Pragmatic Trial to Increase Quality of Care in State Veterans Homes

Background VA is highly invested in a struggling system in urgent need of help: State Veterans Homes. State Veterans Home nursing homes (SVHs) care for 51% of all Veterans receiving nursing home care paid for by VA. These 151 homes are owned and operated by states but constructed with 65% VA funding and certified and substantially financially supported by VA. Occupancy must be at least 75% Veterans; current average occupancy is 93% Veterans. VA paid $1.2 billion in FY19 to SVHs in per diem payments, which is expected to rise by FY24 to $1.7 billion. In addition, since the prior submission, the VA announced $1 billion in additional funding for SVH to aid with responses to the pandemic. And the Facility-Based Care Modernization Plan consolidates VA-led SVH oversight within the Office of Geriatrics and Extended Care, with an additional $13 million annually. SVHs are undoubtedly a critical component of VA's commitment to care for older Veterans. The SVH resident population is older than that in VA's Community Living Centers or VA-contracted community nursing homes: almost 50% of SVH residents are over age 84 (compared with 21% and 24%, respectively). In addition, 78% of SVH residents stay over 100 days (compared with 21% and 30%, respectively). The SVH system is also not centralized. SVHs lack connection to hospital QI infrastructures and, up to now, a national quality improvement (QI) infrastructure. Yet all follow VA SVH-specific regulations and will soon become Centers for Medicare and Medicaid-certified. A national SVH QI is also slowly coming. Falls are a critical safety issue, and huddles show promise as an intervention. The focus for this study-falls of any type-represents an important quality and safety measure that is also meaningful during the COVID pandemic. Falls are consistently the most reported SVH-related issue brief to VA Central Office. In FY20, 75% (255 of 339) of briefs related to falls. For older adults, falls are a leading cause of injury and death. A large database study of nursing home residents' premature deaths from external causes found 82% due to falls. Falls are unintended, not due to acute events (seizure, stroke) or external forces (push). In addition to injuries, falls result in reduced mobility, functional decline, and psychological stress. In nursing homes, they also result in substantial staff time and cost increases. Yet falls are also largely preventable and thus represent a prime target for interventions. Falls are typically caused by multiple, interacting factors, such as age, cognitive impairment, chronic conditions, balance impairment, lower extremity weakness, medications, and environmental hazards. Many of these are modifiable. A 2020 review of fall prevention interventions in nursing homes found evidence for the efficacy of some targeted staff-education and multi-factorial clinical interventions. One staff-focused intervention with a history of success in preventing falls is huddling. Huddles are brief, stand-up meetings to facilitate efficient information exchange. Nursing home and hospital studies have achieved significant reductions in fall rates using post-fall huddles and general fall-prevention huddles. One nursing home study, for example, attributed a 37% reduction in fall rates to the introduction of regular fall-focused huddles. Studies indicate these huddles improve communication and reflection, increase nursing and patient involvement, and reduce task and coordination errors. SVHs have high rates of any fall. This study's primary outcome of "fall during episode of care" had a FY20 Q3 rate, averaged across 4 quarters, of 55.4% of residents. The 10th to 90th percentile range was narrow: 44.4% to 65.9%. On average, over half the residents in an SVH experience at least one fall per quarter. The LOCK program is an evidence-based solution to address critical clinical issues. The LOCK program avoids reliance on a national or even strong local pre-existing QI infrastructure. Instead, the LOCK program itself teaches the fundamental QI structures and processes and helps each nursing home create its own. LOCK stands for (1) Learn from bright spots, (2) Observe, (3) Collaborate in huddles, and (4) Keep it bite-size, with huddle collaboration as the foundational element. The LOCK program is based on 4 evidence-based practices. (1) Strengths-based learning: identifying and learning from exceptional performance (positive deviance) by pinpointing already-existing solutions (bright spots) for overcoming hurdles and achieving success. (2) Observation: using observation to study human factors within a work system and gather data. (3) Relationship-based teams: improving healthcare outcomes by changing organizational practices and relational dynamics using frontline staff huddling. (4) Efficiency: keeping communication focused, direct, and brief and keeping elements of a change effort manageable. The program helps multi-disciplinary frontline staff teams (nursing, physician, housekeeping, physical therapy, etc.) work together to identify and discuss