Transcript Outcome

Quality Measurement
Ira B. Wilson
April 4, 2014
Goals
• What is healthcare quality?
• How do we measure it?
• How are we doing on quality and what are we
doing to improve it?
PHP 310, 2014
2
What is Quality?
Background
• This is not simple … think about how you
would go about measuring the quality of
– Parenting
– Teaching
– Your Brown advisor
– Your priest, mullah, minister, rabbi
– Boyfriend, girlfriend, roommate, classmate
– Government
PHP 310, 2014
4
Value in Health Care
• Intuitively, value has to do with paying a fair or
reasonable amount for something
• Understanding value, and producing value,
requires that we understand and measure quality
• Example: buying a phone or a camera
• Think of it as a ratio: dollars per unit of quality
• Research area: decision analysis, cost
effectiveness analysis
PHP 310, 2014
5
What is Quality in Health Care?
• What are some features of health care that
are good or desirable?
• What are some features of health care that
are bad or undesirable?
PHP 310, 2014
6
Definitions of Quality
• “Quality is the extent to which health services for
individuals and populations increase the
likelihood of desired health outcomes and are
consistent with current professional knowledge”
(IOM, 1990).
• “Quality and safety are ultimately determined by
the degree to which health care improves
important patient outcomes” (AHRQ).
• “Characteristics associated with excellence,
including safety, effectiveness, patient control,
timeliness, efficiency, and equity” (IOM, 2000)
PHP 310, 2014
7
Definitions of Quality
• Abstract, multi-faceted concept that relates to
the improvement of patient health outcomes
• Perspectives vary
– Patient: experience of care, objective outcomes
(e.g., BP control, survival), functional outcomes,
safety
– Provider: technical skills, knowledge, training,
reputation, how their patients fare
– Payer: value proposition, efficiency
PHP 310, 2014
8
Donabedian’s Model (textbook p 505)
• Avedis Donabedian (1966): 3 elements
• Structure: physical and organizational
resources contributing to healthcare delivery
• Process: Activities that constitute care
delivery
• Outcome: results of care processes; recovery,
restoration of function, and survival
PHP 310, 2014
9
Reminder … back in Lecture #2
Wilson and Cleary. JAMA 1995;273:59-65.
PHP 310, Spring 2014
10
Donabedian’s Model
Structure
• Facilities
• Equipment
• Staffing
• Qualifications
• Licensing
• Accreditation
Process
• Technical care
• Screening
• Prevention
• Diagnosis
• Treatment
• Follow up care
Outcome
• Acute recovery
• Restoration of
function
• Survival
• Efficiency
• Interpersonal care
• Respect
• Communication
• Knowledge and
information
PHP 310, 2014
11
Technical vs. Interpersonal Care
• One could argue, at least for the care of
chronic diseases, that this is an artificial
distinction
• High quality interpersonal care, in most cases,
is required for good technical care to be
implemented
PHP 310, 2014
12
Treating Pneumonia in the ED
Structure
• Well trained staff
• Skilled, licensed,
certified MDs
• Proper x-ray
equipment
• Certified clinical
lab
• Available
subspecialty
consultation
• Surgical expertise
Process
•Technical care
•Triage done quickly and
effectively
•Data collection (vital signs,
exam, labs, etc.)
•Timely initiation of antibiotics
•Appropriate discharge
instructions
•Appropriate follow up
Outcome
• Timely recovery
from acute event
• Return to work
• Return to exercise
and normal
physical activity
• Survival
•Interpersonal care
•Shared decision making
•Respectful treatment
•Communication about follow
up
PHP 310, 2014
13
2007
PHP 310, 2014
14
PHP 310, 2014
15
Accountability for Quality
• CASE: Cervical cancer screening in women
with HIV care
• Chain of events that has to happen in order
for this relatively simple even to occur
• Single weak link breaks the chain
• Thus the emphasis on teamwork and systems
PHP 310, 2014
16
Accountability for Quality
• CASE: Cervical cancer screening in women
with HIV care
• Chain of events that has to happen in order
for this relatively simple even to occur
• Single weak link breaks the chain
• Thus the emphasis on teamwork and systems
• THOUGHT QUESTION: How could regulation
be brought to bear to address these low
screening rates?
