Development and Validation of an Inflammatory Bowel ...download.xuebalib.com/2sn1bVMjztGa.pdf · Q1 M. G. H. van Oijen,*,jj and D. W ... from April 2014 through March 2015. RESULTS:

Q12

Q1

Clinical Gastroenterology and Hepatology 2015;-:-–-

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

Development and Validation of an Inflammatory BowelDiseases Monitoring Index for Use With MobileHealth Technologies

68

69

70

71

72

73

74

75

76

Welmoed K. Van Deen,*,‡ A. E. van der Meulen-de Jong,‡ N. K. Parekh,§ E. Kane,*A. Zand,* C. A. DiNicola,* L. Hall,* E. K. Inserra,* J. M. Choi,* Ch. Y. Ha,* E. Esrailian,*M. G. H. van Oijen,*,jj and D. W. Hommes*

*UCLA Center for Inflammatory Bowel Diseases, Melvin and Bren Simon Digestive Diseases Center, David Geffen Schoolof Medicine, University of California, Los Angeles, Los Angeles, California; ‡Department of Gastroenterology andHepatology, Leiden University Medical Center, Leiden, the Netherlands; §Division of Gastroenterology andHepatology, University of California, Irvine, Irvine, California; and jjDepartment of Medical Oncology, Academic MedicalCenter, Amsterdam, the Netherlands

77

78
BACKGROUND & AIMS:
Abbreviations used in this papedisease; CDAI, Crohn’s diseasvisual analogue scale; HBI, HaIBD, inflammatory bowel diseCenter; mHealth, mobile healthpredictive value; pMayo, partivalue; PROs, patient-reportedacteristics; SCCAI, simple clin

F

79

80

81

82

Mobile health technologies are advancing rapidly as smartphone use increases. Patients withinflammatory bowel disease (IBD) might be managed remotely through smartphone applica-tions, but no tools are yet available. We tested the ability of an IBD monitoring tool, which canbe used with mobile technologies, to assess disease activity in patients with Crohn’s disease(CD) or ulcerative colitis (UC).

83

84
METHODS:
85

86

87

88

89

90

We performed a prospective observational study to develop and validate a mobile health indexfor CD and UC, which monitors IBD disease activity using patient-reported outcomes. Wecollected data from disease-specific questionnaires completed by 110 patients with CD and 109with UC who visited the University of California, Los Angeles, Center for IBD from May 2013through January 2014. Patient-reported outcomes were compared with clinical disease activityindex scores to identify factors associated with disease activity. Index scores were validated in301 patients with CD and 265 with UC who visited 3 tertiary IBD referral centers (in Californiaor Europe) from April 2014 through March 2015.

91

92
RESULTS: 93
94

95

96

97

98

99

We assessed activity of CD based on liquid stool frequency, abdominal pain, patient well-being,and patient-assessed disease control, and activity of UC based on stool frequency, abdominalpain, rectal bleeding, and patient-assessed disease control. The indices identified clinical dis-ease activity with area under the receiver operating characteristic curve values of 0.90 in pa-tients with CD and 0.91 in patients with UC. They identified endoscopic activity with area underthe receiver operating characteristic values of 0.63 in patients with CD and 0.82 in patients withUC. Both scoring systems responded to changes in disease activity (P < .003). The intraclasscorrelation coefficient for test-retest reliability was 0.94 for CD and for UC.

100

101
CONCLUSIONS:
102

103

We developed and validated a scoring system to monitor disease activity in patients with CDand UC that can be used with mobile technologies. The indices identified clinical disease activitywith area under the receiver operating characteristic curve values of 0.9 or higher in patientswith CD or UC, and endoscopic activity in patients with UC but not CD.

104

105
Keywords: Patient-Reported Outcomes; Telemedicine; Severity; Management.
106

107

r: AUC, area under the curve; CD, Crohn’se activity index; DA-VAS, disease activityrvey Bradshaw index; Hgb, hemoglobin;ase; LUMC, Leiden University Medical; mHI, mobile health index; NPV, negativeal Mayo score; PPV, positive predictiveoutcomes; ROC, receiver operating char-ical colitis activity index; SES-CD, simple

endoscopic index for Crohn’s disease; UC, ulcerative colitis; UCI,University of California, Irvine; UCLA, University of California, Los Angeles;VAS, visual analogue scale.

© 2015 by the AGA Institute1542-3565/$36.00

http://dx.doi.org/10.1016/j.cgh.2015.10.035

LA 5.4.0 DTD � YJCGH54544_proof � 11 December 2015 � 12:31 pm � ce CJ

108

109

110

111

112

113

114

115

116


Q5 Q6

Q7

Q8

2 Van Deen et al Clinical Gastroenterology and Hepatology Vol. -, No. -

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

The shift from symptom-oriented to prevention-oriented care delivery has accelerated the devel-

opment of mobile health (mHealth) technologies and isthought to radically transform health care delivery.1

Smartphone adoption is increasing rapidly, with 64%of Americans using smartphones in 2014 of which 62%used their telephones to look up health information.2

Many health apps are available, most of which providehealth information or support data collection.3 For pa-tients with inflammatory bowel diseases (IBD), appsare available that assist in tracking symptoms, loggingmeals, and managing medications.4,5 These apps cancreate reports for providers but do not allow for real-time interactions between patient and provider.

Self-monitoring and self-management for chronicdiseases is widely practiced in diabetes care6 and anti-coagulation therapy.7 Additionally, e-technologies forsymptom reporting between patients and providers areincreasingly used in chronic diseases.8 Several systemsfor online symptom reporting and disease managementhave been developed for IBD. In the Danish Constant-Care web system, patients with ulcerative colitis (UC)filled out a clinical symptom score and logged fecal cal-protectin levels weekly; based on these scores the sys-tem made real-time recommendations for adjustingmesalamine dosing. This approach was shown toempower patients and decrease relapse duration.9,10

Similarly, individualization of infliximab dosing in pa-tients with Crohn’s disease (CD) was reported to bepractical and feasible.11 A study evaluating another hometele-management system in UC (UC-HAT) did not showsignificant differences in disease activity and quality oflife between users and control subjects, and more thanone-third of the patients discontinued participation.12 Anongoing multicenter randomized controlled trial istesting the use of a mobile tele-management system us-ing text messaging in IBD. This system sends personal-ized alerts and educational texts, and assesses symptomsand side effects, based on which treatment can bemodified.13

Patient-reported outcomes (PROs) are increasinglyused to evaluate health status, and the importance ofPROs in outcome measurement, symptom management,and quality improvement efforts is increasingly recog-nized.14 Furthermore, the use of PROs as primaryoutcome measures for evaluating effectiveness of IBDinterventions is progressively supported by the Food andDrug Administration.15 Therefore, PROs are promisingfor use in mHealth apps. An example is the Health-PROMISE app, which tracks patient-reported quality oflife scores in patients with IBD and provides decisionsupport to physicians.16 However, accurate e-monitoringtools for disease activity in IBD are yet to be developed.Previous efforts have aimed to develop PRO question-naires by adjusting existing scores.17–19 We aimed toidentify the most optimal PRO score to use on an IBDdisease-monitoring app. The best PROs were selected

FLA 5.4.0 DTD � YJCGH54544_proof � 1

from an exhaustive list of PROs in a prospective cohort ofpatients with IBD. Subsequently, the developed scoreswere tested prospectively in an independent cohort at 3independent IBD centers.

Methods

Design

We performed a prospective, observational study thataimed to develop and validate an mHealth index (mHI)for CD and UC that accurately monitors IBD diseaseactivity using PROs. The study consisted of 2 phases: adevelopment phase and a validation phase. During thedevelopment phase the mHIs were developed usingcollected PROs and clinical disease activity indices.During the validation phase the developed mHIs werevalidated in an independent cohort.

