5. Logistic Regression

Page 1 of 65

Unit 5 Logistic Regression “To all the ladies present and some of those absent” - Jerzy Neyman

What behaviors influence the chances of developing a sexually transmitted disease? Comparing demographics, health education, access to health care, which of these variables are significantly associated with failure to obtain an HIV test? Among the several indicators of risk, including age, co-morbidities, severity of disease, which are significantly associated with surgical mortality among patients undergoing transplant surgery? In all of these examples, the outcome observed for each individual can take on only one of two possible values: positive or negative test, alive or dead, remission or non-remission, and so on. Collectively, the data to be analyzed are proportions. Proportions have some important features that distinguish them from data measured on a continuum. Proportions (1) are bounded from below by the value of zero (or zero percent) and bounded from above by one (or 100 percent); (2) as the proportion gets close to either boundary, the variance of the proportion gets smaller and smaller; thus, we cannot assume a constant variance; and (3) proportions are not distributed normal. Normal theory regression models are not appropriate for the analysis of proportions. In unit 4, Categorical Data Analysis, emphasis was placed on contingency table approaches for the analysis of such data and it was highlighted that these methods should always be performed for at least two reasons: (1) they give a good feel for the data; and (2) they are free of the assumptions required for regression modeling. Unit 5 is an introduction to logistic regression approaches for the analysis of proportions where it is of interest to explore the roles of possibly several influences on the observed proportions.

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 2 of 65

Table of Contents

Topic

Appendix

Nature

Learning Objectives …………………………….…………………………….

3

1. From Linear Regression to Logistic Regression …................................…..

4

2. Use of VDT’s and Spontaneous Abortion …..............................................

5

3. Definition of the Logistic Regression Model .…................................…….

7

4. Estimating Odds Ratios …………………................................……………

11

5. Estimating Probabilities …………….........................……………..………

17

6. The Deviance Statistic …………….........................……….…………….. a. The Likelihood Ratio Test …….........................………………………. b. Model Development …………................................………………….

18 20 23

7. Illustration – Depression Among Free-Living Adults

25

………..………

8. Regression Diagnostics ……………...........................………....………… a. Assessment of Linearity ………………………………….…………… b. Hosmer-Lemeshow Goodness of Fit Test ……..................................…. c. The Linktest ………………………………………………..…………. d. The Classification Table ………………..................................……….. e. The ROC Curve ……………………….................................…………. f. Pregibon Delta Beta Statistic ………….................................…………

37 40 41 44 46 49 51

9. Example - Disabling Knee Injuries in the US Army ...................................

53

Overview of Maximum Likelihood Estimation ………..............................…

61

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 3 of 65

Learning Objectives

When you have finished this unit, you should be able to: §

Explain why a normal theory regression model is not appropriate for a regression analysis of proportions.

§

State the expected value (the mean) of a Bernoulli random variable.

§

Define the logit of the mean of a Bernoulli random variable.

§

State the logistic regression model and, specifically, the logit link that relates the logit of the mean of a Bernoulli random variable to a linear model in the predictors.

§

Explain how to estimate odds ratio measures of association from a fitted logistic regression model.

§

Explain how to estimate probabilities of event from a fitted logistic regression model.

§

Perform and interpret likelihood ratio test comparisons of hierarchical models.

§

Explain and compare crude versus adjusted estimates of odds ratio measures of association.

§

Assess confounding in logistic regression model analyses.

§

Assess effect modification in logistic regression model analyses.

§

Draft an analysis plan for multiple predictor logistic regression analyses of proportions.

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 4 of 65

1. From Linear Regression To Logistic Regression An Organizational Framework In unit 2 (Regression and Correlation), we considered single and multiple predictor regression models for a single outcome random variable Y assumed continuous and distributed normal. In unit 5 (Logistic regression), we consider single and multiple regression models for a single outcome random variable Y assumed discrete, binary, and distributed bernoulli.

Unit 2 Normal Theory Regression

Unit 5 Logistic Regression

- univariate - continuous - Example: Y = cholesterol - one or multiple - discrete or continuous - treated as fixed

- univariate - discrete, binary - Example: Y = dead/alive - one or multiple - discrete or continuous - treated as fixed

Y | X1=x1, .., Xp=xp

- Normal (Gaussian)

- Bernoulli (or binomial)

E(Y| X1=x1, . Xp=xp)

µ Y|X1 ...Xp =β0 +β1x1 +...+β p x p

µ Y|X1 ...Xp = π Y|X1 ...Xp

Y X1, X2, ….., Xp

=

1

1+exp ⎡⎣- (β 0 +β1x1 +...+β p x p )⎤⎦

Right hand side of model

β0 +β1x1 +...+β p x p

β0 +β1x1 +...+β p x p

Link

"natural" or "identity" µ Y|X ...X

"logit" logit[µ Y|X1 ...Xp ]

= β0 +β1x1 +...+β p x p

= logit[π Y|X1 ...Xp ]

1

p

(

)

=ln ⎡ π Y|X1 ...Xp 1 − π Y|X1 ...Xp ⎤ ⎣ ⎦ = β 0 +β1x1 +...+β p x p Estimation Tool Tool

