Peritoneal dialysis catheter risk management method and related equipment

Abstract

The invention discloses a peritoneal dialysis catheter risk management method and related equipment. The method comprises the following steps: acquiring skin temperature information at a catheter outlet reported by a temperature patch worn by a patient, acquiring image data obtained by photographing through a mobile terminal, and acquiring discomfort information of the patient to form a local risk index of the catheter; performing multi-dimensional fusion calculation on the basis of the acquired environment index in the home peritoneal dialysis scene of the patient and the local risk index of the catheter to obtain an environment-infection risk index; and carrying out continuous trend analysis, determining a risk level according to a threshold exceeding degree and a trend change rate when detecting that the environment-infection risk index continuously rises in a preset time window and exceeds a preset safety threshold, and generating corresponding alarm information based on the risk level. The problem that a conventional infection prevention measure mainly depends on subjective recording and visual inspection of a patient, and catheter infection is often found out can be solved.

Description

Technical Field

[0001] This application relates to the field of smart healthcare, and more specifically, to a method and related equipment for risk management of peritoneal dialysis catheters. Background Technology

[0002] With the evolution of treatment models for end-stage renal disease, home peritoneal dialysis has become an important form of renal replacement therapy due to its ability to improve patients' autonomy and quality of life. However, the shift of the treatment location from a highly controlled medical environment to the home introduces significant infection risks. The patient's peritoneal cavity is indirectly connected to the external environment through an indwelling peritoneal dialysis catheter; therefore, the microbial level of the surrounding environment during dialysis fluid exchange directly determines the likelihood of pathogens invading the peritoneal cavity.

[0003] Currently, infection control in home peritoneal dialysis environments relies entirely on patients' self-monitoring. There is a lack of objective and continuous monitoring methods for air quality and surface cleanliness in the operating space. Routine infection prevention measures mainly depend on patients' subjective records and visual observation, often only being discovered when catheter infections occur. Summary of the Invention

[0004] The summary section introduces a series of simplified concepts, which will be further explained in detail in the detailed description section. The summary section of this invention is not intended to limit the key features and essential technical features of the claimed technical solution, nor is it intended to determine the scope of protection of the claimed technical solution.

[0005] To address the issue that conventional infection prevention measures rely primarily on patient records and visual observation, often only being detected when catheter infections occur, this invention proposes, firstly, a peritoneal dialysis catheter risk management method, comprising: The system acquires skin temperature information at the catheter exit site reported by the temperature patch worn by the patient, as well as image data of skin color, local exudate color and turbidity around the catheter exit obtained by taking photos through a mobile terminal, and also acquires information on local pain or discomfort manually entered by the patient, in order to form local risk indicators for the catheter. Based on the environmental indicators obtained in the patient's home peritoneal dialysis scenario and the catheter local risk indicators, a multidimensional fusion calculation is performed to obtain the environment-infection risk index. The multidimensional fusion calculation includes weighting the concentration of air particulate matter, humidity, temperature, local temperature change range, skin discoloration degree and exudate turbidity change according to preset weights. Continuous trend analysis is performed on the environment-infection risk index. If the environment-infection risk index is detected to be continuously rising within a preset time window and exceeding a preset safety threshold, the risk level is determined based on the degree of exceeding the threshold and the rate of trend change, and corresponding alarm information is generated based on the risk level.

[0006] Secondly, the present invention also provides a peritoneal dialysis catheter risk management device, comprising: The acquisition unit is used to acquire skin temperature information at the catheter exit point reported by the temperature patch worn by the patient, as well as image data of skin color, local exudate color and turbidity around the catheter exit point obtained by taking pictures through the mobile terminal, and to acquire local pain or discomfort information manually entered by the patient, so as to form local risk indicators for the catheter. The calculation unit is used to perform multidimensional fusion calculation based on the environmental indicators and local risk indicators of the catheter in the patient's home peritoneal dialysis scenario to obtain the environment-infection risk index. The multidimensional fusion calculation includes weighting the concentration of air particulate matter, humidity, temperature, local temperature change range, skin discoloration degree and exudate turbidity change according to preset weights. The determination unit is used to perform continuous trend analysis on the environment-infection risk index. When the environment-infection risk index is detected to be continuously rising within a preset time window and exceeding a preset safety threshold, the risk level is determined according to the degree of exceeding the threshold and the rate of trend change, and corresponding alarm information is generated based on the risk level.

[0007] Thirdly, an electronic device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program stored in the memory to implement the steps of the peritoneal dialysis catheter risk management method as described in any of the first aspects above.

[0008] Fourthly, the present invention also proposes a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the peritoneal dialysis catheter risk management method of any of the above claims in the first aspect.

[0009] In summary, the peritoneal dialysis catheter risk management method proposed in this application obtains skin temperature information at the catheter exit point reported by the patient's temperature patch, as well as image data of skin color, local exudate color, and turbidity around the catheter exit point obtained through mobile terminal photography, and obtains information on local pain or discomfort manually input by the patient to form a catheter local risk index. Based on the obtained environmental indicators in the patient's home peritoneal dialysis scenario and the catheter local risk index, a multi-dimensional fusion calculation is performed to obtain an environment-infection risk index. The multi-dimensional fusion calculation includes weighting air particulate matter concentration, humidity, temperature, local temperature change amplitude, skin discoloration degree, and exudate turbidity change according to preset weights. Continuous trend analysis is performed on the environment-infection risk index. If the environment-infection risk index continuously increases within a preset time window and exceeds a preset safety threshold, the risk level is determined based on the degree of exceeding the threshold and the rate of trend change, and a corresponding alarm message is generated based on the risk level. This transforms the peritoneal dialysis catheter infection risk management, which previously relied on the patient's subjective feelings and occasional visits, into a continuous, quantitative, and graded intelligent management process. On the one hand, temperature patches and image analysis capture subtle local inflammatory signals in advance and record them in reproducible numerical form. On the other hand, the comprehensive calculation of environmental and local indicators enables the system to identify undamaged areas in poor environments, problematic areas in good environments, and different scenarios where both the environment and the local area deteriorate simultaneously, and provides corresponding estimates. Furthermore, continuous trend analysis effectively filters out occasional interference signals, focusing the system's attention on truly clinically significant, continuously rising risks. Combined with tiered alarm content, it provides patients with a tiered intervention path, from simple cleaning suggestions to remote follow-up and emergency medical arrangements. This significantly reduces the incidence of catheter exit site infection and peritonitis in home peritoneal dialysis patients, reduces hospitalizations and dialysis mode changes due to infection, improves patients' safety and confidence in dialysis treatment at home, and also provides healthcare professionals with a data-supported remote risk monitoring tool, which is beneficial for optimizing follow-up frequency and resource allocation. Attached Figure Description

[0010] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit this specification. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A schematic diagram of a peritoneal dialysis catheter risk management method provided in this application embodiment; Figure 2 A schematic diagram of a peritoneal dialysis catheter risk management device provided in this application embodiment; Figure 3 This is a schematic diagram of an electronic device for risk management of peritoneal dialysis catheters provided in an embodiment of this application. Detailed Implementation

[0011] The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus. The technical solutions of the embodiments of this application will now be clearly and completely described in conjunction with the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them.

