, making it difficult to handle any one patient’s situation in a novel or customized way. Temkin-Greener and Mukamel (2002) found that the percentage of deaths that occurred at home among patients enrolled in the program of all-inclusive care for the elderly (PACE) varied from 25 to 76 percent, depending on the PACE site where patients received care. In a study in 8 counties, Tang and McCorkle (2003) found that patients who died in the county with the most resources available were most likely to die in their preferred location. Tang (2003) also found that many of these same factors, including family caregivers’ ability to provide care at home, might also predict the use of hospice care. These complex issues defy straightforward adjustment, since we have no tools that account for the effects of such factors as the availability of family caregivers or community resources. Broad use of a measure such as site of death, hospital length of stay, hospice referral, or length of hospice use could have adverse consequences. Working to decrease the number of persons who die within a hospital setting without increasing resource availability at home or in the nursing home

OCR for page 287
Performance Measurement: Accelerating Improvement may lead to discharges with uncontrolled symptoms, untrained and over-burdened caregivers, and increased readmissions, or misuse of hospice. In addition, neither palliative care interventions nor those specifically targeted towards improving the rate of home death have shown significant impact on increasing the rate of home death (Higginson et al., 2003). In one trial of hospital at home for the terminally ill, an intent-to-treat analysis showed no effects; but those who actually received the intervention had much higher rates of home death. Hospice in the United States also delivers very high rates of dying at home (50 percent at a private home), compared to the national rate of only 23 percent, but estimating the effect of selection bias would be difficult and has not been done. Denominator Issues/Population Definition A recent systematic review of the end-of-life literature (Lorenz et al., 2004) details the numerous challenges in defining the palliative care population. Most of the practical definition of “end of life” in the United States has relied upon the concept underlying the Medicare hospice benefit, which requires that eligible people would have a discernible phase of dying that reliably lasted less than 6 months. However, other concepts did arise in the literature review: e.g., “readiness” for death, “active dying,” and serious and eventually fatal illness. While many articles address the plausibility of prognosticating the timing of death, the summation of them is that no approach reliably distinguishes those who will die soon from those who will manage to survive for much longer. Most prognoses are ambiguous at a time that turns out to be within a week or two of dying. The inability to create categories by prognostic models affects all of the major causes of death except perhaps the most relentless of cancers. Yet, the other strategies for labeling a group as being at “the end of life” have almost no research base. Quality improvement work has tended to use either an arbitrary category that combines service utilization with diagnosis (e.g., all cancer patients seen in our clinic, or all heart failure patients admitted to the hospital) or the “surprise question,” which requires asking a clinician who knows the patient whether the patient is sick enough that it would be no surprise for the patient to die within 6 months or a year. The “surprise question” captures a much larger population that those thought appropriate for hospice referral, since it focuses upon a high risk of dying, rather than near certainty, and since it does not require also attending to the question of whether the patient will still be under life-prolonging treatment. The measures selected in Table J-1 use a number of different denominator definitions, all of which suffer from lack of validity testing. These include “vulnerable elderly,” or those at high risk of death or reduced functional status; “poor prognosis,” or prognosis of 6 months or less; patients

OCR for page 287
Performance Measurement: Accelerating Improvement considered to be “terminally ill” (MDS); patients currently receiving hospice care; and patients where a provider states that they would not be surprised if the patient died within the next year. Some measures also use denominators identifying all nontraumatic deaths retrospectively. This denominator is particularly problematic for use in performance measures, as many of these patients might not have been identified prospectively as being part of a palliative care population (Bach et al., 2004). Settings We identified two major issues related to the use of measures in different settings. First, due to the fractured nature of our health care system, measures have often been developed specifically for certain settings, often for nonpalliative care measure sets, and therefore cannot be compared across settings. For example, OASIS (home care), MDS (nursing homes), and NDS (hospice) all have very different pain performance measure methodology. Since important portions of palliative care occur in hospitals, providers’ offices, nursing homes, home care, and hospice, and patients will often make multiple transitions among settings, standardization of key measures would be critical to assessing performance and improving care across the continuum. If the hospital or its professionals are not performing well on the treatment of pain, for example, patients admitted to hospice will have a higher frequency and severity of pain on admission, which might affect the hospice’s performance measures adversely. Improving the overall care of these patients would require improvements by nonhospice providers. In working with a population that routinely changes settings often, and for whom improvements might well change the way that different settings are used, measures that are tied to particular settings are likely to be misleading. Use of Surrogates/Missing Data Issues related to the collection of data in palliative care have also been summarized in the recent systematic review (Lorenz et al., 2004) and need further research. Patients who are seriously ill or near the end of life are often unable to report on symptoms or other patient-centered elements of care. Measurement either resorts to proxy measures (such as after-death surveys of families), which often have only moderate congruence with patient reports, or carry high proportions of missing data and are therefore subject to bias. Further research will have to determine how and when to combine patient and proxy reporting and how to account for missing data through methods such as adjustment or repeat assessments.

