A medical instrument related pressure injury prediction and early warning method and system
By collecting and analyzing pressure distribution data of the contact surface between medical devices and skin, calculating the effective contact width and linear load amplification index at the edge points, and generating an instantaneous risk score, the problem of underreporting of changes in the edge compression pattern of medical devices in existing technologies is solved, and accurate early warning of pressure injuries is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGDONG NO 2 PROVINCIAL PEOPLES HOSPITAL
- Filing Date
- 2026-02-11
- Publication Date
- 2026-06-05
Smart Images
Figure CN122140190A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of medical care risk monitoring and early warning technology, and in particular to a method and system for predicting and warning of medical device-related pressure injuries. Background Technology
[0002] In clinical nursing and medical treatment, non-invasive ventilation masks, nasal cannulas, neck collars, fixation straps, and various braces are widely used. To maintain the expected therapeutic effect or fix the body position, these devices usually need to be closely attached to the patient's skin or mucous membrane surface, thus forming a continuous mechanical external force interface on the human body surface. However, this long-term mechanical pressure can easily lead to medical device-related pressure injuries. In actual application scenarios, the contact interface between the device and the skin is not uniformly stressed, and injuries often occur at the contour edges of the device, rigid buckles, pressure fixation strips, and other parts. In these areas, the device is prone to changes in local pressure patterns due to misalignment, improper fixation force, or changes in the patient's body position. In particular, when the contact pattern deteriorates, local stress concentration points are easily formed at the edges.
[0003] Existing monitoring or early warning technologies typically focus on collecting overall pressure data of the contact area and rely primarily on statistical indicators such as regional average pressure, maximum peak pressure, or wearing time to assess damage risk. However, this assessment method based on regional statistics is difficult to accurately capture subtle changes in contact status at the contact edge. Because regional average pressure has a natural smoothing and diluting effect on local data, when embedded pressure at the device edge causes a sharp collapse in the effective contact width, degenerating from a wide surface contact to an extremely narrow linear contact, the rapidly increasing linear load in this extremely narrow area is often masked by the values of the surrounding large area of low pressure. Existing detection methods cannot effectively identify this geometric collapse process of contact width and are unable to issue alarms when the overall fixation tension change is not significant but the local pressure has reached a dangerous threshold. This leads to missed detections or delayed warnings for high-risk points at the edge, failing to meet the clinical need for precise protection. Summary of the Invention
[0004] The purpose of this invention is to address the shortcomings of existing technologies that rely on statistical quantities such as regional average pressure for risk assessment. Due to the smoothing and dilution effect, these technologies struggle to identify the risk of amplified local linear loads caused by the collapse of the effective contact width at the edge of the device when surface contact degenerates into narrow band contact. This results in inaccurate location of high-risk points or untimely warnings. Therefore, this invention proposes a method and system for predicting and warning of pressure-related injuries associated with medical devices.
[0005] To address the problems existing in the prior art, the present invention adopts the following technical solution: A method for predicting and warning of medical device-related pressure injuries includes: S1. Collect pressure distribution data of the contact surface between the medical device and the skin, generate an interface pressure field, and determine the contact mask based on the interface pressure field. S2. Extract the set of contact edge points of the contact mask and determine the inner normal vector pointing to the inside of the contact area for each contact edge point in the set. S3. Perform pressure sampling along the internal normal vector of each contact edge point to construct a normal pressure profile, and calculate the effective contact width of the contact edge point based on the normal pressure profile; S4. Calculate the linear load amplification index at the contact edge points based on the effective contact width and normal pressure profile; S5. Perform statistical analysis on the linear load amplification index of the contact edge point set to generate an instantaneous risk score, and generate an early warning signal based on the instantaneous risk score.
[0006] Preferably, generating an interfacial pressure field and determining a contact mask based on the interfacial pressure field includes: The collected pressure distribution data is mapped to a grid coordinate system and interpolated to obtain the interface pressure field. Zero pressure is set as the contact criterion. When the pressure value of the grid point in the interface pressure field is greater than zero, the value at the corresponding position in the contact mask is set to one; otherwise, it is set to zero.
[0007] Preferably, the process involves extracting a set of contact edge points from the contact mask and determining the inward normal direction of each contact edge point within the contact area, including: Traverse the contact mask and identify grid points with a value of one and a neighboring point with a value of zero as contact edge points, thus forming a set of contact edge points; Perform a distance transformation on the contact mask to generate a distance field that records the distance from each point to the nearest zero point; Calculate the gradient vector of the distance field at each contact edge point, normalize the gradient vector to obtain the inward normal vector of each contact edge point.
[0008] Preferably, a normal pressure profile is constructed, and the effective contact width at each contact edge point is calculated based on the normal pressure profile, including: Within the set upper limit of sampling, sampling is performed along the inner normal vector in the interface pressure field to obtain the normal pressure profile; A half-peak threshold coefficient is set, and the length of the region in the statistical normal pressure profile where the pressure value is greater than or equal to the product of the starting point pressure value and the half-peak threshold coefficient is determined as the effective contact width.
