Home Hazard Removal to Reduce Falls Among Community-Dwelling Older Adults

Key Points Question What is the effect of a home hazard removal program on hazard of a fall in community-dwelling older adults at risk for falling? Findings This randomized clinical trial included 310 older adults. There was no difference in the primary outcome, risk of falling, between the intervention and control groups. The intervention group had a 38% reduction in falls, a secondary outcome, compared with usual care. Meaning This randomized clinical trial found that this low-cost behavioral intervention did not result in a significant difference in risk of falling or the number of home hazards removed.

Falls remain the leading cause of injury, long-term disability, premature institutionalization, and 75 injury-related mortality among older adults and are projected to cost nearly $59.7 billion by 76 2020. [12][13][14][15]71 Fall-related mortality increases with age. 72 Although numerous clinical interventions 77 have demonstrated efficacy in reducing the incidence of falls, 73 translation to the community has 78 failed; fall rates continue to escalate. Indeed, nearly one third of all adults older than 65 years fall 79 each year. The majority of falls experienced by older adults, particularly more frail high risk 80 older adults, occur in the home, 12, 24, 25 and measurable home hazards are associated with an 81 increased risk of older persons falling in the United States. 74,26 82 83 Setting. This intervention will be implemented and evaluated in the urban core of St. Louis, 84 Missouri, an area of high need for an effective fall prevention program. In Missouri, the fall 85 death rate for older adults is consistently higher than the national average. The rate nearly 86 doubled between 2000 and 2009 from 38 to 72.32 per 100,000. 75 Falls are the leading cause of 87 unintentional injury-related hospitalizations and ED visits among older Missourians; older adults 88 account for 64% of all ED visits and hospitalizations resulting from falls, which is higher than 89 national rates. 76, 77 90 91 Our long-term goal is to improve older adults' ability to maintain their independence and safety 92 in the community by translating fall prevention research into effective community programs. We 93 plan to establish the effectiveness of a program that demonstrated a reduction in falls among 94 community-dwelling older adults at high risk for a fall. It is a practical and scalable intervention 95 recommended by the Centers for Disease Control and Prevention(CDC). 1 The intervention will 96 be delivered by OTs in an Area Agency on Aging (AAA)-part of the Aging Services Network 53 97 in the US. We will use community-engaged research approaches 54 and a hybrid design 55 to study 98 both the effect of the intervention in a real world setting and the implementation strategy. We 99 will prepare for dissemination 56 by providing cost estimates, fidelity metrics, and intervention 100 manuals. 101 102 Design of study: 103 To examine the effect of a home modification (home hazard removal) program, we will conduct 104 a hybrid effectiveness/implementation trial of 300 older adults at risk for a fall randomized to a 105 home hazard removal program or usual care, and follow them for 12 months. We will conduct a 106 pragmatic randomized controlled trial/implementation study 107 108 Recruitment: We will recruit participants through the SLAAA's annual assessment of 109 participant's health. . 110 111 Deliver intervention: Interventionists will use a manualized protocol to deliver the intervention. 112 113 iv Follow up: We will follow participants monthly using a highly successful calendar journal 114 technique that has demonstrated a 99% response rate in previous studies. 115 116 Dissemination: We will disseminate the program, manuals and training programs using our 117 advisory network. 118 119 Objectives of study: 120 1. Determine the acceptability costs and feasibility of delivering the program in an Area 121 Agency on Aging. We will conduct a process evaluation 78 of the intervention to aid in 122 dissemination and interpretability of the trial by evaluating acceptability, feasibility, safety, 123 and cost. We will test the hypothesis that home modification interventions will have high 124 acceptability (80% retention), high fidelity by therapists (95% of elements; 90% of dose 125 delivered), low safety risk (no increased rate of falls compared to usual care group), and high 126 adherence (80% of modifications in use) at 12 months. 127 128

Determine if the program is effective in reducing the rate and risk of falls and improving 129
the outcome of other participant-reported outcomes (PROs) that can be effected by the 130 intervention. We will test the hypothesis that older adults who receive the program will have 131 a lower rate and risk of falls, improved self-efficacy, daily activity performance, and quality 132 of life compared with the usual care group. 133 134 Expected outcomes: The central hypothesis of this study is that home modifications (home 135 hazard removal) will be acceptable to older adults, practical for delivery in an existing social 136 service network, and efficacious in reducing falls and maintaining independence for older adults. 137 Our research team has a unique skill set to address this problem, including: a long relationship 138 with an AAA, experience studying home modifications in the community, experience conducting 139 pragmatic trials, and expertise in dissemination and implementation science.

