Estimating Treatment Effects with Observational Data using Instrumental

advertisement
Estimating Treatment Effects with
Observational Data using Instrumental
Variable Estimation: The Extent of
Inference
John M. Brooks, Ph.D.
• Estimate casual relationships between
"treatment" and “outcome” in
healthcare...
→
→
→
Health Effectiveness Research Center (HERCe)
Colleges of Pharmacy and Public Health
University of Iowa
June 26, 2005
Research Goal:
Health
Effectiveness
Research
Center
→
treatment on outcome;
behavior on outcome;
system change on behavior (e.g.
guideline implementation);
system change on outcome.
1
• Written as a linear relationship:
2
• Key research design issues for isolating and
using “T” variation:
Y = a0 + a1• T
our goal is to obtain estimate(s) of “a1”.
• To estimate “a1” T must move or vary.
• To make inferences about “a1” the source
of the variation in T must be scrutinized
relative to your research goal.
1. the manner in which the researcher collects
data; and
2. the approach to deal with “confounding
factors”
confounding factors: factors that vary both
with T and Y.
3
Sources of Treatment Variation in Health Care
Research Environments and Estimation Methods
Statistical “Matching”
Techniques (Propensity Scores)
Secondary
Databases
Quasi-Experimental Designs
1. Randomized Controlled Trials: study of patients with
a given medical condition in which treatment is
randomly assigned.
ANOVA
Logistic Regression
Instrumental Variables
4
Multiple Regression
– “Ex Post Design”
– “Risk Adjustment”
• Why randomly assign treatment to patients?
Statistical Control of
Confounding Factors
Design Control of
Confounding Factors
Weighted Regression
Techniques of Survey
Databases:
• NMES
• MEPS
Entirely Controlled
2
Tests
Experiment - 3
– Randomized
Controlled Trials
Researcher-Collected
Databases
To help ensure that estimated treatment effects
result from the treatment variation and not
unmeasured confounders.
The Gold Standard
5
6
1
• Why not more Randomized Controlled Trials?
→ ethical problems once treatment is
approved
2. Observational Healthcare Databases Containing
Healthcare Treatment Choices:
• Secondary:
→ Claims: medical service treatment claims from
individuals with health insurance
→ expensive and time-consuming
→ little motivation
→ patient sampling problems when comparing
existing treatments (so who wants to be
randomized?)
7
• Strengths:
→ Provider-Specific: databases describing the
utilization of a set of providers.
• Primary:
→ Health Care Surveys: surveys of patients or
providers detailing health
8
care utilization.
• Weaknesses:
→ plenty of variation in treatment choice;
→ often data usually not collected for researcher’s
purpose (secondary);
→ ability to study effects of treatment across a
variety of clinical scenarios;
→ patient enrollment variation;
→ can assess treatments in practice – estimate
“effectiveness”;
→ confounding information may be unobserved.
-
→ often unobtrusively collected;
→ the power of large numbers and time.
care not covered is not observed
care not claimed is not observed
claim form limitations
nuances of illness, treatment, and patient that can’t
be recorded on claims forms
9
Is the Main Weakness with Observational Data
Unmeasured Confounders or Treatment Selection Bias?
10
• Assume true outcome relationship is:
Y = ao +
1. Unmeasured Confounders
a1•T + a2•L + e
where:
Y = measure of outcome (e.g. 1 if survive to a
certain time period, 0 otherwise);
• Unmeasured Confounders argument:
→ homogenous treatment effect (a1 same for all
patients); and
→ unmeasured factors related to both treatment
and outcome is the source of bias.
11
T = 1 if receive treatment, 0 otherwise; and
L = additional factor (e.g. severity, other treatments).
Goal is to estimate a1 – the effect of treatment on outcome.
12
2
• For Estimation Suppose:
• Define the ordinary least squares (ANOVA) estimate of
a1 as â1 .