[0009] Preferably, the linear load amplification index at the contact edge point is calculated based on the effective contact width and the normal pressure profile, including: The line load is obtained by integrating the normal pressure profile. Calculate the ratio of line load to effective contact width to obtain the equivalent edge pressure; The average pressure of the contact area is obtained by averaging the pressure values of the interface pressure field within the contact mask. The ratio of the equivalent edge pressure to the average pressure in the contact area is calculated to obtain the linear load amplification index.
[0010] Preferably, statistical analysis is performed on the linear load amplification index of the contact edge point set to generate an instantaneous risk score, including: Filter out the maximum line load amplification index in the set of contact edge points; Calculate the median and absolute median difference of all line load amplification indices in the set of contact edge points, calculate the difference between the maximum line load amplification index and the median, and divide the difference by the absolute median difference to obtain the instantaneous risk score.
[0011] Preferably, generating an early warning signal based on an instantaneous risk score includes: Calculate the difference between the instantaneous risk score at the current moment and the instantaneous risk score at the previous moment, and add the instantaneous risk score at the current moment to the difference to obtain the one-step forward-looking prediction risk score; Determine whether the instantaneous risk score or the one-step forward-looking prediction risk score exceeds a preset outlier threshold. If it does, generate an early warning signal containing the risk location point and the risk score. Among them, the risk location point is the contact edge point corresponding to the maximum line load amplification index, and the risk score includes the instantaneous risk score at the current moment and the one-step forward-looking prediction risk score.
[0012] To address the above problems, the present invention also provides a medical device-related pressure injury prediction and early warning system, the system comprising: The data acquisition module is used to collect pressure distribution data of the contact surface between the medical device and the skin, generate an interface pressure field, and determine the contact mask based on the interface pressure field. The edge analysis module is used to extract the set of contact edge points of the contact mask and determine the inner normal vector pointing to the inside of the contact area for each contact edge point in the set. The width calculation module is used to sample the pressure along the internal normal vector of each contact edge point to construct a normal pressure profile, and calculate the effective contact width of the contact edge point based on the normal pressure profile; The load assessment module is used to calculate the linear load amplification index at the contact edge points based on the effective contact width and the normal pressure profile. The risk warning module is used to perform statistical analysis on the linear load amplification index of the contact edge point set to generate an instantaneous risk score, and generate a warning signal based on the instantaneous risk score.
[0013] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention explicitly quantifies the geometrical change of the instrument edge from surface contact to near-line contact by constructing an edge-in-the-inner normal pressure profile and calculating the effective contact width. By using the line load amplification index to normalize and compare the local equivalent pressure at the edge with the average pressure of the overall contact area, it effectively overcomes the smoothing and dilution effect of the regional average pressure on the local pressure peak. Even when the overall fixing force does not change much, it can keenly capture the sharp amplification of local line load caused by the collapse of the effective contact width.
[0014] 2. This invention uses a robust statistical analysis method based on the median and absolute median difference to generate an instantaneous risk score. This method can effectively suppress the interference of individual noise points or transient contact jitter of the sensor on the early warning judgment, accurately measure the outlier degree of the maximum amplification point at the edge relative to the overall edge distribution, and combine it with a one-step forward prediction mechanism to use the differential trend of the risk score to predict the risk level at the next moment in advance. This method can issue an early warning signal at the initial stage of deterioration of the contact state, thereby significantly improving the foresight and timeliness of protection against pressure injuries related to medical devices. Attached Figure Description
[0015] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings: Figure 1 This is a flowchart illustrating a method for predicting and warning of medical device-related pressure injuries according to the present invention. Figure 2 This is a functional block diagram of a medical device-related pressure injury prediction and early warning system according to the present invention. Figure 3 This is a schematic diagram of the normal pressure profile and effective contact width of the present invention. Detailed Implementation
[0016] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.
[0017] Example: This example provides a method for predicting and warning of medical device-related pressure injuries. See [link to example]. Figure 1 Specifically, including: S1. Collect pressure distribution data of the contact surface between the medical device and the skin, generate an interface pressure field, and determine the contact mask based on the interface pressure field. In embodiments of the present invention, generating an interface pressure field and determining a contact mask based on the interface pressure field includes: Collect pressure distribution data at the contact surface between medical devices and skin; The collected pressure distribution data is mapped to a grid coordinate system and interpolated to obtain the interface pressure field. Zero pressure is set as the contact criterion. When the pressure value of the grid point in the interface pressure field is greater than zero, the value at the corresponding position in the contact mask is set to one; otherwise, it is set to zero. Specifically, collecting pressure distribution data at the interface between medical devices and skin refers to measuring the normal compressive force per unit area at different spatial locations on the actual contact interface between the device and the skin, forming a numerical set that varies with location. The pressure distribution data reflects the spatial non-uniformity of the contact load. The interface pressure field is a scalar field with spatial coordinates as the independent variable, used to characterize the local pressure at any point on the interface and directly correspond to the degree of pressure on the tissue at that point. Zero pressure indicates that the location does not bear the normal compressive load of the device on the skin and can be regarded as a state of no effective contact. When the grid point pressure value is greater than zero, it indicates that there is a positive pressing effect at that location and actual contact has been formed. The contact mask is a binary distribution map aligned with the interface pressure field on the same coordinates, used to retain only the spatial shape of the contact area. The corresponding position refers to the same spatial point of the mask and the pressure field under the same grid coordinates. A value of one indicates that the point is identified as a contact area and bears the compressive load, while a value of zero indicates that the point does not bear the compressive load or has broken contact, so that subsequent calculations can clearly distinguish the bearing area and the non-bearing area in space.