Population

Development phase. Patients with IBD were identi-fied during clinic visits between May 2013 and January2014 at the University of California, Los Angeles (UCLA)Center for IBD. Patients with esophageal or anal CDinvolvement alone, patients with a pouch or stoma, andpregnant women were excluded. Eligible patients filledout disease-specific questionnaires assessing PROs ofmost common clinical disease activity indices: partialMayo (pMayo), simple clinical colitis activity index(SCCAI), and modified Truelove and Witts index for UC;Harvey Bradshaw index (HBI) and CD activity index(CDAI) including a 7-day diary before the visit(Supplementary Figure 1) for CD. Additionally, patientswere asked to assess their symptoms and perceiveddisease activity using visual analogue scales (VAS). ThePROs were categorized into 10 domains: stool fre-quency, abdominal pain, general well-being, urgency,stool consistency, rectal bleeding, fever, anorexia,nausea/vomiting, and perceived disease activity(Table 1).

During clinic visits, vital signs were measured andphysicians collected the physician-reported outcomesrequired for the clinical disease activity indices(Table 1). Hemoglobin (Hgb), hematocrit, white bloodcell count, platelets, albumin, C-reactive protein, anderythrocyte sedimentation rate were requested.Furthermore, stool testing for calprotectin was reques-ted either at the patient’s preferred laboratory or usinga free stool kit (Genova Diagnostics, Asheville, NC)picked up at the patient’s home. A dedicated studynurse (E.K.I.) followed up with patients via telephone ore-mail to ensure laboratory and stool tests wereperformed.

Validation phase. Eligible patients with IBD wereidentified during clinic and endoscopy visits between

1 December 2015 � 12:31 pm � ce CJ

232

Q9

Table 1. Collected PROs and Clinical Markers in CD and Patients With UC

Domain Question CD UC

1. Stool frequency Number of liquid/very soft stools for each of the last 7 days XHow many stools did you have yesterday during the day? X XHow many stools did you have yesterday during the night? X XHow many stools do you have normally during the day? X XHow many stools do you have normally during the night? X X

2. Abdominal pain Abdominal pain for each of the last 7 days (no/mild/moderate/severe) XAbdominal pain (no/mild/moderate/severe) XAbdominal pain (no abdominal pain/with bowel actions/continuous) X XRate your abdominal pain on a scale from 0 to 10 (VAS) X X

3. General well-being General well-being for each of the last 7 days (very well/slightly below par/poor/very poor/terrible)

X

General well-being (very well/slightly below par/poor/very poor/terrible) XGeneral well-being (perfect/very good/good/average/poor/terrible) XWell-being (no impairment/impaired, but able to continue activities/activities reduced/unable

to work)X X

Rate your well-being on a scale from 0 to 10 (VAS) X X4. Urgency Urgency of defecation (no urgency/hurry/immediate/incontinence) X X5. Stool consistency Stool consistency (normal or variably normal/semiformed/liquid) X X

Do you take opiates or lomotil/imodium for diarrhea? X XHow often do you take antidiarrheals? (0–10 VAS) X X

6. Rectal bleeding What % of bowel movements contains visible blood? (none/less than 50%/50% or more/blood alone)

X X

Amount of blood in stool (none/trace/occasionally frank [bright red]/usually frank) X XHow often do you experience rectal bleeding? (0–10 VAS) X X

7. Fever Fever on each of the last 7 days X8. Anorexia Loss of appetite (yes/no) X X9. Nausea/vomiting Nausea and/or vomiting (yes/no) X X10. Disease activity How well do you feel your disease is under control? (0–10 VAS) X X11. Clinical markers Temperature X X

Weight and height XPulse XAbdominal tenderness XAbdominal mass XExtraintestinal manifestations X XPhysician global assessment of disease activity X XHgb, hematocrit, white blood cell, platelets, albumin, C-reactive protein, erythrocyte

sedimentation rate (blood) and calprotectin (stool)X X

- 2015 mHealth Index for IBD Monitoring 3

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

April 2014 and March 2015 in 3 tertiary IBD referralcenters (UCLA; University of California, Irvine [UCI]; andLeiden University Medical Center [LUMC], theNetherlands). Patients who participated in the devel-opment phase of the study were excluded. For patientswith CD, the developed mHI-CD and HBI werecompleted during clinic visits; during endoscopic visits,the simple endoscopic score for CD (SES-CD) was addi-tionally completed. For patients with UC, the mHI-UCand pMayo were collected during clinic visits, and theMayo endoscopic subscore was additionally obtainedduring endoscopic visits. Patients at LUMC completed aDutch version of the mHI-CD and mHI-UC. After trans-lation to Dutch, the questionnaires were translated backto English by an independent translator; the Dutchquestionnaire was then revised and retested. To assesssensitivity of the mHI to detect changes in clinical dis-ease activity, a subset of patients was included a secondtime during scheduled follow-up visits. To assess test-test reliability, a subset of UCLA patients was asked to


complete a second questionnaire at home after theirclinic visit.

Definitions

For CD, clinical disease activity was defined as an HBI>4 or a CDAI >150. A change of �3 in HBI wasconsidered a clinically relevant change.20 Endoscopicdisease activity was defined as an SES-CD >3. For UC,clinical disease activity was defined as a pMayo >2, a PTI>3, or an SCCAI >2. A change of �3 in the pMayo wasconsidered a clinically relevant change.18 Endoscopicdisease activity was defined as a Mayo endoscopic sub-score >1.

Ethical Considerations

All patients consented to participate in this study.This study was approved by the institutional review



349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

boards of participating centers under the followingprotocol numbers: UCLA, IRB#13-000402; UCI, HS#2014-1231; and LUMC, P14.158.

Statistical Analysis

Descriptive statistics were used for clinical charac-teristics and demographic information. Numeric valuesare presented as mean and standard deviation or medianand range. SAS version 9.2 (SAS Institute, Cary, NC) wasused for statistical analyses.

Development phase. Univariate logistic regressionwas performed using disease activity (HBI >4 for CD orpMayo >2 for UC) as the dependent variable and thePROs as independent variables. For each of the PROs,different cutoffs were used, which roughly created linearassociations between the groups and the chance of activedisease. Because different PROs represented the samedomain (Table 1), the variables with the highest Waldchi-square value for predicting clinical disease activitywere selected within each domain for inclusion in themultivariate logistic regression models. If 2 variableswithin the same domain had a similar predictive value(difference between Wald chi-square values <0.5), bothwere tried in separate models unless the question typewas less preferable. Because of usability on a mobileapplication VAS, yes/no, or numeric questions werepreferred over categorical questions; within those,questions with less response options were preferred.Variables with a P value >.1 in the univariate analysiswere omitted from subsequent analyses. Stepwise for-ward multivariate logistic regression was performedwith clinical disease activity as the dependent variableand the selected PROs as independent variables. A sig-nificance level of P < .1 was required for entry in themodel and a significance level of P < .1 to stay in themodel. Several models were performed using differentclinical disease activity indices to define clinical diseaseactivity as the dependent variable.

Composite scores were created using the regressioncoefficients of independent predictors in the multivariatemodel. Spearman correlation coefficients were estimatedbetween the newly developed mHIs and clinical diseaseactivity indices. Receiver operating characteristic (ROC)curves were used to assess the capability of the mHI todiscriminate active versus nonactive disease usingdifferent clinical disease activity indices, and the areasunder the curves (AUC) were calculated. The compositescore with overall highest AUC using different goldstandards was selected.

Because the main aim of the developed score was toidentify patients at risk for active disease, we defined theoptimal cutoff for disease activity as a negative predic-tive value (NPV) of �95% and a sensitivity of �85%while maintaining maximum specificity. The overallprevalence of active disease was estimated at 22% basedon cross-sectional cohort data from UCLA.


