Opening Pandora's box by generating ICU diaries through artificial intelligence: A hypothetical study protocol

BACKGROUND

Patients and families on Intensive Care Units (ICU) benefit from ICU diaries, enhancing their coping and understanding of their experiences. Staff shortages and a limited amount of time severely restrict the application of ICU diaries. To counteract this limitation, generating diary entries from medical and nursing records using an artificial intelligence (AI) might be a solution.

DESIGN AND PURPOSE

Protocol for a hypothetical multi-center, mixed method study to identify the usability and impact of AI-generated ICU diaries, compared with hand-written diaries.

METHOD

A hand-written ICU diary will be written for patients with expected length of stay ≥ 72 h by trained nursing staff and families. Additionally at discharge, the medical and nursing records are analyzed by an AI software, transformed into understandable, empathic diary entries, and printed as diary. Based on an appointment with patients within 3 months, diaries are read in randomized order by trained clinicians with the patients and families. Patients and families will be interviewed about their experiences of reading both diaries. In addition, usability of diaries will be evaluated by a questionnaire.

EXPECTED FINDINGS AND RESULTS

Patients and families describe the similarities and differences of language and the content of the different diaries. In addition, concerns can be expressed about the generation and data processing by AI.

IMPLICATIONS FOR PRACTICE

Professional nursing involves empathic communication, patient-centered care, and evidence-based interventions. Diaries, beneficial for ICU patients and families, could potentially be generated by Artificial Intelligence, raising ethical and professional considerations about AI's role in complementing or substituting nurses in diary writing.

CONCLUSIONS

Generating AI-based entries for ICU diaries is feasible, but raises serious questions about nursing ethics, empathy, data protection, and values of professional nurses. Researchers and developers shall discuss these questions in detail, before starting such projects and opening Pandora's box, that can never be closed afterwards.

[ChatGPT: curse or blessing in nursing care?]

Artificial intelligence (AI) has been used in healthcare for some years for risk detection, diagnostics, documentation, education and training and other purposes. A new open AI application is ChatGPT, which is accessible to everyone. The application of ChatGPT as AI in education, training or studies is currently being discussed from many perspectives. It is questionable whether ChatGPT can and should also support nursing professions in health care. The aim of this review article is to show and critically discuss possible areas of application of ChatGPT in theory and practice with a focus on nursing practice, pedagogy, nursing research and nursing development.

[The intensive care diary-a communication tool]

Diaries in intensive care units (ICU) are mostly written by families and nurses for patients with impaired consciousness. In the diary, the development of the patients is described in plain language by daily reports. Patients can read the diary later and process their experiences and, if necessary, reframe them differently. ICU diaries reduce the risks of psychosocial sequelae of patients and families and are now used worldwide. Diaries have different purposes and also serve as a communication tool: words are written to a person who will hopefully read them later. This can help families to stay connected and cope better with the situation. However, writing a diary can also be perceived as a burden by some relatives and nurses due to lack of time or because it seems to be too close. ICU diaries can serve as a tool for patientand family-centered care.

The reliability of the German version of the Richards Campbell Sleep Questionnaire

BACKGROUND

The assessment of sleep quality in critically ill patients is a relevant factor of high-quality care. Despite the fact that sleep disturbances and insufficient sleep management contain an increased risk of severe morbidity for these patients, a translated and applicable instrument to evaluate sleep is not available for German-speaking intensive care settings.

AIM

This study aimed to translate the Richards Campbell Sleep Questionnaire (RCSQ), a simple and validated instrument eligible for measuring sleep quality in critically ill patients, and subsequently to evaluate the internal consistency of the German version of the RCSQ. Furthermore, it also aimed to inquire into the perception of sleep in a sample of critically ill patients.

METHODS

The RCSQ was translated following established methodological standards. Data were collected cross-sectionally in a sample of 51 patients at 3 intensive care units at a university hospital in Germany.

RESULTS

The German version of the RCSQ showed an overall internal consistency (Cronbach's alpha) of 0·88. The mean of the RSCQ in the sample was 47·00 (SD ± 27·57). Depth of sleep was rated the lowest and falling asleep again the highest of the RCSQ sleep items.

CONCLUSION

The study demonstrated very good internal consistency of the German version of the RCSQ, allowing for its application in practice and research in German-speaking countries. Quality of sleep perception was generally low in this sample, emphasizing the need for enhanced care concepts regarding the sleep management of critically ill patients. Relevance to clinical practice Assessment of self-perception of sleep is crucial in order to plan an individually tailored care process.