Hierarchical diagnosis and treatment medical resource allocation optimization method and system

Abstract

The invention relates to a medical resource allocation optimization method and system for hierarchical diagnosis and treatment, and relates to the technical field of medical resource management, and the method comprises the steps: collecting to-be-allocated medical resources in hierarchical diagnosis and treatment; acquiring first-level personnel structure information and second-level personnel structure information of the first-level region and the plurality of second-level regions, and performing medical demand analysis; performing override medical probability analysis according to the plurality of pieces of secondary personnel structure information and the plurality of grading distances; and according to the first-level demand error coefficient, the plurality of second-level demand error coefficients and the plurality of override error coefficients, carrying out allocation optimization of the medical resources to obtain a medical resource allocation result, and carrying out medical resource allocation on the first-level region and the plurality of second-level regions. According to the method, the problems of unbalanced and mismatched medical resource distribution in hierarchical diagnosis and treatment are solved, the integrating degree of resources and demands is improved, and the situation that resource waste and shortage exist at the same time is avoided.

Description

Technical Field

[0001] The present application relates to the field of medical resource management, and in particular to a method and system for optimizing the allocation of medical resources for tiered diagnosis and treatment. Background Art

[0002] With the deepening implementation of a tiered diagnosis and treatment system, the rational allocation of medical resources has become a key component in improving primary healthcare service capabilities and optimizing the efficiency of the healthcare system. Currently, there is a significant disparity in medical resources between urban and rural areas, and between regions. High-quality resources are mostly concentrated in large urban hospitals, while resources are relatively scarce in primary and remote areas. This has led to the simultaneous phenomenon of idle primary-level resources and overburdened higher-level hospitals.

[0003] The existing method of allocating medical resources often ignores dynamic factors such as geographical distance, differences in population structure, and skipping levels of medical treatment. It only allocates resources based on experience or static data, resulting in a mismatch between resource allocation and actual needs. This not only affects the efficiency of medical resource utilization, but also hinders the realization of medical equity. Summary of the Invention

[0004] In order to solve the above technical problems, this application provides a medical resource allocation optimization method and system for tiered diagnosis and treatment, which overcomes the problems of existing medical resource allocation that does not consider dynamic factors and mismatches supply and demand, and improves the rationality and fairness of medical resource allocation.

[0005] The embodiments of this application disclose the following technical solutions:

[0006] In a first aspect, an embodiment of the present application provides a method for optimizing allocation of medical resources for hierarchical diagnosis and treatment, the method comprising:

[0007] Collecting medical resources to be allocated in hierarchical diagnosis and treatment, wherein the hierarchical diagnosis and treatment includes a primary region and multiple secondary regions, and multiple hierarchical distances exist between the primary region and the multiple secondary regions;

[0008] Collecting primary personnel structure information and secondary personnel structure information of the primary region and multiple secondary regions, performing medical demand analysis, and obtaining primary medical demand parameters and multiple secondary medical demand parameters, as well as primary demand error coefficients and multiple secondary demand error coefficients;

[0009] Based on multiple secondary personnel structure information and multiple hierarchical distances, a probability analysis of skipping medical care is performed to obtain multiple skipping medical care probabilities and multiple skipping error coefficients;

[0010] According to the first-level demand error coefficient, multiple second-level demand error coefficients and multiple skip-level error coefficients, based on the first-level medical demand parameters, multiple second-level medical demand parameters and multiple skip-level medical probabilities, the allocation of medical resources is optimized to obtain medical resource allocation results, and medical resources are allocated to the first-level region and multiple second-level regions.

[0011] In a second aspect, an embodiment of the present application provides a medical resource allocation optimization system for hierarchical diagnosis and treatment, the system comprising:

[0012] A resource and distance collection module is used to collect medical resources to be allocated in hierarchical diagnosis and treatment, wherein the hierarchical diagnosis and treatment includes a primary region and multiple secondary regions, and there are multiple hierarchical distances between the primary region and the multiple secondary regions;

[0013] a medical demand analysis module, configured to collect primary personnel structure information and secondary personnel structure information of the primary region and multiple secondary regions, perform medical demand analysis, and obtain primary medical demand parameters and multiple secondary medical demand parameters, as well as primary demand error coefficients and multiple secondary demand error coefficients;

[0014] A skip-level medical treatment probability analysis module is used to perform a skip-level medical treatment probability analysis based on multiple secondary personnel structure information and multiple hierarchical distances, and obtain multiple skip-level medical treatment probabilities and multiple skip-level medical treatment error coefficients;

[0015] The resource allocation optimization and decision-making module is used to optimize the allocation of medical resources based on the first-level demand error coefficient, multiple second-level demand error coefficients and multiple skip-level error coefficients, based on the first-level medical demand parameters, multiple second-level medical demand parameters and multiple skip-level medical probabilities, obtain medical resource allocation results, and allocate medical resources to the first-level region and multiple second-level regions.

[0016] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0017] This application proposes a method and system for optimizing the allocation of medical resources for hierarchical diagnosis and treatment. By collecting information on medical resources, hierarchical distances, and personnel structure from a first-level region and multiple second-level regions, a comprehensive analysis of medical demand parameters, the probability of skipping medical care, and its error coefficient is conducted, and multi-source data is combined with a dynamic optimization model. At the same time, based on multiple types of data such as geographical distance, population structure, and medical behavior, an analysis model is constructed according to the number of people in each age range, the distance between the hierarchical levels, and historical configuration data. The resource allocation is dynamically weighted and regulated based on the error coefficient, effectively improving the fit and fairness between the allocation of medical resources and actual needs. This achieves efficient resource allocation in complex scenarios such as large regional differences in medical resources and frequent skipping medical treatment, avoiding the simultaneous occurrence of resource waste and shortage problems. Through the steps of medical demand analysis, skipping probability prediction, and dynamic optimization allocation, multi-source data is integrated and multiple influencing factors are evaluated, effectively avoiding the problem of unreasonable allocation caused by relying on static data and empirical judgment. At the same time, the dynamic error correction mechanism effectively improves the rationality and adaptability of resource allocation.

[0018] The technical solution of this application realizes the precise matching of medical resource allocation with actual regional needs by integrating dynamic data such as geographical distance, population structure, and probability of skipping medical treatment. It solves the problem of irrational resource allocation caused by ignoring dynamic factors in tiered diagnosis and treatment, improves the fit between resources and needs, and avoids the coexistence of idle resources at the grassroots level and overloaded higher-level hospitals. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0020] Figure 1 A flowchart of a method for optimizing allocation of medical resources for hierarchical diagnosis and treatment provided in an embodiment of the present application;

[0021] Figure 2 A schematic diagram of the structure of a medical resource allocation optimization system for hierarchical diagnosis and treatment provided in an embodiment of the present application;

[0022] In the accompanying drawings, the components represented by the reference numerals are described as follows:

[0023] Resource and distance collection module 01, medical demand analysis module 02, skipping probability analysis module 03, resource allocation optimization and decision-making module 04. DETAILED DESCRIPTION

[0024] The present application provides a method and system for optimizing the allocation of medical resources for tiered diagnosis and treatment, which is used to solve the technical problems existing in the prior art of irrational allocation of medical resources in tiered diagnosis and treatment, resulting in idle resources at the grassroots level and overloaded hospitals at higher levels.

[0025] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0026] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature specified as "first" or "second" may explicitly or implicitly include one or more of the described features. In the description of this application, "plurality" means two or more, unless otherwise specifically specified.

