Transcript How well are we doing? Measuring the quality of end

Measuring the quality of endof-life care
David J Casarett MD MA
Division of Geriatrics
University of Pennsylvania
Did this patient receive good care?
84 year old man with heart failure and
pneumonia.
Admitted to the hospital, treated with
antibiotics and oxygen.
Discharged 3 days later.
What is good care for pneumonia?
How do you measure the quality of
care for pneumonia?
Patient received:
» Antibiotics
» Oxygen
» Intravenous fluids
Patient was discharged from the hospital
alive
Did this patient receive good care?
A 69 year old man with severe heart failure.
Admitted to the hospital with pneumonia, and
treated with antibiotics and oxygen.
Despite the best treatment, his condition
becomes worse.
He dies 3 days later.
What is good care for dying patients?
How do you measure the quality of
care for dying patients?
Measuring the quality of care for
dying patients:
Why measure the quality of end-of-life care?
A framework for measurement
» Structures of care
» Processes of care
» Outcomes
Measuring outcomes
» Prospective data
» Retrospective data
One example: Measuring pain management
Good end-of-life care includes
palliative care:
Palliative care is patient and family-centered
care that focuses upon effective management
of pain and other distressing symptoms, while
incorporating psychosocial and spiritual care
according to patient/family needs, values,
beliefs, and culture(s). (NCCN)
Why measure the quality of end-oflife care?
Benchmarking
Quality improvement
Identification of best practices
“Benchmarking”
Identifying goals and targets for high quality
Examples:
» We should assess the pain of all patients
» We should use strong opioids (morphine) for
severe pain in all patients
“Quality improvement”
Interventions to improve the quality of care
that patients receive.
Examples:
» Training for physicians and nurses about how to
assess pain
» Reminders to assess pain in all patients
» Pharmacy suggestions about which opioids to use
“Best practices”
Procedures and rules that lead to better care
Examples:
» One hospital uses a pain assessment form as part
of the admissions process, ensuring that all new
patients have their pain assessed
» One hospital created rules to guide physicians in
using strong opioids (morphine) when appropriate.
Measuring quality of end-of-life care
Measuring quality of end-of-life care makes it
possible to do benchmarking, quality
improvement, and to identify best practices.
But: How do we measure the quality of endof-life care?
3 ways to measure quality
Structures of care
Processes of care
Outcomes of care
Structures, Processes, and
Outcomes
STRUCTURES
PROCESSES
OUTCOMES
“Structures of care”
Aspects of the health care system that
contribute to better quality of care
Examples:
»
»
»
»
Staffing
Training
Protocols
Procedures
What structures would lead to better
quality end-of-life care?
» Staffing—Dedicated physicians and nurses for
palliative care
» Training of all staff in palliative care
» Palliative care teams
» Protocols for pain management
» Procedures for pain assessment
Advantages of measuring structures
Easily measured through surveys or existing
data
Easy to compare hospitals
Quick indicators of needs/problems
Disadvantages of measuring
structures
Not usually responsive to change
Not always related to quality
Generally not sufficient to demonstrate
effectiveness or success
Processes
STRUCTURES
PROCESSES
OUTCOMES
“Processes of care”
Care that is delivered to patients in order to
improve the quality of care
Examples:
» Assessing pain and other symptoms
» Prescribing medications
» Prescribing non-medical treatment
What processes would lead to better
quality end-of-life care?
»
»
»
»
Assessing pain
Having patient seen by a palliative care service
Prescribing opioids (morphine)
Using a bowel regimen alongside opioids
Advantages of measuring processes
Easily measured (in some settings)
Responsive to improvement efforts/funding
Disadvantages of measuring
processes
 May be more associated with better care in some
hospitals than others
 Positive results may offer false sense of improvement
 Processes not always associated with better care*
*Processes and outcomes
One physician prescribes morphine to all of
his patients with cancer
Another physician prescribes morphine to half
of his patients with cancer
Which physician is providing better care?
?
Outcomes
STRUCTURES
PROCESSES
OUTCOMES
“Outcomes of care”
The effect that structures and processes have
on the patient
Examples
»
»
»
»
Control of symptoms
Quality of life
Comfort
Being with family
What outcomes of care are important
to patients near the end of life?
Pain
Other symptoms
Adequate control over health care decisions
(patient’s perspective)
Family adjustment after death
Advantages of measuring outcomes
Ideal measure of needs
Measured from the patient’s perspective
Disadvantages of measuring
outcomes
Change may be slow
Positive/negative results may be due to other
factors
Accurate measurement is often:
» Difficult
» Expensive
Which outcomes are most important?