→ L is not measured and the estimation model is:
Y = ao +
u
a1•T + u
=
→ It can be shown that under these assumptions â1 is
a biased estimate of a1 through its expected value:
where:
(a2•L + e)
E [ aˆ1 ] = a1 + Cov(T,L)•a2
→ L is related to Y (a2 ≠ 0); and
→ Also note that E [ aˆ1 ] will equal a1 if either:
→ T and L are related (Cov(T,L) ≠ 0).
-- Cov(T,L) = 0; or
Cov(T,L) – covariance of T & L. Cov(T,L) ≠ 0 essentially
means that T & L move together.
-- a2 = 0.
13
• Suppose theory about the unmeasured variable “L”
suggests:
14
• Problem with the Unmeasured Confounders argument
to describe bias in observational data:
→ “a2 < 0” (patients with higher severity are less likely
to survive).
→ No theoretical foundation linking treatments to
unmeasured factors....
→ Cov(T,L) > 0 (treated patients are generally more
severe).
• Plug in “signs” into our expected value formula to find:
E [ aˆ1 ] = a1 + (+)(−)
E [ aˆ1 ]
<
→ In the example above, if treatment effect (a1) is the
same for all patients, why would Cov(T,L) > 0?
Perhaps patients getting treated:
( −)
→
Why is Cov(T,L) ≠ 0?
-- live in areas with high/low poverty;
-- live in areas with more pollution; or
-- also tend to get other unmeasured treatments.
a1.
15
2. Treatment Selection Bias (the gestalt underlying most
negative reviewer’s comments)
16
• Assume true outcome relationship is:
Y = bo + (b1•L) •T + b2•L + e
• Treatment Selection Bias argument:
→ Heterogeneous treatment effect -- Cov(T,L) is a
reflection of decision-maker’s beliefs about the
treatment effectiveness across patients related
to unmeasured factors “L”.
→ “Bias” comes from unmeasured factors (L) being
related to the treatment choice and outcome.
→ Researcher must address both bias and ability
to generalize (to whom do the results apply?).
17
where:
Y
= measure of outcome (e.g. 1 if survive to a
certain time period, 0 otherwise);
T
= 1 if receive treatment, 0 otherwise;
L
= unmeasured factor (e.g. severity, other
treatment);
b2
= the direct effect of L on Y; and
(b1•L) = effect of T on Y that depends on L.
18
3
→ L is now related to T through theory linking "treatment
choice" to the decision-maker’s expectations of
treatment benefits across patients with different “L”.
T = co + c1•L + c2•W +
v
where:
• Ultimate goal should be to estimate (b1•L) – the
effect of treatment T on outcome Y across levels
of L.
• For estimation suppose:
T = 1 if receive treatment, 0 otherwise;
L = unmeasured factor (e.g. severity, other
treatment) affecting treatment choice through
expected treatment effectiveness; and
→ L is not measured and it is wrongly assumed
by the researcher that the effect of T is
homogenous, and the estimation model is:
Y = ao +
W = other factors affecting treatment choice.
If decision makers use L in treatment
decisions, c1 ≠ 0 and Cov(T,L) ≠ 0.
19
→ It can be shown that the expected value of â1 is:
E [â ] ≈ b ⋅ E [ L | T =1 ] + c ⋅ b
1
1
2
Yields an average estimate of the treatment effect for
“the treated” in the sample. Result can be generalized
21
only to those with L similar to those treated.
→ If treatment benefit is greater for more severe cases
(e.g. antibiotics for otitis media) then:
b > 0 ⇒ c > 0 ⇒ c ⋅ b = (+ )⋅ (− ) < 0
1
→ Assume that L is unmeasured illness severity
and that higher L means more severe illness.
→ Higher L lowers survival which implies b2 < 0.