[0018] In detail, a flexible pressure sensor array covering the actual contact surface between the medical device and the skin was selected as the acquisition component. Before acquisition, the sensor array underwent zero-point calibration and range calibration. Zero-point calibration was performed by continuously acquiring no less than fifty frames of raw output under no external load conditions and taking the average value of each sensor unit as the zero-point offset to compensate for static drift. Range calibration was performed by applying a known standard pressure load to a flat, rigid reference surface and establishing a linear conversion coefficient from the raw output to the pressure value to ensure comparability of the pressure values. Subsequently, the sensor array was attached to the contact interface between the medical device and the skin or to the internal components of the device. The device is fixed in one piece, ensuring that the effective measurement surface of the array completely covers the expected pressure area. After the device is worn stably according to the clinical fixation method, pressure distribution data is continuously collected at a sampling frequency of not less than 10 Hz. The pressure distribution data is a set of pressure values output by multiple sensing units at the same time. During the acquisition process, zero-point offset and range conversion are performed on each frame of pressure distribution data to obtain the physical pressure value of each sensing unit. Isolated abnormal points appearing for more than three consecutive frames are replaced by the neighborhood median to suppress spike noise caused by contact transient jitter, thereby obtaining a stable pressure distribution data sequence.
[0019] When mapping the collected pressure distribution data to the grid coordinate system, the correspondence between the physical location of the sensing unit and the grid coordinate is established in advance, and the grid resolution is defined. The grid resolution is preferably one to two millimeters to balance the ability to express the edge pressure gradient and the computational cost. The pressure value of each sensing unit is written into its corresponding grid point. If multiple sensing units are mapped to the same grid point, their average value is taken as the pressure value of that grid point. If there are grid points that are not directly assigned a value, interpolation is performed to form the interface pressure field. The interpolation is preferably bilinear interpolation or inverse distance weighted interpolation. Bilinear interpolation is suitable for sensing units that are regularly arranged and can maintain the local continuity of the pressure field. Inverse distance weighted interpolation is suitable for areas with many missing measurements and can reflect the objective distribution of the contact load in space through the distance attenuation law, thereby obtaining the interface pressure field with the grid coordinate as the independent variable.
[0020] To avoid misjudgments caused by sensor zero drift and minor noise when setting zero pressure as the contact criterion, it is preferable to set the contact criterion to a positive pressure threshold of not less than 5 mmHg or not less than 0.7 kPa. This value is based on the fact that the zero-point noise amplitude of common flexible pressure sensors in clinical wearing environments is usually on the order of several mmHg. Using a threshold higher than the noise amplitude can significantly reduce the probability of non-contact points being misjudged as contact points. When the grid point pressure value in the interface pressure field is greater than the contact criterion, the value at the corresponding position in the contact mask is set to one; otherwise, it is set to zero. The contact mask and the interface pressure field use the same grid coordinates and resolution to ensure spatial alignment, thereby forming a contact mask that can clearly distinguish between the load-bearing area and the non-load-bearing area and providing reliable input for subsequent edge extraction and line load amplification analysis.
[0021] S2. Extract the set of contact edge points of the contact mask and determine the inner normal vector pointing to the inside of the contact area for each contact edge point in the set. In an embodiment of the present invention, extracting the set of contact edge points of the contact mask and determining the inward normal vector pointing to the interior of the contact region for each contact edge point in the set includes: Traverse the contact mask and identify grid points with a value of one and a neighboring point with a value of zero as contact edge points, thus forming a set of contact edge points; Perform a distance transformation on the contact mask to generate a distance field that records the distance from each point to the nearest zero point; Calculate the gradient vector of the distance field at each contact edge point, normalize the gradient vector, and obtain the inward normal vector of each contact edge point. Specifically, the contact edge point set refers to a set of grid points determined to be located on the outer contour of the contact area within the contact mask. Its members satisfy the condition that the mask value at that grid point is one and that there is at least one grid point in its neighborhood with a value of zero. Therefore, this set spatially depicts the transition boundary between the actual contact area between the medical device and the skin, from contact to non-contact, and is used for subsequent line load amplification analysis only near the boundary to avoid invalid calculations of the internal uniformly loaded area. The distance field refers to a continuous or discrete scalar distribution defined in the same grid coordinate system as the contact mask, where the value of each grid point represents the distance from that point to the nearest... The shortest geometric distance of non-contact grid points indicates that the point is located further inside the contact area and farther from the boundary, thus providing a stable geometric reference for direction determination at the boundary. The inner normal vector of the contact edge point is a unit direction vector defined at each contact edge point, pointing towards the inside of the contact area. Its direction is determined by the direction of increasing distance of the distance field at that edge point and obtained by normalization. It reflects the shortest geometric direction away from the non-contact area when advancing from the boundary into the contact area. It is used to extract the normal pressure profile along this direction to quantify the effective contact width and linear load amplification of the edge bearing zone.
