Improving safety and preventing falls using an evidence-based, front-line staff huddling practice: protocol for a pragmatic trial to increase quality of care in State Veterans Homes

Abstract

Introduction

Falls in nursing homes are a major cause for decreases in residents’ quality of life and overall health. This study aims to reduce resident falls by implementing the LOCK Falls Programme, an evidence-based quality improvement intervention. The LOCK Falls Programme involves the entire front-line care team in (1) focusing on evidence of positive change, (2) collecting data through systematic observation and (3) facilitating communication and coordination of care through the practice of front-line staff huddles.

Methods and analysis

The study protocol describes a mixed-methods, 4-year hybrid (type 2) effectiveness-implementation study in State Veterans Homes in the USA. The study uses a pragmatic stepped-wedge randomised trial design and employs relational coordination theory and the Reach, Effectiveness, Adoption, Implementation and Maintenance framework to guide implementation and evaluation. A total of eight State Veterans Homes will participate and data will be collected over an 18-month period. Administrative data inclusive of all clinical assessments and Minimum Data Set assessments for Veterans with a State Veterans Home admission or stay during the study period will be collected (8480 residents total). The primary outcome is a resident having any fall. The primary analysis will be a partial intention-to-treat analysis using the rate of participants experiencing any fall. A staff survey (n=1200) and qualitative interviews with residents (n=80) and staff (n=400) will also be conducted. This research seeks to systematically address known barriers to nursing home quality improvement efforts associated with reducing falls.

Ethics and dissemination

This study is approved by the Central Institutional Review Board (#167059-11). All participants will be recruited voluntarily and will sign informed consent as required. Collection, assessment and managing of solicited and spontaneously reported adverse events, including required protocol alterations, will be communicated and approved directly with the Central Institutional Review Board, the data safety monitoring board and the Office of Research and Development. Study results will be disseminated through peer-reviewed publications and conference presentations at the Academy Health Annual Research Meeting, the Gerontological Society of America Annual Scientific Meeting and the American Geriatrics Society Annual Meeting. Key stakeholders will also help disseminate lessons learnt.

Trial registration number

NCT05906095.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• A mixed-method design using implementation research strategies enables researchers to examine in real-time modifications to the training on fall reduction.

• By working to continuously adapt implementation materials to site-specific needs, our study will work in real time to best meet the needs of individual long-term care facilities.

• The implementation curriculum will be vetted by experts in the field through an expert panel review process.

• The study protocol limits enrolled sites to eight, which could potentially limit geographic distribution across the USA.

Introduction

Falls are a significant source of morbidity and mortality in older adults and are the leading cause of fatal and nonfatal injury in adults over the age of 65 years.¹ Adverse consequences of falls include trauma, loss of independence and decreased quality of life.^{1 2} Approximately 10%–19% of falls in older adults result in a major injury, such as a fracture, soft tissue injury or traumatic brain injury.^{3 4} About half of those who reside in long-term care (LTC) settings in the USA will experience a fall each year.^5–8

Falls for older adults are also expensive. In 2020, unintended falls accounted for US$1.54 billion in direct healthcare costs in the USA, more than any other medical event.^{9 10} For the individual, one-third of those with a fall-related injury will require additional help with activities of daily living; over half of these individuals will require assistance for at least 6 months, with 25% requiring admission to a costly skilled nursing facility.^{4 11}

Falls result from the interaction of modifiable and non-modifiable risk factors.^{12 13} These risk factors include age over 65 years, dementia or impaired judgement, multiple chronic conditions, polypharmacy, balance impairment, lower extremity weakness, concern or worry about falling and environmental hazards.4 6 12 14 The more risk factors present, the higher the associated risk of falling.^{12 15} Research has shown that the risk of falling can be reduced by as much as 30% with single or multiple interventions.^16–18 While there is extensive research documenting ways falls negatively impact older adults, less is known about effective ways to decrease falls in specific LTC settings.^19–21 Targeting residents’ specific risk factors in the context of also understanding and working with their personal needs and preferences represents a person-centred approach to falls prevention that holds promise.²²

State Veterans Homes (SVHs) are subacute, hospice and LTC facilities operated by individual states providing care to eligible Veterans, their spouses and parents who lost a child who served in the US Military.²³ SVHs are operated by individual states, with the Department of Veteran’s Affairs providing guidance and assistance with regulatory oversight for these facilities as well as funding for Veterans who have a high level of service-connected disability. There are currently 163 SVHs in the USA; of these, 153 facilities include nursing homes.^{24 25} Fall rates across SVHs are variable. Partly through work with the Department of Veterans Affairs, SVHs are beginning to focus on individualising resident care through interdisciplinary front-line staff huddles. To help SVHs that struggle with high fall rates, this study uses an evidence-based individualised care intervention known as the LOCK Falls Programme²⁶ that aligns with current SVH principles and practices.