1
1
1
• How does c1 • b2 affect this estimate?
b < 0 ⇒ c < 0 ⇒ c ⋅ b = (− )⋅ (− ) > 0
E [â ] ≈ b ⋅ E [ L | T =1 ]
1
20
→ If treatment benefit is less for more severe cases
(e.g. surgery for heart attacks) then:
→ If b2 = 0 (L has no direct effect on Y) or c1 = 0 (no
selection based on L), then E [â1 ] becomes:
1
where:
u = f(L,T, e, b1,b2)
• Define the ordinary least squares (ANOVA) estimate of
a1 as â1 .
1
a1•T + u
2
benefit falls
with higher
severity
1
1
2
less treatment
in more
severe cases
Estimate of the effect of the treatment on the treated
22
will be biased high.
• So what do we have here?
→ Observational data contains treatment variation.
→ If treatment benefits are heterogeneous the best you
can get is an estimate of the treatment effect on the
treated (Does this address the benefits from
expanding treatments?).
benefit increases more treatment
with higher
in more
severity
severe cases
Estimate of the effect of the treatment on the treated
will be biased low.
→ Treatment selection may be based on unmeasured
factors related to both treatment effectiveness and
outcomes.
→ If unmeasured factors affecting selection also
effect outcomes directly, estimate will be biased.
23
Do we have any alternatives?
24
4
Instrumental Variables (IV) Estimation and “Subset B”
• Where do Marginal Patients come from?
• IV estimation offers consistent estimates for a subset of
patients (McClellan, Newhouse 1993):
Marginal Patients: patients whose treatment choices vary
with measured factors called instruments
that do not directly affect outcomes.
Distribution of Patients by Prior Assessment of
the Certainty of Treatment Benefit
A
0%
• McClellan and Newhouse argued that estimates of treatment
effects for Marginal Patients are useful.
→ Estimates may be more suitable than RCT estimates to
address the question of whether existing treatment rates
should change.
B
C
50%
100%
More certainty
about treatment
benefits
Less certainty
about treatment
benefits
A = subset of patients all providers agree to treat.
C = subset of patients all providers agree not to treat.
B = subset of patients whose treatment choice is
situation/provider dependent.
25
• Patients in Subset B are interesting because:
26
• Size and location of Subset B varies with clinical scenario.
→ the “best” treatment choice (treat or don’t treat) is
least certain;
→ treatment or no-treatment for a patient in this subset
is not considered bad medicine – the “art” of
medicine;
→ the possibility of gaining new RCT evidence for
patients in this subset is remote (ethics, motivation);
→ McClellan et al. 1994 argue that (1) policy
interventions and (2) non-clinical factors (e.g.
provider access, market pressures) affect mainly the
treatment choices of patients in this subset.
Ý treatment with little consensus (e.g. aggressive treatment
for early-stage prostate cancer):
A
C
B
0%
50%
100%
More
Certainty
Less
Certainty
Ý off-label use for new treatment (e.g. new anti-cancer
drugs used in non-tested cancer populations):
B
0%
C
50%
100%
More
Certainty
Less
Certainty
27
• Changes in the underlying population definition will affect
the location of Subset B.
Ý aggressive treatment for early-stage prostate cancer for
50-60 year-olds with no comorbidities:
A
0%
B
50%
C
100%
More
Certainty
A
More
Certainty
50%
1. Finding measured variables or “instruments” (Z) that:
a. are related to the possibility of a patient receiving
treatment (cov(T,Z) ≠ 0); and
The theoretical basis for “Z” variables should come from
a model of treatment choice – the “W” variables in:
T = co + c1•L + c2•W +
C
B
• IV estimation involves:
b. are assumed (through theory) unrelated directly to Y
or to unmeasured confounding variables (cov(Z,L) = 0).
Less
Certainty
Ý aggressive treatment for early-stage prostate cancer for
70-80 year-olds with one comorbidity:
0%
28
100%
v
where:
W = other factors affecting treatment choice.
Less
Certainty
29
30
5
• IV estimation involves con’t:
• For example, if an instrument divides patients into two
groups, a simple IV estimate can be found by calculating:
1. the overall treatment rate in each group (ti = treatment
rate in group “i”); and
2. Grouping patients using values of the
“instrument”.
