Download presentation
Presentation is loading. Please wait.
Published byClyde Flynn Modified over 9 years ago
1
1 STA 617 – Chp11 Models for repeated data Analyzing Repeated Categorical Response Data Repeated categorical responses may come from repeated measurements over time on each individual or from a set of measurements that are related because they belong to the same group or cluster (e.g., measurements made on siblings from the same family, measurements made on a set of teeth from the same mouth). Observations within a cluster are not usually independent of each other, as the response from one child of a family, say, may influence the response from another child, because the two grew up together. Matched-pairs are the special case of each cluster having two members.
2
2 STA 617 – Chp11 Models for repeated data Using repeated measures within a cluster can be an efficient way to estimate the mean response at each measurement time without estimating between-cluster variability. Many times, one is interested in the marginal distribution of the response at each measurement time, and not substantially interested in the correlation between responses across times. Estimation methods for marginal modeling include maximum likelihood estimation and generalized estimating equations (GEE). Maximum likelihood estimation is difficult because the likelihood is written in terms of the I T multinomial joint probabilities for T responses with I categories each, but the model applies to the marginal probabilities. Lang and Agresti give a method for maximum likelihood fitting of marginal models in Section 11.2.5. Modeling a repeated multinomial response or repeated ordinal response is handled in the same way.
3
3 STA 617 – Chp11 Models for repeated data Topics In Section 11.1 we compare marginal distributions in T- way tables. The remaining sections extend models to include explanatory variables. In Section 11.2 we use ML methods for fitting marginal models. In Section 11.3 we use generalized estimating equations (GEE), a multivariate version of quasi- likelihood that is computationally simpler than ML. Section 11.4 covers technical details about the GEE approach. In the final section we introduce a transitional approach that models observations in terms of previous outcomes.
4
4 STA 617 – Chp11 Models for repeated data 11.1 COMPARING MARGINAL DISTRIBUTIONS: MULTIPLE RESPONSES Please review 10.1-10.3. Example: in treating a chronic condition with some treatment, the primary goal might be to study whether the probability of success increases over the T weeks of a treatment period. The T success probabilities refer to the T first-order marginal distributions We want to compare marginal distributions.
5
5 STA 617 – Chp11 Models for repeated data 11.1.1 Binary Marginal Models and Marginal Homogeneity T binary responses Marginal logit model with All possible outcomes where Let the joint distribution of is Mult (n, ( 1, 2, …, 2^T ))
6
6 STA 617 – Chp11 Models for repeated data Marginal homogeneity Likelihood The likelihood-ratio test of marginal homogeneity where sample proportions and is maximized likelihood estimate assuming marginal homogeneity. asymptotic null chi-squared distribution with DF=T-1
7
7 STA 617 – Chp11 Models for repeated data 11.1.2 Crossover Drug Comparison Example each subject used each of three drugs for treatment of a chronic condition at three times. The response measured the reaction as favorable or unfavorable. (binary) assume that the drugs have no carryover effects and that the severity of the condition remained stable for each subject throughout the experiment.
8
8 STA 617 – Chp11 Models for repeated data Test marginal homogeneity Sample proportions favorable (n=46) [(6+2+16+4)/46=0.61, 28/46=0.61, 16/46=0.35] for drug A, B, C Clearly, from the sample proportion, A and B are similar, and better than C The likelihood-ratio test statistic is 5.95 (DF=2). P-value=0.05.
9
9 STA 617 – Chp11 Models for repeated data SAS
10
10 STA 617 – Chp11 Models for repeated data simultaneous confidence intervals The confidence interval for the true difference is (0.00133, 0.520) between B and C
11
11 STA 617 – Chp11 Models for repeated data CATMOD Suppose the dependent variable A has three levels and is the only response-effect in the MODEL statement.
12
12 STA 617 – Chp11 Models for repeated data Design Matrix p_A=alpha+beta1+beta2 P_B=alpha+beta1 P_C=alpha Alpha=intercept Beta1=p_B-p_C Beta2=p_A-p_B
13
13 STA 617 – Chp11 Models for repeated data Design Matrix p_A=parameter1 P_B=parameter2 P_C=parameter3 Analysis of Weighted Least Squares Estimates EffectParameterEstimateStandard Error Chi- Square Pr > ChiSq Model10.60870.072071.56<.0001 20.60870.072071.56<.0001 30.34780.070224.53<.0001
14
14 STA 617 – Chp11 Models for repeated data 11.1.3 Modeling Margins of a Multicategory Response Saturated model marginal homogeneity Test
15
15 STA 617 – Chp11 Models for repeated data Ordinal response marginal homogeneity Test Model fitting 11.2.5
16
16 STA 617 – Chp11 Models for repeated data 11.1.4 Wald and Generalized CMH Score Tests of Marginal Homogeneity Similar with paired data in Chapter 10 SAS
17
17 STA 617 – Chp11 Models for repeated data 11.2 MARGINAL MODELING: MAXIMUM LIKELIHOOD APPROACH compared marginal distributions, but accounting for explanatory variables.
