A method and system for early dynamic early warning of ards based on conventional non-invasive parameters

Abstract

The invention discloses an ARDS early dynamic early warning method and system based on conventional non-invasive parameters. The method includes: acquiring non-invasive parameters of patients, preprocessing the non-invasive parameters, and constructing feature data; standardizing the feature data to generate standardized feature data; The standardized feature data is input into the pre-trained intelligent dynamic early warning model to obtain the original probability; the corrected early warning probability is calculated; when the corrected early warning probability of the patient is greater than or equal to 0.5, it is judged that the patient will get ARDS at this time. The present invention only uses commonly used non-invasive parameters, does not require laboratory data, expands the scope of application of the present invention, makes it possible to use it in remote areas, public health emergencies and front-line situations on the battlefield, and eliminates frequent collection of laboratory parameters The harm caused to individual patients reduces the cost of use, and can provide better early warning capabilities for the onset of ARDS. The early warning interval reserved for clinical intervention can provide sufficient time for doctors to design patient treatment plans.

Description

technical field [0001] The invention relates to the field of artificial intelligence technology and medical health, in particular to an ARDS early dynamic early warning method and system based on conventional non-invasive parameters. Background technique [0002] Acute respiratory distress syndrome (ARDS, Acute respiratory distress syndrome) is a clinical syndrome caused by non-cardiogenic pulmonary edema caused by various acute stimuli such as mechanical hyperventilation, sepsis, trauma, shock and disseminated intravascular coagulation. , with refractory progressive hypoxemia as the main clinical manifestation. ARDS disease is characterized by high morbidity and high mortality. ARDS occurs in 10% of all ICU patients and 23% of all mechanically ventilated patients. ARDS has a rapid onset and can occur within 24 to 48 hours. , can also be as long as 5 to 7 days. Although ARDS mortality is decreasing in clinical trials, it remains around 40% in major observational studies. ...

AI Technical Summary

Problems solved by technology

First of all, although the underlying algorithms of these methods are all public, in practice some risk adjustment algorithms contain laboratory data, and the acquisition of these parameters will result in high medical costs, and it is even more difficult to obtain in remote areas

In addition, although chest x-rays are permitted for early stage diagnosis of ARDS, they expose patients to radiation doses

Early detection of ARDS can be achieved with airway aspiration or lavage, the disadvantage of these methods is that they are invasive and harmful to the patient

All of the above methods are not suitable for frequent verification of whether patients suffer from ARDS or may develop ARDS; second, these algorithms do not reserve enough time for clinical intervention between the learning window and the early warning window (onset time), and cannot Better meet the actual clinical needs

Third, the traditional risk adjustment algorithm is limited by static low-frequency data, ignoring a large amount of information contained in dynamic time series data, the accuracy of early warning is low, and the guiding significance of early warning results is not strong

Fourth, these methods usually require medical staff to frequently record and calibrate patient data. The scoring of these algorithms often relies on the medical experience and clinical observations of medical staff, which is highly subjective and requires a large amount of medical human resources. It also increases the difficulty of applying the risk adjustment algorithm

Patent CN111407250A discloses a system for monitoring the development of ARDS in patients, but the system input parameters include respiratory mechanics parameters, pulse index continuous cardiac output, central venous pressure and pulmonary artery pressure, oxygen metabolism power, end-tidal carbon dioxide, etc. It is difficult to obtain hardware support in public health emergencies, frontline battlefields or remote areas, etc.

[0005] Although the above patents can realize the possible prediction of the patient's ARDS onset through the patient's physiological parameters, these are not designed for the whole-process monitoring of the patient's ARDS dynamic early warning, lack of difficulty in data collection, time series data mining, reserved early warning model running time point and onset Clinical intervention time between early warning time points, consideration of the possibility of dynamic monitoring throughout the hospitalization of patients

At present, there is no patent related to the dynamic early warning system for acute respiratory distress syndrome based on conventional non-invasive parameters

Images