[0012] To address the issue that routine infection prevention measures rely primarily on patient records and visual observation, often only being detected when catheter-related infections occur, please refer to [link to relevant documentation]. Figure 1 This is a schematic flowchart of a peritoneal dialysis catheter risk management method provided in an embodiment of this application, which may specifically include steps S110 to S130.

[0013] S110 acquires skin temperature information at the catheter exit point reported by the temperature patch worn by the patient, as well as image data of skin color, local exudate color and turbidity around the catheter exit point obtained by taking pictures through the mobile terminal, and acquires local pain or discomfort information manually entered by the patient to form local catheter risk indicators.

[0014] S120, Based on the environmental indicators obtained in the patient's home peritoneal dialysis scenario and the local risk indicators of the catheter, a multidimensional fusion calculation is performed to obtain the environment-infection risk index. The multidimensional fusion calculation includes weighting the concentration of air particulate matter, humidity, temperature, local temperature change range, skin discoloration degree and exudate turbidity change according to preset weights.

[0015] S130, perform continuous trend analysis on the environment-infection risk index. If the environment-infection risk index is detected to be continuously rising within a preset time window and exceeding a preset safety threshold, determine the risk level based on the degree of exceeding the threshold and the rate of trend change, and generate corresponding alarm information based on the risk level.

[0016] Understandably, peritoneal dialysis catheter exit site infection is essentially a local inflammatory response and microbial invasion. One of the earliest and most stable signals of inflammation is increased skin temperature due to increased local microcirculation, accompanied by capillary dilation and tissue edema, which manifests grossly as skin redness, darkening, and even linear red streaks. As the local infection progresses, the exudate around the catheter gradually changes from clear to turbid, darkening in color, and may even contain purulent components. These changes often begin with subtle shifts in quantitative indicators before becoming grossly visible redness, swelling, pain, and purulent exudate. Therefore, high-temporal-resolution temperature monitoring and objective image analysis, combined with the patient's subjective pain reports, can transform early infection signals that previously relied on visual and experience-based judgment into quantifiable and traceable local catheter risk indicators.

[0017] An exemplary disposable or reusable skin-attached temperature sensor is applied to the skin near the catheter exit point, for example, one to two centimeters above the catheter exit point, avoiding direct coverage of the catheter exit point to reduce the impact of dressing changes on patch adhesion. The temperature patch communicates with the patient's mobile phone or home gateway device via Bluetooth Low Energy, continuously collecting skin temperature data at sampling intervals of five or ten minutes. Upon initial use, the system records an individual's baseline temperature over a period of time, such as the average value and fluctuation range over the past one to two weeks under infection-free and environmentally stable conditions. Subsequent temperature measurements are compared to this baseline to obtain characteristics such as the magnitude of local temperature increases and intraday fluctuations. During data collection, the system can smooth or remove data with abrupt changes. For example, if a temperature reading suddenly increases by several degrees Celsius compared to the previous reading but then returns to its original level, it can be determined that the increase may be due to a temporary lifting of the patch or the influence of an external heat source, and the interference can be reduced through a filtering algorithm.

[0018] For example, before and after each dialysis session, patients take photos of the catheter exit area following the application's guidance. The application displays a viewfinder on the screen, guiding the patient to include the catheter exit, surrounding skin, and dressing in the frame, and provides image stabilization and exposure prompts to help the patient take photos under suitable lighting conditions. Images uploaded to the system are first automatically cropped, establishing a region of interest centered on the catheter exit, removing excessive background, clothing, and other unwanted elements. Subsequently, the image is converted from its original color space to a color space more suitable for representing color changes, such as using red saturation, brightness, and contrast as features to calculate the average redness score, redness distribution range, and erythema area of ​​the skin region. For exudate areas, a segmentation algorithm identifies wet areas on the dressing, calculating their average brightness, hue, and uniformity to determine whether they are clear, slightly turbid, or significantly turbid. Patterns can also be established for certain typical manifestations; for example, when radial red streaks appear around the skin, the algorithm can identify high-redness bands extending roughly to one side and record them as linear red streaks. To reduce the impact of different mobile phones and lighting conditions, the system can identify a standard color patch or dressing mark in the image. By comparing the actual value of the standard color patch with the value in the image, the system can perform color correction on the entire image, making images taken at different times and with different devices comparable in value.

[0019] For example, before and after each dialysis session, the application displays a simple pain and discomfort assessment interface. Patients can slide a scale line to select the current level of localized pain, from zero to ten, and indicate the presence of auxiliary symptoms such as itching, pulling, or burning sensations. The system converts these subjective descriptions into standardized pain scores and symptom labels. For instance, a pain score of seven accompanied by burning and pulling sensations is marked as a high-risk subjective signal. In this way, even if patients have limited expressive abilities, the system can collect subjective experiences in a structured manner.

[0020] For example, the system integrates features from multiple dimensions, such as the degree of temperature increase, skin redness score, erythema area, presence of linear red streaks, turbidity level of exudate, pain score, and symptom labels, to form a set of catheter-specific risk indicators containing multiple elements. Structurally, this can be represented as a record containing several numerical fields and several categorical fields, each with a clear meaning and value range, facilitating its subsequent use in conjunction with environmental indicators in the risk assessment model.

[0021] It is understandable that catheter exit site infection is not determined by a single factor, but rather is the result of the long-term interplay between environmental pollution levels and local defense mechanisms. Airborne particulate matter, relative humidity, temperature, and residual contamination on object surfaces affect the rate at which microorganisms suspend in the air and grow on objects; local temperature at the catheter exit site, the degree of skin congestion, and the nature of exudate reflect whether the local tissues have been invaded. Looking at local indicators alone, even a minor abrasion or dressing friction can cause temporary redness and swelling; looking at environmental indicators alone, short-term window opening or visitors may increase airborne particulate matter, but not necessarily lead to immediate infection. Therefore, it is necessary to construct a multidimensional risk index that integrates environmental and local risks in a weighted and hierarchical manner, ensuring that the assessment results do not overly rely on a single indicator at a particular moment, while also reflecting the true importance of each factor in infection formation. The core idea of ​​multidimensional integration is to first convert various indicators with different units and dimensions into standardized risk contributions on the same evaluation scale, then determine their weights according to clinical experience and statistical learning results, and finally obtain an environmental-infection risk index ranging from zero to one hundred through a unified calculation rule. The higher the index, the closer the environmental conditions and local manifestations within the same time window are to the combined characteristics of previously infected cases.

