Medical health service work order knowledge base collaborative processing method and system

Abstract

The application provides a medical health service work order knowledge base collaborative processing method and system, relates to the technical field of medical data processing, and the method comprises the following steps: receiving a medical health service work order; analyzing the chief complaint and symptoms and signs of a patient in the medical health service work order to obtain an analysis result; obtaining structured processing work order characteristic data according to the analysis result; searching disease positioning detection data in a preset knowledge base according to the symptoms and signs in the structured processing work order characteristic data to obtain a search result; obtaining a matched disease characteristic set according to the search result; extracting historical health evaluation results in a health trend map according to the matched disease characteristic set; and obtaining multi-dimensional health state distribution characteristics according to the extraction result. The application improves the efficiency and quality of medical health services.

Description

Technical Field

[0001] This invention relates to the field of medical data processing technology, and in particular to a collaborative processing method and system for a medical and health service work order knowledge base. Background Technology

[0002] In the existing healthcare service system, the healthcare service work order is the core carrier of a patient's chief complaint and symptoms. The work order processing workflow typically involves entities such as a pre-set knowledge base, a health trend map, and a biobank. The pre-set knowledge base stores disease localization and detection data. The health trend map records historical health assessment results. The biobank is used to associate standard treatment pathways. The system parses the work order information and retrieves relevant data to ultimately determine the corresponding assigned doctor and department group, completing the work order's flow and dispatch.

[0003] In the current process of processing medical and health service work orders, when generating work order transfer instructions, existing technologies usually only allocate resources for a single department or local treatment segment. This approach lacks overall cumulative accounting of standard treatment pathways in the biobank across the entire space. Because it is impossible to comprehensively assess the cumulative resource consumption and transfer load of each node throughout the entire treatment cycle, the accuracy of work order transfer instructions generated by existing technologies is insufficient. Summary of the Invention

[0004] This invention provides a collaborative processing method and system for a medical and health service work order knowledge base, which improves the efficiency and quality of medical and health services.

[0005] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows: Firstly, a collaborative processing method for a medical and health service work order knowledge base, the method comprising: Step 1: Receive medical and health service work orders; parse the patient's chief complaint and symptoms in the medical and health service work orders to obtain the parsing results; based on the parsing results, obtain structured work order feature data; Step 2: Based on the symptoms and signs in the structured work order feature data, retrieve the disease localization detection data from the preset knowledge base to obtain the retrieval results; based on the retrieval results, obtain the matching disease feature set; Step 3: Extract historical health assessment results from the health trend map based on the matched disease feature set; obtain multi-dimensional health status distribution features based on the extraction results; Step 4: Based on the multi-dimensional distribution characteristics of health status, deduce the correlation between corresponding preventive healthcare measures and preliminary treatment recommendations segment by segment to obtain the deduction results; based on the deduction results, obtain a preliminary collaborative treatment plan. Step 5: Based on the preliminary collaborative treatment plan and the standard treatment pathways in the biobank, perform a comprehensive summary and calculation to obtain the calculation results; Step 6: Based on the calculation results, obtain the final medical and health service work order transfer instructions; Step 7: Based on the final medical and health service work order flow instructions, match the corresponding assigned doctor and department group to obtain the allocation result; based on the matching result, obtain the work order dispatch parameters; Step 8: Update the processing status and flow nodes of the medical and health service work orders according to the work order dispatch parameters to obtain the update results; based on the update results, obtain the work order records that have been processed.

[0006] Secondly, the collaborative processing system for the medical and health service work order knowledge base includes: The work order receiving and parsing module is used to receive medical and health service work orders; parse the patient's chief complaint and symptoms and signs in the medical and health service work orders to obtain the parsing results; and obtain structured work order feature data based on the parsing results. The knowledge base retrieval and matching module is used to retrieve disease localization detection data from a preset knowledge base based on the symptoms and signs in the structured work order feature data to obtain retrieval results; and to obtain a set of matching disease features based on the retrieval results. The health trend analysis module is used to extract historical health assessment results from the health trend map based on the matched disease feature set; and to obtain multi-dimensional health status distribution characteristics based on the extraction results. The collaborative diagnosis and treatment derivation module is used to deduce the correlation between corresponding preventive health care measures and preliminary diagnosis and treatment suggestions segment by segment based on the multi-dimensional health status distribution characteristics, so as to obtain the derivation results; based on the derivation results, a preliminary collaborative diagnosis and treatment plan is obtained. The global summary and accounting module is used to perform global summary and accounting based on the preliminary collaborative diagnosis and treatment plan and the standard treatment pathway in the biobank to obtain the accounting results; The workflow instruction generation module is used to obtain the final medical and health service work order workflow instructions based on the calculation results; The doctor and department matching module is used to match the corresponding doctor and department group according to the final medical and health service work order flow instructions to obtain the matching result; and to obtain the work order dispatch parameters based on the matching result. The work order status archiving module is used to update the processing status and flow nodes of medical and health service work orders according to the work order dispatch parameters to obtain the update results; based on the update results, the work order records that have been processed are obtained.

[0007] The above-described solution of the present invention has at least the following beneficial effects: By extracting historical health assessment results from the health trend map, multi-dimensional health status distribution characteristics are obtained. Based on this, the correlation between preventive healthcare measures and preliminary treatment recommendations is deduced. Then, combined with standard treatment pathways in the biobank, a comprehensive summary and calculation is performed to generate a final work order flow instruction that takes into account both individual health dynamics and standardized treatment requirements. This improves the accuracy of medical work order assignment and collaborative treatment plans, reduces triage errors caused by ignoring historical health trends and standard pathway deviations, and improves the efficiency and quality of medical and health services. Attached Figure Description

[0008] Figure 1 This is a flowchart illustrating the collaborative processing method for a medical and health service work order knowledge base provided in an embodiment of the present invention.

