ICU-Talk, A Communication Aid for Intubated Intensive Care Patients
F. MacAulay, A. Judson, M. Etchels, S. Ashraf, I. W. Ricketts, A. Waller, J. K. Brodie, N. Alm, A. Warden, A. J. Shearer, B. Gordon · 2002 · Proceedings of the Fifth International ACM Conference on Assistive Technologies (Assets 02) · doi:10.1145/638249.638290
Summary
This paper describes ICU-Talk, a computer-based communication aid developed specifically for intubated patients in hospital intensive care units (ICUs) who are temporarily unable to speak due to breathing tubes inserted through their mouth or throat. The three-year multidisciplinary project brought together nursing, computer science, and speech and language therapy personnel at the University of Dundee and Ninewells Hospital in Scotland. ICU patients face a unique communication challenge: they are acutely ill, often cognitively compromised by medication, physically weak, and may have restricted hand movement due to drips and monitors — yet their inability to communicate is one of the most distressing aspects of their ICU stay, with lasting psychological effects. Existing AAC devices were unsuitable because they were designed for long-term use by people with permanent disabilities, not for the temporary, acute needs of ICU patients. ICU-Talk was designed to be quick to learn and easy to use, with two interface styles: a "Boxes" interface (grid of rectangular buttons, accommodating up to 10 phrases per screen) and a "Bubbles" interface (a 2D Cone-Trees visualization with up to 6 phrases per screen, better suited for single switch scanning). Both interfaces organized approximately 250 pre-stored phrases across eight color-coded topics (General Questions, Chit-Chat, Statements, Family/Visitors, Feelings, Questions About Me, Interests). The phrase database was developed in two stages: first, 75% of nursing staff surveyed identified common patient communication topics; then, observation of actual patients attempting to communicate refined the phrase set into core (universal) and personalized categories.
Key findings
The device was trialled with real patients at Ninewells Hospital ICU starting May 2001. Six patients had successfully used the system to communicate with nursing staff, family, and friends at the time of the paper. Both planned admissions and emergency admissions were included in the trial. An automated event recorder tracked all patient selections, revealing conversation patterns and frequently used phrases. However, some patients repeatedly selected the same phrase in very short time frames, either accidentally, due to tremor, or because their communication partner did not hear the first selection — making simple frequency statistics unreliable for determining the most useful phrases. A mid-point questionnaire completed by 32 of 44 nurses yielded strong support: 97% felt nurses should be involved with using ICU-Talk, 90% felt ICU patients need a computer-based communication aid, 88% felt communication ability affects patient well-being, and 74% felt patients fail to communicate effectively through mouthing and gesture alone. However, nurses also raised concerns: 71% felt ICU-Talk obstructed their view of the patient, 55% found the device difficult to manoeuvre in the ICU environment, and the phrase database organization was identified as the most significant problem — with 250 phrases across hierarchical topics, patients struggled to find relevant phrases or did not know what was available.
Relevance
This paper addresses an underserved population in AAC research: people with temporary communication disabilities in acute medical settings. While most AAC research focuses on individuals with permanent conditions (cerebral palsy, ALS, aphasia), ICU patients represent a large group who need communication support for days to weeks during a critical period. The research highlights design challenges unique to temporary AAC: users have no time to learn complex systems, their cognitive abilities are compromised, the physical environment is hostile to technology (cleaning requirements, space constraints, patient positioning), and the device must be quickly personalized for each new patient. For practitioners, the key lessons include the critical importance of involving nursing staff in design, the challenge of organizing large phrase databases for cognitively impaired users, and the tension between device functionality and practical ICU constraints like sightlines and maneuverability. The finding that phrase organization and retrieval is the primary usability barrier anticipates ongoing challenges in AAC research around vocabulary organization and prediction.
Tags: augmentative and alternative communication · AAC · intensive care · intubation · hospital communication · touch screen · switch access · patient communication