The patient’s spiritual and psychological wellbeing
Continuity of care across providers and care
settings
The patient’s physical comfort
Information and control over treatment
Family adjustment after death
Family psychological, spiritual, and social wellbeing
The patient’s physical comfort
Pain
Nausea
Pruritis
Constipation
Dyspnea
Thirst
Dry mouth
…
The patient’s spiritual and
psychological well-being
Psychological:
» Depression
» Anxiety
» Confusion
Spiritual/emotional
» Peacefulness
» Sense of community
» Reconciliation with friends/family
Information and control over
treatment
Culturally appropriate understanding of
treatment options
Culturally appropriate understanding of
prognosis and illness trajectory
Treatment consistent with preferences
Site of death consistent with patients’ and
families’ goals
Family psychological, spiritual, and
social well-being
Family’s acceptance of death
Reconciliation
Caregiver burden
Provisions for family members and children
Continuity of care across providers
and care settings
Continuity of information
Continuity of treatment
Continuity of health care providers
Family adjustment after death
Adjustment
Contribution of grief support (formal/informal)
Guilt/acceptance
“Prospective” vs. “Retrospective”
measurement?
Prospective: Going forward
Retrospective: Asking patients or families to
remember
Prospective data: Advantages and
disadvantages
Advantages:
» “Gold standard” of data collection (from
patients)
Disadvantages:
» Challenges of identifying patients near the
end of life
» Challenges of data collection
What proportion of patients near the
end of life can provide data?
65% (Fowler 1999)
65% (Lynn 1997)
50-63% (Hospice)(Casarett 2004, Casarett
2005, Rickerson 2005)
Declines over time (Stromgren 2005)
• 82% at baseline
• 70% at week 1
• 69%-52% at weeks 2 – 13
The problem of attrition in a hospice
population
Number of Responses and Deaths/Terminations
90
80
70
60
50
40
any response
patient response
known dead
dead or left study
30
20
10
0
1
2
3
Visit Number
4
5
Trajectories of functional decline
CHF/COP
D
Cancer
Dementia
Retrospective (patient-level) data:
Advantages and disadvantages
Advantages:
» Does not require interaction with patients who are
seriously ill
» “end-of-life” patients are already defined
» Best way to assess family’s experience, especially
after patient’s death
» Economical
Disadvantages:
» Biased/inaccurate responses
» Not all patients have surrogates
How accurate are families’
recollections of care?
 Ahmedzai (1988) poor agreement (patient-family pre-death) in
all symptoms except dyspnea. (Higher scores from families)
 Milne (2005) better agreement for physical symptoms than for
psychosocial well-being (pre-death)
--------------------------------------------------------------------------- Higginson (1996) better agreement (pre- vs. post- death) for
service provision and communication than for symptom control.
 Klinkenberg (2003) better agreement for physical symptoms
than for psychological symptoms (pre- vs. post-death)
 Hinton (1996) Patient-family ratings during illness and then after
death (families). Good agreement about incidence, less for
severity of physical/psychological symptoms.
Finding a surrogate
No surrogate
» Not available
» Moved/no forwarding address
» None
“Wrong” surrogate
Changing surrogate
Finding a surrogate
Consent: 55%
Knowledgeable: 64%
Reachable: 70%
Surrogates: 85%
Finding a knowledgeable surrogate
79% (Klinkenberg 2003)
37% (Curtis 2002)
90% (Casarett 2004)
69% (Seale 1995)
38%(?) (Teno 2003)
Potential response bias? Patients
with surrogates:
Differences:
»
»
»
»
Older
Fewer ADL dependencies
More likely to be white
More likely to have private insurance and/or
Medicare
BUT: Generally small effect
Outline:
Why measure the quality of end-of-life care?
A framework for measurement
» Structures of care
» Processes of care
» Outcomes
Measuring outcomes
» Prospective data
» Retrospective data
One example: Pain measurement
Outcome measurement: Pain
Pain is often undertreated
Pain is often not recognized
Pain reduces quality of life and is an
important source of discomfort
Measuring pain
For which patients should pain be measured?
How should we measure pain?
For which patients should pain be
measured?
All patients
How should we measure pain?
Is the patient cognitively impaired?
»
»
»
»
No
Mild
Moderate
Severe
Can patients with dementia report
pain?
Is the patient cognitively impaired?