Validation phase. To validate the mHI against clinicaland endoscopic disease activity indices, the mHI-UC wascompared with the pMayo and the endoscopic subscoreof the Mayo; the mHI-CD was compared with the HBI andSES-CD. Spearman correlation coefficients between thescores were calculated and ROC curves to assess theability to predict clinical and endoscopic disease activitywere generated. To assess sensitivity to change, wecompared patients who clinically improved, remainedstable, and deteriorated. A Kruskal-Wallis test was usedto compare groups. Test-retest reliability was assessedby the intraclass correlation coefficient. The performanceof the VAS for patient-reported disease activity (DA-VAS)as single predictor for clinical and endoscopic diseaseactivity also was assessed.

Results

Development Phase

In total, 219 patients (110 CD and 109 UC) wereincluded in the development phase of the study(Figure 1A, Table 2). In 108 out of 110 patients with CDthe HBI was calculated, whereas the CDAI could only becalculated in 93 out of 110. The pMayo, SCCAI, andmodified Truelove and Witts index were calculated in allpatients with UC. Complete laboratory and stool testswere obtained from only 48% of patients. Despiteintensive follow-up by a dedicated research nurse(E.K.I.), 39% of patients did not perform stool testing and13% did not have laboratory values drawn. Additionally,14% of patients had blood drawn, but laboratory setswere incomplete because of protocol deviations.

Univariate logistic regression was performed forPROs and blood and stool tests (Supplementary Tables 1and 2). Stool frequency, abdominal pain, general well-being, urgency, and patient-reported disease activitywere all strong predictors for clinical disease activity inboth CD and UC (P < .0001). Incontinence was onlypresent in 3% of patients and did not predict diseaseactivity in either CD (P ¼ .54) or UC (P ¼ .99). In CD theuse of antidiarrheals was predictive for disease activity(P ¼ .019) but not in UC (P ¼ .96), whereas the VASassessing frequency of antidiarrheal use was a predictorfor neither CD (P ¼ 1.00) nor UC (P ¼ .26). Rectalbleeding was a predictor for disease activity in both UC(P < .0001) and CD (P ¼ .019). Anorexia was predictivein both CD (P ¼ .019) and UC (P ¼ .0025), whereasnausea and vomiting predicted only CD disease activity(P ¼ .035) and not UC disease activity (P ¼ .14). High C-reactive protein (P ¼ .0009), high erythrocyte sedimen-tation rate (P ¼ .0022), low Hgb (P ¼ .022), and lowalbumin (0.034) were predictors for clinical disease ac-tivity in CD. High calprotectin was not a significant pre-dictor for CD disease activity (P ¼ .054), althoughcalprotectin as continuous variable had predictive value


464

Figure 1. Inclusion flow-charts for developmentphase (A) and validationphase (B).


465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

552

553

554

555

556

557

558

559

560

561

562

563

564

565

566

567

568

569

570

571

572

573

574

575

576

577

578

579

(P ¼ .011). Low platelets (P ¼ .98), low hematocrit (P ¼.28), and high white blood cell (P ¼ .11) were not pre-dictive in CD (Supplementary Table 1). In UC highC-reactive protein (P ¼ .0067), high calprotectin (P ¼.022), high white blood cell (0.013), high erythrocytesedimentation rate (P ¼ .028), and low hematocrit (P ¼.0047) were all predictive for clinical disease activity.Low albumin (P ¼ .98) was not predictive for clinicalactivity in UC, although albumin as continuous variablewas (P ¼ .02). Low platelets (P ¼ .99) and low Hgb (P ¼.13) were not predictive for disease activity in UC,whereas Hgb as continuous variable (P ¼ .032) was(Supplementary Table 2).


The most representative PROs were selected for in-clusion in the multivariate regression model(Supplementary Tables 1 and 2). Because of lowcompletion rates of laboratory testing despite intensivefollow-up, laboratory tests were initially excluded fromthe multivariate analysis and score development. In CD, 4composite scores were developed (SupplementaryTable 3); in UC, 11 composite scores were evaluated(Supplementary Table 4). Addition of laboratory vari-ables to the selected models decreased the sample sizeand therefore the power of the regression models;addition of the laboratory variables to the model did notresult in inclusion of these variables in the models,


580

Table 2. Demographics of Included Patients in Development and Validation Phase of the Study

Development phase Validation phase

CD UC CD UC

N 110 109 301 265Median age, median (range) 33 (19–79) 35 (18–81) 33 (18–75) 42 (18–86)Male, n (%) 56 (51) 57 (52) 144 (48) 132 (50)a

Smoking, n (%) 9 (8) 7 (6)a 19 (6)b 12 (5)c

Age at diagnosis, median (range) 24 (8–68) 28 (10–81) 25 (8–66) 29 (2–76)Disease duration, median (range) 8 (0–46) 6 (0–52) 8 (0–52) 7 (0–59)Surgical history, n (%) 48 (44) 1 (1) 132 (44)d 13 (5)d

CD fistulizing disease, n (%) 5 (5)a — 59 (20)a —

Active disease (HBI >4 /pMayo >2), n (%) 32 (30)a 37 (34) 82 (27) 82 (31)

an ¼ 1 unknown.bn ¼ 8 unknown.cn ¼ 9 unknown.dn ¼ 2 unknown.


581

582

583

584

585

586

587

588

589

590

591

592

593

594

595

596

597

598

599

600

601

602

603

604

605

606

607

608

609

610

611

612

613

614

615

616

617

618

619

620

621

622

623

624

625

626

627

628

629

630

631

632

633

634

635

636

637

638

639

640

641

642

643

644

645

646

647

648

649

650

651

652

653

654

655

656

657

658

659

660

661

662

663

664

665

666

667

668

669

670

671

672

673

674

675

676

677

678

because they did not reach the required significance levelof P < .1 for entry in the model.

For CD the selected composite score (Table 3) had anAUC of >0.95 for predicting clinical disease activity,using both CDAI and HBI as gold standards (0.951 and0.963, respectively). The Spearman correlation co-efficients were 0.837 and 0.830, respectively(Supplementary Table 3). The optimal cutoff for themHI-CD to predict clinical disease activity was set at�5.5, resulting in an NPV of 96%, positive predictivevalue (PPV) of 63%, sensitivity of 88%, and specificityof 85%. For UC the selected composite score (Table 3)had an AUC of >0.91 to predict disease activity usingpMayo, SCCAI, and modified Truelove and Witts indexas gold standards (0.960, 0.915, and 0.913, respec-tively). The Spearman correlation coefficients were0.820, 0.832, and 0.790, respectively (SupplementaryTable 4). The optimal cutoff for the mHI-UC to predictclinical disease activity was set at �4.99, resulting in anNPV of 97%, PPV of 72%, sensitivity of 89%, andspecificity of 90%.

Table 3. Calculation of the mHI-CD and mHI-UC

mHI-CD questions Answer Score

Number of liquid/very soft stools/day 0 0.0000 How mhave1–2 1.6983

�3 3.3966Abdominal pain No 0.0000 Rate yo

scaleYes 2.3868

Rate your well-being on a scale from0 to 10 (0 ¼ worst, 10 ¼ best)

8–10 0.0000 How of(0 ¼4–7 2.1336

0–3 4.2672How well do you feel your disease is

under control (0 ¼ no disease activity,10 ¼ worst disease activity)

0–2 0.0000 How w(0 ¼3–6 2.1175

7–10 4.2350Total score (SUM) Total s


Validation Phase

A total of 301 patients with CD (UCLA, n ¼ 127; UCI,n ¼ 82; LUMC, n ¼ 92) and 265 patients with UC (UCLA,n ¼ 119; UCI, n ¼ 67; LUMC, n ¼ 79) were analyzed inthe validation phase (Figure 1B, Table 2). For CD theSpearman correlation coefficient was 0.75 (P < .0001)between HBI and mHI-CD, the AUC of the ROC for pre-dicting clinical disease activity was 0.90, and a sensitivityof 94% and specificity of 67% were achieved using acutoff of �5.5 (Figures 2A and 2B). To achieve theoptimal NPV of 95% and sensitivity of 85%, the cutoff inthis cohort would be �6.38, which would result in 85%sensitivity, 80% specificity, 95% NPV, and 55% PPV. ThemHI-CD was poorly correlated with the SES-CD (r ¼ .30;P ¼ .0039), with an AUC of 0.63 (Figures 2A and 2C).