Study intervention overview: the LOCK Falls Programme

The LOCK Falls Programme is based on the LOCK programme, a bundle of quality improvement tools that can be applied in various settings. The LOCK programme engages nursing home staff using four principles: (1) ‘Learn from bright spots’ (focus on evidence of positive change); (2) ‘Observe’ (collect data through systematic observation); (3) ‘Collaborate in huddles’ (conduct front-line staff huddles) and (4) ‘Keep it bite-size’ (limit activities to 5–15 min). The LOCK programme has been used in nursing homes to enable front-line staff to improve resident outcomes.^26–28

In the LOCK programme, front-line staff huddles (brief, stand-up meetings to facilitate efficient information exchange) play a fundamental role. Through the huddles, the LOCK programme works to ensure that person-centred interventions are appropriate for the setting, manageable for staff and acceptable for the patient.^29–31 The LOCK programme is distinguished from other quality improvement approaches in its heavy emphasis on infrastructure to support extensive front-line staff engagement in all phases of the quality improvement process.³⁰ The LOCK programme complements and extends existing quality improvement infrastructures and can be easily integrated into front-line staff routines.

When applied to falls, the LOCK Falls Programme addresses gaps in communication among staff by ensuring that all levels of multidisciplinary front-line staff team members (nursing, physician, housekeeping, physical therapy, etc) work together in huddles to identify residents at risk of falling, discuss fall risk factors and develop and test appropriate, individualised interventions. The LOCK Falls Programme increases staff members’ awareness of what may precipitate a fall and enables integration of personalised fall-prevention interventions and systematic falls measurement. In the front-line staff huddles, staff use rapid-cycle quality improvement techniques to pilot and monitor their actions (See table 1).

Table 1.

LOCK elements, concepts and examples

LOCK element	Evidence-based concept and explanation	Example related to falls
Learn from the bright spots	Strengths-based learning: When finding solutions to an issue, look for positive outliers to identify instances of success from which to learn.	Jane Smith has a history of nighttime falls, but for 2 weeks she had no falls. During the investigation, staff found that Jane received a medication with a scheduled dose at midnight. During a huddle, staff who gave the medication say that after taking the medicine they helped her walk to the bathroom.
Observe	Observation: Staff briefly step back from regular routines to conduct unstructured or structured observations. Observations provide data for huddle dialogues.	The huddle helped staff identify that a main contributor to Jane’s falls was her needing to use the bathroom at night. They decide nursing and nutrition will work together to observe and informally track her fluid intake. For two nights, night staff will observe what time she gets out of bed by unobtrusively checking her room every 30 min, helping if they see her up, and documenting her movements.
Collaborate in huddles	Relationship-based teams: Conduct brief, collaborative, strength-based front-line staff huddles to discuss risk factors for an issue, bright spots, results of observations and action planning.	The next week, the huddle facilitator moves the conversation about Jane through the steps of bright spot exploration, hypothesis generation based on observation and action planning, making sure everyone is heard. Staff decide to pilot and track (1) offering Jane a schedule where staff help her to the bathroom at 23:00 hours and (2) reducing her liquids after dinner and (3) reviewing her medications. They agreed to huddle Thursday to discuss what worked and other options to pilot.
Keep it bite sized	Efficiency: Keep all lock components to 5–15 min. Incremental changes, rather than systemic overhauls, are easier to integrate	To make room for their front-line staff huddles, staff shortened existing meetings by 10 min. For timely information exchange, they huddle more frequently for shorter intervals. Huddle facilitators ensure action plans include changes that are both meaningful and small enough to fit into staff routines.

Study conceptual framework: relational coordination

The study’s conceptual framework is relational coordination. Relational coordination (figure 1) is an evidence-based framework emphasising 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. For staff working in nursing home settings, high relational coordination is associated with greater job satisfaction, greater job engagement and less burn-out.^{32 33} Nursing homes with high relational coordination achieve better resident clinical outcomes: higher quality of care, shorter length of stay, lower pain and higher functioning.^34–36 Relational coordination theory will guide data collection and analysis methods.

Figure 1.

Relational coordination. The figure outlines the connection between high-quality relationships and high-quality interactions. The relationships and interactions impact performance quality, performance quantity and employee well-being.

Study overview and aims

This study is a mixed-methods, hybrid (type 2) effectiveness-implementation study.³⁷ The study approach assesses both the effectiveness of the clinical innovation (LOCK Falls Programme) and the implementation process itself. The unit of analysis is the SVH resident clustered within SVHs, for evaluating the LOCK Falls Programme’s effectiveness (aim 1) and is the SVH itself for evaluating the programme’s implementation and sustainment (aims 2 and 3). The study employs an incomplete stepped-wedge cohort design³⁸ with each SVH acting as its own control.

Administrative and primary data collection will occur at eight SVH sites using a two-wave stepped-wedge design³⁹ with four sites participating per wave and the second wave lagged 6 months behind the first. Initiation within each SVH will be lagged slightly within each wave to facilitate implementation by the study team. Residents and staff within participating homes will be followed longitudinally from preintervention through sustainment measurement periods (see figure 2).

Figure 2.

Stepped-wedge design with measurement periods. The figure outlines the intervention (light blue line) and sustainment period (dark blue line) over 24 months. Black diamonds represent in-person data collection and grey diamonds represent remote data collection. SVH, State Veterans Home.