Nature

Least squares (= maximum likelihood) Residual sum of squares Partial F Test

Population/ Sample

Observation/ Data

Maximum Likelihood Deviance statistic Likelihood Ratio Test

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 5 of 65

2. Use of Video Display Terminals and Spontaneous Abortion Consider the following published example of logistic regression. Source: Schnorr et al (1991) Video Display Terminals and the Risk of Spontaneous Abortion. New England Journal of Medicine 324: 727-33. Background: Adverse pregnancy outcomes were correlated with use of video display terminals (VDT’s) beginning in 1980. Subsequent studies were inconsistent in their findings. Previous exposure assessments were self-report or derived from job title descriptions. Electromagnetic fields were not previously measured. Research Question: What is the nature and significance of the association, as measured by the odds ratio, between exposure to electromagnetic fields emitted by VDTs and occurrence of spontaneous abortion, after controlling for -

History of prior spontaneous abortion Cigarette Smoking History of thyroid condition

Design: Retrospective cohort investigation of two groups of full-time female telephone operators.

882 Pregnancies: Exposed Unexposed

Nature

Population/ Sample

N 366 516

Observation/ Data

Spontaneous Abortion n % 54 14.8% 82 15.9%

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 6 of 65

The Data:

Variable

Label

Range/Codes

AVGVDT

average hours vdt in 1st trimester

continuous

NUMCIGS

# cigarettes/day

continuous

PRIORSAB

prior spontaneous abortion

1=yes, 0=no

SAB

spontaneous abortion

1=yes, 0=no

SMOKSTAT

smoker

1=yes, 0=no

PRTHYR

prior thyroid condition

1=yes, 0=no

VDTEXPOS

VDT exposure

1=yes, 0=no

AVGVDT 0.000 0.000 0.000 27.764 28.610 0.000 19.717 0.000 25.022 … 0.000

Nature

NUMCIGS PRIORSAB SAB SMOKSTAT PRTHYR VDTEXPOS 15 0 0 1 0 0 10 0 0 1 0 0 20 0 0 1 0 0 20 0 0 1 0 1 20 0 1 1 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 … … … … … … 0 1 0 0 0 0

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 7 of 65

3. Definition of the Logistic Regression Model We suspect that multiple factors, especially use of video display terminals, contribute to an individual’s odds of spontaneous abortion. The outcome or dependent variable is Y=sab. Its value is y and = 1 if spontaneous abortion occurred 0 otherwise The predictors that might influence the odds of SAB are several: X1 = avgvdt X2 = numcigs X3 = priorsab X4 = smokstat X5 = prthyr, and X6 = vdtexpos We are especially interested in X6 = vdtexpos and X1 = avgvdt Among the N=882 in our sample, we have potentially N=882 unique probabilities of spontaneous abortion. π1, π2, …, πN. For the ith person πi = Function ( X1i, X2i, X3i, X4i, X5i, X6i) Pr [ Yi = 1] = πi Pr [ Yi = 0] = (1 - πi)

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 8 of 65

How do we model the N=882 individual probabilities π i in relationship to the predictors? Review of normal theory linear regression analysis: Y is assumed to be distributed normal (Gaussian) with mean=µ x and variance= σ Y | X . 2

The mean of Y is modeled linearly in x = [X1, X2, X3, X4, X5, X6] Thus mean of Y = E [Y] = µ x In normal theory linear regression:

E[Y] = µ x

“natural link”

= β0 + β1 X 1 + β2 X 2 + β3 X 3 + β4 X 4 + β5 X 5 + β6 X 6

“right hand side is linear in the predictors”

In a logistic model regression analysis, the framework is a little different: Y is assumed to be distributed Bernoulli with mean=πx and variance= πx (1-πx) We do not model the mean of Y = πx linearly in x = [X1 … X6]. Instead, we model the logit of the mean of Y = πx linearly in x = [X1 … X6].

⎡ π ⎤ Logit [ E(Y) ] = logit[ πx] = ln ⎢ x ⎥ = β0 + β1X1 + β2X2 + β3X3 + ...+ β5X5 + β6X6 ⎣ 1− π x ⎦

“logit link”

Nature

“right hand side is linear in the predictors”

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 9 of 65

Solution for Probability [Y=1| X1=x1, X2=x2, …, X6=x6] = E[Y | X1=x1, X2=x2, …, X6=x6 ] : Pr [ Y = 1| X1=x1, X2=x2, …, X6=x6 ] is

exp (β 0 +β1x1 +β 2 x 2 +...+β 6 x 6 ) 1+exp (β 0 +β1x1 +β 2 x 2 +...+β 6 x 6 )

πx =

1 1 + exp ⎡⎣ − (β 0 +β1x1 +β 2 x 2 +...+β 6 x 6 )⎤⎦

=

Pr [ Y = 0 | X1=x1, X2=x2, …, X6=x6 ] is

(1-π ) x

=

1 1+exp (β 0 +β1x1 +β 2 x 2 +...+β 6 x 6 )

Two other names for this model are “log-linear odds” and “exponential odds” The logistic regression model focuses on the odds of event (in this case event of spontaneous abortion, SAB).

