Effect of Text Messaging on Bowel Preparation and Appointment Attendance for Outpatient Colonoscopy

Key Points Question Do automated text messages improve outpatient colonoscopy attendance rate and bowel preparation quality? Findings In this randomized clinical trial including 753 patients, usual care patient instructions were compared with an automated text messaging intervention in the week prior to outpatient colonoscopy. There was no significant difference between groups in appointment attendance rate or bowel preparation quality. Meaning This randomized clinical trial found that automated text message reminders and instructions did not improve outpatient colonoscopy adherence, although future studies may identify patient subgroups that benefit from this approach.


6
This supplement provides additional information about the work. It contains the following items: 7 8 Initial Protocol ….………………………………………………………………………………………………………………………………. 1 9 10 Final Protocol ……………………………………………………………………………………………………………………………………. 10 11 12 Summary of protocol changes ………………………………………………………………………………………………………….. 20 13 14 Original Secondary outcomes of the study will include colonoscopy show rate, bowel preparation quality (poor, 62 fair, good, excellent), colonoscopy cancellation rate, timing of advance cancellation notification (days), 63 and colonoscopy appointment no-show rate. 64 65 Background 66 Colorectal cancer (CRC) is the second leading cause of cancer death in the US, yet there are effective 67 screening and treatment strategies that allow for early detection and treatment. CRC screening is 68 recommended for all individuals aged 50-75, which could include stool testing or colonoscopy, but 69 national rates are still suboptimal at 59-65%. Colonoscopy is an essential component of CRC screening,70 as it is also required if stool testing is positive. However, colonoscopy requires a complex process to 71 identify an escort, purchase the preparation, take a day off from work, adhere to a clear liquid diet, and 72 complete the split-dose preparation as recommended. This results in a significant no-show and 73 cancellation rate, along with suboptimal preparation quality, which can lead to non-adherence and 74 incomplete screening. 75 76 Current approaches to engaging patients include having nurses call patients before the procedure or 77 patient navigators. However, it is often difficult to get patients on the phone, and these interventions 78 can be costly, making it less scalable for clinical practices. Other interventions such as videos or mobile 79 apps have been limited by poor user experience or limited engagement with the patient. There is an 80 opportunity to leverage an automated text message navigation intervention using the Way to Health 81 (WTH) platform to improve patient engagement prior to colonoscopy completion. The WTH platform is a 82 Penn Medicine platform that is hosted on site at the University of Pennsylvania. The platform allows 83 custom text messages to automatically be sent to patients, in addition to bidirectional message 84 capabilities. WTH is protected by a secure firewall and is a HIPAA compliant platform. 85

86
In the past year, our team conducted a quality improvement pilot initiative using WTH that tested the 87 feasibility and impact of a one-week text messaging protocol for patients who were scheduled for 88 outpatient colonoscopy. The text messages sent to patients contained information about the 89 preparation process and instructions, expectations about the procedure, and reminders about location 90 and timing. Among the 21 patients enrolled in the pilot, we found high user acceptability and higher 91 colonoscopy show rates as compared to baseline values at PPMC (90% versus ~50%). As such we believe 92 that the texting intervention is feasible for testing in the context of a randomized controlled trial. 93 94

Study Design 95
Design 96 We will perform a pragmatic randomized controlled trial evaluating the impact of the text message-97 based intervention (arm 2) as compared to usual care alone (arm 1). Patient enrollment will be 98 performed with assistance from the call center at the Pennsylvania Presbyterian Medical Center (PPMC), 99 and/or through phone calls from a clinical research coordinator. After enrollment, patients will be 100 randomized 1:1 to the arms listed above through the Way to Health platform. Although patients in arm 101 2 may be enrolled weeks or months in advance of their colonoscopy, they will only receive intervention 102 text messages in the 7 days prior to the scheduled colonoscopy. Usual care consists of a phone call from 103 the PPMC endoscopy staff in the week prior to colonoscopy, if the patient is able to be reached. Patients 104 are also given the endoscopy phone number, and may call to speak to staff if they have specific 105 questions about their colonoscopy or need to reschedule. A detailed outline of study design can be 106 found in the procedures sections. 107 108

