Clinical Investigation and Outcomes Research Health Outcomes Research

1
Clinical Investigation and Outcomes Research
Health Outcomes Research

• Marcia A. Testa, MPH, PhD
• Department of Biostatistics
• Harvard School of Public Health

2
The Discipline of Health Outcomes Research

• Scientific inquiry evaluating the results of
  medical interventions and health care services.

• Outcomes data are direct measures of whether
  medical treatments are beneficial

3
How Are Outcomes Results Used?

• Outcomes research seeks to understand the end
  results of particular health care practices and
  interventions.
• Used to provide information on the quality of
  care so that it can be improved by determining
• which health care services influence the
  probability of optimal patient outcomes
• which produce optimal improvement in the
  patient's physiologic status, physical function,
  emotional and intellectual performance and
  comfort (comparative effectiveness research)

1. Outcomes Research Resource Guide, 1996/97
Edition, American Medical Association
4
Measuring Outcomes

• Outcomes are a function of
• baseline health status
• patient clinical characteristics
• patient demographics
• psychosocial characteristics
• treatment
• health care setting
• The GOAL of health outcomes research analysis is
  to isolate the relationship between outcomes and
  treatment.

5
Outcome Measures

• Mortality Survival time, Death event
• Morbidity Time to an Occurrence of a Clinical
  Events (e.g., Stroke, MI, Cancer)
• Health Status Physical, Mental and Emotional
  Health Functioning
• Quality of Life Health Status as Perceived by
  the Individual
• Patient Satisfaction Distance between Quality
  of Life and Individual Expectations
• Health Economic Outcomes Cost utility, cost
  effectiveness

6
Hematologic Malignancies Anemia Comparative
Effectiveness Trial
7
Hb Changes
Figure 2. Mean hemoglobin (Hb) change
(intent-to-treat n 269). aP lt .0001 early
versus late group. Postrandomization Months 1, 2,
3, and 4 values correspond with mean Hb between
Weeks 0 (baseline) to Weeks 4, 5-9, 10-14, and
15-20, respectively.
a P lt 0.0001
8
Treatment, Fatigue, Symptoms
bP lt .01 late vs. early treatment. cP lt .05 late
vs. early treatment.
9
Treatment and Health Status
bP lt .01 late vs. early treatment. cP lt .05 late
vs. early treatment.
10
Treatment and QOL
bP lt .01 late vs. early treatment. cP lt .05 late
vs. early treatment.
11
Measuring Outcomes
Outcome
Full Health
Death

12
Measuring Outcomes, Measuring Performance
Improving Process (how we perform)
Measurement
Improving Outcomes (the results of our
performance)
13
The Consequences of Health Care and Medical
Intervention
14
The Consequences of Health Care and Medical
Intervention Types of Outcomes
Labs, Clinical Events, Physician Assessments
15
The Consequences of Health Care and Medical
Intervention Types of Outcomes
Symptom reports, health status, quality of life,
patient satisfaction
16
Multi-faceted QOL Domains
17
Outcomes Research - 1996QOL is Recognized as
Important
18
Outcome MeasuresFunctional Health Coverage

• Generic health instruments are address larger
  health constructs and hence their causal links to
  specific treatment events may be more difficult
  to detect
• Condition-specific instruments will vary with the
  condition being treated, and hence are typically
  more sensitive to treatment effects

19
Outcome MeasuresFormat Influencing Coverage

• Fixed-Length or Static number of questions is
  fixed greater coverage requires greater number
  of questions
• Dynamic instruments - use computer adaptive
  testing to restrict items based upon a Bayesian
  approach which selects the next question based
  upon the answer to the previous question
• Combines the practicality of short form
  instruments with the sensitivity and target
  coverage of condition-specific instruments

20
Questionnaires and SurveysGeneric Instruments

• Some outcomes survey questions, commonly referred
  to as instruments, focus on describing how
  individuals rate their health overall or
  generic instruments
• General health surveys such as the SF-36 are now
  used in research studies, population surveys, and
  some health plans to assess patients' overall
  level of functioning.
• Translation of SF-36 into Arabic and the
  translation methods references are given in the
  Additional Resources Section of the Website.