⎡ πx ⎤ … ⎥ = β0 + + β6X6 is a log-linear odds model. ⎣1 − π x ⎦

ln [ odds (πx) ] = ln ⎢

1)

⎡ πx ⎤ … ⎢ ⎥ = exp { β0 + + β6X6 } is an exponential odds model. ⎣1 − π x ⎦

2)

We do not model E[Y | X ] = π x= β 0 + β 1X1 + β 2X2 + β 3X3 + β 4X4+ β 5X5 + β 6X6? 1)

β0 + β1X1 + β2X2 + β3X3 + β4X4+ β5X5 + β6X6 can range from -∞ to +∞ but πx ranges from 0 to 1. 2) πx= β0 + β1X1 + β2X2 + β3X3 + β4X4+ β5X5 + β6X6 is often not a good description of nature.

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 10 of 65

Assumptions:

1) Each Yi follows a distribution that is Bernoulli with parameter E[Y | X ] =

π

xi

.

2) The Y1, Y2, … , YN are independent.

3) The values of the predictors, Xi1=xi1

…

Xi6=xi6, are treated as fixed.

4) The model is correct (this is also referred to as “linearity in the logit”). logit[ E(Y)| X1=x1, X2=x2, …, X6=x6 ] = logit [ πx] = β0 + β1X1 + β2X2 + β3X3 + β4X4+ β5X5 + β6X6

5)

No multicollinearity

6) No outliers 7) Independence of Errors

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 11 of 65

3. Estimating Odds Ratios For now, assume that we have a fitted model. We’ll get to the details of estimation later. Once a logistic regression model has been fit, the prediction equation can be used to estimate odds ratio (OR) measures of association. Example 1: What is the estimated crude relative odds (OR) of spontaneous abortion (SAB) associated with any exposure to a video display terminal (VDTEXPOS)? Step 1: To obtain crude odds ratios, either a 2x2 table can be used or a one predictor logistic regression model can be fit. Here, it is given by

logit { probability [SAB=1] } = β 0 + β 1 VDTEXPOS Stata . use “http://people.umass.edu/biep640w/datasets/vdt.dta”, clear . logit sab vdtexpos Logit estimates

Number of obs LR chi2(1) Prob > chi2 Pseudo R2

Log likelihood = -379.08045

= = = =

882 0.21 0.6443 0.0003

-------------------------------------------------------------------------------------sab | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+------------------------------------------------------------------------vdtexpos |

-.0876939

_cons |

-1.666325

= βˆ 1 = βˆ 0

.1903232 .1204129

-0.46 -13.84

0.645 0.000

-.4607204 -1.90233

.2853327 -1.43032

---------------------------------------------------------------------------------------

Yielding the following prediction equation

Fitted logit { pr[sab=1] } = -1.66633 - 0.08769*vdtexpos

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 12 of 65

Step 2: Recognize a wonderful bit of algebra.

For a single exposure variable (1=exposed, 0=not exposed) OR 1 versus 0

= exp{ β } where β = regression parameter for the exposure variable = exp { logit (π1) – logit (π0) }

Proof (read if you are interested!): OR = exp {ln [OR ]}

⎪⎧ ⎡ π /(1 − π1 ) ⎤ ⎪⎫ = exp ⎨ln ⎢ 1 ⎥⎬ ⎪⎩ ⎣ π 0 /(1 − π 0 ) ⎦ ⎪⎭ ⎡ π ⎤ ⎪⎫ ⎪⎧ ⎡ π ⎤ = exp ⎨ln ⎢ 1 ⎥ − ln ⎢ 0 ⎥ ⎬ ⎣1 − π 0 ⎦ ⎭⎪ ⎩⎪ ⎣1 − π1 ⎦

= exp {logit ( π1 ) -logit(π0 )} “1” is the comparison and is vdtexpos=1:

Estimated logit { prob[SAB=1|vdtexpos=1] } = βˆ0 = -1.66633 - 0.08769 “0” is the reference and is vdtexpos=0: Estimated logit { prob[SAB=1| vdtexpos=0] } =

+ βˆ1

βˆ0 =

-1.66633

Step 3: Apply. The odds ratio measure of association comparing the exposed telephone operator (“1”) to the unexposed telephone operator (“0”) is = = = = =

Nature

exp { logit (π1) – logit (π0) } exp { [ β0 + β1 ] - [ β0] } exp { β1 } exp { -0.08769} 0.9160

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 13 of 65

Stata Illustration – Obtaining estimated odds ratios after logistic regression Method 1. Command logit with option or . logit sab vdtexpos, or

Method 2. Command logistic . logistic sab vdtexpos

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 14 of 65

The two profiles being compared can differ on several predictors! Let’s try another one. Here is another bit of wonderful algebra.

For two profiles of predictor variable values, “comparison” versus “reference” OR comparison

versus reference

= exp { logit (πcomparison) – logit (πreference) }

Example 2 - What is the estimated relative odds (OR) of spontaneous abortion (SAB) for a person who is not exposed to a VDT, smokes 10 cigarettes per day, has no history of prior SAB, and no thyroid condition relative to a person who has an average of 20 hours exposure to a VDT, is a nonsmoker, has a history of prior SAB and does have a thyroid condition?