2.
3. Estimate treatment effects for marginal
patients by exploiting treatment rate
differences across patient groups.
the overall outcome rate in each group (yi = outcome
rate in group “i”); and estimate:
aˆ1IV =
difference in outcome rate
y − y2
= 1
difference in treatment rate
t1 − t 2
where:
Local Average Treatment Effect -(Imbens & Angrist 1994)
aˆ1IV
31
• Hypothetical Treatment Choices Across Patients
Grouped by Access to Providers Required for Treatment
treated
B M
Suppose we also measured “cure” rates in both groups:
C
0%
More Certainty
• We have treatment rates for each group:
Closer Group Treatment Rate: .60
Further Group Treatment Rate: .50
Patient Group Closer to Providers Required for Treatment:
A
= average treatment effect for the “marginal patients”
specific to the instrument used in the analysis –
only those patients whose treatment choices were
affected by the instrument who must have come
32
from Subset B.
100%
Less Certainty
Closer Group Cure Rate: .40
Further Group Cure Rate: .38
Patient Group Further From Providers Required for Treatment:
• Four numbers lead to the following IV estimate:
treated
A
0%
More Certainty
M
B M
50% 60%
C
100%
Less Certainty
= patients within Subset B whose treatment choices
are affected by the instrument – the Marginal
Patients for that instrument.
33
â =
1IV
.40 −.38
.02
=
= .2
.6 − .5
.1
34
• Strict Interpretation:
→ If the treatment rate in the Further Group was increased .01
percentage point (e.g. .50 to .51) by increasing treatment
for the M patients in the Further Group, the Cure rate in the
Further Group would increase .002 (.01 • .2) – from .38 to
.382.
• Stretched “Policy-Relevant” Interpretation (McClellan et al.
1994)
→ A behavioral intervention that increases the overall
treatment rate by .01 percentage point (e.g. .55 to .56)
would lead to an increase in the cure rate of .002 (.01 • .2).
35
• Stretched interpretation assumes that the treatment effect
for patients in Subset B is fairly homogenous and an IV
estimate from a single instrument can be generalized to all
patients in Subset B.
• Stretched interpretation may not be accurate if treatment
effects are heterogeneous within Subset B and different
instruments affect treatment choices from different patients
within Subset B.
→ Results from a single instrument may still be more
appropriate than assuming RCT results apply to Subset B.
→ Ability to generalize results may increase if more than one
instrument is used in an IV analysis.
36
6
• IV qualifiers to remember:
Hypothetical Example to Demonstrate “4-Number” Result
Suppose:
→ second property of IV variables (cov(Z,L) = 0) is
forever an assumption (unless more data are
obtained);
• 2100 children with Otitis Media (OM) in a population.
• Two treatment possibilities:
→ unmeasured but correlated treatments may still bias
estimated treatment benefits; and
→ ability to generalize is limited.
1.
2.
antibiotics;
watchful waiting.
• The patients in our sample are in one of three severity
types “low”, “medium”, and “high”
Researchers should fully qualify their IV estimates –
don't oversell.
• Severity type is observed by the provider/patient but is
not observed by the researcher.
37
• The 2100 patients are distributed across severity type in the
following manner:
number of patients
High
800
severity type
Medium
800
Low
500
• The actual underlying cure rates for each severity type by
treatment are:
treatment
antibiotics
watchful waiting
High
.95
.80
severity type
Medium
.97
.90
38
→ Higher severity means a lower the cure rate in general
(b2 < 0).
→ Treatment effects are heterogeneous and antibiotics have
a higher curative effect in more severe patients and offer
no advantage to the less severe (b1 > 0).
→ All providers have inclination that antibiotics work well in
the "high" severity patients; have little effect on the "low"
severity patients; but the effect in the "medium" type is
unknown.
Low
.98
.98
→ Leads to treatment selection bias...the more severe kids
are treated (c1 > 0) and more severe kids are less likely
cured (b2 < 0).
