Document Actions
ICUTalk
Latest release: ICUTalk 1.3.32
You must read the disclaimer before downloading.
Until I sort out a subversion area, you can download a zip file of the binary and source for ICUTalk. I've just spent several hours fixing the scanning code for the bubbles view (this is the only view with scanning capability), it appears to be reasonably stable but I've not done any thorough testing. If there are any critical bugs / changes that you'd like me to fix then please contact me, otherwise help yourself but please also let me know me about them.
Personnally, I doubt I have the sanity to try to maintain code I started writing 8 years ago... a rewrite is very much needed! If you'd like to help out then please contact me.
-
Read more
Before you use ICUTalk
It is important to note that the ICUTalk software is just part of a solution, i.e. you need hardware to run it on, and someone to support the software & hardware, the users and the nursing staff etc. The work to date was solely done as part of a research project, of which the Ninewellls Hospital ICU and SLT (Speech & Language Therapy) were partners, hence we had the staff to help facilitate as needed. If you are a family member of a patient who is in ICU and thinks it may help, then you should mention it to the ICU staff, and ask them to involve the SLT team.
Due to the nature of the ICU environment, you can't just use a typical pc or laptop. The hardward needs to be robust and cleanable, ideally a ruggedised tablet pc should be used. Thus you need to think carerfully about the hardware you choose to run the system. We have used the ix104c2 dual mode tablet pc from xplore technologies which are extremely rugged and were sufficiently powerful, although the volume / power of the speakers isn't ideal.
News
03/12/2007 by Andy Judson...
Although ICUTalk, as a funded research project ended in 2002, I continued to develop the software as part of my PhD research between 2002 and 2005. ICUTalk was tested with ~20 ICU patients from Ninewells Hospital in Dundee over a period of 1 year during the initial phase. After much debate and odd notes of interest we have decided to release ICUTalk under an open source license. At this time, no extra research funding has been saught but we might in the future. Want to find out more about ICUTalk... well you can read my PhD thesis (contact me for a electronic copy), read the main journal article (care of the critically ill; february 2003; vol 19 no 1 pp 4-9) or just contact me.
ICUTalk was implemented in Java several years ago, much of the original code still exists with several rounds of tidying / refactoring over the years... but it is not a production piece of software, and code wise I'm sure it could be vastly improved. What are the plans (roughly) for ICUTalk ??? 1. get it hosted on oats (as is) and recruit some other developers; 2. fix critical bugs (e.g. scanning and speech output volume); 4. freeze java branch and rewrite in c#... as this is what i tend to develop in nowadays (.NET3 and WPF look to be of particularly cool for this task).
Background
The ICU patient is critically ill and in the majority of cases the patient has the support of an artificial ventilator to augment their breathing. A side effect of artificial ventilation is the incapacity for the patient to speak; combine this with the affects of their condition, the treatment they receive, the ICU environment and the array of drugs controlling their pain and sedation,then a patient's visit to ICU becomes a highly emotional and frightening experience.
In 2002, Scotland 's 8748 ICU patients stayed in the unit on average 2 days, 21.9% of them died as ICU patients, and a further 7.5% died as patients on another ward. Communication is believed to be an important factor in improving ICU patients' overall experience of intensive care, their well-being and their chances of survival. To date, though, little has been successfully achieved with Augmentative and Alternative Communication (AAC) methods in ICU. Current methods used by patients include: mouthing, gesture, and alphabet charts. However, these have been shown to be time consuming and frustrating. The investigation of ICU patients' usage of VOCAs has really just begun. In 2000, positive results were reported on the usage of a commercial VOCA where patients were trained prior to their visit to ICU. A study in 2004 reported further promising results with VOCA usage by patients who were not trained prior to their visit; this study did however find that device complexity, among other factors, to be a significant drawback. In both of these studies the VOCAs used were designed for the long-term AAC user, that is, they used abstract symbolic images to link to pre-stored conversational items. Existing VOCAs are designed for the long term user and in many instances use symbolic representations of language concepts or symbolic representation to link to pre-stored conversational items. Although off-the-shelf VOCAs can be used for non-elective ICU patients, the environment, impact of medication on the patient's ability to concentrate and the limited duration of stay means that the patient is not able to receive the time needed for training. Training ICU patients to use VOCAs that require hours of training is thus difficult, if not impossible to deliver.
Following a user-centred design methodology, the development of ICUTalk provided patients with 3 possible interfaces: the first and second interfaces provide alternative layouts to approximately 200 pre-stored text-based utterances, organised into 8 categories; whilst the third provides an alphabet board. Following the development and pilot evaluation stages, the system was introduced to the Ninewells Hospital ICU (Dundee, Scotland), for evaluation by patients over a 1 year period. Based on the evidence collated through various questionnaires and automated logging, ICUTalk was shown to be successfully used by patients after about 5 minutes of training. The results of the ICUTalk research show that a (bespoke) VOCA could be used by ICU patients, though there is still a lot to be learnt and improved. Conventional VOCAs, however, may not be the most appropriate solution as they are usually designed for a vastly different user group.
Acknowledgements
ICUTalk was the work of a really solid muliti-disciplinary team (listed below); I hope we can build upon what this original team achieved.
- University of Dundee [School of Computing]: Andy Judson, Ian Ricketts, Annalu Waller, Norman Alm, Saqib Ashraf
-
University of Dundee [School of Nursing & Midwifery]: Brian Gordon
Ninewells Hospital [Intensive Care Unit]: Maria Etchels, Alf Shearer, Audrey Warden
Ninewells Hospital [Speech & Language Therapy]: Fiona MacAulay, Jan Brodie
