Hospital management cost control evaluation method

Abstract

The invention relates to the technical field of hospital management, and discloses a hospital management cost control evaluation method, which comprises the following steps of: performing multi-granularity behavior slicing on process state data flow, and performing context sensing analysis on the sliced data to obtain a microscopic cost cause; performing diachronic analysis on the result state data stream to construct a dynamic mapping relation; according to the dynamic mapping relationship, analyzing the conduction relationship between the microcosmic cost motives, and constructing a cost dynamic conduction map by taking the conduction relationship as an edge and the microcosmic cost motives as nodes; based on the cost dynamic conduction atlas, strategy anti-fact deduction is carried out on the management intervention measures, and cost influence panoramic simulation is obtained; analyzing the deviation between the equilibrium state change of the whole network and a historical normal interval, and evaluating the cost structure robustness in combination with external market pressure test data; formulating a composite cost control strategy according to the cost influence panoramic simulation and the cost structure robustness; according to the invention, the accuracy of hospital management cost can be improved.

Description

Technical Field

[0001] This invention relates to the field of hospital management technology, and in particular to an evaluation method for cost control in hospital management. Background Technology

[0002] Existing cost control assessment methods for hospital management have significant shortcomings in data processing. They lack standardized multi-granularity breakdown and in-depth context-aware analysis of process data flows generated during hospital operations, making it difficult to accurately locate micro-cost drivers distributed in various business links. This extensive data processing leads to one-sided and delayed identification of cost influencing factors, resulting in insufficient basic data support for cost control assessment and directly restricting the overall efficiency and accuracy of the assessment work.

[0003] Traditional cost control assessment methods have failed to establish a dynamic correlation mechanism between micro-cost drivers and outcome data flow, and lack a systematic analysis of the transmission relationship between cost drivers. They cannot fully present the complete logic of cost formation and transmission. Existing methods have not conducted effective panoramic simulation and robustness verification for management intervention measures, resulting in cost control strategies that are difficult to adapt to complex business processes and external market fluctuations. The strategies lack pertinence and adaptability, making it difficult to achieve refined and sustainable management of hospital costs. Therefore, how to improve the precision of hospital management cost control has become an urgent problem to be solved. Summary of the Invention

[0004] This invention provides an evaluation method for cost control in hospital management to address the problems mentioned in the background section.

[0005] To achieve the above objectives, the present invention provides an evaluation method for cost control in hospital management, comprising:

[0006] S1. Perform multi-granularity behavioral slicing on the process state data stream of the target hospital, and perform context-aware analysis on the sliced ​​data to obtain the micro cost drivers of the target hospital;

[0007] S2. Based on the micro cost drivers, perform diachronic analysis on the outcome data stream of the target hospital to construct a dynamic mapping relationship between the micro cost drivers and the outcome data stream;

[0008] S3. Based on the dynamic mapping relationship, analyze the transmission relationship between the micro cost drivers, and construct the cost dynamic transmission map of the target hospital by taking the transmission relationship as the edge and the micro cost drivers as the node.

[0009] S4. Based on the cost dynamic transmission map, perform counterfactual analysis of the management intervention measures for the target hospital to obtain a panoramic simulation of the cost impact on the target hospital;

[0010] S5. Analyze the deviation between the overall network equilibrium state change caused by the management intervention measures and the historical normal range of the target hospital, and evaluate the cost structure robustness of the target hospital under the management intervention measures by combining the external market stress test data of the target hospital.

[0011] S6. Based on the panoramic simulation of cost impact and the robustness of the cost structure, formulate a composite cost control strategy for the target hospital.

[0012] In a preferred embodiment, the step of performing multi-granularity behavioral slicing on the process-state data stream of the target hospital and performing context-aware analysis on the sliced ​​data to obtain the micro-cost drivers of the target hospital includes:

[0013] Collect the medical order execution trajectory, equipment usage logs, and personnel movement records of the target hospital to obtain the process state data stream of the target hospital;

[0014] The process-state data stream is aligned with a unified spatiotemporal reference to obtain the synchronous process-state data of the target hospital;

[0015] According to the time window in which the business occurs in the target hospital, the synchronous process state data is initially segmented to obtain the primary data slice of the target hospital;

[0016] The primary data slice is further refined to obtain the secondary data slice of the target hospital;

[0017] Pattern matching is performed on the secondary data slices, and causal reasoning is performed on the matched patterns to obtain the micro cost drivers of the target hospital.

[0018] In a preferred embodiment, the step of performing diachronic analysis on the outcome-state data stream of the target hospital based on the micro-cost drivers to construct a dynamic mapping relationship between the micro-cost drivers and the outcome-state data stream includes:

[0019] Obtain the financial settlement records and material consumption list of the target hospital to obtain the result state data stream of the target hospital;

[0020] Time alignment is performed on the resulting state data stream and the micro cost drivers to obtain the dual data set of the target hospital;

[0021] Cooperative fluctuation pattern analysis was performed on the dual data set to obtain the lag correlation and pattern matching degree of the dual data set.

[0022] The impact intensity of the micro cost drivers is determined based on the lag correlation and pattern matching degree.

[0023] The intensity of the influence is tracked over time to obtain the variation pattern of the influence intensity;

[0024] Based on the intensity of the influence and the pattern of change, a correlation analysis is performed on the micro-cost drivers and the result state data stream to obtain the dynamic mapping relationship between the micro-cost drivers and the result state data stream.

[0025] In a preferred embodiment, the step of analyzing the transmission relationship between the micro-cost drivers based on the dynamic mapping relationship, and constructing a cost dynamic transmission map of the target hospital by using the transmission relationship as edges and the micro-cost drivers as nodes, includes:

[0026] Based on the dynamic mapping relationship, the micro cost drivers with the same time sequence are selected to obtain the target hospital's driver pairs to be analyzed;

[0027] Perform synergistic fluctuation analysis on the pairs of drivers to be analyzed to identify the potential transmission links of the micro cost drivers;

[0028] The potential transmission links are globally networked and verified, and the verified transmission links are reverse-constrained according to the dynamic mapping relationship to obtain the transmission relationship between the micro cost drivers.

[0029] The transmission relationship is treated as a directed edge, and the micro cost driver is treated as a node;

[0030] Based on the business logic of the target hospital, the directed edges and the nodes are connected to construct a dynamic cost transmission graph for the target hospital.

[0031] In a preferred embodiment, the step of performing synergistic fluctuation analysis on the pair of drivers to be analyzed to identify potential transmission links of the micro-cost drivers includes:

[0032] Frequency statistics are performed on the pairs of drivers to be analyzed to obtain the frequency index sequence of the micro cost drivers;

[0033] The frequency index sequence is decomposed into a trend term to obtain the fluctuation component of the micro cost driver.

[0034] By performing time-lag estimation on the fluctuation components, the lead-lag relationship and time difference of the micro cost drivers are obtained;

[0035] Based on the aforementioned lead-lag relationship, a state transition analysis is performed on the pair of drivers to be analyzed to obtain the fluctuation pattern transformation law of the micro cost drivers.

[0036] By combining the aforementioned lead-lag relationship, the aforementioned time difference, and the aforementioned fluctuation pattern transformation law, and by quantitatively evaluating the causal influence direction and intensity of the driver pair to be analyzed, the potential transmission links of the aforementioned micro cost drivers are obtained.