18
18 STA 617 – Chp11 Models for repeated data Longitudinal Mental Depression Example comparing a new drug with a standard drug Outcome: mental depression (normal, abnormal) Stratified randomization by severity of depression (was mild or severe). Four arms n=80, 70, 100, 90 Follow up 1 week, 2 weeks, and 4 weeks
19
19 STA 617 – Chp11 Models for repeated data explanatory variables: treatment type and severity of initial diagnosis T=3 12 marginal distributions result from three repeated observations for each of the four groups. Let s denote the severity of the initial diagnosis, with s=1 for severe and s=0 for mild. Let d denote the drug, with d=1 for new and d=0 for standard. Let t denote the time of measurement. Use score (0, 1, 2), the logs to base 2 of the week (1, 2, 4).
20
20 STA 617 – Chp11 Models for repeated data Descriptive statistics (sample proportions) the sample proportion of normal responses after week 1 for subjects with mild initial diagnosis using the standard drug was
21
21 STA 617 – Chp11 Models for repeated data data depress; input case diagnose treat time outcome ; * outcome=1 is normal; datalines; 1 0 0 0 1 1 0 0 1 1 1 0 0 2 1 2 0 0 0 1 2 0 0 1 1 2 0 0 2 1 3 0 0 0 1 3 0 0 1 1 3 0 0 2 1 4 0 0 0 1 4 0 0 1 1 4 0 0 2 1 5 0 0 0 1 5 0 0 1 1 5 0 0 2 1 6 0 0 0 1 6 0 0 1 1 6 0 0 2 1 7 0 0 0 1 7 0 0 1 1 7 0 0 2 1 8 0 0 0 1 8 0 0 1 1 8 0 0 2 1 9 0 0 0 1 9 0 0 1 1 9 0 0 2 1 10 0 0 0 1 10 0 0 1 1 10 0 0 2 1 11 0 0 0 1 11 0 0 1 1 11 0 0 2 1 12 0 0 0 1 12 0 0 1 1 12 0 0 2 1 13 0 0 0 1 13 0 0 1 1 13 0 0 2 1 14 0 0 0 1 14 0 0 1 1 14 0 0 2 1 15 0 0 0 1 15 0 0 1 1 15 0 0 2 1 16 0 0 0 1 16 0 0 1 1 16 0 0 2 1 17 0 0 0 1 17 0 0 1 1 17 0 0 2 0 18 0 0 0 1 18 0 0 1 1 18 0 0 2 0 19 0 0 0 1 19 0 0 1 1 19 0 0 2 0 20 0 0 0 1 20 0 0 1 1 20 0 0 2 0 21 0 0 0 1 21 0 0 1 1 21 0 0 2 0 22 0 0 0 1 22 0 0 1 1 22 0 0 2 0 23 0 0 0 1 23 0 0 1 1 23 0 0 2 0 24 0 0 0 1 24 0 0 1 1 24 0 0 2 0 25 0 0 0 1 25 0 0 1 1 25 0 0 2 0 26 0 0 0 1 26 0 0 1 1 26 0 0 2 0 27 0 0 0 1 27 0 0 1 1 27 0 0 2 0 28 0 0 0 1 28 0 0 1 1 28 0 0 2 0 29 0 0 0 1 29 0 0 1 1 29 0 0 2 0 30 0 0 0 1 30 0 0 1 0 30 0 0 2 1 31 0 0 0 1 31 0 0 1 0 31 0 0 2 1 32 0 0 0 1 32 0 0 1 0 32 0 0 2 1 33 0 0 0 1 33 0 0 1 0 33 0 0 2 1 34 0 0 0 1 34 0 0 1 0 34 0 0 2 1 35 0 0 0 1 35 0 0 1 0 35 0 0 2 1 