[0009] Figure 2 This is a schematic diagram of a collaborative processing system for a medical and health service work order knowledge base provided in an embodiment of the present invention. Detailed Implementation

[0010] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0011] like Figure 1 As shown, embodiments of the present invention propose a collaborative processing method for a medical and health service work order knowledge base, the method comprising the following steps: Step 1: Receive medical and health service work orders; parse the patient's chief complaint and symptoms in the medical and health service work orders to obtain the parsing results; based on the parsing results, obtain structured work order feature data; Step 2: Based on the symptoms and signs in the structured work order feature data, retrieve the disease localization detection data from the preset knowledge base to obtain the retrieval results; based on the retrieval results, obtain the matching disease feature set; Step 3: Extract historical health assessment results from the health trend map based on the matched disease feature set; obtain multi-dimensional health status distribution features based on the extraction results; Step 4: Based on the multi-dimensional distribution characteristics of health status, deduce the correlation between corresponding preventive healthcare measures and preliminary treatment recommendations segment by segment to obtain the deduction results; based on the deduction results, obtain a preliminary collaborative treatment plan. Step 5: Based on the preliminary collaborative treatment plan and the standard treatment pathways in the biobank, perform a comprehensive summary and calculation to obtain the calculation results; Step 6: Based on the calculation results, obtain the final medical and health service work order transfer instructions; Step 7: Based on the final medical and health service work order flow instructions, match the corresponding assigned doctor and department group to obtain the allocation result; based on the matching result, obtain the work order dispatch parameters; Step 8: Update the processing status and flow nodes of the medical and health service work orders according to the work order dispatch parameters to obtain the update results; based on the update results, obtain the work order records that have been processed.

[0012] In this embodiment of the invention, by analyzing the patient's chief complaint and physical signs and retrieving the knowledge base, the precise structuring of patient symptoms and rapid identification of disease characteristics are achieved, effectively improving the efficiency and accuracy of preliminary diagnosis. Furthermore, by combining historical assessment results with a health trend map, a multi-dimensional distribution of health status characteristics is constructed, overcoming the limitations of single-symptom judgment and providing data support for personalized health tracking. Based on this, the correlation between preventive healthcare and treatment recommendations is derived segment by segment, generating a preliminary collaborative treatment plan, realizing a forward-looking layout for medical intervention. Through comprehensive aggregation and calculation of standard treatment pathways in the biobank, the preliminary plan is validated for authority and feasibility, ensuring the scientific rigor of the final treatment recommendations. Finally, by generating transfer instructions and accurately matching and assigning doctors and department groups, the automated transfer and closed-loop update of medical service work orders in the professional resource pool are realized, shortening patient waiting time and improving the collaborative processing efficiency and overall service quality of medical and health services.

[0013] In a preferred embodiment of the present invention, step 1 above may include: Step 1.1: Based on the received medical and health service work order, extract the patient's chief complaint text field, symptom description text field, and vital sign value field from the work order form to obtain the original field extraction results. Specifically, this includes: according to the received medical and health service work order, following the predefined field mapping rules of the work order form, which refers to a set of pre-configured correspondence specifications between field names and actual stored variables in the system, this specification is based on the hierarchical structure of the data exchange format adopted by the work order form, mapping each field label in the work order form to its corresponding storage variable according to its data area, sequentially locating and extracting the text content contained in the patient's chief complaint text field, the text content contained in the symptom description text field, and the measurement values ​​contained in the vital sign value field of the work order form, and storing the three into their respective field variables. The predefined field mapping rules of the work order form specify the correspondence between each field name and the actual stored variable in the work order data exchange format. For the extracted patient complaint text field and symptom description text field, punctuation marks, redundant whitespace characters, and non-textual noise characters are filtered out, while continuous semantic text content is retained. For the extracted vital sign index numerical fields, the numerical format is verified to be consistent with the preset numerical format template. Abnormal values ​​with non-compliant formats are removed, and vital sign index values ​​that conform to the numerical format template are retained. Thus, the original field extraction results containing the cleaned patient complaint text, the cleaned symptom description text, and the verified list of vital sign values ​​are obtained.

[0014] Step 1.2: Based on the original field extraction results, perform medical terminology segmentation and medical entity recognition processing on the patient's chief complaint text field and the symptom description text field to obtain preliminary parsing results containing medical entity labels and entity boundaries. Specifically, based on the obtained original field extraction results, perform the following processing on the patient's chief complaint text field and the symptom description text field respectively. First, perform medical terminology segmentation processing on the two text fields: using a preset medical segmentation dictionary as the segmentation basis, adopt a bidirectional maximum matching segmentation algorithm to segment the text into words in both directions from left to right and from right to left. Compare the segmentation results in the two directions, and prioritize the result with fewer segmented words as the segmentation output; if the number of segmented words in both directions is the same, prioritize the segmentation result with fewer single-character words, thereby obtaining the segmentation sequence.

[0015] Based on the obtained word segmentation sequence, medical entity recognition processing is performed on the word segmentation sequence: each word in the word segmentation sequence is compared one by one with a preset medical entity type dictionary. If the current word completely matches a standard entity word in the medical entity type dictionary, the current word is labeled as the entity type label corresponding to the standard entity word, including one of the following: chief complaint entity type, symptom entity type, anatomical location entity type, time description entity type, and degree description entity type. The start and end positions of the word in the original text are recorded as entity boundaries. If the current word does not completely match any entry, the edit distance similarity between the current word and each standard entity word in the medical entity type dictionary is calculated. The standard entity word with the highest similarity greater than the preset entity matching threshold is selected, and the current word is labeled as the entity type corresponding to the standard entity word. At the same time, the entity boundaries are recorded.