[0022] For example, air quality sensors and temperature and humidity sensors can be deployed in the patient dialysis operating space. This could include a laser scattering sensor capable of detecting fine particulate matter concentration and a combined temperature and humidity sensor, reporting the concentration of particulate matter, ambient temperature, and relative humidity in real time. If necessary, a surface cleanliness detection module can also be deployed to periodically inspect key locations such as the dialysis machine casing and work surface, generating a surface contamination index. The system statistically analyzes environmental indicators over the past few hours or even one or two days, calculating average levels, maximum values, and fluctuation ranges. Based on preset safety ranges, these are converted into environmental risk contribution scores; for example, excessively high humidity is assigned a higher risk contribution during humid seasons and a relatively lower risk contribution during dry seasons.

[0023] For example, for numerical or grading indicators such as temperature rise, skin redness score, exudate turbidity level, and subjective pain score in catheter-specific risk indicators, the system maps and standardizes them according to preset rules, ensuring that each indicator has a clear meaning between zero and one. For instance, a local temperature rise of 0.5 degrees Celsius corresponds to a medium level of risk contribution, while a rise of more than one degree Celsius corresponds to a high level of risk contribution; skin redness ranges from no obvious redness and mild redness to obvious diffuse redness and streaks of redness, each corresponding to a different gradient of risk score. Based on this, clinical experts and statistical data analysis jointly determine the weight of each indicator. Generally, catheter exit temperature, skin redness, and exudate turbidity are relatively direct precursors to infection and can be given higher weights; environmental particulate matter concentration and surface contamination index reflect the baseline level of bacterial load and belong to background risk, so they can be given medium weights; temperature and humidity, as promoting factors for microbial reproduction in the environment, have slightly lower weights, but can be significantly amplified under extreme conditions. The weights can be initially set based on clinical consensus and literature. Subsequently, as the system accumulates enough infection case data, they can be readjusted through statistical regression or machine learning models to make the weights more closely reflect the importance of each factor in reality.

[0024] For example, at each assessment point in time, the system standardizes and weights the environmental indicators and local duct risk indicators within the current time window before inputting them into the risk assessment engine. The assessment engine can employ a form similar to a scoring card or probabilistic prediction model, internally combining various features linearly or non-linearly using pre-trained coefficients to obtain a raw risk score, which is then mapped to a range of zero to one hundred using a function. For instance, when environmental particulate matter levels are consistently high but local indicators are generally normal, the system provides a moderately high risk index, indicating an environmental problem but not yet at the stage of local damage. When environmental conditions fluctuate slightly but local temperature and redness show a significant increase, the system will emphasize the weight of local anomalies, raising the risk index to a higher level. During long-term operation, the system continuously records the correlation between each risk index calculation result and whether subsequent infection occurs. If it finds that certain feature combinations frequently appear before actual infection occurs, and the current weight settings are insufficient to reflect their importance, the machine learning module readjusts the weights to make future risk indices more sensitive to these feature patterns.

[0025] Understandably, infection is usually a gradual process, progressing from asymptomatic to mild discomfort, and then to noticeable redness, swelling, and foul odor, often taking several days or even longer. Relying solely on a single risk assessment is susceptible to occasional anomalies, such as measurement errors or brief periods of environmental contamination, which can cause a momentary spike in the index. Furthermore, focusing only on absolute values ​​without considering trends can lead to too many meaningless alerts and miss genuinely dangerous situations where the index, while not immediately crossing the high-risk threshold, continues to rise. Therefore, analyzing the changes in the environmental-infection risk index over time, using methods such as moving averages and slope assessments to identify continuous upward trends, and then issuing tiered alerts accordingly, balances sensitivity and specificity.

[0026] For example, the system uses a preset time window as the unit of analysis, such as the past 48 or 72 hours, and obtains multiple historical values ​​of the environmental-infection risk index within this window. These historical values ​​are smoothed, and a moving average and rate of change are calculated. The moving average reflects the recent overall risk level, while the rate of change reflects whether the index is rising, falling, or remaining roughly flat. To reduce the interference of random fluctuations, the system can set time windows of different lengths. For example, a short window is used to quickly capture sharp changes in the last few hours, while a long window is used to observe a slow rise over several days. When both the short and long window results show that the index is rising, and the moving average gradually approaches or crosses a preset safety threshold, a continuous upward trend can be identified.

[0027] For example, based on clinical experience and historical data, the system sets multiple threshold levels for the environment-infection risk index. For instance, a moving average below a first threshold is considered a safe zone; between the first and second thresholds is considered a low-risk zone; exceeding the second threshold with a positive rate of change is considered a moderate risk zone; and exceeding a third threshold or a rate of change exceeding a certain rate of increase threshold is considered a high-risk zone. The concepts of consecutive exceedances and trend duration can be introduced. For example, if more than three assessments within the past 24 hours are above the low-risk threshold and show a gradual increase, it is considered a low-risk, persistent risk; if the index has steadily risen from 30 to 80 within the past three days with almost no significant decline, it is considered a high-risk trend.

[0028] For example, after the system determines the risk level based on trend analysis, it will automatically generate corresponding alarm content: If it is a mild risk, a notification will be pushed in the application to remind the patient to pay attention to improving the dialysis operating space environment in the near future, such as opening windows for ventilation, reducing family visits, using air purification equipment, and regularly wiping the dialysis machine and operating table, etc. At the same time, it is recommended that the patient carefully observe the catheter outlet condition during the next fluid change; If it is a moderate risk, the system will give more specific operation suggestions, such as suggesting that the patient temporarily move to a less contaminated room for dialysis, change dressings more frequently, and perform local disinfection. If necessary, the system can suggest that the patient upload more photos online or communicate with nursing staff via video; If it is a severe risk, in addition to issuing a high-priority alarm to the patient, the system can also synchronize the alarm information to the relevant nursing center or doctor's end, with the recent risk index curve, changes in environmental indicators, and local image analysis results, so that medical staff can conduct remote assessment and decide whether immediate medical attention or home visit is required.

[0029] In some examples, it also includes: After the alarm information triggers the patient to perform at least one intervention measure among environmental cleaning, catheter care, or changing the dialysis operation space, the updated environmental indicators and catheter local risk indicators continue to be collected. The changes in the corresponding environmental-infection risk index before and after the intervention are compared. When it is determined that the risk index is gradually decreasing, the intervention measure is recorded as an effective intervention, and the individualized risk threshold for the target patient is updated. When it is determined that the risk index is continuously increasing, the risk level is increased and a referral or follow-up visit recommendation is output.