36 0 0 0 1 36 0 0 1 0 36 0 0 2 1 37 0 0 0 1 37 0 0 1 0 37 0 0 2 1 38 0 0 0 1 38 0 0 1 0 38 0 0 2 1 39 0 0 0 1 39 0 0 1 0 39 0 0 2 0 40 0 0 0 1 40 0 0 1 0 40 0 0 2 0 41 0 0 0 1 41 0 0 1 0 41 0 0 2 0 42 0 0 0 0 42 0 0 1 1 42 0 0 2 1 43 0 0 0 0 43 0 0 1 1 43 0 0 2 1 44 0 0 0 0 44 0 0 1 1 44 0 0 2 1 45 0 0 0 0 45 0 0 1 1 45 0 0 2 1 46 0 0 0 0 46 0 0 1 1 46 0 0 2 1 47 0 0 0 0 47 0 0 1 1 47 0 0 2 1 48 0 0 0 0 48 0 0 1 1 48 0 0 2 1 49 0 0 0 0 49 0 0 1 1 49 0 0 2 1 50 0 0 0 0 50 0 0 1 1 50 0 0 2 1 51 0 0 0 0 51 0 0 1 1 51 0 0 2 1 52 0 0 0 0 52 0 0 1 1 52 0 0 2 1 53 0 0 0 0 53 0 0 1 1 53 0 0 2 1 54 0 0 0 0 54 0 0 1 1 54 0 0 2 1 55 0 0 0 0 55 0 0 1 1 55 0 0 2 1 56 0 0 0 0 56 0 0 1 1 56 0 0 2 0 57 0 0 0 0 57 0 0 1 1 57 0 0 2 0 58 0 0 0 0 58 0 0 1 1 58 0 0 2 0 59 0 0 0 0 59 0 0 1 1 59 0 0 2 0 60 0 0 0 0 60 0 0 1 0 60 0 0 2 1 61 0 0 0 0 61 0 0 1 0 61 0 0 2 1 62 0 0 0 0 62 0 0 1 0 62 0 0 2 1 63 0 0 0 0 63 0 0 1 0 63 0 0 2 1 64 0 0 0 0 64 0 0 1 0 64 0 0 2 1 65 0 0 0 0 65 0 0 1 0 65 0 0 2 1 66 0 0 0 0 66 0 0 1 0 66 0 0 2 1 67 0 0 0 0 67 0 0 1 0 67 0 0 2 1 68 0 0 0 0 68 0 0 1 0 68 0 0 2 1 69 0 0 0 0 69 0 0 1 0 69 0 0 2 1 70 0 0 0 0 70 0 0 1 0 70 0 0 2 1 71 0 0 0 0 71 0 0 1 0 71 0 0 2 1 72 0 0 0 0 72 0 0 1 0 72 0 0 2 1 73 0 0 0 0 73 0 0 1 0 73 0 0 2 1 74 0 0 0 0 74 0 0 1 0 74 0 0 2 1 75 0 0 0 0 75 0 0 1 0 75 0 0 2 0 336 0 0 0 0 336 0 0 1 0 336 0 0 2 0 337 0 0 0 0 337 0 0 1 0 337 0 0 2 0 338 0 0 0 0 338 0 0 1 0 338 0 0 2 0 339 0 0 0 0 339 0 0 1 0 339 0 0 2 0 340 0 0 0 0 340 0 0 1 0 340 0 0 2 0 76 0 1 0 1 76 0 1 1 1 76 0 1 2 1 77 0 1 0 1 77 0 1 1 1 77 0 1 2 1 78 0 1 0 1 78 0 1 1 1 78 0 1 2 1 79 0 1 0 1 79 0 1 1 1 79 0 1 2 1 80 0 1 0 1 80 0 1 1 1 80 0 1 2 1 81 0 1 0 1 81 0 1 1 1 81 0 1 2 1 82 0 1 0 1 82 0 1 1 1 82 0 1 2 1 83 0 1 0 1 83 0 1 1 1 83 0 1 2 1 84 0 1 0 1 84 0 1 1 1 84 0 1 2 1 85 0 1 0 1 85 0 1 1 1 85 0 1 2 1 86 0 1 0 1 86 0 1 1 1 86 0 1 2 1 87 0 1 0 1 87 0 1 1 1 87 0 1 2 1 88 0 1 0 1 88 0 1 1 1 88 0 1 2 1 89 0 1 0 1 89 0 1 1 1 89 0 1 2 1 90 0 1 0 1 90 0 1 1 1 90 0 1 2 1 91 0 1 0 1 91 0 1 1 1 91 0 1 2 1 92 0 1 0 1 92 0 1 1 1 92 0 1 2 1 93 0 1 0 1 93 0 1 1 1 93 0 1 2 1 94 0 1 0 1 94 0 1 1 1 94 0 1 2 1 95 0 1 0 1 95 0 1 1 1 95 0 1 2 1 96 0 1 0 1 96 0 1 1 1 96 0 1 2 1 97 0 1 0 1 97 0 1 1 1 97 0 1 2 1 98 0 1 0 1 98 0 1 1 1 98 0 1 2 1 99 0 1 