[0016] Edit distance similarity Calculate as follows: ; in: This represents the edit distance between the current term and the standard entity term, which is the minimum number of editing operations required to convert the current term into a standard entity term using the three editing operations of insertion, deletion, and replacement. Indicates the character length of the current term; Indicates the character length of a standard entity term; This indicates that the larger of the two values ​​is taken.

[0017] Based on the original field extraction results, after word segmentation and entity recognition processing, preliminary parsing results containing the medical entity labels and entity boundaries corresponding to each term are obtained.

[0018] Step 1.3: Based on the preliminary analysis results, each identified medical entity is sequentially mapped to the corresponding standard term in the preset standard medical terminology lexicon to obtain the terminology standardization processing results. Specifically, this includes: based on the obtained preliminary analysis results, mapping the original text form entity designation of each labeled medical entity to the preset standard medical terminology lexicon.

[0019] For each medical entity in the preliminary analysis results, its entity referential text and entity type label are extracted. Using the entity type label as the retrieval condition, the terminology sub-table corresponding to the entity type is located in the standard medical terminology thesaurus. In the terminology sub-table, the semantic matching degree between the entity referential text and each standard term in the terminology sub-table is calculated. The semantic matching degree is calculated by comprehensively considering the following two dimensions: the first dimension is the literal matching degree, which is measured by the ratio of the length of the longest common subsequence of the two strings to the length of the longer of the two strings; the second dimension is the semantic vector matching degree, which is obtained by inputting the entity referential text and the standard term into a pre-trained medical text semantic encoding model, obtaining their respective semantic vector representations, and then calculating the cosine distance between the two semantic vectors as the semantic vector matching degree.

[0020] Overall matching degree Calculate as follows: ; in: The weighting coefficient represents the literal matching degree, and its value ranges from zero to one; The weight coefficients representing the semantic vector matching degree range from 0 to 1 and satisfy the following conditions: and The sum of them equals one; The character length of the longest common subsequence between the entity reference text and the standard terminology entry; Indicates the length of the text referring to the entity; The character length of a standard term entry; Represents entity referential text semantic vectors semantic vectors of standard terminology entries The cosine distance value; This represents the semantic vector output after the entity referential text is encoded by the medical text semantic coding model; This represents the semantic vector output after standard terminology entries are encoded by the same medical text semantic coding model.

[0021] Select overall matching degree The standard term entry with the highest value is used as the standardized mapping result of the current medical entity. The original entity reference text is replaced with the standard term entry, while retaining the original entity type label and entity boundary information, thus obtaining the terminology standardization processing result.

[0022] Step 1.4: Based on the terminology standardization processing results, the standardized complaint entities, symptom entities, and sign values ​​are assembled into key-value pair feature data according to a preset structured template. Each feature item is labeled with a data type tag to obtain structured work order feature data. Specifically, this includes assembling the standardized entity data according to the preset structured template based on the obtained terminology standardization processing results. The preset structured template defines multiple feature slots, each containing three attributes: slot name, expected entity type, and value type. Specifically, the slot names include complaint entity slots, symptom entity slots, anatomical location entity slots, time description entity slots, severity description entity slots, and sign value slots; the value types are divided into text type and numerical type.

[0023] During the assembly process, each medical entity in the terminology standardization results is filled into the corresponding feature slot in the structured template according to its labeled entity type. For text slots, the standard terminology text is directly filled as the slot value; for vital sign value slots, the verified vital sign values ​​are paired with the standardized vital sign indicator names and filled into the corresponding slots according to the vital sign indicator names. After the slot filling is completed, each valid filled feature item is labeled with a data type label, which includes discrete text type and continuous numerical type. All filled feature items are assembled into key-value pair feature data in slot order, where the key is the slot name and the value is the filled text or numerical value of that slot. This yields the structured work order feature data.

[0024] In this embodiment of the invention, by accurately extracting multi-source fields such as chief complaint, symptoms, and signs from medical and health service work orders, comprehensive capture of patient medical information is achieved, avoiding the omission of key medical data. Furthermore, deep medical terminology segmentation and entity recognition are performed on unstructured chief complaint and symptom texts to accurately locate the boundaries of core medical information and effectively eliminate the ambiguity of natural language expressions. The identified medical entities are mapped to a standard medical terminology list for standardization, unifying clinical expression norms and completely overcoming data matching errors caused by patients' colloquial descriptions or differences in synonyms. Finally, the standardized multidimensional medical entities are assembled into key-value pair feature data with data type labels according to a preset template, realizing the deep structured fusion of heterogeneous medical information.

[0025] In a preferred embodiment of the present invention, step 2 above may include: Step 2.1 involves extracting labeled symptom entities and vital sign indicators from the structured work order feature data as search query conditions, combining them into a symptom and vital sign query vector. Specifically, this includes: based on the obtained structured work order feature data, selecting feature items corresponding to the vital sign value slots with data type labels of continuous numerical types, and feature items corresponding to the complaint entity slots and symptom entity slots with entity type labels of symptom entity types. The text content of each selected symptom entity slot is used as the query text dimension, and the value of each vital sign value slot, along with its corresponding vital sign indicator name, is used as the query value dimension. For each symptom entity text in the query text dimension, its unique term code in the standard medical terminology glossary is used as its vectorized index; for each vital sign value in the query value dimension, its standard term code for the vital sign indicator name is paired with the specific value to form a value feature pair. All term codes from the query text dimensions and all value feature pairs from the query value dimensions are combined into a multi-dimensional query structure to obtain the symptom and vital sign query vector.

