Does anemia contribute to end-organ dysfunction in ICU patients Statistical Analysis
|
|
- Donna Walker
- 5 years ago
- Views:
Transcription
1 Does anemia contribute to end-organ dysfunction in ICU patients Statistical Analysis Xue Han, MPH and Matt Shotwell, PhD Department of Biostatistics Vanderbilt University School of Medicine November 7, 2014 Contents 1 Analysis 1: Transitional Model to model next day s effect Outcome1: Daily Delirium Descriptive Statistics Multinomial logistic regression with Normal as the reference level for outcome variable Delirium Outcome2: Daily troponin (greater than 0.1) as an index of cardiac dysfunction Descriptive Statistics Outcome3: Daily renal SOFA score Descriptive Statistics Logistic regression, Daily renal SOFA 2(Y/N) Proportional odds logistic regression regression, Ordinal response (0-4) Outcome4: Daily resp SOFA score ( 2) Descriptive Statistics Logistic regression, Daily RESP SOFA 2(Y/N) Proportional odds logistic regression, Ordinal response (0-4) Proportional odds logistic regression, Ordinal response (0-4),without three way interaction Proportional odds logistic regression, Ordinal response (0-4),Sensitivity Analysis Proportional odds logistic regression, Ordinal response (0-4),Sensitivity Analysis2, without three way interaction Proportional odds logistic regression, Ordinal response (0-4),Marginal Model Proportional odds logistic regression, Ordinal response (0-4),Marginal Model with hgb.icu and resp.sofa interaction Tomorrow resp.sofa Tomorrow resp.sofa Tomorrow resp.sofa Outcome5: Daily Ventilator (Y/N) as a measure of success of extubation Descriptive Statistics Proportional odds logistic regression, Ordinal response (0-4),Marginal Model Logistic regression (Daily Ventilation (Y/N)) Analysis2: Hb during ICU stay and predicted outcomes Outcome1: ICU mortality Descriptive Statistics Logistic regression for ICU.mortality (Y/N) Time dependent covariate survival analysis for Time to ICU.mortality
2 2.2 Outcome2: Cardiac dysfunction(tropnin greater than 0.1) Correlated time dependent covariate survival analysis for Time to Tropnin greater than Outcome3: Renal dysfunction Outcome: Time to renal dysfunction Survival analysis with correlated time dependent covariate Outcome4: Time to resolution of delirium Outcome5: Time to successful extubation
3 1 Analysis 1: Transitional Model to model next day s effect 1.1 Outcome1: Daily Delirium Descriptive Statistics Table 1: Descriptive by Icu.status 3 N Normal Discharged Delirious Comatose Deceased Combined Test Statistic N = 1976 N = 697 N = 1994 N = 1917 N = 127 N = 6711 icu.status.today 6711 χ 2 8 = 4326, P < Normal 73% (1451) 66% ( 462) 11% ( 217) 5% ( 88) 9% ( 11) 33% (2229) Delirious 21% ( 410) 28% ( 193) 61% (1210) 20% ( 380) 22% ( 28) 33% (2221) Comatose 6% ( 115) 6% ( 42) 28% ( 567) 76% (1449) 69% ( 88) 34% (2261) Age at enrollment F 4,6706 = 18, P < ICU type 6711 Surgical 28% ( 552) 35% ( 242) 42% ( 844) 38% ( 736) 19% ( 24) 36% (2398) χ 2 4 = 112, P < APACHE APS at enrollment F 4,6706 = 43, P < Charlson score F 4,6706 = 8, P < Framingham stroke risk score F 4,6706 = 24, P < Lowest mean arterial pressure, current 24h F 4,6699 = 51, P < Was patient in ICU today? ICU 100% (1976) 100% ( 697) 100% (1994) 100% (1917) 100% ( 127) 100% (6711) Study day F 4,6706 = 35, P < Septic in the current 24h? 6703 Septic today 60% (1182) 49% ( 344) 67% (1330) 70% (1337) 73% ( 91) 64% (4284) χ 2 4 = 117, P < sevsepsis.l Severely septic today 47% ( 926) 34% ( 240) 65% (1302) 69% (1323) 71% ( 89) 58% (3880) χ 2 4 = 406, P < Lowest hemoglobin, current 24h F 4,5832 = 4.4, P = a b c represent the lower quartile a, the median b, and the upper quartile c for continuous variables. N is the number of non missing values. Numbers after percents are frequencies. Tests used: 1 Pearson test; 2 Kruskal-Wallis test
4 1.1.2 Multinomial logistic regression with Normal as the reference level for outcome variable Delirium 4 Level Discharged vs. Level Normal betahat se pval RRR RRR.lo RRR.up icu.status.todaynormal icu.status.todaydelirious icu.status.todaycomatose age.enroll icu.typesurgical apache.aps charlson.score stroke.risk map.low.icu study.day sepsis.l24septic today hgb.icu study.day:hgb.icu icu.status.todaydelirious:age.enroll icu.status.todaycomatose:age.enroll icu.status.todaydelirious:icu.typesurgical icu.status.todaycomatose:icu.typesurgical icu.status.todaydelirious:apache.aps icu.status.todaycomatose:apache.aps icu.status.todaydelirious:charlson.score icu.status.todaycomatose:charlson.score icu.status.todaydelirious:stroke.risk icu.status.todaycomatose:stroke.risk icu.status.todaydelirious:map.low.icu icu.status.todaycomatose:map.low.icu icu.status.todaydelirious:study.day icu.status.todaycomatose:study.day icu.status.todaydelirious:sepsis.l24septic today icu.status.todaycomatose:sepsis.l24septic today icu.status.todaydelirious:hgb.icu icu.status.todaycomatose:hgb.icu icu.status.todaydelirious:study.day:hgb.icu icu.status.todaycomatose:study.day:hgb.icu
5 5 Level Delirious vs. Level Normal betahat se pval RRR RRR.lo RRR.up icu.status.todaynormal icu.status.todaydelirious icu.status.todaycomatose age.enroll icu.typesurgical apache.aps charlson.score stroke.risk map.low.icu study.day sepsis.l24septic today hgb.icu study.day:hgb.icu icu.status.todaydelirious:age.enroll icu.status.todaycomatose:age.enroll icu.status.todaydelirious:icu.typesurgical icu.status.todaycomatose:icu.typesurgical icu.status.todaydelirious:apache.aps icu.status.todaycomatose:apache.aps icu.status.todaydelirious:charlson.score icu.status.todaycomatose:charlson.score icu.status.todaydelirious:stroke.risk icu.status.todaycomatose:stroke.risk icu.status.todaydelirious:map.low.icu icu.status.todaycomatose:map.low.icu icu.status.todaydelirious:study.day icu.status.todaycomatose:study.day icu.status.todaydelirious:sepsis.l24septic today icu.status.todaycomatose:sepsis.l24septic today icu.status.todaydelirious:hgb.icu icu.status.todaycomatose:hgb.icu icu.status.todaydelirious:study.day:hgb.icu icu.status.todaycomatose:study.day:hgb.icu Level Comatose vs. Level Normal betahat se pval RRR RRR.lo RRR.up
6 6 icu.status.todaynormal icu.status.todaydelirious icu.status.todaycomatose age.enroll icu.typesurgical apache.aps charlson.score stroke.risk map.low.icu study.day sepsis.l24septic today hgb.icu study.day:hgb.icu icu.status.todaydelirious:age.enroll icu.status.todaycomatose:age.enroll icu.status.todaydelirious:icu.typesurgical icu.status.todaycomatose:icu.typesurgical icu.status.todaydelirious:apache.aps icu.status.todaycomatose:apache.aps icu.status.todaydelirious:charlson.score icu.status.todaycomatose:charlson.score icu.status.todaydelirious:stroke.risk icu.status.todaycomatose:stroke.risk icu.status.todaydelirious:map.low.icu icu.status.todaycomatose:map.low.icu icu.status.todaydelirious:study.day icu.status.todaycomatose:study.day icu.status.todaydelirious:sepsis.l24septic today icu.status.todaycomatose:sepsis.l24septic today icu.status.todaydelirious:hgb.icu icu.status.todaycomatose:hgb.icu icu.status.todaydelirious:study.day:hgb.icu icu.status.todaycomatose:study.day:hgb.icu Level Deceased vs. Level Normal betahat se pval RRR RRR.lo RRR.up icu.status.todaynormal icu.status.todaydelirious icu.status.todaycomatose
7 7 age.enroll icu.typesurgical apache.aps charlson.score stroke.risk map.low.icu study.day sepsis.l24septic today hgb.icu study.day:hgb.icu icu.status.todaydelirious:age.enroll icu.status.todaycomatose:age.enroll icu.status.todaydelirious:icu.typesurgical icu.status.todaycomatose:icu.typesurgical icu.status.todaydelirious:apache.aps icu.status.todaycomatose:apache.aps icu.status.todaydelirious:charlson.score icu.status.todaycomatose:charlson.score icu.status.todaydelirious:stroke.risk icu.status.todaycomatose:stroke.risk icu.status.todaydelirious:map.low.icu icu.status.todaycomatose:map.low.icu icu.status.todaydelirious:study.day icu.status.todaycomatose:study.day icu.status.todaydelirious:sepsis.l24septic today icu.status.todaycomatose:sepsis.l24septic today icu.status.todaydelirious:hgb.icu icu.status.todaycomatose:hgb.icu icu.status.todaydelirious:study.day:hgb.icu icu.status.todaycomatose:study.day:hgb.icu Level Withdrawn vs. Level Normal betahat se pval RRR RRR.lo RRR.up icu.status.todaynormal icu.status.todaydelirious icu.status.todaycomatose age.enroll icu.typesurgical apache.aps
8 8 charlson.score stroke.risk map.low.icu study.day sepsis.l24septic today hgb.icu study.day:hgb.icu icu.status.todaydelirious:age.enroll icu.status.todaycomatose:age.enroll icu.status.todaydelirious:icu.typesurgical icu.status.todaycomatose:icu.typesurgical icu.status.todaydelirious:apache.aps icu.status.todaycomatose:apache.aps icu.status.todaydelirious:charlson.score icu.status.todaycomatose:charlson.score icu.status.todaydelirious:stroke.risk icu.status.todaycomatose:stroke.risk icu.status.todaydelirious:map.low.icu icu.status.todaycomatose:map.low.icu icu.status.todaydelirious:study.day icu.status.todaycomatose:study.day icu.status.todaydelirious:sepsis.l24septic today icu.status.todaycomatose:sepsis.l24septic today icu.status.todaydelirious:hgb.icu icu.status.todaycomatose:hgb.icu icu.status.todaydelirious:study.day:hgb.icu icu.status.todaycomatose:study.day:hgb.icu
9 1.2 Outcome2: Daily troponin (greater than 0.1) as an index of cardiac dysfunction Descriptive Statistics Table 2: Descriptive by Troponin (greater than 0.1) 9 N No Yes Combined Test Statistic N = 5710 N = 137 N = 5847 tro.imp.cat 6704 χ 2 1 = 369, P < Yes 5% ( 271) 43% ( 59) 6% ( 330) Highest troponin, current 24h (missing = <0.1) F 1,5840 = 80, P < trop.tomo F 1,5845 = 431, P < Age at enrollment F 1,5845 = 3.6, P = ICU type 6711 Surgical 36% (2067) 36% ( 49) 36% (2116) χ 2 1 = 0.01, P = APACHE APS at enrollment F 1,5845 = 3.5, P = Charlson score F 1,5845 = 4.8, P = Framingham stroke risk score F 1,5845 = 8.1, P = Lowest mean arterial pressure, current 24h F 1,5840 = 18, P < Was patient in ICU today? ICU 100% (5710) 100% ( 137) 100% (5847) Septic in the current 24h? 6703 Septic today 65% (3732) 65% ( 89) 65% (3821) χ 2 1 = 0.01, P = sevsepsis.l Severely septic today 60% (3443) 60% ( 82) 60% (3525) χ 2 1 = 0.01, P = Lowest hemoglobin, current 24h F 1,5157 = 2.1, P = a b c represent the lower quartile a, the median b, and the upper quartile c for continuous variables. N is the number of non missing values. Numbers after percents are frequencies. Tests used: 1 Pearson test; 2 Wilcoxon test Table 3: Outcome:Tropnin ( 0.1).. P-values less than 0.05 are in red. tro.imp.cat=yes ( 6.862, ) age.enroll ( 0.980, 1.027) icu.type=surgical ( 0.607, 1.641) apache.aps ( 0.986, 1.046) charlson.score ( 0.953, 1.177) stroke.risk ( 1.011, 1.117) map.low.icu ( 0.965, 1.002) study.day ( 0.839, 1.443) sepsis.l24=septic today ( 0.584, 1.565) hgb.icu ( 0.850, 1.345) study.day * hgb.icu ( 0.957, 1.014) tro.imp.cat=yes * age.enroll ( 0.944, 1.023) tro.imp.cat=yes * icu.type=surgical ( 0.999, 5.229) tro.imp.cat=yes * apache.aps ( 0.933, 1.030) 0.439
10 Table 3: (continued) tro.imp.cat=yes * charlson.score ( 0.768, 1.105) tro.imp.cat=yes * stroke.risk ( 0.836, 0.995) tro.imp.cat=yes * map.low.icu ( 0.957, 1.021) tro.imp.cat=yes * study.day ( 0.248, 0.908) tro.imp.cat=yes * sepsis.l24=septic today ( 0.738, 3.859) tro.imp.cat=yes * hgb.icu ( 0.617, 1.229) tro.imp.cat=yes * study.day * hgb.icu ( 1.017, 1.163) Outcome3: Daily renal SOFA score Descriptive Statistics Table 4: Descriptive by next day s Renal SOFA score 2 10 N No Yes Combined Test Statistic N = 4319 N = 1528 N = 5847 sofa.renal.cat 6704 χ 2 1 = 4270, P < Yes 4% ( 185) 90% (1380) 27% (1565) SOFA renal component score % (2971) 1% ( 11) 51% (2982) χ 2 4 = 4375, P < % (1161) 9% ( 134) 22% (1295) 2 4% ( 172) 48% ( 726) 15% ( 898) 3 0% ( 6) 21% ( 317) 6% ( 323) 4 0% ( 7) 22% ( 337) 6% ( 344) renal.tomo % (3061) 0% ( 0) 52% (3061) χ 2 4 = 5847, P < % (1258) 0% ( 0) 22% (1258) 2 0% ( 0) 58% ( 886) 15% ( 886) 3 0% ( 0) 20% ( 303) 5% ( 303) 4 0% ( 0) 22% ( 339) 6% ( 339) Age at enrollment F 1,5845 = 22, P < ICU type 6711 Surgical 39% (1668) 29% ( 448) 36% (2116) χ 2 1 = 42, P < APACHE APS at enrollment F 1,5845 = 158, P < Charlson score F 1,5845 = 307, P < Framingham stroke risk score F 1,5845 = 20, P < Lowest mean arterial pressure, current 24h F 1,5840 = 213, P < Was patient in ICU today? ICU 100% (4319) 100% (1528) 100% (5847) Septic in the current 24h? 6703 Septic today 65% (2823) 65% ( 998) 65% (3821) χ 2 1 = 0, P = sevsepsis.l Severely septic today 59% (2533) 65% ( 992) 60% (3525) χ 2 1 = 19, P < Lowest hemoglobin, current 24h F 1,5157 = 34, P < a b c represent the lower quartile a, the median b, and the upper quartile c for continuous variables. N is the number of non missing values. Numbers after percents are frequencies. Tests used: 1 Pearson test; 2 Wilcoxon test
11 Table 5: Descriptive by next day s Renal SOFA score (0-4) 11 N Combined Test Statistic N = 3061 N = 1258 N = 886 N = 303 N = 339 N = 5847 sofa.renal.cat 6704 χ 2 4 = 4379, P < Yes 0% ( 12) 14% ( 173) 85% ( 756) 98% ( 295) 97% ( 329) 27% (1565) SOFA renal component score % (2820) 12% ( 151) 1% ( 7) 1% ( 2) 1% ( 2) 51% (2982) χ 2 16 = 11626, P < % ( 227) 74% ( 934) 14% ( 122) 2% ( 5) 2% ( 7) 22% (1295) 2 0% ( 7) 13% ( 165) 73% ( 642) 22% ( 66) 5% ( 18) 15% ( 898) 3 0% ( 2) 0% ( 4) 10% ( 88) 59% ( 179) 15% ( 50) 6% ( 323) 4 0% ( 3) 0% ( 4) 3% ( 26) 17% ( 50) 77% ( 261) 6% ( 344) Age at enrollment F 4,5842 = 22, P < ICU type 6711 Surgical 40% (1215) 36% ( 453) 35% ( 310) 22% ( 68) 21% ( 70) 36% (2116) χ 2 4 = 77, P < APACHE APS at enrollment F 4,5842 = 52, P < Charlson score F 4,5842 = 93, P < Framingham stroke risk score F 4,5842 = 19, P < Lowest mean arterial pressure, current 24h F 4,5837 = 84, P < Was patient in ICU today? ICU 100% (3061) 100% (1258) 100% ( 886) 100% ( 303) 100% ( 339) 100% (5847) Septic in the current 24h? 6703 Septic today 66% (2032) 63% ( 791) 67% ( 589) 61% ( 184) 67% ( 225) 65% (3821) χ 2 4 = 8.4, P = sevsepsis.l Severely septic today 58% (1781) 60% ( 752) 66% ( 583) 61% ( 184) 67% ( 225) 60% (3525) χ 2 4 = 23, P < Lowest hemoglobin, current 24h F 4,5154 = 11, P < a b c represent the lower quartile a, the median b, and the upper quartile c for continuous variables. N is the number of non missing values. Numbers after percents are frequencies. Tests used: 1 Pearson test; 2 Kruskal-Wallis test