[0027] In the description of this application, the term "for example" is used to mean "used as an example, illustration or explanation". Any embodiment described in this application as "for example" is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any person skilled in the art to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in this application.

[0028] Example 1, as shown in the attached Figure 1 As shown, the present application provides a method for optimizing the allocation of medical resources for hierarchical diagnosis and treatment, the method comprising the following steps:

[0029] S100: Collecting medical resources to be allocated in hierarchical diagnosis and treatment, wherein the hierarchical diagnosis and treatment includes a primary region and multiple secondary regions, and multiple hierarchical distances exist between the primary region and the multiple secondary regions;

[0030] In the embodiment of the present application, during the medical resource allocation process, in order to achieve the reasonable call and accurate allocation of medical resources between the primary region and multiple secondary regions, it is necessary to use the hierarchical diagnosis and treatment information management platform to use data collection technology to comprehensively collect various aspects and types of medical resource data to be allocated, including the number of medical manpower (such as the number of doctors, the ratio of nursing staff), medical material reserves (such as the number of testing equipment, the number of emergency medicine reserves), medical space resources (such as the number of beds, diagnosis and treatment areas), etc., to build a standardized resource information database to provide reliable data support for subsequent resource allocation.

[0031] Step S100 of the method provided in the embodiment of the present application includes:

[0032] Collecting medical resources to be allocated in hierarchical diagnosis and treatment, wherein the medical resources include multiple types of resources, and the hierarchical diagnosis and treatment includes a primary region and multiple secondary regions;

[0033] The geographical distances between the first-level region and the plurality of second-level regions are acquired and collected to obtain a plurality of hierarchical distances.

[0034] In the embodiment of the present application, during the medical resource allocation process, in order to achieve reasonable allocation and accurate matching of medical resources between the primary region and multiple secondary regions, it is necessary to comprehensively collect the medical resource data to be allocated.

[0035] Among them, the tiered diagnosis and treatment system includes first-level regions (such as tertiary hospitals in provincial capital cities and regional medical centers) and multiple second-level regions (such as county-level people's hospitals, township health centers, and community health service centers). Due to geographical and spatial differences, different levels of tiered distances are formed between regions (such as 5 kilometers, 10 kilometers, 30 kilometers, etc.).

[0036] Specifically, the collection of medical resources is completed through the hierarchical diagnosis and treatment management platform. By connecting to the resource management systems of medical institutions at all levels through data interfaces, real-time information on medical resources to be allocated is obtained, including the number of medical staff, nursing staff ratios, the types and quantities of medical equipment, the number of beds, the amount of medicine in stock, and other resource data.

[0037] Furthermore, based on the collected hierarchical distances, the spatial location coordinates of the first-level regions and each second-level region are obtained through the geographic information system (GIS), and the straight-line distances and traffic distances between different regions are calculated and generated, forming a hierarchical distance dataset containing multi-dimensional spatial parameters, providing spatial data support for subsequent medical resource allocation.

[0038] Among them, the collection of medical resources is completed by real-time synchronization of resource data of medical institutions at all levels through the API interface of the tiered diagnosis and treatment management platform. After data unification and standardization (such as unifying units and eliminating outliers), they are classified and stored according to resource type, regional level and other information.

[0039] The hierarchical distance calculation is used to measure and analyze the spatial distance between the first-level region and the second-level region through the Geographic Information System (GIS).

[0040] Specifically, we first use GIS to obtain the spatial coordinates of regions at all levels, and then calculate the geographic straight-line distance between them based on map projection (for example, the straight-line distance between a city central hospital and a township health center). Next, we use route planning to further analyze the actual traffic conditions between the two locations and calculate the travel time between regions, comprehensively considering the impact of dynamic factors such as road grade and congestion on traffic efficiency. Finally, we integrate parameters such as straight-line distance and travel time into a hierarchical distance dataset, which reflects the spatial correlation data between regions at all levels.

[0041] Among them, the hierarchical distance dataset provides a spatial decision-making basis for the allocation of medical resources. It can optimize the priority of resource allocation according to the distance, reserve emergency resource allocation for remote areas with long roads and inconvenient transportation, and improve the spatial adaptability and rationality of resource scheduling in the hierarchical diagnosis and treatment system.

[0042] For example, taking a certain place as an example, the first-level area is Hospital A, and the second-level areas include Community Health Service Center B 3 kilometers in the urban area, Hospital C 15 kilometers in the suburbs, and Health Center D 40 kilometers in the suburbs. The municipal platform collects resources at all levels through data interfaces, such as the number of specialists in tertiary hospitals, the inventory of conventional medicines in the community, and the basic equipment list of townships. According to GIS calculations, it takes 15 minutes to drive from the hospital to the community, 30 minutes to Hospital C, and 1.5 hours (including mountain roads) to get to Health Center D. When allocating resources, chronic disease drugs are replenished in the nearby areas first, remote consultation equipment is configured in the central areas, and due to inconvenient transportation in the remote areas, additional first aid kits are stored and mobile medical vehicles are sent for weekly rounds to improve the efficiency of medical resource allocation.

[0043] S200: Collecting primary personnel structure information and secondary personnel structure information of the primary region and multiple secondary regions, performing medical demand analysis, and obtaining primary medical demand parameters and multiple secondary medical demand parameters, as well as primary demand error coefficients and multiple secondary demand error coefficients;

[0044] In the embodiment of the present application, during the medical resource allocation process, in order to achieve accurate matching of resources between the primary region and multiple secondary regions, it is necessary to collect personnel structure information of the primary region and multiple secondary regions through the hierarchical diagnosis and treatment information management platform.

[0045] Specifically, the first-level personnel structure information includes the age distribution, occupational characteristics, and proportion of chronic disease history of residents in the jurisdiction; the second-level personnel structure information includes the number of special populations in each region (such as the proportion of pregnant women and the elderly population), common disease types, etc.

[0046] By analyzing this data for medical demand, we obtain primary and secondary medical demand parameters reflecting the intensity of medical resource demand in each region. Furthermore, by combining the data's fluctuation range with historical errors, we calculate primary and secondary demand error coefficients, providing a quantitative basis for the dynamic adjustment of subsequent resource allocation.

[0047] Step S200 of the method provided in the embodiment of the present application includes:

[0048] Collecting primary personnel structure information and secondary personnel structure information of the primary region and multiple secondary regions, wherein each personnel structure information includes the number of people in multiple different age ranges;

[0049] Input the primary personnel structure information and the multiple secondary personnel structure information into a medical demand classifier sequence respectively, and output a predicted primary medical demand parameter set and multiple predicted secondary medical demand parameter sets;

[0050] Calculating means based on the predicted primary medical demand parameter set and the multiple predicted secondary medical demand parameter sets to obtain a primary medical demand parameter and multiple secondary medical demand parameters;

[0051] Calculate the average error between other predicted first-level medical demand parameters in the predicted first-level medical demand parameter set and the first-level medical demand parameter as the first-level demand error coefficient;

[0052] The average error margins between other predicted secondary medical demand parameters and the secondary medical demand parameters in the plurality of predicted secondary medical demand parameter sets are calculated respectively as a plurality of secondary demand error coefficients.

[0053] In the embodiment of the present application, the personnel structure information of the first-level region and multiple second-level regions is comprehensively collected through the hierarchical diagnosis and treatment information management platform.