Step 1: Here the model fit is the 4 predictor model logit { probability [sab=1] } = β0 + β1 avgvdt + β2 numcigs + β3 priorsab + β4 prthyr Estimation now yields (output not shown). fitted logit { prob[sab=1] } = -1.95958 + 0.00508(avgvdt) + 0.04267(numcigs) + 0.38500(priorsab) + 1.27420(prthyr)

Step 2: Calculate the two logits and compute their difference.

avgvdt numcigs priorsab prthyr

Nature

Population/ Sample

Value of Predictor for Person “comparison” “reference” 0 20 10 0 0 1 0 1

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 15 of 65

“comparison” logit [ πcomparison ] = -1.95958 + 0.00508(0) + 0.04267(10) + 0.38500(0) + 1.27420(0) = -1.5329 “reference”: logit [πreference] = -1.95958 + 0.00508(20) + 0.04267(0) + 0.38500(1) + 1.27420(1) = -0.1988 logit [πcomparison ] - logit [πreference ] = -1.5329 – [-0.1988] = -1.3341

Step 3: Exponentiate.

ORcomparison versus reference = exp { logit [πcomparison ] - logit [πreference] } = exp { -1.3341 } = 0.2634

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 16 of 65

In General:

ˆ ) of association with outcome The Odds Ratio estimate ( OR accompanying a unit change in the predictor X is a function of the estimated regression parameter

βˆ

ORˆ UNIT change in X = exp { βˆ } Tip – OR10 unit change in X = exp [ 10*β ]

A hypothesis test of OR=1 Is equivalent to A hypothesis test of β = 0

ˆ ) estimate of For a rare outcome (typically disease), the relative risk ( RR association with outcome accompanying a unit change in the predictor X is reasonably estimated as a function of the estimated regression parameter β RRˆ UNIT change in X = exp { βˆ }, approximately

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 17 of 65

5. Estimating Probabilities Again, let’s assume that we have a fitted model. We’ll get to the details of estimation later.

Once a logistic regression model has been fit, the prediction equation can also be used to estimate probabilities of event occurrence. The prediction equation can be used to estimate probabilities of event of disease if the study design is a cohort or probabilities of history of exposure if the study design is case-control. Reminder …– it is not possible to estimate probability of disease from analyses of case-control studies. Recall that for Y distributed Bernoulli E [ Y ] = π = Probability of event occurrence Example 1- Under the assumption of a cohort study design, what is estimated probability of spontaneous abortion (sab) for a person with any exposure to a video display terminal? Consider the single predictor model containing the predictor vdtexpos)

Step 1: Recall that we obtained the following equation for the fitted logit for the one predictor model containing VDTEXPOS: Predicted logit { prob[SAB=1| vdtexpos] }

= βˆ 0 + βˆ1[vdtexpos]

= -1.66633 - 0.08769*VDTEXPOS

Step 2: Utilizing the algebra on page 9, we have:

Estimated pr[SAB=1] = πˆ VDTEXPOS=1 =

(

exp βˆ 0 +βˆ 1[vdtexpos]

(

)

1+exp βˆ 0 +βˆ 1[vdtexpos]

)

=

(

exp βˆ 0 +βˆ 1

(

)

1+exp βˆ 0 +βˆ 1

)

Step 3:

Set VDTEXPOS=1, β0 = -1.66633, β1 =-0.08769 and solve

Estimated pr[SAB=1]=

=

Nature

Population/ Sample

exp ( -1.66633 - 0.08769[1]) 1+exp ( -1.66633 - 0.08769[1])

0.1731 = 0.148 1.1731

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 18 of 65

6. The Deviance Statistic ”G Statistic”, “Log likelihood Statistic”, “Scaled Deviance”, Residual Deviance”” Where are we now? Recall the concept of “analysis of variance” introduced in Unit 2, Regression and Correlation. Analysis of variance is about the total variability of the observed outcome, and its partitioning into portions that are explained by the fitted model (due model/due regression) versus what’s left over as unexplained (due residual/due error). The deviance statistic in logistic regression is a measure of what remains left over as unexplained by the fitted model, analogous to the residual sum of squares in normal theory regression.

But first, a few words about likelihood, L. Lsaturated :

We get the largest likelihood of the data when we fit a model that allows a separate predictor for every person. This is called the likelihood of the saturated model. Lsaturated is a large number.

Lcurrent:

We get an estimated likelihood of the data when we fit the current model. Lcurrent is a smaller number.

The deviance statistic in logistic regression is related to the two likelihoods, Lcurrent and Lsaturated in the following way.

The current model explains a lot Lcurrent ≈ Lsaturated

L L

current

The current model does NOT explain a lot Lcurrent < < Lsaturated

L L

≈ 1

current

saturated

> 0 ⎣ Lsaturated ⎦

A number close to 0

A large positive number

Evidence that the current model explains a lot of the variability in outcome Deviance ≈ small p-value ≈ large

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

⎡ L

Page 19 of 65

⎤

Deviance Statistic, D = -2 ln ⎢ current ⎥ ⎣ Lsaturated ⎦ = (-2) ln (Lcurrent) - (-2) ln (Lsaturated) Deviance df = [Sample size] – [# fitted parameters] where Lcurrent = likelihood of data using current model Lsaturated = likelihood of data using the saturated model

Notes (1) By itself, the deviance statistic does not have a well defined distribution (2) However, differences of deviance statistics that compare hierarchical models do have well defined distributions, namely chi square distributions.

A Feel for the Deviance Statistic (1) Roughly, the deviance statistic D is a measure of what remains unexplained. Hint – The analogue in normal theory regression is the residual sum of squares (SSQ error)

(2) A deviance statistic value close to zero says that a lot is explained and, importantly, that little remains unexplained. à The current model with its few predictors performs similarly to the saturated model that permits a separate predictor for each person. (3) WARNING! The deviance statistic D is NOT a measure of goodness-of-fit. Recall that we said the same thing about the overall F-statistic in normal theory regression. (4) The deviance statistic D is the basis of the likelihood ratio test . (5) The likelihood ratio test is used for the comparison of hierarchical models.