Study duration 109
The duration of participation for patients in the intervention will be from the time of enrollment to the 110 date of scheduled colonoscopy. However, the patient will only receive intervention text messages in the 111 week prior to the scheduled colonoscopy. We will plan to recruit patients from November 2018 through 112 January 2019 Participants in the study will be planned for outpatient colonoscopy at the Pennsylvania Presbyterian 149 Medical Center (PPMC). Once a patient is confirmed to meet eligibility criteria, patient enrollment will 150 be performed with assistance from the call center at PPMC, and/or through phone calls from a clinical 151 research coordinator. When patients have a physician order for a colonoscopy, the patient must contact 152 the call center in order to schedule the colonoscopy. The call center staff have agreed to discuss the 153 research project and enroll patients using a script to be approved by the IRB. Importantly, we are 154 requesting a waiver of consent for several reasons: (1) the study presents no more than minimal risk of 155 harm to participants, (2) the patients who are randomized to either arm will not be deprived of any 156 standard care available in the status quo and thus the study will not violate their rights or welfare, (3) 157 the study could not be practicably completed without such a waiver, as the scheduling process is 158 handled over the phone by call center staff, and (4) requiring standard informed consent would require 159 an additional touchpoint by the research coordinator and the consent process would preclude us from 160 evaluating the intended intervention by introducing selection bias and altering the intervention that 161 patients receive. After enrollment, patients will be 1:1 randomized to arm 1 or arm 2 using the Way to 162 Health Screening -Phase 1: Patients who have a physician order for a colonoscopy will call the PPMC call center 180 in order to schedule their procedure. At this time, the call center staff will screen the patient for 181 eligibility (age criteria and text messaging ability) and will explain the details of the study following a pre-182 determined script. Patients who agree to participate in the study will have their cell phone number 183 recorded. A list of these patients will be forward to the clinical research coordinator on a weekly basis, 184 who will confirm patient eligibility and then manually enter patients into the WTH platform. Based on 185 our experience with the QI pilot described above, we estimate needing to screen between 500 and 600 186 patients to meet our enrollment targets (400 total patients). 187 Randomization -Phase 2: using a random number generator in the WTH platform, patients will be 1:1 188 randomized to arm 1 or arm 2, until the target sample size of 200 patients per arm is reached. 189 Randomization will be performed and recorded in WTH. 190 Chart review -Phase 3: after randomization and in tandem with phase 4 (below), the research 191 coordinator will perform a chart review to obtain patient demographic and comorbidity data 192 (hypertension, hyperlipidemia, diabetes, inflammatory bowel disease, obesity, active opiate 193 prescription). All data will be inputted into and stored in a secured RedCap database. Only key study 194 staff will have access to the RedCap database, which is stored on a secure firewall-protected server. 195 Intervention -Phase 4: Patients in arm 2 will begin receiving text messages per a pre-determined 196 protocol starting 7 days prior to the date of scheduled colonoscopy (the text messaging protocol is 197 attached to this application). Patients in arm 1 and arm 2 will both receive usual care, which includes (1) 198 bowel preparation instructions that are delivered via mail or through a secure online messaging portal, 199 (2) a phone call from the endoscopy staff in the week prior to colonoscopy, and (3) the option to call the 200 endoscopy staff during business hours to have any questions answered on demand. 201 Outcomes data collection -Phase 5: After a patient's scheduled colonoscopy date has passed, the 202 research coordinator will review the medical record to gather additional data from the endoscopy 203 procedure. This will include the indication for the procedure, quality of bowel preparation (poor, fair, 204 good, excellent), and completeness (cecum reached). Procedure status will also be recorded (canceled, 205 no-show, completed). 206 Statistical Analysis and Manuscript Preparation -Phase 6: The statistical analysis (detailed below) will 207 commence after completion of outcomes ascertainment for the entire cohort. 208 209

Analysis Plan 210
Descriptive analysis will include comparison of baseline demographics and major medical comorbidities 211 between study arms. Continuous and categorical data will be compared using the Wilcoxon Rank-sum 212 and Chi-squared tests, respectively. An alpha threshold of 0.05 will be used for statistical significance, 213 with two-tailed testing performed in all instances. For the primary analysis, we will perform a Chi-214 squared test of independence to compare the proportion of colonoscopies with good/excellent bowel 215 preparation between groups. In the secondary analysis, Chi-squared tests of independence will be 216 performed to compare colonoscopy status (show, no-show, cancelation) as well as bowel preparation 217 quality (excellent, good, fair, poor) between groups. Finally, among canceled colonoscopies, median lead 218 time of cancellation will be compared between groups using the Wilcoxon Rank-sum test. 219 Consent 220

Consent Process 221
Overview 222 We are requesting a waiver of informed consent for this study. The reason for our request is explained 223 in the Waiver of Consent section. Verbal consent for text messaging will still be obtained from all 224 participants in the study. Please see the script templates attached to this application. 225 226

Children and Adolescents 227
Not applicable. We are only enrolling subjects 18 years of age and older. 228 229

Adult Subjects Not Competent to Give Consent 230
We are requesting a waiver of informed consent for this study but patients will still need to verbally opt 231 in to text messaging. We plan to enroll only those patients who are competent at time of enrollment to 232 opt-in for themselves. 233 234

Waiver or Alteration of Informed Consent* 236
Waiver or alteration of required elements of consent. 237 238

Minimal Risk* 239
This study is minimal risk as all participants will be receiving standard clinical care. Colonoscopy is the 240 standard of care for colorectal cancer screening, and all patients will receive standard bowel preparation 241 instructions in addition to optional phone communication with endoscopy staff prior to the procedure. 242 The only research-related activity is the randomization of participants to a text messaging program that 243 will supplement routine practice when an individual schedules a colonoscopy at our institution. Since 244 this study is intended to promote the standard of care for colorectal cancer screening, and the receipt of 245 text messages poses negligible risk to the patient, we believe that waiver of consent is appropriate. 246 247

Impact on Subject Rights and Welfare* 248
Participation in the intervention is completely voluntary and subject rights and welfare will not be 249 adversely affected by the waiver of authorization and consent. Participants randomized to the 250 intervention arms still need to opt in to text messaging: Arm 1 will receive standard clinical care and Arm 251 2 will be enrolled into the text-messaging program in addition to standard clinical care. Placement in 252 either arm for the purposes of research does not adversely affect the rights and welfare of the subjects, 253 as each arm has the opportunity to engage in colorectal cancer screening, receive bowel preparation 254 instructions, and communicate with the endoscopy staff over the phone. Randomly assigning patients to 255 intervention and control does not alter patient's rights any differently than conducting an uncontrolled 256 pilot study where some patients receive the intervention and some don't. We believe it is appropriate 257 to obtain waiver of consent because learning the impact of different modalities for improving colorectal 258 screening has significant potential clinical value for the practice (i.e. significantly increasing the rate of 259 successful screening and thus decreasing the rate of mortality related to nonuse of screening). 260 261 Waiver Essential to Research* 262 The informed consent process itself may influence the outcome of our study, as it would require an 263 additional phone call by the research coordinate to ask for participation. The purpose of the Way to 264 Health text messaging program is to provide timely reminders regarding colonoscopy preparation. 265 Obtaining waiver of consent would allow us to avoid the potential selection/volunteer bias for inclusion 266 of patients who answer the phone and may be particularly motivated to complete colon cancer 267 screening. Since our main objective is to understand the potential influence of the text messaging 268 intervention on colonoscopy show rate and adequate bowel preparation quality among all patients who 269 agree to text messaging, we believe that obtaining informed consent separately from scheduling would 270 not allow us to evaluate the impact of the intended intervention, which is text messaging navigation 271 offered to all patients that agree during scheduling. Additionally, participants will be recruited via verbal 272 communication when they call our institution call center to schedule their colonoscopy. We believe the 273 research presents no more than minimal risk of harm to subjects and involves no procedures for which 274 written consent is normally required outside of the research context. The risks associated with this study are no more than minimal. There is the potential risk of breach of 286 confidentiality involving medical records reviews or text messaging which will be maintained on the Way 287 to Health platform. We will minimize this risk by using de-identified information whenever possible and 288 by maintaining all identifiable information on a secure drive and/or in a HIPAA-compliant system (e.g. 289 REDCap). In addition, all personnel will be held to high standards of upholding confidentiality and 290 safeguarding patient privacy. 291 292