0 1 99 0 1 1 1 99 0 1 2 1 100 0 1 0 1 100 0 1 1 1 100 0 1 2 1 101 0 1 0 1 101 0 1 1 1 101 0 1 2 1 102 0 1 0 1 102 0 1 1 1 102 0 1 2 1 103 0 1 0 1 103 0 1 1 1 103 0 1 2 1 104 0 1 0 1 104 0 1 1 1 104 0 1 2 1 105 0 1 0 1 105 0 1 1 1 105 0 1 2 1 106 0 1 0 1 106 0 1 1 1 106 0 1 2 1 107 0 1 0 1 107 0 1 1 0 107 0 1 2 1 108 0 1 0 1 108 0 1 1 0 108 0 1 2 1 109 0 1 0 1 109 0 1 1 0 109 0 1 2 1 110 0 1 0 1 110 0 1 1 0 110 0 1 2 1 111 0 1 0 1 111 0 1 1 0 111 0 1 2 1 112 0 1 0 1 112 0 1 1 0 112 0 1 2 1 113 0 1 0 0 113 0 1 1 1 113 0 1 2 1 114 0 1 0 0 114 0 1 1 1 114 0 1 2 1 115 0 1 0 0 115 0 1 1 1 115 0 1 2 1 116 0 1 0 0 116 0 1 1 1 116 0 1 2 1 117 0 1 0 0 117 0 1 1 1 117 0 1 2 1 118 0 1 0 0 118 0 1 1 1 118 0 1 2 1 119 0 1 0 0 119 0 1 1 1 119 0 1 2 1 120 0 1 0 0 120 0 1 1 1 120 0 1 2 1 121 0 1 0 0 121 0 1 1 1 121 0 1 2 1 122 0 1 0 0 122 0 1 1 1 122 0 1 2 1 123 0 1 0 0 123 0 1 1 1 123 0 1 2 1 124 0 1 0 0 124 0 1 1 1 124 0 1 2 1 125 0 1 0 0 125 0 1 1 1 125 0 1 2 1 126 0 1 0 0 126 0 1 1 1 126 0 1 2 1 127 0 1 0 0 127 0 1 1 1 127 0 1 2 1 128 0 1 0 0 128 0 1 1 1 128 0 1 2 1 129 0 1 0 0 129 0 1 1 1 129 0 1 2 1 130 0 1 0 0 130 0 1 1 1 130 0 1 2 1 131 0 1 0 0 131 0 1 1 1 131 0 1 2 1 132 0 1 0 0 132 0 1 1 1 132 0 1 2 1 133 0 1 0 0 133 0 1 1 1 133 0 1 2 1 134 0 1 0 0 134 0 1 1 1 134 0 1 2 1 135 0 1 0 0 135 0 1 1 1 135 0 1 2 0 136 0 1 0 0 136 0 1 1 1 136 0 1 2 0 137 0 1 0 0 137 0 1 1 0 137 0 1 2 1 138 0 1 0 0 138 0 1 1 0 138 0 1 2 1 139 0 1 0 0 139 0 1 1 0 139 0 1 2 1 140 0 1 0 0 140 0 1 1 0 140 0 1 2 1 141 0 1 0 0 141 0 1 1 0 141 0 1 2 1 142 0 1 0 0 142 0 1 1 0 142 0 1 2 1 143 0 1 0 0 143 0 1 1 0 143 0 1 2 1 144 0 1 0 0 144 0 1 1 0 144 0 1 2 1 145 0 1 0 0 145 0 1 1 0 145 0 1 2 1 146 1 0 0 1 146 1 0 1 1 146 1 0 2 1 147 1 0 0 1 147 1 0 1 1 147 1 0 2 1 148 1 0 0 1 148 1 0 1 1 148 1 0 2 0 149 1 0 0 1 149 1 0 1 1 149 1 0 2 0 150 1 0 0 1 150 1 0 1 0 150 1 0 2 1 151 1 0 0 1 151 1 0 1 0 151 1 0 2 1 152 1 0 0 1 152 1 0 1 0 152 1 0 2 1 153 1 0 0 1 153 1 0 1 0 153 1 0 2 1 154 1 0 0 1 154 1 0 1 0 154 1 0 2 1 155 1 0 0 1 155 1 0 1 0 155 1 0 2 1 156 1 0 0 1 156 1 0 1 0 156 1 0 2 1 157 1 0 0 1 157 1 0 1 0 157 1 0 2 1 158 1 0 0 1 158 1 0 1 0 158 1 0 2 0 159 1 0 0 1 159 1 0 1 0 159 1 0 2 0 160 1 0 0 1 160 1 0 1 0 160 1 0 2 0 161 1 0 0 1 161 1 0 1 0 161 1 0 2 0 162 1 0 