PHP 310, 2014
17
Measuring Quality
Donabedian’s Model
Structure
Process
Outcome
• Technical care
• Interpersonal
care
PHP 310, 2014
19
Measuring Technical Care
• Where do you find process measures?
– Claims databases (if it is something that is
reimbursed)
– Medical records, e.g.
• Smoking cessation discussions
• Examining the feet of diabetic patients
– Medical records can be incomplete and if not
electronic can be difficult to review
PHP 310, 2014
20
Measuring Technical Care
• How do you decide what processes to assess?
– Strong evidence of process-outcome linkage
PHP 310, 2014
21
Evidence Based Medicine
• Is there such a thing as non-evidence based
medicine?
• Refers to a set of methods, developed in the
last 20 years, that are used to evaluate and
summarize the information in the published
literature
• Ranking the quality and generalizability of
evidence
PHP 310, 2014
22
Evidence Based Medicine
• Spectrum of scientific rigor
– Meta analyses of randomized trials
– Individual randomized clinical trials
– Observational studies
– Case reports
– Conventional wisdom
– “In my experience …”
PHP 310, 2014
23
Center Director: Tom Trikalinos
PHP 310, 2014
24
Measuring Technical Care
• Many processes contribute to any outcome:
think of hospitalization for heart attack or
treatment of lung cancer
• Quality of evidence about process-outcome
links varies
• For any process measure there are practical
issues related to the cost, convenience, and
validity of the measurement process
PHP 310, 2014
25
Advantages of Measuring
Technical Processes
• Simplicity:
– All women between 50 and 75 should have a
mammogram every 2 years
– All diabetics should be screened yearly for diabetic
retinopathy
– All patients who have had a heart attack should be
on daily aspirin (absent counterindications)
• Most accepted quality measures are measures
of technical processes
PHP 310, 2014
26
Measuring Interpersonal Care
• You have to ask the patient
• What do you ask them about?
– Service quality: office and telephone wait times
– Communication quality: information exchange
about medications, diagnostic testing, diagnosis,
prognosis
– More complex concepts: trust, respect, overall
satisfaction with care, willingness to recommend
PHP 310, 2014
27
Measuring Interpersonal Care
• How do you ask them?
– Telephone surveys
– Mail surveys
– Internet surveys
PHP 310, 2014
28
Issues with Measuring Outcomes
• Goal of health care is to improve health
outcomes
• It follows that the ultimate goal of quality
measurement and quality improvement is to
improve health outcomes
• Then why don’t we forget structure and
process and focus on outcomes?
PHP 310, 2014
29
Issues with Measuring Outcomes
• “Hard”, objective, outcomes like death rates
are often irrelevant (i.e., in any situation
where mortality is rare)
• Functional outcomes
– Objective measures: stress test, 6-minute walk
– Self-reported functioning (see 2nd lecture)
– Good measures exist, but they are less intuitive
• Risk adjustment is necessary
PHP 310, 2014
30
Risk Adjustment = Fair Comparisons
• Pay for performance for teachers and school
administrators
• Performance measure: student SAT scores
PHP 310, 2014
31
Question
Would it be fair to compare the performance of
teachers in Central Falls, RI, and Barrington, RI,
using student SAT scores ?
A. Yes
B. No
PHP 310, 2014
32
Risk Adjustment
“Risk factors”
Genetic endowment
Childhood nutrition
Family background
Quality of previous education
Etc.
SAT SCORE
Teacher Performance
PHP 310, 2014
33
Risk Adjustment
“Risk factors”
Genetic endowment
Childhood nutrition
Family background
Quality of previous education
Etc.