[0022] In detail, after obtaining the contact mask aligned with the interface pressure field, the contact mask is regarded as a binary grid array whose spatial position is determined by row and column indices. All grid points are traversed in row-first or column-first order. For each grid point, its mask value is read and its neighboring point values are checked to determine whether it belongs to the contact edge point. The neighborhood range preferably adopts eight neighborhoods to avoid the oblique boundary being missed. Specifically, when the value of the current grid point is one and there is at least one grid point with a value of zero in its eight neighborhoods, the current grid point is marked as a contact edge point and written into the contact edge point set. After the traversal is completed, a complete contact edge point set is obtained to characterize the outer contour of the contact area.
[0023] To construct a stable direction pointing inwards from the contact area at each contact edge point, a distance transformation is performed on the contact mask to generate a distance field. The input of the distance transformation is the region with a value of one in the contact mask, and the reference set is the region with a value of zero. The distance is preferably defined using Euclidean distance to match the actual spatial geometry. The distance value at any point in the distance field represents the shortest Euclidean distance from that point to the nearest zero point, thus forming a continuous scalar distribution that gradually increases from the boundary inwards within the contact area. Subsequently, the gradient vector of the distance field is calculated at each contact edge point to obtain directional information pointing in the direction of increasing distance. The gradient vector is preferably calculated using a central difference discretization method, specifically by calculating the gradient vector at the row direction and the distance at the edge direction. The gradient component is obtained by taking the distance difference between two adjacent grid points in the column direction and dividing it by twice the grid spacing. If the contact edge point is located at the mask boundary and the complete neighborhood cannot be obtained, forward difference or backward difference is used to ensure computability. The resulting gradient vector naturally points to the inside of the contact area because the distance field monotonically increases along the direction away from the zero boundary. In order to eliminate the scale effect caused by the gradient magnitude changing with the local shape and retain only the direction information, the gradient vector is normalized. It is preferable to divide each component of the gradient vector by its magnitude to obtain the inner normal vector of each contact edge point. The inner normal vector corresponds one-to-one with the contact edge point and is used for subsequent pressure sampling along the inner normal to construct the normal pressure profile.
[0024] S3. Perform pressure sampling along the internal normal vector of each contact edge point to construct a normal pressure profile, and calculate the effective contact width of the contact edge point based on the normal pressure profile; In embodiments of the present invention, a normal pressure profile is constructed, and the effective contact width at the contact edge points is calculated based on the normal pressure profile, including: Within the set upper limit of sampling, sampling is performed along the inner normal vector in the interface pressure field to obtain the normal pressure profile; Set a half-peak threshold coefficient, and determine the length of the region in the statistical normal pressure profile where the pressure value is greater than or equal to the product of the starting point pressure value and the half-peak threshold coefficient as the effective contact width. Specifically, the normal pressure profile refers to the sequence of pressure changes as it advances inward, obtained by sampling gradually along the contact area from the contact edge point and with the inward normal vector as the direction. This sequence reflects the attenuation or fluctuation of the bearing pressure as it extends inward from the contact boundary, and is used to characterize whether there is a high-pressure narrow band near the edge when the surface contact degenerates into an approximate line contact. The effective contact width refers to the length of the continuous segment in the normal pressure profile where the pressure value is not lower than the threshold and is connected to the starting point, starting from the pressure value at the beginning of the normal pressure profile at the same contact edge point. This width is equivalent to the actual width of the edge bearing band that can bear the main compaction load. When the edge bearing degenerates from surface bearing to narrow band bearing, this width will decrease significantly, thus providing a basis for subsequently dividing the line load by the effective contact width to obtain the equivalent edge pressure and constructing the line load amplification index.
[0025] In detail, after obtaining the interface pressure field and the inward normal vector corresponding to each contact edge point, a normal pressure profile is constructed in the interface pressure field using the grid coordinates of each contact edge point as the sampling starting point and the inward normal vector as the sampling direction. Specifically, a sampling step size is set and the pressure value is read point by point along the inward normal direction. The sampling step size is preferably the grid resolution of the interface pressure field to ensure that adjacent sampling points correspond to adjacent actual spatial positions without introducing additional spatial aliasing errors. At the same time, a sampling upper limit is set to limit the maximum inward depth of the profile coverage. The sampling upper limit is preferably determined by the inward available distance of the distance field at the contact edge point, that is, the distance value of the contact edge point in the distance field is taken as the sampling upper limit to ensure that the sampling does not exceed the limit. The pressure information of the internal region, which is independent of the edge bearing capacity, is introduced by the geometric center of the contact area. When advancing along the inner normal direction, the sampling point at each advancing distance is mapped to the grid coordinate position of the interface pressure field through coordinate transformation and the corresponding pressure value is read. When the sampling point falls between two grid points, the pressure value is read by the interpolation method consistent with the interface pressure field to maintain the continuity of the profile. This forms a pressure sequence arranged according to the advancing distance and defines the pressure sequence as the normal pressure profile. The pressure value at the starting point of the profile is taken as the pressure value at the sampling starting point and recorded as the starting point pressure value. In order to avoid threshold drift caused by noise at the starting point, it is preferable to take the arithmetic mean of the starting point and a small number of adjacent sampling points along the normal direction as the starting point pressure value, so that it can better represent the true peak level of the edge bearing zone.
