Identifying Predictors of Cognitive Change When the Outcome Is Measured With a Ceiling
Download
1 / 28

Identifying Predictors of Cognitive Change When the Outcome Is Measured With a Ceiling - PowerPoint PPT Presentation


  • 60 Views
  • Uploaded on

Identifying Predictors of Cognitive Change When the Outcome Is Measured With a Ceiling. Gerontological Society of America 2004 Annual Meeting Maria Glymour, Jennifer Weuve, Lisa F. Berkman, James M. Robins Harvard School of Public Health. Outline. The question Why it’s difficult to answer

loader
I am the owner, or an agent authorized to act on behalf of the owner, of the copyrighted work described.
capcha
Download Presentation

PowerPoint Slideshow about ' Identifying Predictors of Cognitive Change When the Outcome Is Measured With a Ceiling' - vance-hunter


An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript

Identifying Predictors of Cognitive Change When the Outcome Is Measured With a Ceiling

Gerontological Society of America

2004 Annual Meeting

Maria Glymour, Jennifer Weuve, Lisa F. Berkman, James M. Robins

Harvard School of Public Health


Outline
Outline Is Measured With a Ceiling

  • The question

  • Why it’s difficult to answer

  • How CLAD regression helps

  • An example with HRS data


The question
The Question Is Measured With a Ceiling

  • Does education affect cognitive change in old age?

  • Earl attended 10 years of school and declined by 2 points on a cognitive test score from age 70 to 75.

    Would Earl have experienced more or less cognitive change if he had, counter to fact, completed more schooling?


Indirect measurement of cognition
Indirect Measurement of Cognition Is Measured With a Ceiling

  • Test is an indirect measure of our primary interest (cognitive function):

    Test Score=g(cognition) + e

  • But the test has a maximum possible score:

    Test Score=min(15, g(cognition) + e)


Scaling challenges
Scaling Challenges Is Measured With a Ceiling

True Cognitive Status Values

Low

High

Maximum text score

Measured Test Score


Measurement ceilings
Measurement Ceilings Is Measured With a Ceiling

A ceiling on the dependent variable will bias the regression coefficient away from the coefficient for the true outcome variable.


Ceilings with longitudinal data

36 Is Measured With a Ceiling

36

34

34

32

32

30

30

28

28

26

26

24

24

22

22

20

20

18

18

16

16

0

1

Time

Ceilings with Longitudinal Data

Difference in True D = 0

Observed = -3


Ceilings with longitudinal data1

36 Is Measured With a Ceiling

36

34

34

32

32

30

30

28

28

26

26

24

24

22

22

20

20

18

18

16

16

0

1

Time

Ceilings with Longitudinal Data

Difference in True D = 0

Observed = 3


Medians vs means

800 Is Measured With a Ceiling

Test Score

600

400

200

0

Median

Mean

Medians vs Means

800

600

400

Cognitive Status

200

0

Mean, Median


Clad regression
CLAD Regression Is Measured With a Ceiling

  • The median is more robust to ceiling effects than the mean, so contrast medians by level of exposure

  • Use CLAD if believe the relationship between X and Y does not differ above (vs below) the ceiling

    • Calculate the median regression coefficients

    • Drop observations with a predicted value of Y over ceiling

    • Repeat steps 1 and 2 until all predicted values are below the ceiling.

  • Standard errors are messy: bootstrap.

  • Can use any quantile


Data set
Data Set Is Measured With a Ceiling

  • AHEAD cohort of HRS

    • Enrolled in 1993

    • National sample of non-institutionalized survivors born pre-1924

    • n=7,542, Observations=23,752

  • Self-report years of education: dichotomized at <12 years

  • Telephone Interview for Cognitive Status (modified)

    • Possible range 0 (bad) -15 (good)

    • ~20% scored max at each interview

    • Assessed 1-5 times


Analysis

TICS Is Measured With a Ceilingti = b00 + b1Timeti + b2Educationi

+ b3Timeti*Educationi

+ bkOther Covariatesti + ei

Bootstrap (500 resamples) for standard errors, resampling on the individual (rather than the observation)

Analysis


Analysis1

Other covariates: Is Measured With a Ceiling

Age at enrollment, mother’s education, father’s education, Hispanic ethnicity

Stratify by sex and race (black vs all other)

Up to 5 cognitive assessments

Initial models treat time flexibly

Impose a linear model of time

Analysis


Summary of ahead data
Summary of AHEAD Data Is Measured With a Ceiling


Predicted median tics score
Predicted Median TICS Score Is Measured With a Ceiling

From CLAD models, adjusted for sex, race, age at baseline, Hispanic ethnicity, mother’s and father’s education


Baseline education effect estimates
Baseline Education Effect Estimates Is Measured With a Ceiling


Slope education effect estimates
Slope Education Effect Estimates Is Measured With a Ceiling


Loss to follow up
Loss to Follow-Up Is Measured With a Ceiling


Effect at alternative quantiles
Effect at Alternative Quantiles Is Measured With a Ceiling


Less desirable alternatives
(Less Desirable) Alternatives Is Measured With a Ceiling

  • Baseline adjustment

    • Introduces new (and larger) biases

  • Add the scales

    • Hides the ceiling

    • Hides the bias

  • Tobit models

    • Stronger assumptions about the distribution


Conclusions
Conclusions Is Measured With a Ceiling

  • More educated respondents had much higher average cognitive scores for the duration of the study.

  • Education associated with better evolution of cognitive function for white women.

  • Ceilings introduced bias of unknown direction.


Limitations future work
Limitations & Future Work Is Measured With a Ceiling

  • Discrete outcomes

  • Missing data

  • Complex sampling design

  • Unequal scale intervals not due to ceilings


Acknowledgements
Acknowledgements Is Measured With a Ceiling

  • Dean Jolliffe, CLAD ado

  • Funding:

    • National Institute of Aging

    • Office for Behavioral and Social Science Research

    • “Causal Effects of Education on Elder Cognitive Decline”

    • AG023399

    • NIA Training grant: AG00138


END Is Measured With a Ceiling


Unequal scale intervals

True Cognitive Status Values Is Measured With a Ceiling

Low

High

Measured MMSE

Unequal Scale Intervals


Unequal scale intervals1

6 Is Measured With a Ceiling

4.0

3.5

5

3.0

4

2.5

function=square root of errors

Function=ln(score)

3

2.0

1.5

2

1.0

1

0.5

0

-

0

5

10

15

20

25

30

Test Score

Unequal Scale Intervals

Do similar size increments have the same “meaning” across all levels of the test?


Unequal scale intervals2
Unequal Scale Intervals Is Measured With a Ceiling

Do similar size increments have the same “meaning” across all levels of the test?


Unequal scale intervals3
Unequal Scale Intervals Is Measured With a Ceiling

Do similar size increments have the same “meaning” across all levels of the test?


ad