The study team will provide SVH staff with the necessary training and coaching to provide the intervention. The study will include an initial 3-month preintervention period (phase 1) in which each SVH builds their front-line staff huddling practices across all units and all shifts. Following this period, each SVH will begin the 3-month LOCK intervention+facilitation period (phase 2) on all units. A 12-month sustainment period (phase 3) will follow (see figure 3).

Figure 3.

Timing and spacing of intervention implantation and outcomes. Each blue step corresponds to an implementation period (preintervention, intervention and facilitation, and the sustainment). Each black wedge outlines the data collection type and time point.

In addition to the relational coordination conceptual framework, we will use the Replicating Effective Programmes framework⁴⁰ to guide implementation and track primary and secondary outcome measures. We will also use the Reach, Effectiveness, Adoption, Implementation and Maintenance (RE-AIM)⁴¹ evaluation framework to assess RE-AIM of the LOCK Falls Programme and implementation processes.

Aim 1 will investigate the effectiveness of the LOCK Falls Programme on improving our primary outcome: the proportion of residents within each SVH experiencing a fall during each reporting period. For the secondary outcome, we will investigate resident clinical outcomes, including mobility, medication changes, restraint use, alarm use and hospitalisations. Primary and secondary outcomes will be collected for all eligible residents at each participating SVH using minimum data set (MDS) data and are described in more detail below.⁴² We will perform semistructured interviews with five residents recruited from each SVH (see table 2). Work-process outcomes will be assessed through staff surveys (job satisfaction, work engagement and burn-out). Researchers will also conduct an environmental assessment during site visits.

Table 2.

Recruitment and time points

Aim	Group	Total no	Time points
Aim 1 Investigate effectiveness of the LOCK programme	Staff	30 staff members per unit (average of 5 units per SVH) 150 respondents max per SVH 1200 across all sites per time point	Preintervention Baseline Postintervention 12-month follow-up
	Residents	5 residents per SVH 40 residents across all sites	Baseline Postintervention
Aim 2 Evaluate implementation	Staff	30 staff members per unit (average of 5 units per SVH) 50 respondents max per SVH 1200 across all sites per time point	Preintervention 6-month follow-up
	Staff	15 participants per SVH 120 across all sites per time point	Baseline Postintervention
Aim 3 Assess programme sustainment	Staff	10 participants per sites 80 respondents max across all sites at each time point	Preintervention Baseline Postintervention 3-month follow-up 6-month follow-up 12-month follow-up

SVH, State Veterans Home.

Aim 2 will evaluate factors related to variation in the LOCK Falls Programme’s implementation. To support successful programme implementation, we will use the Replicating Effective Programmes framework and multimodal implementation facilitation strategies.⁴³ We will use mixed methods to evaluate the programme’s reach, adoption and implementation. Aim 3 assesses the extent of programme sustainment (maintenance) and sustainment variability at 3, 6 and 12 months postintervention and sustainment variability across sites.

Intervention tailoring

Prior to initiating the study, our team will recruit a LOCK leadership team at each participating SVH. This team will comprise the medical director (or designee), the director of nursing, the MDS data coordinator, and the quality assurance-performance improvement nurse and/or falls coordinator. Along with membership from the National Association of State Veterans Homes (NASVH), these stakeholders will work with us for 3 months prior to study initiation to tailor existing intervention materials to the SVH context using user-centred experience design. They will, for example, provide input on engaging nursing leadership, huddle logistics, technological issues and motivating front-line staff. We will also involve SVH stakeholders to make further adjustments prior to implementation at each site.

Intervention strategies

The LOCK Falls Programme uses several strategies to improve the likelihood of intervention success, including (1) supporting participant buy-in, (2) training a leadership team, (3) focusing on front-line staff and (4) just-in-time teaching.^{44 45} These strategies are conceptualised as follows.

1. Supporting participant buy-in: Participant buy-in is an integral part of all implementation studies. Participant buy-in ensures that all key players are not only invested in the initial start-up and implementation but also the sustainment of the project.

2. Training a leadership team: Having a stable and engaged leadership team is an essential component to ensure sustainment. Providing training and resources for the leadership team adds an additional layer of buy-in and accountability.

3. Focus on front-line staff: Front-line staff in LTC facilities have the most interaction with the residents and can provide the important information regarding resident preferences, routines, history and present condition. The LOCK Falls Programme leverages front-line staff knowledge by ensuring the front-line expertise is included and acted on in the staff huddles.

4. Just-in-time teaching: Just-in-time teaching provides key knowledge and is focused on active learning.

Methods and analysis

Setting

The study will identify an initial pool of at least 14 SVH from the existing 153 SVHs nursing homes located in the USA. We will recruit potential sites through a direct email solicitation with support from the NASVH. Homes that are interested will complete an online screening tool and have an initial phone conversation with study staff to identify homes that meet the following criteria.