[0026] Step 2.2: Based on the symptom and sign query vector, perform multi-level retrieval of the disease localization detection data in the preset knowledge base. This multi-level retrieval sequentially executes three retrieval levels: exact matching retrieval, fuzzy matching retrieval, and semantic association retrieval, obtaining the disease entries matched at each retrieval level and their corresponding matching scores. Specifically, this includes performing multi-level retrieval of the disease localization detection data in the preset knowledge base based on the obtained symptom and sign query vector. The disease localization detection data in the preset knowledge base uses disease entries as the basic storage unit. Each disease entry contains six fields: disease name, disease code, disease category, typical symptom feature set, typical sign indicator range set, and clinical feature description text. The typical symptom feature set is stored as a set of standardized symptom term codes, and the typical sign indicator range set is stored as a pair of sign indicator name codes and numerical ranges.

[0027] The multi-level search is executed sequentially at the following three search levels: The first retrieval level is exact match retrieval: The set of symptom term codes in the obtained symptom and sign query vector is intersected with the set of typical symptom features for each disease entry in the knowledge base. The ratio of the number of intersection elements to the total number of symptom term codes in the symptom and sign query vector is recorded as the exact match score for that disease entry. Next, the sign numerical features in the symptom and sign query vector are compared one by one with the typical sign indicator range set for each disease entry to determine whether each sign value falls within the numerical range defined by the same indicator name in the corresponding disease entry. The ratio of the number of sign indicators that match the numerical range to the total number of sign numerical feature pairs in the symptom and sign query vector is calculated and used as the exact match score for the sign dimension. The exact match scores for the symptom dimension and the sign dimension are weighted and averaged to obtain the comprehensive exact match score for each disease entry, and all disease entries matched at this level are recorded.

[0028] The second retrieval level is fuzzy matching retrieval: The set of symptom term codes in the obtained symptom and sign query vector is compared with the typical symptom feature set of each disease entry in the knowledge base using Jaccard similarity calculation. The Jaccard similarity coefficient is used as the fuzzy matching score for the disease entry in the symptom dimension. Simultaneously, normalized vectors of each sign value in the symptom and sign query vector and normalized vectors of the median values ​​of typical signs in the disease entry are constructed. The Euclidean distance between the two vectors is calculated, and the normalized value of this Euclidean distance after inverse transformation is used as the fuzzy matching score in the sign dimension. The fuzzy matching scores in the symptom and sign dimensions are weighted and averaged to obtain the comprehensive fuzzy matching score for each disease entry, and all disease entries matched by this layer are recorded.

[0029] The third retrieval level is semantic association retrieval: the symptom entity text and the sign indicator name text in the obtained symptom and sign query vector are concatenated to form the query description text; the clinical feature description text of each disease entry in the knowledge base is used as candidate matching text. The query description text and candidate matching text are respectively input into a pre-trained medical text semantic encoding model to obtain their respective text-level semantic vectors. The cosine distance between the two text-level semantic vectors is calculated as the semantic association score, and all disease entries matched by this layer are recorded.

[0030] Thus, through three levels of retrieval processing, the disease entries matched at each retrieval level and the corresponding matching scores for each entry are obtained.

[0031] Step 2.3: Based on the disease entries matched at each retrieval level and the matching score of each entry, the disease entries returned by different retrieval levels are deduplicated and merged, and then sorted in descending order of matching score to obtain a sorted list of candidate diseases. Specifically, this includes: based on the disease entries matched at each retrieval level and the matching score of each entry, the disease entries returned by different retrieval levels are deduplicated, merged, and sorted.

[0032] The disease entries returned from the three search levels are deduplicated and merged according to their disease codes: if the same disease code is matched in multiple search levels, only one record of that disease entry is retained, and its fusion matching score is calculated based on the matching score of the disease entry in different search levels as follows. : ; in: This represents the overall score obtained by the disease entry at the exact match search level. If the disease entry is not found at the exact match search level, then... The value is zero; This represents the overall score obtained by the disease entry in the fuzzy matching search hierarchy. If the disease entry is not found in the fuzzy matching search hierarchy, then... The value is zero; This represents the score obtained by the disease entry in the semantic association retrieval hierarchy. If the disease entry is not found in the semantic association retrieval hierarchy, then... The value is zero; This represents the weight coefficient for the exact match retrieval level. The value is greater than and greater than Positive numbers; Represents the weight coefficients for the fuzzy matching retrieval level; This represents the weight coefficient of the semantic association retrieval hierarchy, and , and The sum of the three equals 1.

[0033] After completing the deduplication, merging, and matching score calculation for all disease entries, rank all disease entries according to their matching scores. Sort the candidate diseases in descending order from highest to lowest to obtain a sorted list of candidate diseases.

[0034] Step 2.4: Based on the sorted candidate disease list, extract candidate diseases whose matching scores rank within a preset threshold range, and extract the clinical feature description information corresponding to each candidate disease to obtain a matching disease feature set. Specifically, this includes: based on the obtained sorted candidate disease list, starting from the first item in the list, traversing each candidate disease item in descending order of fusion matching score until the fusion matching score of the current candidate disease item is less than the preset screening threshold. All candidate disease items with fusion matching scores greater than or equal to the preset screening threshold are selected as included diseases. For each included disease, extract its clinical feature description text, typical symptom feature set, and typical sign indicator range set from the preset knowledge base to form the feature information block of the disease. Merge the feature information blocks of all included diseases into one set to obtain the matching disease feature set.