Potential Study Benefits 293
The benefits of this study for participants include a platform that may improve the bowel preparation 294 experience prior to colonoscopy and increase the likelihood that a colonoscopy will not be hindered by 295 inadequate preparation or a missed/canceled appointment. This will occur through a set of curated text 296 messages with timely reminders as well as information with online links and a phone number to improve 297 accessibility for questions to be answered on a timely basis. It is also possible that the benefits for some 298 participants will be minimal. However, as mentioned, we believe the risks are also minimal. The control 299 group is unlikely to directly benefit, as this group will continue to receive usual care. The potential public 300 health impact of a successful intervention to improve colonoscopy show rates and bowel preparation 301 quality is significant and could increase the chances of identifying colorectal cancer at an early stage and 302 reduce the number of repeat colonoscopies and related costs due to inadequate bowel preparation. 303 Information learned from this study may benefit society through a better understanding of how to 304 effectively increase the rate of adequate colonoscopies which could increase the rate of colorectal 305 cancer screening and reduce the rate of mortality. 306 307 Alternatives to Participation (optional) 308 309

Data and Safety Monitoring 310
The study is minimal risk to participants and therefore the Principal Investigators and study team will 311 monitor the safety of this study on an ongoing basis. 312

Risk / Benefit Assessment 314
The risks associated with this study are no more than minimal. Better knowledge on how to increase 315 colonoscopy show rates and improve bowel preparation quality could potentially address one of the 316 major barriers to appropriate colorectal cancer screening, which is the second leading cause of cancer 317 death in the US. This study is designed to test an intervention with demonstrated feasibility and 318 successful preliminary results in a small QI pilot. For these reasons and those outlined in the above 319 benefits section, the investigators believe that the potential benefits outweigh the risks of participating 320 in the study. Colonoscopy is an effective screening technique for colorectal cancer (CRC) prevention, but many 341 patients either do not show up or have poor bowel preparation for the procedure. We plan to evaluate 342 the impact and feasibility of a text message-based program to provide patients with timely educational 343 and reminder messages regarding their upcoming colonoscopy and bowel preparation process. In this 344 pragmatic randomized controlled trial, we aim to (1) to provide text message-based educational and 345 reminder messages to patients or support partners regarding a scheduled colonoscopy, and (2) to 346 evaluate the impact of the texting intervention on colonoscopy show rate and bowel preparation. We 347 will include patients who are scheduled for outpatient colonoscopy at the Pennsylvania Presbyterian 348 Medical Center. After enrollment, patients will be randomized 1:1 to usual care (arm 1) or the text 349 message-based intervention (arm 2, one week in duration). We will measure colonoscopy show rate 350 with good/excellent bowel preparation as the primary outcome. 351

Overall objectives 354
The specific aims of the study are to (1) to provide text message-based educational and reminder 355 messages to patients or support partners regarding a scheduled colonoscopy, and (2) to evaluate the 356 impact of the texting intervention on colonoscopy show rate and bowel preparation. 357

Primary outcome variable(s) 359
The primary outcome of the study will be colonoscopy show rate with good/excellent bowel preparation 360 (binary). 361