Recall – In normal theory regression, hierarchical models are compared using the Partial F-test.

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 20 of 65

a. The Likelihood Ratio (LR) Test

Likelihood Ratio (LR) Test Under the assumptions of a logistic regression model and the comparison of the hierarchical models:

Reduced: logit[π | X1 ,X 2 ...,X p ] = β 0 +β1X1 +...+β p X p Full: logit[π | X1 ,X 2 ...,X p ,X p+1 ,X p+2 ,...,X p+k ] = β 0 +β1X1 +...+β p X p + β p+1X p+1 +...+β p+k X p+k For testing:

HO : β p+1 = β p+2 = ... = β p+k = 0 H A : not A Likelihood Ratio Test Statistic LR, defined LR = DevianceREDUCED - DevianceFULL = [ (-2) ln (L) REDUCED – (-2) ln(L)SATURATED ] = [ (-2) ln (L) REDUCED ] -

[ (-2) ln (L) FULL – (-2) ln(L)SATURATED]

[ (-2) ln (L) FULL ]

has null hypothesis distribution that is Chi SquareDF=k Thus, rejection of the null hypothesis occurs for Test statistic values, LR = large and accompanying p-value= small Tip – In practice, we obtain LR using the 2nd formula; it says: LR = [ (-2) ln (L) REDUCED ] -

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

[ (-2) ln (L) FULL ]

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 21 of 65

Example: Controlling for prior spontaneous abortion (PRIORSAB), is 0/1 exposure to VDT associated with spontaneous abortion? The idea here is similar to the idea of the partial F test in normal theory linear regression. Two models that are hierarchical are compared: a “reduced/reference” versus a “full/comparison”.

Step 1: Fit the “reduced/reference” model, defined as containing the control variable(s) only. (Note – The available sample size here is 881) It estimates that logit {pr [sab=1]} = β0 + β1 PRIORSAB (-2) ln Lreduced = 754.56 Deviance DFreduced = 881 – 1 = 880

Step 2: Fit the “full/comparison” model, defined as containing the control variable(s) + predictor(s) of interest. It estimates that logit {pr [sab=1]} = β0 + β1 PRIORSAB + β2 VDTEXPOS (-2) ln Lfull = 753.81 Deviance DFfull = 881 – 2 = 879

Step 3: Compute the change in deviance and the change in deviance df, remembering that in logistic regression the subtraction is of the form “reduced” - “full”. Likelihood Ratio Test LR = (-2) ln Lreduced - (-2) ln Lfull = 754.56 - 753.81 = 0.75 Δ Deviance Df = Deviance DFreduced - Deviance DFfull = 880 - 879 = 1

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 22 of 65

Example – continued. H0: VDTEXPOS, controlling for PRIORSAB, is not associated with SAB βVDTEXPOS = 0 in the model that also contains PRIORSAB HA: VDTEXPOS, controlling for PRIORSAB, is associated with SAB βVDTEXPOS ≠ 0 in the model that also contains PRIORSAB Suppose we obtain: Likelihood Ratio Statistic χ2(df=1) = 0.75 p-value = .39 Interpretation. Assumption of the null hypothesis βVDTEXPOS = 0 and its application to the observed data yields a result that is reasonably plausible (p-value=.39). The null hypothesis is NOT rejected. Conclude that there is not statistically significant evidence that exposure to VDT, after controlling for prior spontaneous abortion, is associated with spontaneous abortion. Note - A little algebra (not shown) reveals that there are two, equivalent, formulae for the LR test: Solution #1 LR Test = Δ Deviance Statistic [ Deviance (reduced model) ] - [ Deviance (full model) ] Solution #2: this works because ln likelihood (saturated) drops out LR Test = Δ Deviance = Δ { (-2) ln (likelihood) ] = [ (-2) ln likelihood (reduced model) ] - [ (-2) ln likelihood (full model) ]

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 23 of 65

b. Model Development Goal:

The approach to model development in multiple logistic regression analysis is similar to the approach in normal theory multiple linear regression. Hierarchical models are compared to assess the statistical significance of the extra predictors in the larger model, controlling for the predictors in the smaller model. In logistic regression, this is done using the likelihood ratio test. If the likelihood ratio statistic is statistically significant (small p-value), we say that the added variables are statistically significant after adjustment for the control variables.

Example – Depression Among Free-Living Adults. Among free-living adults of Los Angeles County, what is the prevalence of depression and what are its correlates? In particular, in a given data set containing information on several candidate predictors, which predictors are the significant ones?

A reasonable analysis approach for this particular example is the following: Step 1. Fit single predictor models. Retain for further consideration: • Predictors with crude significance levels of association p 15% or so.

Step 6. Investigate effect modification • Begin with the “near final” model identified in step 5 • Fit, one at a time, enhanced models that contain each pairwise interaction • Assess statistical significance of each interaction using the LR test

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 25 of 65

7. Illustration Depression Among Free-Living Adults Source: Frerich RR, Aneshensel CS and Clark VA (1981) Prevalence of depression in Los Angeles County. American Journal of Epidemiology 113: 691-99.