[0035] In this embodiment of the invention, by combining symptom entities and sign indicators into a query vector, the effective transformation of medical features into computable data is achieved, laying the foundation for accurate retrieval. Furthermore, an innovative multi-level retrieval mechanism combining precise, fuzzy, and semantic association is adopted, effectively overcoming the limitations of a single retrieval method. This ensures the accuracy of core symptom matching while improving the tolerance and recall rate for non-standard descriptions and potential medical associations. The entries returned at different levels are deduplicated, merged, and quantified in descending order, eliminating data redundancy and intuitively reflecting the degree of fit between the disease and the current signs. Finally, high-probability candidate diseases are locked by thresholding and their clinical feature descriptions are extracted, effectively filtering out low-relevance noise.

[0036] In a preferred embodiment of the present invention, step 3 above may include: Step 3.1: Based on the matched disease feature set, using the patient's unique identifier as the index, query the patient's historical health assessment records in the health trend map to obtain the historical assessment query results. Specifically, this includes: based on the obtained matched disease feature set, using the patient's unique identifier as the primary index, performing a query operation on the historical health assessment records in the health trend map. The health trend map is a multi-layered data index structure with the patient's unique identifier as the primary key. Its underlying layer stores complete records of each patient's previous health assessments. Each health assessment record includes an assessment timestamp, a list of test item names for this assessment, the test indicator values ​​for each test item, the upper and lower limits of the reference range for each test indicator, and the overall health risk level comment for this assessment.

[0037] The specific process of the query operation is as follows: The unique identifier of the patient is precisely matched with the primary key index of the health trend map to locate the data layer corresponding to that patient. In this data layer, all historical health assessment records are extracted and sorted in ascending order by assessment timestamp. If no historical health assessment records exist in the data layer corresponding to that patient, a null value is returned, and in subsequent steps, the current work order data is used as the sole basis for processing. This yields a historical assessment query result containing all historical health assessment records or null value identifiers for that patient.

[0038] Step 3.2: Based on the historical assessment query results, arrange the detection indicator values ​​in each assessment record in ascending order by assessment timestamp, and calculate the direction and magnitude of change of the same detection indicator between two adjacent assessments to obtain the health trend characteristics in the time dimension. Specifically, this includes: based on the obtained historical assessment query results, if the result is not a null value, calculate the health trend characteristics in the time dimension as follows: arrange each assessment record in the historical assessment query results in ascending order by assessment timestamp, take the detection data of two adjacent assessment records as a pair of comparison samples, match the detection indicators with the same detection item name in the two assessments one by one for each pair of adjacent assessment records, and calculate the magnitude and direction of change of the same detection indicator between two adjacent assessments.

[0039] range of change Calculate as follows: ; in: This indicates the value of the current detection indicator in a later assessment; This indicates the value of the current detection indicator in an earlier assessment; The timestamp indicating the later assessment is expressed in days from the base date; The timestamp of the earlier assessment is indicated by the number of days from the base date; This represents the preset unit time length, which is a constant.

[0040] Direction of change Determine as follows: If If the value is greater than zero and the absolute value exceeds the preset trend judgment threshold, the direction of change is marked as an upward trend; if If the value is less than zero and the absolute value exceeds the preset trend judgment threshold, the direction of change is marked as a downward trend; if If the absolute value is less than or equal to the preset trend determination threshold, the direction of change is marked as a stable trend.

[0041] For each detection indicator, the sequence of its change direction markers in all adjacent assessment pairs is summarized, and the frequency of each upward trend, downward trend and stationary trend is counted. The change direction with the highest frequency is selected as the overall change direction of the detection indicator in the current time window. The cumulative change amplitude of the detection indicator is the ratio of the difference between the value of the detection indicator in the most recent assessment and the value in the earliest assessment to the value in the earliest assessment. The overall change direction and cumulative change amplitude of each detection indicator are combined to form the time dimension health trend characteristics of the patient.

[0042] Step 3.3: Based on the health trend characteristics in the time dimension, compare the current values ​​of each test indicator of the patient with the baseline reference values ​​of the same age group stored in the health trend map, calculate the deviation of each indicator, and obtain the deviation analysis results. Specifically, this includes: based on the health trend characteristics in the time dimension obtained in Step 3.2, comparing the current values ​​of each test indicator of the patient with the baseline reference values ​​of the same age group stored in the health trend map.

[0043] The baseline reference values ​​for the same age group include two statistics: the baseline mean of each test indicator for healthy individuals in the same age group. and baseline standard deviation Baseline mean and baseline standard deviation The Health Trend Map statistically calculates each test indicator based on the test data of all healthy people of the same age group stored in it and then stores it in the baseline data table.

[0044] For each detection indicator involved, extract the current value of that indicator in the most recent evaluation. Read the baseline mean of this detection indicator in the same age group from the baseline data table. and baseline standard deviation Calculate the deviation : ; in: This indicates the current value of the tested indicator for the patient in the most recent assessment. This represents the baseline mean of this test indicator among healthy individuals of the same age group; This represents the baseline standard deviation of this test indicator among healthy individuals of the same age group.

[0045] Deviation The symbol indicates deviation from the direction: A value greater than zero indicates that the current value is higher than the baseline mean. A value less than zero indicates that the current value is below the baseline mean. Deviation The absolute value indicates the degree of deviation: the larger the absolute value, the greater the deviation of the current value from the baseline mean. After performing the above deviation calculation for each detection indicator involved, the deviation value corresponding to each detection indicator is obtained and summarized into the deviation analysis results.