[0054] Among them, the core content collected is the number of people in different age ranges in each region, including group size data such as infants (0-6 years old), teenagers (7-18 years old), young and middle-aged people (19-59 years old), and the elderly (60 years old and above), so as to construct a detailed personnel structure data set and provide basic data support in the population structure dimension for the precise allocation of medical resources.

[0055] Specifically, information collection is completed through a combination of data interface docking and collaborative reporting. Through standardized data interfaces, it directly connects to the electronic medical record system and health archive of the tertiary hospitals in the first-level area, as well as the public health management system of the primary medical institutions in the second-level area, to automatically obtain structured data such as residents' age, gender, and underlying diseases.

[0056] Furthermore, for some areas that are not connected to the Internet or have missing data, staff from community health service centers and township health centers will use the platform's exclusive reporting port to enter the number of people in each age range and their health characteristics according to a unified template. At the same time, a data verification function will be set up to manually screen for duplicates and logically contradictory items to ensure the accuracy and completeness of the collected information.

[0057] The training steps of “the medical need classifier sequence” in the method provided in the embodiment of the present application include:

[0058] Based on the historical data of medical resource allocation, a set of sample personnel structure information is collected, and the medical resources required for each sample personnel structure information are marked to obtain a set of sample medical demand parameters;

[0059] randomly selecting a portion of data with replacement from the sample personnel structure information set and the sample medical demand parameter set to obtain first medical demand classification training data, constructing a first medical demand classifier using machine learning, and performing supervised training on the first medical demand classifier using the first medical demand classification training data until convergence;

[0060] Continue to randomly select multiple groups of medical demand classification training data with replacement, train multiple medical demand classifiers, and obtain a medical demand classifier sequence.

[0061] In an embodiment of the present application, by consulting historical medical resource configuration files, personnel structure data under different time periods and different public health conditions are screened out from the historical database of the hierarchical diagnosis and treatment information management platform to form a sample personnel structure information set.

[0062] Among them, personnel structure data includes the number of people in different age groups in each region, the proportion of special groups, etc.

[0063] For example, in addition to the number of people in different age ranges in each region, it also involves the distribution of people in different occupations, such as people in high-risk operations and people who work at a desk for a long time. These data reflect the characteristics of the population structure in each region and provide an accurate basis for the precise allocation of subsequent medical resources.

[0064] Furthermore, based on historical resource allocation, medical management experts and statisticians analyzed the personnel structure corresponding to each sample and marked the required medical resource categories and quantities, such as pediatric outpatient clinic reception capacity, elderly chronic disease drug reserves, emergency equipment, etc., and further summarized them into a sample medical demand parameter set to provide a reference basis for medical resource allocation.

[0065] After completing the construction of the sample personnel structure information set and the sample medical needs parameter set, in order to further improve the accuracy of medical needs analysis, it is necessary to randomly select some data with replacement from the two sets to obtain training data for the first medical needs classification.

[0066] Specifically, professional data personnel set a reasonable sampling ratio based on the sample size and analysis needs to ensure that the extracted data can cover various scenarios such as the population structure characteristics of different regions (such as urban-rural differences, uneven age distribution), changes in medical needs during different public health events, etc., so as to obtain training data for the first medical needs classification.

[0067] For example, in sample data from a district in Xi'an, analysis of the population structure information set revealed that 65% of the population was aged 19-59. Of these, 12% worked in high-risk occupations such as manufacturing, 20% worked in offices that worked at desks for long periods of time, and 20% were aged 60 and above. Furthermore, 45% of the elderly population suffered from chronic diseases such as hypertension and diabetes. Combined with the annotation of the sample medical demand parameter set, it can be seen that for people working in high-risk occupations, trauma emergency equipment and emergency medicines should be prioritized; for office workers, investment in cervical and lumbar spine disease diagnosis and treatment resources should be strengthened; and for the elderly with chronic diseases, a stockpile of commonly used medications should be ensured. This data was randomly sampled with replacement, such as selecting the population structure and corresponding medical demand data of three communities as the first medical demand classification training data to construct and train the medical demand classifier.

[0068] Furthermore, through the collaboration of a team of medical experts and information technology personnel, the first medical demand classification model was jointly constructed by adopting statistical decision tree algorithms, regional health planning standards and tiered diagnosis and treatment implementation specifications, and combining them with the actual medical resource allocation logic.

[0069] Among them, the core of the statistical decision tree method is to construct a tree-shaped classification model. This model takes personnel structure information as input conditions, starts from the root node, and divides the data into different branches (such as "the proportion of elderly population ≥20%" and "<20%") according to key features such as age distribution and the proportion of special populations; for each branch, the next key feature (such as the proportion of chronic disease patients) is selected to continue splitting until the leaf node forms a clear medical demand classification result and quantity prediction result (such as "the need to increase the reserve of chronic disease drugs for the elderly" and "the need to supplement 15 sets of emergency equipment for high-risk groups"); finally, the pruning parameters are optimized through cross-validation to complete the construction of the first medical demand classifier, which is used to realize accurate classification and prediction of medical needs based on personnel structure.

[0070] Furthermore, during the model training and optimization phase, a team of medical experts and statisticians closely monitor the entire process, inputting demographic information from the training data (such as age distribution and the number of special populations) into the model. The model's output is then verified against pre-labeled medical demand parameter data. By calculating metrics such as accuracy and error rate, if any deviations between the model's predictions and actual needs are detected, the expert team will discuss and analyze the results, adjust the model's calculation parameters, and further optimize the classification rules.

[0071] Furthermore, after multiple rounds of repeated verification and correction, until the model output results become stable and conform to the actual medical resource allocation rules, the first medical demand classifier is successfully constructed. The construction of this classifier provides reliable support for the subsequent accurate judgment of the type and quantity of medical resource demand in each region.

[0072] In order to effectively improve the accuracy and comprehensiveness of medical resource demand analysis, based on the obtained sample personnel structure information set and sample medical demand parameter set, a random sampling method with replacement was adopted to repeatedly extract data at a uniform sampling ratio (such as 70% of the total sample size each time) to form 10 independent sets of medical demand classification training data.

[0073] Furthermore, for each data set, we also combine the decision tree algorithm with industry standards to adjust parameters to train a separate medical needs classifier. Specifically, a team of medical experts and statisticians adjust parameters such as tree depth and splitting rules to optimize the model, ensuring that its predictions are more aligned with actual medical resource allocation needs.

[0074] Furthermore, after multiple rounds of sampling modeling and the construction of 10 classifiers, cross-validation was used to screen classifiers with excellent performance in indicators such as accuracy and recall rate. Consistency testing was performed to ensure complementarity, and finally a reliable medical demand classifier sequence was obtained, providing effective support for the multivariate analysis and accurate judgment of medical resource needs.

[0075] Cross-validation involves dividing data into multiple groups and repeatedly training and testing them to identify classifiers that are accurate (over 85% accuracy) and consistently predict the best outcomes (over 80% recall). Consistency testing uses the Kappa coefficient (a statistical indicator used to assess classification consistency) to check whether the predictions of the classifiers are consistent. If the Kappa coefficient is above 0.75, it indicates that the two classifiers can complement each other and work together to analyze medical needs.

[0076] After the medical demand classifier is constructed and trained, the primary and secondary personnel structure information is fed into the medical demand classifier sequence. Each classifier performs analysis and outputs a set of predicted primary medical demand parameters, as well as multiple sets of predicted secondary medical demand parameters tailored to different population characteristics, providing rich data for medical resource allocation.