39
40
1. Randomize Patients Across Population – ANOVA.
Potential Methods to Get Treatment Variation for Analysis:
1. Randomize Patients Into Treatments -- ANOVA
Patient Treatment Assignments After Randomization
by Severity Type
2. Providers Assign Treatments -- ANOVA
patient groups
antibiotics
watchful waiting
High
400
400
severity type
Medium
400
400
Low
250
250
3. Instrumental Variable Grouping
41
42
7
Expected average cure rates for each group:
Antibiotic Cure Rate =
W .W .Cure Rate =
2.
400
400
250
× .95 +
× .97 +
× .98 = .965
1050
1050
1050
Providers Assign Treatments -- ANOVA
If providers follow “inclinations”, we may end up with
something like:
Number of Patients Assigned by Providers to Each
Treatment Group by Severity Type
400
400
250
× .80 +
× .90 +
× .98 = .881
1050
1050
1050
• Unbiased average antibiotic treatment effect for the
entire population (.965-.881 = .084), but
patient group
antibiotics
watchful waiting
High
800
0
severity type
Medium
400
400
Low
0
500
• Estimate will vary with the average severity in the
population...E[L|T=1].
• To whom does it apply? A patient randomly chosen
43
from an urn? Are patients chosen from urns?
Expected average cure rates for each group:
3. Instrumental Variable Grouping – Further assume:
800
400
0
Antibiotic Cure Rate =
× .95 +
×.97 +
×.98 = .957
1200
1200
1200
W .W .Cure Rate =
44
0
400
500
× .80 +
× .90 +
×.98 = .944
900
900
900
• For this population the average treatment effect is on the
treated (800/1200*.15 + 400/1200*.07=.123).
a.
Information is available to the researcher to
approximate distances from patients to providers
• address of patient
• supply of providers in area around patients
b. Evidence suggests that patients in areas with more
physicians per capita have a higher probability of being
treated with antibiotics for their OM than patients in
areas with fewer physicians per capita.
• We
find a biased low estimate of the antibiotic treatment
effect for the average treated patient (.957 - .944 = .013 < .123).
• “Biased low” follows our theory as...
45
46
If “b” is true, divide 2100 patients into two groups based on
the physicians per capita in the area around their home:
Using our assumptions, does this grouping qualify as an
instrument?
Group 1: the group of patients living in areas with a higher
number of physicians per capita.
1. Doc supply related to treatment? Yes, if patients tend to go to
the closest provider for
treatment.
Group 2: the group of patients living in areas with a lower
number of physicians per capita.
47
If true, and providers follow inclinations we may see treatment
patterns something like:
Patient Treatment Assignments by Severity Type
patient
group
Group 1
High
100% antibiotics
Group 2
100% antibiotics
severity type
Medium
80% antibiotics
20% W.W.
30% antibiotics
70% W.W.
Low
100% W.W.
100% W.W.
48
8
Expected average estimated cure rates for these groups:
2. Is grouping related to unmeasured confounding variables
(e.g. severity)? Related to severity only if parents chose
residences in expectation of the severity of a future acute
condition.
If not related to severity, we assume equivalent severity
distributions across groups:
Number of Patients in Each Group by Severity Type
patient group
Group 1
Group 2
High
400
400
severity type
Medium
400
400
Group 1 Cure Rate =
250
400
320
80
×.95 +
×.97 +
×.90 +
×.98 = .959428
1050
1050
1050
1050
Group 2 Cure Rate =
400
120
280
250
×.95 +
×.97 +
×.90 +
×.98 = .946092
1050
1050
1050
1050
Well, (.959428 - .946092) = .013336 doesn't appear to reveal
much of anything…!
Low
250
250
49
• Remember the actual “unknown” cure rates for each
group by treatment are:
Now look at the antibiotic treatment rate in each group:
720/1050 = .68571 in Group 1
520/1050 = .4952381 in Group 2
treatment
antibiotics
watchful waiting
These differences also don't look very informative….