[0046] Step 3.4: Based on the deviation analysis results, and combined with the typical indicator anomaly patterns of each candidate disease in the matched disease feature set, a three-dimensional cross-summary is performed according to the indicator category dimension, time change dimension, and risk level dimension to obtain multi-dimensional health status distribution characteristics. Specifically, this includes: based on the obtained deviation analysis results, and combined with the typical indicator anomaly patterns of each candidate disease in the obtained matched disease feature set, a three-dimensional cross-summary is performed according to the indicator category dimension, time change dimension, and risk level dimension.

[0047] The first dimension is the indicator category dimension: each detection indicator is categorized into its corresponding indicator category group according to the physiological system it belongs to. The physiological system categories include the cardiovascular system indicator group, the endocrine and metabolic system indicator group, the liver and kidney function system indicator group, the blood system indicator group, and the respiratory system indicator group.

[0048] The second dimension is the time change dimension: the time dimension health trend characteristics of each detection indicator are referenced, and the overall change direction and cumulative change magnitude of each detection indicator are marked.

[0049] The third dimension is the risk level dimension: For each detection indicator, its calculated absolute value of deviation is compared with a preset multi-level deviation threshold to classify its risk level. If the absolute value of deviation is less than the first deviation threshold, the risk level is marked as low risk; if the absolute value of deviation is greater than or equal to the first deviation threshold and less than the second deviation threshold, the risk level is marked as medium risk; if the absolute value of deviation is greater than or equal to the second deviation threshold, the risk level is marked as high risk.

[0050] Based on the obtained set of matched disease features, typical indicator anomaly patterns are extracted for each candidate disease. A typical indicator anomaly pattern defines the abnormal combination of test indicators that typically occurs during the diagnosis of the disease, including the set of test indicator names involved, the expected deviation direction of each indicator, and the expected deviation level range. The typical indicator anomaly pattern for each candidate disease is cross-compared with the actual deviation analysis results of each test indicator for the patient: if the actual deviation level of a patient's test indicator falls within the expected deviation level range of the corresponding indicator in the typical indicator anomaly pattern of the candidate disease, and the actual deviation direction is consistent with the expected deviation direction, then that indicator is marked as a matching abnormal indicator for the current candidate disease.

[0051] For each candidate disease, the ratio of the number of matching abnormal indicators to the total number of indicators involved in the typical abnormal indicator pattern of that disease is calculated, and this ratio is recorded as the comprehensive abnormal matching degree of the candidate disease under the three-dimensional cross-summary. Overall anomaly matching degree Calculate as follows: ; in: This indicates the number of detection indicators in the actual deviation of the patient that match the abnormal pattern of typical indicators for the candidate disease; This indicates the total number of detection indicators involved in the abnormal pattern of typical indicators of the candidate disease; express The number of indicators marked as high-risk among the matched anomaly indicators; The penalty amplification factor representing the number of high-risk indicators is a preset normal number.

[0052] according to The values ​​of are used to summarize the matching results of each candidate disease under the three-dimensional cross-summary, together with the detailed annotation information of each dimension, into a multi-dimensional health status distribution feature. The multi-dimensional health status distribution feature includes: each detection indicator and its deviation value under the indicator category grouping with the patient's unique identifier as the main body, the time change direction and time change magnitude of each detection indicator, the risk level of each detection indicator, and the comprehensive abnormal matching degree with each candidate disease.

[0053] In this embodiment of the invention, by using the patient's unique identifier as an index to accurately query historical assessment records in a health trend map, the rapid extraction of personalized longitudinal health data for patients is achieved. Furthermore, by deeply analyzing the dynamic evolution of detection indicators between adjacent assessments according to time series, the temporal development trajectory of health status is effectively captured, overcoming the limitations of static single assessments. The current individual indicators are quantitatively compared with the baseline reference values ​​of the same age group to accurately calculate the degree of deviation, successfully eliminating assessment bias caused by group physiological differences and ensuring the objectivity of health status positioning. Finally, the individual deviation degree is integrated with typical abnormal patterns of candidate diseases, and a comprehensive, multi-dimensional health status distribution feature is constructed from three dimensions: indicator category, time change, and risk level, comprehensively depicting the patient's overall health risk profile.

[0054] In a preferred embodiment of the present invention, step 4 above may include: Step 4.1: Map each abnormal indicator in the multi-dimensional health status distribution characteristics to the preset preventive healthcare measures knowledge base, find the intervention measures corresponding to each abnormal indicator, and obtain the candidate preventive healthcare measures set corresponding to each abnormal indicator. Specifically, this includes: extracting all detection indicators marked as medium or high risk based on the obtained multi-dimensional health status distribution characteristics to form the abnormal indicator set of the current patient. For each abnormal indicator in the abnormal indicator set, a precise search is performed in a pre-defined preventive healthcare measures knowledge base, coded according to its standardized name. This knowledge base uses the standard name of the detection indicator as an index key, and each index key is associated with a list of intervention measures. Each intervention measure entry includes: the name identifier of the intervention measure, the category of the intervention measure, the specific operational text of the intervention measure, the direction of the abnormal deviation to which the intervention measure applies, and a pre-defined correlation strength value between the intervention measure and the corresponding detection indicator. For each abnormal indicator in the abnormal indicator set, after retrieving the corresponding list of intervention measures in the preventive healthcare measures knowledge base, intervention measures are selected based on the overall change direction of the abnormal indicator over a defined time dimension. These intervention measures either have an applicable abnormal deviation direction consistent with the overall change direction or have a bidirectional applicable abnormal deviation direction. This yields a set of candidate preventive healthcare measures for each abnormal indicator.