Primary Objective 150
Determine the acceptability and feasibility of delivering the program in an AAA. To 151 examine the effect of the adapted program, we will conduct a hybrid 152 effectiveness/implementation trial of 300 older adults at high risk for a fall who will be 153 randomized to the adapted program or usual agency care and then followed for 12 months. 154 We will conduct a process evaluation 78 of the intervention to aid in dissemination and 155 interpretability of the trial by evaluating acceptability, feasibility, safety, and cost. We will 156 test the hypothesis that home-modification interventions will have high acceptability (80% 157 retention), high fidelity by therapists (95% of elements; 90% of dose delivered), low safety 158 risk (no increased rate of falls compared with the usual care group), and high adherence (80% 159 of modifications in use) at 12 months. 160 161

Secondary Objectives 162
Determine whether the adapted program is effective in reducing the rate and risk of falls 163 and improving the outcome of other participant-reported outcomes that can be affected by the 164 intervention. We will test the hypothesis that older adults who receive the program will have a 165 lower rate and risk of falls and improved self-efficacy, programs. This study will be a substantial departure from previous studies. The research 226 proposed in this application is innovative, in our opinion, because it represents the first time an 227 effective home-modification intervention to prevent falls will be implemented through a US 228 social service delivery system. Successful translation into practice requires an understanding of 229 the interaction among an intervention, the systems in which it is implemented, and the 230 stakeholders. 42 We will use community-engaged research approaches 43-46 to determine 231 adaptations necessary to deploy a home-modification intervention to prevent falls in the US 232 within a reliable aging services network. We will reduce barriers to program participation by 233 involving key stakeholders as part of the research team 47, 48 and clarifying the benefits of the 234 research to the community. 49 235 236 The goal of this study is to implement an evidence-based home-modification intervention for 237 older adults that is designed to reduce the incidence of falls through a community program 238 delivered through the aging services network with high acceptance. The program will be 239 designed to reduce the incidence of falls among community-dwelling older adults so they may 240 independently remain in their own homes as long as possible. 241

STUDY DESIGN 242
To examine the effect of a home-modification 243 (home-hazard removal) program, we will conduct 244 a hybrid effectiveness/implementation trial of 245 300 older adults at risk for a fall who will be 246 randomized to a home-hazard removal program 247 or usual care and then followed for 12 months. 248 Figure 5 outlines the milestones of a pragmatic 249 randomized controlled trial and implementation 250 study. Study milestones are further detailed in the 251 timeline (under Section E.1, Table 2). 252 Recruitment: We will recruit participants 253 through the SLAAA's annual assessment of 254 participants' health. We estimate more than 800 255 older adults will meet eligibility criteria. 256 Deliver intervention: Interventionists will use a 257 manualized protocol to deliver the intervention. 258 Follow-up: We will follow participants monthly 259 using a highly successful calendar journal 260 technique that has demonstrated a 99% response rate in previous studies. Dissemination: We 261 will disseminate the program, manuals, and training programs using our advisory network. 262 263