Background The data for this illustration is a subset of n=294 observations from the original study of 1000 adult residents of Los Angeles County. The purpose of the original study was to estimate the prevalence of depression and to identify the predictors of, and outcomes associated with, depression. The study design was a longitudinal one that included four interviews In this illustration, only data from the first interview are used. Thus, this example is a cross-sectional analysis to identify the correlates of prevalent depression. Among these n=294, there are 50 events of prevalent depression.

Codebook: Variable

Label

Range/Codes

depressed

Case of depression

1=yes, 0 =no

age

Age, years

continuous

income

Income, thousands of dollars

continuous

female

Female gender

1=female, 0=male

unemployed

Unemployed

1=unemployed, 0=other

chronic

Chronic illness in past year

1=yes, 0=no

alcohol

Current alcohol use

1=yes, 0=no

Goal Perform a multiple logistic regression analysis of these data to identify the correlates of prevalent depression.

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 26 of 65

Preliminary. Describe the analysis sample. . use "http://people.umass.edu/biep640w/datasets/depress_small.dta" (Depression Data Small Version) . codebook, compact Variable Obs Unique Mean Min Max Label ---------------------------------------------------------------------------------------age 294 66 44.41497 18 89 age in years at last birthday alcohol 294 2 .7959184 0 1 chronic 294 2 .5068027 0 1 depressed 294 2 .170068 0 1 female 294 2 .622449 0 1 income 294 30 20.57483 2 65 thousands of dollars per year unemployed 294 2 .047619 0 1 ---------------------------------------------------------------------------------------Looks reasonable. There are no missing data. All of the binary variables are coded 0/1. The 2 continuous variables have reasonable ranges. . * Continuous variable distributions: . sort depressed

by depression status

. tabstat age, by(depressed) col(stat) stat(n mean sd min q max) format(%8.2f) longstub depressed variable | N mean sd min p25 p50 p75 max -----------------------+-------------------------------------------------------------------------------normal age | 244.00 45.24 18.15 18.00 29.00 43.50 59.00 89.00 depressed age | 50.00 40.38 17.40 18.00 26.00 34.50 51.00 79.00 -----------------------+-------------------------------------------------------------------------------Total age | 294.00 44.41 18.09 18.00 28.00 42.50 59.00 89.00 --------------------------------------------------------------------------------------------------------

Depressed persons tend to be younger.

Variability is comparable.

. tabstat income, by(depressed) col(stat) stat(n mean sd min q max) format(%8.2f) longstub depressed variable | N mean sd min p25 p50 p75 max -----------------------+-------------------------------------------------------------------------------normal income | 244.00 21.68 15.98 2.00 9.00 17.00 28.00 65.00 depressed income | 50.00 15.20 9.84 2.00 7.00 13.00 23.00 45.00 -----------------------+-------------------------------------------------------------------------------Total income | 294.00 20.57 15.29 2.00 9.00 15.00 28.00 65.00 --------------------------------------------------------------------------------------------------------

Depressed persons tend to be lower income. Also, the variability in income is less (sd=9.84 for depressed, sd=15.98 for non-depressed).

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

. * . * Discrete variable distributions:

Page 27 of 65

by depression status

. tab2 alcohol depressed, row exact | depressed alcohol | normal depressed | Total ------------+----------------------+---------non-drinker | 51 9 | 60 | 85.00 15.00 | 100.00 ------------+----------------------+---------drinker | 193 41 | 234 | 82.48 17.52 | 100.00 ------------+----------------------+---------Total | 244 50 | 294 | 82.99 17.01 | 100.00 Fisher's exact = 1-sided Fisher's exact =

Depression is more prevalent among drinkers. but this is not statistically significant.

0.705 0.402

. tab2 chronic depressed, row exact | depressed chronic | normal depressed | Total ----------------+----------------------+---------other | 126 19 | 145 | 86.90 13.10 | 100.00 ----------------+----------------------+---------chronic illness | 118 31 | 149 | 79.19 20.81 | 100.00 ----------------+----------------------+---------Total | 244 50 | 294 | 82.99 17.01 | 100.00 Fisher's exact = 1-sided Fisher's exact =

Depression is slightly more prevalent among the ill.

0.089 0.054

. tab2 female depressed, row exact | depressed female | normal depressed | Total -----------+----------------------+---------male | 101 10 | 111 | 90.99 9.01 | 100.00 -----------+----------------------+---------female | 143 40 | 183 | 78.14 21.86 | 100.00 -----------+----------------------+---------Total | 244 50 | 294 | 82.99 17.01 | 100.00 Fisher's exact = 1-sided Fisher's exact =

Nature

Population/ Sample

Depression is more prevalent among females.

0.004 0.003

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 28 of 65

. tab2 unemployed depressed, row exact | depressed unemployed | normal depressed | Total -----------+----------------------+---------other | 236 44 | 280 | 84.29 15.71 | 100.00 -----------+----------------------+---------unemployed | 8 6 | 14 | 57.14 42.86 | 100.00 -----------+----------------------+---------Total | 244 50 | 294 | 82.99 17.01 | 100.00 Fisher's exact = 1-sided Fisher's exact =

Depression is more prevalent among the unemployed.