[0026] After obtaining the normal pressure profile, a half-peak threshold coefficient is set to characterize the effective width of the edge bearing zone. The half-peak threshold coefficient is preferably set to 0.5, based on the fact that when the edge bearing gradually decreases from high pressure to the inward side, the cutoff corresponding to the half-peak can stably reflect the width of the main bearing zone and has good robustness to changes in the overall pressure scale. Then, the half-peak threshold is calculated as the product of the starting point pressure value and the half-peak threshold coefficient. In the normal pressure profile, the pressure value is judged point by point to see if it is greater than or equal to the half-peak threshold. The continuous sampling segment that meets the condition is regarded as the effective bearing segment, and the length of the effective bearing segment is counted. The length is obtained by multiplying the number of sampling points that meet the condition by the sampling step size. If there are multiple discrete segments that meet the condition in the profile, it is preferable to take only the segment containing the starting point as the effective bearing segment to ensure that the effective contact width reflects the bearing zone near the edge without being disturbed by the internal secondary peak. Finally, the length of the effective bearing segment is determined as the effective contact width of the contact edge point and output for subsequent calculation of the line load amplification index.
[0027] S4. Calculate the linear load amplification index at the contact edge points based on the effective contact width and normal pressure profile; In an embodiment of the present invention, the calculation of the linear load amplification index at the contact edge point based on the effective contact width and the normal pressure profile includes: The line load is obtained by integrating the normal pressure profile. Calculate the ratio of line load to effective contact width to obtain the equivalent edge pressure; Specifically, line load refers to the load-bearing capacity per unit edge length obtained by accumulating the pressure value along the inward depth range covered by the normal pressure profile at a certain contact edge point. It is formed by numerical integration of the pressure sequence of the normal pressure profile with the advancement distance, reflecting the comprehensive result of all the pressure contributions in the inward direction when extending from the boundary to the inside near the edge. Therefore, it can be used to characterize the overall load-bearing level at the edge point. Equivalent edge pressure refers to the representative pressure level obtained by converting the line load according to the effective contact width. It is calculated by the ratio of line load to effective contact width, indicating the equivalent pressure that the main load-bearing band needs to bear within the actual effective width. When the effective contact width narrows while the line load remains at a similar level, the equivalent edge pressure will increase significantly, thus directly characterizing the pressure amplification effect generated when the edge bearing degenerates from surface contact to narrow band bearing. Narrow band bearing refers to a bearing pattern where, when a medical device comes into contact with the skin, the main force is no longer distributed across a wider contact surface, but is concentrated on a very narrow, continuous bearing band formed along the edge of the device or a local hard edge to bear the compressive load. Spatially, this bearing band is characterized by a high-pressure area near the contact edge that is significantly narrower than the typical width of the contact area and extends continuously along the edge. Mechanically, this is equivalent to the resultant force per unit edge length being compressed into a smaller effective contact width for transmission. This results in the same fixing force or similar overall linear load corresponding to a higher local pressure level, making it easier to form indentations in superficial skin tissues, restrict local perfusion, and increase the risk of tissue damage. This constitutes a common high-risk stress state in medical device-related pressure injuries.
[0028] In detail, after obtaining the normal pressure profile and effective contact width corresponding to each contact edge point, the normal pressure profile is regarded as the pressure distribution along the inner normal advance distance and used to estimate the load per unit length near the edge. Specifically, the grid resolution of the interface pressure field is used as the sampling step size and denoted as the sampling step size. The normal pressure profile is discretized into a set of pressure sample values within the advance range from the starting point to the upper limit of the sampling according to the sampling step size. Then, the set of pressure sample values is numerically integrated to obtain the line load. Trapezoidal integral is preferably used to balance accuracy and stability. The implementation method is to take the arithmetic mean of the pressure sample values of two adjacent points, multiply it by the sampling step size, and apply the result across the entire profile. The values are accumulated to obtain the line load value representing the cumulative effect of the bearing pressure along the normal direction at the edge point. The physical meaning of the line load is the resultant force intensity per unit edge length obtained by integrating the pressure along the inward depth direction, which is used to characterize the overall bearing level near the edge. Subsequently, in order to link the line load with the effective bearing bandwidth, the line load is divided by the effective contact width to obtain the equivalent edge pressure. The effective contact width is used to characterize the actual width of the main bearing band. The ratio of the line load to the effective contact width can convert the resultant force intensity per unit edge length into the equivalent pressure level in the main bearing band, so that the equivalent edge pressure increases accordingly when the effective contact width collapses.