[0037] In a preferred embodiment, the step of performing counterfactual analysis of management intervention measures for the target hospital based on the cost dynamic transmission map to obtain a panoramic simulation of the cost impact on the target hospital includes:

[0038] The management intervention measures of the target hospital are encoded into intervention instructions for the target hospital;

[0039] The intervention command is mapped to the initial state adjustment amount of the target node in the cost dynamic transmission map;

[0040] Starting from the target node, traverse the directed edges of the cost dynamic transmission graph to determine the transmission path and transmission sequence of the initial state adjustment amount, and obtain the state change transmission sequence of the target hospital;

[0041] Based on the state change propagation sequence, the downstream nodes of the target node are iteratively updated, and the cyclic feedback path generated during the iterative update process is tracked in real time.

[0042] According to the cyclic feedback path, adjust the state values ​​of relevant nodes in the cost dynamic transmission graph until the state changes of the relevant nodes tend to stabilize;

[0043] By integrating the temporal evolution records of the relevant nodes and the steady-state distribution of the nodes after stabilization, a panoramic simulation of the cost impact on the target hospital is obtained.

[0044] In a preferred embodiment, the step of analyzing the deviation between the network-wide equilibrium change caused by the management intervention and the historical normal range of the target hospital, combined with the external market stress test data of the target hospital, to assess the cost structure robustness of the target hospital under the management intervention includes:

[0045] From the panoramic simulation of the cost impact of the target hospital, extract the node stability values ​​of the cost dynamic transmission map in the target hospital to obtain the current equilibrium vector of the target hospital;

[0046] By comparing and analyzing the current equilibrium vector with the historical normal range of the target hospital, the degree and direction of the state deviation of the target hospital can be obtained.

[0047] Based on the cost dynamic transmission map, a panoramic mapping is performed on the degree of state deviation and the direction of deviation to obtain the multidimensional deviation spectrum of the target hospital;

[0048] The external market stress test data of the target hospital is analyzed and mapped to the cost dynamic transmission map to obtain the standardized stress impact scenario of the target hospital.

[0049] Applying the standardized stress impact scenario to the current equilibrium vector yields a multi-scenario stress test response sequence for the target hospital;

[0050] The cost structure robustness of the target hospital is determined based on the multidimensional deviation spectrum and the multi-scenario stress test response sequence.

[0051] In a preferred embodiment, determining the cost structure robustness of the target hospital based on the multidimensional deviation spectrum and the multi-scenario stress test response sequence includes:

[0052] The standard deviation of the multidimensional deviation spectrum is obtained by performing scalar convergence on the multidimensional deviation spectrum.

[0053] The stress scenarios in the multi-scenario stress test response sequence are quantified using resilience parameters to obtain the scenario elasticity ratio of the multi-scenario stress test response sequence.

[0054] The robustness of the initial cost structure of the target hospital is calculated based on the standard deviation value and the scenario elasticity ratio.

[0055] The robustness of the initial cost structure is verified for consistency to obtain the robustness of the target hospital's cost structure.

[0056] In a preferred embodiment, the formula for calculating the robustness of the initial cost structure is:

[0057] ;

[0058] in, This indicates the robustness of the initial cost structure. This represents the total number of nodes in the multidimensional deviation spectrum. Indicates the first The standard deviation value of each node, This represents the total number of the aforementioned stress scenarios. Indicates the first The scenario elasticity ratio of the aforementioned stress scenarios, Represents the natural constant. This represents the natural logarithm function.

[0059] In a preferred embodiment, the step of formulating a composite cost control strategy for the target hospital based on the cost impact panoramic simulation and the cost structure robustness includes:

[0060] By performing inverse analysis of the strategy impact on the panoramic simulation of cost impact, the core control nodes of the target hospital can be obtained;

[0061] Based on the cost structure robustness assessment results and the core control nodes, a set of candidate action nodes for the target hospital is obtained;

[0062] The candidate action node set is deconstructed to obtain the basic strategy atoms of the target hospital;

[0063] Based on the cost dynamic transmission map of the target hospital, the basic strategy atoms are combined and spliced ​​to obtain the composite strategy scheme of the target hospital;

[0064] The composite strategy scheme is optimized through multi-objective strategy search to obtain the composite cost control strategy for the target hospital.

[0065] Compared with the prior art, the present invention has the following beneficial effects:

[0066] 1. This invention performs multi-granularity behavioral slicing and context-aware analysis on the process-state data flow of the target hospital to accurately extract micro-cost drivers. Then, it performs time-based analysis to construct a dynamic mapping relationship between micro-cost drivers and outcome-state data flow, thereby building a complete cost dynamic transmission map. This makes the capture and correlation analysis of cost influencing factors more systematic and accurate, significantly improving the overall efficiency of hospital management cost control assessment, and making the sorting and mining of cost-related core information more in-depth and comprehensive.

[0067] 2. This invention uses a cost dynamic transmission map to conduct counterfactual analysis of management intervention strategies, enabling a panoramic simulation of cost impacts. Simultaneously, it combines external market stress test data to assess the robustness of the cost structure, providing scientific and comprehensive support for the formulation of complex cost control strategies. This effectively enhances the adaptability and feasibility of cost control strategies, helps hospitals accurately control key cost management nodes, ensures the stability and optimization of the cost structure, and further improves the actual effectiveness of cost control. Attached Figure Description

[0068] Figure 1 A flowchart illustrating a method for evaluating cost control in hospital management, provided as an embodiment of the present invention;

[0069] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0070] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0071] This application provides a method for evaluating cost control in hospital management. The executing entity of this method includes, but is not limited to, at least one of the following electronic devices: a server, a terminal, or any other electronic device configured to execute the method provided in this application. In other words, the method for evaluating cost control in hospital management can be executed by software or hardware installed on a terminal device or a server-side device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cluster of cloud servers. The server can be a standalone server or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0072] Reference Figure 1 The diagram shown is a flowchart illustrating a method for evaluating cost control in hospital management according to an embodiment of the present invention. In this embodiment, the method for evaluating cost control in hospital management includes:

[0073] S1. Perform multi-granularity behavioral slicing on the process state data stream of the target hospital, and perform context-aware analysis on the sliced ​​data to obtain the micro cost drivers of the target hospital;

[0074] In this embodiment of the invention, the step of performing multi-granularity behavioral slicing on the process-state data stream of the target hospital and performing context-aware analysis on the sliced ​​data to obtain the micro-cost drivers of the target hospital includes:

[0075] Collect the medical order execution trajectory, equipment usage logs, and personnel movement records of the target hospital to obtain the process state data stream of the target hospital;

[0076] The process-state data stream is aligned with a unified spatiotemporal reference to obtain the synchronous process-state data of the target hospital;

[0077] According to the time window in which the business occurs in the target hospital, the synchronous process state data is initially segmented to obtain the primary data slice of the target hospital;

[0078] The primary data slice is further refined to obtain the secondary data slice of the target hospital;

[0079] Pattern matching is performed on the secondary data slices, and causal reasoning is performed on the matched patterns to obtain the micro cost drivers of the target hospital.

[0080] When collecting medical order execution trajectories, equipment usage logs, and personnel movement records from the target hospital, complete medical order execution-related information, such as the time of medical order issuance, the executing department, the executing personnel, and the completion status, is extracted from the hospital's medical order management system. Equipment usage-related data, such as equipment startup time, runtime, operators, and usage scenarios, are collected through the equipment's built-in operation record module. Personnel movement-related information, such as the movement paths, dwell time, and frequency of round trips of medical staff and related personnel in different areas, is obtained with the help of the hospital's internal positioning system. The extracted medical order execution trajectories, collected equipment usage logs, and obtained personnel movement records are integrated and summarized to form a process-state data stream for the target hospital.