To fairly compare teacher
performance in Central
Falls and Barrington you
have to account for or
adjust for these sociodemographic differences
SAT SCORE
Teacher Performance
PHP 310, 2014
34
Risk Adjustment
Barrington
SAT SCORE
Central Falls
Statistically adjust for:
• % college educated parents
• Income
• Employment
• Crime rates
• % English speaking
• Rates of incarceration
• $ spend on school system
Teacher Performance
PHP 310, 2014
35
Risk Adjustment
Barrington
SAT SCORE
Central Falls
Statistically adjust for:
• % college educated parents
• Income
• Employment
• Crime rates
• % English speaking
• Rates of incarceration
• $ spend on school system
Teacher Performance
PHP 310, 2014
36
Risk Adjustment
Hospital A
30 Day Mortality
Hospital B
Statistically adjust for:
• Patient age
• Patient gender
• Previous cardiac disease
• Cardiac ejection fraction
• Insurance
• Income
• Social and family support
Heart Attack Care
PHP 310, 2014
37
Case Mix or Risk Adjustment
• Medical care is, if anything, more complex than
educational achievement
• Outcome measurement, to be fair, should take
account of or adjust for patient characteristics
related to their baseline risks for the outcome in
question
• Comparing apples to apples, “all other things
being equal”
• Risk adjustment can be technically complex and
expensive
• Critical in public reporting and pay-forperformance
PHP 310, 2014
38
Summary: Measuring Quality
• Measurement science is well developed, for
structure, process, and outcome
• Science of risk adjustment well developed for
some outcomes, but imperfect
• Drive for standardization
• Examples of organizations that review and
certify measurement tools
– NCQA (National Committee on Quality Assurance)
– NQF (National Quality Forum)
PHP 310, 2014
39
Take Home Lessons
• Including concepts from the last lecture … in
the last 30 years the science related to quality
has rapidly advanced
– We can define quality … but it isn’t simple
– We can measure quality … but it isn’t easy
• Quality improvement methods successfully
used in industry are increasingly being applied
to health care … but changing complex
systems is hard
PHP 310, 2014
40
NON-REQUIRED
READING
PHP 310, 2014
41
NON-REQUIRED
READING
PHP 310, 2014
42
Quality Performance
How are we doing?
PHP 310, 2014
44
Overall Quality of Care
PHP 310, 2014
45
PHP 310, 2014
46
PHP 310, 2014
47
PHP 310, 2014
48
PHP 310, 2014
49
PHP 310, 2014
50
PHP 310, 2014
51
PHP 310, 2014
52
PHP 310, 2014
53
PHP 310, 2014
54
PHP 310, 2014
55
PHP 310, 2014
56
PHP 310, 2014
57
PHP 310, 2014
58
PHP 310, 2014
59
Reading
60
Reading
• Accountability (nice review!)
– Physicians and physician organizations
– Hospitals
– Insurers
• Competition, consumerism, and market forces
– Consumerism slow to arrive to health care
– Ratings in news magazines and websites
– Increasing public interest in accountability
PHP 310, 2014
61
PHP 310, 2014
62
PHP 310, 2014
63
Leapfrog Top Hospitals 2013
PHP 310, 2014
64
Leapfrog Top Childrens’ Hospitals 2013
PHP 310, 2014
65
Selected Houston, Tx Hospitals
PHP 310, 2014
66
PHP 310, 2014
67
PHP 310, 2014
68
PHP 310, 2014
69
PHP 310, 2014
70
Quality Focused Organizations
• Institute for Health Care Improvement (IHI):
http://www.ihi.org
• Massachusetts Health Quality Partnership
(MHQP): http://mhqp.org
• Pacific Business Group on Health (PBGH):
http://pbgh.org/
• Many others
PHP 310, 2014
71
Take Home Messages
• Critical trend: using market forces to drive
quality improvement
– Public reporting
– Pay for performance (P4P)
• Who is driving this evolution?
– AHRQ: collecting and disseminating data
– BOTH private and public payers
– New private organizations that have developed to
meet the needs of payers
PHP 310, 2014
72