For UC the Spearman correlation coefficient was 0.72(P < .0001) between pMayo and mHI-UC, the AUC of theROC for predicting clinical disease activity was 0.91, anda sensitivity of 73% and specificity of 90% specificitywere achieved using a cutoff of �4.99 (Figures 2D and

mHI-UC questions Answer Score

any stools did youyesterday?

�2 0.00003–4 1.4428>4 2.8856

ur abdominal pain on afrom 0 to 10 (0 ¼ none, 10 ¼ worst)

0–2 0.00003–6 1.03927–10 2.0784

ten do you experience rectal bleeding?none, 10 ¼ always)

0–3 0.00004–10 2.2019

ell do you feel your disease is under controlno disease activity, 10 ¼ worst disease activity)

0–2 0.00003–5 1.75576–10 3.5114

core (SUM)


679

680

681

682

683

684

685

686

687

688

689

690

691

692

693

694

695

696

Figure 2. Performance of the developed mHI scores to detect disease activity in CD (A–C) and UC (D–F). ROC curves of themHI to predict clinical and endoscopic disease activity (A and D), scatter plot of the mHI versus clinical disease activity scores(B and E), and endoscopic disease activity scores (C and F).


697

698

699

700

701

702

703

704

705

706

707

708

709

710

711

712

713

714

715

716

717

718

719

720

721

722

723

724

725

726

727

728

729

730

731

732

733

734

735

736

737

738

739

740

741

742

743

744

745

746

747

748

749

750

751

752

753

754

755

756

757

758

759

760

761

762

763

764

765

766

767

768

769

770

771

772

773

774

775

776

777

778

779

780

781

782

783

784

785

786

787

788

789

790

791

792

793

794

795

796

797

798

799

800

801

802

803

804

805

806

807

808

809

810

811

812

2E). The optimal cutoff in this cohort would be �3.2,which would result in 85% sensitivity, 80% specificity,95% NPV, and 55% PPV. The mHI-UC was stronglycorrelated with the Mayo endoscopic subscore (r ¼ .60;P < .0001), with an AUC of 0.82 (Figures 2D and 2F).

Sensitivity to change was assessed in a subset of 50patients with CD and 44 patients with UC. Median timebetween questionnaires was 46 days (range, 2–352) forCD and 57.5 days (range, 3–275) for UC. Four (8%) pa-tients with CD deteriorated, 31 (62%) had stable diseaseactivity, and 15 (30%) improved. Of the patients with UC,5 (11%) deteriorated, 27 (61%) remained stable, and 12(27%) improved. There was a significant difference inmHI between patients who clinically improved, remainedstable, or worsened, with either CD (P ¼ .0030) or UC(P ¼ .0025) (Figures 3A and 3C). Test-retest reliabilitywas assessed in a subset of 40 patients with CD and 37patients with UC. The median time to second question-naire completion was 21 hours (range, 7–36) for the mHI-CD and 23 hours (range, 11–144) for the mHI-UC. Theintraclass correlation coefficient was 0.94 (confidencelimits, 0.89–0.97) for the mHI-CD and 0.94 (confidencelimits, 0.89–0.97) for the mHI-UC (Figures 3B and 3D).

One question in both the mHI-CD and the mHI-UCassessed patient-reported disease activity using a VAS(DA-VAS). In patients with CD (n ¼ 301) the DA-VAS had


a Spearman correlation of 0.63 (P < .0001) with the HBI,and the AUC for predicting clinical disease activity was0.83 (compared with r ¼ .75 and AUC ¼ 0.90 for the fullmHI-CD). The CD DA-VAS was weakly correlated with theSES-CD (r ¼ .21; P ¼ .040), and had an AUC to predictendoscopic disease activity of 0.59 (compared with r ¼0.30 and AUC ¼ 0.63 for the full mHI-CD). The DA-VASwas not significantly different among patients whodeteriorated, remained stable, or improved (P ¼ .12). Inpatients with UC (n ¼ 265) the DA-VAS had a Spearmancorrelation of 0.67 (P < .0001) with the pMayo, and theAUC for predicting clinical disease activity was 0.86(compared with r ¼ .72 and AUC ¼ 0.91 for the full mHI-UC). The UC DA-VAS was also correlated with theendoscopic component of the Mayo score (r ¼ .55; P <.0001), and had an AUC to predict endoscopic diseaseactivity of 0.79 (compared with r ¼ .60 and AUC ¼ 0.82for the full mHI-UC). A significant difference between theDA-VAS of patients with UC that deteriorated, remainedstable, or improved was observed (P ¼ .0052).

Discussion

We developed 2 questionnaires of 4 items consistingsolely of PROs for remote monitoring of patients with


Figure 3. Sensitivity tochange (A and C) and test-retest reliability (C and D)of the mHI-CD (A and B)and mHI-UC (C and D).


813

814

815

816

817

818

819

820

821

822

823

824

825

826

827

828

829

830

831

832

833

834

835

836

837

838

839

840

841

842

843

844

845

846

847

848

849

850

851

852

853

854

855

856

857

858

859

860

861

862

863

864

865

866

867

868

869

870

871

872

873

874

875

876

877

878

879

880

881

882

883

884

885

886

887

888

889

890

891

892

893

894

895

896

897

898

899

900

901

902

903

904

905

906

907

908

909

910

911

912

913

914

915

916

917

918

919

920

921

922

923

924

925

926

927

928

IBD, which can be used on mobile technology. Thequestionnaires were validated in a multicenter validationstudy and showed excellent characteristics to monitorclinical disease activity and symptom changes. As previ-ously shown, UC clinical disease activity highly correlateswith endoscopic disease activity, whereas correlationbetween CD symptoms and endoscopic findings is poor.21

Although previous studies have aimed to identifyPROs for disease monitoring either by adjusting existingquestionnaires,16 or by using subcomponents of existingquestionnaires,18,19 we were able to prospectively iden-tify PROs relevant for clinical disease monitoring andvalidate those in an independent cohort. Interestingly,patient-reported disease activity was shown to be anindependent predictor for clinical disease activity inpatients with UC and patients with CD, even after in-clusion of common IBD symptoms, such as stool fre-quency and abdominal pain. Patient-reported diseaseactivity alone had a comparable performance with thecomplete mHI in both CD and UC for detecting clinicaland endoscopic activity, although responsiveness tochanges in disease activity was reduced in particular inpatients with CD.


A limitation of this study is the potential for recallbias in CDAI calculation. Although 7-day diary formswere sent out in advance, we did not log whether diarieswere filled out daily or by recall. Additionally, in thevalidation cohort, optimal cutoffs of the mHI for detec-tion of disease activity were higher than expected for CDand lower than expected for UC. This might be caused bythe reduction of questionnaire items from >20 PROs tojust 4, or by differences in the patient population. Thevalidation phase of the study most accurately representsthe real-life situation. Therefore, we implemented thecutoff for optimal sensitivity and specificity as observedin the validation cohort in clinical practice.

