VitalPAD: an Intelligent Monitoring and Communication Device to Optimize Safety in the PICU

1. Purpose

   Critically ill patients in the intensive care unit (ICU) need constant monitoring by nurses, respiratory therapists, and physicians. To make decisions and plan therapies, these clinicians must observe and reconcile a vast array of information originating from devices located at the bedsides of multiple patients. Their efforts are hindered by: information being physically limited to the bedside; information overload from the number of devices and separate displays; lack of integration and interaction between devices; displays which fail to present information in a way that is easy to understand; alarm fatigue; lacking historical data or clinical context to support clinical decision making; difficulty locating a team member for consultation or to provide an extra pair of hands

   This project aims to integrate information from multiple monitoring and therapy devices (multi-parameter monitor, infusion pumps, mechanical ventilator) onto one application, the VitalPAD, hosted on a mobile device (tablet or smartphone). The VitalPAD will process information from multiple bedside monitors in an intelligent decision-support system, and provide an intuitive display of vital signs information, alerts and reminders to ICU healthcare team members.

   Assigning priority of care and early detection of deterioration of critically ill patients represents an enormous source of preventable morbidity, mortality and cost. Enhanced efficiency in patient monitoring can help reduce the time to detection of critical conditions and improve priority assignment, contributing to enhanced patient safety.

   The goal of this project is to demonstrate that a mobile application (VitalPAD) can provide dynamic, meaningful, integrated information to intensive care clinicians in real-time. (also see above)

2. Hypothesis / Aims

   Part I - Participatory Design. The participatory design process, as proposed by Holtzblatt, for system development is a mixed-method approach. The aim is to develop a prototype that is easy to use, supports decision making, improves awareness of the patient's condition by ICU healthcare team members, and reduces mental workload.

   Part II - Simulation Experiment. The hypothesis is that in simulated ICU scenarios, critical care physicians, nurses, and respiratory therapists will prioritize the care of critically ill patients more more rapidly and more accurately with the use of VitalPAD than without it. Specifically, using VitalPAD to triage 5 patients will be 30 seconds faster (moderate effect) than performing the same task with the control display. Expected secondary outcomes are a decrease in mental workload as well as an increase in situational awareness.

3. Justification This research represents a significant opportunity to enhance patient safety in ICUs in Canada and elsewhere by directly advancing the conceptual, theoretical and clinical knowledge required to optimize the use of information in the ICU. It will leverage advances in mobile computing to provide real time support to healthcare providers on the move and reduce cognitive demands.

   This integrated system aims to enhance performance of the clinicians in real-time, optimize therapeutic interventions, reduce adverse outcomes and ultimately save lives. The aim is to provide healthcare team members with a one-stop solution for their patient monitoring needs, which will improve the way work is done in the ICU and will decrease worker stress, especially for nurses.

4. Research Design Part I - Participatory design: an inter-professional team of nurses, physicians, designers, engineers, and human factors experts will follow a user-centered design process to refine the system requirements, and obtain feedback during the rapid prototyping phase of prototype development.

   In Part II - Simulation Experiment: the prototype device will be evaluated in a simulated ICU setting using a triaging task. Answer accuracy and time taken to complete task will be measured. Participants will complete questionnaires (NASA-TLX workload assessment and a PSSUQ usability evaluation).

5. Statistical Analysis Part I - Participatory design: qualitative methods will be used to analyze and summarize data. Specific themes will be identified and similar themes merged. Specific obstacles to usability, design preferences, and opportunities for information integration will be collated and summarized. In the usability experiment, error rates and task completion times will be used to detail opportunities for additional usability improvements.

Part II - Simulation Experiment:

• Analysis of Response Time. Mixed linear effect modeling will be used to explore the effects of display, triaging set, and presentation order, and their interactions. Differences between groups (increase or decrease in time taken to triage) will be expressed using the nonparametric Hodges-Lehmann estimator and its 95% confidence limit.
• Analysis of Answer Accuracy. As a secondary outcome, the number of correct responses for each participant will be recorded. The two proportions will be compared using Fisher's exact test. False positive and false negative detection rates will be compared.
• Analysis of Workload Scores. The NASA-TLX scores will be analyzed using two approaches: a Friedman's ANOVA, a non-parametric version of the ANOVA, of each of the 6 individual sub-scores to test for individual effects; and an analysis of the weighted composite score using the same statistical test.
• Analysis of Usability Questionnaire. The PSSUQ results will be tabulated and specific problematic usability issues identified. Friedman's ANOVA will be used to test for individual effects. Feedback provided by the subjects will be screened for improvement of both user interface and decision support system rules.