 Mild impairment (Mini Mental State score of >22 out
of 30)
» Some problems with memory or complex tasks
» Often impairment is not noticeable
 Moderate impairment (Mini Mental State score of 1222 out of 30)
» Knows name/location
» Able to follow commands
» Impaired memory
 Moderate or severe impairment (Mini Mental State
score of <12 out of 30)
»
»
»
»
May not know name/location
Limited understanding
Able to follow simple commands only
Severely impaired memory
Cognitive function and pain
assessment
Mild impairment
» No different assessment techniques
Moderate impairment
» Additional teaching
» Use simple scales with pain descriptions
Moderate or severe impairment
» Use scale, but also rely on behaviors
» Include assessments of family members and
caregivers
Numeric rating scale: acceptable
for mild impairment
0 1
No
pain
2
3
4
5
6
7
8 9 10
Worst
possible
pain
Assessing pain in patients with moderate
cognitive impairment
Rely on patient
Surrogate report only if patient cannot reliably
communicate
Use terms synonymous with pain
Use standard pain scale:
» 0-10 Numeric Rating Scale
» Better: Verbal Descriptor/Pain Thermometer
Assessing pain in patients with moderate
cognitive impairment
Ask about present pain
Ensure understanding of tool use
» Do a “practice” assessment
» Compare pain now with best and worst pain
» Check for discrepancies between pain rating and
function to identify possible underreporting of
pain
Verbal Descriptor Scales
Verbal Descriptor Scale (VDS)
___
___
___
___
___
___
___
Most Intense Pain Imaginable
Very Severe Pain
Severe Pain
Moderate Pain
Mild Pain
Slight Pain
No Pain
(Herr et al., 1998)
Present Pain Inventory
(PPI)
0 = No pain
1 = Mild
2 = Discomforting
3 = Distressing
4 = Horrible
5 = Excruciating
(Melzack & Katz, 1992)
Pain Thermometer
Pain as bad as it could be
Extreme pain
Severe pain
Moderate pain
Mild pain
Slight pain
No pain
(Herr and Mobily, 1993)
Assessing pain in patients with severe cognitive
impairment
Ask the patient
Also ask:
Family members
Caregivers
Particular attention to factors that could increase
pain (movement, bathing, transfers, dressing
changes)
Use simple scale and/or behaviors to assess
severity
Facial Pain Scales
Faces Pain Scale
Bieri D et al. Pain. 1990;41:139-150.
Behaviors and non-verbal clues
Direct Observation or History from Caregiver for evidence of
pain-related behaviors (during movement, not just at rest)
 Facial expressions of pain (grimacing)
» Less Obvious: slight frown, rapid blinking, sad/frightened face, any
distorted expression
 Vocalizations (crying, moaning, groaning)
» Less obvious: grunting, chanting, calling out, noisy breathing,
asking for help)
 Body movements (guarding)
» Less obvious: rigid, tense posture, guarding, fidgeting, increased
pacin, rocking, restricted movement, gait/mobility changes such as
limping, resistance to moving
The importance of recognizing changes as clues
to pain
Changes in Interpersonal
Interactions






Combative/ aggressive
Resisting care
Decreased social interactions
Socially inappropriate
Disruptive
Withdrawn
Changes in Activity
Patterns/Routines
 Sudden cessation of common
routines
 Increased wandering
 Difficulty sleeping
 Increase in rest periods
 Refusing food/appetite change
Changes in Mental Status
Irritability or distress
Increased confusion
Agitation
Crying or
tears
Algorithm for assessing pain in severe cognitive
impairment
1.
Unusual behavior should trigger assessment of pain as a
potential cause
But: Some patients exhibit little or no pain-related behaviors
associated with severe pain
2.
3.
4.
5.
6.
Are Comfort Needs Being Met?
Evidence of pathology that may be causative (e.g. infection,
constipation, fracture)?
Assess patient report (if possible)
Assess behaviors
Attempt an analgesic trial
1.
2.
If in doubt, analgesic trial may be diagnostic
Acetaminophen 500mg TID, (titrate up to 3-4G/day)
Summary of pain measurement
 Overview
 Challenges and approaches
 Mild to moderate cognitive impairment
 Moderate to severe cognitive impairment
 Examples of tools to assist assessment
»
»
»
»
Intensity Scales
Interview Guide
Pain-related Behaviors
Approach to Nonverbal Cognitively-Impaired Older Adult
Outline:
Why measure the quality of end-of-life care?
A framework for measurement
» Structures of care
» Processes of care
» Outcomes
Measuring outcomes
» Prospective data
» Retrospective data
One example: Pain measurement