Changing Talk Online (CHATO) Study

Background on Condition, Disease, or Other Primary Study Focus: The population afflicted with
Alzheimer's disease and other dementias will expand from 5 to 16 million by 2050, increasing
dementia care costs from $259 billion to $1.1 trillion. Of today's 1.4 million nursing home
(NH) residents, 61% have moderate to severe dementia, and up to 90% of them exhibit
behavioral and psychological symptoms of dementia (BPSD) such as physical and verbal
aggression, agitation, and wandering. These behaviors are associated with depression as well
as reduced quality of life and lower survival rates in persons with dementia (PWD). BPSD also
stress family caregivers and precipitate NH placement. In the NH, BPSD increase time to
provide care, and NH staff, primarily Certified Nursing Assistants (CNAs) who provide most
direct care, report that BPSD represent the most stressful aspect of their job. Considering
additional costs for CNA burnout and turnover, It is estimated that BPSD increase costs of
dementia care by 25 to 35%. With national NH rates for a semi-private room of $82,200 per
year, reducing BPSD may save up to $20,000 per resident annually.

As cognitive and communication abilities decline due to dementia, NH residents become unable
to convey care preferences and needs and staff communication becomes infantilizing,
impersonal, and task-oriented resulting in BPSD. As verified in the investigator's past
research using behavioral coding and sequential analyses of video-recorded care, staff
elderspeak (communication that sounds like baby talk) is linked to resident resistiveness to
care (RTC), a subset of BPSD that disrupt nursing care. NH residents were more than twice as
likely to be resistive to care when staff used elderspeak compared to normal communication.
Thus, improving communication has great potential as a nonpharmacological intervention to
reduce BPSD in NH care.

Study Rationale: The Communication Predicament of Aging theory establishes the link between
elderspeak and BPSD. Elderspeak derives from stereotypical views of older adults as less
competent than younger persons. When younger people talk with older adults, they modify their
speech by simplifying, clarifying, and altering the underlying affective quality of messages.
The resulting implicit message of incompetence begins a negative feedback loop for older
persons, who react with depression, withdrawal, and dependency. Elderspeak is especially
threatening to self-concept and personhood, critical to the wellbeing of PWD who are likely
to respond with BPSD. The Need-driven Dementia-compromised Behavior model recognizes BPSD as
the expression of unmet needs of PWD. Communication, that staff can modify to prevent BPSD,
is an essential constant part of the environment connecting PWD to others and affirming their
self-concept.

Psychotropic medications are often used inappropriately to control BPSD in NH residents with
dementia. Alarmingly high rates persist, despite negative outcomes, an FDA black box warning
of increased mortality for older adults with dementia, and a recent Centers for Medicare and
Medicaid Services (CMS) mandate to reduce off-label prescribing of antipsychotics. CMS and
the National Partnership to Improve Dementia Care target reductions in psychotropic drug use
as top priority. Despite reductions in antipsychotic rates (one type of psychotropic
medication) ranging from 3 to 12% from 2011-2016, up to 20% of NH residents received
inappropriate antipsychotic medication in 2017. Research demonstrates that educating direct
care providers in behavioral interventions to control BPSD also reduces psychotropic drug use
(antipsychotics, hypnotics, antidepressants, antianxiety, sedative, anticonvulsant and mood
stabilizers), although evidence is limited by lack of rigorous clinical trials that also
evaluate approaches that influence intervention effects.

The recently completed R01 clinical trial that tested Changing Talk (CHAT) communication
training (NR011455) provides preliminary data for this CHATO pilot. CHAT decreased staff
elderspeak that reduced resident RTC and increased staff awareness of elderspeak's negative
effects. CHAT NHs also saw a significant reduction in psychotropic medication use after the
training versus averaged state rates. Despite the success of CHAT in reducing elderspeak and
RTC, investigators found challenges to educating NH staff that limited participation in CHAT
including turnover, absenteeism, heavy workloads, and personal conflicts. Each CHAT session
was held multiple days and times. Still, as few as 44% of staff in one NH completed at least
two of the three sessions, although this rate is higher than that noted for other NH staff
training programs. Although successful in reducing RTC, the classroom format limits staff
access and participation and feasibility for widespread dissemination. Creative, efficient
approaches are needed to overcome NH staff education barriers. An online web conference
training with multiple NHs as an alternative format was first evaluated for increasing access
and dissemination. However, engagement of individual staff was limited with this approach. To
facilitate dissemination, online CHAT modules (CHATO) were developed to provide the same CHAT
content with asynchronous and independent access for busy NH staff.

The PI worked with an instructional designer, item writer, and media team to transition CHAT
content, including 20 video clips of NH staff-resident interactions, to the online CHATO
modules. Scripts from the original CHAT were narrated to maintain content, integrating adult
learning theories and principles for online learning, and eliminating a need for advanced
literacy skills. Interactive scenario and game-based activities engage staff. For example,
participants watch a video clip, select problem communication in the transcript, type their
improved communication, and compare it to suggested corrections. Moderated online discussions
are included in the modules that are supported on Training-Source.org, a free and publicly
available learning portal. IT functionality and content equivalency of the newly developed
CHATO online modules was demonstrated by a convenient university-affiliated sample of nurses,
CNAs, and students with NH experience. Although the group testing CHATO may not represent all
NH staff, findings confirm that CHATO is feasible and comparable in content and effects.

While CHAT effectively reduced RTC, the in-person classroom format required an onsite
interventionist, which limits accessibility and feasibility for dissemination. A pilot test
of acceptability and preliminary efficacy of online CHATO modules is the next logical step.
The pilot will prepare for a pragmatic clinical trial that will test the effects of improved
staff communication (from CHATO) on resident BPSD and effects on psychotropic medication use.
This research will address the gap in rigorous trials testing nonpharmacological
interventions to decrease BPSD that also identify strategies to improve intervention
dissemination. The goal is to increase access to CHATO training, as a tool to reduce BPSD and
inappropriate psychotropic medication use to improve dementia care.

STUDY DESIGN

The R61 will prepare for the R01 pragmatic trial by establishing feasibility of online
modules and preliminary efficacy of CHATO with NH staff. The research design is a randomized
clinical trial. One NH will provide initial feasibility testing. Any modifications to the
modules will be made. Then 6 nursing homes (estimated N=150 staff) will be randomly assigned
to intervention or wait-list control groups. The primary outcome will be knowledge gain for
staff completing CHATO training. Additional outcomes include resident quality measures
related to behavioral and psychological symptoms of dementia (BPSD) on both resident and
facility levels and facility level data related to inappropriate use of psychotropic
medications to control BPSD.

Staff Knowledge Gain. At time 1, nursing home staff in both groups will complete baseline
pre-tests of knowledge and rating of communication. The immediate intervention group will
then complete the online CHATO modules within a one-month period. This includes assessments
of post-test knowledge, communication rating, program evaluation, and diffusion of innovation
surveys at Time 2. Also, at Time 2, the wait-list control group will repeat the pre-tests of
knowledge and communication rating and will then complete the CHATO online module training
over a one-month period, followed by Time 3 collection of post-test knowledge, communication
rating, program evaluation, and diffusion of innovation surveys. Knowledge gain and
communication rating data will be compared between the intervention and wait-list control
groups and within nursing homes before and after the CHATO training. The immediate and
wait-list groups will be compared at baseline to identify important covariates for analyses
of changes in outcomes. Changes in knowledge from Time 1 (baseline) to Time 2 will be
compared between immediate and wait-list groups using model estimates obtained with a linear
mixed model (LMM) approach to account for repeated measures and clustering within nursing
homes. Next, pre- to post-training changes in knowledge will be combined for immediate and
wait-list groups and also tested using LMM approach.

Resident Quality Measure Outcomes. In addition, nursing homes in both groups will provide
monthly summary reports for behavioral symptom occurrence for the facility (in aggregate as
well as for individual residents [deidentified]). A nursing home aggregate antipsychotic
medication use report will also be collected and analyzed. The BPSD and medication reports
will be provided for the one-month period before baseline data collection and for each month
thereafter until the post-training assessment of the wait-listed nursing homes is completed
(a total of seven months for both groups). Resident outcomes data will be compared between
the intervention and wait-list control groups and within nursing homes before and after the
CHATO training. Initially, seven months of data will be plotted to examine changes in nursing
homes and resident outcomes. Pre- to post-training differences will be compared between
immediate and wait-listed groups with a LMM approach.

Implementation Strategies and Process Evaluation. An Implementation Toolkit for NHs and a
CHATO Training Manual have been created to provide support and implementation suggestions to
pilot nursing homes. Several consultants provided feedback and additional resources to be
included in these materials. Each NH will be given a three-month period to complete the CHATO
training. One month for orientation, team development, and planning, one month to complete
the three-week training, and one month for staff recognition and follow up. The CHATO
Research Team will meet with the NH leadership at the beginning and end of this three-month
period, manage CHATO training virtual discussion board, and provide technical assistance as
necessary. The process evaluation includes an online implementation survey which identifies
the strategies NHs used to implement the training and includes the Artifacts of Culture
survey. The survey uses the Diffusion of Innovation framework and mirrors the Implementation
Toolkit. Additional process evaluation activities include: Leadership phone interviews
completed by the consultants and external evaluators at LeadingAge and open-ended questions
asked of direct care staff in the CHATO virtual discussion.

Inclusion Criteria

- Nursing Homes that serve people with dementia.

- Nursing Homes that have internet available for staff to complete the CHATO training.

- Nursing Homes that are willing to complete leadership interviews and surveys.

- CNAs and nurses who are permanent employees and who provide direct care at least 8
hours weekly will be invited to complete the CHATO training, available by URL link.

- Other personnel, such as housekeeping, dietary, and administrators may also
participate. Participation by as many staff as possible is desired to achieve
facility-wide communication change.

- Aggregate, deidentified data for residents with Alzheimer's disease or non-Alzheimer's
dementia documented on the MDS Active Diagnoses list.

Exclusion Criteria

- Assisted Living facilities or other types of facilities are excluded due to lack of
MDS data as well as NHs that previously participated in the CHAT study.