This study was not primarily designed to evaluatecorrelations between PROs and endoscopic healing. In at-risk patients, clinical assessments remain warranted,which may lead to further endoscopic evaluation. Thistool offers an optimal screening method to monitor andevaluate disease activity in and outside of clinical prac-tice with a high NPV. The mHIs are currently imple-mented in the UCLA eIBD patient app (SupplementaryFigure 2) and automated messages are sent to a nursecoordinator when the mHI indicates disease activity. The


Q10


929

930

931

932

933

934

935

936

937

938

939

940

941

942

943

944

945

946

947

948

949

950

951

952

953

954

955

956

957

958

959

960

961

962

963

964

965

966

967

968

969

970

971

972

973

974

975

976

977

978

979

980

981

982

983

984

985

986

987

988

989

990

991

992

993

994

995

996

997

998

999

1000

1001

1002

1003

1004

1005

1006

1007

1008

1009

app is currently available to patients treated at the UCLACenter for IBD and can be downloaded in iOS or Androidby searching for “UCLA eIBD.”

The calculation of the mHI is complex. However,simplifying the score would most likely result in loss ofaccuracy. Because the index is meant to be automatedand implemented on digital platforms, we believe thatusing the more complex calculations is justified.

Cloud-based health technologies are predicted torevolutionize care delivery and patient engagement. Pa-tients can participate in their care by signaling meaningfulhealth outcomes during year-round monitoring. Barriersfor more widespread implementation of mHealth in IBDcare include policies affecting reimbursement and regula-tory requirements,22 and privacy and security concerns.1

In summary, we developed the mHI-CD and mHI-UCfor remote monitoring of patients with CD and patientswith UC. The scores are specifically designed for imple-mentation on a mobile application and are currentlyavailable to patients with IBD treated at the UCLA Centerfor IBD. Prospective randomized studies need to assessthe effect of remote monitoring on disease control,quality of life, patient satisfaction, and health care costs.

1010

1011

1012

1013

1014

1015

1016

1017

Supplementary Material

Note: To access the supplementary material accom-panying this article, visit the online version of ClinicalGastroenterology and Hepatology at www.cghjournal.org,and at http://dx.doi.org/10.1016/j.cgh.2015.10.035.

1018

1019
References
Q2

Q3

Q4

1020

1021

1022

1023

1024

1025

1026

1027

1028

1029

1030

1031

1032

1033

1034

1035

1036

1037

1038

1039

1040

1041

1042

1. Steinhubl SR, Muse ED, Topol EJ. The emerging field of mobilehealth. Sci Transl Med 2015;7:283rv3.

2. The Smartphone Difference. Pew Research Center, April2015. Available at: http://www.pewinternet.org/2015/04/01/us-smartphone-use-in-2015/. Accessed: May 11, 2015.

3. Kuehn BM. Is there an app to solve app overload? JAMA 2015;313:1405–1407.

4. Available at: http://www.medivo.com/gi/.

5. Available at: http://www.ibdetermined.org/Tracker.aspx.Accessed: May 13, 2015.

6. Haas L, Maryniuk M, Beck J, et al. National standards for dia-betes self-management education and support. Diabetes Care2013;36(Suppl 1):S100–108.

7. Heneghan C, Ward A, Perera R, et al. Self-monitoring of oralanticoagulation: systematic review and meta-analysis of indi-vidual patient data. Lancet 2012;379:322–334.

8. Johansen MA, Henriksen E, Horsch A, et al. Electronic symptomreporting between patient and provider for improved health careservice quality: a systematic review of randomized controlledtrials. part 1: state of the art. J Med Internet Res 2012;14:e118.

9. Elkjaer M, Shuhaibar M, Burisch J, et al. E-health empowers pa-tients with ulcerative colitis: a randomised controlled trial of theweb-guided ’Constant-care’ approach. Gut 2010;59:1652–1661.

10. Pedersen N, Thielsen P, Martinsen L, et al. eHealth: individual-ization of mesalazine treatment through a self-managed


web-based solution in mild-to-moderate ulcerative colitis.Inflamm Bowel Dis 2014;20:2276–2285.

11. Pedersen N, Elkjaer M, Duricova D, et al. eHealth: individuali-sation of infliximab treatment and disease course via a self-managed web-based solution in Crohn’s disease. AlimentPharmacol Ther 2012;36:840–849.

12. Cross RK, Cheevers N, Rustgi A, et al. Randomized, controlledtrial of home telemanagement in patients with ulcerative colitis(UC HAT). Inflamm Bowel Dis 2012;18:1018–1025.

13. Cross RK, Jambaulikar G, Langenberg P, et al. TELEmedicinefor Patients with Inflammatory Bowel Disease (TELE-IBD):design and implementation of randomized clinical trial. ContempClin Trials 2015;42:132–144.

14. Jensen RE, Rothrock NE, DeWitt EM, et al. The role of technicaladvances in the adoption and integration of patient-reportedoutcomes in clinical care. Med Care 2015;53:153–159.

15. Williet N, Sandborn WJ, Peyrin-Biroulet L. Patient-reportedoutcomes as primary end points in clinical trials of inflam-matory bowel disease. Clin Gastroenterol Hepatol 2014;12:1246–1256.e6.

16. Atreja A, Khan S, Rogers JD, et al. Impact of the MobileHealthPROMISE platform on the quality of care and quality oflife in patients with inflammatory bowel disease: study protocolof a pragmatic randomized controlled trial. JMIR Res Protoc2015;4:e23.

17. Evertsz FB, Hoeks CC, Nieuwkerk PT, et al. Development of thepatient Harvey Bradshaw index and a comparison with aclinician-based Harvey Bradshaw index assessment of Crohn’sdisease activity. J Clin Gastroenterol 2013;47:850–856.

18. Lewis JD, Chuai S, Nessel L, et al. Use of the noninvasivecomponents of the Mayo score to assess clinical response inulcerative colitis. Inflamm Bowel Dis 2008;14:1660–1666.

19. Khanna R, Zou G, D’Haens G, et al. A retrospective analysis: thedevelopment of patient reported outcome measures for theassessment of Crohn’s disease activity. Aliment Pharmacol Ther2015;41:77–86.

20. Vermeire S, Schreiber S, Sandborn WJ, et al. Correlation be-tween the Crohn’s disease activity and Harvey-Bradshawindices in assessing Crohn’s disease severity. Clin Gastro-enterol Hepatol 2010;8:357–363.

21. Falvey JD, Hoskin T, Meijer B, et al. Disease activity assessmentin IBD: clinical indices and biomarkers fail to predict endoscopicremission. Inflamm Bowel Dis 2015;21:824–831.

22. Adler-Milstein J, Kvedar J, Bates DW. Telehealth among UShospitals: several factors, including state reimbursement andlicensure policies, influence adoption. Health Aff 2014;33:207–215.

Reprint requestsAddress requests for reprints to: Welmoed K. Van Deen, MD, Center forInflammatory Bowel Diseases, Melvin and Bren Simon Digestive DiseasesCenter, David Geffen School of Medicine, University of California at LosAngeles, 10945 Le Conte Avenue, #2114, Los Angeles, California 90095.e-mail: [email protected]; fax: (310) 206-9906.

Conflicts of interestThe authors disclose no conflicts.

FundingThis study was partly supported by a grant from Genova Diagnostics. GenovaDiagnostics provided stool collection kits and performed fecal calprotectintesting during the study.