1. Are licensed by the Centers for Medicare and Medicaid Services.

2. Have electronic health records

3. Have at least 60 Veteran residents.

4. Struggle with falls

5. Have some successful experience with QI.

6. Have a stable and engaged leadership team.

7. Are willing to have researchers come on site for data collection/training.

Once an eligible SVH has expressed interest in participation, the study team will communicate with the home’s administrator to explain the study further and answer questions. Items discussed will include an explanation of the benefits of study participation; the procedures for recruitment of staff and residents and any site-specific considerations, including union notification; logistics for research team members coming on site; and a request to allow staff to participate in the study survey and interviews during their work hours, with the assurance that it will not interfere with resident care. Homes that reaffirm interest in the study after this will be added to the initial pool.

From the initial pool, a total of eight sites will be selected based on having a four-quarter-average rate of residents with any fall of over 44% (the current 10th percentile range) based on SVH resident data from the MDS. We will use sequential balancing³⁹ to select the final eight sites using (1) administrative data (fall rates, bed size/volume and case mix based on MDS resource utilisation group) and (2) site-level demographic information (administrative leadership stability/tenure, nurse manager stability/tenure and nurse staffing ratios).

Study population

Administrative data inclusive of all clinical assessments and MDS assessments for Veterans with an SVH admission or stay during the study period will be collected (8480 residents total).

Resident sample

To recruit SVH residents for our structured interviews, we will use MDS data to identify residents at SVHs that have a history of falling in the past 6 months and length of stay greater than 100 days (long stay). During midpoint and postintervention site visits, we will recruit up to 5 SVH residents per site (80 residents total) to complete interviews. We will exclude Veterans who are less than 18 years of age, as well as those who are in a comatose state or admitted for hospice or respite care.

Leadership and staff sample

All SVH leadership and staff members over the age of 18 will be eligible to participate in surveys and interviews. We will exclude staff who identify as volunteers or contractors. SVH leadership will provide the names and emails of all front-line staff, interdisciplinary team members and leaders to the study team. We estimate 30 staff members per unit and an average of five units per SVH. We, thus, estimate a maximum of 150 respondents per SVH and 1200 across all sites per time point for staff surveys. During the 3-month and 6-month time points, we will recruit up to 400 staff to complete interviews.

The LOCK Falls Programme

Leadership team development

We will ask the SVHs to identify a LOCK Falls Programme leadership team and a LOCK Falls Programme champion. The study team will train the leadership team and champion on the LOCK Programme front-line staff huddling practice during the 3-month preintervention period. The leadership team will support the champion in training front-line staff and implementing a front-line staff huddling practice on the units, focused on huddling about residents at risk of clinical decline. After baseline data collection, the research team will help the site leadership team and champion focus the existing huddles on using a whole-team approach to prevent falls. To obtain broad buy-in, we will initially hold an 8-hour training for home leadership teams at each home at the beginning of the preintervention period (see figure 3, first black arrow). It also will include 1 hour where the teams run a front-line staff huddle and come back to debrief.

Implementation plan development

Consistent with the Replicating Effective Programmes framework, our training incorporates time for participants to create a detailed implementation plan. We will also support the home leadership teams in piloting front-line staff huddles on one unit and then spreading it to other units using the following external facilitation activities: (1) weekly 1-hour research staff office hours for all SVHs to call in with questions, (2) biweekly 30 min coaching calls led by a research team member for all leadership team members from all sites in the current wave, (3) sharing videos on huddle facilitation tips and (4) as needed additional emails and phone calls.

The LOCK Falls Programme training

At the baseline visit (see figure 3, second black arrow), we will begin the intervention. We will hold a 4-hour training for each site’s leadership team on risk factors for falling and how to guide staff to use their front-line staff huddles to continually identify residents at highest risk of falling, explore residents’ personal histories and develop individualised action plans based on risk factors. Staff will continually target a fluid ‘watch’ list of residents at highest risk. Consistent with the Replicating Effective Programmes framework, we 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 assessments⁴⁶ and postfall huddling to identify fall prevention intervention opportunities.

The fall prevention training includes evidence-based information on why falls are important, their aetiology, evidence-based fall prevention intervention suggestions and how to use huddles to integrate such interventions into individualised treatment plans. Similar to the preintervention phase, we will support the SVH leadership teams in piloting the LOCK Falls Programme on one unit and then spreading it to other units through external facilitation activities of (1) weekly office hours, (2) biweekly coaching calls, (3) sharing videos and (4) as needed calls and emails. Given communication and education that takes places within huddles and prior literature on the effects of huddles on fall prevention, we expect spillover effects to other residents, even those not specifically targeted through the huddles.

Multiple implementation facilitation strategies are necessary to successfully support change in nursing homes.⁴⁷ We will, therefore, use many of Waltz et al’s strategies.⁴³ Specifically, we will use the following: (1) blended facilitation (using both research team members and site-based programme leadership to support implementation); (2) tailoring strategies (tailoring VA materials to SVH context); (3) identifying and preparing champions (the programme leadership team); (4) on-site and remote training; (5) ongoing consultation (office hours and phone coaching); (6) creating learning collaboratives (multiple study waves in which SVHs help each other); (7) creating new clinical teams (front-line staff huddles); (8) intervening with SVH residents (through the action plans resulting from huddles) and (9) helping stage implementation scale up (teach sites how to pilot and determine when to scale up).