0 1 162 1 0 1 0 162 1 0 2 0 163 1 0 0 1 163 1 0 1 0 163 1 0 2 0 164 1 0 0 1 164 1 0 1 0 164 1 0 2 0 165 1 0 0 1 165 1 0 1 0 165 1 0 2 0 166 1 0 0 1 166 1 0 1 0 166 1 0 2 0 167 1 0 0 0 167 1 0 1 1 167 1 0 2 1 168 1 0 0 0 168 1 0 1 1 168 1 0 2 1 169 1 0 0 0 169 1 0 1 1 169 1 0 2 1 170 1 0 0 0 170 1 0 1 1 170 1 0 2 1 171 1 0 0 0 171 1 0 1 1 171 1 0 2 1 172 1 0 0 0 172 1 0 1 1 172 1 0 2 1 173 1 0 0 0 173 1 0 1 1 173 1 0 2 1 174 1 0 0 0 174 1 0 1 1 174 1 0 2 1 175 1 0 0 0 175 1 0 1 1 175 1 0 2 1 176 1 0 0 0 176 1 0 1 1 176 1 0 2 0 177 1 0 0 0 177 1 0 1 1 177 1 0 2 0 178 1 0 0 0 178 1 0 1 1 178 1 0 2 0 179 1 0 0 0 179 1 0 1 1 179 1 0 2 0 180 1 0 0 0 180 1 0 1 1 180 1 0 2 0 181 1 0 0 0 181 1 0 1 1 181 1 0 2 0 182 1 0 0 0 182 1 0 1 1 182 1 0 2 0 183 1 0 0 0 183 1 0 1 1 183 1 0 2 0 184 1 0 0 0 184 1 0 1 1 184 1 0 2 0 185 1 0 0 0 185 1 0 1 1 185 1 0 2 0 186 1 0 0 0 186 1 0 1 1 186 1 0 2 0 187 1 0 0 0 187 1 0 1 1 187 1 0 2 0 188 1 0 0 0 188 1 0 1 1 188 1 0 2 0 189 1 0 0 0 189 1 0 1 1 189 1 0 2 0 190 1 0 0 0 190 1 0 1 1 190 1 0 2 0 191 1 0 0 0 191 1 0 1 0 191 1 0 2 1 192 1 0 0 0 192 1 0 1 0 192 1 0 2 1 193 1 0 0 0 193 1 0 1 0 193 1 0 2 1 194 1 0 0 0 194 1 0 1 0 194 1 0 2 1 195 1 0 0 0 195 1 0 1 0 195 1 0 2 1 196 1 0 0 0 196 1 0 1 0 196 1 0 2 1 197 1 0 0 0 197 1 0 1 0 197 1 0 2 1 198 1 0 0 0 198 1 0 1 0 198 1 0 2 1 199 1 0 0 0 199 1 0 1 0 199 1 0 2 1 200 1 0 0 0 200 1 0 1 0 200 1 0 2 1 201 1 0 0 0 201 1 0 1 0 201 1 0 2 1 202 1 0 0 0 202 1 0 1 0 202 1 0 2 1 203 1 0 0 0 203 1 0 1 0 203 1 0 2 1 204 1 0 0 0 204 1 0 1 0 204 1 0 2 1 205 1 0 0 0 205 1 0 1 0 205 1 0 2 1 206 1 0 0 0 206 1 0 1 0 206 1 0 2 1 207 1 0 0 0 207 1 0 1 0 207 1 0 2 1 208 1 0 0 0 208 1 0 1 0 208 1 0 2 1 209 1 0 0 0 209 1 0 1 0 209 1 0 2 1 210 1 0 0 0 210 1 0 1 0 210 1 0 2 1 211 1 0 0 0 211 1 0 1 0 211 1 0 2 1 212 1 0 0 0 212 1 0 1 0 212 1 0 2 1 213 1 0 0 0 213 1 0 1 0 213 1 0 2 1 214 1 0 0 0 214 1 0 1 0 214 1 0 2 1 215 1 0 0 0 215 1 0 1 0 215 1 0 2 1 216 1 0 0 0 216 1 0 1 0 216 1 0 2 1 217 1 0 0 0 217 1 0 1 0 217 1 0 2 1 218 1 0 0 0 218 1 0 1 0 218 1 0 2 0 219 1 0 0 0 219 1 0 1 0 219 1 0 2 0 220 1 0 0 0 220 1 0 1 0 220 1 0 2 0 221 1 0 0 0 221 1 0 1 0 221 1 0 2 0 222 1 0 0 0 222 1 0 1 0 222 1 0 2 0 223 1 0 0 0 223 1 0 1 0 223 1 0 2 0 224 1 0 0 0 224 1 0 1 0 224 1 0 2 0 225 