1043

1044

http://www.cghjournal.org


http://www.pewinternet.org/2015/04/01/us-smartphone-use-in-2015/

http://www.pewinternet.org/2015/04/01/us-smartphone-use-in-2015/

http://www.medivo.com/gi/

http://www.ibdetermined.org/Tracker.aspx

mailto:[email protected]

SupplementaryFigure 1. Seven-day diaryform, which was sent topatients to log symptomsduring the 7 days beforethe clinic visit for calcula-tion of the CDAI.

web4C=FPO

9.e1 Van Deen et al Clinical Gastroenterology and Hepatology Vol. -, No. -

1045

1046

1047

1048

1049

1050

1051

1052

1053

1054

1055

1056

1057

1058

1059

1060

1061

1062

1063

1064

1065

1066

1067

1068

1069

1070

1071

1072

1073

1074

1075

1076

1077

1078

1079

1080

1081

1082

1083

1084

1085

1086

1087

1088

1089

1090

1091

1092

1093

1094

1095

1096

1097

1098

1099

1100

1101

1102

1103

1104

1105

1106

1107

1108

1109

1110

1111

1112

1113

1114

1115

1116

1117

1118

1119

1120

1121

1122

1123

1124

1125

1126

1127

1128

1129

1130

1131

1132

1133

1134

1135

1136

1137

1138

1139

1140

1141

1142

1143

FLA 5.4.0 DTD � YJCGH54544_proof � 11 December 2015 � 12:31 pm � ce CJ

1144

1145

1146

1147

1148

1149

1150

1151

1152

1153

1154

1155

1156

1157

1158

1159

1160

web4C=FPO

Supplementary Figure 2. Screenshots of the mHI as implemented in the UCLA eIBD patient app available to patients treatedat the UCLA IBD Center.


- 2015 mHealth Index for IBD Monitoring 9.e2

1161

1162

1163

1164

1165

1166

1167

1168

1169

1170

1171

1172

1173

1174

1175

1176

1177

1178

1179

1180

1181

1182

1183

1184

1185

1186

1187

1188

1189

1190

1191

1192

1193

1194

1195

1196

1197

1198

1199

1200

1201

1202

1203

1204

1205

1206

1207

1208

1209

1210

1211

1212

1213

1214

1215

1216

1217

1218

1219

1220

1221

1222

1223

1224

1225

1226

1227

1228

1229

1230

1231

1232

1233

1234

1235

1236

1237

1238

1239

1240

1241

1242

1243

1244

1245

1246

1247

1248

1249

1250

1251

1252

1253

1254

1255

1256

1257

1258

1259

1260

1261

1262

1263

1264

1265

1266

1267

1268

1269

1270

1271

1272

1273

1274

1275

1276

Supplementary Table 1. CD Variables, Univariate Logistic Regression Outcomes for Prediction of Active Disease (HBI>4)

Domain Variable n Cutoffs MLE SE Chi-square P value Modela Var#

Stool frequency Number of liquid/very soft stoolsper day

108 Continuous 1.63 0.35 22.34 <.0001 —

0/1–3/>3 3.49 0.73 23.11 <.0001 —

0/1–2/>2 3.55 0.66 28.58 <.0001 1 V1.10/>0 3.36 0.77 19.10 <.0001 —

How many stools did you haveyesterday during the day?

108 Continuous 0.85 0.19 20.09 <.0001 —

0–1/>1 2.00 0.65 9.53 .0020 —

Number of stools more thannormal per day

108 Continuous 0.87 0.36 5.88 .015 —

0/>0 0.84 0.51 2.69 .10 —

Total number of stools per day 108 Continuous 0.97 0.20 22.75 <.0001 —

�1/>1 1.75 0.77 5.08 .024 —

0–2/>2 2.58 0.55 22.19 <.0001 —

0–1/2–4/>4 2.64 0.58 20.79 <.0001 —

0–2/3–4/>4 2.16 0.40 28.49 <.0001 2 V1.2Stools at night How many stools did you have

last night?108 Continuous 0.88 0.26 11.59 .0007 —

0/>0 1.15 0.44 6.90 .0086 —

0/1/>1 0.99 0.30 11.01 .0009 —

0/1–2/>2 1.28 0.36 12.29 .0005 1, 2 V2Abdominal pain Abdominal pain (no–severe) 108 No/mild/moderate/severe 2.68 0.50 28.11 <.0001 —

No/yes 3.72 0.62 35.35 <.0001 1, 2 V3Abdominal pain (no pain–continuous) 108 No abdominal pain/with bowel actions/continuous 1.52 0.31 23.97 <.0001 —

Abdominal pain VAS 108 Continuous 0.51 0.10 24.99 <.0001 —

0–3/>3, <8/8–10 2.02 0.41 24.17 <.0001 —

<3/3–6/>6 1.94 0.39 25.03 <.0001 —

General well-being General well-being (very well– terrible) 108 Very well slightly below par/poor/very poor/terrible 1.31 0.30 18.93 <.0001 —

Well-being VAS 108 Continuous 0.47 0.11 18.44 <.0001 —

<4/�4, <7/�7 1.42 0.35 16.76 <.0001 —

�2/>2, <7/�7 1.80 0.40 20.09 <.0001 1, 2 V4Well-being (no impairment–unable

to work)108 No impairment/impaired but able to continue

activities/activities reduced/unable to work1.21 0.83 17.90 <.0001 —

Urgency Urgency of defecation(no urgency–incontinence)

108 No urgency/hurry/immediately/incontinence 1.30 0.32 16.99 <.0001 1, 2 V5No incontinence/incontinence 0.88 1.43 0.38 .54 —

Stool consistency Stool consistency 108 Normal or variably normal/semiformed/liquid 2.09 0.48 19.06 <.0001 1, 2 V6Do you take opiates or lomotil/imodium

for diarrhea?108 No/yes 1.22 0.52 5.47 .019 —

Antidiarrheals VAS 108 Continuous 0.00 0.09 0.00 1.00 —

Rectal bleeding Rectal bleeding VAS 108 Continuous 0.16 0.09 3.06 .080 —

�3/>3 1.22 0.52 5.47 .019 1, 2 V7What % of bowel movements contains

visible blood?108 None/<50%/�50%/blood alone 0.36 0.34 1.08 .30 —

Amount of blood in stool 108 None/trace/occasionally frank/usually frank 0.61 0.28 4.72 .030 —

Fever Did you have fever yesterday? 108 No/yes 14.53 913.40 0.00 .99 —

Anorexia Loss of appetite 108 No/yes 1.04 0.45 5.51 .019 1, 2 V8Nausea/vomiting Nausea and/or vomiting 108 No/yes 1.03 0.49 4.44 .035 1, 2 V9

FLA

5.4.0

DTD

�YJC

GH54544_proof�

11Decem

ber

2015�

12:31pm

�ce

CJ

9.e3Van

Deen

etal

Clinical

Gastroenterology

andHepatology

Vol.

-,No.

-

1277

1278

1279

1280

1281

1282

1283

1284

1285

1286

1287

1288

1289

1290

1291

1292

1293

1294

1295

1296

1297

1298

1299

1300

1301

1302

1303

1304

1305

1306

1307

1308

1309

1310

1311

1312

1313

1314

1315

1316

1317

1318

1319

1320

1321

1322

1323

1324

1325

1326

1327

1328

1329

1330

1331

1332

1333

1334

1335

1336

1337

1338

1339

1340

1341

1342

1343

1344

1345

1346

1347

1348

1349

1350

1351

1352

1353

1354

1355

1356

1357

1358

1359

1360

1361

1362

1363

1364

1365

1366

1367

1368

1369

1370

1371

1372

1373

1374

1375

1376

1377

1378

1379

1380

1381

1382

1383

1384

1385

1386

1387

1388

1389

1390

1391

1392

Disease activity Disease control VAS 108 Continuous 0.42 0.09 21.88 <.0001 —

�3, >3, <7/�7 1.60 0.32 24.41 <.0001 —

�2, >2, <7/�7 1.80 0.36 24.58 <.0001 1, 2 V10Laboratory studies C-reactive protein (mg/dL) 93 Continuous 0.27 0.14 3.80 .051 —