Secondary outcome variable(s) 362
Secondary outcomes of the study will include colonoscopy show rate, bowel preparation quality (rescue,  363 poor, fair, adequate, good, excellent; we will also collect Boston Bowel Prep Score if available [scale 0-364 9]), colonoscopy cancellation rate, timing of advance cancellation notification (days), colonoscopy 365 reschedule rate, and colonoscopy appointment no-show rate. A colonoscopy "show" will be defined as 366 a patient who attends their originally scheduled colonoscopy appointment. A "cancellation" will be 367 defined as an appointment that is canceled at least one day prior to the originally scheduled 368 colonoscopy appointment date. A "reschedule" will be defined as an appointment that is canceled 369 and rescheduled (for a future date) on the same day, at least one day prior to the originally scheduled 370 colonoscopy appointment date. Finally, as an additional secondary outcome, we will also collect the 371 proportion of patients who opt out of the texting program in the late phase of enrollment ( as it is also required if stool testing is positive. However, colonoscopy requires a complex process to 380 identify an escort, purchase the preparation, take a day off from work, adhere to a clear liquid diet, and 381 complete the split-dose preparation as recommended. This results in a significant no-show and 382 cancellation rate, along with suboptimal preparation quality, which can lead to non-adherence and 383 incomplete screening. 384 Current approaches to engaging patients include having nurses call patients before the procedure or 385 patient navigators. However, it is often difficult to get patients on the phone, and these interventions 386 can be costly, making it less scalable for clinical practices. Other interventions such as videos or mobile 387 apps have been limited by poor user experience or limited engagement with the patient. There is an 388 opportunity to leverage an automated text message navigation intervention using the Way to Health 389 (WTH) platform to improve patient engagement prior to colonoscopy completion. The WTH platform is a 390 Penn Medicine platform that is hosted on site at the University of Pennsylvania. The platform allows 391 custom text messages to automatically be sent to patients, in addition to bidirectional message 392 capabilities. WTH is protected by a secure firewall and is a HIPAA compliant platform. 393 In the past year, our team conducted a quality improvement pilot initiative using WTH that tested the 394 feasibility and impact of a one-week text messaging protocol for patients who were scheduled for 395 outpatient colonoscopy. The text messages sent to patients contained information about the 396 preparation process and instructions, expectations about the procedure, and reminders about location 397 and timing. Among the 21 patients enrolled in the pilot, we found high user acceptability and higher 398 colonoscopy show rates as compared to baseline values at PPMC (90% versus ~50%). As such we believe 399 that the texting intervention is feasible for testing in the context of a randomized controlled trial. 400 401

Study Design 402
Design 403 We will perform a pragmatic randomized controlled trial evaluating the impact of the text message-404 based intervention (arm 2) as compared to usual care alone (arm 1). Patient enrollment will proceed 405 through two pathways (early phase and late phase). In the early phase, 250 patients will be enrolled 406 through be performed with assistance from the call center at the Pennsylvania Presbyterian Medical 407 Center (PPMC), and/or through phone calls from a clinical research coordinator, using a script pre-408 approved by the University of Pennsylvania Institutional Review Board. After enrollment, patients will 409 be randomized 1:1 to the arms listed above through the Way to Health platform. Although patients in 410 arm 2 may be enrolled more than one weeks or months in advance of their colonoscopy, they will only 411 receive intervention text messages in the 7 days prior to the scheduled colonoscopy, in addition to two 412 text messages at the time of enrollment explaining the texting program. Usual care consists of a phone 413 call from the PPMC endoscopy staff in the week prior to colonoscopy, if the patient is able to be 414 reached. Patients are also given the endoscopy phone number, and may call to speak to staff if they 415 have specific questions about their colonoscopy or need to reschedule. In the late phase, an additional 416 500 patients will be automatically enrolled in the Way to Health system using verified cell phone 417 numbers from the electronic medical record system, where they will be randomized in a 1:1 ratio to 418 the study arms in variable blocks of 8 and 4. Patients in arm 2 will receive two enrollment text 419 messages explaining the texting program and providing the opportunity to opt out (by replying with 420 the text STOP). A detailed outline of the study design can be found in the procedures sections. 421 422