1 0 0 0 225 1 0 1 0 225 1 0 2 0 226 1 0 0 0 226 1 0 1 0 226 1 0 2 0 227 1 0 0 0 227 1 0 1 0 227 1 0 2 0 228 1 0 0 0 228 1 0 1 0 228 1 0 2 0 229 1 0 0 0 229 1 0 1 0 229 1 0 2 0 230 1 0 0 0 230 1 0 1 0 230 1 0 2 0 231 1 0 0 0 231 1 0 1 0 231 1 0 2 0 232 1 0 0 0 232 1 0 1 0 232 1 0 2 0 233 1 0 0 0 233 1 0 1 0 233 1 0 2 0 234 1 0 0 0 234 1 0 1 0 234 1 0 2 0 235 1 0 0 0 235 1 0 1 0 235 1 0 2 0 236 1 0 0 0 236 1 0 1 0 236 1 0 2 0 237 1 0 0 0 237 1 0 1 0 237 1 0 2 0 238 1 0 0 0 238 1 0 1 0 238 1 0 2 0 239 1 0 0 0 239 1 0 1 0 239 1 0 2 0 240 1 0 0 0 240 1 0 1 0 240 1 0 2 0 241 1 0 0 0 241 1 0 1 0 241 1 0 2 0 242 1 0 0 0 242 1 0 1 0 242 1 0 2 0 243 1 0 0 0 243 1 0 1 0 243 1 0 2 0 244 1 0 0 0 244 1 0 1 0 244 1 0 2 0 245 1 0 0 0 245 1 0 1 0 245 1 0 2 0 246 1 1 0 1 246 1 1 1 1 246 1 1 2 1 247 1 1 0 1 247 1 1 1 1 247 1 1 2 1 248 1 1 0 1 248 1 1 1 1 248 1 1 2 1 249 1 1 0 1 249 1 1 1 1 249 1 1 2 1 250 1 1 0 1 250 1 1 1 1 250 1 1 2 1 251 1 1 0 1 251 1 1 1 1 251 1 1 2 1 252 1 1 0 1 252 1 1 1 1 252 1 1 2 1 253 1 1 0 1 253 1 1 1 1 253 1 1 2 0 254 1 1 0 1 254 1 1 1 1 254 1 1 2 0 255 1 1 0 1 255 1 1 1 0 255 1 1 2 1 256 1 1 0 1 256 1 1 1 0 256 1 1 2 1 257 1 1 0 1 257 1 1 1 0 257 1 1 2 1 258 1 1 0 1 258 1 1 1 0 258 1 1 2 1 259 1 1 0 1 259 1 1 1 0 259 1 1 2 1 260 1 1 0 1 260 1 1 1 0 260 1 1 2 0 261 1 1 0 1 261 1 1 1 0 261 1 1 2 0 262 1 1 0 0 262 1 1 1 1 262 1 1 2 1 263 1 1 0 0 263 1 1 1 1 263 1 1 2 1 264 1 1 0 0 264 1 1 1 1 264 1 1 2 1 265 1 1 0 0 265 1 1 1 1 265 1 1 2 1 266 1 1 0 0 266 1 1 1 1 266 1 1 2 1 267 1 1 0 0 267 1 1 1 1 267 1 1 2 1 268 1 1 0 0 268 1 1 1 1 268 1 1 2 1 269 1 1 0 0 269 1 1 1 1 269 1 1 2 1 270 1 1 0 0 270 1 1 1 1 270 1 1 2 1 271 1 1 0 0 271 1 1 1 1 271 1 1 2 1 272 1 1 0 0 272 1 1 1 1 272 1 1 2 1 273 1 1 0 0 273 1 1 1 1 273 1 1 2 1 274 1 1 0 0 274 1 1 1 1 274 1 1 2 1 275 1 1 0 0 275 1 1 1 1 275 1 1 2 1 276 1 1 0 0 276 1 1 1 1 276 1 1 2 1 277 1 1 0 0 277 1 1 1 1 277 1 1 2 1 278 1 1 0 0 278 1 1 1 1 278 1 1 2 1 279 1 1 0 0 279 1 1 1 1 279 1 1 2 1 280 1 1 0 0 280 1 1 1 1 280 1 1 2 1 281 1 1 0 0 281 1 1 1 1 281 1 1 2 1 282 1 1 0 0 282 1 1 1 1 282 1 1 2 1 283 1 1 0 0 283 1 1 1 1 283 1 1 2 1 284 1 1 0 0 284 1 1 1 1 284 1 1 2 1 285 1 1 0 0 285 1 1 1 1 285 1 1 2 1 286 1 1 0 0 286 1 1 1 1 286 1 1 2 1 287 1 1 0 0 287 1 1 1 1 287 1 1 2 1 