�0.8/>0.8 1.77 0.53 11.11 .0009 —

Calprotectin (mg/g) 63 Continuous 0.00 0.00 6.45 .011 —

<163/�163 1.22 0.63 3.71 .054 —

<50/�50 0.28 0.66 0.18 .67 —

White blood cell count (*103 cells/mL) 96 Continuous 0.15 0.08 3.20 .074 —

�9.95/>9.95 1.06 0.66 2.60 .11 —

Albumin (g/dL) 94 Continuous -1.55 0.61 6.47 .011 —

�3.7/<3.7 1.92 0.90 4.51 .034 —

Platelets (*103 cells/mL) 96 Continuous 0.00 0.00 1.43 .23 —

�143/<143 15.17 770.20 0.00 .98 —

Erythrocyte sedimentation rate (mm/hr) 85 Continuous 0.05 0.02 7.28 .0070 —

�22 (F) or �10 (M)/>22 (F) or >10 (M) 1.62 0.53 9.38 .0022 —

Hemoglobin (g/dL) 95 Continuous -0.10 0.14 0.59 .44 —

>11.6 (F) or >13.5 (M)/�11.6 (F) or �13.5 (M) 1.15 0.50 5.24 .022 —

Hematocrit 96 Continuous -3.34 5.39 0.38 .54 —

>0.349 (F) or >0.385 (M)/�0.349 (F) or �0.385 (M) 0.61 0.57 1.15 .28 —

F, female; M, male; MLE, maximum likelihood estimate, SE, standard error.aVariables selected for multivariate regression models (Supplementary Table 4).

FLA

5.4.0

DTD

�YJC

GH54544_proof�

11Decem

ber

2015�

12:31pm

�ce

CJ

-2015

mHealth

Indexfor

IBDMonitoring

9.e4

1393

1394

1395

1396

1397

1398

1399

1400

1401

1402

1403

1404

1405

1406

1407

1408

1409

1410

1411

1412

1413

1414

1415

1416

1417

1418

1419

1420

1421

1422

1423

1424

1425

1426

1427

1428

1429

1430

1431

1432

1433

1434

1435

1436

1437

1438

1439

1440

1441

1442

1443

1444

1445

1446

1447

1448

1449

1450

1451

1452

1453

1454

1455

1456

1457

1458

1459

1460

1461

1462

1463

1464

1465

1466

1467

1468

1469

1470

1471

1472

1473

1474

1475

1476

1477

1478

1479

1480

1481

1482

1483

1484

1485

1486

1487

1488

1489

1490

1491

1492

1493

1494

1495

1496

1497

1498

1499

1500

1501

1502

1503

1504

1505

1506

1507

1508

Supplementary Table 2. UC Variables, Univariate Logistic Regression Outcomes for Prediction of Active Disease (pMayo >2)

Domain Variable n Cutoffs MLE SE Chi-square P value Modela Var#

Stool frequency How many stools did you haveyesterday during the day?

109 Continuous 0.87 0.79 20.57 <.0001 —

<3/3–4/5–6/7–9/>9 1.61 0.33 23.11 <.0001 —

<4/4–6/>6 2.24 0.50 19.69 <.0001 —

<4/4–6/>6 2.32 0.49 22.35 <.0001 —

Stools more than normal 109 Continuous 0.99 0.33 9.23 .0024 —

0/1–2/3–4/>4 1.65 0.41 16.17 <.0001 —

0/1–2/>2 1.78 0.41 19.18 <.0001 —

Total number of stools/day 109 Continuous 0.78 0.16 23.03 <.0001 —

<3/3–6/>6 2.24 0.42 28.86 <.0001 —

0–4/>4 3.57 0.62 32.79 <.0001 1, 2 V1.10–2/3–4/>4 1.95 0.34 32.46 <.0001 3, 4 V1.2

Stools at night How many stools did youhave last night?

109 Continuous 0.95 0.26 12.77 .0004 —

0/>0 1.48 0.43 11.79 .0006 —

0/1–3/>3 1.49 0.38 15.01 .0001 1, 2, 3, 4 V2Abdominal pain Abdominal pain (no– severe) 109 No/mild/moderate/severe 0.78 0.25 9.60 .0019 —

No/yes 1.47 0.45 10.62 .0011 —

Abdominal pain (no– continuous) No abdominal pain 1.02 0.30 11.44 .0001 —

Abdominal pain VAS 109 Continuous 0.40 0.09 19.21 <.0001 —

<3/�3, �6/>6 1.75 0.37 21.84 <.0001 1, 2, 3, 4 V30/>0, �4/>4 1.40 0.33 18.20 <.0001 —

General well-being General well-being? (very well–terrible) 109 Very well/slightly below par/poor/very poor/terrible 1.27 0.34 14.12 .0002 —

General well-being (perfect–terrible) 109 Perfect/very good/good/average/poor/terrible 0.91 0.23 16.08 <.0001 1, 3 V4.1Well-being VAS 109 Continuous 0.43 0.11 15.30 <.0001 —

�3/>3 1.62 0.45 13.31 .0003 —

�3/>3, �6/>6 1.42 0.36 15.73 <.0001 2, 4 V4.2Well-being (no impairment–unable

to work)109 No impairment/impaired but able to continue

activities/activities reduced/unable to work1.14 0.36 15.93 <.0001 —

Urgency Urgency of defecation 109 No urgency/hurry/immediate/incontinence 2.01 0.40 24.82 <.0001 —

No urgency/hurry/immediate 2.03 0.40 25.13 <.0001 1, 2, 3, 4 V5No incontinence/incontinence 14.13 826.90 0.00 .99 —

Stool consistency Stool consistency 109 Normal or variably normal/semiformed/liquid 2.13 0.46 21.68 <.0001 1, 2, 3, 4 V6Do you take opiates or

lomotil/imodium for diarrhea?109 No/yes -0.03 0.59 0.00 .96 —

Antidiarrheals VAS 109 Continuous 0.05 0.09 0.26 .26 —

Rectal bleeding Rectal bleeding VAS 109 Continuous 0.59 0.11 27.82 <.0001 —

�3/>3 2.71 0.49 29.97 <.0001 —

<2/�2, <5/�5, <8/�8 1.66 0.31 28.15 <.0001 1, 2, 3, 4 V7What % of bowel movements

contains visible blood?109 None/<50%/�50%/blood alone 2.84 0.59 22.89 <.0001 —

None/<50%/�50% 2.95 0.60 24.26 <.0001 —

None/>0% 3.86 1.04 13.76 .0002 —

Amount of blood in stool 109 None/trace/occasionally frank/usually frank 1.83 0.35 27.22 <.0001 —

None/trace/more than a trace 1.95 0.37 27.53 <.0001 —

Anorexia Loss of appetite 109 No/yes 1.44 0.48 9.10 .0025 1, 2, 3, 4 V8Nausea/vomiting Nausea and/or vomiting 109 No/yes 0.72 0.49 2.12 .14 —

FLA

5.4.0

DTD

�YJC

GH54544_proof�

11Decem

ber

2015�

12:31pm

�ce

CJ

9.e5Van

Deen

etal

Clinical

Gastroenterology

andHepatology

Vol.

-,No.