Inclusion/Exclusion Criteria 266
Inclusion criteria: 1) aged 65 years or older; 2) self-report of one previous fall or more in the 267 preceding 12 months or self-report as "worried about falling;" and currently receiving services 268 from SLAAA. 269 Exclusion criteria: 1) residents of nursing homes; 2) individuals with severe cognitive 270 impairment who are unable to give consent to participate (as determined by a score of greater 271 than 10 on the SBT for memory 10 ) 272 percent are female and 74% are African American. More than 25% reported a fall in the past 12 280 months, and 40% are "worried about falling." The average score on ADL performance was 4 281 (range, 1-7); indicating moderate ADL impairment). We will recruit from this cohort of older 282 adults who are "at high risk for a fall." 79 We estimate more than 800 older adults will meet 283 eligibility criteria in the cohort. 2, 3 The SLAAA staff will identify potential participants and 284 assist in recruiting participants at no cost 285 286 We will recruit 300 older adults, conduct baseline assessments, and randomize participants to 287 receive the adapted program or usual care. Outcomes (falls, fall self-efficacy) will be assessed by 288 a blinded evaluator at baseline and 12 months after intervention. Flow through the study is 289 illustrated in Figure 5. All assessments and study visits will occur in participants' homes. 290 291

Study Enrollment Procedures
Recruitment. We will randomly select a sample of older adults from the large observational 292 cohort of older adults assessed annually by SLAAA via NAPIS as part of service delivery 80 and 293 invite them to participate by telephone. Refusals will be tracked for attribution of cause. 294 295 Consent and enrollment. Consent will be obtained at the baseline home assessment visit. 296 Potential participants will be provided with a study brochure and a copy of the large-print 297 consent form. Scripted details of the project including the randomization procedure, study 298 interventions, and follow-up procedures will be discussed (under Section C.2.c). 299 300 Randomization sequence generation, allocation concealment, and implementation. 301 Participants will be allocated using a 1:1 ratio via randomization sequences generated a priori 302 using a computerized formal probability model. The allocation ratio will be maintained at 303 periodic intervals. Groups will be balanced with regard to race (self-reported African American 304 or "other") and sex. Randomization sequence concealment will be achieved by the Research 305 Electronic Data Capture (REDCap) system. 81 After baseline assessment, the study coordinator 306 will verify that all eligibility requirements are satisfied and then elicit treatment assignment. 307 6 STUDY INTERVENTIONS 308