[0029] The average pressure of the contact area is obtained by averaging the pressure values of the interface pressure field within the contact mask. The ratio of the equivalent edge pressure to the average pressure in the contact area is calculated to obtain the line load amplification index. It should be noted that the line load amplification index is a dimensionless ratio index used to quantify the narrow band load-bearing capacity at the contact edge. It is obtained by dividing the equivalent edge pressure at a certain contact edge point by the average pressure of the contact area within the contact mask at the same time. The equivalent edge pressure represents the representative pressure level that the main load-bearing band at that edge point needs to bear within the actual effective contact width, while the average pressure of the contact area represents the average load-bearing scale of the device and skin within the overall contact area. The ratio of the two can directly characterize the amplification factor of the edge load-bearing pressure relative to the overall load-bearing scale. When the edge load degenerates from a wider surface load-bearing capacity to a narrow band load-bearing capacity extending along the edge, the effective contact width decreases significantly while the line load remains at a similar level. This results in a significant increase in the equivalent edge pressure relative to the average pressure, causing the line load amplification index to increase significantly and reflecting the high-risk edge morphology of pressure-related injuries associated with medical devices.
[0030] In detail, after obtaining the equivalent edge pressure of each contact edge point and the interface pressure field and contact mask aligned with it in space, in order to form a normalized index that can be compared across different wearing forces and different device scenarios, the average pressure of the contact area is first calculated within the load-bearing area defined by the contact mask. Specifically, all grid points of the interface pressure field are traversed and the values of the contact mask at the corresponding positions are read simultaneously. Only grid points with a contact mask value of one are included in the statistics. The pressure values of these grid points are summed to obtain the total contact pressure, and the number of contact points with a contact mask value of one is counted. Then, the total contact pressure is divided by the total number of contact points to obtain the average pressure of the contact area. This average pressure reflects the average load level of the device and skin in the overall contact area and can characterize the overall load scale under the current wearing state. To avoid numerical instability caused by the total number of contact points being zero or extremely small, it is preferable to add an extremely small positive number to the denominator and directly determine that there is no contact when the total number of contact points is zero, thus skipping subsequent calculations.
[0031] After obtaining the average pressure of the contact area, the equivalent edge pressure of each contact edge point is divided by the average pressure of the contact area to obtain the line load amplification index. This index is used to characterize the degree of amplification of the pressure of the main load-bearing band at the edge relative to the overall load-bearing scale. When the effective contact width collapses at the contact edge, resulting in narrow band load, the equivalent edge pressure will increase significantly relative to the average pressure of the contact area, thereby increasing the line load amplification index and using it for subsequent statistical outlier analysis and early warning judgment.
[0032] S5. Perform statistical analysis on the linear load amplification index of the contact edge point set to generate an instantaneous risk score, and generate an early warning signal based on the instantaneous risk score; In an embodiment of the present invention, generating an instantaneous risk score and generating an early warning signal based on the instantaneous risk score includes: Filter out the maximum line load amplification index in the set of contact edge points; Calculate the median and absolute median difference of all line load amplification indices in the set of contact edge points, calculate the difference between the maximum line load amplification index and the median, and divide the difference by the absolute median difference to obtain the instantaneous risk score; Specifically, the instantaneous risk score is a single scalar obtained by robustly measuring the linear load amplification index of the contact edge point set at the same moment. It is used to characterize the degree of anomalousness of the strongest narrow band load-bearing point at the edge relative to the overall load-bearing state of the edge. It is obtained by subtracting the median of the linear load amplification index of all edge points from the maximum linear load amplification index, and then dividing the outlier by the absolute median difference of the linear load amplification index of all edge points. Thus, it simultaneously considers the prominence of the strongest amplification point at the edge and the dispersion level of the overall edge distribution. When the effective contact width collapses at the contact edge and forms a local narrow band load with significantly amplified linear load, the maximum linear load amplification index will be significantly higher than the typical level of the edge as a whole and the deviation will exceed the typical fluctuation range of the edge distribution, thus increasing the instantaneous risk score and serving as an immediate early warning criterion for medical device-related pressure injuries.
[0033] In detail, after obtaining the linear load amplification index corresponding to each contact edge point in the set of contact edge points at the same time, the linear load amplification index is regarded as a spatial distribution quantity characterizing the strength of the narrow band load at the edge. First, the set of contact edge points is traversed once, and the linear load amplification indices of each contact edge point are compared. The linear load amplification index with the largest value is selected and recorded to represent the amplification degree of the most significant narrow band load point at the current edge. Subsequently, to avoid interference from a single abnormal noise point or a small amount of sensing error on the overall judgment, robust statistical analysis is performed on all linear load amplification indices in the set of contact edge points. Specifically, all linear load amplification indices are sorted from smallest to largest, and the median value after sorting is taken as the median. When the number of edge points... When the number of deviations is even, the arithmetic mean of the two middle values is taken as the median, so that the median can represent the typical amplification level of the edge as a whole. Then, the absolute value of the difference between each line load amplification index and the median is taken to form a deviation set. The deviation set is sorted from smallest to largest and the median value is taken as the absolute median difference. When the number of deviations is even, the arithmetic mean of the two middle values is also taken as the absolute median difference, so that the absolute median difference can represent the typical dispersion of the edge amplification index. Then, the difference between the maximum line load amplification index and the median is calculated as the maximum outlier amplitude. The maximum outlier amplitude is then divided by the absolute median difference to obtain the instantaneous risk score, where the instantaneous risk score represents the outlier intensity of the strongest narrowband bearing point of the edge relative to the overall amplification level of the edge.