OCR for page 287
Performance Measurement: Accelerating Improvement The Effect of Altering Survival Time Survival time has a troublesome interaction with most of the other elements that one might measure to estimate quality of care. With many outcome indicators of quality care, the patient is more at risk of adverse experience with longer survival, both from longer exposure and from more fragile health. Thus, for example, a care pattern that secured two months longer survival with emphysema would seem to have higher rates of dyspnea, more caregiver burnout, higher costs, and generally more adverse indicators. Since policymakers and researchers do not pay attention to this possibility and do not have metrics that would allow adjustment, this acts as an unmeasured confounder. This potential effect is one that is particularly difficult to discuss, since putting it into words risks allegations of having interest in foreshortening life (or, for that matter, of prolonging dying and inflicting suffering while increasing the bills). CONCLUSION While the costs of care at the end of life probably use about one-third of Americans’ lifetime health care, and while disapproval of the quality and reliability of that care is widespread, the indicators of quality, the measures to estimate quality, and the benchmarks and practical approaches to ensuring quality show longstanding inattention. Within a year, the NQF could probably field measures of physical pain and advance care planning that would be good enough for comparing health care delivery systems as to the quality of care. With more deliberate development over just a few years, life closure, caregiver experience, and some other symptoms (depression, dyspnea, chemotherapy-associated nausea and vomiting, for example) could be in the field. Some composite measures like knowing and delivering on the preferred place of death show promise precisely because high rates require a number of generally beneficial steps to have been taken. Having practical approaches to identifying the “end of life” population more usefully will require focused attention; finding clinical and administrative triggers that can concurrently identify the patients who face serious illness through to death is a task that will be essential for improvement activities. A recent State of the Science conference documented research priorities for end of life care (http://consensus.nih.gov/ta/024/EndofLifeStatementDRAFThtml.htm). REFERENCES American Academy of Hospice and Palliative Medicine and Americans for Better Care of the Dying. 2004. Testimony Concerning MDS 3.0. [Online]. Available: http://www3.cms.hhs.gov/quality/mds30/ [accessed January 4, 2005].

OCR for page 287
Performance Measurement: Accelerating Improvement Asch SM, Kerr EA, Hamilton EG, Reifel JL, McGlynn EA, eds. 2000. Quality Of Care for Oncologic Conditions and HIV: A Review of the Literature and Quality Indicators. MR-1281-AHRQ, 2000. [Online]. Available: http://www.rand.org/publications/MR/MR1281/ [accessed January 3, 2005]. Asch SM, McGlynn EA, Hogan MM, et al. 2004. Comparison of quality of care for patients in the Veterans Health Administration and patients in a national sample. Annals of Internal Medicine 14(12):938–945. Bach PB, Schrag D, Begg CB. 2004. Resurrecting treatment histories of dead patients: A study design that should be laid to rest. Journal of the American Medical Associaton 292(22): 2765–2770. Baier RR, Gifford DR, Patry G, Banks SM, Rochon T, DeSilva D, Teno JM. 2004. Ameliorating pain in nursing homes: A collaborative quality-improvement project. Journal of the American Geriatrics Society 42:1988–1995. Beal AC, Co JPT, Dougherty D, Jorsling T, Kam J, Perrin J, Palmer RH. 2004. Quality measures for children’s health care. Pediatrics 133:199–209. Centers for Medicare and Medicaid Services’ National Nursing Home Quality Initiative (NHQI). 2004. Nursing Home Compare. [Online]. Available: http://www.medicare.gov/nhcompare/home.asp [accessed January 3, 2005]. Clarke EB, Curtis JR, Luce JM, Levy M, Danis M, Nelson J, Solomon MZ. 2003. Robert Wood Johnson Foundation Critical Care End-of-Life Peer Workgroup Members. Quality indicators for end-of-life care in the intensive care unit. Critical Care Medicine 31(9): 2255–2262. Cleeland CS, Reyes-Gibby CC, Schall M, Nolan K, Paice J, Rosenberg JM, Tollett JH, Kerns RD. 2003. Rapid improvement in pain management: The Veterans Health Administration and the Institute for Healthcare Improvement Collaborative. Clinical Journal of Pain 19(5):298–305. Connor SR, Tecca M, LundPerson J, Teno J. 2004. Measuring hospice care: The National Hospice and Palliative Care Organization National Hospice Data Set. Journal of Pain and Symptom Management 28(4):316–328. Dokken DL, Heller KS, Levetown, M, et al. for The Initiative for Pediatric Palliative Care (IPPC). 2001. Quality Domains, Goals, and Indicators of Family-Centered Care of Children Living with Life-Threatening Conditions. Newton, MA: Education Development Center, Inc., 2001. [Online]. Available: http://www.ippcweb.org [accessed December 27, 2004]. Du Pen SL, Du Pen AR, Polissar N, Hansberry J, Kraybill BM, Stillman M, Panke J, Everly R, Syrjala K. 1999. Implementing guidelines for cancer pain management: Results of a randomized controlled clinical trial. Journal of Clinical Oncology 17:361–370. Earle CC, Park ER, Lai B, Weeks JC, Ayanian JZ, Block S. 2003. Identifying potential indicators of the quality of end-of-life cancer care from administrative data. Journal of Clinical Oncology 21(6):1133–1138. Fisher ES, Wennberg DE, Stukel TA, Gottlieb DJ, Lucas FL, Pinder EL. 2003a. The implications of regional variations in Medicare spending. Part 1: The content, quality, and accessibility of care. Annals of Internal Medicine 138(4):273–287. Fisher ES, Wennberg DE, Stukel TA, Gottlieb DJ, Lucas FL, Pinder EL. 2003b. The implications of regional variations in Medicare spending. Part 2: Health outcomes and satisfaction with care. Annals of Internal Medicine 138(4):288–298. Forte AL, Hill M, Pazder R, Feudtner C. 2004. Bereavement care interventions: A systematic review. BioMed Central Palliative Care 3(1):3. Fried TRE, O’Leary JH, Drickamer MA. 1999. Older persons’ preferences for site of terminal care. Annals of Internal Medicine 131:109–112.