[0077] Furthermore, after obtaining the set of predicted first-level medical demand parameters, the mean is calculated to obtain the first-level medical demand parameters, and the difference between each predicted first-level medical demand parameter and the first-level medical demand parameter in the set is calculated one by one. The absolute value of these differences is taken and the arithmetic average is performed. The ratio of the average value to the first-level medical demand parameter is used as the first-level demand error coefficient to measure the degree of deviation of multiple prediction results from the average situation.

[0078] Similarly, after obtaining multiple sets of predicted secondary medical demand parameters, the mean is calculated respectively to obtain multiple secondary medical demand parameters, and the absolute mean of the difference between the parameters in each set and the corresponding secondary medical demand parameters is calculated, and the ratio of the difference to the secondary medical demand parameter is taken as the secondary demand error coefficient.

[0079] Furthermore, the mean absolute value of the error between each parameter in the predicted first-level medical demand parameter set and the first-level medical demand parameter is calculated, and its ratio to the first-level medical demand parameter is used as the first-level demand error coefficient; the mean absolute value of the error between each parameter in multiple predicted second-level medical demand parameter sets and the corresponding second-level medical demand parameter is simultaneously calculated, and used as multiple second-level demand error coefficients respectively.

[0080] For example, the primary medical demand parameter is 80 general practitioners, and the parameters in the prediction set are 76, 82, and 79. The absolute values ​​of the differences between each predicted value and 80 are calculated (4, 2, and 1). The arithmetic mean is 2.33 ((4+2+1) / 3=2.33). The ratio of this to the primary medical demand parameter is 2.91% (2.33 / 80=2.19%), which is the primary demand error coefficient. The secondary medical demand is 1000 boxes of chronic disease medication for the elderly. The predicted values ​​are 950 and 1020 boxes. The average absolute value of the difference is 35 (((1000-950)+(1020-1000)) / 2=35). The ratio of this to the actual value of 1000 is 3.5% (35 / 1000=3.5%), which is the secondary demand error coefficient.

[0081] S300: performing a skip-level medical treatment probability analysis based on multiple secondary personnel structure information and multiple hierarchical distances to obtain multiple skip-level medical treatment probabilities and multiple skip-level medical treatment error coefficients;

[0082] In an embodiment of the present application, in the process of medical resource management, in order to accurately assess the possibility of patients seeking medical treatment at a higher level and avoid the problem of rural personnel going directly to towns for medical treatment, which leads to waste of rural resources and insufficient urban resources, it is necessary to collect multiple secondary personnel structure information and multiple hierarchical distances.

[0083] Specifically, the secondary personnel structure information includes the health status and medical habits of patients in various regions; the hierarchical distance refers to the spatial distance and service capacity gap between medical institutions at all levels.

[0084] Furthermore, by analyzing these data, we obtained multiple probabilities of skipping medical care, and combined with data change trends and historical deviations, we calculated multiple skipping error coefficients, providing data support for optimizing the distribution of medical resources and balancing urban and rural medical pressures.

[0085] Step S300 in the method provided in the embodiment of the present application includes:

[0086] Combining the plurality of secondary personnel structure information and the plurality of hierarchical distances respectively, inputting the plurality of levels of medical skipping probability predictors into a sequence, and outputting a plurality of predicted levels of medical skipping probability sets;

[0087] Calculating the means of the plurality of predicted skip-level medical treatment probability sets respectively to obtain a plurality of skip-level medical treatment probabilities;

[0088] In each set of predicted skip-level medical treatment probabilities, the average error margins between other predicted skip-level medical treatment probabilities and the skip-level medical treatment probabilities are calculated as multiple skip-level error coefficients.

[0089] In an embodiment of the present application, based on the obtained multiple secondary personnel structure information and multiple hierarchical distance data, the two are combined and input into the skipping probability predictor sequence. After model calculation and analysis, multiple predicted skipping medical probability sets are output to provide a quantitative reference for medical resource allocation.

[0090] The training steps of the "skip-level probability predictor sequence" in the method provided in the embodiment of the present application include:

[0091] Based on the historical data of medical resource allocation, we collected the sample personnel structure information set and the sample classification distance set, and collected the proportion of skipping medical treatment under different sample personnel structure information and sample classification distance, and annotated the sample skipping medical treatment probability set;

[0092] Randomly selecting some data with replacement from the sample personnel structure information set, the sample hierarchical distance set, and the sample skip-level medical treatment probability set to obtain first skip-level medical treatment probability prediction training data, using machine learning to construct a first skip-level medical treatment probability predictor, and using the first skip-level medical treatment probability prediction training data to supervise the training of the first skip-level medical treatment probability predictor until convergence;

[0093] Continue to randomly select multiple groups of skipping probability prediction training data with replacement, train multiple skipping probability predictors, and obtain a skipping probability predictor sequence.

[0094] In the embodiment of the present application, in order to ensure the reliability and comprehensiveness of the sample data, it is necessary to collect data on the sample personnel structure, graded distance and probability of skipping medical care, so that these data can effectively reflect the actual situation.

[0095] Among them, when collecting the sample personnel structure information set, the tiered diagnosis and treatment information management platform can be used to retrieve the health records of residents in various regions and the census data to obtain data such as age distribution, occupational composition, and the proportion of special groups. After data screening and integration, the data can be labeled to form the sample personnel structure information set.

[0096] When collecting sample graded distance sets, the geographic information system (GIS) can be used to measure the physical distances between medical institutions at all levels, and the service differences can be evaluated in combination with service items, reception capabilities, etc. The distance data and service capacity differences can be comprehensively quantified, labeled, and formed into a sample graded distance set.

[0097] When calculating the proportion of times of skipping medical treatment, we can analyze the medical records of the hospital information system (HIS), mark the specific data of cross-level medical treatment, calculate its proportion in the total number of medical treatments, and thus mark and form a sample skipping medical treatment probability set.

[0098] Furthermore, from the sample personnel structure information set, the sample hierarchical distance set and the sample skip-level medical probability set, some data are randomly selected with replacement as the first skip-level probability prediction training data, and machine learning is used to construct the first skip-level probability predictor, and it is brought to a convergence state through supervised training.

[0099] Among them, the construction of the first leapfrog probability predictor is completed through three steps: data preparation, model building and model training.

[0100] First, during the data preparation phase, based on statistical sampling principles, random sampling with replacement was performed from a set of sample personnel structure information, a set of sample hierarchical distances, and a set of sample skipping-level medical probabilities to form a representative first training dataset for predicting skipping-level medical probabilities. This dataset includes multidimensional features such as age distribution, occupational composition, spatial distance between medical institutions, and service capacity gaps, as well as the corresponding actual values ​​of skipping-level medical probabilities.

[0101] Secondly, during the model building phase, a decision tree algorithm was used to construct the model architecture. Specifically, sub-datasets were generated from the training data using sampling with replacement, and multiple decision trees were constructed on each of these sub-datasets. During construction, the decision tree splits the data based on key features, such as age and distance, until the conditions are met and a prediction result is output. The results of multiple trees are then combined to derive a predicted value for the probability of skipped medical care, forming the basic framework of the predictive model.

[0102] Finally, during the model training phase, supervised learning methods are used, based on the training dataset, to continuously adjust the model's internal parameters to optimize the model's predictions, minimizing the error between the final predicted value and the actual value. During this phase, a validation dataset can be used to further verify the model's accuracy. When the model's error on the validation dataset falls within a preset range and the prediction accuracy gradually stabilizes, the model converges, indicating that the first level-skipping probability predictor has been constructed. This predictor can be used to accurately predict the probability of skipping medical treatment.