-

1509

1510

1511

1512

1513

1514

1515

1516

1517

1518

1519

1520

1521

1522

1523

1524

1525

1526

1527

1528

1529

1530

1531

1532

1533

1534

1535

1536

1537

1538

1539

1540

1541

1542

1543

1544

1545

1546

1547

1548

1549

1550

1551

1552

1553

1554

1555

1556

1557

1558

1559

1560

1561

1562

1563

1564

1565

1566

1567

1568

1569

1570

1571

1572

1573

1574

1575

1576

1577

1578

1579

1580

1581

1582

1583

1584

1585

1586

1587

1588

1589

1590

1591

1592

1593

1594

1595

1596

1597

1598

1599

1600

1601

1602

1603

1604

1605

1606

1607

1608

1609

1610

1611

1612

1613

1614

1615

1616

1617

1618

1619

1620

1621

1622

1623

1624

Disease activity Disease control VAS 109 Continuous 0.46 0.10 21.67 <.0001 —

�3/>3, <7, �7 1.44 0.31 21.37 <.0001 —

�2/>2, �5/>5 1.82 0.36 25.86 <.0001 1, 2, 3, 4 V10�2/>2, <7/�7 1.68 0.36 21.76 <.0001 —

Labs C-reactive protein (mg/dL) 91 Continuous 1.03 0.42 5.87 .015 —

�0.8/>0.8 1.50 0.55 7.35 .0067 —

Calprotectin (mg/g) 70 Continuous 0.00 0.00 4.31 .038 —

<163/�163 1.29 0.56 5.22 .022 —

<50/�50 1.96 0.80 6.06 .014 —

White blood cell count (*103 cells/mL) 93 Continuous 0.20 0.08 5.91 .015 —

�9.95/>9.95 1.39 0.56 6.19 .013 —

Albumin (g/dL) 88 Continuous -1.52 0.66 5.39 .02 —

�3.7/<3.7 14.36 610.50 0.00 .98 —

Platelets (*103 cells/mL) 93 Continuous 0.00 0.00 2.54 .11 —

�143/<143 -12.66 805.50 0.00 .99 —

Erythrocyte sedimentation rate (mm/hr) 85 Continuous 0.04 0.01 8.38 .0038 —

�22 (F) or �10 (M)/>22 (F) or >10 (M) 1.04 0.47 4.81 .028 —

Hemoglobin (g/dL) 93 Continuous -0.31 0.15 4.59 .032 —

>11.6 (F) or >13.5 (M)/�11.6 (F) or �13.5 (M) 0.75 0.50 2.25 .13 —

Hematocrit 93 Continuous -9.46 5.91 2.66 .10 —

>0.349 (F) or >0.385 (M)/�0.349 (F) or �0.385 (M) 1.84 0.65 7.99 .0047 —

F, female; M, male; MLE, maximum likelihood estimate, SE, standard error.aVariables selected for multivariate regression models (Supplementary Table 4).

FLA

5.4.0

DTD

�YJC

GH54544_proof�

11Decem

ber

2015�

12:31pm

�ce

CJ

-2015

mHealth

Indexfor

IBDMonitoring

9.e6

1625

1626

1627

1628

1629

1630

1631

1632

1633

1634

1635

1636

1637

1638

1639

1640

1641

1642

1643

1644

1645

1646

1647

1648

1649

1650

1651

1652

1653

1654

1655

1656

1657

1658

1659

1660

1661

1662

1663

1664

1665

1666

1667

1668

1669

1670

1671

1672

1673

1674

1675

1676

1677

1678

1679

1680

1681

1682

1683

1684

1685

1686

1687

1688

1689

1690

1691

1692

1693

1694

1695

1696

1697

1698

1699

1700

1701

1702

1703

1704

1705

1706

1707

1708

1709

1710

1711

1712

1713

1714

1715

1716

1717

1718

1719

1720

1721

1722

1723

1724

1725

1726

1727

1728

1729

1730

1731

1732

1733

1734

1735

1736

1737

1738

1739

1740

Supplementary Table 3. Development and Evaluation of Composite Scores for CD

Composite score development AUC ROC curves P value

Dependent variable Model Variables in composite scoresa HBI CDAI HBI CDAI

HBI 1 V1.1; V3 0.981 0.912 0.903 0.7392 V1.2; V3; V10 0.956 0.898 0.750 0.702

CDAI 1 V1.1; V3; V4; V10 0.951 0.963 0.837 0.8302 V3; V4; V10 0.900 0.942 0.731 0.771

V1.1, number of liquid/very soft stools per day; V1.2, total number of stools per day; V3, abdominal pain; V4, well-being VAS; V10, disease control VAS.aSee Supplementary Table 1 for full details on each variable.

Supplementary Table 4. Development and Evaluation of Composite Scores for UC

Composite score development AUC ROC curves P value

Dependent variable Model Variables in composite scoresa pMayo SCCAI MTWI pMayo SCCAI MTWI

pMayo 1, 2 V1.1; V3; V7; V8; V10 0.960 0.865 0.849 0.769 0.769 0.7031, 2b V1.1; V3; V7; V10 0.957 0.879 0.883 0.797 0.803 0.762

3 V1.2; V3; V7; V8; V10 0.964 0.908 0.890 0.808 0.812 0.7483b V1.2; V3; V7; V10 0.960 0.915 0.913 0.820 0.832 0.7904 V1.2; V 4.2; V7; V10 0.956 0.920 0.909 0.809 0.838 0.787

SCCAI 1 V2; V4.1; V5; V7 0.872 0.974 0.896 0.711 0.907 0.8362, 4 V2; V3; V4.2; V5; V7; V10 0.904 0.971 0.883 0.765 0.911 0.810

3 V1.2; V2; V4.1; V5; V7 0.885 0.981 0.923 0.721 0.914 0.869MTWI 1 Unbalanced model NA NA NA NA NA NA

2 V1.1; V2; V6; V7; V8 0.928 0.879 0.937 0.801 0.806 0.8553 V1.2; V2; V4.1 0.880 0.907 0.984 0.712 0.797 0.9334 V1.2; V2; V4.2; V7; V8 0.906 0.894 0.950 0.777 0.810 0.866

MTWI, modified Truelove and Witts index; NA Q11; V1.1 and V1.2, number of stools per day; V2, number of stools at night; V3, abdominal pain VAS; V4.1, generalwell-being; V4.2, well-being VAS; V5, urgency of defecation; V6, stool consistency; V7, rectal bleeding VAS; V8, anorexia; V10, disease control VAS.aSee Supplementary Table 2 for full details on each variable.bIn these models loss of appetite was excluded as independent variable because of a clinically irrelevant negative value in the model.


9.e7 Van Deen et al Clinical Gastroenterology and Hepatology Vol. -, No. -

1741

1742

1743

1744

1745

1746

1747

1748

1749

1750

1751

1752

1753

1754

1755

1756

1757

1758

1759

1760

1761

1762

1763

1764

1765

1766

1767

1768

1769

1770

1771

1772

1773

1774

1775

1776

1777

1778

1779

1780

1781

1782

1783

1784

1785

1786

1787

1788

1789

1790

1791

1792

1793

1794

1795

1796

1797

1798

1799

1800

1801

1802

1803

1804

1805

1806

1807

1808

1809

1810

1811

1812

1813

1814

1815

1816

1817

1818

1819

1820

1821

1822

1823

1824

1825

1826

1827

1828

1829

1830

1831

1832

1833

1834

1835

1836

1837

1838

1839

1840

1841

1842

1843

1844

1845

1846

1847

1848

1849

1850

1851

1852

1853

1854

1855

1856

1857

1858

1859

1860

1861

1862

1863

1864

1865

1866

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页文献云下载图书馆入口外文数据库大全疑难文献辅助工具

http://www.xuebalib.com/cloud/

http://www.xuebalib.com/

http://www.xuebalib.com/cloud/


http://www.xuebalib.com/vip.html

http://www.xuebalib.com/db.php

http://www.xuebalib.com/zixun/2014-08-15/44.html


Documents

Development and Validation of an Inflammatory Bowel ...download.xuebalib.com/2sn1bVMjztGa.pdf · Q1 M. G. H. van Oijen,*,jj and D. W ... from April 2014 through March 2015. RESULTS: