Вы находитесь на странице: 1из 10

Homework 4 (Due by 11:59 on 2/25/2014)

In Table D.3 of Appendix D, data is reported on the death times of 863 kidney transplant patients (see section 1.7). Here, patients can be classified by race and sex into one of four groups. Answer the following questions by both hand calculation and Proc lifetest.

(a) Test the hypothesis that blacks have a higher mortality rate than white without adjusting sex;

Code for Answer (a):

libname dog "C:\Users\andresg\Dropbox\USF\2014 Spring\Survival analysis\HW4";

/*to obtain Yi1 and di1*/ proc lifetest data=dog.sec1_7 atrisk; time time*delta(0); where race=1; run;

/*to obtain Yi2 and di2*/ proc lifetest data=dog.sec1_7 atrisk; time time*delta(0); where race=2; run;

/*to obtain Yi and di*/ proc lifetest data=dog.sec1_7 atrisk; time time*delta(0); run;

/* combine the three pieces of data above together */ data table7_2; input ti Yi1 di1 Yi2 di2 Yi di ; datalines;

2

151

0

710

1

861

1

3

151

0

709

1

860

1

7

151

0

706

2

857

2

10

151

0

702

2

853

2

17

150

0

698

1

848

1

21

150

0

694

1

844

1

26

150

0

693

2

843

2

28

150

0

690

1

840

1

37

150

1

686

0

836

1

40

149

1

686

0

835

1

43

148

1

686

1

834

2

44

147

0

682

1

829

1

45

147

1

681

0

828

1

50

146

0

680

1

826

1

52

146

0

675

1

821

1

56

146

0

674

1

820

1

57

146

1

673

1

819

2

59

145

0

672

1

817

1

62

145

0

671

2

816

2

68

145

0

665

2

810

2

69

145

0

663

1

808

1

78

145

0

662

1

807

1

79

145

0

661

1

806

1

88

143

0

655

1

798

1

91

143

0

654

1

797

1

97

141

0

652

1

793

1

98

141

0

651

2

792

2

104

141

0

649

1

790

1

106

141

1

645

1

786

2

119

135

0

638

1

773

1

121

134

1

637

0

771

1

135

133

0

634

1

767

1

143

133

0

631

1

764

1

150

133

0

629

1

762

1

154

132

0

626

1

758

1

158

132

1

625

0

757

1

162

131

0

625

1

756

1

190

130

0

616

1

746

1

206

129

1

610

0

739

1

209

128

0

610

1

738

1

228

127

0

605

1

732

1

229

127

1

604

0

731

1

242

125

0

599

1

724

1

248

125

0

595

1

720

1

249

125

0

594

1

719

1

252

124

0

591

1

715

1

273

122

0

579

1

701

1

291

121

0

573

1

694

1

297

121

0

572

1

693

1

311

120

1

570

1

690

2

334

117

0

564

1

681

1

340

117

0

561

1

678

1

344

117

1

556

0

673

1

346

116

0

556

1

672

1

354

116

0

553

1

669

1

366

116

0

551

1

667

1

391

116

0

543

1

659

1

402

115

1

540

0

655

1

421

113

0

536

1

649

1

439

113

0

530

1

643

1

450

111

1

526

0

637

1

470

109

0

524

1

633

1

478

109

0

523

1

632

1

481

108

0

522

1

630

1

490

108

0

519

1

627

1

495

108

0

516

1

624

1

570

105

0

504

1

609

1

583

105

0

499

1

604

1

614

105

0

494

1

599

1

615

105

0

493

1

598

1

621

105

0

492

1

597

1

652

102

0

487

1

589

1

697

97

0

473

1

570

1

730

97

0

467

1

564

1

773

95

0

455

1

550

1

776

95

0

454

1

549

1

790

93

0