[0103] Furthermore, on the basis of completing the construction of a single first skip-level probability predictor, the random sampling operation with replacement is repeatedly performed to select multiple groups of different skip-level probability prediction training data, and independently train multiple predictor models, and finally form a skip-level probability predictor sequence composed of multiple predictors, providing a model set for subsequent combined prediction.

[0104] Furthermore, the same test data set is used to perform predictions based on the multiple trained skip-level care probability predictors to obtain multiple sets of predicted skip-level care probabilities. The arithmetic mean of the probability values ​​within each set is then calculated to generate multiple skip-level care probability values.

[0105] Furthermore, within each set of predicted skip-level medical treatment probabilities, the corresponding skip-level medical treatment probability value of the set is taken as the benchmark value, and the absolute errors of other predicted skip-level medical treatment probabilities and the benchmark value are calculated one by one. Then the arithmetic mean of all absolute errors is taken to obtain the average error amplitude corresponding to each set, which is defined as multiple skip-level error coefficients, which are used to measure the degree of deviation of the output results of different predictors from the benchmark values, providing data support for the comprehensive evaluation and optimization of subsequent models.

[0106] For example, taking a community in a certain city as an example, the personnel structure data such as 30% of the population are over 60 years old and 15% of the population are chronic disease patients are obtained from health records; GIS calculates that the distance between the community health station and the tertiary hospital is 5 kilometers, and the latter's reception capacity is 5 times that of the former, forming hierarchical distance data; statistics from the hospital system show that cross-level medical treatment accounts for 20%, which is used as the probability data of skipping medical care.

[0107] Furthermore, a single predictor was trained on a random sample of data, and parameters were adjusted to stabilize the model error within 5%. Ten models were trained repeatedly and tested on the remaining data, yielding 10 sets of probability values. The average of each set was taken to generate a baseline probability. The average error between the predicted values ​​of each model and the baseline was calculated, and the optimal model was selected to predict the probability of residents seeking medical care at a higher level, thus assisting in the rational allocation of medical resources.

[0108] S400: Optimizing the allocation of medical resources based on the primary demand error coefficient, the multiple secondary demand error coefficients, and the multiple skip-level error coefficients, the primary medical demand parameters, the multiple secondary medical demand parameters, and the multiple skip-level medical probabilities, obtaining a medical resource allocation result, and allocating medical resources to the primary region and the multiple secondary regions;

[0109] In the embodiment of the present application, when allocating medical resources, in order to solve the problem of mismatch between medical resource allocation and regional demand, it is necessary to comprehensively evaluate the demand error coefficients at all levels and the possibility of skipping levels of medical treatment.

[0110] Specifically, the primary demand error coefficient reflects the deviation in overall medical demand within a primary region, the secondary demand error coefficient corresponds to the variation in local demand across multiple secondary regions, and the skip-level error coefficient measures the likelihood of patients seeking care at a different level. Furthermore, the primary medical demand parameter and multiple secondary medical demand parameters intuitively reflect the amount of basic medical demand in each region, while the multiple skip-level medical probabilities reflect the likelihood of patients seeking care across regions.

[0111] Furthermore, the above-mentioned error coefficients are integrated with demand parameters and probability data, and weights are allocated according to the size of the errors. More medical resources are invested in areas with large demand errors and high possibility of skipping levels. The equipment configuration, personnel structure and resource investment of medical institutions at all levels are dynamically adjusted, and finally reliable medical resource allocation results are output to achieve efficient and effective allocation of medical resource supply and demand between the first-level region and multiple second-level regions, thereby improving the efficiency of medical resource utilization and service rate.

[0112] Step S400 in the method provided in the embodiment of the present application includes:

[0113] According to the multiple skip-level medical treatment probabilities, the first-level medical demand parameter and the multiple second-level medical demand parameters are corrected and calculated to obtain a corrected first-level medical demand parameter and multiple corrected second-level medical demand parameters;

[0114] Randomly allocating the medical resources to the first-level region and the multiple second-level regions to obtain a first-level region medical resource parameter and multiple first-level second-level region medical resource parameters as a first medical resource allocation result, and calculating a first medical resource allocation fitness based on the first-level demand error coefficient, the multiple second-level demand error coefficients, and the multiple leapfrog error coefficients;

[0115] Continue to randomly allocate medical resources for optimization until convergence, output the medical resource allocation result with the maximum medical resource allocation fitness, and allocate medical resources to the first-level region and multiple second-level regions.

[0116] In the embodiment of the present application, in order to make the medical demand parameters more consistent with the actual medical situation, the first-level medical demand parameters and multiple second-level medical demand parameters are corrected and calculated based on multiple skip-level medical probabilities.

[0117] Specifically, the obtained probability of skipping medical treatment levels was used as an adjustment coefficient, and the scale of basic medical demand was recalculated based on the trend of patients seeking medical treatment at different levels in each region. By weighting the overall demand in the first-level region and the local demand in multiple second-level regions, the adjusted first-level medical demand parameters and multiple adjusted second-level medical demand parameters were ultimately obtained, providing a reliable data foundation for the precise allocation of subsequent medical resources.

[0118] Among them, the weighted correction method is to use the probability of skipping medical care as the weight coefficient. For the secondary region, taking into account the outflow of patients, the initial demand parameter is multiplied by (1-the probability of skipping medical care) for correction; for the primary region, the total amount of medical demand reduction after correction in multiple secondary regions is first calculated (that is, the sum of the initial demand parameters of each secondary region × the probability of skipping medical care), and then the total reduction is added to the primary medical demand parameter to obtain the corrected primary medical demand parameter.

[0119] For example, if the initial demand in Level 2 Region A is 1,000 people and the initial demand in Level 2 Region B is 1,200 people, and the probability of skipping levels of care is 5%, then the adjusted demand in Level 2 Region A is 950 (1,000 × (1-5%) = 950) people, and the adjusted demand in Level 2 Region B is 1,140 (1,200 × (1-5%) = 1,140) people, for a total reduction of 110 ((1,000 - 950) + (1,200 - 1,140) = 110) people. If the initial demand in Level 1 Region is 5,000 people, the adjusted demand is 5,110 (5,000 + 110 = 5,110) people.

[0120] In the method provided in the embodiment of the present application, the step of “randomly allocating the medical resources to the first-level region and the plurality of second-level regions, obtaining a first-level region medical resource parameter and a plurality of first-level second-level region medical resource parameters as a first medical resource allocation result, and calculating a first medical resource allocation fitness in combination with the first-level demand error coefficient, the plurality of second-level demand error coefficients, and the plurality of leapfrog error coefficients” includes:

[0121] Calculating and obtaining a plurality of secondary error coefficients according to the plurality of secondary demand error coefficients and the plurality of leapfrog error coefficients;

[0122] Calculate the average of the multiple leapfrog error coefficients to obtain an average leapfrog error coefficient, and combine it with the first-level demand error coefficient to calculate a first-level error coefficient;

[0123] Allocating and obtaining a plurality of regional weights according to the plurality of secondary error coefficients and the primary error coefficients;

[0124] Respectively calculating the ratios of the first-level regional medical resource parameter and the plurality of first-level and second-level regional medical resource parameters to the corrected first-level medical demand parameter and the plurality of corrected second-level medical demand parameters to obtain a first-level allocation coefficient and a plurality of first-level and second-level allocation coefficients;

[0125] The first first-level allocation coefficient and the multiple first-level second-level allocation coefficients are weightedly calculated using the multiple regional weights to obtain a first medical resource allocation adaptability.