risk factors and use rapid-cycle QI techniques to pilot and monitor their actions. Actions that staff might take are individualized to the resident and may include medication review, occupational therapy, exercise, socialization, and a host of other options. Relational Coordination represents the critical ingredient for achieving lasting improvement Relational coordination is an evidence-based framework emphasizing that staff across all job types must be connected by high quality interactions and supportive relationships to create an environment in which teams achieve improved results. Healthcare organizations-including nursing homes-with high relational coordination achieve better clinical outcomes: higher quality of care, shorter length of stay, lower pain, and higher functioning. For staff working in hospital and nursing home settings, high relational coordination is associated with greater job satisfaction, greater job engagement, and less burnout. Objectives This 4-year hybrid (Type 2) effectiveness-implementation study uses a pragmatic stepped-wedge randomized trial design in 8 SVHs and relies on relational coordination theory and the RE-AIM framework. Aim 1: This study will use primary and secondary data collection to investigate the effectiveness of the LOCK program at improving resident falls and other resident clinical outcomes (mobility, medication changes, restraint, and alarm use) and work-process outcomes for staff (job satisfaction, work engagement, burnout). Aim 2: To support successful program implementation, this study will use the replicating effective programs framework and multi-modal implementation facilitation strategies. The program's reach, adoption, and implementation will be evaluated using mixed methods. Aim 3: This study will employ mixed methods to assess the extent of program sustainment (maintenance) at 3-, 6-, and 12-months post intervention and sustainment variability across sites. Study Design Replicating effective programs framework This study will use the replicating effective programs framework to examine the site-specific learning context and implementation barriers, focusing on the pre-implementation and implementation phases. Pre-implementation focuses on implementation preparation and obtaining broad buy-in. It encompasses tailoring, orienting, planning, training, and technical assistance. Implementation comprises ongoing support, training, evaluation, and feedback. Recruitment Methods SVH recruitment and eligibility The study team will work with the National Association of State Veterans Homes (NASVH) to recruit SVHs. NASVH leadership will send an email to SVHs describing the study. The study team will send a follow-up email to SVH administrators, asking them to participate. The study team will continue to send emails until a pool of at least 14 SVHs is reached, from which the research team will choose a sample of 8 SVHs. Staff recruitment SVH employees will be recruited for survey participation and qualitative interview involvement through emails from study team members, in-person during site visits, and at staff meetings using scripts written by research personnel. Survey recruitment The research team will invite all SVH staff from the 8 participating SVHs to complete online surveys. This survey will be administered at four time points: pre-intervention, baseline, post-intervention, 6- and 12-month follow up. Qualitative interview recruitment LOCK program implementation interviews: The research team will recruit the LOCK program leadership team members as well as a sample of frontline staff who participated in the huddles to participate in semi-structured qualitative interviews at two time points: baseline and post-intervention. LOCK Program Sustainment interviews: In addition, at the 3-month and 6-month follow up time points, the study team will conduct semi-structured qualitative phone interviews with key SVH staff involved in the intervention. These include the LOCK program leadership team members and staff who participated in frontline staff huddles. Resident recruitment At two time points (baseline and post-intervention) the research team will conduct semi-structured interviews with cognitively intact Veteran residents living in the participating SVHs who have a history of falling. Intervention implementation Pre-intervention phase - Training on and implementing frontline staff huddles: The leadership team will serve as the study's internal facilitators, responsible for spearheading and supporting staff in implementing the LOCK program during both the pre-intervention and intervention periods. In the pre-intervention phase, the investigators will use the investigators' team's extensive VA experience to teach the leadership team how to develop and run interdisciplinary, relationally coordinated frontline staff huddles. Intervention - LOCK program focused on fall prevention: The intervention begins at the baseline visit. The research team will hold a 4-hour training for each SVH's leadership team on risk factors for falling and how to guide staff to use their frontline staff huddles to continually