12 1.3.2 Logistic regression, Daily renal SOFA 2(Y/N) Table 6: Outcome:Daily renal SOFA =2.. P-values less than 0.05 are in red. 12 sofa.renal.cat=yes 3e ( 1.184, ) age.enroll 9e ( 0.992, 1.026) icu.type=surgical -3e ( 0.509, 1.073) apache.aps 2e ( 1.001, 1.047) charlson.score 1e ( 1.019, 1.189) stroke.risk -3e ( 0.930, 1.010) map.low.icu -4e ( 0.950, 0.977) <0.001 study.day 1e ( 0.848, 1.202) sepsis.l24=septic today 1e ( 0.694, 1.471) hgb.icu -4e ( 0.806, 1.153) study.day * hgb.icu -1e ( 0.981, 1.017) sofa.renal.cat=yes * age.enroll 7e ( 0.978, 1.024) sofa.renal.cat=yes * icu.type=surgical 5e ( 0.634, 1.741) sofa.renal.cat=yes * apache.aps -1e ( 0.957, 1.015) sofa.renal.cat=yes * charlson.score 1e ( 0.905, 1.130) sofa.renal.cat=yes * stroke.risk 4e ( 0.985, 1.097) sofa.renal.cat=yes * map.low.icu 4e ( 1.020, 1.059) <0.001 sofa.renal.cat=yes * study.day -9e ( 0.717, 1.159) sofa.renal.cat=yes * sepsis.l24=septic today 3e ( 0.820, 2.254) sofa.renal.cat=yes * hgb.icu -7e ( 0.738, 1.185) sofa.renal.cat=yes * study.day * hgb.icu 1e ( 0.987, 1.038) Proportional odds logistic regression regression, Ordinal response (0-4) Table 7: Outcome:Daily renal SOFA (0-4).. P-values less than 0.05 are in red. sofa.renal (11.321, ) <0.001 age.enroll ( 0.999, 1.020) icu.type=surgical ( 0.676, 1.079) apache.aps ( 1.008, 1.037) charlson.score ( 1.009, 1.119) stroke.risk ( 0.964, 1.014) map.low.icu ( 0.965, 0.983) <0.001 study.day ( 0.872, 1.086) sepsis.l24=septic today ( 0.823, 1.319) hgb.icu ( 0.879, 1.093) study.day * hgb.icu ( 0.991, 1.014) 0.722
13 Table 7: (continued) sofa.renal * age.enroll ( 0.991, 1.003) sofa.renal * icu.type=surgical ( 0.860, 1.149) sofa.renal * apache.aps ( 0.981, 0.997) sofa.renal * charlson.score ( 0.960, 1.019) sofa.renal * stroke.risk ( 0.990, 1.019) sofa.renal * map.low.icu ( 1.004, 1.015) <0.001 sofa.renal * study.day ( 0.924, 1.047) sofa.renal * sepsis.l24=septic today ( 0.866, 1.146) sofa.renal * hgb.icu ( 0.932, 1.059) sofa.renal * study.day * hgb.icu ( 0.996, 1.009) Outcome4: Daily resp SOFA score ( 2) Descriptive Statistics Table 8: Descriptive by next day s RESP SOFA score ( 2) 13 N No Yes Combined Test Statistic N = 276 N = 5571 N = 5847 sofa.resp.cat 6704 χ 2 1 = 2576, P < Yes 39% ( 107) 99% (5510) 96% (5617) SOFA respiratory component score % ( 0) 0% ( 9) 0% ( 9) χ 2 4 = 2717, P < % ( 169) 1% ( 47) 4% ( 216) 2 34% ( 95) 51% (2852) 50% (2947) 3 4% ( 12) 45% (2505) 43% (2517) 4 0% ( 0) 3% ( 153) 3% ( 153) resp.tomo % ( 8) 0% ( 0) 0% ( 8) χ 2 4 = 5847, P < % ( 268) 0% ( 0) 5% ( 268) 2 0% ( 0) 57% (3152) 54% (3152) 3 0% ( 0) 41% (2285) 39% (2285) 4 0% ( 0) 2% ( 134) 2% ( 134) Age at enrollment F 1,5845 = 0.13, P = ICU type 6711 Surgical 23% ( 64) 37% (2052) 36% (2116) χ 2 1 = 21, P < APACHE APS at enrollment F 1,5845 = 33, P < Charlson score F 1,5845 = 0.71, P = Framingham stroke risk score F 1,5845 = 2.4, P = Lowest mean arterial pressure, current 24h F 1,5840 = 9.6, P = Was patient in ICU today? ICU 100% ( 276) 100% (5571) 100% (5847) Septic in the current 24h? 6703 a b c represent the lower quartile a, the median b, and the upper quartile c for continuous variables. χ 2 1 = 157, P < N is the number of non missing values. Numbers after percents are frequencies. Tests used: 1 Pearson test; 2 Wilcoxon test
14 Table 8: (continued) N No Yes Combined Test Statistic N = 276 N = 5571 N = 5847 Septic today 30% ( 84) 67% (3737) 65% (3821) sevsepsis.l Severely septic today 22% ( 61) 62% (3464) 60% (3525) χ 2 1 = 177, P < Lowest hemoglobin, current 24h F 1,5157 = 38, P < a b c represent the lower quartile a, the median b, and the upper quartile c for continuous variables. N is the number of non missing values. Numbers after percents are frequencies. Tests used: 1 Pearson test; 2 Wilcoxon test Table 9: Descriptive by next day s RESP SOFA score (0-4) 14 N Combined Test Statistic N = 8 N = 268 N = 3152 N = 2285 N = 134 N = 5847 sofa.resp.cat 6704 χ 2 4 = 2633, P < Yes 88% ( 7) 37% ( 100) 99% (3102) 100% (2275) 99% ( 133) 96% (5617) SOFA respiratory component score % ( 0) 0% ( 0) 0% ( 7) 0% ( 2) 0% ( 0) 0% ( 9) χ 2 16 = 5156, P < % ( 1) 63% ( 168) 1% ( 40) 0% ( 6) 1% ( 1) 4% ( 216) 2 38% ( 3) 34% ( 92) 76% (2404) 19% ( 435) 10% ( 13) 50% (2947) 3 50% ( 4) 3% ( 8) 22% ( 685) 76% (1744) 57% ( 76) 43% (2517) 4 0% ( 0) 0% ( 0) 0% ( 13) 4% ( 96) 33% ( 44) 3% ( 153) renal.tomo % ( 6) 44% ( 119) 55% (1718) 51% (1158) 45% ( 60) 52% (3061) χ 2 16 = 101, P < % ( 0) 16% ( 44) 20% ( 644) 23% ( 534) 27% ( 36) 22% (1258) 2 25% ( 2) 15% ( 40) 14% ( 441) 17% ( 380) 17% ( 23) 15% ( 886) 3 0% ( 0) 7% ( 19) 6% ( 182) 4% ( 95) 5% ( 7) 5% ( 303) 4 0% ( 0) 17% ( 46) 5% ( 167) 5% ( 118) 6% ( 8) 6% ( 339) Age at enrollment F 4,5842 = 11, P < ICU type 6711 Surgical 75% ( 6) 22% ( 58) 37% (1177) 37% ( 841) 25% ( 34) 36% (2116) χ 2 4 = 39, P < APACHE APS at enrollment F 4,5842 = 11, P < Charlson score F 4,5842 = 6.4, P < Framingham stroke risk score F 4,5842 = 11, P < Lowest mean arterial pressure, current 24h F 4,5837 = 8.9, P < Was patient in ICU today? ICU 100% ( 8) 100% ( 268) 100% (3152) 100% (2285) 100% ( 134) 100% (5847) Septic in the current 24h? 6703 Septic today 75% ( 6) 29% ( 78) 64% (2024) 71% (1611) 76% ( 102) 65% (3821) χ 2 4 = 192, P < sevsepsis.l Severely septic today 62% ( 5) 21% ( 56) 60% (1876) 66% (1505) 62% ( 83) 60% (3525) χ 2 4 = 205, P < Lowest hemoglobin, current 24h F 4,5154 = 12, P < a b c represent the lower quartile a, the median b, and the upper quartile c for continuous variables. N is the number of non missing values. Numbers after percents are frequencies. Tests used: 1 Pearson test; 2 Kruskal-Wallis test Logistic regression, Daily RESP SOFA 2(Y/N)
