In this study cameras placed at the bedside will be evaluated for their ability to safely and accurately measure vital signs, such as heart rate and breathing, continuously after heart surgery. Camera-based measurements will be compared with the usual checks that nurses perform several times a day using sensors on the skin or finger clip. The goal is to see whether camera monitoring can help notice changes in a patient's condition earlier. Another aim of the study is to find out whether the camera monitoring can predict if a patient's health is improving or worsening. Patients and healthcare staff will be asked by the investigators about their experience to learn whether this type of monitoring is acceptable in daily care.
Rationale: Over the past decades cardiothoracic surgery has advanced substantially resulting in reduced complication rates. However, when postoperative complications do occur they potentially can have a life-changing impact and sometimes even have death as a result. Therefore, timely detection of clinical deterioration is essential as early treatment may limit the severity of the complication and sometimes prevent their occurrence. The current method of vital signs monitoring, by nurses, usually 3 times a day, often leads to delayed detection of deterioration since early warning signs are missed and to standardized discharge protocols. Objective: The primary objective of the ADVANCE FORSEE project is to evaluate the feasibility and acceptability of continuous video-based monitoring for early detection of postoperative deterioration as well as confirmation of uneventful recovery in cardiothoracic surgery patients. Secondary objectives are (1) to timely detect specific complications such as: atrial fibrillation, tamponade and acute coronary syndrome, (2) to assess the robustness of the video-based monitoring system, (3) to explore if additional physiological and other parameters can reliably be extracted and (4) to develop and evaluate a data management solution capable of efficiently handling large-scale video data. Study design: Observational study, prospective for the data generation and retrospective for the data analysis. User-Centered Design with action research, allowing design and practice-oriented research to iteratively-reinforce each other. Study population: 200-300 cardiac surgery patients. Main study parameters/endpoints: The primary study endpoints are the performance of the deterioration detection and uneventful recovery algorithm assessed using a ROC-analysis. The secondary endpoints are (1) sensitivity, specificity and the time interval between algorithm-based deterioration detection and clinical confirmation by medical staff, (2) the accuracy of heart rate and respiratory rate for varying conditions, (3) the accuracy and feasibility of extracting additional parameters, (4) assessing the data management solution using RMSE as a function of bitrate across compression and storage models, (5) assessment of data loss, incorrect ROI detection and occlusion reported as percentages with 95% CI and compared to reference standards and (6) acceptability of the video-based monitoring system assessed by questionnaires, focus group and interviews. Nature and extent of the burden and risks associated with participation, benefit and group relatedness: Participation to the study is not accompanied by clinical risk to the patient as the proposed technology is completely unobtrusive. Given the objective and hypothesis, the use of video-based monitoring in cardiac surgery patients has potential benefits to patients that outweigh the ethical constraints. Furthermore, the technique has the potential to support Value-Based Health Care (VBHC) principles since it is expected to reduce the number of unexpected adverse events, better risk prediction, reduce hospital stay, reduce the workload of the nursing staff and hereby increase patient wellbeing by freeing up staff time for the patient.
Study Type
OBSERVATIONAL
Enrollment
250
Participants will be continuously monitored with multiple video cameras placed in their room on cardiothoracic surgery ward. The cameras record RGB, NIR and thermal video to allow extraction of vital signs and other non-regulated parameters. No physical contact or additional procedures are required and all usual clinical care continues as normal.
Catharina Ziekenhuis Eindhoven
Eindhoven, North Brabant, Netherlands
The primary study endpoints are the performance of the deterioration detection and uneventful recovery algorithms and the acceptability of the video-based monitoring system among healthcare staff and patients.
Algorithm performance will be assessed using receiver operating characteristics (ROC) analysis expressed as the area under the curve (AUC) with corresponding 95% confidence intervals. As a comparison a ROC curve will be made based on the conventional EWS data. Sensitivity, specificity, positive and negative predictive value will be calculated accordingly.
Time frame: From arrival on the cardiothoracic surgery ward until discharge (typically 2-7 days after surgery).
Sensitivity, specificity and time interval between algorithm-based detection and clinical confirmation.
Agreement between algorithm-based detection and clinical confirmation of postoperative complications. Sensitivity, specificity, and the time interval between the algorithm detection and confirmation by medical staff for events such as atrial fibrillation, acute coronary syndrome, tamponade, and other complications will be assessed.
Time frame: From arrival on the cardiothoracic surgery ward until discharge (typically 2-7 days after surgery).
Accuracy and data coverage of video-derived heart rate and respiratory rate measurements under varying conditions.
Accuracy of video-derived heart rate and respiratory rate are compared with telemetry measurements under varying conditions. Data coverage is defined as the proportion of the monitoring period with usable video-based data.
Time frame: From arrival on the cardiothoracic surgery ward until discharge (typically 2-7 days after surgery).
Accuracy and feasibility of extracting physiological and behavioral parameters such as oxygen saturation, eye movement, paleness and patient movement using the camera monitoring system.
Accuracy and feasibility of video-derived metrics including oxygen saturation, eye movement, paleness and patient movement using the camera monitoring system will be evaluated by comparing camera-based measurements with reference clinical measurements (telemetry-based monitoring) and clinical staff assessments. Measurement validity will be expressed using metrics such as sensitivity, specificity, correlation coefficients and agreement rates. Feasibility will be assessed based on the proportion of usable recordings and successful parameters extractions during postoperative monitoring.
Time frame: From arrival on the cardiothoracic surgery ward until discharge (typically 2-7 days after surgery).
Performance of the data management solution
Performance is evaluated by calculating the RMSE of vital sign measurements using various data-management techniques.
Time frame: From arrival on the cardiothoracic surgery ward until discharge (typically 2-7 days after surgery).
Quantitative assessment of acceptability of the video-based monitoring
Quantitative assessment of the acceptability of the video-based monitoring in terms of privacy, comfort, feeling of safety, ease of use and trust in the system will be assessed using structured questionnaires (Likert scale).
Time frame: At postoperative day 5 or at hospital discharge whichever occurs first.
Qualitative assessment of acceptability of the video-based monitoring
Qualitative assessment of the acceptability of the video-based monitoring in terms of privacy, comfort, feeling of safety, ease of use and trust in the system using semi-structured interviews and focus groups.
Time frame: At postoperative day 5 or at hospital discharge whichever occurs first.
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.