identify residents at highest risk of falling, explore residents' personal histories, and develop individualized action plans based on risk factors. Staff will continually target a fluid "watch" list of residents at highest risk of falling. Consistent with the replicating effective programs framework, The research team will guide the teams to teach staff to use continual QI methods to monitor the impact of their action plans. They will use observation and data collection, including environmental assessments, and post-fall huddling to identify fall prevention intervention opportunities. AIM 1: Primary data collection, staff and resident Staff surveys: SVH staff will be recruited for research surveys via email. Surveys will be administered electronically using a VA-approved online survey portal electronic data capture. Residents discussed in huddles: As part of the huddling process, staff will keep a running huddle log that contains information on which residents are discussed and why. Data from the huddle logs will be used as process measures to determine how huddles are used to focus on residents potentially at risk for falling. Resident interviews: The study team will aim to recruit 3-5 residents per site. Consented residents will be interviewed; audio recordings will be collected via IRB-approved methods and transcribed verbatim. AIM 2: Data collection instruments RE-AIM evaluation overview For Aim 2, LOCK program's implementation will be evaluated using RE-AIM's first 4 (reach, effectiveness, adoption, and implementation) constructs. FRAME To assess any modifications to the intervention, research staff will use a Framework for Reporting Adaptations and Modifications-Enhanced (FRAME) checklist when they learn from sites that modifications were made. Implementation Facilitation Time Tracking Log: This tool, developed through a VA QUERI grant, documents the time, activities, and personnel involved in implementation facilitation efforts. External facilitators will record, on an at least weekly basis, facilitation event type, mode of communication, with whom they interacted, and specific activities, thereby also gathering information on tailoring and uptake of the intervention. Relational Coordination Survey: The survey asks participants to rate their experiences with their own and other workgroups on the 7 key dimensions of relational coordination (frequency, timeliness, accuracy, and problem-solving nature of communication; shared goals; shared knowledge; and mutual coworker respect). Qualitative interviews: LOCK Program implementation interviews with leadership team and staff: The research team developed a semi-structured interview guide, with undergirding from relational coordination and the RE-AIM framework. The guide will assess participants' impressions of the content, structure, and implementation of the LOCK program; how it affected staff interactions; who did/did not participate and why; how it affected resident outcomes, including unexpected results; barriers to or facilitators of implementing it; and how the SVH sustained it. Field notes: The research team will collect field notes during site visits using a structured template. It captures general impressions from informal conversations and overall impressions of the physical layout and atmosphere, staff-staff interactions, staff-resident interactions, processes, etc. AIM 2: Primary data collection, staff LOCK Program Implementation interviews: The research team will recruit the LOCK program leadership team members as well as a sample of frontline staff who participated in the huddles. The research team will use a semi-structured qualitative interview guide informed by relational coordination theory and RE-AIM to examine variation in implementation and sustainment and for analysis of barriers. AIM 3: Additional sustainment-specific instruments Clinical Sustainability Assessment Tool: To assess sustainment (maintenance), site program leadership and research team members will use the Clinical Sustainability Assessment Tool independently to rate sites at pre-intervention, baseline, post-intervention, and 3-, 6-, and 12-months post intervention. AIM 3: Data collection Site program self-assessment: Each site will complete a Self-Assessment Tool to assess team performance at pre-intervention, baseline, post-intervention, and 3, 6, and 12 months post intervention. LOCK Program sustainment interviews: The research team will conduct semi-structured qualitative phone interviews with key SVH staff involved in the intervention: site leadership team members and staff who participated in frontline staff huddles .. Data Analysis AIM 1: Quantitative data analysis The investigators' primary outcome will be measured using the MDS indicator of "any falls since admission/entry or reentry or prior assessment" for long-stay residents. The investigators will predict the likelihood of falling and test for differences in the likelihood of falling given the presence or absence of prior falls. The research team will evaluate change in the likelihood of falling, expressed as an aggregate likelihood for SVHs in