Quantitative measures

Primary outcome

Our primary outcome will be measured using the MDS indicator of ‘any falls since admission/entry or reentry or prior assessment’ for long-stay residents, (which has an average rate of approximately 55% of residents per quarter across all SVHs). This study will collect data from a variety of sources, at various time points in alignment with each specific aims (see figure 4). Historically, falls have been under-reported on the MDS among nursing home residents.^{17 48} Capturing residents with any fall(s) is less sensitive to under-reporting since it is not affected by the absolute number of falls.

Figure 4.

Aims, study instruments and study phase. The figure outlines the three primary study aims, the study measures associated with each aim as well as the source of the study measure and the data target. Data collection time points are illustrated by black arrows and black check marks. MDS, Minimum Data Set; SVH, State Veterans Home.

Secondary outcomes

Mobility

The LOCK Falls Programme can address mobility through staff incorporating mobility exercises into daily resident interactions.^6–8 The MDS includes multiple items related to mobility. We will use the following: ‘Balance during transitions and walking’ measures impaired balance and unsteadiness during transitions and walking (assessed from 0=steady at all times to 2=not steady, only able to stabilise with staff assistance). ‘Mobility devices’ lists the types of mobility aids the resident uses. ‘Functional rehabilitation potential’ assesses both resident and staff member perspectives on whether the resident is capable of improvement.

Medication changes

We will assess changes in medications associated with a risk of falling: cardiovascular drugs (loop diuretics, antiarrhythmics, antihypertensives),^{49 50} psychotropic medications (antipsychotics, antidepressants, benzodiazepines and non-benzodiazepine hypnotics),⁵¹ what is commonly known as ‘Z-drugs’ (zolpidem, zopiclone, eszopiclone and zaleplon),⁵² opioids⁵³ and antiepileptics.⁵³ We will also look for reductions in polypharmacy (four or more medications).⁵³

Restraint and alarm use

Restraint use in nursing homes is highly regulated and rarely allowable. Restraints and bed alarms also have no evidence base in fall prevention.⁵⁴ Thus, as a balance measure, we will assess restraint and alarm use (MDS has indicators for each), as both are still inappropriately used to prevent falls with the hope that this intervention will reduce use as has been previously shown.⁵⁵

Hospitalisation

We will use the MDS discharge assessment to identify if a resident is discharged to an acute care hospital. One benefit MDS discharge records is that they identify discharges to a community or VA hospital. In previous research comparing the MDS discharge records with Medicare claims, the MDS discharge records were shown to have a high degree of accuracy.⁵⁶

Additional quantitative measures

Environmental assessment

At the baseline and postintervention data collection visits, a research assistant will assess the environment using a checklist based on CDC STEADI checklist for safety and inpatient fall prevention resources.⁵⁷ Assessment will include both hazards (eg, chords, furniture placement, flooring issues) and accessibility (eg, bathroom safety equipment availability, bed height, meal trays delivered within reach) appropriate for the Veteran.

Delirium and urinary tract infections

For an exploratory heterogeneity of treatment effects analysis, MDS data will be used. The analysis will provide preliminary information on how huddling can help manage increased risks associated with both.

Resident-level covariates

From the MDS, we will include a measure for cognitive function, the Cognitive Function Scale, created and validated by members of the study team to assess residents’ cognitive performance.⁵⁸ The Cognitive Function Scale is a single-item, 4-point scale that integrates the MDS’s Brief Interview for Mental Status and the Cognitive Performance Scale; we have demonstrated that it has high levels of construct validity.⁵⁸

Staff outcomes

Job satisfaction

We will use one global item to measure job satisfaction: ‘On the whole, how satisfied are you with your present primary job?’ (1=very dissatisfied to 4=very satisfied). This item has been used in numerous prior nursing home and hospital studies.^{59 60} Higher scores on this global measure are negatively associated with intent to leave one’s current position.⁶¹

Work engagement

Work engagement is defined as having motivation manifested as vigour, dedication and absorption.⁶² Work engagement also plays an important role in preventing burnout and turnover among employees.⁶³ We will use the validated, nine-item Utrecht Work Engagement Scale.⁶⁴ This scale will be used as a composite to measure the separate dimensions of absorption, vigour and dedication of staff.⁶²

Burn-out

Burn-out is prevalent in healthcare and potentially related to worse patient safety.⁶⁵ We will use the validated Copenhagen Burnout Inventory.⁶⁶ It has been used extensively to measure burnout in healthcare^{67 68} and, unlike the Maslach Burnout Inventory, is in the public domain. It has high Cronbach’s levels for its three domains (personal burnout α=0.892, work-related burnout α=0.896 and client-related burnout α=0.897). The scale will be used as a composite and to measure the separate domains.