[0081] When aligning the process data stream with a unified spatiotemporal reference, the time reference adopts the hospital's unified standard time format. All time-related information in the medical order execution trajectory, equipment usage log, and personnel movement records is checked one by one. Time data in different formats are all adjusted to this standard time to ensure that the timestamps of various types of data are consistent. The spatial reference uses the clearly defined physical areas of the hospital as a unified standard to determine the specific physical location corresponding to each data, such as a ward, examination room, or equipment storage area. Through the calibration of references in both time and space dimensions, the unified spatiotemporal reference alignment of the process data stream is completed, and the synchronous process data of the target hospital is obtained.

[0082] When initially segmenting synchronous process data according to the time windows in which business activities occur in the target hospital, the division criteria for time windows are first determined based on the cycle and frequency of routine business activities in the hospital. For example, outpatient business is set with a time window every two hours, and inpatient nursing business is set with a time window every four hours. The start and end times of each time window are clearly defined. Then, synchronous process data is filtered one by one, and all business-related data that occur within the same time window are grouped together to complete the initial segmentation operation of synchronous process data and obtain the primary data slices of the target hospital.

[0083] When refining the primary data slices, each primary data slice is first divided into three categories according to the business type it corresponds to: data related to medical order execution, data related to equipment usage, and data related to personnel movement. Then, within each category, it is further subdivided according to specific business processes. For example, data related to medical order execution is divided into data related to medical order review, drug dispensing, and treatment implementation; data related to equipment usage is divided into data related to equipment preparation, equipment operation, and equipment maintenance; and data related to personnel movement is divided into data related to entering the area, carrying out work, and leaving the area. Through this hierarchical decomposition operation, secondary data slices of the target hospital are obtained.

[0084] When performing pattern matching on secondary data slices, the data characteristics of each secondary data slice are first analyzed, including the corresponding business processes, time distribution, and associated objects. Then, all secondary data slices are classified according to these data characteristics. Secondary data slices with the same business processes, similar time distributions, and the same associated personnel or equipment are grouped into one category to complete pattern matching. After that, for each matched pattern, the generation order and interrelationships between various types of data are analyzed, and the mutual influence between data is analyzed. For example, it is clear that the generation of equipment usage data is to support the execution of medical orders, and the changes in personnel movement data are closely related to equipment usage and the execution of medical orders. Through such analysis and deduction, the specific factors affecting costs are identified, and the micro cost drivers of the target hospital are obtained.

[0085] The beneficial effects are that by comprehensively collecting relevant data on the core business of hospitals and performing standardized spatiotemporal alignment, hierarchical segmentation, and refined decomposition, followed by precise pattern matching and rigorous causal reasoning, it is possible to systematically and completely uncover the micro-cost drivers in the cost formation process of the target hospital. This ensures that the obtained micro-cost drivers truly reflect the actual business situation, laying a solid data foundation for subsequent work such as constructing a dynamic mapping relationship between micro-cost drivers and outcome data streams and building a dynamic cost transmission map, thus ensuring the smooth progress of hospital cost control assessment.

[0086] S2. Based on the micro cost drivers, perform diachronic analysis on the outcome data stream of the target hospital to construct a dynamic mapping relationship between the micro cost drivers and the outcome data stream;

[0087] In this embodiment of the invention, the step of performing diachronic analysis on the outcome data stream of the target hospital based on the micro-cost drivers to construct a dynamic mapping relationship between the micro-cost drivers and the outcome data stream includes:

[0088] Obtain the financial settlement records and material consumption list of the target hospital to obtain the result state data stream of the target hospital;

[0089] Time alignment is performed on the resulting state data stream and the micro cost drivers to obtain the dual data set of the target hospital;

[0090] Cooperative fluctuation pattern analysis was performed on the dual data set to obtain the lag correlation and pattern matching degree of the dual data set.

[0091] The impact intensity of the micro cost drivers is determined based on the lag correlation and pattern matching degree.

[0092] The intensity of the influence is tracked over time to obtain the variation pattern of the influence intensity;

[0093] Based on the intensity of the influence and the pattern of change, a correlation analysis is performed on the micro-cost drivers and the result state data stream to obtain the dynamic mapping relationship between the micro-cost drivers and the result state data stream.

[0094] When obtaining the target hospital's financial settlement records and material consumption list, all financial settlement-related information, including outpatient billing records, inpatient settlement records, drug sales settlement records, and examination and testing item billing records, are extracted from the financial system corresponding to the hospital's financial management system. This includes information such as billing amount, settlement time, billing item name and code, and corresponding patient identification. From the hospital's material management system, consumption data for various materials such as medical consumables, drugs, medical equipment accessories, and office supplies are collected to form a material consumption list. The list includes detailed information such as material name, specifications, quantity consumed, consuming department, consumption time, receiving personnel, and material code. The collected complete financial settlement records are integrated with the comprehensive material consumption list, and duplicate, erroneous, and invalid data are removed to obtain the target hospital's result data stream.

[0095] When aligning the outcome data stream and micro-cost drivers with time, the hospital's unified standard time format is used as the time alignment benchmark. The actual occurrence time of each record in the outcome data stream is extracted one by one, and the generation time or effect time of each micro-cost driver is extracted at the same time. The time of each record in the outcome data stream is compared with the time of the micro-cost driver one by one. The outcome data records with completely consistent time or time intervals within 10 minutes are bound and paired with the corresponding micro-cost drivers to ensure that each pair of paired data has a direct correlation in the time dimension, and finally the dual data group of the target hospital is obtained.

[0096] When performing cooperative fluctuation pattern analysis on the dual data sets, the changes in micro cost drivers in each dual data set are sequentially analyzed in chronological order, such as increases or decreases in frequency, expansion or contraction of the scope of influence, etc. At the same time, the changes in related data in the result state data stream are recorded, such as increases or decreases in amount, increases or decreases in consumption quantity, etc. The time difference between the changes in micro cost drivers and the changes in corresponding data in the result state data stream is observed to clarify the lag correlation between the two. The change patterns of micro cost drivers and corresponding data in the result state data stream are compared. If they both show the same or similar change characteristics such as continuous growth, synchronous fluctuation, and initial increase followed by decrease, the total duration and frequency percentage of the same change pattern are statistically analyzed to obtain the pattern matching degree of the dual data sets.

[0097] When determining the impact intensity of micro-cost drivers based on lag correlation and pattern matching degree, the correspondence between lag correlation and impact intensity is first clarified: the shorter the lag time, the more timely the impact of the micro-cost driver on the resulting data stream, and the higher the corresponding impact intensity. Then, the correspondence between pattern matching degree and impact intensity is clarified: the higher the proportion of pattern matching degree, the closer the correlation between the two, and the higher the corresponding impact intensity. Combining the specific numerical performance of these two indicators, a clear impact intensity level is assigned to each micro-cost driver through comprehensive judgment, with levels divided from high to low, thereby determining the impact intensity of the micro-cost driver.

[0098] When tracking the impact intensity over time, the historical operating time of the target hospital is divided into multiple consecutive and equal-length time periods, such as one month per period. The impact intensity level of each micro cost driver within each time period is recorded. The changes in the impact intensity level of the same micro cost driver in different time periods are compared. The specific time periods in which the impact intensity level increases, decreases, or remains unchanged are analyzed. At the same time, the hospital's business operations during the time period, such as changes in business volume and departmental operational adjustments, are combined to summarize the changing trend of the impact intensity of micro cost drivers over time, thus obtaining the changing pattern of the impact intensity.

[0099] When performing correlation analysis on micro-cost drivers and outcome data streams based on the intensity and variation patterns of their influence, the intensity level and variation pattern of each micro-cost driver are matched one-to-one with the corresponding specific data records in the outcome data stream. This clarifies which data items in the outcome data stream will change and the specific range of change caused by each micro-cost driver at different intensity levels. Simultaneously, based on the variation patterns, the initiation time, duration, and decay trend of the influence of the micro-cost driver on the outcome data stream at different time stages are determined. The system systematically sorts out the dynamic correlation logic between micro-cost drivers and outcome data streams, thereby obtaining the dynamic mapping relationship between the micro-cost drivers and outcome data streams.