288 1 1 0 0 288 1 1 1 1 288 1 1 2 1 289 1 1 0 0 289 1 1 1 1 289 1 1 2 1 290 1 1 0 0 290 1 1 1 1 290 1 1 2 1 291 1 1 0 0 291 1 1 1 1 291 1 1 2 1 292 1 1 0 0 292 1 1 1 1 292 1 1 2 1 293 1 1 0 0 293 1 1 1 1 293 1 1 2 0 294 1 1 0 0 294 1 1 1 1 294 1 1 2 0 295 1 1 0 0 295 1 1 1 1 295 1 1 2 0 296 1 1 0 0 296 1 1 1 1 296 1 1 2 0 297 1 1 0 0 297 1 1 1 1 297 1 1 2 0 298 1 1 0 0 298 1 1 1 0 298 1 1 2 1 299 1 1 0 0 299 1 1 1 0 299 1 1 2 1 300 1 1 0 0 300 1 1 1 0 300 1 1 2 1 301 1 1 0 0 301 1 1 1 0 301 1 1 2 1 302 1 1 0 0 302 1 1 1 0 302 1 1 2 1 303 1 1 0 0 303 1 1 1 0 303 1 1 2 1 304 1 1 0 0 304 1 1 1 0 304 1 1 2 1 305 1 1 0 0 305 1 1 1 0 305 1 1 2 1 306 1 1 0 0 306 1 1 1 0 306 1 1 2 1 307 1 1 0 0 307 1 1 1 0 307 1 1 2 1 308 1 1 0 0 308 1 1 1 0 308 1 1 2 1 309 1 1 0 0 309 1 1 1 0 309 1 1 2 1 310 1 1 0 0 310 1 1 1 0 310 1 1 2 1 311 1 1 0 0 311 1 1 1 0 311 1 1 2 1 312 1 1 0 0 312 1 1 1 0 312 1 1 2 1 313 1 1 0 0 313 1 1 1 0 313 1 1 2 1 314 1 1 0 0 314 1 1 1 0 314 1 1 2 1 315 1 1 0 0 315 1 1 1 0 315 1 1 2 1 316 1 1 0 0 316 1 1 1 0 316 1 1 2 1 317 1 1 0 0 317 1 1 1 0 317 1 1 2 1 318 1 1 0 0 318 1 1 1 0 318 1 1 2 1 319 1 1 0 0 319 1 1 1 0 319 1 1 2 1 320 1 1 0 0 320 1 1 1 0 320 1 1 2 1 321 1 1 0 0 321 1 1 1 0 321 1 1 2 1 322 1 1 0 0 322 1 1 1 0 322 1 1 2 1 323 1 1 0 0 323 1 1 1 0 323 1 1 2 1 324 1 1 0 0 324 1 1 1 0 324 1 1 2 1 325 1 1 0 0 325 1 1 1 0 325 1 1 2 1 326 1 1 0 0 326 1 1 1 0 326 1 1 2 1 327 1 1 0 0 327 1 1 1 0 327 1 1 2 1 328 1 1 0 0 328 1 1 1 0 328 1 1 2 1 329 1 1 0 0 329 1 1 1 0 329 1 1 2 1 330 1 1 0 0 330 1 1 1 0 330 1 1 2 0 331 1 1 0 0 331 1 1 1 0 331 1 1 2 0 332 1 1 0 0 332 1 1 1 0 332 1 1 2 0 333 1 1 0 0 333 1 1 1 0 333 1 1 2 0 334 1 1 0 0 334 1 1 1 0 334 1 1 2 0 335 1 1 0 0 335 1 1 1 0 335 1 1 2 0 ; proc sort; by diagnose treat time; proc means n mean std; class diagnose treat time; var outcome; run;
22
22 STA 617 – Chp11 Models for repeated data The sample proportion of normal responses increased over time for each group; increased at a faster rate for the new drug than the standard, for each fixed initial diagnosis; and was higher for the mild than the severe initial diagnosis, for each treatment at each occasion. The company would hope to show that patients have a significantly higher rate of improvement with the new drug.