[0126] In the embodiment of the present application, in order to comprehensively evaluate the matching degree between medical needs and resource allocation and avoid skipping the level of medical treatment, it is necessary to integrate and calculate the secondary demand error coefficient and multiple skipping error coefficients, scientifically set the weights, and build a reliable error assessment system.

[0127] Specifically, multiple secondary demand error coefficients and multiple skip-level error coefficients are calculated uniformly. Using a summation method, the errors caused by fluctuations in medical demand within the secondary region are added to the risk errors caused by patients skipping levels of care to calculate the secondary error coefficient. That is, the secondary error coefficient = the secondary demand error coefficient + the multiple skip-level error coefficients. This coefficient can fully reflect the actual deviations in medical demand in each region.

[0128] Furthermore, multiple skip-level error coefficients are aggregated and averaged to obtain the average skip-level error coefficient, which is then combined with the first-level demand error coefficient for unified calculation. This summation method is used to integrate the errors caused by skip-level medical treatment risks with the overall demand deviation in the first-level region to obtain the first-level error coefficient: the first-level error coefficient = the average skip-level error coefficient + the first-level demand error coefficient. This coefficient comprehensively reflects the overall demand deviation of medical resources.

[0129] Furthermore, regional weight distribution is performed based on the multiple secondary error coefficients and primary error coefficients obtained.

[0130] Specifically, different weight allocation rules (the larger the error, the greater the weight) are determined based on the degree of medical demand deviation reflected by the coefficient, combined with actual conditions such as regional population size and the carrying capacity of medical institutions. Based on the size of the error coefficient, a proportional allocation method is used to decompose the overall weight, ultimately obtaining multiple regional weights that reflect the shortage and abundance of medical resources in each region, laying the foundation for the subsequent rational allocation of medical resources.

[0131] Furthermore, we will conduct specific calculations of medical resource parameters. Specifically, we will collect resource data such as the number of beds, medical staff, and medical equipment in first-tier regions. This data will be combined with operational indicators such as the average daily number of patients seen and the population covered by medical services. These data will be standardized and summarized to form the first-tier region's medical resource parameters.

[0132] Similarly, multiple first- and second-level regions use the same data collection standards to separately count the scale of primary hospitals, specialized diagnosis and treatment capabilities, public health resource reserves, etc. in each region. After data standardization and weighted calculation, multiple corresponding first- and second-level regional medical resource parameters are generated.

[0133] Furthermore, the calculated first-level regional medical resource parameters are divided by the corrected first-level medical demand parameters to obtain the ratio of the two. This ratio is the first-level allocation coefficient, which is used to measure the degree of matching between the supply and demand of medical resources in the first-level region.

[0134] At the same time, the medical resource parameters of each first- and second-level region are divided by the corresponding corrected second-level medical demand parameters to obtain multiple corresponding first- and second-level allocation coefficients. These coefficients can clearly reflect the supply and demand adaptation of medical resources in the second-level region, and provide a quantitative reference for the formulation of subsequent resource allocation strategies.

[0135] Furthermore, the obtained multiple regional weights are used as weighting basis to perform weighted calculation on the first-level distribution coefficient and multiple first-level and second-level distribution coefficients.

[0136] Specifically, the first-level allocation coefficient is multiplied by the corresponding weight of the first-level region. Each first-level and second-level allocation coefficient is multiplied by the weight of the corresponding second-level region. These products are then summed up to obtain the final value, which is the first medical resource allocation adaptability. This value intuitively reflects the degree of fit between the current medical resource allocation plan and actual needs, providing a reliable assessment basis for optimizing resource allocation.

[0137] Furthermore, based on the obtained first medical resource allocation fitness, multiple rounds of iterative adjustment and optimization are further carried out. By randomly adjusting the medical resource allocation parameters of various levels of regions and recalculating the allocation fitness, more optimal resource allocation plans are continuously obtained.

[0138] Specifically, the basic parameters are first set, that is, the initial value of the iteration count is set to 0, and the convergence judgment criteria are constructed, including setting the fitness change threshold (such as 0.001) as the basis for judging small changes, and establishing the maximum number of consecutive invalid iterations (such as 50 times) as the termination condition of the optimization. At the same time, the fitness value calculated for the first time is marked as the initial optimal solution.

[0139] During this iterative process, regional medical resource allocation parameters at all levels are dynamically adjusted through a "random adjustment" approach. Resource indicators such as bed capacity and medical staffing in first-level regions are randomly increased or decreased by 5%, while resource allocation structures for categories like medical equipment procurement and drug reserves in second-level regions are adjusted. After each adjustment, the allocation coefficients for each level of region are recalculated based on the updated resource allocation parameters. Combined with the established regional weighting system, the fitness for the new round of medical resource allocation is calculated using a weighted summation algorithm.

[0140] After each round of iteration is completed, the latest calculated fitness value is compared and analyzed with the historical optimal value. If the latest fitness value is better than the historical optimal value, the optimal solution is updated to the latest comfort and the current resource allocation plan is saved; if the latest fitness value does not exceed the historical optimal value, the number of invalid iterations is recorded.

[0141] Furthermore, when the number of consecutive invalid iterations reaches a preset upper limit (≥50 times) and the deviation between the current fitness value and the historical optimal value is less than the set threshold (≤0.001), the optimization process is judged to have reached a convergence state and the iteration is terminated immediately.

[0142] Furthermore, a resource allocation plan is output to maximize the adaptability of medical resource allocation, clarify the specific allocation amounts of various resource categories such as medical equipment, human resources, drugs and consumables in the first-level region and multiple second-level regions, and use this as the final implementation plan to complete the precise allocation of medical resources at all levels and achieve efficient adaptation between medical resource allocation and regional medical service needs.

[0143] For example, taking a community in a certain city as an example, after processing data such as personnel structure and hierarchical distance, the secondary demand error coefficients for the three secondary areas A, B, and C are 0.09, 0.07, and 0.05, respectively, and the leapfrog error coefficients are 0.04, 0.03, and 0.02, respectively. Calculated by the summation method, the secondary error coefficient of area A is 0.13 (0.09 + 0.04 = 0.13), and similarly, it is 0.1 for area B and 0.07 for area C. Summarizing the leapfrog error coefficients and calculating the average leapfrog error coefficient is 0.03 ((0.04 + 0.03 + 0.02) / 3 = 0.03). When summed with the primary demand error coefficient of 0.05, the primary error coefficient is 0.08 (0.03 + 0.05 = 0.08).

[0144] Furthermore, after steps such as parameter measurement, coefficient calculation, and 50 iterative optimizations, the final resource allocation plan was output. For example, the first-level regional tertiary hospital added 2 ventilators and 5 medical staff, Community A added 3 sets of chronic disease equipment, Community B added 4 beds, and Community C increased its drug reserves by 20%.