Other quantitative covariates

We will email sites to collect administrative data about SVH specialisation, SVH size/volume and casemix.⁶⁹ We will also gather data on SVH on QI participation and will flag intervals in which care may have been severely impacted by procedures in reaction to COVID-19 or other crises.

Implementation facilitation time tracking log

This tool⁷⁰ documents the time, activities and personnel involved in implementation facilitation efforts. External facilitators will regularly record the facilitation event type, mode of communication, with whom they interacted and specific activities, thereby also gathering information on tailoring and uptake of the LOCK Falls Programme using an access database. It also includes a free text notes section in which we specifically ask research team members to reflect on what worked and did not work regarding engagement, facilitation, implementation, adaptation and partnering. Conducting time sampling (eg, 1 week/month) will also enable for estimates of necessary resources prior to operational dissemination.

Relational coordination survey

We will document implementation of relational coordination using the relational coordination survey.⁷¹ The survey asks staff participants to rate their experiences with their own and other workgroups on the seven key dimensions of relational coordination (frequency, timeliness, accuracy, and problem-solving nature of communication; shared goals; shared knowledge; and mutual coworker respect). It uses a 5-point Likert scale. We will work with the participating SVH to determine the relevant workgroups based on the job types.

Clinical Sustainability Assessment Tool

To assess sustainment (maintenance), we will use the Clinical Sustainability Assessment Tool,⁷² a variant of the Programme Sustainability Assessment Tool.⁷³ It has 35 Likert scale items evaluating 7 domains: engaged staff and leadership, engaged stakeholders, organisational readiness, workflow integration, implementation and training, monitoring and evaluation, and outcomes and effectiveness. Site programme leadership and research team members will independently rate sites at preintervention, baseline, postintervention, and 3, 6 and 12 months postintervention.

Site programme self-assessment

We will ask each site to complete the Self-Assessment Tool, which is used nationally for VA nursing homes. This tool allows sites to assess their current implementation level of the intervention (fully, partially and not at All) and identify top priorities to focus on for improvement. Site programme leadership will rate their site at preintervention, baseline, postintervention, and 3, 6 and 12 months postintervention.

Qualitative measures

Resident interviews

We will conduct semistructured interviews with up to five residents per SVH during the baseline and postintervention data collection visits to gather their impressions of fall prevention and fall risk-related activities and interventions in the SVH. The interview guide will focus on 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 and (e) other. To enable informed consent, we will target cognitively intact long-stay residents who have fallen at least once.

Leadership team and staff interviews

We developed a semistructured interview guide, with undergirding from Relational Coordination and the RE-AIM framework. The guide assesses staff impressions of the content, structure and implementation of the LOCK programme; 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. We also ask them to reflect on what did and did not work regarding engagement, facilitation, implementation and adaptation. For aim 3, we will modify the postintervention interview guide as necessary for the 3-month, 6-month and 12-month postintervention sustainment interviews by, for example, adding additional questions based on the implementation analysis findings.

Field notes

We will collect field notes during site visits using a structured template that captures general impressions from informal conversations and overall impressions of the physical layout and atmosphere, staff-staff interactions, staff-resident interactions, processes, etc.

Data collection

The study received institutional review board approval for the study. Prior to initiating study activities, informed consent (see online supplemental material) will be obtained as follows. The study received approval for a modified, verbal informed consent and a waiver of documentation of informed consent for the semistructured virtual staff and family member interviews. Research team members will conduct consent discussions and record participants verbal consent. The study received a waiver of documentation of informed consent for the anonymous staff survey. The study received approval to gather written informed consent from all enrolled resident participants.

Staff surveys

SVH staff will be recruited for research surveys via email and virtual presentations by researchers at staff meetings, as has been done successfully in our other studies (CRE 11-349; IIR 14-008). Surveys will be administered electronically. Staff will be emailed a link with three reminders and opt-out information in each email. No identifiers will be obtained; anonymity will be maintained to facilitate high response rates.

Resident interviews

Prior to site visits, the study team will use VA’s Corporate Data Warehouse data, SVHs’ electronic health record and MDS data to identify cognitively intact residents who have a history of falling. We will take extra precautions to allow for maximum resident participation with minimal risk, including researcher administration of a structured cognitive assessment tool to assure informed consent competence. Interviews with consented residents will be audiorecorded and will average 20 min each.

Leadership team and staff interviews

We will recruit the LOCK Falls Programme leadership team members as well as a sample of front-line staff who participated in the huddles, with an approximate sample size of 15 per SVH, a sample size we base on our prior work and the literature.^{74 75} To the extent possible, we will conduct semistructured qualitative interviews in person during the baseline and postintervention site visits, prioritising nursing assistants and individuals in other job types that do not have control over their own schedule/access to private spaces with phones. Remaining interviews will be conducted via phone or an IRB-approved video conferencing mechanism, with participants recruited through email, as described for the staff surveys. Interviews will average 30 min each, to minimise impact on clinical staff. Interviews will be transcribed verbatim. For aim 3, we will conduct semistructured qualitative phone interviews (average of 20 min) with key SVH staff involved in the intervention: site leadership team members and staff who participated in front-line staff huddles (n=10 per site).

On-site data collection

Research team members will collect environmental assessment and SVHs health record data at preintervention and postintervention site visits.

Patient and public involvement

Veteran input has been a central tenant of this work from its inception. We conducted resident interviews as part of the pilot study that developed the LOCK programme. For this proposal, we recruited two older Veterans with experience with VA services to serve as advisors. They will participate in twice-yearly meetings with the study team to provide suggestions on intervention plans, progress and outcomes. In addition, many SVH directors are retired military, so we will benefit from their input as we proceed. Finally, assessing State Veteran Home Veteran resident perspectives on the intervention is a critical component of the study design and will be done using semistructured interviews preintervention and postintervention.

Method to reduce imbalance

Our incomplete, stepped-wedge cohort design helps mitigate the risk of imbalance over time by having each site serve as its own control. But stepped-wedge designs also risk imbalance on key site characteristics. We will thus use sequential balancing³⁹ of site characteristics over the steps to strengthen internal validity and minimise potential confounders. Six probable factors for balancing are listed here. Each SVH will complete a short readiness survey comprising three close-ended items on (1) administrative leadership stability/tenure, (2) nurse manager stability/tenure and (3) nurse staffing ratios. Three other factors will be extracted from administrative data⁷⁶ specifically, (4) fall rates, (5) size/volume (number of beds, number of admissions per bed) and (6) casemix, as indicated by average MDS Resource Utilisation Groups score (rated 1–44, with higher scores indicating greater severity). Following methods described by Lew et al,³⁹ we will calculate imbalance scores and identify the most balanced assignment schemes to designate start times.

Should any SVHs drop out prior to their baseline data collection visit, we will replace them with an alternate SVH. To ensure adequate power, SVH sites with significantly low enrolment (less than 70% of the number of target staff participants in huddles at baseline) will be replaced with an alternate site. Dropped sites will receive two facilitation calls at 3 months and 6 months postbaseline; no additional data collection or facilitation will take place.

Data management

The co-principal investigators will be responsible for ensuring participants’ safety throughout the course of the study. All personally identifiable information from the study will be housed behind the firewall in password-protected files. Deidentified data sets will be created that do not include any identifying information.

Audiorecording transcripts will not include participant names or other identifying information. Interviews and other qualitative data will be coded. Identifying information linked to study numbers will be stored in secure, limited access electronic files. Data collected for both quantitative and qualitative analysis will be subjected to rigorous quality reviews by study staff.

Analyses

Investigate effectiveness of the LOCK Falls Programme (aim 1)

We will use quantitative methods to evaluate within-cluster and between-cluster data to determine effectiveness of the LOCK Falls Programme over time. As described in figure 3, data sources include primary data collection (staff surveys, staff surveys and researcher-collected data) and administrative data (MDS, health record).

We will examine differences in our primary outcome (any fall) across the preintervention through sustainment periods using a multi-level mixed-models approach and either generalised linear mixed models or generalised estimating equations. Both approaches enable the estimation of models with random and fixed effects and the use of outcome measures having the range of distribution types likely to be observed in our primary and secondary outcomes.⁷⁷ The level of analysis will be the individual SVH resident nested within SVH. By beginning with a preintervention period and implementing the interventions in ‘steps’ at set times, each SVH serves as its own control, with preintervention observations used to estimate temporal trends so these can be distinguished from intervention effects. Standard stepped-wedge designs treat time in discrete blocks (the steps), but this trial will provide more robust estimates of secular trends by linking observations to each participant’s baseline assessments, providing finer temporal resolution.

We will use data from each resident’s last available MDS assessment preintervention and the first available MDS postintervention. Additional MDS data may be included, as available, to extend the preimplementation and postimplementation assessment windows, increase our ability to distinguish individual variability from intervention-related change, and provide greater statistical power for these analyses. The form of the models tested will depend on the completeness of MDS data available for each participant over the course of the study. Data will enable estimation of (a) within-individual variation over time, (b) between-individual variation incorporating individual characteristics (eg, demographics) and (c) site-level variation incorporating SVH characteristics (eg, size, casemix). The postintervention period will enable us to test for longer-term intervention effects (sustainment). Plots of outcomes over time (overall and stratified by SVH) will be used to identify temporal fluctuations that may indicate unanticipated changes in processes or personnel.

Our 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. We will include all residents with preintervention assessments. For residents who leave the SVH during the study for whatever reason, we will use multiple imputation 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. We will apply our model approach described above to the multiply imputed datasets. Model results will be combined using Rubin’s rule.⁷⁸ We will also conduct a complete-case analysis (individuals with data from all measurement periods) to test the sensitivity of our primary analysis results.

Participating SVHs will be treated as random effects with residents clustered within SVHs. By clustering within SVH, we can control for site-level effects and plausibly treat the individual residents as independent.⁷⁹ We will estimate both unadjusted and adjusted models. We will include individual-level and SVH-level characteristics in our adjusted models, as described above. We will use a type I error rate of 5% (α<0.05) to identify statistically significant associations. We will use a similar approach to test the intervention’s impact on our secondary outcomes. We will also conduct 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.

We estimated our necessary sample size based on our primary outcome of any fall, using Woertman et al’s⁷⁹ methods for estimating the design effect associated with our stepped-wedge design. While SVHs have an average size of 158 residents, we estimated power using the more conservative average of 60 Veterans per SVH, meaning our actual power should be greater. We estimated approximately 20% attrition across our study period. Assuming the average intracluster correlation of 0.10 within SVHs, we anticipate greater than 90% power at α=0.01 for detecting a decline of 10 absolute percentage points in the likelihood that a resident will be reported as having experienced a fall in the MDS reporting period (eg, from 50% to 40%). This means we will be able to detect a magnitude of improvement smaller than what has been seen in other nursing home fall interventions.⁸⁰ In addition, the range of our primary outcome variable is narrow enough (10th to 90th percentile range of approximately 44–66) that we can argue for our ability to detect an intervention effect even among SVHs at the lower end of the range.

Qualitative data analysis (aims 1 and 2)

For the qualitative data, we will use a content analytic approach⁸¹ facilitated by a qualitative research software in a shared VA research folder accessible across research team members. One qualitative team member will be the primary data analyst for each data source.

In the case of the aim 1 resident interviews, a primary analyst will first read through three documents, highlighting keywords 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 and (e) other.

In the case of the aim 2 analyses, a primary analyst will first read through three documents, highlighting keywords 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 Falls Programme, (b) barriers to or facilitators of implementing the programme, (c) system-level strategies for sustainment, (d) suggestions for improvements, (e) mechanisms and mediators of change⁸² and (f) other.

In both cases, the primary analyst will then create a preliminary codebook of these data. A second researcher will use identical methods and create a similar codebook. These codebooks will be compared. If there are discrepancies, they will be resolved by discussion with the entire team. The analysts will then code the rest of the documents independently. When seven documents are complete, the pair will review the seventh transcript in the same manner as before. If substantial disagreement (˃25%) exists, the analyst pair will conduct the remaining coding together; if not, each will progress alone. When all data have been analysed for a source, the primary analyst will create a content-analytical summary table.

Relational coordination survey data analysis (aim 1)

To assess effectiveness, we will use data from the relational coordination survey collected at three time points: preintervention, immediately after the intervention ends (postintervention) and 6 months after intervention ends. On the survey, respondents indicate how frequently other individuals from each workgroup coordinated care with them. We will use relational coordination survey standard analysis techniques to calculate SVH total scores as well as scores for within and between each SVH workgroup (eg, within nursing as well as between nursing and physical therapy, nursing and dietary). We will assess implementation using methods described in the literature, employing RE-AIM dimensions to summarise aim 2 findings across data sources into a summary matrix by and across sites.⁸³

Assess program sustainment (aim 3)

We will combine the qualitative summary tables with quantitative results to determine patterns in variation of implementation and impacts on effectiveness.^{84 85} The template enables raters to combine qualitative and quantitative data, using the examples and definitions within the template, which we will refine based on the study experiences prior to use. We will develop a protocol for rating and conduct full-team consensus-reaching discussions. This will enable integration of our mixed-methods data on LOCK Falls Programme progress over time at the site level. We will also assess the coherence between the quantitative and qualitative results and look for areas of confirmation, expansion and discordance.⁸⁶ The primary goal will be to integrate quantitative and qualitative findings to identify facilitators of and barriers to sustainment. We will hold a series of team meetings, using relational coordination as a guide for clustering facilitating and hindering practices.

Ethics and dissemination

Ethical considerations

This intervention works to reduce falls by implementing and improving huddling practices among staff in SVHs. Resident engagement is minimal and targeted on gathering their input via qualitative interviews, thus the risks to residents are minimal. Different measures occur at varying time points throughout the study and collect data from staff, residents, as well as researcher observations. All participants will be recruited voluntarily and will sign informed consent as required. Resident and staff recruitment will start on 1 February 2024.

This study was granted full board approval by the Department of Veterans Affairs Central Institutional Review Board (CIRB; IRB reference number: 1670549-6). Collection, assessment and managing of solicited and spontaneously reported adverse events, including required protocol alterations, will be communicated and approved directly with the CIRB, the data safety monitoring board and the Office of Research and Development.

Dissemination

The publication plan was developed in collaboration with study team members and includes study staff, key stakeholders and expert reviewers. We will work in tandem with the study’s key stakeholders to disseminate results. The publication and dissemination plan will evolve throughout the study but includes the following activities:

• Results of the study will be disseminated through peer-reviewed publications.

• Conference presentation at the Academy Health Annual Research Meeting.

• Conference presentation at the Gerontological Society of America Annual Scientific Meeting.

• Conference presentation at the American Geriatrics Society annual meeting.

Adherence to reporting guidelines

This work adheres to the quality reporting guidelines applicable to protocols of randomised controlled trials. Achecklist was submitted with this manuscript documenting compliance with the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) guidelines.

Ethics statements

Patient consent for publication

Not applicable.