Journal of clinical monitoring and computing
-
J Clin Monit Comput · Oct 2024
Early prediction of ventricular peritoneal shunt dependency in aneurysmal subarachnoid haemorrhage patients by recurrent neural network-based machine learning using routine intensive care unit data.
Aneurysmal subarachnoid haemorrhage (aSAH) can lead to complications such as acute hydrocephalic congestion. Treatment of this acute condition often includes establishing an external ventricular drainage (EVD). However, chronic hydrocephalus develops in some patients, who then require placement of a permanent ventriculoperitoneal (VP) shunt. ⋯ At that point, the accuracy of the prediction was 76% (CI: 75.98-83.09%), with a sensitivity of 85% (CI: 83-88%) and a specificity of 74% (CI: 71-78%). RNN-based machine learning has the potential to predict VP shunt dependency on Day 4 after ictus in aSAH patients using routine data collected in the ICU. The use of machine learning may allow early identification of patients with specific therapeutic needs and accelerate the execution of required procedures.
-
J Clin Monit Comput · Oct 2024
Effect of vertical stopcock position on start-up fluid delivery in syringe pumps used for microinfusions.
The purpose of this in vitro study was to evaluate the impact of the vertical level of the stopcock connecting the infusion line to the central venous catheter on start-up fluid delivery in microinfusions. Start-up fluid delivery was measured under standardized conditions with the syringe outlet and liquid flow sensors positioned at heart level (0 cm) and exposed to a simulated CVP of 10 mmHg at a set flow rate of 1 ml/h. Flow and intraluminal pressures were measured with the infusion line connected to the stopcock primarily placed at vertical levels of 0 cm, + 30 cm and - 30 cm or primarily placed at 0 cm and secondarily, after connecting the infusion line, displaced to + 30 cm and - 30 cm. Start-up fluid delivery 10 s after opening the stopcock placed at zero level and after opening the stopcock primarily connected at zero level and secondary displaced to vertical levels of + 30 cm and - 30 cm were similar (- 10.52 [- 13.85 to - 7.19] µL; - 8.84 [- 12.34 to - 5.33] µL and - 11.19 [- 13.71 to - 8.67] µL (p = 0.469)). ⋯ Start-up fluid delivery with the stopcock primarily placed at + 30 cm and - 30 cm resulted in large anterograde and retrograde fluid volumes of 34.39 [33.43 to 35.34] µL and - 24.90 [- 27.79 to - 22.01] µL at 10 s, respectively (p < 0.0001). Fluid delivered with the stopcock primarily placed at + 30 cm and - 30 cm resulted in 140% and 35% of calculated volume at 360 s, respectively (p < 0.0001). Syringe infusion pumps should ideally be connected to the stopcock positioned at heart level in order to minimize the amounts of anterograde and retrograde fluid volumes after opening of the stopcock.
-
Prediction and avoidance of intraoperative hypotension (IOH) can lead to less postoperative morbidity. Machine learning (ML) is increasingly being applied to predict IOH. We hypothesize that incorporating demographic and physiological features in an ML model will improve the performance of IOH prediction. ⋯ The multivariate model yielded average AUC = 0.97 in the static context of a single prediction made up to 8 min before a possible IOH event, and it outperformed a univariate model based on MAP-only (average AUC = 0.83). The MMV model demonstrated AUC = 0.96, PPV = 0.89, and NPV = 0.98 within the challenging context of a dynamic sliding window across 40 min prior to a possible IOH event. We present a novel ML model to predict IOH with a distinctive "dial" on sensitivity and specificity to predict first IOH episode during liver resection surgeries.
-
J Clin Monit Comput · Sep 2024
ReviewA review of machine learning methods for non-invasive blood pressure estimation.
Blood pressure is a very important clinical measurement, offering valuable insights into the hemodynamic status of patients. Regular monitoring is crucial for early detection, prevention, and treatment of conditions like hypotension and hypertension, both of which increasing morbidity for a wide variety of reasons. ⋯ Non-invasive techniques, in contrast, reduce these risks and can provide intermittent or continuous blood pressure readings. This review explores modern machine learning-based non-invasive methods for blood pressure estimation, discussing their advantages, limitations, and clinical relevance.
-
J Clin Monit Comput · Sep 2024
Left ventricular end-diastolic pressure response to spinal anaesthesia in euvolaemic vascular surgery patients.
Regional anaesthesia techniques provide highly effective alternative to general anaesthesia. Existing evidence on the effect of spinal anaesthesia (SA) on cardiac diastolic function is scarce. This study aimed to evaluate the effects of a single-injection, low-dose SA on left ventricular end-diastolic pressures (LVEDP) using echocardiography in euvolaemic patients undergoing elective vascular surgery. ⋯ This study in euvolaemic patients undergoing elective vascular surgery provides evidence that SA improved LVEDP.