[0030] In some examples, it also includes: The concentrations of suspended particulate matter and volatile organic compounds (VOCs) in the dialysis operating space are monitored using an air quality sensor. The VOC concentration is used to characterize the levels of harmful gases associated with bacterial or fungal growth. Temperature and relative humidity within the dialysis operating space are monitored using temperature and humidity sensors; The surface cleanliness detection module periodically detects the surfaces of the dialysis machine, the operating table, and frequently touched objects near the catheter outlet to generate corresponding surface contamination indices. The concentration of suspended particulate matter, the concentration of volatile organic compounds, the ambient temperature, the ambient humidity, and the surface pollution index are all used as the environmental indicators.

[0031] In some examples, the acquisition of skin temperature information at the catheter exit site reported by the patient's temperature patch, as well as image data of skin color, local exudate color and turbidity around the catheter exit obtained by taking photos via a mobile terminal, and acquisition of local pain or discomfort information manually entered by the patient, are used to form catheter local risk indicators, including: The wearable temperature patch collects the skin temperature around the catheter exit point at a preset sampling period and calculates the temperature rise compared to the patient's baseline temperature. The local skin and dressing area at the catheter exit point are photographed using a mobile terminal imaging device. Based on the image analysis model, the range of skin erythema, color saturation, and possible distribution of linear red streaks are identified. The color, transparency, and turbidity of the exudate area are graded and evaluated. The temperature rise, the area and color change of skin erythema, the turbidity of exudate, and the description of local pain and discomfort manually entered by the patient are used as indicators of local catheter risk.

[0032] In some examples, the multidimensional fusion calculation based on the environmental indicators obtained from the patient's home peritoneal dialysis scenario and the catheter local risk indicators includes: An environmental-infection risk assessment rule base is established, in which influence coefficients and threshold ranges are configured for air particulate matter concentration, environmental humidity, environmental temperature, surface contamination index, local temperature rise, skin color change grading, exudate turbidity grading, and subjective pain score, respectively. In the initial stage of the assessment, each indicator was weighted and summed according to the aforementioned impact coefficients to obtain the standardized environment-infection risk baseline index. The regularized environment-infection risk baseline index is compared with the infection occurrences marked in historical cases, and the influence coefficient is iteratively adjusted based on a statistical model or machine learning model to obtain an optimized environment-infection risk index calculation model for different patient groups.

[0033] For example, when performing multidimensional fusion calculations based on the environmental indicators and catheter local risk indicators obtained from the patient's home peritoneal dialysis scenario, the system can first establish an environment-infection risk assessment rule base on the server side. In this rule base, nephrologists, peritoneal dialysis specialists, and infection control personnel summarize the correlation between indicators such as airborne particulate matter concentration, environmental humidity, environmental temperature, surface contamination index, local temperature increase, skin color change grading, exudate turbidity grading, and subjective pain scores and infection occurrence based on past cases of catheter exit infection and peritonitis. Corresponding risk levels and influence coefficients are then assigned to different value ranges for each type of indicator. For example, airborne particulate matter within a safe range is configured as low risk and assigned a lower influence coefficient, while in the case of... The significantly elevated intervals are configured as high-risk and assigned a higher impact coefficient. The weights of significantly higher local temperatures than the individual baseline, diffuse erythema or linear red streaks on the skin, turbid exudate, and moderate to severe pain scores on the risk of infection are set higher than those for mild environmental fluctuations. The rule base is stored in the form of data tables or configuration files. Each rule includes fields such as indicator name, threshold range, risk level, impact coefficient, and applicable population label. Through preset thresholds and coefficients, raw monitoring values ​​from different sources and with different scales are converted into structured risk contribution units, enabling the system to have interpretable initial assessment basis. Even in the early stages where large-scale training data is lacking, it can reasonably quantify environmental and local anomalies based on clinical experience and provide clear parameter entry points for subsequent model training and automatic optimization.

[0034] In the initial assessment phase, when calculating the infection risk within a specific time window, the system obtains environmental indicators such as air particulate matter concentration, ambient humidity, ambient temperature, and surface contamination index from air quality sensors, temperature and humidity sensors, and surface cleanliness detection modules in the home dialysis environment. Simultaneously, it obtains catheter-specific risk indicators such as the degree of local temperature increase, skin color change grading, erythema area, exudate turbidity level, and subjective pain score from the catheter local monitoring module. These raw values ​​or grading results are then searched for corresponding threshold ranges and influence coefficients in the environment-infection risk assessment rule base. If necessary, instantaneous outliers are smoothed or limited to reduce the impact of occasional noise. Finally, the risk contribution of each indicator is weighted and summed according to the configured influence coefficients to obtain the final result. The system generates a standardized environmental-infection risk baseline index within the current time window. This baseline index can be compressed into a fixed scoring range, such as 0 to 100, using linear or piecewise mapping methods. This index represents the patient's overall infection risk intensity during this time period under predetermined rule assumptions. This step is equivalent to aggregating environmental background risk and catheter-specific abnormal signals within a unified scoring framework, making different variables comparable on a uniform scale. At the same time, strategies such as amplitude limiting and smoothing are used to prevent a single extreme indicator from completely dominating the overall assessment. The system can generate stable and comparable baseline risk scores for each collection cycle in the early deployment stage, facilitating doctors' longitudinal observation of risk changes for the same patient and the system's horizontal comparison of risk levels for different patients or different time periods, providing a foundation for subsequent trend analysis and alarm grading.

[0035] After the system has been running continuously and accumulated sufficient monitoring data and clinical outcomes, the aforementioned standardized environment-infection risk baseline index can be compared and analyzed with the actual infection occurrences marked in historical cases to construct a training dataset. Each sample record includes environmental indicator values ​​within a specific time window, catheter local risk indicator values, the corresponding baseline index, and information indicating whether catheter exit site infection or peritonitis occurred during the subsequent preset observation period. During the training phase, statistical models or machine learning models such as logistic regression, gradient boosting trees, or neural networks can be used. Input features include the location of airborne particulate matter, humidity and temperature combinations, surface contamination index level, local temperature rise level, skin color change grading, exudate turbidity level, and pain score. The occurrence of infection is used as the output label. By iteratively optimizing the model parameters, the true contribution of each feature to the infection risk and the interaction between different features can be calculated. Based on this, the influence coefficients in the environmental-infection risk assessment rule base can be automatically or semi-automatically adjusted, or the trained model can be directly packaged into an environmental-infection risk index calculation engine. This allows the engine to output a risk index that more closely approximates the actual infection probability during daily assessments, based on input features. This is equivalent to using a large amount of real-world data to correct the weights initially set based on experience, so that the assessment model gradually converges to a parameter region that can better distinguish between feature combinations that will develop into infection and feature combinations that will not ultimately lead to infection. This can significantly improve the accuracy of the risk index in predicting the occurrence of infection and reduce false positives and false negatives caused by setting weights based solely on experience. By continuously learning the real outcomes of patients in different regions, seasons, underlying diseases, and living environments, the same method can adapt to different patient groups, forming an environmental-infection risk index calculation model optimized for different patient groups. This allows the model to maintain interpretability while having the ability to continuously evolve.