The IV change in the cure rates resulting from a one unit
increase in the drug treatment rate equals:
aˆ1IV =
50
.959428 − .946092
.013336
=
= .07
.68571 − .4952381
.190471905
• This estimate is the average difference in the antibiotic cure
rate for the marginal or in this example the “Medium”
severity patients.
High
.95
.80
severity type
Medium
.97
.90
.07
Low
.98
.98
• This estimate was found using only measured treatment
rates and outcome rates across “groups” that are
defined by the instruments.
• Which of the estimates above is the most important for
policy-makers wondering about over/underutilization of
a treatment?
51
52
• General IV Estimation Model
IV Brass Tacks
Treatment Choice Equation (1st stage):
• Where do instruments come from?
T = c + c ⋅ X + c ⋅Z + (v + c ⋅L
i
→ Theory on what motivated choices, not theory on
how choices can be motivated.
0
2
i
3
i
Outcome Equation (2nd stage):
i
1
i
)
Yi = a0 + a1 ⋅ Tˆi + a2 ⋅ X i + (ei + a3 ⋅ Li
→ Observed differences in:
-- guideline implementation (timing/interpretation)
-- product approval rules across payers
-- reimbursement differences across payers/geography
-- area provider “treatment signatures”
-- geographic access to relevant providers
-- provider market structure/competition
→ Generally, “Natural Experiments” (Angrist and Krueger,
2001)
53
)
Yi = 1 if health outcome occurs, 0 otherwise;
Xi = measured patient clinical characteristics;
Ti = 1 if patient received treatment, 0 otherwise;
Tˆi = predicted treatment from 1st stage;
Zi = a set of binary variables grouping patients based on
values of instrumental variables (from W); and
Li = unmeasured confounding variables assumed related
to both Y and T but not Z.
The only variation in T used to estimate a1 comes from Z.
54
9
• Define the IV estimate of a1 as
aˆ1IV
.
→ The estimate of a1 can only be definitively generalized
to the patients whose treatment choices were affected
by Z (Angrist, Imbens, Rubin 1996).
→ It can be shown that the expected value of aˆ1IV is:
E [â
1 IV
] ≈ b ⋅ E [ L |T ( Z )]
→ F-test of whether the parameters within c3 are
simultaneously equal to zero provides a test of the
first instrumental variable criterion:
1
Yields an average estimate of the treatment effect for
the set of patients whose treatment choices were
dependent on their value of Z.
Finding measured variables or “instruments” (Z) that:
a. are related to the possibility of a patient receiving
treatment (cov(T,Z) ≠ 0)
55
56
• How many groups?
→ Model can be estimated via:
-- Two-Stage Least Squares (2SLS) – PROC
SYSLIN in SAS.
-- Bivariate Probit – BIPROBIT function in STATA.
-- Two-Stage Replacement (e.g. Beenstock &
Rahav, 2002).
→ 2SLS offers consistent estimates that are
asymptotically normal with the fewest assumptions
(Angrist 2001).
-- essentially regressing group-level outcome rate
changes on group-level treatment rate changes.
→ Z can be specified as continuous variables, but results
are then conditional on this assumption and are less
interpretable.
→ Creating many groups from an instrument (more binary
variables in Z) uses more information and yields a
weighted average of many two-group comparisons, e.g.
-- low/high groups using the median of the instrument
VS
-- low/med low/med high/high groups using the
quartiles of the instrument.
→ Too many groups may introduce bias.
→ Best to report estimates for several grouping strategies.
57
58
Introduction
In a meta-analysis of observational studies,
Devereaux et al (1) found that patient survival at
for-profit dialysis centers was poorer than nonprofit centers.
Effect of Dialysis Center ProfitStatus on Patient Survival: An
Instrumental Variables Approach
Objective
Brooks, Irwin, Pendergast, Chrischilles, Flanigan,
Hunsicker
Compare estimates of the effect of dialysis center
profit status on patient survival using riskadjustment and IV estimation.