[0034] Calculate the difference between the instantaneous risk score at the current moment and the instantaneous risk score at the previous moment, and add the instantaneous risk score at the current moment to the difference to obtain the one-step forward-looking prediction risk score; Determine whether the instantaneous risk score or the one-step forward-looking prediction risk score exceeds a preset outlier threshold. If it does, generate an early warning signal containing the risk location point and the risk score. Specifically, the one-step forward-looking predictive risk score refers to the predictive quantity obtained by extrapolating the risk level of the next sampling moment based on the instantaneous risk score obtained at the current moment and the changing trend of the instantaneous risk score between adjacent sampling moments. The calculation method is to first calculate the difference between the instantaneous risk score at the current moment and the instantaneous risk score at the previous moment to represent the increase or decrease of the risk score within a sampling interval, and then add the difference to the instantaneous risk score at the current moment to obtain the one-step forward-looking predictive risk score. Therefore, this predictive quantity can reflect the outlier intensity that may occur at the next moment in advance when the risk score continues to rise but has not yet reached the trigger threshold, so as to identify earlier the risk of further aggravation of narrow band load at the contact edge and the potential for medical device-related pressure injury.
[0035] In detail, after continuously acquiring the pressure field at the interface and generating an instantaneous risk score at each sampling moment using the same operation, the instantaneous risk score at the current moment is recorded as the current instantaneous risk score, and the instantaneous risk score at the previous sampling moment is recorded as the previous instantaneous risk score. Then, the difference between the two is calculated to characterize the change in risk score within adjacent sampling intervals. The difference is obtained by subtracting the previous instantaneous risk score from the current instantaneous risk score. The current instantaneous risk score is then added to the difference to obtain a one-step forward-looking predicted risk score. This predicted risk score is used to anticipate the risk level that may be reached at the next sampling moment when the risk score shows an upward trend, thereby improving the foresight of the early warning.
[0036] After calculating the risk score and the predicted risk score, an outlier threshold is set as the criterion for triggering an early warning. The outlier threshold is preferably three, based on the fact that the instantaneous risk score adopts a robust standardized form of median and absolute median difference. When the score reaches three, it indicates that the deviation of the maximum line load amplification point from the typical level of the edge has exceeded the typical fluctuation range of the edge distribution and belongs to a significant outlier state. Then, it is determined whether the instantaneous risk score at the current moment or the one-step forward-looking predicted risk score exceeds the outlier threshold. If either exceeds, an early warning signal is generated. The early warning signal includes at least a risk location point and a risk score. The risk location point is the contact edge point corresponding to the maximum line load amplification index to clearly indicate the instrument edge position that needs to be checked. The risk score preferably includes both the instantaneous risk score at the current moment and the one-step forward-looking predicted risk score to distinguish between immediate risk and trend risk, thereby realizing timely warning and traceable recording of the line load amplification risk caused by the collapse of the effective contact width of the edge.
[0037] like Figure 2 The diagram shown is a functional block diagram of a medical device-related pressure injury prediction and early warning system provided in an embodiment of the present invention.
[0038] In this embodiment, the functions of each module / unit are as follows: The data acquisition module is used to collect pressure distribution data of the contact surface between the medical device and the skin, generate an interface pressure field, and determine the contact mask based on the interface pressure field. The edge analysis module is used to extract the set of contact edge points of the contact mask and determine the inner normal vector pointing to the inside of the contact area for each contact edge point in the set. The width calculation module is used to sample the pressure along the internal normal vector of each contact edge point to construct a normal pressure profile, and calculate the effective contact width of the contact edge point based on the normal pressure profile; The load assessment module is used to calculate the linear load amplification index at the contact edge points based on the effective contact width and the normal pressure profile. The risk warning module is used to perform statistical analysis on the linear load amplification index of the contact edge point set to generate an instantaneous risk score, and generate a warning signal based on the instantaneous risk score.
[0039] like Figure 3 As shown, the horizontal axis represents the inward normal advance distance along the contact edge point, and the vertical axis represents the pressure value. The blue curve represents the normal pressure profile obtained by sampling the interface pressure field point by point along the inward normal direction. It obtains the initial pressure at the starting point and gradually decreases as it advances inward. The gray dashed line represents the half-peak threshold line obtained by multiplying the initial pressure by the half-peak threshold coefficient. By comparing the positional relationship between the normal pressure profile and the half-peak threshold line, the length of the continuous segment from the starting point where the normal pressure profile is still not lower than the half-peak threshold can be determined. This continuous segment length is marked by a red double-headed arrow and is determined as the effective contact width, which is used to characterize the actual width of the main load-bearing zone near the contact edge. The light-colored filled area in the figure is the integral area of the normal pressure profile within the advance distance range, which is used to represent the line load at the contact edge point. This provides an intuitive basis for calculating the equivalent edge pressure based on the line load and the effective contact width and further constructing the line load amplification index.