[0036] In some examples, the influence coefficients are iteratively adjusted based on statistical or machine learning models to obtain an optimized environment-infection risk index calculation model for different patient groups, including: A labeled dataset containing confirmed cases of catheter exit site infection or peritonitis was used as the training sample. The labeled dataset includes time series of environmental indicators and time series of catheter local risk indicators within the corresponding monitoring period, as well as a label indicating whether infection occurred. The time series of environmental indicators and the time series of local duct risk indicators are converted into statistical features, including mean, maximum, minimum, magnitude of change and trend slope, and used as input features for the model. The input features are trained using at least one of a logistic regression model, a gradient boosting tree model, or a neural network model to obtain a predictive model whose output is the probability of infection, and the probability of infection output by the predictive model is mapped to the environment-infection risk index. During the model's deployment and operation, the parameters of the prediction model are updated periodically based on subsequent real case feedback data.

[0037] For example, during the model training phase, the system first uses a labeled dataset containing confirmed cases of catheter exit site infection or peritonitis as training samples. This labeled dataset is compiled from multiple hospitals or long-term follow-up projects. Each sample record corresponds to complete monitoring information for a specific patient within a specific observation window. This includes time series of environmental indicators such as air particulate matter concentration, ambient humidity, ambient temperature, and surface contamination index, as well as time series of catheter-related local risk indicators such as local temperature, skin color change grading, exudate turbidity level, and pain score, along with a label indicating whether catheter exit site infection or peritonitis occurred within a subsequent preset time period. Based on this, the system performs feature engineering on each time series, converting the original continuous curve into a set of statistically significant features. For example, it calculates the mean, maximum, minimum, daily variation range, and trend slope of air particulate matter concentration within the time window. Similar statistics are performed on humidity, temperature, and surface contamination index. Simultaneously, it calculates catheter-related local indicators such as the magnitude of local temperature increase, skin redness score, exudate turbidity level, and pain score. The system calculates the overall level, fluctuation range, and upward or downward trend over the past few hours, integrating these statistical features into a multi-dimensional feature vector as the input feature of the model. This extracts the dynamic changes over a certain period into a static description that is easy for the model to understand, preserving the trend information while reducing the complexity of directly processing long-term series. Subsequently, the system selects at least one of logistic regression, gradient boosting tree, or neural network models as the basic prediction model. The system performs supervised training on the above multi-dimensional input features and the label of whether infection has occurred. Through repeated iterations to optimize the weight parameters and structure in the model, the model learns to give the corresponding infection probability under different feature combinations. After training, the model outputs an infection probability value between zero and one for each time window in daily evaluation. The system then converts this probability into an environment-infection risk index through a preset mapping relationship. For example, the probability is compressed or stretched to a scoring range of zero to one hundred through a monotonically increasing function. This index is used to supplement or replace the basic index calculated by the rule base, transforming complex high-dimensional features into a single interpretable risk score.During the model's deployment, the system does not fix the training parameters indefinitely. Instead, it periodically collects new real-case feedback, such as every three months or after accumulating a certain number of new and uninfected samples. New data is added to the training set or used as incremental samples to retrain or fine-tune the model, allowing the parameters to continuously absorb new population characteristics, seasonal changes, and changes in medical practice. This gradually corrects the actual contribution of different characteristics to infection risk, achieving adaptive optimization for patient groups in different regions, with different underlying conditions, and living environments. On the one hand, this significantly improves the accuracy of the environment-infection risk index in predicting real infection events and its early warning capabilities, reducing false positives and false negatives caused by setting influence coefficients solely based on experience. On the other hand, it enables the model to learn new risk patterns and the performance of vulnerable populations over time, ensuring the same assessment framework remains relevant and effective in long-term use. This is beneficial for continuously improving the accuracy and reliability of peritoneal dialysis catheter infection risk management in home settings.

[0038] In some examples, the continuous trend analysis includes: Calculate the moving average and slope of change of the environment-infection risk index within a sliding time window; If the moving average exceeds a first threshold and the slope of the change is positive, the risk is determined to be an upward trend; If the risk index exceeds the preset threshold for a short period of time but the slope of change is close to zero or negative, the situation will be judged as an occasional fluctuation and only an observation prompt will be generated. If both the moving average and the slope of the risk index exceed the second threshold, the risk is judged to be a sustained and significant upward trend, triggering a high-level risk warning.

[0039] In some examples, the step of generating alarm information based on risk level includes: When the environmental-infection risk index is slightly above the safe threshold, a first type of alarm message is generated. The first type of alarm message includes suggestions for patients to improve the environment, such as strengthening ventilation in the dialysis operation space, closing unnecessary sources of pollution, wiping the dialysis machine and operating table, and changing bed sheets and curtains in a timely manner. When the environmental-infection risk index is in the range of moderately exceeding the safety threshold and the duration exceeds the preset time, a second type of alarm information is generated. The second type of alarm information includes suggesting that the patient temporarily change the dialysis operation space, increase the frequency of cleaning and disinfection of the skin around the catheter outlet, and suggest performing additional environmental disinfection operations before the next dialysis cycle. When the environmental-infection risk index is in a range that is severely exceeding the safety threshold or the trend slope exceeds the preset rate of increase, a third type of alarm information is generated and synchronized to the nursing center information system to prompt nursing staff to arrange remote follow-up visits for patients.

[0040] In some examples, the updated individualized risk threshold includes: If the environmental-infection risk index steadily falls below the safe threshold in multiple samplings after the patient undergoes environmental improvement or catheter care, the intervention is recorded as an effective intervention. Based on the peak value and rate of decline of the risk index before and after the intervention, as well as the type of intervention used, intervention response characteristics are generated for the patient. Within multiple monitoring cycles of the same patient, the intervention response characteristics of different intervention measures were accumulated. Based on statistical analysis, the thresholds for mild, moderate, and severe environmental-infection risk index of the patient were adjusted so that the thresholds could reflect the patient's sensitivity to environmental changes and local abnormalities. In subsequent assessments, the environmental-infection risk index will be graded based on the updated individualized thresholds to achieve risk warnings tailored to each individual.

[0041] In some examples, the surface cleanliness detection module includes optical or chemical sensors for detecting residual organic matter, sebum, or biofilm precursors. These sensors are positioned on the dialysis machine housing, operating table surface, near catheter fixation devices, or on surfaces frequently touched by patients to detect the presence of contaminated areas conducive to bacterial or fungal attachment and reproduction, and to generate a surface contamination index based on the detection results.

[0042] In some examples, the system further considers external seasonal factors and patient dialysis protocol information when analyzing the environment-infection risk index, including: The external seasonal factors include local seasonal temperature and humidity changes and information on peak periods of infectious respiratory infections; The dialysis protocol information includes the patient’s dialysis frequency, the volume of dialysis fluid per session, the duration of catheter indwelling time, and whether there is a history of multiple catheter replacements. During periods of high seasonal humidity or high temperature, or in cases of high dialysis frequency or long catheter indwelling time, the classification threshold of the environmental-infection risk index should be appropriately lowered to enhance the sensitivity of early warning for high-risk periods and high-risk patients.