Study Interventions 309 310
Control group (usual care). During the treatment phase of the study, each participant 311 randomized to control will receive usual care from SLAAA, which includes annual NAPIS 312 evaluation (in home) and individualized referrals to services based on identified need. Referrals 313 are generated by case managers or SLAAA staff based on the annual assessment or if older 314 adults (or their family members) call and request a referral for services. 315 316 Procedures for intervention (treatment). Participants who are randomized to treatment will 317 receive usual care plus the home-modification (home-hazard removal) intervention. The 318 intervention, based on a competence/press theoretical framework, 96 will be provided by 319 registered and licensed OT. The intervention will be delivered via a manualized intervention 320 using a visit by visit checklist. Home modifications 79 include a 2-hour home assessment 321 conducted by an OT in the participant's home using the WeHSA. A comprehensive clinical 322 assessment will be performed on each participant to determine functional limitations that may 323 interact with hazards in the home and result in a fall. Environmental hazards and unsafe 324 behaviors will be identified. A barrier-removal plan will be provided to the older adult. The OT 325 will facilitate home modifications through housing services programs and provided follow-up 2 326 weeks after the initial home visit to ensure that the participant adopted the recommended 327 modifications. It is anticipated that 2-3 visits of 90 minutes each will be provided. At about 6 328 months participants will receive a booster session (in home visit by an OT) to reevaluate 329 performance in the home and modify the intervention plan as needed. 330 331 Interventionist Training 332 Each interventionist will receive a 3 hour didactic training and additional 2 hours of guided self-333 directed learning. 334 335 5.2 Process Evaluation 336 We will evaluate a set of implementation outcomes distinct from clinical treatment outcomes 99 337 that are relevant for decision makers and important for implementation and dissemination of the 338 intervention. The components and data sources of the process evaluation are presented in Table  339 1. We will conduct a process evaluation to understand the fidelity, safety, cost, and quality of the 340 intervention. 100 We will monitor recruitment metrics for all eligible participants, including 341 demographic characteristics, enrollment status, and reason for decline. We will determine the 342 fidelity of the intervention by determining whether the intervention was delivered as planned and 343 calculating the dose of the intervention received by both groups (number of minutes and number 344 of sessions delivered). OTs will rate adherence at the final visit as the number of 345 recommendations implemented per recommendations suggested. 6 Long-term adherence will be 346 calculated as the number of recommendations used at 12 months per recommendations 347 suggested. We will examine reasons for abandonment of strategies to inform the final program. 348 We will explore the safety of the intervention by determining the rate and severity of falls 349 (calculated with an algorithm). 101 Environmental modifications are provided through a 350 patchwork of services in the US; therefore, reliable cost data are not available. We will track the 351 cost of service provision (in the treatment group). We will record OT time to provide services 352 (direct and indirect) 102 and costs of equipment issued to participants. Costs of contracting 353 services will include installation (labor and materials) and maintenance or repair during follow-354 up. All OT time will be recorded using current procedural terminology codes. The hourly wage of the 355 contractor will be estimated using wage rates for the St. Louis area as reported by the U.S. Bureau of 356 Labor Statistics. 103 The cost of all materials will be captured from invoices. Costs will be estimated from 357 the payor perspective (e.g., Medicare allowable costs will be used to estimate covered clinical 358 encounters    Informed Consent Process. All elements of consent will be reviewed with older adults prior to 382 enrolling in the study, including the purpose of the study, risks, benefits, alternatives to the study, 383 how confidentiality will be maintained, the PI's contact information, no consequences to 384 withdrawal, and how study results will be shared. All study staff have undergone IRB training. In 385 addition, all staff members have participated in cultural competence training and are trained to 386 interview older adults. All staff members are trained to obtain assent, which is particularly 387 important for older adults with low vision, low literacy levels, or limited English proficiency. All 388 study documents are available in multiple formats, are developed in 16-point Arial font, and are 389 written in plain language. All informed consent documents are reviewed by the laboratory 390 advisory board of older adults (involved in the design of the study) to ensure clarity for older 391 adults. 392 Steps to understand the elements of informed consent. Written informed consent to participate in 393 the study will be obtained before any test or measurement is performed. Eligibility will be 394 confirmed prior to enrollment. All potential participants will be screened using a script by the 395 coordinator. Screening will take place in person or over the phone. Potential participants will be 396 given a copy of the large-print informed consent document and a study review sheet written at 397 the appropriate reading level to review. During screening, all participants (1) will have a 398 detailed, plain-language explanation of the study and what is expected of them; (2) will discuss 399 potential problems that could interfere with participation; (3) will have their questions answered; 400 and (4) will receive a plain-language summary of the study and contact information for the PI 401 and study coordinator. Consent will occur during the first home assessment visit. The consent 402 form will be signed by a witness and stored in the office of the PI under double locks. 403 Participants will be advised in the consent form that there is a possibility that their medical 404 research record, including identifying information, may be inspected and photocopied by 405 Participants will be allocated using a 1:1 ratio via randomization sequences generated a priori 413 using a computerized formal probability model. The allocation ratio will be maintained at 414 periodic intervals. Groups will be balanced with regard to race (self-reported African American 415 or "other") and sex. Randomization sequence concealment will be achieved by the Research 416 Electronic Data Capture (REDCap) system. 81 The study coordinator will verify that all eligibility 417 requirements are satisfied and then elicit treatment assignment prior to baseline assessment. 418

