A Speech Recognition Application as a Communication Aid for Acute and Critical Care Patients With Tracheostomies

Study Design: This is a mixed-method, multi-centre prospective observational cohort study to establish the feasibility of SRAVI. Mixed methods include:

• A multi-centre prospective observational cohort study of SRAVI as an addition to usual communication aids.
• Qualitative interviews/focus groups to inform future study design by exploring patients', their significant others and the critical care MDTs' subjective experiences of the study intervention and outcome measures.

Sample size: The investigators aim to recruit a minimum target of 55 patients. For every patient recruited, the investigators anticipate that three qualitative interviews will be undertaken (i.e., with the patient, their significant other, and at least one healthcare professional). Therefore, a feasible target recruitment is 165 participants.

Consent: Written informed consent will be obtained from participants or a personal or professional consultee in accordance with ethical approval.

Data collection: Data will be collected by the investigators and recorded in the study case report form (CRF).

Baseline data

• Inclusion/exclusion criteria and eligibility screen
• Sex
• Age on admission to hospital
• Date of acute/critical care admission
• Admission diagnosis
• Date of tracheostomy insertion during acute/critical care

Daily data collected

• CAM-ICU score (critical care patients only)
• Frequency of SRAVI delivery
• Compliance with SRAVI use
• Adverse events

Feasibility and clinical outcome data collected in acute/critical care

• Screening
• Recruitment rates
• Adverse events
• Delirium (critical care patients only)

Data collected after critical care discharge

• Duration of critical care stay

Data collected during virtual/telephone/in-person follow-up 3 months following acute/critical care discharge:

• Health related quality of life (HRQoL)
• Post-Traumatic Stress Disorder (PTSD)
• Anxiety and depression
• Cognitive status

Data analysis: Descriptive analysis will be used to analyse data from the observational cohort study. Numbers of patients screened, eligible, recruited, consented, and withdrawn from the study will be reported. Baseline demographic and clinical data will be summarised for study participants. Continuous variables will be summarised as mean (standard deviation) and median (interquartile range) and categorical variables will be summarised as number (percent).

Qualitative interviews/focus groups: Interviews/focus groups will be conducted by the investigator. Interview schedules/focus group topic guides will be developed from discussions within the research team, the patient advisory group, and from literature around study participation. The schedules will include a pre-defined list of questions that will be informed by the objectives of the study. The schedule will enable all participants to be asked similar questions and thus permit comparison of themes across each subject during data analysis.

Qualitative data collection: Interviews/focus groups will be audio recorded and professionally transcribed verbatim using an authorised transcription service. All identifying information will be removed prior to analysis. Transcriptions will be reviewed and verified prior to analysis by the research fellow by comparing the audio and written versions to identify errors.

Qualitative analysis: Qualitative data will be subjected to a thematic analysis, using Newell and Burnard's framework. This approach will permit an inductive process of drawing out important data-driven themes and a deductive process relating the major themes that emerge to the pre-defined objectives of the research. A process of constant comparison, reading, and re-reading of the data will enable identification of emerging themes. The process will be facilitated by using a computer-assisted qualitative data software package, NVIVO.