21
Questionnaires and Surveys Condition-Specific
Instruments

• Developing outcome instruments for specific
  diseases has been an especially prolific research
  area
• Such instruments are more likely than general
  health survey measures to be able to detect
  changes in the disease due to treatment

22
Steps in Designing an Outcome Research Study

• Define a researchable question
• Develop a conceptual model
• Identify the critical dependent and independent
  variables
• Identify appropriate measures for each
• Develop an analysis plan
• Indicate what is believed to cause the outcome
• Identify critical pathways and what other factors
  are likely to affect the outcome
• Identify which variables (in the outcomes
  function equation) are relevant to your
  hypothesis

23
The Conceptual Health Model

• Environmental Behavioral, Social
• Income
• Social Support
• Education
• Health Access
• Lifestyle

• Medical Interventions
• Specific medications
• Surgery
• Diet and Exercise
• Case Management

• Outcome Measures
• Lab values
• Symptoms
• Functioning
• Quality of life
• Employment/Work

• Patient Factors
• Age
• Gender
• Occupation

24
The Outcomes Model
Risk Adjustment
Patient Characteristics
Change
Structure (Setting)
Process (Treatment)
Measurement
25
Outcomes Research - 1998Health Economic Benefits
26
Diabetes Outcomes Model

• Treatment
• Specific medications
• Diet
• Exercise
• Case Management

• Clinical Factors
• Severity
• Duration
• Etiology
• Comorbidity

• Outcomes
• HbA1c
• Symptoms
• Function
• Complications
• Quality of life
• Employment/Work

• Patient Factors
• Age
• Race
• Gender
• Occupation

27
Understanding Causal Pathways
HEALTH ECONOMICS Productivity Health Care
Utilization
28
Study Design
62 sites in the United States Glipizide GITS
diet vs Placebo diet
16-week multicenter, randomized, double-blind,
placebo-controlled, parallel titration study
1-week washout
o
o
o
o
3-week single-blind placebo
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
4-week dose titration
o
o
o
o
o
o
o
o
o
o
8-week maintenance
o
o
o
o
o
o
o
o
Testa MA, Simonson DC. JAMA 1998 2801490-1496.
29
Study Design
Maintenance
Screening /
Single-blind
Dose titration
Maintenance
washout
placebo
5-20 mg






?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
-1 0 1 2 3 4 5 6
7 8 9 10 11 12 13 14 15

Week
Randomization to placebo diet or
Glipizide GITS diet (12)
Clinical and Laboratory Assessment
?
Health Economic Assessment

Testa MA, Simonson DC. JAMA 1998 2801490-1496.
30
(No Transcript)
31
Patient Population
Baseline Clinical Characteristics
Placebo
Glipizide GITS
Number of Patients
201
393
Gender (M / F) ()
60 / 40
55 / 45
Race (W / B / O) ()
70 / 18 / 12
70 / 18 / 12
Emp / Ret / Unemp()
50 / 38 / 12
48 / 40 / 12
Age (yrs)
58 12
59 11
Duration DM (yrs)
5 5
6 6
FPG (mg / dL)
231 66
218 62
HbA1c ()
8.7 1.4
8.5 1.5
Testa MA, Simonson DC. JAMA 1998 2801490-1496.
32
Clinical Results
End of Week 15
Placebo
Glipizide GITS
P-Value
FPG (mg / dL)
224 66
161 41
lt 0.001
HbA1c ()
9.3 1.9
7.5 1.2
lt 0.001
Hypoglycemic
4.8
6
NS
Symptoms ()
Glucose lt 55
0
0
NS
mg / dL ()
Testa MA, Simonson DC. JAMA 1998 2801490-1496.
33
HbA1c and Symptom Distress