23
23 STA 617 – Chp11 Models for repeated data Modeling The marginal logit model 1 (main effects model) Time (t) is continuous The natural sampling assumption is multinomial for the eight cells in the 2 3 cross-classification of the three responses A check of model fit compares the 32 cell counts in Table 11.2 to their ML fitted values. Since the model describes 12 marginal logits using four parameters, residual df=8. The deviance G 2 =34.6. Lack of fit, since model assumes a common rate of improvement (should be higher for new drug)
24
24 STA 617 – Chp11 Models for repeated data Model 2
25
25 STA 617 – Chp11 Models for repeated data For each drug-time combination, the estimated odds of a normal response when the initial diagnosis was severe equal exp(-1.29)=0.27 times the estimated odds when the initial diagnosis was mild. The estimate indicates an insignificant difference between the drugs after 1 week. At time t, the estimated odds of normal response with the new drug are exp(-0.06+1.01 t) times the estimated odds for the standard drug, for each initial diagnosis level. Conclusion: severity of initial diagnosis, drug treatment, and time all have substantial effects on the probability of a normal response.
26
26 STA 617 – Chp11 Models for repeated data 11.2.2 Modeling a Repeated Multinomial Response At observation t, the marginal response distribution has I-1 logits. nominal responses, baseline-category logit models describe the odds of each outcome relative to a baseline. For ordinal responses, one might use cumulative logit models. checking for interaction is crucial.
27
27 STA 617 – Chp11 Models for repeated data 11.2.3 Insomnia Example randomized, double-blind clinical trial comparing an active hypnotic drug with a placebo in patients who have insomnia problems. response is the patient’s reported time in minutes to fall asleep after going to bed.
28
28 STA 617 – Chp11 Models for repeated data Proportional odds model Sample marginal distributions proc sort; by treat time; proc freq; tables treat*time*outcome /nocol NOFREQ NOPERCENT; run;
29
29 STA 617 – Chp11 Models for repeated data ML model fitting G2=8.0 (df=6) shows evidence of interaction At the initial observation, the estimated odds that time to falling asleep for the active treatment is below any fixed level equal exp(0.046)=1.04 times the estimated odds for the placebo treatment; at the follow-up observation, the effect is exp(0.046+0.662)=2.03. In other words, initially the two groups had similar distributions, but at the follow-up those with the active treatment tended to fall asleep more quickly. Follow-up with placebo or treatment, both tended to fall sleep more quickly (exp(1.07)=2.9)
30
30 STA 617 – Chp11 Models for repeated data 11.2.4 Comparisons That Control for Initial Response Model assumption: the marginal distributions for initial response are identical for the treatment groups. This is true if random assignment of subjects to the groups (one of the principles in experimental design: randomization, other two: replication, blocking) If the initial marginal distributions are not identical, however, the difference between follow-up and initial marginal distributions may differ between treatment groups, even though their conditional distributions for follow-up response are identical. In such cases, although marginal models can be useful, they may not tell the entire story. It may be more informative to construct models that compare the follow-up responses while controlling for the initial response.
31
31 STA 617 – Chp11 Models for repeated data transitional model Let Y 2 denote the follow-up response, for treatment x with initial response y 1.
32
32 STA 617 – Chp11 Models for repeated data 11.2.5 ML Fitting of Marginal Logit Models* For T observations on an I-category response, at each setting of predictors the likelihood refers to I T multinomial joint probabilities, but the model applies to T sets of marginal multinomial parameters The marginal multinomial variates are not independent. Marginal logit models have the generalized loglinear model form where denote the complete set of multinomial joint probabilities for all settings of predictors.
33
33 STA 617 – Chp11 Models for repeated data Example, model (11.1) the model of marginal homogeneity (T=2)
34
34 STA 617 – Chp11 Models for repeated data likelihood The likelihood function for a marginal logit model is the product of the multinomial mass functions from the various predictor settings. Usually, no continuous predictor is allowed if U denote a full column rank matrix such that the space spanned by the columns of U is the orthogonal complement of the space spanned by the columns of X. maximizing the likelihood incorporates these model constraints as well as identifiability constraints
35
35 STA 617 – Chp11 Models for repeated data ML Joseph Lang ( jblang@stat.uiowa.edu) has R and S-Plus functions for ML fitting of marginal models through the generalized loglinear model (11.8), using the constraint approach with Lagrange multipliers. http://www.stat.uiowa.edu/~jblang/mph.fitting/mph.fit.documentation.2.0.htm http://www.stat.uiowa.edu/~jblang/mph.fitting/mph.fit.documentation.2.0.htm The program MAREG (Kastner et al. 1997) provides GEE fitting and ML fitting of marginal models with the Fitzmaurice and Laird (1993) approach, allowing multicategory responses.
36
36 STA 617 – Chp11 Models for repeated data Generalized Estimating Equation (GEE)
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.