[0145] The embodiments of the present application achieve the following technical effects through the above specific implementation methods:

[0146] The embodiment of the present application provides a method for optimizing the allocation of medical resources for hierarchical diagnosis and treatment. First, through the hierarchical diagnosis and treatment information management platform, the hierarchical diagnosis and treatment information management platform and the geographic information system are used to collect data on medical manpower, materials and other resources, and a standardized resource information database is established. At the same time, by measuring data such as geographical straight-line distance and traffic travel distance, a hierarchical distance data set is formed, which provides a spatial decision-making basis for resource allocation. Secondly, by connecting the medical institution information system and the health archive, the personnel structure data of each region are collected, and the medical demand classifier sequence and the skipping probability predictor sequence are used to obtain the medical demand parameters, error coefficient and skipping medical probability, accurately quantify the specific values ​​of medical demand in each region, and scientifically evaluate the possibility of patients skipping medical treatment. Finally, based on the error coefficient and demand parameters, the demand scale is corrected, and combined with the random allocation and iterative optimization mechanism, the regional medical resource configuration parameters at all levels are dynamically adjusted to obtain the optimal solution for the medical resource allocation plan.

[0147] The methods and systems provided in the embodiments of this application address the challenges of unbalanced and inappropriate allocation of traditional medical resources. By collecting diverse analytical data and taking into account dynamic data such as geographic distance, population structure, and skipped-level medical consultations, the system avoids a mismatch between medical resource allocation and actual needs, improves the accuracy and adaptability of allocation, and effectively promotes efficient and balanced allocation of medical resources.

[0148] Example 2, as shown in the attached Figure 2 As shown, based on the inventive concept of a medical resource allocation optimization method for hierarchical diagnosis and treatment provided in Example 1, this application also provides a medical resource allocation optimization system for hierarchical diagnosis and treatment, specifically including:

[0149] Resource and distance collection module 01 is used to collect medical resources to be allocated in hierarchical diagnosis and treatment, wherein the hierarchical diagnosis and treatment includes a primary region and multiple secondary regions, and there are multiple hierarchical distances between the primary region and the multiple secondary regions;

[0150] Medical demand analysis module 02 is used to collect the primary personnel structure information and the secondary personnel structure information of the primary region and multiple secondary regions, perform medical demand analysis, and obtain primary medical demand parameters and multiple secondary medical demand parameters, as well as primary demand error coefficients and multiple secondary demand error coefficients;

[0151] The skip-level medical treatment probability analysis module 03 is used to perform a skip-level medical treatment probability analysis based on multiple secondary personnel structure information and multiple hierarchical distances to obtain multiple skip-level medical treatment probabilities and multiple skip-level medical treatment error coefficients;

[0152] The resource allocation optimization and decision module 04 is used to optimize the allocation of medical resources based on the first-level demand error coefficient, multiple second-level demand error coefficients and multiple skip-level error coefficients, based on the first-level medical demand parameters, multiple second-level medical demand parameters and multiple skip-level medical probabilities, obtain medical resource allocation results, and allocate medical resources to the first-level region and multiple second-level regions.

[0153] In one embodiment, the resource and distance collection module 01 is further configured to:

[0154] Collecting medical resources to be allocated in hierarchical diagnosis and treatment, wherein the medical resources include multiple types of resources, and the hierarchical diagnosis and treatment includes a primary region and multiple secondary regions;

[0155] The geographical distances between the first-level region and the plurality of second-level regions are acquired and collected to obtain a plurality of hierarchical distances.

[0156] In one embodiment, the medical needs analysis module 02 is further configured to:

[0157] Collecting primary personnel structure information and secondary personnel structure information of the primary region and multiple secondary regions, wherein each personnel structure information includes the number of people in multiple different age ranges;

[0158] Input the primary personnel structure information and the multiple secondary personnel structure information into a medical demand classifier sequence respectively, and output a predicted primary medical demand parameter set and multiple predicted secondary medical demand parameter sets;

[0159] Calculating means based on the predicted primary medical demand parameter set and the multiple predicted secondary medical demand parameter sets to obtain a primary medical demand parameter and multiple secondary medical demand parameters;

[0160] Calculate the average error between other predicted first-level medical demand parameters in the predicted first-level medical demand parameter set and the first-level medical demand parameter as the first-level demand error coefficient;

[0161] The average error margins between other predicted secondary medical demand parameters and the secondary medical demand parameters in the plurality of predicted secondary medical demand parameter sets are calculated respectively as a plurality of secondary demand error coefficients.

[0162] In one embodiment, the skip-level probability analysis module 03 is further configured to:

[0163] Combining the plurality of secondary personnel structure information and the plurality of hierarchical distances respectively, inputting the plurality of levels of medical skipping probability predictors into a sequence, and outputting a plurality of predicted levels of medical skipping probability sets;

[0164] Calculating the means of the plurality of predicted skip-level medical treatment probability sets respectively to obtain a plurality of skip-level medical treatment probabilities;

[0165] In each set of predicted skip-level medical treatment probabilities, the average error margins between other predicted skip-level medical treatment probabilities and the skip-level medical treatment probabilities are calculated as multiple skip-level error coefficients.

[0166] In one embodiment, the resource allocation optimization and decision module 04 is further configured to:

[0167] According to the multiple skip-level medical treatment probabilities, the first-level medical demand parameter and the multiple second-level medical demand parameters are corrected and calculated to obtain a corrected first-level medical demand parameter and multiple corrected second-level medical demand parameters;

[0168] Randomly allocating the medical resources to the first-level region and the multiple second-level regions to obtain a first-level region medical resource parameter and multiple first-level second-level region medical resource parameters as a first medical resource allocation result, and calculating a first medical resource allocation fitness based on the first-level demand error coefficient, the multiple second-level demand error coefficients, and the multiple leapfrog error coefficients;

[0169] Continue to randomly allocate medical resources for optimization until convergence, output the medical resource allocation result with the maximum medical resource allocation fitness, and allocate medical resources to the first-level region and multiple second-level regions.

[0170] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0171] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application.

[0172] This specification and drawings are merely illustrative of the present application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Obviously, those skilled in the art may make various modifications and variations to this application without departing from the scope of this application. Thus, this application is intended to include such modifications and variations as fall within the scope of this application and its equivalents.

Claims

1. A method for optimizing the allocation of medical resources for hierarchical diagnosis and treatment, characterized in that: The method comprises: Collecting medical resources to be allocated in hierarchical diagnosis and treatment, wherein the hierarchical diagnosis and treatment includes a primary region and multiple secondary regions, and multiple hierarchical distances exist between the primary region and the multiple secondary regions; Collecting primary personnel structure information and secondary personnel structure information of the primary region and multiple secondary regions, performing medical demand analysis, and obtaining primary medical demand parameters and multiple secondary medical demand parameters, as well as primary demand error coefficients and multiple secondary demand error coefficients; Based on multiple secondary personnel structure information and multiple hierarchical distances, a probability analysis of skipping medical care is performed to obtain multiple skipping medical care probabilities and multiple skipping error coefficients; According to the first-level demand error coefficient, multiple second-level demand error coefficients and multiple skip-level error coefficients, based on the first-level medical demand parameters, multiple second-level medical demand parameters and multiple skip-level medical probabilities, the allocation of medical resources is optimized to obtain medical resource allocation results, and medical resources are allocated to the first-level region and multiple second-level regions.

2. The method for optimizing the allocation of medical resources for hierarchical diagnosis and treatment according to claim 1, characterized in that: Collect medical resources to be allocated in hierarchical diagnosis and treatment, including: Collecting medical resources to be allocated in hierarchical diagnosis and treatment, wherein the medical resources include multiple types of resources, and the hierarchical diagnosis and treatment includes a primary region and multiple secondary regions; The geographical distances between the first-level region and the plurality of second-level regions are acquired and collected to obtain a plurality of hierarchical distances.