451

1

544

1

793

93

0

450

1

543

1

806

93

0

449

1

542

1

840

93

0

446

1

539

1

852

93

0

445

1

538

1

864

91

1

444

0

535

1

875

88

0

442

1

530

1

929

82

1

431

0

513

1

939

80

0

430

1

510

1

943

80

1

429

0

509

1

945

79

0

429

2

508

2

946

79

0

427

1

506

1

1001

75

1

419

0

494

1

1013

73

0

418

1

491

1

1016

73

1

417

0

490

1

1105

72

0

407

1

479

1

1164

67

0

390

1

457

1

1186

66

0

388

1

454

1

1191

66

0

387

1

453

1

1196

66

1

386

0

452

1

1210

65

0

382

1

447

1

1275

62

1

368

0

430

1

1326

59

0

356

1

415

1

1331

59

0

355

1

414

1

1357

58

0

352

1

410

1

1384

58

1

344

0

402

1

1388

57

0

343

1

400

1

1418

57

0

341

1

398

1

1473

55

0

328

1

383

1

1509

54

0

323

1

377

1

1734

48

0

281

1

329

1

1777

46

0

267

1

313

1

1820

43

0

252

1

295

1

1835

43

0

251

1

294

1

1877

43

0

246

1

289

1

1940

42

0

238

1

280

1

2034

39

0

215

1

254

1

2056

38

0

204

1

242

1

2108

36

0

191

1

227

1

2171

36

1

185

0

221

1

2276

33

1

167

0

200

1

2291

32

0

167

1

199

1

2301

31

0

165

1

196

1

2313

31

0

163

1

194

1

2369

31

1

152

0

183

1

2414

28

1

148

0

176

1

2421

27

0

147

1

174

1

2489

22

0

134

1

156

1

2557

22

1

124

0

146

1

2567

21

0

124

1

145

1

2650

19

1

112

0

131

1

2795

12

0

85

1

97

1

3146

6

0

33

1

39

1

;

run;

data table7_2; set table7_2;

col8=Yi1*(di/Yi);

col9=di1-Yi1*(di/Yi);

col10=di*(yi1/Yi)*(1-Yi1/Yi)*((Yi-di)/(Yi-1));

run;

proc means data=table7_2 sum; var col9 col10; run;

/* the chi-square value from the hand calculation above can be obtained by proc lifetest*/ proc lifetest data=dog.sec1_7 ; time time*delta(0); strata race; run;

Answer (a)

Answer (a) Rank Statistics race Log-Rank Wilcoxon 1 -4.6468 -1285.0 2 4.6468 1285.0 Covariance

Rank Statistics

race

Log-Rank Wilcoxon

1

-4.6468

-1285.0

2

4.6468

1285.0

Covariance Matrix for the Log-Rank Statistics

race

1

2

1

19.4303 -19.4303

2

-19.4303

19.4303

Variable

Sum

col9

4.6468274

col10

19.4303000

So the test statistic will be:

4.64

2 = √19.43 = 1.0526

Which represent a P-value greater than 0.2 (SAS= 0.2918) which for a level of significance of 5% provides a high probability of reject the null hypothesis being this true, so we end up accepting the null hypothesis. Then we conclude that there is no significant difference between the races for the time to event.

(b) Test the hypothesis that blacks have a higher mortality rate than white without adjusting sex;

Code for answer (b)

data Male;

set dog.sec1_7;

if gender=1;

run;

/*to obtain Yi1 and di1*/ proc lifetest data=Male atrisk ; time time*delta(0); where race=1; run;

/*to obtain Yi2 and di2*/ proc lifetest data=Male atrisk; time time*delta(0); where race=2; run;

/*to obtain Yi and di*/ proc lifetest data=Male atrisk; time time*delta(0); run;

/* combine the three pieces of data above together */ data example7_8; input ti Yi1 di1 Yi2 di2 Yi di ; datalines;