Baseline Assessment 419
Demographics. Demographic and other pertinent information, including assistance received, 420 adaptive equipment usage, and falls, will be collected. 421 422 OARS ADL scale. The OARS ADL scale [51] will be used to screen for functional performance. 423 The OARS ADL has been validated for the older adult ED population [52] , is brief and easy to 424 administer. Potential participants are asked about their ability to perform each of 14 activities. 425 Responses are scored on a 0-2 scale, with higher scores indicating greater independence. 426 Individuals who receive a score of 0 on one or more activities or a score of 1 on two or more 427 activities will qualify for the study. 428 429 Falls Efficacy Scale-International (FES-I). The Falls Efficacy Scale-International (FES-I) [53] will 430 be used to assess participants' self-efficacy in performing daily activities without falling. The 431 FES-I has good test-retest reliability and validity and consists of a list of 16 daily activities (e.g., 432 getting in and out of bed) on which respondents rate their confidence doing each activity without 433 falling on a scale from 1 (very confident) to 10 (no confidence at all). Total FES-I score is the 434 sum of each activity score, with higher scores indicating greater fear of falling. Short Michigan Alcoholism Screening Test -Geriatric Version (S-MAST-G). The S-MAST-450 G [58] will be used to screen for the presence of an alcohol problem in potential participants. The 451 S-MAST-G has been validated for the older adult population [59] . Potential participants will 452 answer 10 yes/no questions about their alcohol consumption in the past year; two or more "yes" 453 responses indicate a likely alcohol problem.. 454 455 Geriatric Depression Scale Short Form (GDS-SF). The GDS-SF [60] will be used to assess 456 depression levels in participants. The GDS-SF is a 15-item self-report questionnaire specifically 457 designed and validated with the older population, and correlates highly with the original 30-item 458 GDS (r=.84) [60] . Participants answer yes-or-no questions about their feelings in the past week. 459 Total scores range from 0 to 15, and scores of 5 or more indicate probable depression. 460 461 Short Blessed Test (SBT). The SBT [61] will be used to screen for severe cognitive impairment. 462 The SBT has good validity for older adults [61] and correlates well with the Mini-Mental State 463 Exam [62] . This brief, six-item questionnaire assesses cognition, memory, and orientation. 464 Incorrect items are scored as the product of the number of errors made and a weighting factor; 465 the sum of incorrect item scores comprises the total score. Total scores range from 0-28; scores 466 of 10 or more are consistent with dementia and other cognitive difficulties. Individuals who 467 receive a SBT score 10 or more will not be allowed to participate in the proposed study. Participants randomized to the intervention group will receive the intervention over 2 visits and 490 one booster session at 6 months according to a manualized intervention. Briefly, for the 491 intervention group, the occupational therapist will begin the intervention during the WeHSA. 492 During the WeHSA the occupational therapist will assist the older adult to identify home 493 hazards, work through potential solutions to reduce or remove the hazard, if possible, implement 494 solutions (such as removing a throw rug), create plan for additional hazard removal, begin self-495 management training for behavioral and environmental fall risks, and provide training to 496 complete the calendar journal. Between visits, the occupational therapist will assist the older 497 adult to obtain equipment or home modifications. At the second visit, the occupational therapist 498 will provide any necessary installation of equipment (such as a tub transfer bench), explain the 499 purpose of each new equipment, teach techniques to safely use the equipment, and observe the 500 older adult safely using the equipment. The occupational therapist will also continue the self-501 management training to improve daily activity performance and safety. 502 7.3.3.2 5 month adherence 503 A phone call to evaluate adherence will be made at approximately five months (in anticipation of 504 the 6 month booster session). 505 7.3.3.3 Six month Booster 506 A 6-month booster session will also be provided to assist the older adults with any new fall 507 hazards and to reinforce self-management/mindfulness training techniques. To identify 508 additional fall hazards, the occupational therapist will complete the WeHSA with the older adult 509 and make recommendations for solutions or provide additional training. To record falls, an individually customized calendar journal will be designed for each participant 512 for daily recording of falls [48] . Our method of tailoring the calendar journal includes life event 513 date anchors, incentives and personalization [49] , we will construct the participant-tailored 514 calendar journals [50] . There will be space to record the primary outcome of falls (fall or no fall) 515 each day. If a participant falls, they will be instructed to complete a fall form. Calendar journal 516 pages and fall forms will be collected monthly via U.S. mail. Falls will be verified with a follow-517 up telephone interview by a community coordinator. Follow up assessments will conducted by trained raters blind to group allocation. The follow-up 520 sessions will repeat assessments conducted at the first study home visit and first phone call. 521