59
60
10
Sample
• N =
Instrumental Variable Strategy
101,669 incident ESRD patients from United States
Renal Data System (USRDS) from 1996-1999 that:
-- were between 67 and 100 years old at dialysis initiation;
-- had hemodialysis as initial modality;
-- obtained dialysis in a non-government dialysis facility;
-- had complete information on all model variables;
-- zip codes linked to 1990 census data.
Key Variable Definitions
• Outcome: one-year survival after dialysis initiation = 1,
0 otherwise.
• Treatment Setting: patient initiated dialysis in a for-profit
dialysis center = 1, 0 otherwise.
• Followed McClellan et al. (1994) and grouped patients
based on Differential Distance (DD) to various hospital
classifications:
DD
=
(DFP - DNP)
where
DFP = distance from patient residence to the nearest for
profit dialysis center; and
DNP = distance from patient residence to the nearest
non-profit dialysis center.
• Assessed whether IV estimates were robust to the
number of patient groups defined using differential
distance.
61
Percent Initial For-Profit and Number of Comorbidities
by Patient Differential Distance
% for-profit
comorbidities
0.8
% for-profit
4
0.6
3
0.4
0.2
2
0.0
for-profit closer
not for-profit closer
1
-100
-50
0
50
100
Miles a not for-profit is closer
63
“Marginal” End Stage Renal Disease Patients,
1996-1999
48.1%
92.8%
M
0%
More Likely
to For-Profit
50%
Table 1: Attributes of Dialysis Patient Groups, 1996-1999
Patient
Treatment Setting
Characteristics For-Profit Non-Profit
For Profit %
100
0
White %
70.6
73.3**
Black %
23.5
20.0**
Cardiac Failure % 42.6
44.2**
Diabetes %
45.1
41.3**
CerebroVasc Dis% 11.9
12.5**
Isch Heart Disc % 32.1
36.1**
AMI
11.6
13.2**
a
Reside in:
High Hlth State % 61.2
47.4**
Med Hlth State % 16.4
40.3**
Low Hlth State % 22.4
12.3**
Number
73,480
30,678
5
Average number of comorbidities
1.0
62
100%
Less Likely
to For-Profit
M = patients whose dialysis center choice is dependent on
the relative distance to for-profit and non-profit dialysis
centers – Marginal Patients.
65
Differential Distance (DD)
Patient Closer to:
For-Profit
Non-Profit
92.8
48.1**
74.9
67.9**
19.8
25.1**
43.1
43.1
44.8
43.1**
12.0
12.1
33.5
33.1
12.2
12.0
61.4
13.7
24.9
52,443
52.9**
33.3
13.9**
51,715
a.Subramanian, S, Kawacki I, et al. (2001). “Does the state you live in make a difference? A multilevel analysis of
self-rated health in the US.” Social Science & Medicine 53(1): 9-19.
**,* statistically significant at the .01 and .05 levels, respectively
64
Table 2: F-Statistics Testing Factors in Center Choice
Model are Related to the Use of For-Profit
Dialysis Facilities, 1996-1999.
Factora
Differential Distance (instrument)
Year
Gender
Age
Race
Comorbidity
Previous Healthcare Use
State of Residence
Distance to Nearest Center
Area Socioeconomic Status
Partial F-Statistics .
2150.53**
59.53**
5.81**
7.29**
6.26**
8.55**
2.63*
212.00**
75.41**
21.83**
a. specified using binary variables reflecting differences in respective characteristic.
Differential distance was used to group patients into 20 separate groups.