15 Table 10: Outcome:Daily resp SOFA =2.P-values less than 0.05 are in red. 15 sofa.resp.catyes ( , ) <0.001 age.enroll ( 0.955, 1.041) icu.typesurgical ( 1.022, 6.859) apache.aps ( 0.991, 1.088) charlson.score ( 0.801, 1.107) stroke.risk ( 0.952, 1.130) map.low.icu ( 0.956, 1.004) study.day ( 0.482, 1.373) sepsis.l24septic today ( 0.826, 4.261) hgb.icu ( 0.554, 1.329) study.day:hgb.icu ( 0.963, 1.068) sofa.resp.catyes:age.enroll ( 0.954, 1.050) sofa.resp.catyes:icu.typesurgical ( 0.150, 1.233) sofa.resp.catyes:apache ( 0.930, 1.035) sofa.resp.catyes:charlson.score ( 0.801, 1.155) sofa.resp.catyes:stroke.risk ( 0.893, 1.085) sofa.resp.catyes:map.low.icu ( 0.975, 1.034) sofa.resp.catyes:study.day ( 0.730, 2.243) sofa.resp.catyes:sepsis.l24septic today ( 0.611, 3.908) sofa.resp.catyes:hgb.icu ( 0.632, 1.624) sofa.resp.catyes:study.day:hgb.icu ( 0.928, 1.036) Proportional odds logistic regression, Ordinal response (0-4) Table 11: Outcome:Daily resp SOFA (0-4).. P-values less than 0.05 are in red. sofa.resp ( 2.842, ) <0.001 age.enroll ( 0.964, 1.013) icu.type=surgical ( 3.157, 9.969) <0.001 apache.aps ( 1.000, 1.064) charlson.score ( 0.852, 1.091) stroke.risk ( 1.042, 1.173) <0.001 map.low.icu ( 0.958, 0.998) study.day ( 0.753, 1.283) sepsis.l24=septic today ( 2.562, 8.005) <0.001 hgb.icu ( 0.557, 0.949) study.day * hgb.icu ( 0.970, 1.024) sofa.resp * age.enroll ( 0.995, 1.014) sofa.resp * icu.type=surgical ( 0.408, 0.635) <0.001 sofa.resp * apache.aps ( 0.974, 0.997) sofa.resp * charlson.score ( 0.942, 1.039) 0.662
16 Table 11: (continued) sofa.resp * stroke.risk ( 0.940, 0.983) <0.001 sofa.resp * map.low.icu ( 0.998, 1.013) sofa.resp * study.day ( 0.894, 1.099) sofa.resp * sepsis.l24=septic today ( 0.499, 0.775) <0.001 sofa.resp * hgb.icu ( 1.000, 1.218) sofa.resp * study.day * hgb.icu ( 0.992, 1.013) Proportional odds logistic regression, Ordinal response (0-4),without three way interaction Table 12: Outcome:Daily resp SOFA (0-4).. P-values less than 0.05 are in red. 16 sofa.resp ( 3.029, ) <0.001 age.enroll ( 0.964, 1.013) icu.type=surgical ( 3.148, 9.935) <0.001 apache.aps ( 1.000, 1.064) charlson.score ( 0.852, 1.091) stroke.risk ( 1.043, 1.174) <0.001 map.low.icu ( 0.958, 0.998) study.day ( 0.854, 0.994) sepsis.l24=septic today ( 2.561, 8.000) <0.001 hgb.icu ( 0.573, 0.840) <0.001 sofa.resp * age.enroll ( 0.996, 1.014) sofa.resp * icu.type=surgical ( 0.409, 0.636) <0.001 sofa.resp * apache.aps ( 0.974, 0.997) sofa.resp * charlson.score ( 0.942, 1.039) sofa.resp * stroke.risk ( 0.939, 0.983) <0.001 sofa.resp * map.low.icu ( 0.998, 1.013) sofa.resp * study.day ( 1.001, 1.033) sofa.resp * sepsis.l24=septic today ( 0.499, 0.775) <0.001 sofa.resp * hgb.icu ( 1.051, 1.202) <0.001 study.day * hgb.icu ( 0.997, 1.010) Proportional odds logistic regression, Ordinal response (0-4),Sensitivity Analysis
17 Table 13: Outcome:Daily resp SOFA (0-4).. P-values less than 0.05 are in red. 17 sofa.resp ( 3.508, ) <0.001 age.enroll ( 0.966, 1.019) icu.type=surgical ( 3.240, ) <0.001 apache.aps ( 0.994, 1.060) charlson.score ( 0.898, 1.169) stroke.risk ( 1.020, 1.156) map.low.icu ( 0.956, 0.997) study.day ( 0.696, 1.353) sepsis.l24=septic today ( 2.624, 8.658) <0.001 hgb.icu ( 0.587, 1.095) study.day * hgb.icu ( 0.966, 1.032) sofa.resp * age.enroll ( 0.993, 1.013) sofa.resp * icu.type=surgical ( 0.396, 0.632) <0.001 sofa.resp * apache.aps ( 0.975, 1.000) sofa.resp * charlson.score ( 0.920, 1.022) sofa.resp * stroke.risk ( 0.945, 0.991) sofa.resp * map.low.icu ( 0.998, 1.014) sofa.resp * study.day ( 0.878, 1.141) sofa.resp * sepsis.l24=septic today ( 0.487, 0.773) <0.001 sofa.resp * hgb.icu ( 0.947, 1.195) sofa.resp * study.day * hgb.icu ( 0.989, 1.015) Proportional odds logistic regression, Ordinal response (0-4),Sensitivity Analysis2, without three way interaction Table 14: Outcome:Daily resp SOFA (0-4).. P-values less than 0.05 are in red. sofa.resp ( 4.315, ) <0.001 age.enroll ( 0.966, 1.018) icu.type=surgical ( 3.237, ) <0.001 apache.aps ( 0.993, 1.060) charlson.score ( 0.898, 1.169) stroke.risk ( 1.021, 1.156) map.low.icu ( 0.956, 0.997) study.day ( 0.852, 1.024) sepsis.l24=septic today ( 2.617, 8.625) <0.001 hgb.icu ( 0.624, 0.980) sofa.resp * age.enroll ( 0.993, 1.013) sofa.resp * icu.type=surgical ( 0.397, 0.633) <0.001 sofa.resp * apache.aps ( 0.975, 1.000) 0.056
18 Table 14: (continued) sofa.resp * charlson.score ( 0.920, 1.022) sofa.resp * stroke.risk ( 0.945, 0.990) sofa.resp * map.low.icu ( 0.998, 1.014) sofa.resp * study.day ( 1.000, 1.034) sofa.resp * sepsis.l24=septic today ( 0.488, 0.774) <0.001 sofa.resp * hgb.icu ( 0.993, 1.164) study.day * hgb.icu ( 0.994, 1.011) Proportional odds logistic regression, Ordinal response (0-4),Marginal Model Table 15: Outcome:Daily resp SOFA (0-4).. P-values less than 0.05 are in red. 18 sofa.resp= ( , ) <0.001 sofa.resp= ( , ) <0.001 sofa.resp= ( , ) <0.001 age.enroll ( 0.994, 1.006) icu.type=surgical ( 0.875, 1.141) apache.aps ( 0.987, 1.003) charlson.score ( 0.913, 0.972) <0.001 stroke.risk ( 0.983, 1.012) map.low.icu ( 0.986, 0.996) <0.001 study.day ( 0.917, 1.037) sepsis.l24=septic today ( 1.207, 1.589) <0.001 hgb.icu ( 0.898, 1.015) study.day * hgb.icu ( 0.996, 1.008) Proportional odds logistic regression, Ordinal response (0-4),Marginal Model with hgb.icu and resp.sofa interaction Table 16: Outcome:Daily resp SOFA (0-4).. P-values less than 0.05 are in red. sofa.resp= ( , ) <0.001 sofa.resp= ( , ) <0.001 sofa.resp= ( , ) <0.001 hgb.icu ( 0.728, 1.118) age.enroll ( 0.995, 1.007) 0.866
19 Table 16: (continued) icu.type=surgical ( 0.871, 1.135) apache.aps ( 0.986, 1.002) charlson.score ( 0.909, 0.968) <0.001 stroke.risk ( 0.984, 1.012) map.low.icu ( 0.987, 0.997) study.day ( 0.898, 1.017) sepsis.l24=septic today ( 1.209, 1.593) <0.001 sofa.resp=2 * hgb.icu ( 0.762, 1.181) sofa.resp=3 * hgb.icu ( 0.880, 1.354) sofa.resp=4 * hgb.icu ( 0.997, 1.831) hgb.icu * study.day ( 0.998, 1.011)
20 20
21 1.4.9 Tomorrow resp.sofa 2 Today level 34 Today level P(y>=2) 0.6 Today level 1 1 Today level 2 2 Today level 3 3 Today level Today level Lowest hemoglobin, current 24h ted to:age.enroll=61.76 icu.type=medical apache.aps=22 charlson.score=2 stroke.risk=10 map.low.icu=62 study.day=6 sepsis.l24=septic today
22 22
23 Tomorrow resp.sofa 3 Today level Today level P(y>=3) Today level 1 1 Today level 2 2 Today level 3 3 Today level Today level 2 Today level Lowest hemoglobin, current 24h ted to:age.enroll=61.76 icu.type=medical apache.aps=22 charlson.score=2 stroke.risk=10 map.low.icu=62 study.day=6 sepsis.l24=septic today
24 24
25 Tomorrow resp.sofa Today level 1 1 Today level 2 2 Today level 3 3 Today level P(y>=4) 0.4 Today level Today level 3 Today level Lowest hemoglobin, current 24h ted to:age.enroll=61.76 icu.type=medical apache.aps=22 charlson.score=2 stroke.risk=10 map.low.icu=62 study.day=6 sepsis.l24=septic today