[0043] In some cases, considering that temperature increases truly caused by local inflammation often exhibit spatial asymmetry—that is, the skin temperature at the site of inflammation is significantly higher than that of symmetrical locations on the same patient's body or adjacent unaffected areas—while temperature fluctuations due to changes in air conditioning temperature, bathing, exercise, or overall fever are more likely to show bilateral or systemic synchronous increases, it is easy to misinterpret overall fever as local inflammation if only the absolute temperature at a single point at the catheter exit point is considered. Based on this, some examples also include: A second temperature patch is placed on the opposite or relatively symmetrical part of the skin around the catheter exit. The skin temperature at the catheter exit and the skin temperature at the opposite symmetrical position are obtained respectively. The temperature difference and the trend of temperature difference change are calculated within a preset time window to form a thermal asymmetry index to characterize the degree of local inflammatory thermal asymmetry. When performing multidimensional fusion calculation based on the environmental index and the local risk index of the catheter, the thermal asymmetry index is used to replace or take priority over the absolute temperature rise of the single catheter exit in the weighted processing, so as to reduce the interference of the overall environmental temperature change on the inflammation judgment and improve the accuracy of identifying the true risk of local fever at the catheter exit.

[0044] Understandably, by attaching a second temperature patch at a position symmetrical to the catheter exit point—for example, on a skin area at the same height on the opposite side of the abdomen or on the same side but farther from the catheter exit point and without catheter stimulation—skin temperature data is collected synchronously from both patches under the same environment. Within a preset time window, the temperature difference between the two patches and its trend over time are calculated, thus obtaining a thermal asymmetry index to characterize the degree of local inflammatory thermal asymmetry: when the overall environment becomes hot, the temperatures of both patches rise, but the difference does not change significantly, and the thermal asymmetry index approaches zero; when the catheter exit point becomes hot due to infection or irritation, the temperature rise of the patch on the catheter side is significantly greater than that on the opposite side, the temperature difference and its slope increase significantly, and the thermal asymmetry index rises. In practice, during the initial deployment phase, the system collects skin temperature data from both patches under relatively stable environmental conditions and without infection over a period of time. It statistically analyzes the baseline and normal fluctuation range of temperature difference under normal conditions; for example, the temperature difference is mostly within 0.3 degrees Celsius, occasionally fluctuating briefly to 0.5 degrees Celsius. During subsequent real-time monitoring, it calculates the mean, maximum, and rate of increase of the temperature difference between the catheter side and the contralateral side using a sliding window, such as the last 24 hours or the last 48 hours, and combines these statistics into a thermal asymmetry index. In the multi-dimensional fusion calculation phase of the environment-infection risk index, the system no longer uses the absolute temperature increase at the catheter exit as the sole local temperature feature. Instead, it prioritizes the use of thermal asymmetry in the weighted processing. For example, if the thermal asymmetry index remains within the baseline range for a long period, even if the external high temperature or air conditioning causes a slight increase in the catheter side temperature, it does not significantly improve the infection risk score. Conversely, when the thermal asymmetry index gradually increases under stable environmental conditions and is accompanied by mild erythema, the system determines that local inflammation may be developing, thus increasing the weight of local fever in the comprehensive assessment. Thus, on the one hand, it effectively offsets the interference of overall environmental temperature changes and systemic fever on the judgment of local infection, reducing false alarms caused by factors such as summer heat, winter heating, or hot baths; on the other hand, by observing how much hotter the right side is than the left side, rather than how hot the right side is absolutely, the system can capture the trend of asymmetrical temperature rise even in the early stages of mild local inflammation. For example, when the temperature on both sides of a patient rises from 34 degrees to 34.5 degrees, the thermal asymmetry index remains unchanged and the risk index is stable. However, when the temperature on the catheter side continues to rise to 35 degrees while the temperature on the opposite side remains around 34.5 degrees, the thermal asymmetry index rises from 0.2 to 0.5. The system then issues a local inflammation trend warning even when the environmental temperature appears stable, thus providing an early risk signal before the patient feels obvious pain or sees redness and swelling.

[0045] In some cases, a significant source of infection in home peritoneal dialysis is the patient's lack of intuitive understanding of how contamination travels from their fingers to the catheter exit point. Many subtle touches and cross-regional movements, seemingly harmless to the patient, are crucial from a microbial transmission perspective. Traditional textual or video education struggles to cultivate stable spatial intuition in patients. Augmented reality technology, however, can visualize unseen contamination overlaid on a real-world scenario in real-time during simulated procedures, allowing patients to practice repeatedly in a safe environment and develop muscle memory. Based on this, some examples also include: A training mode is provided in non-dialysis treatment states. The imaging device of the mobile terminal captures images of the patient's table and hands during a simulated peritoneal dialysis procedure. An augmented reality layer with virtual contamination sources and contamination diffusion effects is overlaid on the images. Based on the detected hand touch sequence, touch frequency, and cross-regional movement trajectory, a simulated contamination diffusion heatmap is generated in real time, and a contamination diffusion index is calculated to quantify the potential cross-contamination level of the patient under standard operating procedures. The contamination diffusion index is used as a behavioral risk feature to adjust the aseptic operation-related weights and / or individual risk thresholds in the environmental-infection risk index calculation model, making the environmental-infection risk index of patients with higher contamination diffusion indices more sensitive during actual monitoring.

[0046] Understandably, the system provides a training mode in non-dialysis treatment states. Patients place dialysis bags, catheters, sterilization supplies, and dressings on the operating table according to standard procedures. The mobile terminal's camera continuously captures images of the table and hands. The system uses hand detection and object recognition algorithms to determine the outline of the hand, the location of each key consumable, and each contact event. It then overlays an augmented reality layer on the screen with virtual contamination sources and contamination diffusion effects. For example, after a hand touches an unsterilized area of ​​the table, the outline of the palm is painted with a light-colored contamination cloud. When the hand touches the area near the catheter connector again, the contamination color in that area deepens and spreads along the subsequent contact path on the operating table, forming a diffusion effect similar to a heat map. After the entire simulation process is completed, the system counts factors such as the contact sequence, the number of touches, the cross-area movement trajectory, and the frequency of crossing boundaries from highly contaminated areas to key sterile areas. It then calculates a contamination diffusion index, with a higher value indicating more potential cross-contamination paths under the standard procedure. This contamination diffusion index is not only used to provide feedback to patients at the end of training, such as indicating which steps repeatedly lead to contamination diffusion, but also fed back as a behavioral risk characteristic into the environmental-infection risk index calculation model. This is used to adjust the weights related to aseptic technique and individual risk thresholds. For example, for patients whose contamination diffusion index is consistently high during multiple training sessions, the system can lower the local abnormality threshold required to trigger an alarm or increase the weight of the aseptic technique score in the overall risk during actual monitoring. This allows the system to remind the patient to pay more attention when there are slight abnormalities in the environment and local indicators. Conversely, for patients whose contamination diffusion index has significantly decreased and whose action patterns are stable after multiple training sessions, the model can appropriately increase the alarm threshold to avoid causing excessive interference. Therefore, by visualizing the originally abstract process of microbial diffusion through augmented reality, patients can intuitively see the consequences of touching their phone and then the catheter tip during practice. This will spontaneously reduce cross-area touching and unnecessary contact paths in future real-world operations. As the behavior itself improves, even if there are occasional slight fluctuations in environmental and local indicators, the overall risk of infection will be reduced. At the same time, the system establishes a quantifiable operational risk profile for each patient through the contamination diffusion index. The risk model no longer simply assumes that all patients have the same operational level, but dynamically adjusts the model parameters based on training performance to achieve individualized early warning driven by behavioral characteristics.

[0047] In some cases, considering that the movement of dust particles and fibrous particles in the home environment is not only affected by airflow and gravity but also closely related to electrostatic fields, high electrostatic potentials tend to accumulate on the body surface, especially in dry winter climates or when patients wear synthetic clothing. This can promote the adsorption of fine dust onto clothing and skin, and may also cause momentary discharges when touching metal equipment or duct parts, causing particles originally attached to the clothing surface to detach and enter the air layer around the duct outlet in a short-distance jet, thus forming a pollution channel that is difficult to detect with conventional air particulate matter indicators. Based on this, some examples also include: An electrostatic potential detection device is worn on the patient's wrist, waist, or other suitable location to continuously collect electrostatic potential data on the patient's body surface before, during, and after dialysis. Within a preset time window, the amplitude, fluctuation range, and frequency of discharge events of the electrostatic potential are calculated to form an electrostatic load index. When performing multidimensional fusion calculations based on the environmental indicators and catheter local risk indicators, the electrostatic load index is used as one of the environmental indicators to characterize the potential risk of contamination around the catheter caused by electrostatic adsorption of dust and instantaneous desorption of particles. When the electrostatic load index exceeds a preset threshold, the corresponding risk weight is increased, and when an alarm message is generated, the patient is prompted to adjust the material of their clothing or take electrostatic control measures such as grounding.

[0048] Understandably, an electrostatic potential detection device, such as a miniature potential sensor integrated into a smart bracelet or belt buckle, is worn on the patient's wrist, waist, or other suitable location. This device periodically measures the electrostatic potential on the patient's body surface via a high-impedance input circuit, continuously sampling before, during, and after dialysis procedures. Within a preset time window, the amplitude, fluctuation range, and frequency of events exceeding a certain discharge threshold of the electrostatic potential are statistically analyzed to calculate the electrostatic load index. This index reflects both the degree of body charge under current environmental and clothing conditions and the activity level of discharge events. In the multi-dimensional fusion stage of the environment-infection risk index, the system integrates the electrostatic load... As a special type of electrical risk among environmental indicators, the index should be included in the assessment model. For example, when the electrostatic load index is at a low level for a long time, even if there are occasional slight dust fluctuations, the system considers the additional pollution risk caused by electrostatic adsorption and spraying to be low and will not significantly adjust the overall risk. However, when the electrostatic load index is high for a long time in winter or in air-conditioned rooms and is accompanied by frequent discharge events, the system will increase the weight of dust-related risks and provide specific suggestions when generating an alarm, such as reminding patients to change to cotton or antistatic clothing, touch metal faucets or grounding floors to release static electricity before critical operations, and add grounding mats near the operating table if necessary. Building upon this foundation, firstly, by explicitly monitoring electrostatic states and discharge events, a previously invisible and intangible pollution-promoting factor is brought into the monitoring field, compensating for the difficulty in identifying risk sources solely through traditional environmental indicators such as air particulate matter concentration and humidity. Secondly, by linking the electrostatic load index with the environmental-infection risk index, the system provides early warnings and behavioral adjustment suggestions even if particulate matter sensor readings are normal for a short period when electrostatic risk is high. This prevents scenarios where electrostatic charge attracts dust to cuffs, and during operation, instantaneous discharge sprays dust into the catheter area. For example, if a patient is wearing synthetic fiber pajamas in winter, and the particulate matter concentration appears acceptable, but the electrostatic load index repeatedly exceeds the set threshold and multiple discharge events are recorded, the system will therefore increase the behavioral risk in the risk assessment and suggest changing clothing and performing grounding operations. This electrostatic-based risk control is more comprehensive than the traditional approach of completely ignoring the impact of electrostatics. Thirdly, with long-term data accumulation, the model can also discover the correlation between days with high electrostatic load indices, abnormal local indicators, and infection events, thereby further improving the weight settings and making electrostatic control measures a truly effective tool for reducing infections during home peritoneal dialysis.

[0049] Please see Figure 2 One embodiment of the peritoneal dialysis catheter risk management device in this application includes: The acquisition unit 21 is used to acquire the skin temperature information at the catheter exit point reported by the temperature patch worn by the patient, as well as the image data of skin color, local exudate color and turbidity around the catheter exit point obtained by taking pictures through the mobile terminal, and to acquire the local pain or discomfort information manually entered by the patient, so as to form a local risk indicator for the catheter. The calculation unit 22 is used to perform multidimensional fusion calculation based on the environmental indicators and catheter local risk indicators obtained in the patient's home peritoneal dialysis scenario to obtain an environment-infection risk index. The multidimensional fusion calculation includes weighting the concentration of air particulate matter, humidity, temperature, local temperature change range, skin discoloration degree and exudate turbidity change according to preset weights. The determination unit 23 is used to perform continuous trend analysis on the environment-infection risk index. When the environment-infection risk index is detected to be continuously rising within a preset time window and exceeding a preset safety threshold, the risk level is determined according to the degree of exceeding the threshold and the rate of trend change, and corresponding alarm information is generated based on the risk level.

[0050] like Figure 3 As shown, this application embodiment also provides an electronic device 300, including a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor. When the processor 320 executes the computer program 311, it implements the steps of any of the above-mentioned methods for peritoneal dialysis catheter risk management.

[0051] Since the electronic device described in this embodiment is the device used to implement the peritoneal dialysis catheter risk management device in the embodiments of this application, those skilled in the art can understand the specific implementation method and various variations of the electronic device in this embodiment based on the method described in the embodiments of this application. Therefore, how the electronic device implements the method in the embodiments of this application will not be described in detail here. Any device used by those skilled in the art to implement the method in the embodiments of this application is within the scope of protection of this application.

[0052] In practical implementation, when the computer program 311 is executed by the processor, it can achieve the following: Figure 1 Any of the corresponding implementation methods in the embodiments.

Claims

1. A method for risk management of peritoneal dialysis catheters, characterized in that, include: The system acquires skin temperature information at the catheter exit site reported by the temperature patch worn by the patient, as well as image data of skin color, local exudate color and turbidity around the catheter exit obtained by taking photos through a mobile terminal, and also acquires information on local pain or discomfort manually entered by the patient, in order to form local risk indicators for the catheter. Based on the environmental indicators obtained in the patient's home peritoneal dialysis scenario and the catheter local risk indicators, a multidimensional fusion calculation is performed to obtain the environment-infection risk index. The multidimensional fusion calculation includes weighting the concentration of air particulate matter, humidity, temperature, local temperature change range, skin discoloration degree and exudate turbidity change according to preset weights. Continuous trend analysis is performed on the environment-infection risk index. If the environment-infection risk index is detected to be continuously rising within a preset time window and exceeding a preset safety threshold, the risk level is determined based on the degree of exceeding the threshold and the rate of trend change, and corresponding alarm information is generated based on the risk level.

2. The method as described in claim 1, characterized in that, Also includes: After the alarm information triggers the patient to perform at least one intervention measure among environmental cleaning, catheter care, or changing the dialysis operating space, the updated environmental indicators and catheter local risk indicators continue to be collected. The changes in the corresponding environmental-infection risk index before and after the intervention are compared. When it is determined that the risk index is gradually decreasing, the intervention measure is recorded as an effective intervention, and the individualized risk threshold for the target patient is updated. When it is determined that the risk index is continuously increasing, the risk level is increased and a referral or follow-up visit recommendation is output.

3. The method as described in claim 1, characterized in that, Also includes: The concentrations of suspended particulate matter and volatile organic compounds (VOCs) in the dialysis operating space are monitored using an air quality sensor. The VOC concentration is used to characterize the levels of harmful gases associated with bacterial or fungal growth. Temperature and relative humidity within the dialysis operating space are monitored using temperature and humidity sensors; The surface cleanliness detection module periodically detects the surfaces of the dialysis machine, the operating table, and frequently touched objects near the catheter outlet to generate corresponding surface contamination indices. The concentration of suspended particulate matter, the concentration of volatile organic compounds, the ambient temperature, the ambient humidity, and the surface pollution index are all used as the environmental indicators.

4. The method as described in claim 1, characterized in that, The system acquires skin temperature information at the catheter exit site reported by the patient's temperature patch, as well as image data of skin color, local exudate color, and turbidity around the catheter exit site obtained through mobile terminal photography, and acquires information on local pain or discomfort manually input by the patient, to form local catheter risk indicators, including: The wearable temperature patch collects the skin temperature around the catheter exit point at a preset sampling period and calculates the temperature rise compared to the patient's baseline temperature. The local skin and dressing area at the catheter exit point are photographed using a mobile terminal imaging device. Based on the image analysis model, the range of skin erythema, color saturation, and possible distribution of linear red streaks are identified. The color, transparency, and turbidity of the exudate area are graded and evaluated. The temperature rise, the area and color change of skin erythema, the turbidity of exudate, and the description of local pain and discomfort manually entered by the patient are used as indicators of local catheter risk.

5. The method as described in claim 1, characterized in that, The multidimensional fusion calculation based on the environmental indicators and catheter local risk indicators obtained from the patient's home peritoneal dialysis scenario includes: An environmental-infection risk assessment rule base is established, in which influence coefficients and threshold ranges are configured for air particulate matter concentration, environmental humidity, environmental temperature, surface contamination index, local temperature rise, skin color change grading, exudate turbidity grading, and subjective pain score, respectively. In the initial stage of the assessment, each indicator was weighted and summed according to the aforementioned impact coefficients to obtain the standardized environment-infection risk baseline index. The regularized environment-infection risk baseline index is compared with the infection occurrences marked in historical cases, and the influence coefficient is iteratively adjusted based on a statistical model or machine learning model to obtain an optimized environment-infection risk index calculation model for different patient groups.

6. The method as described in claim 1, characterized in that, The iterative adjustment of the influence coefficient based on a statistical model or machine learning model to obtain an optimized environment-infection risk index calculation model for different patient groups includes: A labeled dataset containing confirmed cases of catheter exit site infection or peritonitis was used as training samples. The labeled dataset includes time series of environmental indicators and time series of catheter local risk indicators within the corresponding monitoring period, as well as labels indicating whether infection occurred. The time series of environmental indicators and the time series of local duct risk indicators are converted into statistical features, including mean, maximum, minimum, magnitude of change and trend slope, and used as input features for the model. The input features are trained using at least one of a logistic regression model, a gradient boosting tree model, or a neural network model to obtain a predictive model whose output is the probability of infection, and the probability of infection output by the predictive model is mapped to the environment-infection risk index. During the model's deployment and operation, the parameters of the prediction model are updated periodically based on subsequent real case feedback data.

7. The method according to any one of claims 1 to 6, characterized in that, Also includes: Temperature patches were placed on the opposite or relatively symmetrical part of the skin around the catheter exit to obtain the skin temperature at the catheter exit and the skin temperature at the opposite symmetrical part. The temperature difference and its trend within a preset time window are calculated to form a thermal asymmetry index for characterizing the degree of local inflammatory thermal asymmetry. When performing multidimensional fusion calculations based on the aforementioned environmental indicators and catheter local risk indicators, the thermal asymmetry indicators are used to replace or take priority over the absolute temperature rise at the single catheter outlet in the weighting process.

8. A peritoneal dialysis catheter risk management device, characterized in that, include: The acquisition unit is used to acquire skin temperature information at the catheter exit point reported by the temperature patch worn by the patient, as well as image data of skin color, local exudate color and turbidity around the catheter exit point obtained by taking pictures through the mobile terminal, and to acquire local pain or discomfort information manually entered by the patient, so as to form local risk indicators for the catheter. The calculation unit is used to perform multidimensional fusion calculation based on the environmental indicators and local risk indicators of the catheter in the patient's home peritoneal dialysis scenario to obtain the environment-infection risk index. The multidimensional fusion calculation includes weighting the concentration of air particulate matter, humidity, temperature, local temperature change range, skin discoloration degree and exudate turbidity change according to preset weights. The determination unit is used to perform continuous trend analysis on the environment-infection risk index. When the environment-infection risk index is detected to be continuously rising within a preset time window and exceeding a preset safety threshold, the risk level is determined according to the degree of exceeding the threshold and the rate of trend change, and corresponding alarm information is generated based on the risk level.

9. An electronic device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program stored in the memory, implements the steps of the peritoneal dialysis catheter risk management method as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the peritoneal dialysis catheter risk management method as described in any one of claims 1-7.

Images