[0040] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.
Claims
1. A method for predicting and warning of medical device-related pressure injuries, characterized in that, include: S1. Collect pressure distribution data of the contact surface between the medical device and the skin, generate an interface pressure field, and determine the contact mask based on the interface pressure field. S2. Extract the set of contact edge points of the contact mask and determine the inner normal vector pointing to the inside of the contact area for each contact edge point in the set. S3. Perform pressure sampling along the internal normal vector of each contact edge point to construct a normal pressure profile, and calculate the effective contact width of the contact edge point based on the normal pressure profile; S4. Calculate the linear load amplification index at the contact edge points based on the effective contact width and normal pressure profile; S5. Perform statistical analysis on the linear load amplification index of the contact edge point set to generate an instantaneous risk score, and generate an early warning signal based on the instantaneous risk score.
2. The method for predicting and warning of medical device-related pressure injuries according to claim 1, characterized in that, Generating an interfacial pressure field and determining a contact mask based on the interfacial pressure field includes: The collected pressure distribution data is mapped to a grid coordinate system and interpolated to obtain the interface pressure field. Zero pressure is set as the contact criterion. When the pressure value of the grid point in the interface pressure field is greater than zero, the value at the corresponding position in the contact mask is set to one; otherwise, it is set to zero.
3. The method for predicting and warning of medical device-related pressure injuries according to claim 1, characterized in that, Extract the set of contact edge points of the contact mask, and determine the inward normal direction pointing to the interior of the contact area for each contact edge point in the set, including: Traverse the contact mask and identify grid points with a value of one and a neighboring point with a value of zero as contact edge points, thus forming a set of contact edge points; Perform a distance transformation on the contact mask to generate a distance field that records the distance from each point to the nearest zero point; Calculate the gradient vector of the distance field at each contact edge point, normalize the gradient vector to obtain the inward normal vector of each contact edge point.
4. The method for predicting and warning of medical device-related pressure injuries according to claim 1, characterized in that, Construct a normal pressure profile, and calculate the effective contact width at each contact edge point based on the normal pressure profile, including: Within the set upper limit of sampling, sampling is performed along the inner normal vector in the interface pressure field to obtain the normal pressure profile; A half-peak threshold coefficient is set, and the length of the region in the statistical normal pressure profile where the pressure value is greater than or equal to the product of the starting point pressure value and the half-peak threshold coefficient is determined as the effective contact width.
5. The method for predicting and warning of medical device-related pressure injuries according to claim 1, characterized in that, The linear load amplification index at the contact edge points is calculated based on the effective contact width and the normal pressure profile, including: The line load is obtained by integrating the normal pressure profile. Calculate the ratio of line load to effective contact width to obtain the equivalent edge pressure; The average pressure of the contact area is obtained by averaging the pressure values of the interface pressure field within the contact mask. The ratio of the equivalent edge pressure to the average pressure in the contact area is calculated to obtain the linear load amplification index.
6. The method for predicting and warning of medical device-related pressure injuries according to claim 1, characterized in that, Statistical analysis of the line load amplification index of the contact edge point set is performed to generate an instantaneous risk score, including: Filter out the maximum line load amplification index in the set of contact edge points; Calculate the median and absolute median difference of all line load amplification indices in the set of contact edge points, calculate the difference between the maximum line load amplification index and the median, and divide the difference by the absolute median difference to obtain the instantaneous risk score.
7. The method for predicting and warning of medical device-related pressure injuries according to claim 6, characterized in that, Early warning signals are generated based on instantaneous risk scores, including: Calculate the difference between the instantaneous risk score at the current moment and the instantaneous risk score at the previous moment, and add the instantaneous risk score at the current moment to the difference to obtain the one-step forward-looking prediction risk score; Determine whether the instantaneous risk score or the one-step forward-looking prediction risk score exceeds a preset outlier threshold. If it does, generate an early warning signal containing the risk location point and the risk score. Among them, the risk location point is the contact edge point corresponding to the maximum line load amplification index, and the risk score includes the instantaneous risk score at the current moment and the one-step forward-looking prediction risk score.
8. A medical device-related pressure injury prediction and early warning system, characterized in that, The system includes: The data acquisition module is used to collect pressure distribution data of the contact surface between the medical device and the skin, generate an interface pressure field, and determine the contact mask based on the interface pressure field. The edge analysis module is used to extract the set of contact edge points of the contact mask and determine the inner normal vector pointing to the inside of the contact area for each contact edge point in the set. The width calculation module is used to sample the pressure along the internal normal vector of each contact edge point to construct a normal pressure profile, and calculate the effective contact width of the contact edge point based on the normal pressure profile; The load assessment module is used to calculate the linear load amplification index at the contact edge points based on the effective contact width and the normal pressure profile. The risk warning module is used to perform statistical analysis on the linear load amplification index of the contact edge point set to generate an instantaneous risk score, and generate a warning signal based on the instantaneous risk score.