26 1.5 Outcome5: Daily Ventilator (Y/N) as a measure of success of extubation Descriptive Statistics Proportional odds logistic regression, Ordinal response (0-4),Marginal Model Table 17: Descriptive by ventilator (Y/N) 26 N No Yes Combined Test Statistic N = 1580 N = 4269 N = 5849 on.vent.l χ 2 1 = 3661, P < Yes 25% ( 399) 98% (4198) 79% (4597) Age at enrollment F 1,5847 = 8.7, P = ICU type 6711 Surgical 32% ( 500) 38% (1616) 36% (2116) χ 2 1 = 19, P < APACHE APS at enrollment F 1,5847 = 110, P < Charlson score F 1,5847 = 16, P < Framingham stroke risk score F 1,5847 = 0.06, P = Lowest mean arterial pressure, current 24h F 1,5842 = 83, P < Was patient in ICU today? ICU 100% (1580) 100% (4269) 100% (5849) Septic in the current 24h? 6703 Septic today 54% ( 858) 69% (2963) 65% (3821) χ 2 1 = 116, P < sevsepsis.l Severely septic today 37% ( 587) 69% (2938) 60% (3525) χ 2 1 = 483, P < Lowest hemoglobin, current 24h F 1,5157 = 29, P < a b c represent the lower quartile a, the median b, and the upper quartile c for continuous variables. N is the number of non missing values. Numbers after percents are frequencies. Tests used: 1 Pearson test; 2 Wilcoxon test Logistic regression (Daily Ventilation (Y/N)) Table 18: Outcome:Daily ventilation (Y/N).. P-values less than 0.05 are in red. on.vent.l24=yes 9e e+04 ( , ) <0.001 age.enroll 4e e+00 ( 0.977, 1.031) icu.type=surgical 7e e+00 ( 1.123, 3.430) apache.aps 1e e+00 ( 0.978, 1.043) charlson.score 2e e+00 ( 0.906, 1.154) stroke.risk -7e e+00 ( 0.938, 1.052) map.low.icu -3e e+00 ( 0.979, 1.016) study.day 2e e+00 ( 0.939, 1.579) sepsis.l24=septic today 6e e+00 ( 1.016, 3.119) hgb.icu 3e e+00 ( 0.989, 1.703) study.day * hgb.icu -2e e+00 ( 0.952, 1.004) 0.093
27 Table 18: (continued) on.vent.l24=yes * age.enroll 4e e+00 ( 0.975, 1.033) on.vent.l24=yes * icu.type=surgical -9e e-01 ( 0.224, 0.750) on.vent.l24=yes * apache.aps 2e e+00 ( 0.966, 1.036) on.vent.l24=yes * charlson.score -8e e-01 ( 0.812, 1.054) on.vent.l24=yes * stroke.risk -1e e+00 ( 0.926, 1.049) on.vent.l24=yes * map.low.icu -3e e+00 ( 0.953, 0.993) on.vent.l24=yes * study.day -6e e-01 ( 0.709, 1.262) on.vent.l24=yes * sepsis.l24=septic today -2e e-01 ( 0.452, 1.520) on.vent.l24=yes * hgb.icu -2e e-01 ( 0.584, 1.045) on.vent.l24=yes * study.day * hgb.icu 1e e+00 ( 0.982, 1.042)
28 2 Analysis2: Hb during ICU stay and predicted outcomes 2.1 Outcome1: ICU mortality Descriptive Statistics Table 19: Descriptive by Mortality (Y/N) unique subject) N No Yes Combined Test Statistic N = 685 N = 128 N = 813 Age at enrollment F 1,811 = 4.3, P = ICU type 813 Surgical 35% (238) 19% ( 24) 32% (262) χ 2 1 = 13, P < APACHE APS at enrollment F 1,811 = 22, P < Charlson score F 1,811 = 6.5, P = Framingham stroke risk score F 1,811 = 1.9, P = lowest.hgb.icu F 1,801 = 4.8, P = auc.hgb.icu F 1,811 = 0.57, P = mean.sofa F 1,811 = 134, P < duration.sepsis F 1,811 = 12, P < max.sofa.renal % (285) 15% ( 19) 37% (304) χ 2 4 = 44, P < % (163) 23% ( 30) 24% (193) 2 19% (132) 37% ( 47) 22% (179) 3 6% ( 39) 12% ( 16) 7% ( 55) 4 10% ( 66) 12% ( 16) 10% ( 82) a b c represent the lower quartile a, the median b, and the upper quartile c for continuous variables. N is the number of non missing values. Numbers after percents are frequencies. Tests used: 1 Wilcoxon test; 2 Pearson test Logistic regression for ICU.mortality (Y/N) Two different functional forms of hemoglobin during ICU was considered in the below analysis (two different models) Table 20: Outcome:Mortality.. P-values less than 0.05 are in red. lowest.hgb.icu (0.857, 1.157) age.enroll (1.019, 1.065) <0.001 icu.type=surgical (0.217, 0.633) <0.001 apache.aps (0.964, 1.024) charlson.score (0.892, 1.090) stroke.risk (0.907, 1.000) mean.sofa (1.401, 1.670) <0.001 duration.sepsis (0.982, 1.063) Table 21: Outcome:Mortality.. P-values less than 0.05 are in red. auc.hgb.icu (0.991, 1.004) age.enroll (1.020, 1.065) <0.001 icu.type=surgical (0.223, 0.655) <0.001 apache.aps (0.964, 1.024) charlson.score (0.897, 1.094) stroke.risk (0.905, 0.998) mean.sofa (1.407, 1.674) <0.001 duration.sepsis (0.978, 1.114)
29 2.1.3 Time dependent covariate survival analysis for Time to ICU.mortality Variable hgb.icu, daily.sofa, severe sepsis were time variang covariates. charlson.score,stroke.risk did not change over time. Age.enroll, icu.type, apache.aps, Call: coxph(formula = Surv(time1, time2, mortality.censor) ~ hgb.icu + age.enroll + icu.type + apache.aps + charlson.score + stroke.risk + daily.sofa + sevsepsis.l24 + cluster(id), data = brain.dailyobs) coef exp(coef) se(coef) robust se z p hgb.icu e-01 age.enroll e-02 icu.typesurgical e-03 apache.aps e-01 charlson.score e-02 stroke.risk e-02 daily.sofa e-11 sevsepsis.l24severely septic today e-01 Likelihood ratio test=78.3 on 8 df, p=1.05e-13 n= 5848, number of events= 84 (876 observations deleted due to missingness) For every unit increase of hgb.icu, the hazard of mortality in ICU increased by 6.4%. Given the p value equals to 0.4, no evidence to conclude hgb.icu has statistically significant effect on ICU.mortality after controlling for all other covariates. The robust standard error is used due to repeated measurement of hgb.icu on each subject. 29
30 2.2 Outcome2: Cardiac dysfunction(tropnin greater than 0.1) Correlated time dependent covariate survival analysis for Time to Tropnin greater than 0.1 Variable hgb.icu, daily.sofa, severe sepsis were time variang covariates. charlson.score,stroke.risk did not change over time. Call: coxph(formula = Surv(time1, time2, tropnin.censor) ~ hgb.icu + age.enroll + icu.type + apache.aps + charlson.score + stroke.risk + daily.sofa + sevsepsis.l24 + cluster(id), data = brain.dailyobs) Age.enroll, icu.type, apache.aps, coef exp(coef) se(coef) robust se z p hgb.icu age.enroll icu.typesurgical apache.aps charlson.score stroke.risk daily.sofa sevsepsis.l24severely septic today Likelihood ratio test=33.4 on 8 df, p= n= 5848, number of events= 129 (876 observations deleted due to missingness) 30
31 2.3 Outcome3: Renal dysfunction Outcome: Time to renal dysfunction Survival analysis with correlated time dependent covariate Variable hgb.icu, daily.sofa, severe sepsis were time variang covariates. charlson.score,stroke.risk did not change over time. Call: coxph(formula = Surv(time1, time2, renal.censor) ~ hgb.icu + age.enroll + icu.type + apache.aps + charlson.score + stroke.risk + daily.sofa + sevsepsis.l24 + cluster(id), data = brain.dailyobs) Age.enroll, icu.type, apache.aps, coef exp(coef) se(coef) robust se z p hgb.icu age.enroll icu.typesurgical apache.aps charlson.score stroke.risk daily.sofa sevsepsis.l24severely septic today Likelihood ratio test=508 on 8 df, p=0 n= 5848, number of events= 557 (876 observations deleted due to missingness) 2.4 Outcome4: Time to resolution of delirium To be continue 2.5 Outcome5: Time to successful extubation To be continue 31
Lecture 12. Multivariate Survival Data Statistics Survival Analysis. Presented March 8, 2016
Statistics 255 - Survival Analysis Presented March 8, 2016 Dan Gillen Department of Statistics University of California, Irvine 12.1 Examples Clustered or correlated survival times Disease onset in family
More informationThe coxvc_1-1-1 package
Appendix A The coxvc_1-1-1 package A.1 Introduction The coxvc_1-1-1 package is a set of functions for survival analysis that run under R2.1.1 [81]. This package contains a set of routines to fit Cox models
More informationTurning a research question into a statistical question.