OCR for page 287
Performance Measurement: Accelerating Improvement Hammes BJ, Rooney BL. 1998. Death and end-of-life planning in one Midwestern community. Archives of Internal Medicine 158(4):383–390. Higginson IJ, Finlay IG, Goodwin DM, Hood K, Edwards AG, Cook A, Douglas HR, Normand CE. 2003. Is there evidence that palliative care teams alter end-of-life experiences of patients and their caregivers? Journal of Pain and Symptom Management 25(2):150–168. Lorenz K, Lynn J, Morton SC, Dy S, Mularski R, Shugarman L, Sun V, Wilkinson A, Maglione M, Shekelle PG. 2004. End-of-Life Care and Outcomes. Evidence Report/Technology Assessment No. 110. (Prepared by the Southern California Evidence-based Practice Center, under Contract No. 290-02-0003). AHRQ Publication No. 05-E004-2. Rockville, MD: Agency for Healthcare Research and Quality, 2004. Lunney JR, Foley KM, Smith TJ, Gelband H, eds. 2003. Describing Death in America: What We Need to Know. Washington, DC: The National Academies Press. Lynn J, Schuster JL, Kabcenell A. 2000. Improving Care for the End of Life: A Sourcebook for Heatlh Care Managers and Clinicians. New York: Oxford University Press. Mitchell SL, Kiely DK, Hamel MB, Park PS, Morris JN, Fries BE. 2004. Estimating prognosis for nursing home residents with advanced dementia. Journal of the American Medical Association 291(22):2734–2740. Murray SA, Boyd K, Sheikh A, Thomas K, Higginson IJ. 2004. Developing primary palliative care. British Medical Journal 329(7474):1056–1057. [Online]. Available: http://www.macmillan.org.uk/healthprofessionals/disppage.asp?id=6875 [accessed January 8, 2005]. Nakajima GA, Wenger NS. 2003. Quality Indicators for the Care of Depression in Vulnerable Elders. [Online]. Available: http://www.rand.org/health/tools/vulnerable.elderly.html [accessed January 7, 2005]. National Consensus Project for Quality Palliative Care. 2004. Clinical Practice Guidelines for Quality Palliative Care. [Online]. Available: http://www.nationalconsensusproject.org [accessed December 15, 2004]. Pritchard RS, Fisher ES, Teno JM et al. 1998. Influence of patient preferences and local health system characteristics on the place of death. Journal of the American Geriatric Society 46:1242–1250. Rosenfeld K, Wenger NS. 2000. Measuring quality in end-of-life care. Clinical Geriatric Medicine; 16(2):387–400. Ryndes T, Connor S, Cody C, Merriman M, Bruno S, Fine P, Dennis J. 2000. Report on the Alpha and Beta Pilots of End Result Outcome Measures Constructed by the Outcomes Forum. A joint effort of the National Hospice and Palliative Care Organization and the National Hospice Work Group. [Online]. Available: http://www.nhpco.org [accessed January 3, 2005]. Solomon DH, Wenger NS, Saliba D, et al. 2003. Appropriateness of quality indicators for older patients with advanced dementia and poor prognosis. Journal of the American Geriatric Society 51:902–907. Steel N, Melzer D, Shekelle PG, Wenger NS, Forsyth D, McWilliams BC. 2004. Quality and Safety of Health Care 13(4):260–264. Steinhauser KE, Christakis NA, Clipp EC, McNeilly M, McIntyre L, Tulsky JA. 2000. Factors considered important at the end of life by patients, family, physicians, and other providers. Journal of the American Medical Association 284:2476–2482. Tang ST. 2003. Determinants of hospice home care use among terminally ill cancer patients. Nursing Research 52:217–225. Tang ST, McCorkle R. 2003. Determinants of congruence between the preferred and actual place of death for terminally ill cancer patients. Journal of Palliative Care 19:230–237.