**,* statistically significant at the .01 and .05 levels, respectively
66
11
Table 3: 2SLS/IV and Ordinary Least Squares (OLS) Estimates of the
Effect of Initial For-Profit Initial Dialysis Provider Relative to a
Non-Profit Provider on 1-Year Patient Survival
Estimation Model
Number of Instrument
and Specification
Groups Specified
OLS no covariates
na
OLS Devereaux covariatesa
na
b
OLS Devereuax covariates plus na
2SLS/IVb
2
2SLS/IVb
5
2SLS/IVb
10
b
2SLS/IV
20
2SLS/IVb
40
Estimate (P-value)
-0.0031c (0.3450)
-0.0122c (<.0001)
-0.0071c (0.0511)
0.0009 (0.9264)
0.0025 (0.7373)
-0.00004 (0.9953)
-0.0002 (0.9823)
0.0006 (0.9349)
a. Factors consistently controlled for in the studies within the Devereaux meta-analysis – age,
gender, race, comorbidities.
b. Factors consistently controlled for in the studies within the Devereaux meta-analysis – age,
gender, race, comorbidities, plus dialysis year, state of residence, previous healthcare utilization,
provider access (distance to nearest dialysis center), socioeconomic status (patient zip percent
rural, percent poverty, and per capita income).
c. Logistic regression estimates were consistent in both magnitude and statistical significance.
67
OLS estimates were reported because their interpretation is more consistent with IV estimates.
Summary
• The foundation of IV estimation is theory that suggests
instruments – what factors motivated treatment choices.
• Ability to generalize is limited, but IV estimates offer a
more natural estimate of the effects of rate changes than
RCT estimates.
• Estimates can vary by sample and instrument used.
• Estimates are conditional on the truth (and acceptance)
of a known identification restriction. The source of the
treatment variation is known. The relationship between
this variation source and unmeasured confounders can
be debated.
68
References
Angrist JD, 2001. Estimation of Limited Dependent Variable Models with Dummy Endogenous Regressors: Simple Strategies for
Empirical Practice. Journal of Business & Economic Statistics. 19(1):2-16
Angrist, JD, Imbens GW, Rubin, DB. 1996. Identification of Causal Effects Using Instrumental Variables. Journal of the American
Statistical Association. 91:444-454.
Angrist JD, Krueger AB. 2001. Instrumental Variables and the Search for Identification: From Supply and Demand to Natural
Experiments. Journal of Economic Perspectives. 15(4): 69-85.
Beenstock M, Rahav G. Testing Gateway Theory: do cigarette prices affect illicit drug use? Journal of Health Economics
2002;21:679-98.
Brooks JM, Chrischilles E, Scott S, Chen-Hardee S. 2003. Was Lumpectomy Underutilized for Early Stage Breast Cancer? –
Instrumental Variables Evidence for Stage II Patients from Iowa. Health Services Research, 38(6):1385-1402.
Brooks JM, McClellan M, Wong H. 2000. The Marginal Benefits of Invasive Treatment for Acute Myocardial Infarction: Does
Insurance Coverage Matter? Inquiry, 37(1):75-90.
Devereaux, P., H. Schunemann, et al. (2002). “Comparison of Mortality Between Private For-Profit and Private Not-For-Profit
Hemodialysis Centers. A Systematic Review and Meta-analysis.” JAMA 288(19): 2449-2457.
Imbens GW, Angrist, JD. 1994. Identification and Estimation of Local Average Treatment Effects, Econometrica. 62(2):467-475.
McClellan M, McNeil BJ, Newhouse JP. 1994. Does More Intensive Treatment of Acute Myocardial Infarction in the Elderly Reduce
Mortality: Analysis Using Instrumental Variables", Journal of the American Medical Association. 272:859-866.
McClellan M, Newhouse JP. 1993. The Marginal Benefits of Medical Treatment Intensity. Cambridge, Mass: National Bureau of
Economic Research: Working Paper.
McClellan M, Newhouse JP. 1997. The Marginal Cost-Effectiveness of Medical Technology - a Panel Instrumental Variables
Approach, Journal of Econometrics. 77:39-64.
Subramanian, S, Kawacki I, et al. (2001). “Does the state you live in make a difference? A multilevel analysis of self-rated health
in the US.” Social Science & Medicine 53(1): 9-19.
69
12
Download