Study duration 423
The duration of participation for patients in the intervention will be from the time of enrollment to the 424 date of scheduled colonoscopy. However, the patient will only receive intervention text messages in the 425 week prior to the scheduled colonoscopy, with the exception of the two enrollment messages in the 426 late phase. We will plan to recruit patients from November 2018 through January June 2019. We 427 estimate that the study will be completed by Participants in the study will be planned for outpatient colonoscopy at the Pennsylvania Presbyterian 466 Medical Center (PPMC). Once a patient is confirmed to meet eligibility criteria, patient enrollment will 467 be performed proceed based on one of two pathways (early and late phase). In the early phase (first 468 250 patients), enrollment will be performed with assistance from the call center at PPMC, and/or 469 through phone calls from a clinical research coordinator. When patients have a physician order for a 470 colonoscopy, the patient must contact the call center in order to schedule the colonoscopy. The call 471 center staff have agreed to discuss the research project and enroll patients using a script to be approved 472 by the IRB. Similarly, The research coordinator will call newly scheduled patients to discuss the 473 research project and enroll patients using a script to approved by the University of Pennsylvania IRB.

474
If patients are scheduled for colonoscopy in person while at the office, they will receive an 475 informational flyer about the study that gives them the opportunity to contact a study team member 476 for additional information or to enroll. It also informs them that someone from the study may contact 477 them directly about participating. After enrollment, patients will be 1:1 randomized to arm 1 or arm 2 478 using the Way to Health platform. 479

480
In the late phase (additional 500 patients), patients with upcoming colonoscopy appointments will be 481 screened for eligibility as per the selection criteria. As in the early phase, if patients are scheduled for 482 colonoscopy in person while at the office, they will receive an informational flyer about the study that 483 lets them know they may be contacted by text message about their upcoming procedure. Those who 484 qualify will have their validated cell phone numbers imported into Way to Health, where they will 485 subsequently be randomized. Those in the intervention arm will receive two enrollment messages 486 describing the study and providing the opportunity to opt out (by replying with the text STOP). Of 487 note, if a patient in the late phases chooses to opt out and not receive the texting intervention, their 488 data will still be analyzed in the intervention group (as an intention-to-treat approach). Importantly, 489 we are requesting a waiver of consent for several reasons: (1) the study presents no more than minimal 490 risk of harm to participants, (2) the patients who are randomized to either arm will not be deprived of 491 any standard care available in the status quo and thus the study will not violate their rights or welfare, 492 (3) the study could not be practicably completed without such a waiver, as the scheduling process is 493 handled over the phone by call center staff, and (4)  Screening -Phase 1: As above, patients will be screened and enrolled through one of two pathways 517 (early and late phase). In the early phase (first 250 patients), Patients who have a physician order for a 518 colonoscopy will call the PPMC call center in order to schedule their procedure. At this time, the call 519 center staff will screen the patient for eligibility (age criteria and text messaging ability) and will explain 520 the details of the study following a pre-determined script. Patients who agree to participate in the study 521 will have their cell phone number recorded. A list of these patients will be forward to the clinical 522 research coordinator on a weekly regular basis, who will confirm patient eligibility and then manually 523 enter patients into the WTH platform for randomization. Alternatively, a PPMC practice manager will 524 forward a list of scheduled colonoscopies to the clinical research coordinator (CRC) on a regular basis.