Completion/Final Evaluation 522
List each assessment to be performed at the participant's final visit. 523 Falls behavioral Scale for Older People (FaB). Protective behavioral factors to prevent falls will 524 be assessed using the FaB. This instrument includes ten behavioral dimensions: cognitive 525 adaptations, protective mobility, avoidance, awareness, pace, practical strategies, displacing 526 activities, being observant, changes in level, and getting to the phone [66] . 527 528 OARS ADL scale. The OARS ADL scale [51] will be used to screen for functional performance. 529 The OARS ADL has been validated for the older adult ED population [52] , is brief and easy to 530 administer. Potential participants are asked about their ability to perform each of 14 activities. 531 Responses are scored on a 0-2 scale, with higher scores indicating greater independence. 532 Individuals who receive a score of 0 on one or more activities or a score of 1 on two or more 533 activities will qualify for the study. 534 535 Falls Efficacy Scale-International (FES). The Falls Efficacy Scale-International (FES-I) [53] will 536 be used to assess participants' self-efficacy in performing daily activities without falling. The 537 FES-I has good test-retest reliability and validity and consists of a list of 16 daily activities (e.g., 538 getting in and out of bed) on which respondents rate their confidence doing each activity without 539 falling on a scale from 1 (very confident) to 10 (no confidence at all). Total FES-I score is the 540 sum of each activity score, with higher scores indicating greater fear of falling. 541 542 Patient Reported Outcomes Measurement Information System (PROMIS) Global Scale. Self-543 rated health will be measured using the Patient Reported Outcomes Measurement Information 544 System (PROMIS) Global Scale [54] which measures general perceptions of health. We will use 545 two scales: global physical health (overall physical health, physical function, pain, and fatigue) 546 and global mental health (quality of life, mental health, satisfaction with social activities, and 547 emotional problems implemented and reasons that any recommendations were not implemented. Initial adherence 576 will be a proportion of the number of recommendations implemented /recommendations 577 suggested. Finally, we will determine the number of recommendations used/recommendations 578 suggested at 6 months. Reasons for abandoning strategies will be examined using the Adherence 579 Log; in which the participant will report on current level of use for each modification: very 580 useful, somewhat useful, not at all useful, no longer use equipment. Any independence that was 581 regained by improved sensorimotor performance will not count against the adherence rating. 582

Safety 583
To explore the safety of our study, we will measure the number of falls and the circumstances 584 surrounding the falls with a self-report Fall Form used at the 6 month follow up visit. The rate 585 and severity of the falls will be calculated using a standardized algorithm established by Tinetti, 586 et al, 1988[ 38 ]. We will also track both health care utilization for participants in both groups (# of 587 emergency room and out patient visits, # of hospitalizations, and number of days in therapy) 588 using the Health Care Utilization Form. 589 7.4.6 Cost 590 The cost of the treatment will be measured by tracking cost of modifications and adaptive 591 equipment for each participant. This will be tracked using the Invoice Form which includes both 592 costs from the contractor as well as costs of any equipment ordered from a medical supply 593 company or obtained from a community resource (medical equipment lending program hospitalization, results in persistent or significant disability/incapacity, or is a congenital 620 anomaly. 621

Reporting Procedures 622
All serious adverse events will be reported to the HRPO in the falling time frames: a) death -623 immediately; b) life-threatening within 7 calendar days; c) all other SAEs within 15 calendar days using 624 the Electronic Serious Adverse Event Reporting System. Should there be a serious adverse event that 625 occurs that increases the risks to the participants, the study will be stopped, an investigation will be 626 conducted, and a findings report will be generated before the study is resumed.