0.018 0.018

Step 1. Fit single predictor models - Using Wald Z-score, retain predictors with significance levels < .25 or that are of a priori interest. . logit depressed age Logistic regression

Number of obs LR chi2(1)

= =

294 3.10

= Likelihood Ratio Statistic for current model (“full”) v intercept only model (“reduced”) Analogous to Overall F

Prob > chi2 Pseudo R2

Log likelihood = -132.51436 (-2) ln L = 265.50287

Wald Z

= =

0.0785 0.0115

Wald Z p-value (2 sided)

-----------------------------------------------------------------------------depressed | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------age | -.0156211 .0090668 -1.72 0.085 -.0333917 .0021495 _cons | -.9171994 .4043128 -2.27 0.023 -1.709638 -.1247608 -----------------------------------------------------------------------------. logit depressed alcohol Logistic regression

Number of obs LR chi2(1) Prob > chi2 Pseudo R2

= = = =

294 0.22 0.6387 0.0008

Log likelihood = -133.95203 (-2) ln L = 267.90406 -----------------------------------------------------------------------------depressed | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------alcohol | .1854829 .400363 0.46 0.643 -.5992142 .97018 _cons | -1.734601 .3615508 -4.80 0.000 -2.443228 -1.025975 ------------------------------------------------------------------------------

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

. logit depressed chronic Logistic regression

Page 29 of 65

Number of obs LR chi2(1) Prob > chi2 Pseudo R2

= = = =

294 3.12 0.0775 0.0116

Number of obs LR chi2(1) Prob > chi2 Pseudo R2

= = = =

294 8.73 0.0031 0.0325

Number of obs LR chi2(1) Prob > chi2 Pseudo R2

= = = =

294 8.72 0.0031 0.0325

Number of obs LR chi2(1) Prob > chi2 Pseudo R2

= = = =

294 5.46 0.0195 0.0204

Log likelihood = -132.50414 (-2) ln L = 265.00828 -----------------------------------------------------------------------------depressed | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------chronic | .5551455 .3182777 1.74 0.081 -.0686675 1.178958 _cons | -1.891843 .2461058 -7.69 0.000 -2.374201 -1.409484 -----------------------------------------------------------------------------. logit depressed female Logistic regression

Log likelihood = -129.69883 (-2) ln L = 259.39766 -----------------------------------------------------------------------------depressed | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------female | 1.03857 .3766882 2.76 0.006 .3002749 1.776866 _cons | -2.312535 .3315132 -6.98 0.000 -2.962289 -1.662782 -----------------------------------------------------------------------------. logit depressed income Logistic regression

Log likelihood = -129.70102 (-2) ln L = 259.40204 -----------------------------------------------------------------------------depressed | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------income | -.0358267 .0134794 -2.66 0.008 -.0622458 -.0094076 _cons | -.9375673 .2658415 -3.53 0.000 -1.458607 -.4165276 -----------------------------------------------------------------------------. logit depressed unemployed Logistic regression

Log likelihood = -131.33315 (-2) ln L = 262.6663 -----------------------------------------------------------------------------depressed | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------unemployed | 1.39196 .5644743 2.47 0.014 .2856108 2.498309 _cons | -1.679642 .1642089 -10.23 0.000 -2.001486 -1.357799 ------------------------------------------------------------------------------

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 30 of 65

Step 1 – Summary Predictor age alcohol chronic female income unemployed

Significance of Wald Z .085 .643 .081 .006 .008 .014

Remark Consider further – pvalue is < .25 Drop Consider further. pvalue is < .25 Consider further. pvalue is < .25 Consider further. pvalue is < .25 Consider further. pvalue is < .25.

Step 2 – Assess candidate predictors for evidence of multicollinearity Note – This assumes you have downloaded and installed collin.ado . collin age alcohol chronic female income unemployed

Collinearity occurs when the predictors are themselves inter-related If extreme, this is a problem for at least 2 reasons 1. Model is unstable 2. Model is uninterpretable Multicollinearity problem is suggested if VIF > 10 or Tolerance < .10 Here, things look reasonable.

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 31 of 65

Step 3. Fit multiple predictor model using step 1 predictors having crude significance < .25 . logit depressed age chronic female income unemployed Logistic regression

Number of obs LR chi2(5) Prob > chi2 Pseudo R2

Log likelihood = -121.04134

= = = =

294 26.04 0.0001 0.0971

(-2) ln L = 242.08268 Deviance df = 294-(5) = 289

-----------------------------------------------------------------------------depressed | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------age | -.0219383 .009494 -2.31 0.021 -.0405462 -.0033305 chronic | .594859 .3508664 1.70 0.090 -.0928265 1.282545 female | .8121316 .3968805 2.05 0.041 .0342602 1.590003 income | -.0320672 .0141399 -2.27 0.023 -.0597809 -.0043534 unemployed | 1.069739 .5989254 1.79 0.074 -.1041334 2.243611 _cons | -1.031844 .6121359 -1.69 0.092 -2.231608 .1679207 ------------------------------------------------------------------------------

Step 3 – Summary Predictor Adjusted Significance (Wald) age .021 chronic .090 female income unemployed

.041 .023 .074

Remark Retain – pvalue is < .10 For illustration purposes, let’s consider dropping this variable, despite pvalue < .10 (it’s close!)

Retain – pvalue is < .10 Retain – pvalue is < .10 Retain – pvalue is < .10.