[0055] Step 4.2: Based on the set of candidate preventive healthcare measures corresponding to each abnormal indicator, deduce the causal chain between each preventive healthcare measure and the abnormal indicator it targets in descending order of risk level, and obtain the deduction results with causal labels. Specifically, this includes: based on the set of candidate preventive healthcare measures corresponding to each abnormal indicator, deduce the causal chain between each preventive healthcare measure and the abnormal indicator it targets in descending order of risk level of each labeled abnormal indicator.

[0056] For each pair of abnormal indicators and its corresponding candidate preventive healthcare measure, the following causal derivation process is performed: extract the pre-defined association strength value of the intervention measure from the preventive healthcare measure knowledge base. This value indicates the strength of the intervention's effect on the abnormal indicator as recognized in academic or clinical guidelines. The cumulative change in the abnormal indicator is extracted from the obtained time-dimensional health trend characteristics, and the product of the current absolute value of the abnormal indicator's deviation and the cumulative change is calculated as a measure of the degree of deterioration of the abnormal indicator. : ; in: This represents the absolute value of the deviation calculated for this abnormal indicator; This represents the absolute value of the cumulative change in the calculated abnormal indicator.

[0057] Calculate the causal strength of the current intervention on this abnormal indicator. : ; in: This indicates the pre-defined association strength value of the intervention in the preventive healthcare measures knowledge base; This is a measure of the degree of deterioration of the abnormal indicator. This represents a priority coefficient based on risk level; when the risk level of this abnormal indicator is high... The value is a constant value for high-risk priority, and for medium-risk... The value is a constant value for medium-risk priority, and a constant value for high-risk priority is greater than a constant value for medium-risk priority.

[0058] The name of the intervention, the code of the abnormal indicator it targets, and the strength of causation are identified. and the correlation strength values ​​referenced in the derivation process and measures of deterioration Record them together to form the causal labeling entries corresponding to the intervention measures. After completing the above causal derivation for each intervention measure, summarize the derivation results with causal labels to obtain the segmented results.

[0059] Step 4.3: Based on the step-by-step derivation results with causal annotations, calculate the confidence score for each candidate disease in the matched disease feature set by combining the symptom and sign matching degree and the deviation degree of historical health trends. Select candidate diagnostic directions from high to low confidence scores, and generate corresponding preliminary diagnosis and treatment suggestions for each candidate diagnostic direction to obtain a list of preliminary diagnosis and treatment suggestions with priority. Specifically, based on the obtained step-by-step derivation results with causal annotations, calculate the diagnostic confidence score and generate preliminary diagnosis and treatment suggestions for each candidate disease in the obtained matched disease feature set in sequence.

[0060] For each candidate disease, its diagnostic confidence score Calculate as follows: ; in: This represents the fusion matching score calculated for the candidate disease, which reflects the degree of matching between the symptoms and signs of the current work order and the typical characteristics of the disease. This represents the overall abnormality matching degree calculated for the candidate disease. This value reflects the degree of agreement between the patient's historical test data and the abnormal patterns of typical indicators of the disease. This represents the mean of the causal strengths calculated from all abnormal indicators related to the candidate disease; that is, for each abnormal indicator related to the candidate disease, the causal strength of all corresponding interventions is taken. The maximum value in the range is then summed, and the sum of the maximum causal strengths corresponding to each abnormal indicator is divided by the number of abnormal indicators involved in the candidate disease. This reflects the potential efficacy of existing preventive healthcare interventions in improving this candidate disease; The weighting coefficients represent the fusion matching score; The weighting coefficients represent the overall anomaly matching degree; Weighting coefficients representing the mean strength of causal effects. , and The values ​​are all positive numbers between zero and one, and the sum of the three is equal to one.

[0061] All candidate diseases were ranked by diagnostic confidence score. The candidate diseases are sorted in descending order of diagnostic confidence score, and those with a score greater than a preset diagnostic confidence threshold are selected as alternative diagnostic directions. For each alternative diagnostic direction, a standardized preliminary treatment suggestion template corresponding to the disease is extracted from the preset knowledge base, including a list of recommended examination items, a recommended department type, and a recommended treatment timeliness level. The extracted standardized preliminary treatment suggestion template is bound to the diagnostic confidence score of the alternative diagnostic direction to generate a corresponding preliminary treatment suggestion item. The preliminary treatment suggestion items corresponding to each alternative diagnostic direction are sorted in descending order of diagnostic confidence score to obtain a priority-sorted list of preliminary treatment suggestions.

[0062] Step 4.4: Based on the set of candidate preventive healthcare measures and the list of preliminary treatment recommendations with priority, merge and organize them according to the hierarchical structure of preventive measures taking precedence over treatment recommendations, and remove redundancy from the merged items to obtain a preliminary collaborative treatment plan. Specifically, this includes: merging and organizing the obtained set of candidate preventive healthcare measures and the obtained list of preliminary treatment recommendations with priority, according to the hierarchical structure of preventive measures taking precedence over treatment recommendations.