OCR for page 287
Performance Measurement: Accelerating Improvement Temkin-Greener H, Mukamel DB. 2002. Predicting place of death in the program of all-inclusive care for the elderly (PACE): Participant versus program characteristics. Journal of the American Geriatric Society 50:125–135. Teno JM. 2000. TIME: Toolkit of Instruments to Measure End-of-Life Care. [Online]. Available: http://www.chcr.brown.edu/pcoc/toolkit.htm [accessed January 7, 2005]. Teno JM. 2001. Quality of care and quality indicators for end-of-life cancer care: Hope for the best, yet prepare for the worst. In: Foley KM, Gelband H, eds. Improving Palliative Care for Cancer. Washington, DC: National Academy Press. Pp. 96–131. Teno, JM. 2004. The Brown Atlas of Dying. Brown University Center for Gerontology and Health Care Research. [Online]. Available: http://www.chcr.brown.edu/dying [accessed December 27, 2004]. Teno JM, Clarridge B, Casey V, Edgman-Levitan S, Fowler J. 2001. Validation of toolkit after-death bereaved family member interview. Journal of Pain and Symptom Management 22(3):752–758. Teno JM, Weitzen S, Wetle T, Mor V. 2002. Persistent pain in nursing home residents. Journal of the American Medical Association 285(16):2081. Teno JM, Clarridge BR, Casey V, Welch LC, Wetle T, Shield R, Mor V. 2004. Family perspectives on end-of-life care at the last place of care. Journal of the American Medical Association 291(1):88–93. Walter LC, Davidowitz NP, Heineken PA, Covinsky KE. 2004. Pitfalls of converting practice guidelines into quality measures: Lessons learned from a VA performance measure. Journal of the American Medical Association 291(20):2466–2470. Wenger NS, Rosenfeld K. 2001. Quality indicators for end-of-life care in vulnerable elders. Annals of Internal Medicine 135(8 Pt. 2):677–685. Wenger NS, Young RT. 2003. Quality Indicators of Continuity and Coordination of Care for Vulnerable Elder Persons. [Online]. Available: http://www.rand.org/health/tools/vulnerable.elderly.html [accessed January 7, 2005]. Wenger NS, Solomon DH, Roth CP, et al. 2003. The quality of medical care provided to vulnerable community-dwelling older patients. Annals of Internal Medicine 139(9): 740–747. Wennberg JE, Cooper MM, eds. 1999. The Quality of Medical Care in the United States: A Report on the Medicare Program. The Dartmouth Atlas of Health Care 1999. Chicago, IL: American Hospital Association Press. Wennberg JE, Fisher ES, Stukel TA, Skinner JS, Sharp SM, Bronner KK. 2004. Use of hospitals, physician visits, and hospice care during last six months of life among cohorts loyal to highly respected hospitals in the United States. British Medical Journal 328:607. World Health Organization (WHO). 2002. Summary Measures of Population Health: Concepts, Ethics, Measurement and Application. Geneva, Switzerland: World Health Organization.

OCR for page 287
Performance Measurement: Accelerating Improvement LIST OF ABBREVIATIONS ACOVE Assessing Care of Vulnerable Elders CMS Center for Medicare and Medicaid Services EOL end of life ICU intensive care unit IHI Institute for Healthcare Improvement LOS length of stay MDS Minimum Data Set (CMS) NDS National Discharge Sample (NHPCO) NHPCO National Hospital and Palliative Care Organization UHC University Health Consortium VHA Veterans Health Administration