Follow-up for Adverse Events 628
AEs will be followed until resolved or considered stable for the study period. 629

INTERVENTION DISCONTINUATION 630
Safety data will be collected on any participant discontinued due to an AE or SAE. In any case, 631 every effort must be made to undertake protocol-specified safety follow-up procedures. If 632 voluntary withdrawal occurs, the participant will be asked to continue scheduled evaluations, 633 complete an end-of-study evaluation, and be given appropriate therapy until the symptoms of any 634 AE resolve or the participant's condition becomes stable. 635 10 STATISTICAL CONSIDERATIONS 636

Sample Size and Randomization 637
Power computations reflect two general principles. 638 The first is that we will be conservative by 639 underestimating and overstating power. Second, 640 estimates of statistical power will be consistent with 641 the method of analyses. Our primary outcome is time 642 to event; therefore, we will determine the sample size 643 and power within the survival data analyses 644 framework. and because our intervention can be defined as a high-intensity home modification, we project a 659 30% or greater reduction in fall risk at 12 months after treatment. Figure 6 displays the power as 660 a function of total sample size given a fall rate of 42% (event-free rate of 0.58) in the 661 intervention group and a fall rate of 0.61 (event-free rate of 0.39) in the control group, with a 662 20% attrition rate. A two-sided log-rank test with an overall sample size of 300 participants (150 663 in each group) achieves 84% power at a 0.05 significance level to detect a difference of 0.19 664 between 0.39 and 0.58-the proportions without falls in the control and intervention groups, 665 respectively. This corresponds to a hazard ratio of 0.5785 of fall in intervention versus control 666 groups. These results are based on the assumption that the hazard rates are proportional. 667

Data Analyses 668
We will examine the process data before the efficacy analyses. 107 We will use descriptive 669 statistics to describe recruitment rates and reasons for not enrolling. To explore the extent and 670 quality of the implementation of the intervention, we will compare the characteristics of patients 671 who complete the assigned intervention with those who do not for differences in fall risks. 672 Adherence and cost estimates will be conducted for participants in the intervention group. 673 Descriptive statistics will be used for adherence. We will test the hypothesis that the intervention 674 will have high acceptability (at least 80% retention), high fidelity by therapists (at least 95% of 675 elements and at least 90% of dose delivered), and high adherence to the program (at least 80% of 676 modifications in use) at 12 months. These hypotheses will be tested using a 1-sample 1-sided 677 proportion test. For costs, initial analysis will be conducted to determine what type of approach 678 will be necessary to make unbiased and relevant estimates of the cost of the intervention. We 679 will compare the effectiveness of the program and usual care with the time to the first fall over 680 12 months as the primary outcome of interest. Survival analysis techniques will be used where 681 time zero is the date on which the intervention is completed, and the ending time is 12 months 682 after the intervention. Participants who do not have a fall will be censored at the end of the study. 683 Participants who move to a nursing home will be censored on the date of moving. For those who 684 fall during follow-up, the time to the first fall will be the event time. Kaplan-Meier product limit 685 estimates will be used to describe the fall experiences for the intervention and control groups, 686 with the log-rank statistic testing for significant difference. Prespecified covariates (e.g., age, 687 race, sex, number of reported medications, mobility, ADL scores, and visual acuity) will be 688 adjusted using the Cox proportional hazards model. We will test the interventions effect using 689 the Cox proportional hazards model and adjusting for covariates. The regression coefficient of 690 the indicator variable quantifies the difference in the log hazard of time to first fall between 691 intervention and control groups. We will determine whether fall-prevention treatment is superior 692 to usual care on secondary outcomes (total number of falls, number of injurious falls) and 693 participant-reported outcomes (daily-activity performance, fall self-efficacy, and global health). 694 We expect that the intervention group will experience fewer total falls, fewer injurious falls, 695 greater daily-activity performance, greater self-efficacy, and greater quality of life compared 696 with the usual care group. 697

Data Collection Forms 699
Data will be collected at each follow up point by a blinded rater over the phone. Data will be 700 directly entered into an online data management software (see 11.2) Back up paper forms will 701 be available (See MOP). 702

Data Management 703
Data-entry and data-management systems have been built to ensure accurate and complete data 704 sets. Data will be entered via Web-based screens uploaded to a secure server. REDCap 705 provides secure, Web-based applications, including real-time validation rules with automated 706 data type and range checks at the time of data entry 81 . Data will be automatically backed up 707