7

92

0

430

1

522

1

17

92

0

426

1

518

1

26

92

0

424

2

516

2

28

92

0

422

1

514

1

37

92

1

419

0

511

1

43

91

1

419

1

510

2

44

90

0

417

1

507

1

56

90

0

413

1

503

1

57

90

1

412

1

502

2

59

89

0

411

1

500

1

62

89

0

410

1

499

1

68

89

0

406

1

495

1

69

89

0

405

1

494

1

79

89

0

404

1

493

1

88

87

0

400

1

487

1

91

87

0

399

1

486

1

98

86

0

397

2

483

2

106

86

0

393

1

479

1

119

83

0

388

1

471

1

135

82

0

385

1

467

1

150

82

0

383

1

465

1

158

81

1

381

0

462

1

162

80

0

381

1

461

1

190

79

0

376

1

455

1

206

78

1

372

0

450

1

228

76

0

368

1

444

1

242

75

0

365

1

440

1

248

75

0

363

1

438

1

249

75

0

362

1

437

1

252

75

0

360

1

435

1

291

74

0

348

1

422

1

311

74

1

347

1

421

2

334

72

0

342

1

414

1

340

72

0

340

1

412

1

346

72

0

337

1

409

1

354

72

0

335

1

407

1

391

72

0

329

1

401

1

402

72

1

327

0

399

1

421

70

0

324

1

394

1

439

70

0

320

1

390

1

450

69

1

319

0

388

1

478

67

0

318

1

385

1

481

66

0

317

1

383

1

495

66

0

312

1

378

1

570

64

0

303

1

367

1

583

64

0

298

1

362

1

615

64

0

295

1

359

1

621

64

0

294

1

358

1

652

62

0

290

1

352

1

697

58

0

283

1

341

1

730

58

0

278

1

336

1

773

57

0

274

1

331

1

776

57

0

273

1

330

1

790

56

0

272

1

328

1

806

56

0

271

1

327

1

875

52

0

266

1

318

1

939

51

0

259

1

310

1

945

51

0

258

2

309

2

946

51

0

256

1

307

1

1001

49

1

251

0

300

1

1105

47

0

244

1

291

1

1186

42

0

228

1

270

1

1191

42

0

227

1

269

1

1210

42

0

223

1

265

1

1275

39

1

215

0

254

1

1357

37

0

208

1

245

1

1384

37

1

203

0

240

1

1388

36

0

202

1

238

1

1418

36

0

200

1

236

1

1509

35

0

187

1

222

1

1734

31

0

159

1

190

1

1820

28

0

140

1

168

1

2056

25

0

113

1

138

1

2291

22

0

91

1

113

1

2313

21

0

89

1

110

1

2369

21

1

81

0

102

1

2414

19

1

80

0

99

1

2421

18

0

79

1

97

1

2489

15

0

71

1

86

1

2557

15

1

66

0

81

1

3146

3

0

17

1

20

1

;

run;

data example7_8; set example7_8;

col8=Yi1*(di/Yi);

col9=di1-Yi1*(di/Yi);

col10=di*(yi1/Yi)*(1-Yi1/Yi)*((Yi-di)/(Yi-1));

run; proc means data=example7_8 sum; var col9 col10; run;

data Female;

set dog.sec1_7;

if Gender=2;

run;

/*to obtain Yi1 and di1*/ proc lifetest data=Female atrisk ; time time*delta(0); where race=1; run;

/*to obtain Yi2 and di2*/ proc lifetest data=female atrisk; time time*delta(0); where race=2; run;

/*to obtain Yi and di*/ proc lifetest data=female atrisk; time time*delta(0); run;

/* combine the three pieces of data above together */ data example7_8; input ti Yi1 di1 Yi2 di2 Yi di ; datalines;

2

59

0

279

1

338

1

3

59

0

278

1

337

1

7

59

0

276

1

335

1

10

59

0

274

2

333

2

21

58

0

270

1

328

1

40

58

1

267

0

325

1

45

57

1

265

0

322

1

50

56

0

264

1

320

1

52

56

0

262

1

318

1

62

56

0

261

1

317

1

68

56

0

259

1

315

1

78

56

0

258

1

314

1

97

55

0

255

1

310

1

104

55

0

254

1

309

1

106

55

1

252

0

307

1

121

52

1

250

0

302

1

143

51

0

247

1

298

1

154

51

0

245

1

296

1

209

51

0

238

1

289

1

229

51

1

237

0

288

1

273

47

0

227

1

274

1

297

47

0

225

1

272

1

344

45

1

219

0

264

1

366

44

0

218

1

262

1

470

42

0

206

1

248

1

490

42

0

205

1

247

1

614

41

0

199

1

240

1

793

37

0

179

1

216

1

840

37

0

178

1

215

1

852

37

0

177

1

214

1

864

37

1

176

0

213

1

929

30

1

171

0

201

1

943

29

1

171

0

200

1

1013

26

0

167

1

193