terms of rates. Within- and between-cluster data will be used to determine effectiveness of the LOCK program over time. The level of analysis will be the individual SVH resident nested within SVH. The research team will examine differences in the investigators' primary outcome across the pre-intervention through sustainment periods using a multi-level mixed-models approach and either generalized linear mixed models or generalized estimating equations. The aggregate model-adjusted likelihood of falls will be reported at the SVH level as the rate of any fall per 100 residents. The primary analysis will be a partial intention-to-treat analysis using the rate of participants experiencing any fall during an MDS assessment period as the outcome. All residents with pre-intervention assessments will be included. For residents who leave the SVH during the study for whatever reason, multiple imputation will be used to estimate missing data from individuals' existing data, assuming no change from their last available assessment, and incorporating an estimate of random variation based on observed data. The research team will also conduct a complete case analysis (individuals with data from all measurement periods) to test the sensitivity of the primary analysis results. Participating SVHs will be treated as random effects with residents clustered within SVHs. The analysis will include estimates using both unadjusted and adjusted models. Individual-level and SVH-level characteristics will be used in the adjusted models and a Type I error rate of 5% ( \< .05) to identify statistically significant associations. A similar approach will be used to test the intervention's impact on secondary outcomes. The analysis will include an exploratory heterogeneity of treatment effects analysis (difference in difference analysis) examining the relative effect of the intervention on residents with and without delirium and urinary tract infection. AIM 1: Qualitative data analysis Resident interviews: The qualitative analysis will include a content analytic approach facilitated by a qualitative research software (e.g., NVivo). In the case of the resident interviews, a primary analyst will first read through 3 documents, highlighting key words and phrases that pertain to the relevant relational coordination constructs as they relate to the following: (a) experiences related to falling or nearly falling, (b) interactions with staff around falls and fall prevention, (c) falls and the environment, (d) resident-perceived risk factors and fall prevention facilitators, (e) other. AIM 2: Relational Coordination Survey data analysis To assess effectiveness, the investigators will use data from the Relational Coordination Survey collected at 3 time points: pre-intervention, immediately after the intervention ends (post-intervention), and 6 months after intervention ends. The study will use Relational Coordination Survey standard analysis techniques to calculate SVH total scores as well as scores for within and between each SVH workgroup. AIM 2: Qualitative data analysis The qualitative data will be analyzed using a content analytic approach facilitated by a qualitative research software (e.g., NVivo). One qualitative team member will be the primary data analyst for each data source. A primary analyst will first read through 3 documents, highlighting key words and phrases that pertain to the relevant relational coordination and RE-AIM constructs as they relate to the following: (a) content, structure, or implementation of the LOCK program, (b) barriers to or facilitators of implementing the program, (c) system-level strategies for sustainment, (d) suggestions for improvements, (e) mechanisms and mediators of change, and (f) other. The primary analyst will then create a preliminary codebook of these data. A second researcher will use identical methods and create a similar codebook. When all data have been analyzed for a source, the primary analyst will create a content-analytic summary table. AIM 2: Consolidation The research team will assess implementation employing RE-AIM dimensions to summarize Aim 2 findings across data sources into a summary matrix by and across sites. AIM 3: Data Analysis The use of both quantitative data collection and interviews represents a key opportunity to triangulate data. Raters will combine the site level qualitative summary tables combined with quantitative results to examine patterns of variation of implementation and impacts on effectiveness. The research team will develop a protocol for rating and conduct full-team consensus-reaching discussions. This will enable integration of the investigators' mixed-methods data on LOCK program progress over time at the site-level. The investigators will also assess the coherence between the quantitative and qualitative results and look for areas of confirmation, expansion, and discordance. The primary goals will be to: 1) integrate quantitative and qualitative findings to identify facilitators of and barriers to sustainment and 2) identify facilitating and hindering practices using relational coordination as a guide.