• EFFICACY
• HbA1c

• HEALTH STATUS
• Self-Reported Symptom distress

34
Pharmacological SideEffects and Symptom Distress

• EFFICACY
• HbA1c

• HEALTH STATUS
• Self-reported Symptom distress

• SIDE EFFECTS
• Adverse reactions

35
Mean HbA1c 9.3 1.9 7.5 1.2
Symptom Worsening With Glucose Lowering
P N.S.
Symptom Improvement With Glucose Lowering
P lt .05
P lt .01
P lt .001
0
0.2
0.4
0.6
0.8
1
-0.2
SD Units
Testa MA, Simonson DC. JAMA 1998 2801490-1496.
36
Impact on Quality of Life

• EFFICACY
• HbA1c

• HEALTH STATUS
• Self- Reported Symptom distress

• QUALITY OF LIFE
• Functional status
• Physical, Mental, Cognitive, Social

• SIDE EFFECTS
• Adverse reactions

37
Change In QOL Scales With Therapy in Type 2
Diabetes
0.4

Overall Rating Mental Health Cognitive
Function Perceived Health Symptom Distress
0.3



0.2
0.1
QOL Score (SD units)
0
P lt 0.05 P lt 0.01 P lt 0.001
-0.1
-0.2
-0.3
Placebo
Glipizide GITS
Testa MA, Simonson DC. JAMA 1998 2801490-1496.
38
Change in HbA1c and QOL
In Patients with NIDDM
0.2
Favorable QOL

Response


0
No Change in QOL

-0.2
No Change in HbA1c
QOL Global Outcome (Z-score)

-0.4
Unfavorable QOL
-0.6
Response
log-linear regression, r .95, p lt .01
-0.8
gt1.5
1.5 to .5
-.5 to -1.5
.5 to -.5
lt-1.5
Improved
Worsened
HbA
c () Change from Baseline
1
HbA1c
HbA1c
Testa MA, Simonson DC. JAMA 1998 2801490-1496.
39
Understanding Causal Pathways
HEALTH ECONOMICS Productivity Health Care
Utilization
40
Production Losses
Restricted Activity Days (US / 1000 person days)
Bed Days (US / 1000 person days)
Absenteeism (US / worker / month)
Loss
1000's
100
4
1000
50
2
500
P 0.01
Change (Week 15 Baseline)
0
0
0
P lt 0.001
P 0.05
-2
-50
-500
Placebo
Placebo
Placebo
Glipizide GITS
Glipizide GITS
Glipizide GITS
Savings
Testa MA, Simonson DC. JAMA 19982801490-1496.
41
Health Care Utilization
Percent of Patients Reporting 1 or More
Non-study-related Ambulatory Visits (clinic, ER,
physician office) per Month
40
40
P NS
P 0.002
Percent of Patients
30
30
20
20
Baseline
Week 15
Baseline
Week 15
Glipizide GITS
Placebo
Estimated savings 11 per patient per month
(assuming cost of 66 per ambulatory visit)
Testa MA, Simonson DC. JAMA 19982801490-1496.
42
Summary

• Use Outcomes Research to
• Improve the quality of health care by changing
  treatment and services and by promoting
  preventive strategies
• Outcomes are probability statements
• Multifaceted
• Requires integrating and consolidating many
  different components of health functioning
• Outcomes may take time to develop, use
  intermediate outcomes if necessary

43
Summary

• Use Outcomes Research to
• Improve the quality of health care by changing
  treatment and services and by promoting
  preventive strategies
• Outcomes are probability statements
• Multifaceted
• Requires integrating and consolidating many
  different components of health functioning
• Outcomes may take time to develop, use an
  intermediate outcomes

44
Additional Resources and References

• Online Questionnaires SF12, LASA, EQ-5D.
• SF-36 translated into Arabic by Saud Abdulaziz
  bin Al Abdulmoshin, 1997
• Coons SJ, Reliability of an Arabic Version of the
  RAND-36 Health Survey and Its Equivalence to the
  US English Version
• Testa and Simonson, NEJM, 1996
• Testa and Simonson, JAMA, 1998
• Straus, Testa, Sarokhan et al., Cancer 2006