Borja Barrachina1, Raquel Cobos, Noemi Mardones, Angel Castañeda, Cristina Vinuesa. 1. From the Anaesthesia and Resuscitation Unit, (BB); Araba Research Unit, Bioaraba Research Institute (RC) and Intensive Care Unit, (NM, AC, CV) Araba Integrated Health Organization (University Hospital), Vitoria-Gasteiz, Álava, Spain.
Abstract
BACKGROUND: Less invasive and noninvasive methods are emerging for haemodynamic monitoring. Among them is Capstesia, a smartphone app that, from photographs of a patient monitor showing invasive arterial pressure, estimates advanced haemodynamic variables after digitising and analysing the pressure curves. OBJECTIVE: The aim of this study was to compare the level of agreement between the analysis of the signals obtained from the patient monitor and a photograph of the same images using the Capstesia app. DESIGN: Cross-sectional study. SETTING: Araba University hospital (Txagorritxu), Vitoria-Gasteiz, Alava, Spain, from January to February 2015. PATIENTS: Twenty patients (229 images) who had an arterial catheter (radial or femoral artery) inserted for haemodynamic monitoring. INTERVENTION: Snapshots obtained from the patient monitor and a photograph of these same snapshots using the Capstesia application were assessed with the same software (MATLAB, Mathworks, Natick, Massachusetats, USA) for evaluating the level of concordance of the following variables: pulse pressure variation (PPV), cardiac output (CO) and maximum slope of the pressure curve (dP/dt). Comparison was made using interclass correlation coefficients with corresponding 95% confidence intervals, and Bland-Altman plots with the corresponding percentages of error. MAIN OUTCOME MEASURES: (PPV). Secondary outcome: CO and maximum slope of the pressure curve [dP/dt]. RESULTS: The interclass correlation coefficients for PPV, CO and max dP/dt were 0.991 (95% confidence interval 0.988 to 0.993), 0.966 (95% confidence interval 0.956 to 0.974) and 0.962 (95% confidence interval 0.950 to 0.970), respectively. In the Bland-Altman analysis, bias and limits of agreement of PPV were (0.50% ± 1.42) resulting in a percentage of error of 20% for PPV. For CO they were 0.19 ± 0.341, with a 13.8% of error. Finally bias and limits of agreement for max dP/dt were 1.33 ± 77.71, resulting in an error of 14.20% CONCLUSIONS: Photograph of the screenshots obtained with the Capstesia app show a good concordance with analysis of the original screenshots. Either approach could be used to monitor the haemodynamic variables assessed.
BACKGROUND: Less invasive and noninvasive methods are emerging for haemodynamic monitoring. Among them is Capstesia, a smartphone app that, from photographs of a patient monitor showing invasive arterial pressure, estimates advanced haemodynamic variables after digitising and analysing the pressure curves. OBJECTIVE: The aim of this study was to compare the level of agreement between the analysis of the signals obtained from the patient monitor and a photograph of the same images using the Capstesia app. DESIGN: Cross-sectional study. SETTING: Araba University hospital (Txagorritxu), Vitoria-Gasteiz, Alava, Spain, from January to February 2015. PATIENTS: Twenty patients (229 images) who had an arterial catheter (radial or femoral artery) inserted for haemodynamic monitoring. INTERVENTION: Snapshots obtained from the patient monitor and a photograph of these same snapshots using the Capstesia application were assessed with the same software (MATLAB, Mathworks, Natick, Massachusetats, USA) for evaluating the level of concordance of the following variables: pulse pressure variation (PPV), cardiac output (CO) and maximum slope of the pressure curve (dP/dt). Comparison was made using interclass correlation coefficients with corresponding 95% confidence intervals, and Bland-Altman plots with the corresponding percentages of error. MAIN OUTCOME MEASURES: (PPV). Secondary outcome: CO and maximum slope of the pressure curve [dP/dt]. RESULTS: The interclass correlation coefficients for PPV, CO and max dP/dt were 0.991 (95% confidence interval 0.988 to 0.993), 0.966 (95% confidence interval 0.956 to 0.974) and 0.962 (95% confidence interval 0.950 to 0.970), respectively. In the Bland-Altman analysis, bias and limits of agreement of PPV were (0.50% ± 1.42) resulting in a percentage of error of 20% for PPV. For CO they were 0.19 ± 0.341, with a 13.8% of error. Finally bias and limits of agreement for max dP/dt were 1.33 ± 77.71, resulting in an error of 14.20% CONCLUSIONS: Photograph of the screenshots obtained with the Capstesia app show a good concordance with analysis of the original screenshots. Either approach could be used to monitor the haemodynamic variables assessed.