EICU transfer-out patient staging rehabilitation evaluation system

Abstract

The invention relates to an EICU transfer-out patient staging rehabilitation evaluation system, which performs trend analysis on continuous physiological parameters when a patient is transferred out of an EICU, judges whether the patient really enters a rehabilitation feasible window by identifying a dynamic stable interval of an index, and dynamically adjusts a staging evaluation starting point based on a recovery trend drift rule. The misjudgment risk of a false positive stable state is eliminated; dividing the functional state of the patient into a plurality of functional clusters, and dynamically judging whether the plurality of clusters are converted from an intervention dependence state to an autonomous maintenance state or not; recording a recovery rate gain generated by each type of intervention measures to the target function cluster, and reversely correcting a current stage division result when monitoring an intervention effect decreasing or reversing trend, so as to form a dynamic closed-loop logic among intervention, response and stages; high-frequency small-amplitude abnormal changes in the low-intensity fluctuation physiological signals of the patient are detected, and the latent rehabilitation decline trend is recognized by setting the relation between a continuous fluctuation threshold value and response stability changes after intervention.

Description

TECHNICAL FIELD

[0001] The present application relates to a patient rehabilitation evaluation system, in particular to an EICU transfer patient staging rehabilitation evaluation system. BACKGROUND

[0002] In combination with the existing public document CN118948294A "Postoperative intelligent monitoring method and system", it can be seen that the present application mainly focuses on the physiological signal monitoring related to the heart of postoperative patients and the intelligent evaluation of arrhythmia risk. The core lies in the feature extraction of characteristic indexes such as heart rate variability (HRV) and QT interval, and the generation of multiple change rate indexes such as HRV change rate HVR, RMS change rate RVR, and QTV change rate QTR through time series analysis of sliding window, so as to calculate the arrhythmia risk index ARI, and match the preset upper and lower threshold values to form the risk evaluation result and assist the medical decision. Although this scheme provides certain intelligent capability in the field of postoperative cardiac function monitoring, it has obvious deficiencies and limitations in the face of the staging evaluation needs of complex and multi-system recovery state of EICU (intensive care unit) patients transferred to general wards, and it is difficult to meet the fine requirements of multi-dimensional rehabilitation state judgment and stage transition management. First of all, the present application method highly depends on the cardiac single system parameters, and its monitoring range and decision basis mainly revolve around the electrocardiogram signal, especially focusing on the short-term fluctuation trend of indexes such as HRV, RR interval and QT interval, which is suitable for the early warning of single arrhythmia, but in the highly complex population of EICU transfer patients, rehabilitation is not just the stability of a single physiological dimension, but involves the coordinated recovery state of multiple systems such as neurobehavior, motor ability, cardiopulmonary endurance, metabolic homeostasis and cognitive perception. Therefore, the traditional postoperative electrocardiogram monitoring method cannot identify the coupling and decoupling process between multiple functional clusters, and cannot provide support for the judgment of staging transition logic, which is easy to cause the risk of misjudgment caused by excessive dependence on single-point indicators.

[0003] Secondly, the timing analysis of the technology is based on a sliding window for smoothing, mainly used for quantifying short-term fluctuation trends, but does not provide a mechanism for identifying feasible windows and false positive elimination. In the early rehabilitation period after EICU transfer, patients often show transient stability of indicators, but due to the quiet environment or temporary removal of intervention, physiological parameters show false stationary phenomena. At this time, if the rehabilitation stage transition is determined based on the stability of the indicators for a short time, it is easy to cause misclassification and bring the risk of over-rehabilitation rhythm. In addition, the patent does not construct a functional cluster-driven rehabilitation staging structure, and the evaluation result is a risk index, not a distributed functional unit judgment for staging transition. In the rehabilitation process of EICU patients, the recovery rhythm of different systems is often different, for example, the improvement of motor ability may precede cognitive recovery or self-control ability of neurobehavior. Without clustered evaluation and transition judgment logic, the system cannot determine which functions have transition conditions and which are still in the intervention-dependent period.

[0004] In terms of intervention and response mechanisms, the patent lacks recording and feedback correction logic for intervention response, and its algorithm structure is a static scoring strategy that cannot adjust the current evaluation logic based on the effectiveness of intervention measures. Furthermore, the existing technology does not consider the fusion of patient subjective behavior and perception signals, nor does it construct an artificial confirmation mechanism for doctor-patient interaction. For EICU patients, their subjective initiative such as spontaneous behavior, autonomous adjustment intention (such as active breathing, and attempting to sit up) is also important during the rehabilitation process. Therefore, the lack of behavior intention and execution consistency judgment will affect the completeness of the evaluation. SUMMARY

[0005] The purpose of the present application is to provide an EICU transfer patient staging rehabilitation evaluation system, thereby solving some of the problems and deficiencies pointed out in the background art.

[0006] The technical problems solved by the present application are solved by the following technical solution: An EICU transfer patient staging rehabilitation evaluation system, comprising: performing trend analysis on continuous physiological parameters when the patient is transferred out of the EICU, identifying the dynamic stable interval of the indicators to determine whether the patient has truly entered the rehabilitation feasible window, and dynamically adjusting the starting point of the staging evaluation based on the recovery trend drift rule to eliminate the risk of misjudgment caused by false positive stable state; The patient's functional state is divided into multiple functional clusters, including neurobehavior, motor ability, cardiopulmonary load, metabolic self-stability, and cognitive perception. By monitoring the migration state of the underlying indicators in each functional cluster, it is dynamically determined whether multiple clusters have changed from intervention-dependent state to self-maintenance state to serve as the basis for the transition of the patient's rehabilitation stage; The recovery rate gain of each type of intervention measure on the target functional cluster is recorded, and the current stage division result is corrected in reverse when monitoring the decreasing or reversing trend of intervention effect, thereby forming a dynamic closed-loop logic among intervention, response and stage; high-frequency small-amplitude abnormal changes in the patient's low-intensity fluctuating physiological signal are detected, including night respiratory micro-tremor, HRV linear deviation and peripheral microcirculation abnormality, and a latent rehabilitation decline trend is identified by setting a continuity fluctuation threshold and a response stability change relationship after intervention.

[0007] Further, the trend analysis of the continuous physiological parameter includes: preferentially collecting data during the period without artificial intervention after the patient is transferred out of the EICU, and excluding the monitoring data of the high-frequency period of medical operation; the identification of the dynamic stable interval is provided with a buffer observation period before determining the rehabilitation feasible window, and passive physiological monitoring is performed in the buffer observation period, and if the fluctuation of any index exceeds the preset range, the start time of the staging evaluation is delayed.

[0008] Further, in the process of dividing the patient's functional state into multiple functional clusters, the system sets the trigger condition for staging transition as the intervention-dependent state of at least two functional clusters simultaneously transforming into autonomous maintenance state, and maintains the index stable in the two consecutive evaluation periods before determination; the state transformation of the neurobehavioral cluster must be based on the premise that the patient's autonomous behaviors including wakefulness maintenance, active communication or visual tracking are continuously up to standard and the physiological signals are consistently improved, and is used as an effective transition criterion.

[0009] Further, in the evaluation of the recovery rate gain of the intervention measure on the target functional cluster, the subjective rehabilitation perception feedback index of the patient is preferentially referred to, and if the subjective feedback is inconsistent with the physiological improvement trend, the current intervention state is maintained and the staging update is delayed; in the intervention response evaluation process, the system sets that the intervention scheme can be adjusted at most twice within a stage, and if the index does not improve or reverses after continuous adjustment, the stage rollback mechanism is triggered and an abnormal record is generated.

[0010] Further, in the execution process of the dynamic closed-loop logic, the judgment of each stage transition requires confirmation by medical staff, and the system synchronously pushes the transition suggestion to the artificial audit end after generating the transition suggestion, and if the confirmation fails, the evaluation period is extended.

[0011] Further, the method for dynamically judging whether multiple clusters are transformed from intervention-dependent state to autonomous maintenance state is based on the following continuous judgment mechanism: P1, time continuity and intervention reaction weakening condition, the core index of the functional cluster still maintains stable or improves in the case of no new intervention or intervention intensity weakening in a plurality of consecutive evaluation periods, and it is determined to have initial autonomous maintenance ability; P2, recovery response lag shortening condition, after the same intervention is applied to the patient, identify the functional cluster to show a shortened recovery response time or an early spontaneous recovery behavior, indicating that the intervention is no longer the main driving force; P3, functional cluster cooperative decoupling condition, detect that the functional cluster gradually breaks away from the supportive input of other clusters during the staging process and independently maintains its core indicators, which is considered as completed functional decoupling and can be used as an independent transition unit; P4, active behavior replaces passive response condition, identify that the patient has a spontaneous adjustment behavior in the functional cluster, and the behavior persistence is consistent with the improvement trend of the indicator, which is preferentially determined as an active recovery state and is one of the transition bases.

[0012] Further, the time continuity and intervention response weakening condition includes that the core indicators of the functional cluster do not have abnormal fluctuations in at least two consecutive evaluation periods, and still maintain a stable state under the condition that the intervention frequency decreases by more than a preset proportion, and the system marks the functional cluster as a low dependence state candidate accordingly; in the recovery response lag shortening condition, the system compares the response time of the indicators before and after the intervention after performing the same intensity of rehabilitation intervention, and confirms that the functional cluster enters the recovery response acceleration stage when the recovery response is detected to be ahead of the set threshold in two consecutive evaluations; In the above, the dynamic rehabilitation staging transition judgment method based on evaluation period trend judgment and response lag change is used to identify whether the patient's functional cluster is transformed from intervention dependence state to autonomous maintenance state, and the core is: 1. In the consecutive evaluation period, detect whether the core indicator maintains stability under the condition that the intervention intensity decreases, and mark it as a low dependence candidate state; 2. Under the same intensity intervention, judge whether the response time of the indicator is significantly ahead of time, and identify it as a response acceleration stage; In order to unify the trend quantization results of the two dimensions, the following state transformation function formula is used: Among them: The functional cluster transition score value (state transition index) is used to judge whether to enter the autonomous maintenance state; The number of consecutive evaluation periods (for example: 2 or 3 periods); The intervention frequency reduction ratio in the period reflects the degree of intervention weakening; The stability weight of the core indicator in the period (the smaller the fluctuation range, the greater the value); The number of slight abnormalities of the indicator in the period (used to suppress the false stability phenomenon); ​​​Indicates a non-zero balance factor to avoid a zero denominator (set to a very small constant); Indicates the response time shortening index, indicating the ratio of the index response time advance after intervention; Indicates the response time sensitive adjustment coefficient (empirical adjustment value), controls the degree of incentive to the advance recovery trend; The first part is the comprehensive score of multi-cycle intervention-stability: when the intervention frequency decreases (Δt↑) and the index fluctuation is small (R↑), the number of abnormalities is low (H↓), the score of this part is improved; the second part is the advance reward factor of recovery response: if the recovery time is shortened after intervention, Increase, multiplied by the adjustment coefficient , amplify the transition intention; when Exceed the set transition threshold (set by the system or experts), it is judged that the function cluster has transformed from an intervention-dependent state to an autonomous maintenance state, and enters the rehabilitation transition state; The function cluster transition judgment method realizes quantitative evaluation through the state transition function In a continuous number of evaluation cycles, the system calculates the intervention intensity reduction ratio, the index stability degree, the abnormal frequency and the index response time shortening trend, and integrates them into the function calculation result, if the result exceeds the set transition threshold, it is judged that the function cluster has the maintenance ability to get rid of intervention, which is used for dynamic rehabilitation phase transition decision.

[0013] Further, under the condition of cooperative decoupling of the function cluster, the index trigger sequence and dependent behavior between each function cluster are monitored in real time, and a cluster that completes index standard and stable maintenance without the support of other clusters is identified as having independent maintenance ability; in the condition of active behavior replacing passive response, the self-regulation action of the patient in the daily non-induction environment is identified, and the number of periods consistent with the index change trend of the corresponding function cluster is used as the confirmation standard to determine that the patient has the initiative ability of phase transition.

[0014] Further, the system generates a function cluster state conversion mark after executing any one of the judgment conditions of P1 to P4, records the trigger condition, trigger time and corresponding intervention record, which is used for subsequent phase transition decision and path review; when identifying the cooperative decoupling condition P3, if it is found that other function clusters are still in a high dependence state, the decoupled cluster is allowed to enter the transition buffer period preferentially.

[0015] Further, the continuous judgment mechanism supports combined judgment strategy, when both P1 and P2 conditions are met and last for more than two evaluation cycles, the function cluster state is converted to a quasi-transition cluster, and the final confirmation process before transition is enabled; when evaluating the P4 condition, if the patient appears multiple autonomous behaviors but is not accompanied by consistency of the core index improvement trend, the behavior is not counted in the transition judgment.

[0016] The beneficial effects of the present application: through real-time monitoring and dynamic analysis of multiple functional clusters of patients (such as neurobehavior, motor ability, cardiopulmonary load, etc.), personalized rehabilitation paths can be provided according to the specific situation of each patient. The system not only considers the changes of physiological data, but also integrates the subjective feedback and behavior performance of the patient, thereby ensuring the accuracy and individualization of rehabilitation evaluation. Through continuous physiological data analysis and intervention feedback, the system can dynamically determine whether the patient has changed from intervention-dependent state to autonomous maintenance state, and automatically generate rehabilitation stage transition decisions based on this. This automatic decision mechanism reduces the error of manual judgment, improves the efficiency of rehabilitation evaluation, and avoids premature or delayed rehabilitation intervention.

[0017] The system can real-time monitor and analyze the coordination between each functional cluster. When a functional cluster gradually maintains its core indicators independently without relying on other clusters, the system will identify it as decoupled and support its transition to the next rehabilitation stage. This coordinated decoupling mechanism helps to identify the patient's true autonomous maintenance ability and avoid misjudgment caused by single indicator judgment. For functional clusters that have not yet fully developed independent maintenance ability, the system sets a transition buffer period to ensure stable transition to autonomous maintenance state. This flexible adjustment mechanism can reduce the risk of premature transition caused by individual differences in patients. At the same time, if the intervention effect of the patient appears to be reversed or does not meet expectations, the system can trigger a stage rollback mechanism to avoid adverse consequences caused by excessive intervention. The system not only relies on automatic evaluation, but also combines manual confirmation mechanism. When the system generates a transition suggestion, it will be pushed to medical staff for review to ensure that the final decision meets the actual rehabilitation progress of the patient. This closed-loop feedback mechanism ensures the reliability and accuracy of each transition decision and improves the participation and decision transparency of medical staff in patient rehabilitation. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 Simplified flowchart for EICU transfer patient staging rehabilitation evaluation of the present application.

[0019] Figure 2 Functional relationship diagram for EICU transfer patient rehabilitation evaluation of the present application.

[0020] Figure 3 Simplified decision flow for functional cluster transition of the present application.

[0021] Figure 4 Simplified flowchart for rehabilitation process of patient Mr. Zhang after EICU transfer of the present application.

[0022] Figure 5 Simplified flowchart for rehabilitation transition process of patient Mr. Zhang of the present application. DETAILED DESCRIPTION

[0023] The specific embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0024] In conjunction with the accompanying Figure 1 The EICU patient transfer rehabilitation staging evaluation system of the present application can perform real-time and dynamic evaluation and monitoring on the rehabilitation process of the patient after being transferred out of the EICU. The system mainly performs trend analysis on the continuous physiological parameters of the patient, and judges whether the patient has truly entered the rehabilitation feasible window according to the physiological changes in the rehabilitation process of the patient, and dynamically adjusts the starting point of the staging evaluation according to the drift law of the recovery trend. The system first performs real-time monitoring on the physiological data of the patient after being transferred out of the EICU, collects core physiological parameters such as heart rate, blood pressure, blood oxygen saturation, respiratory rate, body temperature, and uses statistical methods to analyze the change trend of these physiological parameters. Through data smoothing and trend identification in the continuous time window, the system can identify whether the physiological parameters appear a continuously stable state, so as to judge whether the patient has reached the rehabilitation feasible window. In order to avoid false positive stable states that may occur in traditional methods, the system sets a dynamic stable interval, and comprehensively considers the change amplitude and rate of each physiological index, to avoid misjudging a short-term physiological fluctuation as a stable state. This effectively eliminates the misjudgment caused by a single index or short-term fluctuation, and ensures the accuracy of the starting point of the rehabilitation evaluation. On this basis, the system analyzes the drift law of the recovery trend of the patient, that is, monitors the recovery curves of the physiological indexes of the patient within a certain time, and judges whether the starting point of the staging evaluation should be advanced or delayed according to the change law of these curves. If the system detects that the physiological indexes of the patient are stable in the preset recovery interval and reach the predetermined target, the system will automatically adjust the starting point of the staging evaluation through an algorithm, and then dynamically promote the evaluation of the rehabilitation stage of the patient. Through this dynamic adjustment mechanism, the real-time state of the patient's rehabilitation can be accurately identified and a timely response can be made, ensuring that the rehabilitation progress of the patient is highly consistent with the physiological state. In addition, the dynamic stable interval identification capability of the system can effectively eliminate false positive stable states, avoid the patient being incorrectly evaluated as having entered the rehabilitation window, and thus reduce unnecessary excessive intervention and optimize the allocation of rehabilitation resources.

[0025] The system dynamically divides and monitors the functional status of patients during the rehabilitation process, enabling precise rehabilitation assessment and stage transition. The system divides the functional status of patients into multiple functional clusters, including neurobehavioral, motor ability, cardiopulmonary load, metabolic homeostasis, and cognitive perception. Each functional cluster covers a series of corresponding underlying indicators, which can be obtained by continuously monitoring the physiological and behavioral data of patients. For each functional cluster, the system monitors the migration state of the underlying indicators based on the changes in the physiological data of patients during the rehabilitation process and assesses their trends. The system analyzes the dynamic changes of these underlying indicators in real time, combines the clinical records and feedback information of patients, and determines whether the rehabilitation status of each functional cluster has changed, especially whether the functional cluster has transitioned from intervention-dependent state to autonomous maintenance state. In the intervention-dependent state, the physiological and behavioral indicators of patients are highly dependent on external intervention measures (such as drug treatment, assistive devices, etc.), while the autonomous maintenance state indicates that patients can independently maintain or control their physiological and behavioral indicators without excessive external intervention. The system tracks the changes of the underlying indicators in each functional cluster in real time, uses multi-level analysis algorithms to perform multi-dimensional evaluation in terms of physiological indicator stability, recovery ability, and intervention response, and dynamically determines whether the autonomous maintenance state criteria are met by combining the cross-feedback mechanism of each cluster. If multiple functional clusters meet the autonomous maintenance conditions, the system automatically identifies that the rehabilitation stage of patients has transitioned and enters the next rehabilitation stage. Through this method, the system can accurately assess the recovery process of each functional cluster and dynamically adjust the rehabilitation plan and intervention scheme of patients according to the state changes of each functional cluster, ensuring that patients enter the next stage of rehabilitation at the most appropriate time. At the same time, the system avoids the subjective judgment errors and excessive intervention phenomena in traditional evaluation methods, thereby optimizing the rehabilitation path and treatment effect of patients.

[0026] By monitoring and recording the recovery rate gain of each type of intervention on the target functional cluster, the system dynamically assesses the patient's rehabilitation status and automatically corrects the current stage division results when the intervention effect decreases or reverses, forming a dynamic closed-loop logic between intervention, response and stage. After each intervention, the system monitors the patient's physiological state, especially the recovery rate of the target functional cluster. This assessment is done by comparing the change rate of the core indicators of the functional cluster before and after the intervention. If the recovery rate after the intervention improves significantly, the system considers the intervention effective and continues to track its effect. As the rehabilitation process progresses, the system monitors whether the intervention effect is decreasing or reversing in real time. When the system detects that the intervention effect is decreasing or reversing, it means that the patient's recovery process may be hindered or negatively affected, and the system will automatically trigger a correction mechanism to reverse the current stage division results and return the patient's rehabilitation stage to the previous more suitable stage, thus avoiding over-promoting the rehabilitation process. The reverse correction process ensures that the patient receives the best intervention at the right time and provides a more personalized and safe rehabilitation path. The system also has the ability to monitor the patient's low-intensity fluctuating physiological signals, including night-time respiratory micro-tremor, heart rate variability (HRV) linear shift and peripheral microcirculation abnormalities, and other minor physiological fluctuations. By setting a continuous fluctuation threshold and a response stability change relationship after intervention, the system can identify potential rehabilitation decline trends. When the patient's physiological signals show minor but continuous abnormal changes without obvious external intervention, the system will automatically detect these changes and mark them as potential rehabilitation decline risks. To this end, the system will adjust the intervention intensity, delay the stage transition or trigger a risk warning to ensure that the patient can be adjusted and supported in a timely manner during the rehabilitation process.

[0027] The accompanying drawings are incorporated in and constitute a part of this specification. Figure 2The trend analysis part of the continuous physiological parameters prioritizes the collection of physiological data from the patient during periods of no artificial intervention and excludes monitoring data from periods of high-frequency medical operations after the patient is transferred out of the EICU. The core of this process is to ensure the accuracy and stability of the data by monitoring the patient's physiological trend without external intervention, avoiding misleading rehabilitation assessments due to short-term fluctuations caused by medical operations. The system automatically identifies and labels periods of no intervention for the patient after being transferred out of the EICU, prioritizes physiological data collection during these periods, and excludes noise data that may be generated during periods of frequent medical operations, such as monitoring data during high-frequency intervention periods such as drug administration, cannulation, and bed transfer, thereby reducing the impact of human operation on rehabilitation assessment results and improving the reliability and accuracy of the assessment. In addition, the identification of the dynamic stable interval is performed before determining whether the patient has entered the rehabilitation feasible window by setting a buffer observation period. During the buffer observation period, the system only performs passive physiological monitoring and tracks the patient's physiological indicators in real time. If the fluctuation of any physiological indicator exceeds the preset range during this period, the system will delay the start time of the staging assessment until the patient's physiological indicators return to a stable state. The setting of this buffer observation period helps to avoid premature determination of the patient's entry into the rehabilitation window and avoid misjudgment caused by short-term fluctuations or temporary stability.

[0028] In this system, the patient's functional status is divided into multiple functional clusters, each representing a different physiological and behavioral dimension in the patient's rehabilitation process, such as neurobehavior, motor ability, and cardiopulmonary load. In the process of dividing functional clusters, the system sets trigger conditions for staging transitions, i.e., the intervention-dependent state of at least two functional clusters must simultaneously change to an autonomous maintenance state, and the indicators of each functional cluster must remain stable during the two consecutive assessment periods before determination. The system monitors whether each functional cluster shows a decrease in intervention dependence and gradually enters an autonomous maintenance state. Only when two or more functional clusters show stability in their core physiological indicators without or with minimal external intervention will the system trigger the patient's rehabilitation stage transition. This approach ensures the scientificity and accuracy of rehabilitation assessment by coordinating the judgment of multiple functional clusters, avoiding the influence of single functional cluster state on staging assessment. For the neurobehavioral cluster, the system requires that the state transition must be based on the patient's sustained compliance with autonomous behaviors. The system monitors the patient's sustained performance in maintaining wakefulness, active communication, and eye tracking, and combines consistent improvements in physiological signals to make judgments. For example, when the patient can maintain a wakeful state for a long time without external help, actively interact with the surrounding environment, or successfully complete eye tracking, and physiological signals such as heart rate and blood oxygen show consistent improvement, the system considers this as a sign of effective transition and confirms that the neurobehavioral cluster has reached an autonomous maintenance state, thereby supporting the patient's transition from the current rehabilitation stage to the next stage.

[0029] In this system, the subjective rehabilitation perception feedback indicators of the patient are given priority when the intervention measures assess the recovery rate gain of the target functional cluster. The system collects the patient's subjective feedback data, such as pain score, fatigue perception, motor ability assessment, etc., as the priority reference for recovery rate assessment. If the patient's subjective feedback shows that his rehabilitation state is not good, although the physiological indicators (such as heart rate, blood oxygen, blood pressure, etc.) show an improvement trend, the system will give priority to the patient's subjective feeling and maintain the current intervention state, delay the staging update, until the patient's subjective perception and physiological improvement trend are consistent, then stage transition can be performed. This mechanism ensures the unity of individual perception and clinical evaluation in the patient's rehabilitation process, and avoids excessive intervention or premature staging update due to relying solely on physiological indicators. In order to further ensure the scientificity and stability of the rehabilitation process, the system sets a maximum of two intervention scheme adjustments in each stage during the intervention response evaluation process. In each stage, if the intervention effect is not improved or the indicators do not fluctuate as expected, the system will try to adjust the intervention scheme, with a maximum of two adjustments. If after two intervention adjustments, the patient's physiological indicators do not improve significantly, or appear to be reversed (for example, physiological indicators worsen or new adverse reactions occur), the system will trigger the stage rollback mechanism. After triggering the rollback mechanism, the system will automatically generate an exception record, recording the patient's intervention history, evaluation results, adjustment process and possible cause analysis, for subsequent clinical tracking and adjustment, while ensuring timely adjustment of the patient's rehabilitation path and personalized intervention. This intervention rollback mechanism effectively avoids the rehabilitation path failure caused by inappropriate intervention, ensures that the patient can receive the most appropriate intervention in each stage, and ensures the accuracy and individualization of the rehabilitation evaluation process.

[0030] During the execution of the dynamic closed-loop logic, the medical staff is required to confirm each stage transition judgment, ensuring that the assessment and adjustment of the patient's rehabilitation stage have sufficient clinical basis and accuracy. In this system, first, the system generates a stage transition suggestion based on real-time monitoring of physiological parameters, behavioral indicators, subjective feedback, and other data, to determine whether the patient meets the conditions for entering the next stage of rehabilitation. This suggestion not only relies on the results of the system's automated algorithm, but the system also requires medical staff to manually review the transition suggestion to ensure that the system's suggestion meets the patient's specific rehabilitation needs and clinical performance. After the system generates the transition suggestion, it automatically pushes the suggestion to the medical staff's review end for the medical staff to evaluate in detail based on the patient's specific situation. The medical staff will check the patient's physiological indicators, intervention effects, and subjective perception data through the interface, and combine clinical experience to confirm whether to agree with the system's recommended transition result. If the medical staff confirms the transition suggestion and approves the patient to enter the next rehabilitation stage, the system will continue to evaluate the stages and implement the next stage of rehabilitation intervention; if the medical staff disagrees with the transition suggestion or believes that the current stage is not suitable for transition, the system will immediately terminate the transition operation and extend the evaluation period of the current stage. After extending the evaluation period, the system will continue to monitor the patient's indicators and generate a new transition suggestion during the new evaluation period until the medical staff confirms that the patient's condition meets the transition criteria. Through this dynamic closed-loop mechanism, the system not only ensures the accuracy of automated evaluation, but also enhances the flexibility of manual review and clinical judgment, effectively avoiding premature or delayed transition caused by algorithm misjudgment or individual differences in patients.

[0031] In combination with the accompanying Figure 3 , the system dynamically determines whether multiple functional clusters have transitioned from intervention-dependent state to self-maintenance state through the following continuous judgment mechanism. Specifically, the system monitors each functional cluster during the patient's rehabilitation process using four key conditions to determine whether it has the ability to maintain independently, thereby providing a basis for the patient's rehabilitation stage transition.

[0032] First, the time continuity and intervention response weakening condition (P1) requires that if the core indicators of the functional cluster remain stable or improve without new intervention or intervention intensity weakening for at least two consecutive assessment periods, the functional cluster is determined to have a preliminary self-maintenance ability. At this time, the system monitors and records the performance of the functional cluster without intervention or intervention weakening, and judges the trend through multiple period evaluation data to ensure that the patient maintains his function independently without external intervention. Second, the recovery response lag shortening condition (P2) requires that when the same intervention measures continue to act on the patient, the system compares the recovery response time before and after the intervention. If the patient's response time to the same intervention measures is shortened, or the patient shows spontaneous recovery behavior in advance under the same intervention, it means that the intervention measures are no longer the main driving force for maintaining the patient's recovery, but are in an auxiliary role, further indicating that the patient's recovery status has begun to shift to self-maintenance. Third, the functional cluster coordination decoupling condition (P3) means that when the functional cluster gradually breaks away from the supportive input of other functional clusters during rehabilitation and can maintain its core indicators independently, the system will determine that the functional cluster has completed decoupling and can be an independent rehabilitation unit for stage transition. For example, if a motor function cluster no longer relies on cardiopulmonary function support and can independently perform autonomous activities, the cluster can be determined to be self-maintained and can undergo rehabilitation stage transition. Finally, the active behavior replaces passive response condition (P4) requires the system to identify whether the patient exhibits active adjustment behavior, and the consistency of the persistence of these spontaneous behaviors and the improvement trend of their physiological indicators. When the patient spontaneously exhibits behaviors such as active turning over and regular breathing in the functional cluster, and the persistence of these behaviors is consistent with the improvement trend of the indicators, the system will preferentially determine that the patient enters an active recovery state and will be used as one of the criteria for determining whether to transition. Through the continuous evaluation of the above four conditions, the rehabilitation status of each functional cluster of the patient can be accurately identified, and it can be determined whether to enter the next rehabilitation stage.

[0033] The system is based on the trend changes of physiological indicators in multiple continuous evaluation cycles, combined with the intervention response ability of the patient, to judge whether the patient's functional cluster is transformed from intervention-dependent state to self-maintaining state, and then to provide basis for the transition of the rehabilitation stage. The system includes two main conditions in the execution process. First, the time continuity and intervention reaction weakening condition, in at least two consecutive evaluation cycles, the system monitors the core indicators of the target functional cluster, if the indicators do not fluctuate abnormally and can maintain a stable state under the condition of intervention intensity reduction, the system will mark the functional cluster as a low dependence state candidate. The basis of this judgment shows that although the intervention intensity is reduced, the patient's functional indicators are still stable, which meets the preliminary requirement of entering the self-maintaining state. Secondly, the recovery response lag shortening condition, when the same intensity intervention is implemented on the patient, the system compares the response time of the indicators before and after the intervention, if the patient's response time to the intervention is significantly shortened or the behavior is recovered in advance in the consecutive two evaluation cycles, it indicates that the patient is no longer dependent on external intervention to maintain the physiological state, and the system confirms that the functional cluster has entered the recovery response acceleration stage, thereby supporting the decision of the transition of the rehabilitation stage.

[0034] Based on the above two conditions, a new dynamic rehabilitation staging transition judgment method is adopted, which quantitatively judges whether the functional cluster is transformed from intervention-dependent state to self-maintaining state by comprehensively considering the evaluation cycle trend and the intervention reaction lag change. It includes: first, in the consecutive evaluation cycle, the system monitors whether the core indicators remain stable under the condition of intervention intensity reduction, and marks the functional cluster as a low dependence candidate state according to the change; second, under the same intensity intervention, the system judges whether the recovery response time of the indicators is significantly advanced, if so, it is identified as entering the response acceleration stage. In order to unify the trend quantization results of the two dimensions, the present application adopts the following state transformation function formula: Among them: The functional cluster transition score value (state transition index) is used to judge whether to enter the self-maintaining state; The number of consecutive evaluation cycles (for example: 2 or 3 cycles); The intervention frequency reduction rate in the first period, reflecting the degree of intervention weakening; The stability weight of the core indicators in the first period (the smaller the fluctuation range, the greater the value); The number of slight abnormalities of the indicators in the first period (used to suppress the false stability phenomenon); The non-zero balance factor avoids zero denominator (set to a very small constant); Response time reduction index, representing the ratio of the response time after intervention to the response time before intervention; Response time sensitivity adjustment coefficient (empirical adjustment value), controlling the degree of incentive to the trend of early recovery. The first part of the formula is the multi-cycle intervention-stability comprehensive score, which is based on the decrease in intervention frequency, the stability weight of the core indicator, and the number of abnormalities. If the intervention frequency decreases ( ), and the indicator fluctuates little ( ), and the number of abnormalities is low ( ), the score of this part will increase, indicating that the functional cluster gradually breaks away from intervention dependence and enters the autonomous maintenance state; the second part is the early reward factor of recovery response. If the recovery time after intervention is shortened ( ), it is multiplied by the adjustment coefficient to amplify the intention of transition, indicating that the recovery ability of the functional cluster is accelerated, promoting the transition of the rehabilitation process. If the state transition function exceeds the set transition threshold (determined by the system or experts), it is determined that the functional cluster has transformed from intervention-dependent state to autonomous maintenance state, and formally enters the rehabilitation transition phase.

[0035] Under the condition of functional cluster synergistic decoupling, the system determines whether the patient has changed from a dependent intervention state to an autonomous maintenance state by monitoring the index triggering sequence and dependent behavior between each functional cluster in real time. The system monitors the interdependence between different functional clusters of the patient (such as neurobehavior, motor ability, cardiopulmonary load, etc.), especially during the patient's rehabilitation process. The system tracks the physiological and behavioral indicators of each functional cluster to determine whether it can independently maintain the core physiological indicators without relying on the support input of other functional clusters. In this process, if a functional cluster can independently complete the index standard and maintain a stable state without the support of other functional clusters, the functional cluster will be identified by the system as having independent maintenance capability and is ready to enter the next stage of rehabilitation. This basis for judgment shows that in the patient's rehabilitation process, a functional cluster is no longer affected by other functional clusters and can self-regulate and maintain its physiological function. The system will trigger a staged transition decision based on this to promote the patient from a dependent intervention state to an autonomous maintenance state and enter the next rehabilitation stage. At the same time, the condition of active behavior replacing passive response further deepens the judgment of the patient's autonomous ability to recover. Under this condition, the system identifies the patient's autonomous adjustment behavior in the daily non-inductive environment to determine whether the patient has the ability to actively recover. The system tracks the patient's spontaneous actions in the daily life environment, such as active turning over, autonomous adjustment of body position, regular breathing, etc., and monitors whether these spontaneous behaviors are consistent with the physiological indicator change trend corresponding to the functional cluster. The system records the number of continuous cycles of the patient's spontaneous behavior in this process and compares it with the improvement trend of the corresponding physiological indicators as a judgment standard. Once the patient shows consistent autonomous adjustment behavior within a certain period and the behavior is consistent with the improvement trend of the physiological indicators, the system will determine that the patient has the active ability to transition to the next stage and further promote the patient to enter the next rehabilitation stage.

[0036] The system generates a functional cluster state transition flag after executing any of the judgment conditions P1 to P4, and records the triggering conditions, triggering time, and intervention records related to the functional cluster, for subsequent stage transition decision and rehabilitation path review. The system monitors and analyzes multiple functional clusters (such as neurobehavior, motor ability, and cardiopulmonary load) in real time to determine whether each functional cluster meets specific rehabilitation transition conditions (such as P1 to P4). Once a functional cluster meets these conditions, the system immediately generates a functional cluster state transition flag and automatically records the conditions, time nodes, and corresponding interventions and execution that triggered the state transition. These records will be used for subsequent review and adjustment of the rehabilitation path, ensuring that the patient's rehabilitation process is scientific, accurate, and flexible. If the patient's rehabilitation status is abnormal in subsequent evaluations, the system can quickly backtrack and analyze the triggering conditions and intervention history to adjust the rehabilitation plan in a timely manner. In addition, when identifying the P3 cooperative decoupling condition, if the system finds that a functional cluster has reached the decoupling standard, but other functional clusters are still in a high dependence state, the system will allow the decoupled functional cluster to enter the transition buffer period first, without immediately making a transition decision. The transition buffer period is set to ensure that the state change of the decoupled cluster is not affected by the dependence of other clusters, and to ensure that the independent maintenance ability of the cluster is fully verified. During the transition buffer period, the system will continue to monitor the physiological indicators and behavior of the functional cluster and other functional clusters, and avoid forced transition during this period, in order to more accurately judge the overall rehabilitation progress of the patient. This buffer period can effectively reduce the risk of misjudgment caused by the mutual dependence between multiple clusters, and ensure that the formal stage transition decision is made only after the independent functional cluster reaches a stable state.

[0037] The continuous judgment mechanism supports a combination judgment strategy to ensure that the functional cluster state is transformed into a quasi-transition cluster with sufficient reliability and accuracy. During the evaluation process, the system combines P1 (time continuity and intervention response weakening condition) and P2 (recovery response lag shortening condition). When both conditions are met and this state lasts for more than two evaluation periods, the system will judge that the functional cluster is a quasi-transition cluster and enable the final confirmation process before transition. During this process, the system continuously observes the functional cluster and records its recovery process in consecutive periods. Only when the long-term stability and gradual recovery criteria are met will the final transition confirmation be triggered. The final confirmation process requires manual review by medical personnel to ensure that the results of the automated evaluation match the overall condition of the patient and to ensure the scientificity and accuracy of the transition decision. On the other hand, when evaluating the P4 condition (active behavior replaces passive response condition), the system requires that the patient exhibit multiple autonomous behaviors and further confirms whether these autonomous behaviors are consistent with the improvement trend of the patient's core physiological indicators. The system monitors the patient's spontaneous behaviors in a daily non-inductive environment, such as autonomous turning over and active activities, and determines whether these behaviors are consistent with the improvement trend of core physiological indicators (such as heart rate, blood oxygen, etc.). If the patient has multiple autonomous behavior performances, but the core physiological indicators do not show an improvement trend consistent with these behaviors, the system will exclude these behaviors from the transition decision and not include them in the transition evaluation factors. This mechanism aims to ensure that only when the patient shows true autonomous recovery ability and their physiological state is consistent with autonomous behavior can the transition into the next stage be confirmed, thereby avoiding false positives due to occasional behavioral changes.

[0038] Embodiment 1: In conjunction with the accompanying Figure 4 Mr. Zhang was admitted to the hospital due to acute respiratory failure. After treatment and monitoring in the EICU, the patient's vital signs have stabilized, and is preparing to transfer out of the EICU for further rehabilitation treatment. At the time of transfer out of the EICU, the patient's main physiological indicators include heart rate (HR), oxygen saturation (SpO2), respiratory rate (RR), and blood pressure (BP), and there are mild neurological and behavioral disorders and limited motor function.

[0039] Within 24 hours of Mr. Zhang's transfer out of the EICU, the system first monitors the patient's physiological parameters in real time. During this period, the system prioritizes data collection during periods without human intervention, such as monitoring the patient's physiological state while resting quietly without external treatment measures. Assuming that during this period without human intervention, Mr. Zhang's heart rate (HR) is maintained at 72 beats per minute, oxygen saturation (SpO2) is 95%, respiratory rate (RR) is 18 beats per minute, and blood pressure (BP) is stable at 130 / 85 mmHg.

[0040] At the same time, the system automatically identifies physiological data generated during high-frequency medical procedures (e.g., medication administration, intubation, blood draw, etc.) and discards these data. Assuming that during the intubation procedure, Mr. Zhang's heart rate experiences a rapid fluctuation from 72 beats per minute to 120 beats per minute within a short period, this data will be flagged as interference data by the system and excluded from analysis to ensure that subsequent rehabilitation assessments rely only on stable data from periods without intervention.

[0041] During the initial assessment phase after the patient is transferred out of the EICU, the system analyzes the collected physiological data to determine whether the patient has entered a rehabilitation feasibility window. The system identifies the dynamic stable interval for each metric and sets a buffer observation period to further confirm the patient's stability. Assuming that the system analyzes Mr. Zhang's heart rate data and finds that the heart rate fluctuates between 60 and 80 beats per minute within a 24-hour period, and in terms of blood oxygen saturation, Mr. Zhang's SpO2 consistently maintains above 95%. Based on these observations, the system labels heart rate and blood oxygen saturation as stable, meeting the conditions for entering the rehabilitation window. However, the system does not immediately transition to the rehabilitation phase because a buffer observation period is set.

[0042] Assuming the buffer observation period is 48 hours, during which the system continues to monitor Mr. Zhang's physiological data and ensures that the fluctuation of each metric does not exceed the preset stable range. If during this period, any one of the metrics (e.g., heart rate or blood oxygen saturation) fluctuates beyond the set warning range (e.g., heart rate fluctuation exceeds 10 beats per minute or SpO2 drops below 92%), the system will delay the start time of the staging assessment and re-evaluate the patient's rehabilitation status.

[0043] At the end of the buffer observation period, Mr. Zhang's heart rate remains at 74 beats per minute, blood oxygen saturation stabilizes at 96%, respiratory rate is 18 times per minute, blood pressure remains at 130 / 85 mmHg, and there is no significant fluctuation. At this time, the system determines that Mr. Zhang's rehabilitation status is stable, meeting the conditions for entering the rehabilitation window, and the evaluation system begins to recommend entering the next stage of rehabilitation treatment.

[0044] During the last few hours of the buffer observation period, Mr. Zhang experiences a short-term heart rate fluctuation, with heart rate jumping from 74 beats per minute to 92 beats per minute, but quickly returning to the normal range. Although this fluctuation does not last long, the system will perform further data analysis to ensure that this fluctuation is not affected by any external intervention and does not represent potential risks to the patient's recovery. In this case, the system may choose to extend the buffer period and continue to observe the patient's physiological state until it is confirmed that the patient's recovery stability is unquestionable before making a transition decision.

[0045] The dynamic stability interval and buffer observation period effectively avoid inaccurate rehabilitation assessment when the patient experiences physiological fluctuations in the short term, ensuring patient safety and ensuring that the patient can enter the next rehabilitation stage at the appropriate time, avoiding premature or delayed rehabilitation intervention.

[0046] After initial stabilization, Mr. Zhang was transferred out of the EICU and began rehabilitation treatment. At different stages, the system gradually presents his rehabilitation path through functional cluster division, intervention response assessment, and manual review processes.

[0047] At the beginning of Mr. Zhang's rehabilitation process, the system divides his functional status into multiple functional clusters, mainly including neurobehavior, motor ability, cardiopulmonary load, metabolic self-stability, and cognitive perception. The system sets the trigger conditions for stage transition, requiring at least two functional clusters to change from intervention-dependent state to self-maintaining state, and these functional clusters to maintain stable indicators for two consecutive assessment periods. This judgment mechanism first relies on Mr. Zhang's physiological data for evaluation. In the two periods after transferring out of the EICU, Mr. Zhang's cardiopulmonary function and motor ability functional clusters showed good stability, with heart rate stabilized at 72 beats per minute and blood oxygen saturation maintained at 96%, and Mr. Zhang's active motor ability gradually recovered. Through continuous monitoring, the system identified that these two functional clusters have begun to maintain these core indicators independently without external intervention, so it judged that these two functional clusters have met the standard of self-maintenance and entered the tentative transition state. At the same time, the state transition of the neurobehavior cluster needs to be based on the sustained performance of the patient's autonomous behavior. The system monitors Mr. Zhang's wakefulness maintenance time and active communication, and within the 48-hour observation period, Mr. Zhang maintained a wakeful state and showed the ability to actively communicate, and his physiological indicators (such as heart rate, blood oxygen) were consistent with the trend of these autonomous behaviors. The system combined these data to confirm that the neurobehavior cluster also met the conditions for transition, marking that Mr. Zhang has changed from intervention-dependent state to self-maintaining state.

[0048] During Mr. Zhang's rehabilitation process, the system assesses the recovery rate gain from the intervention measures. Suppose that during Mr. Zhang's rehabilitation process, the system intervenes on his two functional clusters, motor ability and cardiopulmonary function, and monitors the changes in his core indicators. Through subjective rehabilitation perception feedback, the system finds that Mr. Zhang reports feeling less fatigue and pain during physical therapy, and in the cardiopulmonary load functional cluster, the patient's gait gradually recovers. When the system assesses these feedback data, the physiological data and subjective feedback are not entirely consistent. Although the physiological indicators (such as heart rate, blood oxygen) have improved, Mr. Zhang's subjective feedback shows that he still feels significant physical exertion. Based on this subjective feedback result, the system decides to maintain the current intervention state and delay the staging update until the patient's subjective feedback and physiological improvement trends tend to be consistent. This effectively avoids premature rehabilitation stage transition and ensures the patient's overall recovery.

[0049] The system further evaluates the intervention effect on Mr. Zhang, and allows at most two adjustments of the intervention plan within each rehabilitation stage. Suppose that during the physical therapy process in the second stage, Mr. Zhang's physiological indicators (such as gait, muscle strength) improve after lower limb rehabilitation training, but still do not reach the expected recovery level. The system makes the first adjustment of the intervention effect, increasing the intensity of rehabilitation training, but in the next evaluation period, although the intensity is increased, Mr. Zhang's gait recovery speed is still slow, and the physiological indicators do not show significant improvement, and even in some cases (such as in the morning), there is some slight pain and discomfort. The system triggers the second intervention adjustment according to this change, temporarily reduces the training intensity and modifies the intervention plan. After several adjustments, the system still does not see significant improvement, and the indicators show some reverse changes, so the system automatically triggers the stage rollback mechanism according to the preset rules and returns to the previous rehabilitation stage. At this time, the system will generate an exception record, recording all the adjustment process and the patient's recovery situation, and provide a detailed report for medical staff to review.

[0050] During the entire rehabilitation process, every phase transition judgment performed by the system requires manual confirmation by medical staff. When the system generates a transition suggestion based on the assessment data, it automatically pushes the suggestion to the medical staff's review end. During Mr. Zhang's rehabilitation process, when the system suggested that he transition from the motor ability functional cluster to the next phase, the system synchronously pushed the relevant rehabilitation data, feedback information, and phase transition suggestion to the medical staff for manual confirmation. The medical staff reviewed Mr. Zhang's rehabilitation situation in the past 48 hours through the review interface and compared the physiological data and subjective feedback. They believed that although most indicators met the standards, there was still some uncertainty, especially regarding Mr. Zhang's muscle strength recovery and activity tolerance. The medical staff ultimately did not confirm the transition suggestion, but decided to extend the assessment period and continue to observe Mr. Zhang's recovery to ensure that all indicators are stable and meet the standards before making further decisions. This dynamic closed-loop logic ensures the individualization and accuracy of the rehabilitation path through real-time data feedback, manual review, and system decision-making.

[0051] Example 2: In conjunction with the accompanying Figure 5 Based on Example 1, Mr. Zhang, a patient, was admitted to the EICU due to acute respiratory failure. After several days of treatment, the patient's vital signs gradually stabilized, and he was transferred out of the EICU to begin subsequent rehabilitation treatment. Mr. Zhang's main functional clusters include neurobehavior, motor ability, cardiopulmonary load, and metabolic self-stability. After being transferred out of the EICU, Mr. Zhang's rehabilitation assessment will be dynamically judged based on the following four core conditions (P1 to P4) to determine whether he transitions from an intervention-dependent state to a self-maintaining state.

[0052] P1: Time continuity and intervention response weakening condition: Mr. Zhang's cardiopulmonary load functional cluster and motor ability functional cluster first received rehabilitation intervention. After the first assessment period, Mr. Zhang's heart rate decreased from 88 beats per minute at rest to 80 beats per minute, and his blood oxygen saturation (SpO2) increased steadily from 94% to 97%. In the subsequent 24 hours, Mr. Zhang's core physiological indicators remained within the normal range, and the system detected that the patient's heart rate did not exhibit abnormal fluctuations, and the blood oxygen saturation remained above 95%. In addition, Mr. Zhang's motor ability gradually improved, and he was able to sit up independently and perform light exercise training. Since these indicators did not exhibit significant fluctuations and remained stable with reduced intervention intensity or no additional intervention, the system determined that the functional cluster had preliminary self-maintenance capabilities based on the P1 condition and marked it as a low-dependence candidate state.

[0053] P2: Recovery response lag shortening condition: In the next intervention period, Mr. Zhang continues to receive the same intensity of rehabilitation intervention, including exercise training and respiratory muscle training. The system monitors that after receiving the same intensity of intervention, Mr. Zhang's recovery time is significantly shortened. Specifically, Mr. Zhang's gait recovery time is shortened from an average of 2 minutes to 1.5 minutes, and when performing respiratory training, the respiratory rate is maintained at 20 times / minute, and the blood oxygen level is stable. Compared with the initial intervention, Mr. Zhang's reaction time to the intervention measures is advanced, and the spontaneous recovery behavior is to actively stand and walk after training. After system analysis, it is confirmed that Mr. Zhang's recovery reaction time is significantly shortened, indicating that the intervention is no longer the main driving force, but is in an auxiliary state, and the system determines that the function cluster enters the recovery reaction acceleration stage according to the P2 condition.

[0054] P3: Function cluster cooperative decoupling condition: As the rehabilitation process advances, Mr. Zhang shows a gradual decrease in dependence on other function clusters when performing cardiopulmonary training. For example, when performing gait training, Mr. Zhang's cardiopulmonary load function cluster gradually breaks away from the support of the motor ability cluster and can independently maintain a stable heart rate and blood oxygen level. The system detects that Mr. Zhang does not need to rely on excessive support of the cardiopulmonary function during gait recovery and can independently maintain his core indicators (such as heart rate, blood oxygen), which indicates that the cardiopulmonary load function cluster has completed cooperative decoupling with other function clusters. At this time, the system considers this function cluster to have independent maintenance ability and, according to the P3 condition, regards the cardiopulmonary load function cluster as an independent transition unit, supporting Mr. Zhang's rehabilitation stage transition.

[0055] P4: Active behavior replaces passive response condition: Finally, the system conducts a detailed assessment of Mr. Zhang's neurobehavioral cluster. Mr. Zhang gradually shows active behavior during rehabilitation, actively performing visual tracking, active communication, and sitting and standing behaviors without external intervention, and these behaviors are consistent with the improvement trend of physiological indicators (such as heart rate, blood oxygen). Specifically, Mr. Zhang can maintain wakefulness during active communication and complete a short conversation with medical staff, and the consistency of physiological indicators with these behaviors is manifested as stable heart rate and blood oxygen saturation maintained above 96%. After system analysis, it is judged that Mr. Zhang's neurobehavioral function cluster has changed from passive response to active recovery, meeting the P4 condition, so it is preferentially determined as an active recovery state and provides the basis for transition.

[0056] After the comprehensive judgment of the above four conditions, the system summarizes Mr. Zhang's various data. The specific data is as follows: Cardiopulmonary load function cluster: Heart rate stabilizes at 80 beats / minute, blood oxygen saturation is above 97%, and recovery time is shortened, indicating that the ability to maintain independent function has been achieved; Motor function cluster: Independent sitting up, gait recovery time shortened from 2 minutes to 1.5 minutes, and recovery time is earlier, entering the accelerated recovery phase; Neurobehavioral function cluster: Consistent self-awareness, active communication, and eye-tracking behavior, with stable physiological indicators; Metabolic homeostasis function cluster: Stable blood glucose, body temperature, and metabolic rate, requiring minimal external intervention.

[0057] Based on the above assessment results, the system believes that Mr. Zhang has successfully broken free from intervention dependence and has the ability to maintain his functional clusters independently, meeting the conditions for transitioning from intervention dependence to self-maintenance. Therefore, it is recommended that Mr. Zhang enter the next rehabilitation stage.

[0058] During Mr. Zhang's rehabilitation process, the system first monitored his core indicators, such as heart rate (HR) and oxygen saturation (SpO2). In the first assessment cycle after entering rehabilitation treatment, Mr. Zhang's heart rate decreased from 90 beats / minute to 80 beats / minute, and his oxygen saturation increased from 92% to 96%. During this period, the intervention intensity gradually decreased, and Mr. Zhang's ventilator support frequency also decreased from 6 hours per day to 3 hours. The system monitored fluctuations in indicators such as heart rate and oxygen saturation over two consecutive assessment cycles, finding that these core indicators did not fluctuate abnormally and remained stable or improved even with the reduced intervention intensity. Based on these conditions, the system calculated the intervention frequency reduction rate. The weighting is 30%, and the stability weighting of the heart rate during this cycle is... The number of anomalies is 0.95. A value of 1 indicates that one minor fluctuation occurred during the evaluation period. The system uses the formula: Calculate the functional cluster transition score ( The system determined that Mr. Zhang's cardiopulmonary function cluster had achieved preliminary self-sustaining ability. At this point, based on the P1 condition, the system marked Mr. Zhang's cardiopulmonary load function cluster as a low-dependency candidate state and prepared to continue tracking the next step of rehabilitation assessment.

[0059] In the next intervention period, the system continues to assess Mr. Zhang's rehabilitation progress. Assume that under the same intervention intensity, Mr. Zhang's cardiopulmonary load functional cluster shows a shortened recovery response time. For example, after completing a breathing training, Mr. Zhang's recovery time is shortened from 5 minutes to 3 minutes. The system monitors that Mr. Zhang's recovery time is significantly ahead, indicating that the patient responds more quickly and effectively to the same intensity of intervention. By comparing the response time before and after intervention in the two assessments, the system confirms the advance trend of recovery response, further indicating that intervention is no longer the main driving force, and Mr. Zhang begins to gradually show the ability of autonomous recovery. At this time, the system will confirm that the cardiopulmonary load functional cluster enters the recovery response acceleration stage according to the P2 condition.

[0060] To unify the trend quantification results of multiple dimensions, the system uses the state transition function defined earlier to make the final jump judgment. According to Mr. Zhang's performance under P1 and P2 conditions, the system collects the core indicators of cardiopulmonary load functional cluster and motor ability functional cluster, intervention intensity, and recovery response time changes, and calculates the transition score value.

[0061] Assume that in the next two periods, Mr. Zhang's heart rate and motor ability indicators continue to remain stable, the system calculates: ; ; ; (a small constant to avoid a zero denominator); (reflecting the magnitude of recovery response advance); (response time sensitive adjustment coefficient); Substitute the formula to calculate: In this calculation result, is greater than the transition threshold set by the system (assuming the threshold is set to 0.3), so the system decides to mark the cardiopulmonary load functional cluster as an autonomous maintenance state, and prepares to enter the next rehabilitation stage.

[0062] During Mr. Zhang's rehabilitation process, the system also conducts feedback evaluation on the intervention measures and dynamically adjusts them based on the patient's subjective perception and physiological data. Assuming that after entering the rehabilitation stage, Mr. Zhang reports feeling fatigue in recovery, and the physiological data (such as exercise endurance) do not continue to improve as expected. The system will evaluate the current intervention program through the set rollback mechanism. If the patient's indicators do not improve significantly or even reverse after two consecutive adjustments, the system will automatically trigger the rollback mechanism, record the anomaly, and provide data support for the next step of intervention adjustment. This rollback mechanism ensures that the patient will not have adverse reactions due to excessive intervention, and can flexibly adjust the rehabilitation path according to real-time data.

[0063] Each transition judgment of the system will be pushed to the medical staff end for review. When the system calculates the transition suggestion of Mr. Zhang from intervention-dependent state to self-maintenance state, the relevant data (including the core indicators of each functional cluster, subjective feedback, intervention adjustment, etc.) will be pushed to the medical staff for manual review. The medical staff reviews the patient's rehabilitation data and clinical feedback through the system, and comprehensively judges whether the patient meets the transition criteria. If the medical staff confirms the transition suggestion, the rehabilitation stage transition will continue; if not, the system will extend the evaluation period and continue to observe Mr. Zhang's rehabilitation progress to ensure that all data fully support the patient's rehabilitation transition.

[0064] During Mr. Zhang's rehabilitation process, the system real-time monitors the coordination and dependency relationship between each functional cluster, especially focusing on the relationship between the cardiopulmonary load and exercise capacity functional clusters. In the early stage of Mr. Zhang's rehabilitation treatment, the cardiopulmonary load functional cluster provides significant support to the exercise capacity functional cluster, and Mr. Zhang needs to rely on cardiopulmonary support (such as maintaining normal blood oxygen levels through oxygen therapy) when exercising. However, after a certain period of rehabilitation intervention, Mr. Zhang's cardiopulmonary load function gradually stabilizes, and when he performs lower limb rehabilitation training, his heart rate is maintained at 80 times / minute, and the blood oxygen saturation is stable at 96%, indicating that the cardiopulmonary load has no longer relies on the support of other functional clusters. At this time, the system analyzes and identifies that the cardiopulmonary load functional cluster has broken away from the supportive input of other functional clusters and can independently maintain its core physiological indicators. Based on this situation, the system identifies the cardiopulmonary load functional cluster as "having independent maintenance ability" and prepares to use it as an independent transition unit to enter the next stage of rehabilitation treatment.

[0065] In addition to the cardiopulmonary load functional cluster, Mr. Zhang's neurobehavioral functional cluster also showed significant recovery. During this process, the system further assessed Mr. Zhang's rehabilitation progress by monitoring his autonomous behavior changes. In the early stage, Mr. Zhang's neurobehavioral functional cluster relied on external intervention, such as visual tracking training and active sitting training guided by medical staff. However, as rehabilitation progressed, Mr. Zhang gradually showed more autonomous behavior. After the first evaluation period, Mr. Zhang autonomously completed 10 eye tracking training sessions in a daily non-inductive environment, and his heart rate remained stable at 75 beats per minute, with blood oxygen above 95%. These spontaneous behaviors and consistent improvement trends in physiological indicators indicated that Mr. Zhang no longer relied on external intervention and could autonomously regulate his neurobehavioral function. The system analyzed and identified Mr. Zhang's improvement in autonomous behavior, and marked it as "active ability with phase transition."

[0066] After each functional cluster of Mr. Zhang met the transition conditions, the system automatically generated a state transition label for each functional cluster. For example, when the cardiopulmonary load functional cluster and the neurobehavioral functional cluster successfully completed the transition from intervention dependence to autonomous maintenance state, the system marked these two functional clusters as "pseudo-transition clusters," and recorded the triggering conditions, triggering time, and corresponding intervention records for each transition. These records not only provide support for subsequent rehabilitation phase transitions, but also facilitate medical staff to review and adjust the patient's rehabilitation path. The system's design ensures that each transition decision has data basis and effectively avoids adverse consequences caused by premature or delayed transition.

[0067] During Mr. Zhang's rehabilitation process, although multiple functional clusters have met the transition conditions, when the cardiopulmonary load functional cluster and the neurobehavioral functional cluster transition, the system found that other functional clusters (such as the motor ability functional cluster) were still in a high dependence state and could not completely maintain their core indicators independently. Based on this finding, the system decided to prioritize the decoupled cardiopulmonary load functional cluster into the transition buffer period, rather than immediately transitioning. This buffer period is designed to ensure that the independent maintenance ability of the cardiopulmonary load functional cluster is fully verified and to reduce potential risks caused by other functional clusters still in a high dependence state. During this period, the system continues to monitor Mr. Zhang's overall rehabilitation to ensure that the coordination and independence between his functional clusters are in the best state before making the final transition decision.

[0068] During the rehabilitation process of Mr. Zhang, the system also adopts a combined judgment strategy. According to the comprehensive judgment of P1 and P2 conditions, the system requires that the intervention frequency decreases and the indicators are stable in at least two consecutive evaluation periods before the function cluster state is converted into a quasi-transition cluster. Specifically, the motor function cluster of Mr. Zhang decreased by 40% in intervention frequency in two evaluation periods, and the motor indicators (such as gait recovery time) remained stable. The system marked the motor function cluster as a quasi-transition cluster according to this condition and pushed the transition suggestion to the artificial review end at the same time. After reviewing Mr. Zhang's rehabilitation data, medical staff confirmed that his physiological indicators met the transition conditions, and his subjective perception feedback (such as fatigue perception and pain report) met the rehabilitation standards, and finally agreed to transition, and officially entered the next rehabilitation stage. If the medical staff does not confirm the transition suggestion, the system will extend the evaluation period and continue to observe Mr. Zhang's rehabilitation state to ensure that all indicators are stable and meet the transition requirements.

[0069] The above shows and describes the basic principles, main features and advantages of the present application. Those skilled in the art should understand that the present application is not limited to the above embodiments, and the above embodiments and descriptions in the specification are only to illustrate the principles of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. An EICU discharge patient staging rehabilitation assessment system, characterized by The application relates to a method for dynamically determining the rehabilitation stage of a patient. The method comprises the following steps: trend analysis is performed on continuous physiological parameters of the patient after the patient is transferred out of an EICU, a dynamic stable interval of the indicators is identified, it is judged whether the patient truly enters a rehabilitation feasible window, and the starting point of the staging evaluation is dynamically adjusted in advance or delayed based on the drift law of the recovery trend, so that the risk of misjudgment caused by false positive stable state is eliminated; The functional state of the patient is divided into multiple functional clusters, including five types of neurobehavior, motor ability, cardiopulmonary load, metabolic self-stability and cognitive perception, the migration state of the underlying indicators in each functional cluster is monitored, it is dynamically judged whether multiple clusters are transformed from intervention-dependent state to autonomous maintenance state, and the transformation basis of the rehabilitation stage of the patient is formed; The recovery rate gain of each type of intervention measure on the target functional cluster is recorded, and when the intervention effect is monitored to be reduced or reversed, the current stage division result is corrected in reverse, so that a dynamic closed-loop logic among intervention, response and stage is formed; high-frequency small-amplitude abnormal changes in low-intensity fluctuation physiological signals of the patient are detected, including night respiratory micro-tremor, HRV linear deviation and peripheral microcirculation abnormality, a continuous fluctuation threshold and a response stability change relationship after intervention are set, and a latent rehabilitation recession trend is identified.

2. The EICU step-down patient transitional rehabilitation assessment system of claim 1, wherein The trend analysis of the continuous physiological parameters comprises the following steps: data is preferentially collected in a period without artificial intervention after the patient is transferred out of the EICU, and the monitoring data of the high-frequency period of medical operation is eliminated; the dynamic stable interval is identified, and a buffer observation period is set before the rehabilitation feasible window is determined; passive physiological monitoring is performed in the buffer observation period, and if the fluctuation of any indicator exceeds a preset range, the starting time of the staging evaluation is delayed.

3. The EICU step-down patient transitional rehabilitation assessment system of claim 2, wherein In the process of dividing the functional state of the patient into multiple functional clusters, the trigger condition of the staging transition is set as the intervention-dependent state of at least two functional clusters being simultaneously transformed into the autonomous maintenance state, and the indicators are maintained stable in the two consecutive evaluation periods before the determination; the state transformation of the neurobehavior cluster must be based on the premise that the patient's autonomous behaviors including wakefulness maintenance, active communication or visual tracking are continuously up to the standard, and the physiological signals are consistently improved, and the effective transition criterion is used.

4. The EICU step-down patient transitional rehabilitation assessment system of claim 3, wherein In the evaluation of the recovery rate gain of the intervention measure on the target functional cluster, the subjective rehabilitation perception feedback indicators of the patient are preferentially referred to, if the subjective feedback is inconsistent with the physiological improvement trend, the current intervention state is maintained and the staging update is delayed; in the intervention response evaluation process, the system sets that the intervention scheme is allowed to be adjusted at most twice in the stage, if the indicators are not improved or are reversely changed after continuous adjustment, a stage rollback mechanism is triggered and an abnormal record is generated.

5. The EICU step-down patient transitional rehabilitation assessment system of claim 4, wherein In the execution process of the dynamic closed-loop logic, the judgment of the stage transition is required to be confirmed by medical staff, the system synchronously pushes the transition suggestion to the artificial audit end after the transition suggestion is generated, and the evaluation period is prolonged if the confirmation fails.

6. The EICU step-down patient transitional rehabilitation assessment system of claim 1, wherein The method for dynamically judging whether multiple clusters are transformed from intervention-dependent state to autonomous maintenance state is based on the following continuous judgment mechanism: P1, time continuity and intervention reaction weakening condition, the core indicators of the functional cluster are still maintained stable or improved under the condition that no new intervention or intervention intensity is weakened in a plurality of continuous evaluation periods, and it is determined that the preliminary autonomous maintenance ability is preliminarily determined. P2, recovery response lag shortening condition, when the same intervention is applied to the patient, identify the functional cluster to show a shortened recovery response time or an early spontaneous recovery behavior, indicating that the intervention is no longer the main driving force; P3, functional cluster cooperative decoupling condition, detect that the functional cluster gradually breaks away from the supportive input of other clusters during the staging process and independently maintains its core indicators, which is considered as a completed functional decoupling and can be used as an independent transition unit; P4, active behavior replacing passive response condition, identify that the patient has a spontaneous adjustment behavior in the functional cluster, and the behavior consistency is consistent with the improvement trend of the indicators, which is preferentially determined as an active recovery state and is one of the transition bases.

7. The EICU step-down patient transitional rehabilitation assessment system of claim 6, wherein The time continuity and intervention response weakening condition includes that the core indicators of the functional cluster do not have abnormal fluctuations in at least two consecutive evaluation periods, and still maintain a stable state under the condition that the intervention frequency decreases by more than a preset proportion, and the system marks the functional cluster as a low dependence state candidate accordingly; in the recovery response lag shortening condition, the system compares the response time of the indicators before and after the intervention after performing the same intensity of rehabilitation intervention, and confirms that the functional cluster enters the recovery response acceleration stage when the recovery response is detected to be ahead of the set threshold in two consecutive evaluations.

8. The EICU step-down patient transitional rehabilitation assessment system of claim 6, wherein In the functional cluster cooperative decoupling condition, the index trigger sequence and dependent behavior between each functional cluster are monitored in real time, and when a cluster completes the index standard and stable maintenance without the support of other clusters, it is identified as having independent maintenance ability; in the active behavior replacing passive response condition, the self-regulation action of the patient in the daily non-induction environment is identified, and the number of periods in which the trend of the corresponding indicators of the functional cluster is consistent is used as the confirmation standard to determine that the patient has the active ability of stage transition.

9. The EICU step-down patient transitional rehabilitation assessment system of claim 6, wherein The system generates a functional cluster state conversion mark after executing any one of the judgment conditions P1 to P4, and records the trigger condition, trigger time and corresponding intervention record, which is used for subsequent stage transition decision and path review; when the P3 cooperative decoupling condition is identified, if it is found that other functional clusters are still in a high dependence state, the decoupled cluster is allowed to enter the transition buffer period preferentially.

10. The EICU step-down patient transitional rehabilitation assessment system of claim 6, wherein The continuous judgment mechanism supports a combination judgment strategy, and when both P1 and P2 conditions are met and last for more than two evaluation periods, the functional cluster state is converted to a quasi-transition cluster, and the final confirmation process before transition is enabled; when the P4 condition is evaluated, if the patient has multiple autonomous behaviors but the consistency of the core indicator improvement trend is not consistent, the behavior is not counted in the transition judgment.

Images