525
If patients are scheduled for colonoscopy in person while at the office, they will receive an 526 informational flyer about the study that gives them the opportunity to contact a study team member 527 for additional information or to enroll. It also informs them that someone from the study may contact 528 them directly about participating. The CRC will screen for patients for eligibility and make up to 3 529 phone call attempts to reach the patient. Once on the phone, the CRC will explain the details of the 530 study using a script approved by the IRB and enter patients directly into the WTH platform for 531 randomization. In the late phase (additional 500 patients), patients will be screened for eligibility 532 through the same process as in the early phase. If they are scheduled for colonoscopy in person while 533 at the office, they will also receive an informational flyer about the study that lets them know they 534 may be contacted by text message about their upcoming procedure. For eligible patients, the CRC will 535 then manually enter the validated patient cell phone number from the electronic medical record 536 system into WTH. Randomization will then occur as described below, and intervention arm patients 537 will receive two enrollment messages. These messages will describe the texting program and provide 538 the opportunity for patients to opt out (by texting the word STOP). Of note, phone numbers will be 539 confirmed to represent cell phones by using a publicly-available lookup utility. Importantly, patients 540 who opt out will still be included in the intervention arm as an intention-to-treat analysis. Based on 541 our experience with the QI pilot described above, we estimate needing to screen between 500 and 542 approximately 1,000 600 patients to meet our enrollment targets (750 400 total patients). 543 Randomization -Phase 2: In the early phase, using a random number generator in the WTH platform, 544 patients will be 1:1 randomized to arm 1 or arm 2. Beginning in the late phase of enrollment, patients , 545 until the target will be randomized to arm 1 or arm 2 in a 1:1 ratio in variable blocks of 8 and 4 using 546 the WTH platform, until the target sample size of 375 patients per arm is reached. sample size of 200 547 patients per arm is reached. Randomization will be performed and recorded in WTH. 548 Chart review -Phase 3: after randomization and in tandem with phase 4 (below), the research 549 coordinator will perform a chart review to obtain patient demographic and comorbidity data 550 (hypertension, hyperlipidemia, diabetes, inflammatory bowel disease, obesity, active opiate 551 prescription). We will also collect data on the number of patients opting out in the late phase of 552 enrollment, as well as data on nursing phone calls which are performed as a component of usual care 553 (not called, called but could not reach patient, called and spoke with patient). All data will be inputted 554 into and stored in a secured RedCap database. Only key study staff will have access to the RedCap 555 database, which is stored on a secure firewall-protected server. 556 Intervention -Phase 4: Patients in arm 2 will begin receiving text messages per a pre-determined 557 protocol starting 7 days prior to the date of scheduled colonoscopy (the text messaging protocol is 558 attached to this application). Of note, if a patient in arm 2 cancels or reschedules their colonoscopy 559 after randomization, they will not receive any additional protocol text messages as part of this trial. 560 Patients in arm 1 and arm 2 will both receive usual care, which includes (1) bowel preparation 561 instructions that are delivered via mail or through a secure online messaging portal, (2) a phone call 562 from the endoscopy staff in the week prior to colonoscopy, and (3) the option to call the endoscopy staff 563 during business hours to have any questions answered on demand. 564 Outcomes data collection -Phase 5: After a patient's scheduled colonoscopy date has passed, the 565 research coordinator will review the medical record to gather additional data from the endoscopy 566 procedure. This will include the indication for the procedure, quality of bowel preparation (rescue, poor, 567 fair, adequate, good, excellent; Boston Bowel Prep Score will also be collected if recorded), and 568 completeness (cecum reached). Procedure status will also be recorded (canceled, no-show, 569 rescheduled, completed). If the procedure was canceled, the cancellation lead time (in days) prior to 570 scheduled colonoscopy will be recorded. 571 Statistical Analysis and Manuscript Preparation -Phase 6: The statistical analysis (detailed below) will 572 commence after completion of outcomes ascertainment for the entire cohort. 573 574

Analysis Plan 575
Descriptive analysis will include comparison of baseline demographics and major medical comorbidities 576 between study arms. Continuous and categorical data will be compared using the Wilcoxon Rank-sum 577 and Chi-squared tests, respectively. An alpha threshold of 0.05 will be used for statistical significance, 578 with two-tailed testing performed in all instances. For the primary analysis, we will perform a Chi-579 squared test of independence to compare the proportion of colonoscopies with good/excellent bowel 580 preparation between groups. In the secondary analysis, Chi-squared tests of independence will be 581 performed to compare colonoscopy status (show, no-show, reschedule, cancelation) as well as bowel 582 preparation quality (excellent, good, adequate, fair, poor) between groups. Finally, aAmong canceled 583 colonoscopies, median lead time of cancellation will be compared between groups using the Wilcoxon 584 Rank-sum test. Finally, for both primary and secondary analyses, we will consider multivariable 585 regression modeling to adjust for imbalances in possible confounders, if present despite 586 randomization. 587 Consent 588

Consent Process 589
Overview 590 We are requesting a waiver of informed consent for this study. The reason for our request is explained 591 in the Waiver of Consent section. Verbal consent for text messaging will still be obtained from all 592 participants in the study. Please see the script templates attached to this application. 593 594 Children and Adolescents 595 Not applicable. We are only enrolling subjects 18 years of age and older. 596 597

