Speeko for Elderspeak

SPEEKO for Elderspeak: A Self-Monitoring App to Improve Communication and Reduce Behavioral Symptoms in Care for Persons With Alzheimer's Disease and Other Dementias

This study will test feasibility and preliminary effects of an automated and performance-based feedback app (Speeko for Elderspeak) that detects and reports the use of diminutives (terms of endearment such as honey, dearie, and sweetie), prevalent elderspeak terms linked to BPSD. Building on proof of concept established in the laboratory, the app will demonstrate feasibility at the point-of-care. Next, a clinical trial (N= 6NHs) will be conducted to test preliminary efficacy of the app for amplifying reductions in diminutives use for NH staff completing the CHATO training (3 online modules).

A new person is diagnosed with Alzheimer's disease or other dementia every 66 seconds, and most persons with dementia (PWD) spend the late stages of dementia in nursing homes (NHs) where lack of dementia care skills and staff shortages limit quality of care. Care of PWD in NHs is complicated by behavioral and psychological symptoms of dementia (BPSD) such as aggression, vocal outbursts, wandering, and withdrawal that occur as PWD lose cognitive and communication abilities and cannot express their unmet physical and psychosocial needs. BPSD present to NH staff as resistiveness to care (RTC) that increases staff stress and costly time to complete care, often leading to staff turnover, injury, and inappropriate use of psychotropic medications to control BPSD. Although Center for Medicare and Medicaid Services (CMS) mandates and penalties have reduced antipsychotic medication use slightly, contraindicated use in NH residents remains a pervasive problem, causing harmful side effects and reducing the quality of life for PWD. The PI and other researchers have empirically verified that RTC occurs when NH staff use elderspeak (speech similar to baby talk) that features inappropriately intimate terms of endearment (diminutives such as "honey"), belittling pronoun substitutions that imply dependence ("the investigators" need a bath), and harsh task-oriented commands ("sit down"). Elderspeak conveys a message of disrespect and incompetence to residents who react with withdrawal or BPSD. The R03 study established that residents with dementia are more than twice as likely to exhibit BPSD (measured by coding RTC behavior in videos) when staff use elderspeak instead of normal adult communication. The subsequent CHAT R01 trial verified that staff reduced their use of elderspeak after attending the three-session Changing Talk (CHAT) communication training program, and that this reduced RTC. To facilitate dissemination, interactive online CHAT modules have been developed (CHATO) which provides the same CHAT classroom content with the advantage of flexible access via the internet for busy NH staff, including those in rural areas and small, independent NHs. Recognizing the delay and incomplete adoption and application of evidence-based skills in practice, the PI and colleagues believe additional strategies to optimize CHATO skills implementation are indicated. Performance-based reinforcement of skills is effective in achieving greater immediate implementation and long-term maintenance of new skills use in practice. However, feasibility and costs for individualized expert feedback in NH settings are usually prohibitive and thus not widely used in practice. This study will test feasibility and preliminary effects of an automated and performance-based feedback app that detects and reports the use of diminutives (terms of endearment such as honey, dearie, and sweetie) which are prevalent elderspeak terms linked to BPSD. The PI and colleagues developed the SPEEKO for Elderspeak app using archived NH recordings from previous research to identify the most commonly used diminutives and then to develop algorithms to detect them in speech. Building on proof of concept established in the laboratory, the app will demonstrate feasibility at the point-of-care. Next, a clinical trial (N= 6NHs) will be conducted to test preliminary efficacy of the app for amplifying reductions in diminutives use for NH staff completing the CHATO training. SPECIFIC AIMS: AIM 1. Demonstrate feasibility, acceptability, and validity of the SPEEKO for Elderspeak feedback app use by staff in the NH. Five certified nursing assistants (CNAs) will use the app during NH care and provide feedback about any needed modifications. Hypothesis: The app will be readily used and acceptable to CNAs. Diminutive counts determined by the app will be correlated with psycholinguistic analysis, validating accuracy. AIM 2. Test preliminary effects of an innovative self-monitoring feedback app on staff elderspeak use and compare accuracy to psycholinguistic analyses of audio-recorded staff communication. Hypothesis: Staff who receive immediate app feedback (N= 30 in three NHs) will have greater elderspeak reduction after completing CHATO training compared to delayed feedback controls (N= 30 in three NHs). AIM 3. Evaluate app acceptability and costs. Hypothesis: Staff who use the app will find it acceptable and beneficial for their practice. Process-based costing will be used to determine costs for app use in NHs and other long-term service and support settings. Costs will be compared in relation to each group's outcomes.

Six nursing homes will be randomly assigned to treatment and control conditions. All NHs will received the CHATO training. The treatment NHs will received immediate feedback from the Speeko App after each recording sessions whereas the Control NHs will not receive feedback until all recording sessions are complete.

Change in elderspeak measured by the number of diminutives detected in participant's recorded speech.

The number of diminutives detected by the Speeko for Elderspeak app will be correlated with manual psycholinguistic analysis to determine app accuracy. Diminutive data will be compared between the intervention and control groups and within nursing homes.

A validated, 10-item industry standard survey. Questions are in 5-point Likert scale with responses from strongly agree to strongly disagree. Question scores are converted to a 0-100 total score where lower scores indicate low usability and high scores indicate greater usability.

Survey completed after participants finish all recording sessions. At the end of the 3-month implementation period per nursing home

Evaluate the app cost by measuring wages per hour by NH role and costs associated with the Speeko for Elderspeak application.

The cost associated with elderspeak reduction will be calculated using primary outcomes, wage data, and application costs.

Survey completed by each nursing home after the CHATO training. By the end of the 3-month implementation period per nursing home

Inclusion Criteria: CNAs who are age 18 or older Permanent employees Fluent in English Exclusion Criteria: CNAs younger than 18 or CNAs who are not fluent in English.