Step 4. Fit the multivariable model containing predictors with adjusted significance levels < .10 from step 3. We will then compare the step 3 model with the step 4 model using a likelihood ratio test. . logit depressed age female income unemployed Logistic regression

Number of obs LR chi2(4) Prob > chi2 Pseudo R2

Log likelihood = -122.51896

= = = =

294 23.09 0.0001 0.0861

(-2) ln L = 245.03792 Deviance df = 294-(4) = 290

-----------------------------------------------------------------------------depressed | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------age | -.018802 .0091785 -2.05 0.041 -.0367917 -.0008124 female | .938952 .3887469 2.42 0.016 .177022 1.700882 income | -.0334314 .0141518 -2.36 0.018 -.0611684 -.0056944 unemployed | .9634566 .5921991 1.63 0.104 -.1972324 2.124146 _cons | -.8968284 .5978889 -1.50 0.134 -2.068669 .2750123 -----------------------------------------------------------------------------Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 32 of 65

By Hand: Likelihood ratio test comparing step 3 and step 4 models LR Test

= [ (-2) ln (L) REDUCED ] - [ (-2) ln (L) FULL ] = [ 245.04 ] – [ 242.08 ] = 2.96

LR Test df = Δ Deviance df = Δ # predictors in model = 290-289 = 1 p-value = Pr { Chi square with 1 degree of freedom > 2.96 } = .0853 This is not significant. Possibly, we can drop chronic

Stata: Likelihood ratio test comparing step 2 and step 3 models. . . . .

* REDUCED model using command quietly: to suppress output. Don’t forget the colon. quietly: logit depressed age female income unemployed * Save results using stata command estimates store NAME estimates store reduced

. . . .

* FULL model using command quietly: to suppress output. Don’t forget the colon. quietly: logit depressed age chronic female income unemployed * Save results using stata command estimates store NAME estimates store full

. * Obtain LR test using stata command lrtest . lrtest reduced full Likelihood-ratio test (Assumption: reduced nested in full)

Nature

Population/ Sample

LR chi2(1) = Prob > chi2 =

Observation/ Data

Relationships/ Modeling

2.96 0.0856

match!

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 33 of 65

Step 5. Investigate confounding. Tentatively, a “good” final model is the four predictor model with predictors: age, female, income, and unemployed. Here, we explore possible confounding of the four predictor model by the omitted variable chronic. Specifically, we assess chronic as a potential confounder using 2 criteria: __1. Likelihood Ratio test < .10 ( or .05 or threshold of choice). __2. Relative Change in estimated betas > 15% (or threshold of choice) using the following formula:

ˆ ˆ ˆ ⎛⎜ | β without confounder -β with confounder | ⎞⎟ x100 Δβ= ⎜ ⎟ βˆ with confounder ⎝ ⎠

Fit of tentative “good” final model (shown again…) . logit depressed age female income unemployed -----------------------------------------------------------------------------depressed | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------age | -.018802 .0091785 -2.05 0.041 -.0367917 -.0008124 female | .938952 .3887469 2.42 0.016 .177022 1.700882 income | -.0334314 .0141518 -2.36 0.018 -.0611684 -.0056944 unemployed | .9634566 .5921991 1.63 0.104 -.1972324 2.124146 _cons | -.8968284 .5978889 -1.50 0.134 -2.068669 .2750123 ------------------------------------------------------------------------------

Fit of enhanced model with chronic . logit depressed age chronic female income unemployed -----------------------------------------------------------------------------depressed | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------age | -.0219383 .009494 -2.31 0.021 -.0405462 -.0033305 chronic | .594859 .3508664 1.70 0.090 -.0928265 1.282545 female | .8121316 .3968805 2.05 0.041 .0342602 1.590003 income | -.0320672 .0141399 -2.27 0.023 -.0597809 -.0043534 unemployed | 1.069739 .5989254 1.79 0.074 -.1041334 2.243611 _cons | -1.031844 .6121359 -1.69 0.092 -2.231608 .1679207 ------------------------------------------------------------------------------

Nature

Population/ Sample

Observation/ Data

Relationships/ Modeling

Analysis/ Synthesis

PubHlth 640 - Spring 2015

5. Logistic Regression

Page 34 of 65

Looking for > 15% Change in Betas for Predictors in Model Potential confounding of age, female, income, unemployed By: chronic βˆ age (w/o chronic) = -.018802; βˆ age (w chronic) = -.0219383; Change = 14.30% βˆ female (w/o chronic) = .938952; βˆ female (w chronic) = .8121316; Change = 15.62% βˆ (w/o chronic) = -.0334314; βˆ (w chronic) = -.0320672; Change = 2.32% income

income

βˆ unemployed (w/o chronic) = .9634566; βˆ age (w chronic) = 1.069739; Change = 9.94%

The relative change in the beta for female is borderline at 15.6%. For parsimony, let’s drop chronic.

Step 6. Investigate effect modification. Are individuals who are both unemployed and with low income more likely to be depressed? For this illustration, we will create a new variable called low to capture individuals whose income is less than $10,000. Then we will create an interaction of low and unemployed. Tip – When assessing interaction, it is necessary to include the main effects of both of the variables contributing to the interaction. Thus, this model includes the main effects low and unemployed in addition to the interaction low_unemployed. . . . . . .

* Create new variable low generate low=income recode low (min/10=1) (10/max=0) label define lowf 0 "other" 1 "low (