[0100] The beneficial effects are that by systematically collecting complete outcome data and accurately aligning it with micro-cost drivers in time, and through detailed analysis of collaborative fluctuation patterns, clear determination of impact intensity, and continuous historical tracking, the final constructed dynamic mapping relationship can comprehensively and accurately reflect the correlation characteristics and dynamic change trends between micro-cost drivers and outcome data streams. This provides a solid and reliable foundation for subsequent analysis of the transmission relationship between micro-cost drivers and the construction of a cost dynamic transmission map, effectively ensuring the accuracy and systematic advancement of hospital cost control assessment.

[0101] S3. Based on the dynamic mapping relationship, analyze the transmission relationship between the micro cost drivers, and construct the cost dynamic transmission map of the target hospital by taking the transmission relationship as the edge and the micro cost drivers as the node.

[0102] In this embodiment of the invention, the step of analyzing the transmission relationship between the micro-cost drivers based on the dynamic mapping relationship, and constructing a cost dynamic transmission map of the target hospital by using the transmission relationship as edges and the micro-cost drivers as nodes, includes:

[0103] Based on the dynamic mapping relationship, the micro cost drivers with the same time sequence are selected to obtain the target hospital's driver pairs to be analyzed;

[0104] Perform synergistic fluctuation analysis on the pairs of drivers to be analyzed to identify the potential transmission links of the micro cost drivers;

[0105] The potential transmission links are globally networked and verified, and the verified transmission links are reverse-constrained according to the dynamic mapping relationship to obtain the transmission relationship between the micro cost drivers.

[0106] The transmission relationship is treated as a directed edge, and the micro cost driver is treated as a node;

[0107] Based on the business logic of the target hospital, the directed edges and the nodes are connected to construct a dynamic cost transmission graph for the target hospital.

[0108] The step of performing coordinated fluctuation analysis on the pair of drivers to be analyzed to identify the potential transmission links of the micro-cost drivers includes:

[0109] Frequency statistics are performed on the pairs of drivers to be analyzed to obtain the frequency index sequence of the micro cost drivers;

[0110] The frequency index sequence is decomposed into a trend term to obtain the fluctuation component of the micro cost driver.

[0111] By performing time-lag estimation on the fluctuation components, the lead-lag relationship and time difference of the micro cost drivers are obtained;

[0112] Based on the aforementioned lead-lag relationship, a state transition analysis is performed on the pair of drivers to be analyzed to obtain the fluctuation pattern transformation law of the micro cost drivers.

[0113] By combining the aforementioned lead-lag relationship, the aforementioned time difference, and the aforementioned fluctuation pattern transformation law, and by quantitatively evaluating the causal influence direction and intensity of the driver pair to be analyzed, the potential transmission links of the aforementioned micro cost drivers are obtained.

[0114] When filtering micro-cost drivers with the same time sequence based on the dynamic mapping relationship, first extract the timestamp corresponding to each micro-cost driver in the dynamic mapping relationship. This timestamp is consistent with the time when the micro-cost driver affects the result state data stream. Then, all micro-cost drivers are classified and organized according to their timestamps. Micro-cost drivers with the same timestamp are paired one by one. Every two successfully paired micro-cost drivers form a group, and finally, the driver pairs to be analyzed for the target hospital are formed.

[0115] When performing frequency statistics on the pairs of factors to be analyzed, the natural day is used as a fixed time period. The actual number of times the two micro cost factors in each pair of factors to be analyzed occur in each time period is recorded. According to the order of the time periods, the occurrence frequency of each micro cost factor is arranged in sequence to form a continuous sequence of data, thus obtaining the frequency index sequence of the micro cost factors.

[0116] When performing trend term decomposition on the frequency index sequence, first observe the overall trend of the frequency index sequence over a longer period of time, and distinguish the long-term change parts in the sequence that are continuously rising, continuously falling, or remaining stable. These parts are the trend terms. Then, by stripping away this long-term trend term, the short-term change parts in the sequence that fluctuate around the trend term are retained, and the fluctuation components of the micro cost drivers are obtained.

[0117] When estimating the time lag of the fluctuation components, the change curves of the fluctuation components of the two micro cost drivers in the analysis driver pair are plotted one by one. The order of the occurrence of the fluctuation start points in the two curves is compared to determine that the micro cost driver that fluctuates first is the leading one, and the one that fluctuates later is the lagging one, thus clarifying the leading-lagging relationship between the two. At the same time, the time interval between the fluctuation start point of the leading one and the fluctuation start point of the lagging one is calculated to obtain the time difference of the micro cost driver.

[0118] When performing state transition analysis on the driver pair to be analyzed based on the aforementioned lead-lag relationship, three states of the fluctuation component are first defined: rising state, stable state, and falling state. The state changes of the leading micro cost driver are continuously tracked, and the state response of the lagging micro cost driver after the leading state changes is recorded. The fixed way in which the lagging state follows the leading state change is summarized, and the fluctuation pattern transformation law of the micro cost driver is obtained.

[0119] When combining the aforementioned lead-lag relationship, time difference, and fluctuation pattern transformation law, the lead-lag relationship clarifies the direction of the causal influence of the driver to be analyzed, i.e., from the leading side to the lagging side; the time difference judges the speed of influence transmission, the shorter the time difference, the faster the influence transmission; the fluctuation pattern transformation law verifies the stability of the influence correlation, the more fixed the transformation law, the more stable the correlation; and then, the percentage of times the two fluctuate together within the same time period is statistically analyzed to quantify the intensity of the causal influence. By integrating the information on direction, transmission speed, stability, and intensity, the potential transmission link of the aforementioned micro cost driver is obtained.

[0120] When performing global network-based verification on the potential transmission links, all potential transmission links are integrated to form a preliminary network structure. The structure is then checked for contradictory transmission directions, broken transmission paths, or transmission connections without actual business relevance. Contradictory, broken, and unrelated links are eliminated, while logically coherent transmission links are retained. Then, reverse constraints are applied based on the dynamic mapping relationship. Each verified transmission link is compared to verify whether the impact logic of the micro-cost drivers in the link on the resulting data flow is consistent with the dynamic mapping relationship. Transmission links that do not conform to the logic are adjusted, and finally, the transmission relationship between the micro-cost drivers is obtained.

[0121] When the transmission relationship is used as a directed edge and the micro-cost driver is used as a node, each micro-cost driver is treated as an independent node, and the specific content of the micro-cost driver is clearly marked in the node; each transmission relationship corresponds to a directed edge, and the direction of the arrow of the directed edge is consistent with the direction of the causal influence in the transmission relationship, that is, from the micro-cost driver node that produces the influence to the micro-cost driver node that is affected.

[0122] When connecting the directed edges and nodes according to the business logic of the target hospital, the actual business processes carried out by the hospital are used as the basis, including outpatient treatment processes, inpatient nursing processes, material procurement and usage processes, equipment operation and maintenance processes, etc. The nodes representing micro cost drivers are connected through corresponding directed edges according to the order of business operations and their inherent relationships, ensuring that the connection between nodes and directed edges conforms to the transmission logic of cost impact in actual business, thus constructing a complete cost dynamic transmission map of the target hospital.

[0123] The beneficial effects are that by accurately screening time-consistent pairs of drivers to be analyzed, identifying potential transmission links through detailed collaborative fluctuation analysis, and then determining effective transmission relationships through global network verification and dynamic mapping relationship reverse constraints, the cost dynamic transmission map constructed according to business logic can clearly and systematically present the transmission paths and related logics between micro cost drivers. This provides an intuitive and reliable analytical carrier for counterfactual inference of subsequent management intervention strategies, ensuring the accuracy and comprehensiveness of the panoramic simulation of cost impact.

[0124] S4. Based on the cost dynamic transmission map, perform counterfactual analysis of the management intervention measures for the target hospital to obtain a panoramic simulation of the cost impact on the target hospital;

[0125] In this embodiment of the invention, the step of performing counterfactual analysis of management intervention measures for the target hospital based on the cost dynamic transmission map to obtain a panoramic simulation of the cost impact on the target hospital includes:

[0126] The management intervention measures of the target hospital are encoded into intervention instructions for the target hospital;

[0127] The intervention command is mapped to the initial state adjustment amount of the target node in the cost dynamic transmission map;

[0128] Starting from the target node, traverse the directed edges of the cost dynamic transmission graph to determine the transmission path and transmission sequence of the initial state adjustment amount, and obtain the state change transmission sequence of the target hospital;

[0129] Based on the state change propagation sequence, the downstream nodes of the target node are iteratively updated, and the cyclic feedback path generated during the iterative update process is tracked in real time.

[0130] According to the cyclic feedback path, adjust the state values ​​of relevant nodes in the cost dynamic transmission graph until the state changes of the relevant nodes tend to stabilize;

[0131] By integrating the temporal evolution records of the relevant nodes and the steady-state distribution of the nodes after stabilization, a panoramic simulation of the cost impact on the target hospital is obtained.

[0132] When coding management interventions for the target hospital into intervention instructions, the specific content of the management interventions should be clarified first, including the business processes targeted by the measures, the purpose of implementation, the direction of adjustment, and the scope of execution. Then, a unified coding rule should be formulated to transform the core elements of each intervention into standardized character combinations. Each character combination corresponds to a unique intervention requirement, ensuring that the coded instructions can accurately and completely reflect the core intent of the original management interventions, and ultimately form the intervention instructions for the target hospital.

[0133] When mapping intervention commands to the initial state adjustment amount of target nodes in the cost dynamic transmission map, we first analyze the cost impact area pointed to by the intervention command, locate the micro cost driver node directly related to this area in the cost dynamic transmission map, i.e., the target node, and then determine the specific magnitude or value that the node needs to be adjusted based on the adjustment direction and intensity of the intervention command and the current state value of the target node. This magnitude or value is the initial state adjustment amount of the target node, thus achieving a precise correspondence between the intervention command and the initial state adjustment amount.

[0134] When traversing the directed edges of the cost dynamic transmission graph starting from the target node, the downstream nodes directly connected to the target node are checked sequentially according to the arrows on the directed edges, with the target node as the starting point. Then, these downstream nodes are used as new starting points to continue tracing their subsequent connected nodes until all nodes that may be affected by the initial state adjustment are covered, forming a complete transmission path. At the same time, the order in which each node receives the state adjustment is recorded, identifying the nodes affected first and the nodes affected later. The transmission path and the corresponding time sequence information are organized in order to obtain the state change transmission sequence of the target hospital.

[0135] When iteratively updating downstream nodes of a target node based on a state change propagation sequence, the initial state adjustment is sequentially passed to the first downstream node according to the order of the state change propagation sequence. The state value of that node is then updated based on the adjustment, and the updated node state value is used as the new influence value to be passed to the next downstream node in the sequence. This process is repeated to complete the first round of updates for all downstream nodes. Throughout the iterative update process, it is continuously monitored whether node state changes will trigger secondary effects on the updated nodes. Once such an interaction path is discovered, the propagation direction, involved nodes, and influence mode of the path are immediately recorded, and the generated cyclic feedback path is tracked in real time.

[0136] When adjusting the state values ​​of relevant nodes in the dynamic cost transmission graph based on the cyclic feedback path, the mutual influence logic between nodes in the cyclic feedback path is first analyzed to clarify the role and degree of influence of each node in the feedback path. According to the transmission order of the feedback path, the updated state value of the next node is passed back to the previous node to correct the state value of the previous node. Then, the corrected state value is passed back to the next node to perform a new round of state adjustment. This process is repeated. After each adjustment, the difference in the state value of the node before and after the adjustment is compared. When the difference in the state value of the node remains within a fixed range after three consecutive adjustments, it is determined that the state change of the node tends to be stable, until all relevant nodes reach a stable state.

[0137] When integrating the temporal evolution records of relevant nodes and the steady-state distribution of nodes after stabilization, every change in the state value of each relevant node during the iterative update and cyclic feedback adjustment process is collected and organized into a complete temporal evolution record in chronological order, clearly presenting the entire trajectory of node state changes from the initial state to the steady state; at the same time, the final state values ​​of all nodes after stabilization are statistically analyzed, clarifying the steady-state value and distribution of each node. The temporal evolution records and steady-state distribution are systematically integrated to comprehensively demonstrate the impact process and final result of management intervention measures on each node in the cost dynamic transmission map, resulting in a panoramic simulation of the cost impact on the target hospital.

[0138] The beneficial effects are that standardized coding and mapping enable the precise conversion of management intervention measures into node adjustment quantities. Through systematic path traversal, iterative updates and cyclical feedback adjustments, and then the integration of complete temporal evolution records and steady-state distribution, the resulting panoramic simulation of cost impact can comprehensively and meticulously present the impact path, process and final result of management intervention measures on hospital cost-related nodes. This provides an intuitive, comprehensive and accurate decision-making basis for subsequent assessment of cost structure robustness and formulation of composite cost control strategies, ensuring the depth and effectiveness of cost control assessment work.

[0139] S5. Analyze the deviation between the overall network equilibrium state change caused by the management intervention measures and the historical normal range of the target hospital, and evaluate the cost structure robustness of the target hospital under the management intervention measures by combining the external market stress test data of the target hospital.

[0140] In this embodiment of the invention, the step of analyzing the deviation between the network-wide equilibrium change caused by the management intervention measures and the historical normal range of the target hospital, combined with the external market stress test data of the target hospital, to assess the cost structure robustness of the target hospital under the management intervention measures includes:

[0141] From the panoramic simulation of the cost impact of the target hospital, extract the node stability values ​​of the cost dynamic transmission map in the target hospital to obtain the current equilibrium vector of the target hospital;

[0142] By comparing and analyzing the current equilibrium vector with the historical normal range of the target hospital, the degree and direction of the state deviation of the target hospital can be obtained.

[0143] Based on the cost dynamic transmission map, a panoramic mapping is performed on the degree of state deviation and the direction of deviation to obtain the multidimensional deviation spectrum of the target hospital;

[0144] The external market stress test data of the target hospital is analyzed and mapped to the cost dynamic transmission map to obtain the standardized stress impact scenario of the target hospital.

[0145] Applying the standardized stress impact scenario to the current equilibrium vector yields a multi-scenario stress test response sequence for the target hospital;

[0146] The cost structure robustness of the target hospital is determined based on the multidimensional deviation spectrum and the multi-scenario stress test response sequence.

[0147] The step of determining the cost structure robustness of the target hospital based on the multidimensional deviation spectrum and the multi-scenario stress test response sequence includes:

[0148] The standard deviation of the multidimensional deviation spectrum is obtained by performing scalar convergence on the multidimensional deviation spectrum.

[0149] The stress scenarios in the multi-scenario stress test response sequence are quantified using resilience parameters to obtain the scenario elasticity ratio of the multi-scenario stress test response sequence.

[0150] The robustness of the initial cost structure of the target hospital is calculated based on the standard deviation value and the scenario elasticity ratio.

[0151] The robustness of the initial cost structure is verified for consistency to obtain the robustness of the target hospital's cost structure.

[0152] The formula for calculating the robustness of the initial cost structure is as follows:

[0153] ;

[0154] in, This indicates the robustness of the initial cost structure. This represents the total number of nodes in the multidimensional deviation spectrum. Indicates the first The standard deviation value of each node, This represents the total number of the aforementioned stress scenarios. Indicates the first The scenario elasticity ratio of the aforementioned stress scenarios, Represents the natural constant. This represents the natural logarithm function.

[0155] When extracting the node stability values ​​of the cost dynamic transmission map from the panoramic simulation of the cost impact of the target hospital, the specific values ​​of the final stable state of all nodes recorded in the panoramic simulation of the cost impact are screened one by one. According to the arrangement order of the nodes in the cost dynamic transmission map, these stability values ​​are arranged and combined in sequence to form an ordered set of values, thus obtaining the current equilibrium vector of the target hospital.

[0156] When comparing the current equilibrium vector with the historical normal range of the target hospital, the stable value data of each node in the cost dynamic transmission spectrum of the target hospital over the past 12 consecutive months are retrieved first. Through statistical analysis, the reasonable fluctuation range of the stable value of each node is determined, i.e., the historical normal range. Then, the state value of each node in the current equilibrium vector is compared with the corresponding historical normal range. If the state value is higher than the upper limit of the historical normal range, the deviation direction is positive, and the value above the upper limit is the degree of deviation. If the state value is lower than the lower limit of the historical normal range, the deviation direction is negative, and the value below the lower limit is the degree of deviation. Thus, the degree and direction of the state deviation of the target hospital are obtained.

[0157] When performing a panoramic mapping of the degree and direction of state deviation based on the cost dynamic transmission map, the distribution and connection relationship of the nodes in the cost dynamic transmission map are used as the basis. The degree and direction of state deviation corresponding to each node are marked at the corresponding node position in the map. At the same time, according to the transmission relationship between nodes, the potential transmission path of deviation between nodes is sorted out. The deviation information is classified and integrated according to multiple dimensions such as hospital business modules, cost types, and transmission links to form a multi-dimensional deviation spectrum covering the entire map.

[0158] When analyzing the external market stress test data of the target hospital and mapping it to the cost dynamic transmission map, we first collect stress data related to hospital operations in the external market, including data on fluctuations in medical consumable prices, changes in drug procurement costs, adjustments to medical insurance payment policies, and competition data in the medical service market. We then classify and organize this data, clarifying the impact area and intensity of each type of stress data. Next, we match each type of stress data with the nodes in the cost dynamic transmission map that are involved in that area, clarifying the type of stress, impact magnitude, and mode of action that each node needs to bear, thus obtaining the standardized stress impact scenario for the target hospital.

[0159] When applying a standardized pressure shock scenario to the current equilibrium vector, each standardized pressure shock scenario is applied to the current equilibrium vector. The state value change of each node from the initial state to the stable state again is recorded in real time. According to the type of pressure scenario and the order of time, the state change data of each node is organized into a continuous sequence. The sequence data of all nodes are summarized to obtain the multi-scenario pressure test response sequence of the target hospital.

[0160] When performing scalar convergence on a multidimensional deviation spectrum, the quantification standard of the deviation data in each dimension of the multidimensional deviation spectrum is first defined. The deviation data of different dimensions and units are uniformly converted into dimensionless values. Then, the dimensionless deviation data are aggregated into a single-dimensional quantification result through data integration methods. This result is the standard deviation value of the multidimensional deviation spectrum, which centrally reflects the overall deviation level.

[0161] When quantifying the resilience parameters of stress scenarios in a multi-scenario stress test response sequence, for each stress scenario, the time length from when a node is subjected to a stress shock to when it recovers to a stable state, the maximum fluctuation amplitude of the state value, and the number of fluctuations are statistically analyzed. Based on these data, the node's ability to withstand stress and recover stability is evaluated, and this ability is converted into specific quantitative values ​​to obtain the scenario resilience ratio of the multi-scenario stress test response sequence. The higher the ratio, the stronger the node resilience.

[0162] When calculating the robustness of the initial cost structure of a target hospital based on the standard deviation and scenario elasticity ratio, the internal deviation reflected by the standard deviation and the external pressure coping ability reflected by the scenario elasticity ratio are considered together. The relationship between the two is comprehensively taken into account. If the standard deviation is smaller and the scenario elasticity ratio is higher, the robustness of the initial cost structure is stronger. By comprehensively judging the actual values ​​of the two, the specific robustness result of the initial cost structure is obtained.

[0163] The total number of nodes in the multidimensional deviation spectrum comes from the multidimensional deviation spectrum, which is obtained by panoramic mapping of the state deviation degree and deviation direction based on the cost dynamic transmission map. The state deviation degree and deviation direction are obtained by comparing and analyzing the state value of the current equilibrium vector with the historical normal range of the target hospital. The current equilibrium vector is obtained by extracting the node stability value of the cost dynamic transmission map of the target hospital from the panoramic simulation of cost impact.

[0164] The standard deviation value is the result of scalar convergence processing of the multidimensional deviation spectrum. Scalar convergence processing is the normalization calculation of the relevant data of all nodes in the multidimensional deviation spectrum, and finally obtains a single quantized value corresponding to each node.

[0165] The total number of stress scenarios comes from standardized stress shock scenarios. Standardized stress shock scenarios are formed by first analyzing the external market stress test data of the target hospital, and then mapping the analysis results of the analyzed data onto the cost dynamic transmission map.

[0166] The scenario resilience ratio of a stress scenario is obtained by quantifying the resilience parameter for each stress scenario in the multi-scenario stress test response sequence. The resilience parameter quantification is to extract resilience-related features and perform quantification calculations on the response of the hospital cost structure under each stress scenario, and finally obtain the specific value corresponding to each stress scenario. The multi-scenario stress test response sequence is a sequence formed by recording the response changes of the target hospital cost structure under different stress scenarios after applying a standardized stress impact scenario to the current equilibrium vector.

[0167] This calculation quantifies the robustness of the initial cost structure by integrating relevant data from multidimensional deviation spectra and multi-scenario stress test response sequences.

[0168] The calculation first calculates the natural logarithm of each standard deviation value added to 1 for all nodes, then sums all these natural logarithm results, divides the sum by the total number of nodes in the multidimensional deviation spectrum to obtain the average value, and then takes the reciprocal of the average value.

[0169] Meanwhile, for all stress scenarios, the scenario elasticity ratio is calculated as the negative scenario elasticity ratio power of the natural constant. The result is subtracted from 1, and all these differences are summed. The summation is then divided by the total number of stress scenarios to obtain the average value.

[0170] Multiplying the results of the two average calculations above, the final value can comprehensively reflect the degree to which the target hospital's cost structure deviates from the historical normal range after the implementation of management intervention measures, as well as its resilience in the face of external market pressures, thereby accurately quantifying the robustness of the cost structure.

[0171] When the standard deviation of each node decreases as a whole, it means that the degree to which the cost structure of the target hospital deviates from the historical norm is reduced. The natural logarithm of each standard deviation added to 1 will decrease accordingly. The sum of all such natural logarithms will decrease, and the average value obtained by dividing by the total number of nodes will also decrease. The reciprocal of this average value will increase, which will increase the numerical contribution of the corresponding part in the calculation.

[0172] When the overall scenario elasticity ratio of each stress scenario increases, it indicates that the overall resilience of the target hospital to external market stress shocks is enhanced. The negative scenario elasticity ratio power of the natural constant will decrease accordingly. The difference between 1 and this value will increase. The sum of all such differences will increase, and the average value obtained by dividing by the total number of stress scenarios will also increase, thus increasing the numerical contribution of the corresponding part in the calculation.

[0173] These two numerical contributions jointly affect the final result. When the standard deviation decreases as a whole and the scenario elasticity ratio increases as a whole, both numerical contributions increase, and the final result after multiplication will increase, representing an improvement in the robustness of the initial cost structure.

[0174] When the standard deviation of each node increases as a whole, the degree to which the cost structure of the target hospital deviates from the historical norm increases as a whole. The natural logarithm of each standard deviation added to 1 increases, the sum of all such natural logarithms increases, the average value after dividing by the total number of nodes increases, the reciprocal of the average value decreases, and the numerical contribution of the corresponding part in the calculation decreases.

[0175] When the overall scenario elasticity ratio of each stress scenario decreases, the overall resilience of the target hospital to external market stress shocks weakens, the negative scenario elasticity ratio power of the natural constant increases, the difference after subtracting this value from 1 decreases, the sum of all such differences decreases, the average value after dividing by the total number of stress scenarios decreases, and the numerical contribution of the corresponding part in the calculation decreases.

[0176] The decrease in the contribution of both parts will result in a decrease in the final result after multiplication, indicating a decline in the robustness of the initial cost structure.

[0177] When performing a consistency check on the robustness of the initial cost structure, cost operation data from different business cycles and different business modules of the target hospital are retrieved. The initial cost structure robustness results are compared with these actual operation data to check whether the results are consistent with the actual cost operation patterns of the hospital. If there are abnormal values ​​that are inconsistent with the actual situation, the initial results are corrected in conjunction with the business logic to ensure that the final results can truly reflect the actual robustness level of the hospital's cost structure, thus obtaining the robustness of the target hospital's cost structure.

[0178] The beneficial effects are that by extracting the current equilibrium vector from the system and comparing it with the historical normal range, and combining it with external market stress test data to carry out multi-dimensional analysis and verification, the final cost structure robustness assessment results can comprehensively and objectively reflect the internal deviation of the hospital's cost structure and its ability to cope with external pressures under management intervention measures. This provides an accurate and reliable assessment basis for the subsequent formulation of scientific and reasonable composite cost control strategies, ensuring the pertinence and effectiveness of cost control work.

[0179] S6. Based on the panoramic simulation of cost impact and the robustness of the cost structure, formulate a composite cost control strategy for the target hospital.

[0180] In this embodiment of the invention, the step of formulating a composite cost control strategy for the target hospital based on the panoramic simulation of cost impact and the robustness of the cost structure includes:

[0181] By performing inverse analysis of the strategy impact on the panoramic simulation of cost impact, the core control nodes of the target hospital can be obtained;

[0182] Based on the cost structure robustness assessment results and the core control nodes, a set of candidate action nodes for the target hospital is obtained;

[0183] The candidate action node set is deconstructed to obtain the basic strategy atoms of the target hospital;

[0184] Based on the cost dynamic transmission map of the target hospital, the basic strategy atoms are combined and spliced ​​to obtain the composite strategy scheme of the target hospital;

[0185] The composite strategy scheme is optimized through multi-objective strategy search to obtain the composite cost control strategy for the target hospital.

[0186] When performing reverse analysis of the policy impact of the panoramic simulation of cost impact, we comprehensively sort out the time-series evolution records and steady-state distribution of all nodes in the panoramic simulation, locate the key nodes that play a leading role in cost changes and can trigger chain effects, analyze the core position of these nodes in the cost transmission path and their decisive role in the overall cost outcome, and screen out the nodes that can significantly optimize the cost structure and improve cost performance after adjustment, thus obtaining the core control nodes of the target hospital.

[0187] When combining the results of the cost structure robustness assessment with the core control nodes, we first extract the weak nodes in the cost structure identified in the cost structure robustness assessment, including nodes with large deviations and low scenario elasticity ratios. Then, we integrate these weak nodes with the core control nodes, eliminate duplicate nodes, and supplement key intervention nodes that may have been missed due to insufficient robustness, forming a set of candidate role nodes for target hospitals that cover core control needs and risk prevention and control needs.

[0188] When performing a critical deconstruction of the candidate action node set, the cost impact mechanism, intervention dimensions, intervention boundary conditions and expected control targets of each candidate action node are analyzed one by one. The specific intervention actions, implementation methods and applicable scenarios corresponding to each node are broken down into independent and indivisible basic intervention units. Each unit clearly corresponds to the specific control needs of a single node, thus obtaining the basic strategy atoms of the target hospital.

[0189] When combining basic strategy atoms based on the cost dynamic transmission map of the target hospital, the synergy and compatibility between different basic strategy atoms are judged according to the transmission relationship of the nodes in the map and the order of business logic. This avoids situations where intervention actions conflict with each other or transmission paths contradict each other. Basic strategy atoms that act on upstream and downstream related nodes and can form synergistic effects are combined in a reasonable order, while covering core control nodes and weak nodes, to construct multiple composite strategy schemes for the target hospital with complete intervention logic.

[0190] When optimizing a composite strategy for multiple objectives, the objectives should be clearly defined, including the extent of cost reduction, the level of improvement in cost structure robustness, and the degree of guarantee for the normal operation of the hospital. A unified evaluation standard should be established, and the achievement effect of each composite strategy on each objective should be calculated separately. The importance of each objective should be comprehensively weighed, and schemes with obvious shortcomings or implementation risks should be eliminated. The scheme that performs well in multiple objectives and has the best overall benefits should be selected to obtain the composite cost control strategy for the target hospital.

[0191] The beneficial effects are that by reverse analysis, core control nodes are accurately identified, and key intervention nodes are supplemented by robustness assessment. After deconstruction, precise basic strategy atoms are obtained, and then combined and optimized according to the transmission logic. The resulting composite cost control strategy can accurately meet the hospital's cost management needs, take into account both core control and risk prevention, and achieve synergistic linkage among various strategy units. This effectively improves the pertinence, systematicness, and feasibility of cost control, and helps hospitals ensure a robust cost structure while controlling costs.

[0192] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0193] This application embodiment can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0194] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for evaluating cost control in hospital management, characterized in that, The method includes: S1. Perform multi-granularity behavioral slicing on the process state data stream of the target hospital, and perform context-aware analysis on the sliced ​​data to obtain the micro cost drivers of the target hospital; S2. Based on the micro cost drivers, perform diachronic analysis on the outcome data stream of the target hospital to construct a dynamic mapping relationship between the micro cost drivers and the outcome data stream; S3. Based on the dynamic mapping relationship, analyze the transmission relationship between the micro cost drivers, and construct the cost dynamic transmission map of the target hospital by taking the transmission relationship as the edge and the micro cost drivers as the node. S4. Based on the cost dynamic transmission map, perform counterfactual analysis of the management intervention measures for the target hospital to obtain a panoramic simulation of the cost impact on the target hospital; S5. Analyze the deviation between the overall network equilibrium state change caused by the management intervention measures and the historical normal range of the target hospital, and evaluate the cost structure robustness of the target hospital under the management intervention measures by combining the external market stress test data of the target hospital. S6. Based on the panoramic simulation of cost impact and the robustness of the cost structure, formulate a composite cost control strategy for the target hospital.

2. The method for evaluating cost control in hospital management as described in claim 1, characterized in that, The process of segmenting the target hospital's process-state data stream into multi-granular behavior slices and performing context-aware analysis on the sliced ​​data to obtain the target hospital's micro-cost drivers includes: Collect the medical order execution trajectory, equipment usage logs, and personnel movement records of the target hospital to obtain the process state data stream of the target hospital; The process-state data stream is aligned with a unified spatiotemporal reference to obtain the synchronous process-state data of the target hospital; According to the time window in which the business occurs in the target hospital, the synchronous process state data is initially segmented to obtain the primary data slice of the target hospital; The primary data slice is further refined to obtain the secondary data slice of the target hospital; Pattern matching is performed on the secondary data slices, and causal reasoning is performed on the matched patterns to obtain the micro cost drivers of the target hospital.

3. The method for evaluating cost control in hospital management as described in claim 1, characterized in that, The step of performing a diachronic analysis on the outcome-state data stream of the target hospital based on the micro-cost drivers to construct a dynamic mapping relationship between the micro-cost drivers and the outcome-state data stream includes: Obtain the financial settlement records and material consumption list of the target hospital to obtain the result state data stream of the target hospital; Time alignment is performed on the resulting state data stream and the micro cost drivers to obtain the dual data set of the target hospital; Cooperative fluctuation pattern analysis was performed on the dual data set to obtain the lag correlation and pattern matching degree of the dual data set. The impact intensity of the micro cost drivers is determined based on the lag correlation and pattern matching degree. The intensity of the influence is tracked over time to obtain the variation pattern of the influence intensity; Based on the intensity of the influence and the pattern of change, a correlation analysis is performed on the micro-cost drivers and the result state data stream to obtain the dynamic mapping relationship between the micro-cost drivers and the result state data stream.

4. The method for evaluating cost control in hospital management as described in claim 1, characterized in that, The step of analyzing the transmission relationship between the micro-cost drivers based on the dynamic mapping relationship, and constructing a cost dynamic transmission map of the target hospital by using the transmission relationship as edges and the micro-cost drivers as nodes, includes: Based on the dynamic mapping relationship, the micro cost drivers with the same time sequence are selected to obtain the target hospital's driver pairs to be analyzed; Perform synergistic fluctuation analysis on the pairs of drivers to be analyzed to identify the potential transmission links of the micro cost drivers; The potential transmission links are globally networked and verified, and the verified transmission links are reverse-constrained according to the dynamic mapping relationship to obtain the transmission relationship between the micro cost drivers. The transmission relationship is treated as a directed edge, and the micro cost driver is treated as a node; Based on the business logic of the target hospital, the directed edges and the nodes are connected to construct a dynamic cost transmission graph for the target hospital.

5. The method for evaluating cost control in hospital management as described in claim 4, characterized in that, The step of performing coordinated fluctuation analysis on the pair of drivers to be analyzed to identify the potential transmission links of the micro-cost drivers includes: Frequency statistics are performed on the pairs of drivers to be analyzed to obtain the frequency index sequence of the micro cost drivers; The frequency index sequence is decomposed into a trend term to obtain the fluctuation component of the micro cost driver. By performing time-lag estimation on the fluctuation components, the lead-lag relationship and time difference of the micro cost drivers are obtained; Based on the aforementioned lead-lag relationship, a state transition analysis is performed on the pair of drivers to be analyzed to obtain the fluctuation pattern transformation law of the micro cost drivers. By combining the aforementioned lead-lag relationship, the aforementioned time difference, and the aforementioned fluctuation pattern transformation law, and by quantitatively evaluating the causal influence direction and intensity of the driver pair to be analyzed, the potential transmission links of the aforementioned micro cost drivers are obtained.

6. The method for evaluating cost control in hospital management as described in claim 1, characterized in that, Based on the cost dynamic transmission map, the management intervention measures for the target hospital are subjected to counterfactual inference to obtain a panoramic simulation of the cost impact on the target hospital, including: The management intervention measures of the target hospital are encoded into intervention instructions for the target hospital; The intervention command is mapped to the initial state adjustment amount of the target node in the cost dynamic transmission map; Starting from the target node, traverse the directed edges of the cost dynamic transmission graph to determine the transmission path and transmission sequence of the initial state adjustment amount, and obtain the state change transmission sequence of the target hospital; Based on the state change propagation sequence, the downstream nodes of the target node are iteratively updated, and the cyclic feedback path generated during the iterative update process is tracked in real time. According to the cyclic feedback path, adjust the state values ​​of relevant nodes in the cost dynamic transmission graph until the state changes of the relevant nodes tend to stabilize; By integrating the temporal evolution records of the relevant nodes and the steady-state distribution of the nodes after stabilization, a panoramic simulation of the cost impact on the target hospital is obtained.

7. The method for evaluating cost control in hospital management as described in claim 1, characterized in that, The analysis examines the deviation between the network-wide equilibrium change caused by the management intervention measures and the historical normal range of the target hospital. Combined with external market stress test data from the target hospital, the assessment evaluates the cost structure robustness of the target hospital under the management intervention measures, including: From the panoramic simulation of the cost impact of the target hospital, extract the node stability values ​​of the cost dynamic transmission map in the target hospital to obtain the current equilibrium vector of the target hospital; By comparing and analyzing the current equilibrium vector with the historical normal range of the target hospital, the degree and direction of the state deviation of the target hospital can be obtained. Based on the cost dynamic transmission map, a panoramic mapping is performed on the degree of state deviation and the direction of deviation to obtain the multidimensional deviation spectrum of the target hospital; The external market stress test data of the target hospital is analyzed and mapped to the cost dynamic transmission map to obtain the standardized stress impact scenario of the target hospital. Applying the standardized stress impact scenario to the current equilibrium vector yields a multi-scenario stress test response sequence for the target hospital; The cost structure robustness of the target hospital is determined based on the multidimensional deviation spectrum and the multi-scenario stress test response sequence.

8. The method for evaluating cost control in hospital management as described in claim 7, characterized in that, The step of determining the cost structure robustness of the target hospital based on the multidimensional deviation spectrum and the multi-scenario stress test response sequence includes: The standard deviation of the multidimensional deviation spectrum is obtained by performing scalar convergence on the multidimensional deviation spectrum. The stress scenarios in the multi-scenario stress test response sequence are quantified using resilience parameters to obtain the scenario elasticity ratio of the multi-scenario stress test response sequence. The robustness of the initial cost structure of the target hospital is calculated based on the standard deviation value and the scenario elasticity ratio. The robustness of the initial cost structure is verified for consistency to obtain the robustness of the target hospital's cost structure.

9. The method for evaluating cost control in hospital management as described in claim 8, characterized in that, The formula for calculating the robustness of the initial cost structure is as follows: ； in, This indicates the robustness of the initial cost structure. This represents the total number of nodes in the multidimensional deviation spectrum. Indicates the first The standard deviation value of each node, This represents the total number of the aforementioned stress scenarios. Indicates the first The scenario elasticity ratio of the aforementioned stress scenarios, Represents the natural constant. This represents the natural logarithm function.

10. The method for evaluating cost control in hospital management as described in claim 1, characterized in that, The step of formulating a comprehensive cost control strategy for the target hospital based on the panoramic simulation of cost impact and the robustness of the cost structure includes: By performing inverse analysis of the strategy impact on the panoramic simulation of cost impact, the core control nodes of the target hospital can be obtained; Based on the cost structure robustness assessment results and the core control nodes, a set of candidate action nodes for the target hospital is obtained; The candidate action node set is deconstructed to obtain the basic strategy atoms of the target hospital; Based on the cost dynamic transmission map of the target hospital, the basic strategy atoms are combined and spliced ​​to obtain the composite strategy scheme of the target hospital; The composite strategy scheme is optimized through multi-objective strategy search to obtain the composite cost control strategy for the target hospital.

Images