3. The method for optimizing the allocation of medical resources for hierarchical diagnosis and treatment according to claim 1, characterized in that: Collecting the primary personnel structure information and the secondary personnel structure information of the primary region and the multiple secondary regions, performing medical demand analysis, and obtaining the primary medical demand parameters and the multiple secondary medical demand parameters, as well as the primary demand error coefficient and the multiple secondary demand error coefficients, including: Collecting primary personnel structure information and secondary personnel structure information of the primary region and multiple secondary regions, wherein each personnel structure information includes the number of people in multiple different age ranges; Input the primary personnel structure information and the multiple secondary personnel structure information into a medical demand classifier sequence respectively, and output a predicted primary medical demand parameter set and multiple predicted secondary medical demand parameter sets; Calculating means based on the predicted primary medical demand parameter set and the multiple predicted secondary medical demand parameter sets to obtain a primary medical demand parameter and multiple secondary medical demand parameters; Calculate the average error between other predicted first-level medical demand parameters in the predicted first-level medical demand parameter set and the first-level medical demand parameter as the first-level demand error coefficient; The average error margins between other predicted secondary medical demand parameters and the secondary medical demand parameters in the plurality of predicted secondary medical demand parameter sets are calculated respectively as a plurality of secondary demand error coefficients.

4. The method for optimizing the allocation of medical resources for hierarchical diagnosis and treatment according to claim 3, characterized in that: The training steps of the medical need classifier sequence include: Based on the historical data of medical resource allocation, a set of sample personnel structure information is collected, and the medical resources required for each sample personnel structure information are marked to obtain a set of sample medical demand parameters; randomly selecting a portion of data with replacement from the sample personnel structure information set and the sample medical demand parameter set to obtain first medical demand classification training data, constructing a first medical demand classifier using machine learning, and performing supervised training on the first medical demand classifier using the first medical demand classification training data until convergence; Continue to randomly select multiple groups of medical demand classification training data with replacement, train multiple medical demand classifiers, and obtain a medical demand classifier sequence.

5. The method for optimizing allocation of medical resources for hierarchical diagnosis and treatment according to claim 1, characterized in that: Based on multiple secondary personnel structure information and multiple hierarchical distances, we conduct a probability analysis of skipping medical care and obtain multiple skipping medical care probabilities and multiple skipping error coefficients, including: Combining the plurality of secondary personnel structure information and the plurality of hierarchical distances respectively, inputting the plurality of levels of medical skipping probability predictors into a sequence, and outputting a plurality of predicted levels of medical skipping probability sets; Calculating the means of the plurality of predicted skip-level medical treatment probability sets respectively to obtain a plurality of skip-level medical treatment probabilities; In each set of predicted skip-level medical treatment probabilities, the average error margins between other predicted skip-level medical treatment probabilities and the skip-level medical treatment probabilities are calculated as multiple skip-level error coefficients.

6. The method for optimizing allocation of medical resources for hierarchical diagnosis and treatment according to claim 5, characterized in that: The training steps of the skip-level probability predictor sequence include: Based on the historical data of medical resource allocation, we collected the sample personnel structure information set and the sample classification distance set, and collected the proportion of skipping medical treatment under different sample personnel structure information and sample classification distance, and annotated the sample skipping medical treatment probability set; Randomly selecting some data with replacement from the sample personnel structure information set, the sample hierarchical distance set, and the sample skip-level medical treatment probability set to obtain first skip-level medical treatment probability prediction training data, using machine learning to construct a first skip-level medical treatment probability predictor, and using the first skip-level medical treatment probability prediction training data to supervise the training of the first skip-level medical treatment probability predictor until convergence; Continue to randomly select multiple groups of skipping probability prediction training data with replacement, train multiple skipping probability predictors, and obtain a skipping probability predictor sequence.

7. The method for optimizing allocation of medical resources for hierarchical diagnosis and treatment according to claim 1, characterized in that: According to the multiple secondary demand error coefficients and the multiple skip-level error coefficients, based on the primary medical demand parameters, the multiple secondary medical demand parameters, and the multiple skip-level medical probabilities, the allocation of the medical resources is optimized to obtain a medical resource allocation result, and the medical resources are allocated to the primary region and the multiple secondary regions, including: According to the multiple skip-level medical treatment probabilities, the first-level medical demand parameter and the multiple second-level medical demand parameters are corrected and calculated to obtain a corrected first-level medical demand parameter and multiple corrected second-level medical demand parameters; Randomly allocating the medical resources to the first-level region and the multiple second-level regions to obtain a first-level region medical resource parameter and multiple first-level second-level region medical resource parameters as a first medical resource allocation result, and calculating a first medical resource allocation fitness based on the first-level demand error coefficient, the multiple second-level demand error coefficients, and the multiple leapfrog error coefficients; Continue to randomly allocate medical resources for optimization until convergence, output the medical resource allocation result with the maximum medical resource allocation fitness, and allocate medical resources to the first-level region and multiple second-level regions.

8. The method for optimizing allocation of medical resources for hierarchical diagnosis and treatment according to claim 7, characterized in that: The medical resources are randomly allocated to the first-level region and multiple second-level regions to obtain a first medical resource allocation result, and the first medical resource allocation fitness is calculated based on the first-level demand error coefficient, multiple second-level demand error coefficients, and multiple leapfrog error coefficients, including: Calculating and obtaining a plurality of secondary error coefficients according to the plurality of secondary demand error coefficients and the plurality of leapfrog error coefficients; Calculate the average of the multiple leapfrog error coefficients to obtain an average leapfrog error coefficient, and combine it with the first-level demand error coefficient to calculate a first-level error coefficient; Allocating and obtaining a plurality of regional weights according to the plurality of secondary error coefficients and the primary error coefficients; Respectively calculating the ratios of the first-level regional medical resource parameter and the plurality of first-level and second-level regional medical resource parameters to the corrected first-level medical demand parameter and the plurality of corrected second-level medical demand parameters to obtain a first-level allocation coefficient and a plurality of first-level and second-level allocation coefficients; The first first-level allocation coefficient and the multiple first-level second-level allocation coefficients are weightedly calculated using the multiple regional weights to obtain a first medical resource allocation adaptability.

9. A medical resource allocation optimization system for hierarchical diagnosis and treatment, characterized by: The system is used to execute the medical resource allocation optimization method according to any one of claims 1 to 8, and the system comprises: A resource and distance collection module is used to collect medical resources to be allocated in hierarchical diagnosis and treatment, wherein the hierarchical diagnosis and treatment includes a primary region and multiple secondary regions, and there are multiple hierarchical distances between the primary region and the multiple secondary regions; a medical demand analysis module, configured to collect primary personnel structure information and secondary personnel structure information of the primary region and multiple secondary regions, perform medical demand analysis, and obtain primary medical demand parameters and multiple secondary medical demand parameters, as well as primary demand error coefficients and multiple secondary demand error coefficients; A skip-level medical treatment probability analysis module is used to perform a skip-level medical treatment probability analysis based on multiple secondary personnel structure information and multiple hierarchical distances, and obtain multiple skip-level medical treatment probabilities and multiple skip-level medical treatment error coefficients; The resource allocation optimization and decision-making module is used to optimize the allocation of medical resources based on the first-level demand error coefficient, multiple second-level demand error coefficients and multiple skip-level error coefficients, based on the first-level medical demand parameters, multiple second-level medical demand parameters and multiple skip-level medical probabilities, obtain medical resource allocation results, and allocate medical resources to the first-level region and multiple second-level regions.

Images