Adult Subjects Not Competent to Give Consent 598
We are requesting a waiver of informed consent for this study but patients will still need to verbally opt 599 in to text messaging. We plan to enroll only those patients who are competent at time of enrollment to 600 opt-in for themselves. 601 602

Waiver or Alteration of Informed Consent* 604
Waiver or alteration of required elements of consent. 605 606

Minimal Risk* 607
This study is minimal risk as all participants will be receiving standard clinical care. Colonoscopy is the 608 standard of care for colorectal cancer screening, and all patients will receive standard bowel preparation 609 instructions in addition to optional phone communication with endoscopy staff prior to the procedure. 610 The only research-related activity is the randomization of participants to a text messaging program that 611 will supplement routine practice when an individual schedules a colonoscopy at our institution. Since 612 this study is intended to promote the standard of care for colorectal cancer screening, and the receipt of 613 text messages poses negligible risk to the patient, we believe that waiver of consent is appropriate. 614 615

Impact on Subject Rights and Welfare* 616
Participation in the intervention is completely voluntary and subject rights and welfare will not be 617 adversely affected by the waiver of authorization and consent. Participants randomized to the 618 intervention arms still need to opt in to text messaging: Arm 1 will receive standard clinical care and Arm 619 2 will be enrolled into the text-messaging program in addition to standard clinical care. Placement in 620 either arm for the purposes of research does not adversely affect the rights and welfare of the subjects, 621 as each arm has the opportunity to engage in colorectal cancer screening, receive bowel preparation 622 instructions, and communicate with the endoscopy staff over the phone. Randomly assigning patients to 623 intervention and control does not alter patient's rights any differently than conducting an uncontrolled 624 pilot study where some patients receive the intervention and some don't. We believe it is appropriate 625 to obtain waiver of consent because learning the impact of different modalities for improving colorectal 626 screening has significant potential clinical value for the practice (i.e. significantly increasing the rate of The benefits of this study for participants include a platform that may improve the bowel preparation 662 experience prior to colonoscopy and increase the likelihood that a colonoscopy will not be hindered by 663 inadequate preparation or a missed/canceled appointment. This will occur through a set of curated text 664 messages with timely reminders as well as information with online links and a phone number to improve 665 accessibility for questions to be answered on a timely basis. It is also possible that the benefits for some 666 participants will be minimal. However, as mentioned, we believe the risks are also minimal. The control 667 group is unlikely to directly benefit, as this group will continue to receive usual care. The potential public 668 health impact of a successful intervention to improve colonoscopy show rates and bowel preparation 669 quality is significant and could increase the chances of identifying colorectal cancer at an early stage and 670 reduce the number of repeat colonoscopies and related costs due to inadequate bowel preparation. 671 Information learned from this study may benefit society through a better understanding of how to 672 effectively increase the rate of adequate colonoscopies which could increase the rate of colorectal 673 cancer screening and reduce the rate of mortality. 674 675 Alternatives to Participation (optional) 676 677

Data and Safety Monitoring 678
The study is minimal risk to participants and therefore the Principal Investigators and study team will 679 monitor the safety of this study on an ongoing basis. 680 681

Risk / Benefit Assessment 682
The risks associated with this study are no more than minimal. Better knowledge on how to increase 683 colonoscopy show rates and improve bowel preparation quality could potentially address one of the 684 major barriers to appropriate colorectal cancer screening, which is the second leading cause of cancer 685 death in the US. This study is designed to test an intervention with demonstrated feasibility and 686 successful preliminary results in a small QI pilot. For these reasons and those outlined in the above 687 benefits section, the investigators believe that the potential benefits outweigh the risks of participating 688 in the study. 689 -modify CRC call script -modify "welcome text" message for pts in intervention arm -modify inclusion/exclusion criteria -add description of procedure if arm 2 patient cancels or reschedules colonoscopy -add lead time of cancellation as an outcome measure (in days) 4/4/2019

Summary of Protocol Changes
Modification: -modify sample size target -distinguish early and late phase enrollment