Turning a research question into a statistical question. IGINAL QUESTION: Concept Concept Concept ABOUT ONE CONCEPT ABOUT RELATIONSHIPS BETWEEN CONCEPTS TYPE OF QUESTION: DESCRIBE what s going on? DECIDE
More informationRelative-risk regression and model diagnostics. 16 November, 2015
Relative-risk regression and model diagnostics 16 November, 2015 Relative risk regression More general multiplicative intensity model: Intensity for individual i at time t is i(t) =Y i (t)r(x i, ; t) 0
More informationLecture 7 Time-dependent Covariates in Cox Regression
Lecture 7 Time-dependent Covariates in Cox Regression So far, we ve been considering the following Cox PH model: λ(t Z) = λ 0 (t) exp(β Z) = λ 0 (t) exp( β j Z j ) where β j is the parameter for the the
More informationSurvival analysis in R
Survival analysis in R Niels Richard Hansen This note describes a few elementary aspects of practical analysis of survival data in R. For further information we refer to the book Introductory Statistics
More informationLecture 8 Stat D. Gillen
Statistics 255 - Survival Analysis Presented February 23, 2016 Dan Gillen Department of Statistics University of California, Irvine 8.1 Example of two ways to stratify Suppose a confounder C has 3 levels
More informationβ j = coefficient of x j in the model; β = ( β1, β2,
Regression Modeling of Survival Time Data Why regression models? Groups similar except for the treatment under study use the nonparametric methods discussed earlier. Groups differ in variables (covariates)
More informationMatched Pair Data. Stat 557 Heike Hofmann
Matched Pair Data Stat 557 Heike Hofmann Outline Marginal Homogeneity - review Binary Response with covariates Ordinal response Symmetric Models Subject-specific vs Marginal Model conditional logistic
More informationIntroduction to Statistical Analysis
Introduction to Statistical Analysis Changyu Shen Richard A. and Susan F. Smith Center for Outcomes Research in Cardiology Beth Israel Deaconess Medical Center Harvard Medical School Objectives Descriptive
More informationSurvival analysis in R
Survival analysis in R Niels Richard Hansen This note describes a few elementary aspects of practical analysis of survival data in R. For further information we refer to the book Introductory Statistics
More informationCDA Chapter 3 part II
CDA Chapter 3 part II Two-way tables with ordered classfications Let u 1 u 2... u I denote scores for the row variable X, and let ν 1 ν 2... ν J denote column Y scores. Consider the hypothesis H 0 : X
More informationMAS3301 / MAS8311 Biostatistics Part II: Survival
MAS3301 / MAS8311 Biostatistics Part II: Survival M. Farrow School of Mathematics and Statistics Newcastle University Semester 2, 2009-10 1 13 The Cox proportional hazards model 13.1 Introduction In the
More informationNemours Biomedical Research Statistics Course. Li Xie Nemours Biostatistics Core October 14, 2014
Nemours Biomedical Research Statistics Course Li Xie Nemours Biostatistics Core October 14, 2014 Outline Recap Introduction to Logistic Regression Recap Descriptive statistics Variable type Example of
More informationStat 642, Lecture notes for 04/12/05 96
Stat 642, Lecture notes for 04/12/05 96 Hosmer-Lemeshow Statistic The Hosmer-Lemeshow Statistic is another measure of lack of fit. Hosmer and Lemeshow recommend partitioning the observations into 10 equal
More informationMultistate models and recurrent event models
Multistate models Multistate models and recurrent event models Patrick Breheny December 10 Patrick Breheny Survival Data Analysis (BIOS 7210) 1/22 Introduction Multistate models In this final lecture,
More informationIn contrast, parametric techniques (fitting exponential or Weibull, for example) are more focussed, can handle general covariates, but require
Chapter 5 modelling Semi parametric We have considered parametric and nonparametric techniques for comparing survival distributions between different treatment groups. Nonparametric techniques, such as
More informationHypothesis testing, part 2. With some material from Howard Seltman, Blase Ur, Bilge Mutlu, Vibha Sazawal
Hypothesis testing, part 2 With some material from Howard Seltman, Blase Ur, Bilge Mutlu, Vibha Sazawal 1 CATEGORICAL IV, NUMERIC DV 2 Independent samples, one IV # Conditions Normal/Parametric Non-parametric
More informationIntroduction and Descriptive Statistics p. 1 Introduction to Statistics p. 3 Statistics, Science, and Observations p. 5 Populations and Samples p.
Preface p. xi Introduction and Descriptive Statistics p. 1 Introduction to Statistics p. 3 Statistics, Science, and Observations p. 5 Populations and Samples p. 6 The Scientific Method and the Design of
More informationDETAILED CONTENTS PART I INTRODUCTION AND DESCRIPTIVE STATISTICS. 1. Introduction to Statistics
DETAILED CONTENTS About the Author Preface to the Instructor To the Student How to Use SPSS With This Book PART I INTRODUCTION AND DESCRIPTIVE STATISTICS 1. Introduction to Statistics 1.1 Descriptive and
More informationStatistics in medicine
Statistics in medicine Lecture 4: and multivariable regression Fatma Shebl, MD, MS, MPH, PhD Assistant Professor Chronic Disease Epidemiology Department Yale School of Public Health Fatma.shebl@yale.edu
More informationBeyond GLM and likelihood
Stat 6620: Applied Linear Models Department of Statistics Western Michigan University Statistics curriculum Core knowledge (modeling and estimation) Math stat 1 (probability, distributions, convergence
More informationMarginal versus conditional effects: does it make a difference? Mireille Schnitzer, PhD Université de Montréal
Marginal versus conditional effects: does it make a difference? Mireille Schnitzer, PhD Université de Montréal Overview In observational and experimental studies, the goal may be to estimate the effect
More informationSurvival Regression Models
Survival Regression Models David M. Rocke May 18, 2017 David M. Rocke Survival Regression Models May 18, 2017 1 / 32 Background on the Proportional Hazards Model The exponential distribution has constant
More informationMultistate models and recurrent event models
and recurrent event models Patrick Breheny December 6 Patrick Breheny University of Iowa Survival Data Analysis (BIOS:7210) 1 / 22 Introduction In this final lecture, we will briefly look at two other
More informationSSUI: Presentation Hints 2 My Perspective Software Examples Reliability Areas that need work
SSUI: Presentation Hints 1 Comparing Marginal and Random Eects (Frailty) Models Terry M. Therneau Mayo Clinic April 1998 SSUI: Presentation Hints 2 My Perspective Software Examples Reliability Areas that
More informationFoundations of Probability and Statistics
Foundations of Probability and Statistics William C. Rinaman Le Moyne College Syracuse, New York Saunders College Publishing Harcourt Brace College Publishers Fort Worth Philadelphia San Diego New York
More informationPart [1.0] Measures of Classification Accuracy for the Prediction of Survival Times
Part [1.0] Measures of Classification Accuracy for the Prediction of Survival Times Patrick J. Heagerty PhD Department of Biostatistics University of Washington 1 Biomarkers Review: Cox Regression Model
More informationLecture 12: Effect modification, and confounding in logistic regression
Lecture 12: Effect modification, and confounding in logistic regression Ani Manichaikul amanicha@jhsph.edu 4 May 2007 Today Categorical predictor create dummy variables just like for linear regression
More informationSection IX. Introduction to Logistic Regression for binary outcomes. Poisson regression
Section IX Introduction to Logistic Regression for binary outcomes Poisson regression 0 Sec 9 - Logistic regression In linear regression, we studied models where Y is a continuous variable. What about
More informationLecture 7. Proportional Hazards Model - Handling Ties and Survival Estimation Statistics Survival Analysis. Presented February 4, 2016
Proportional Hazards Model - Handling Ties and Survival Estimation Statistics 255 - Survival Analysis Presented February 4, 2016 likelihood - Discrete Dan Gillen Department of Statistics University of
More informationREGRESSION ANALYSIS FOR TIME-TO-EVENT DATA THE PROPORTIONAL HAZARDS (COX) MODEL ST520
REGRESSION ANALYSIS FOR TIME-TO-EVENT DATA THE PROPORTIONAL HAZARDS (COX) MODEL ST520 Department of Statistics North Carolina State University Presented by: Butch Tsiatis, Department of Statistics, NCSU
More informationSample Size/Power Calculation by Software/Online Calculators
Sample Size/Power Calculation by Software/Online Calculators May 24, 2018 Li Zhang, Ph.D. li.zhang@ucsf.edu Associate Professor Department of Epidemiology and Biostatistics Division of Hematology and Oncology
More informationECLT 5810 Linear Regression and Logistic Regression for Classification. Prof. Wai Lam
ECLT 5810 Linear Regression and Logistic Regression for Classification Prof. Wai Lam Linear Regression Models Least Squares Input vectors is an attribute / feature / predictor (independent variable) The
More informationNonparametric Statistics
Nonparametric Statistics Nonparametric or Distribution-free statistics: used when data are ordinal (i.e., rankings) used when ratio/interval data are not normally distributed (data are converted to ranks)
More informationOutline. Frailty modelling of Multivariate Survival Data. Clustered survival data. Clustered survival data
Outline Frailty modelling of Multivariate Survival Data Thomas Scheike ts@biostat.ku.dk Department of Biostatistics University of Copenhagen Marginal versus Frailty models. Two-stage frailty models: copula
More informationStatistics 262: Intermediate Biostatistics Model selection
Statistics 262: Intermediate Biostatistics Model selection Jonathan Taylor & Kristin Cobb Statistics 262: Intermediate Biostatistics p.1/?? Today s class Model selection. Strategies for model selection.
More informationAnalysis of Categorical Data. Nick Jackson University of Southern California Department of Psychology 10/11/2013
Analysis of Categorical Data Nick Jackson University of Southern California Department of Psychology 10/11/2013 1 Overview Data Types Contingency Tables Logit Models Binomial Ordinal Nominal 2 Things not
More informationOn a connection between the Bradley-Terry model and the Cox proportional hazards model
On a connection between the Bradley-Terry model and the Cox proportional hazards model Yuhua Su and Mai Zhou Department of Statistics University of Kentucky Lexington, KY 40506-0027, U.S.A. SUMMARY This
More informationBooklet of Code and Output for STAD29/STA 1007 Midterm Exam
Booklet of Code and Output for STAD29/STA 1007 Midterm Exam List of Figures in this document by page: List of Figures 1 Packages................................ 2 2 Hospital infection risk data (some).................
More informationProbabilistic Index Models
Probabilistic Index Models Jan De Neve Department of Data Analysis Ghent University M3 Storrs, Conneticut, USA May 23, 2017 Jan.DeNeve@UGent.be 1 / 37 Introduction 2 / 37 Introduction to Probabilistic
More informationSmall n, σ known or unknown, underlying nongaussian
READY GUIDE Summary Tables SUMMARY-1: Methods to compute some confidence intervals Parameter of Interest Conditions 95% CI Proportion (π) Large n, p 0 and p 1 Equation 12.11 Small n, any p Figure 12-4
More informationLogistic Regressions. Stat 430
Logistic Regressions Stat 430 Final Project Final Project is, again, team based You will decide on a project - only constraint is: you are supposed to use techniques for a solution that are related to
More informationTextbook Examples of. SPSS Procedure
Textbook s of IBM SPSS Procedures Each SPSS procedure listed below has its own section in the textbook. These sections include a purpose statement that describes the statistical test, identification of
More informationTABLES AND FORMULAS FOR MOORE Basic Practice of Statistics
TABLES AND FORMULAS FOR MOORE Basic Practice of Statistics Exploring Data: Distributions Look for overall pattern (shape, center, spread) and deviations (outliers). Mean (use a calculator): x = x 1 + x
More informationMaster s Written Examination - Solution
Master s Written Examination - Solution Spring 204 Problem Stat 40 Suppose X and X 2 have the joint pdf f X,X 2 (x, x 2 ) = 2e (x +x 2 ), 0 < x < x 2
More informationLogistic Regression. Continued Psy 524 Ainsworth
Logistic Regression Continued Psy 524 Ainsworth Equations Regression Equation Y e = 1 + A+ B X + B X + B X 1 1 2 2 3 3 i A+ B X + B X + B X e 1 1 2 2 3 3 Equations The linear part of the logistic regression
More informationModeling Prediction of the Nosocomial Pneumonia with a Multistate model
Modeling Prediction of the Nosocomial Pneumonia with a Multistate model M.Nguile Makao 1 PHD student Director: J.F. Timsit 2 Co-Directors: B Liquet 3 & J.F. Coeurjolly 4 1 Team 11 Inserm U823-Joseph Fourier
More informationESP 178 Applied Research Methods. 2/23: Quantitative Analysis
ESP 178 Applied Research Methods 2/23: Quantitative Analysis Data Preparation Data coding create codebook that defines each variable, its response scale, how it was coded Data entry for mail surveys and
More information3 Joint Distributions 71
2.2.3 The Normal Distribution 54 2.2.4 The Beta Density 58 2.3 Functions of a Random Variable 58 2.4 Concluding Remarks 64 2.5 Problems 64 3 Joint Distributions 71 3.1 Introduction 71 3.2 Discrete Random
More informationAPPENDIX B Sample-Size Calculation Methods: Classical Design
APPENDIX B Sample-Size Calculation Methods: Classical Design One/Paired - Sample Hypothesis Test for the Mean Sign test for median difference for a paired sample Wilcoxon signed - rank test for one or
More informationMultivariable Fractional Polynomials
Multivariable Fractional Polynomials Axel Benner September 7, 2015 Contents 1 Introduction 1 2 Inventory of functions 1 3 Usage in R 2 3.1 Model selection........................................ 3 4 Example
More informationTesting Independence
Testing Independence Dipankar Bandyopadhyay Department of Biostatistics, Virginia Commonwealth University BIOS 625: Categorical Data & GLM 1/50 Testing Independence Previously, we looked at RR = OR = 1
More informationStatistics Handbook. All statistical tables were computed by the author.
Statistics Handbook Contents Page Wilcoxon rank-sum test (Mann-Whitney equivalent) Wilcoxon matched-pairs test 3 Normal Distribution 4 Z-test Related samples t-test 5 Unrelated samples t-test 6 Variance
More informationOnline supplement. Absolute Value of Lung Function (FEV 1 or FVC) Explains the Sex Difference in. Breathlessness in the General Population
Online supplement Absolute Value of Lung Function (FEV 1 or FVC) Explains the Sex Difference in Breathlessness in the General Population Table S1. Comparison between patients who were excluded or included
More informationECLT 5810 Linear Regression and Logistic Regression for Classification. Prof. Wai Lam
ECLT 5810 Linear Regression and Logistic Regression for Classification Prof. Wai Lam Linear Regression Models Least Squares Input vectors is an attribute / feature / predictor (independent variable) The
More informationBIOL 51A - Biostatistics 1 1. Lecture 1: Intro to Biostatistics. Smoking: hazardous? FEV (l) Smoke
BIOL 51A - Biostatistics 1 1 Lecture 1: Intro to Biostatistics Smoking: hazardous? FEV (l) 1 2 3 4 5 No Yes Smoke BIOL 51A - Biostatistics 1 2 Box Plot a.k.a box-and-whisker diagram or candlestick chart
More informationHypothesis Testing, Power, Sample Size and Confidence Intervals (Part 2)
Hypothesis Testing, Power, Sample Size and Confidence Intervals (Part 2) B.H. Robbins Scholars Series June 23, 2010 1 / 29 Outline Z-test χ 2 -test Confidence Interval Sample size and power Relative effect
More informationWheel for assessing spinal block study
Wheel for assessing spinal block study Xue Han, xue.han@vanderbilt.edu Matt Shotwell, matt.shotwell@vanderbilt.edu Department of Biostatistics Vanderbilt University December 13, 2012 Contents 1 Preliminary
More informationMissing Covariate Data in Matched Case-Control Studies
Missing Covariate Data in Matched Case-Control Studies Department of Statistics North Carolina State University Paul Rathouz Dept. of Health Studies U. of Chicago prathouz@health.bsd.uchicago.edu with
More informationBasic Medical Statistics Course
Basic Medical Statistics Course S7 Logistic Regression November 2015 Wilma Heemsbergen w.heemsbergen@nki.nl Logistic Regression The concept of a relationship between the distribution of a dependent variable
More informationTABLE OF CONTENTS CHAPTER 1 COMBINATORIAL PROBABILITY 1
TABLE OF CONTENTS CHAPTER 1 COMBINATORIAL PROBABILITY 1 1.1 The Probability Model...1 1.2 Finite Discrete Models with Equally Likely Outcomes...5 1.2.1 Tree Diagrams...6 1.2.2 The Multiplication Principle...8
More informationHomework Solutions Applied Logistic Regression
Homework Solutions Applied Logistic Regression WEEK 6 Exercise 1 From the ICU data, use as the outcome variable vital status (STA) and CPR prior to ICU admission (CPR) as a covariate. (a) Demonstrate that
More informationHow to Present Results of Regression Models to Clinicians
How to Present Results of Regression Models to Clinicians Frank E Harrell Jr Department of Biostatistics Vanderbilt University School of Medicine f.harrell@vanderbilt.edu biostat.mc.vanderbilt.edu/fhhandouts
More informationComparison of Hazard, Odds and Risk Ratio in the Two-Sample Survival Problem
Western Michigan University ScholarWorks at WMU Dissertations Graduate College 8-2014 Comparison of Hazard, Odds and Risk Ratio in the Two-Sample Survival Problem Benedict P. Dormitorio Western Michigan
More informationBooklet of Code and Output for STAD29/STA 1007 Midterm Exam
Booklet of Code and Output for STAD29/STA 1007 Midterm Exam List of Figures in this document by page: List of Figures 1 NBA attendance data........................ 2 2 Regression model for NBA attendances...............
More informationInstrumental variables estimation in the Cox Proportional Hazard regression model
Instrumental variables estimation in the Cox Proportional Hazard regression model James O Malley, Ph.D. Department of Biomedical Data Science The Dartmouth Institute for Health Policy and Clinical Practice
More informationMulti-level Models: Idea
Review of 140.656 Review Introduction to multi-level models The two-stage normal-normal model Two-stage linear models with random effects Three-stage linear models Two-stage logistic regression with random
More informationOutline. Frailty modelling of Multivariate Survival Data. Clustered survival data. Clustered survival data
Outline Frailty modelling of Multivariate Survival Data Thomas Scheike ts@biostat.ku.dk Department of Biostatistics University of Copenhagen Marginal versus Frailty models. Two-stage frailty models: copula
More informationInstantaneous geometric rates via generalized linear models
Instantaneous geometric rates via generalized linear models Andrea Discacciati Matteo Bottai Unit of Biostatistics Karolinska Institutet Stockholm, Sweden andrea.discacciati@ki.se 1 September 2017 Outline
More informationCase-control studies
Matched and nested case-control studies Bendix Carstensen Steno Diabetes Center, Gentofte, Denmark b@bxc.dk http://bendixcarstensen.com Department of Biostatistics, University of Copenhagen, 8 November
More informationFitting Cox Regression Models
Department of Psychology and Human Development Vanderbilt University GCM, 2010 1 Introduction 2 3 4 Introduction The Partial Likelihood Method Implications and Consequences of the Cox Approach 5 Introduction
More informationLecture 9. Statistics Survival Analysis. Presented February 23, Dan Gillen Department of Statistics University of California, Irvine
Statistics 255 - Survival Analysis Presented February 23, 2016 Dan Gillen Department of Statistics University of California, Irvine 9.1 Survival analysis involves subjects moving through time Hazard may
More informationBMI 541/699 Lecture 22
BMI 541/699 Lecture 22 Where we are: 1. Introduction and Experimental Design 2. Exploratory Data Analysis 3. Probability 4. T-based methods for continous variables 5. Power and sample size for t-based
More informationSTAT 7030: Categorical Data Analysis
STAT 7030: Categorical Data Analysis 5. Logistic Regression Peng Zeng Department of Mathematics and Statistics Auburn University Fall 2012 Peng Zeng (Auburn University) STAT 7030 Lecture Notes Fall 2012
More informationA Handbook of Statistical Analyses Using R 2nd Edition. Brian S. Everitt and Torsten Hothorn
A Handbook of Statistical Analyses Using R 2nd Edition Brian S. Everitt and Torsten Hothorn CHAPTER 7 Logistic Regression and Generalised Linear Models: Blood Screening, Women s Role in Society, Colonic
More informationNATIONAL UNIVERSITY OF SINGAPORE EXAMINATION. ST3241 Categorical Data Analysis. (Semester II: ) April/May, 2011 Time Allowed : 2 Hours
NATIONAL UNIVERSITY OF SINGAPORE EXAMINATION Categorical Data Analysis (Semester II: 2010 2011) April/May, 2011 Time Allowed : 2 Hours Matriculation No: Seat No: Grade Table Question 1 2 3 4 5 6 Full marks
More informationStatistical. Psychology
SEVENTH у *i km m it* & П SB Й EDITION Statistical M e t h o d s for Psychology D a v i d C. Howell University of Vermont ; \ WADSWORTH f% CENGAGE Learning* Australia Biaall apan Korea Меяко Singapore
More informationComparison of Two Samples
2 Comparison of Two Samples 2.1 Introduction Problems of comparing two samples arise frequently in medicine, sociology, agriculture, engineering, and marketing. The data may have been generated by observation
More informationUnit 9: Inferences for Proportions and Count Data
Unit 9: Inferences for Proportions and Count Data Statistics 571: Statistical Methods Ramón V. León 1/15/008 Unit 9 - Stat 571 - Ramón V. León 1 Large Sample Confidence Interval for Proportion ( pˆ p)
More informationFunctional Regression Methods for Densely-Sampled. Biomarkers in the ICU. Jonathan E. Gellar
Functional Regression Methods for Densely-Sampled Biomarkers in the ICU by Jonathan E. Gellar A dissertation submitted to The Johns Hopkins University in conformity with the requirements for the degree
More informationGlossary for the Triola Statistics Series
Glossary for the Triola Statistics Series Absolute deviation The measure of variation equal to the sum of the deviations of each value from the mean, divided by the number of values Acceptance sampling
More informationChapter 4 Regression Models
23.August 2010 Chapter 4 Regression Models The target variable T denotes failure time We let x = (x (1),..., x (m) ) represent a vector of available covariates. Also called regression variables, regressors,
More informationLecture 1. Introduction Statistics Statistical Methods II. Presented January 8, 2018
Introduction Statistics 211 - Statistical Methods II Presented January 8, 2018 linear models Dan Gillen Department of Statistics University of California, Irvine 1.1 Logistics and Contact Information Lectures:
More informationNonparametric Methods
Nonparametric Methods Marc H. Mehlman marcmehlman@yahoo.com University of New Haven Nonparametric Methods, or Distribution Free Methods is for testing from a population without knowing anything about the
More informationLecture 01: Introduction
Lecture 01: Introduction Dipankar Bandyopadhyay, Ph.D. BMTRY 711: Analysis of Categorical Data Spring 2011 Division of Biostatistics and Epidemiology Medical University of South Carolina Lecture 01: Introduction
More informationData are sometimes not compatible with the assumptions of parametric statistical tests (i.e. t-test, regression, ANOVA)
BSTT523 Pagano & Gauvreau Chapter 13 1 Nonparametric Statistics Data are sometimes not compatible with the assumptions of parametric statistical tests (i.e. t-test, regression, ANOVA) In particular, data
More informationYou can specify the response in the form of a single variable or in the form of a ratio of two variables denoted events/trials.
The GENMOD Procedure MODEL Statement MODEL response = < effects > < /options > ; MODEL events/trials = < effects > < /options > ; You can specify the response in the form of a single variable or in the
More informationResiduals and regression diagnostics: focusing on logistic regression
Big-data Clinical Trial Column Page of 8 Residuals and regression diagnostics: focusing on logistic regression Zhongheng Zhang Department of Critical Care Medicine, Jinhua Municipal Central Hospital, Jinhua
More informationANALYSING BINARY DATA IN A REPEATED MEASUREMENTS SETTING USING SAS
Libraries 1997-9th Annual Conference Proceedings ANALYSING BINARY DATA IN A REPEATED MEASUREMENTS SETTING USING SAS Eleanor F. Allan Follow this and additional works at: http://newprairiepress.org/agstatconference
More informationClass: Dean Foster. September 30, Read sections: Examples chapter (chapter 3) Question today: Do prices go up faster than they go down?
Class: Dean Foster September 30, 2013 Administrivia Read sections: Examples chapter (chapter 3) Gas prices Question today: Do prices go up faster than they go down? Idea is that sellers watch spot price
More informationBinomial Model. Lecture 10: Introduction to Logistic Regression. Logistic Regression. Binomial Distribution. n independent trials
Lecture : Introduction to Logistic Regression Ani Manichaikul amanicha@jhsph.edu 2 May 27 Binomial Model n independent trials (e.g., coin tosses) p = probability of success on each trial (e.g., p =! =
More informationTMA 4275 Lifetime Analysis June 2004 Solution
TMA 4275 Lifetime Analysis June 2004 Solution Problem 1 a) Observation of the outcome is censored, if the time of the outcome is not known exactly and only the last time when it was observed being intact,
More informationTest of Association between Two Ordinal Variables while Adjusting for Covariates
Test of Association between Two Ordinal Variables while Adjusting for Covariates Chun Li, Bryan Shepherd Department of Biostatistics Vanderbilt University May 13, 2009 Examples Amblyopia http://www.medindia.net/
More informationQuestion. Hypothesis testing. Example. Answer: hypothesis. Test: true or not? Question. Average is not the mean! μ average. Random deviation or not?
Hypothesis testing Question Very frequently: what is the possible value of μ? Sample: we know only the average! μ average. Random deviation or not? Standard error: the measure of the random deviation.
More informationLecture 10: Introduction to Logistic Regression
Lecture 10: Introduction to Logistic Regression Ani Manichaikul amanicha@jhsph.edu 2 May 2007 Logistic Regression Regression for a response variable that follows a binomial distribution Recall the binomial
More informationOn a connection between the Bradley Terry model and the Cox proportional hazards model
Statistics & Probability Letters 76 (2006) 698 702 www.elsevier.com/locate/stapro On a connection between the Bradley Terry model and the Cox proportional hazards model Yuhua Su, Mai Zhou Department of
More informationGoodness-Of-Fit for Cox s Regression Model. Extensions of Cox s Regression Model. Survival Analysis Fall 2004, Copenhagen
Outline Cox s proportional hazards model. Goodness-of-fit tools More flexible models R-package timereg Forthcoming book, Martinussen and Scheike. 2/38 University of Copenhagen http://www.biostat.ku.dk
More informationAnalysis of repeated measurements (KLMED8008)
Analysis of repeated measurements (KLMED8008) Eirik Skogvoll, MD PhD Professor and Consultant Institute of Circulation and Medical Imaging Dept. of Anaesthesiology and Emergency Medicine 1 Day 2 Practical
More information