[0063] All preventive healthcare measures in the candidate preventive healthcare measure set are considered as the first level of preventive healthcare in the program, and are ranked according to the strength of causality. First, the items are arranged in descending order of diagnostic confidence score. Second, all preliminary diagnostic recommendations in the list are considered as the second level of the treatment plan, maintaining the descending order of diagnostic confidence score. After completing the hierarchical division and intra-level sorting, redundancy removal is performed on the items within each of the two levels: within the preventive healthcare layer, if multiple preventive healthcare measures target the same abnormal indicator and belong to the same category, only the causal strength of the measures is retained. The highest-ranking entry is deleted, and all other entries are removed. Within the treatment suggestion layer, if multiple preliminary treatment suggestion entries point to the same type of department, these entries are merged into a single comprehensive treatment suggestion entry. The comprehensive treatment suggestion entry uses the candidate diagnosis direction with the highest diagnostic confidence score as the primary diagnosis direction, and the others as differential diagnosis directions. The diagnostic confidence score for each diagnosis direction is noted in the entry. The preventive healthcare layer and treatment suggestion layer, after redundancy removal, are then assembled into a single overall plan in hierarchical order. A summary of the patient's basic information and the work order number are added to the beginning of the plan to obtain a preliminary collaborative treatment plan.

[0064] In this embodiment of the invention, by accurately associating each abnormal indicator with a knowledge base of preventive healthcare measures, targeted candidate intervention measures are quickly identified, achieving an initial connection between health risks and intervention methods. Furthermore, the causal chain between each measure and the abnormal indicator is deduced segment by segment according to risk level from high to low, clearly revealing the internal logic of health intervention and effectively avoiding blind intervention. Combining the causal deduction results with candidate disease characteristics, the degree of fit between each diagnostic direction is quantitatively assessed by comprehensively evaluating symptom matching and health trend deviation, accurately screening candidate diagnostic directions and generating priority treatment recommendations, ensuring the scientific validity and timeliness of the treatment recommendations. Finally, the candidate preventive measures and the ranked treatment recommendations are deeply integrated and redundancy-removed according to a prevention-first hierarchical structure, eliminating duplicate intervention items and outputting a clearly structured preliminary collaborative treatment plan.

[0065] In a preferred embodiment of the present invention, step 5 above may include: Step 5.1 involves extracting the disease codes and treatment direction identifiers corresponding to each treatment suggestion item in the preliminary collaborative treatment plan. Using the disease codes and treatment direction identifiers as joint search conditions, a matching standardized treatment path template is queried in the biobank to obtain the standardized treatment path matching results. Specifically, this includes extracting the disease codes and treatment direction identifiers bound to each treatment suggestion item within the second-level treatment suggestion layer of the obtained preliminary collaborative treatment plan. The disease code is a unique identifier assigned to each disease item in the preset knowledge base. The treatment direction identifier is a classification label used to distinguish the treatment direction category to which the treatment suggestion belongs, including five categories: drug treatment, surgical treatment, physical rehabilitation treatment, lifestyle intervention, and regular follow-up observation.

[0066] The extracted disease codes and treatment direction identifiers form a combined search condition. The disease codes are used as the first search key, and the treatment direction identifiers as the second search key, to perform a two-level search within the biobank. The biobank uses disease codes as the first-level index key and treatment direction identifiers as the second-level index key. Each second-level index key stores at least one standardized treatment pathway template. Each standardized treatment pathway template includes the following predefined fields: pathway name identifier, applicable disease code, applicable treatment direction identifier, treatment stage division sequence, list of corresponding examination items for each stage, expected treatment cycle for each stage, resource consumption level for each stage, list of collaborating department codes for each stage, and list of contraindications for the pathway.

[0067] The retrieval process is as follows: Using the disease code as the first search key, the path storage area corresponding to the disease is located in the primary index of the biobank; then, using the treatment direction identifier as the second search key, standardized treatment path templates matching the treatment direction identifier are filtered in the secondary index of this path storage area. If the combined search conditions for a certain treatment suggestion item do not match any standardized treatment path template in the biobank, the item is marked as having no path match. This item will not participate in the standardized path calculation in subsequent steps and will instead be processed using a general path template. This yields a standardized treatment path matching result containing a list of standardized treatment path templates matched by each treatment suggestion item.

[0068] Step 5.2: Based on the standardized treatment pathway matching results, extract the predefined treatment stage divisions, required examination items for each stage, and expected treatment cycles for each stage from each standardized treatment pathway template, and compare them item by item at the stage level with the corresponding items in the preliminary collaborative treatment plan to obtain the pathway comparison results. Specifically, based on the standardized treatment pathway matching results obtained in Step 5.1, for each matched standardized treatment pathway template, extract its predefined treatment stage division sequence, list of examination items for each stage, and expected treatment cycle for each stage, and compare them item by item at the stage level with the content of the corresponding treatment suggestion items in the preliminary collaborative treatment plan.

[0069] The specific comparison process is divided into three dimensions: The first dimension is the stage division comparison: The predefined treatment stage division sequence in the standardized treatment pathway template is compared with the treatment stages suggested by the treatment recommendations in the preliminary collaborative treatment plan. The standardized treatment pathway template typically includes four standard stages: initial assessment, diagnostic examination, treatment implementation, and rehabilitation follow-up. The correspondence between these four standard stages and the preventive healthcare measures and treatment recommendations in the preliminary collaborative treatment plan is compared along the timeline, marking the overlapping stages in the template and those not yet covered by the preliminary collaborative treatment plan.

[0070] The second dimension involves comparing the examination items: This involves performing a set operation between the list of examination items corresponding to each stage in the standardized treatment pathway template and the examination items already included in the preventive healthcare measures and treatment recommendations in the preliminary collaborative treatment plan. The ratio of template examination items to those already covered by the preliminary collaborative treatment plan is then calculated. : ; in, This indicates the number of examination items in the standardized treatment pathway template examination item list that have been covered by the corresponding items in the preliminary collaborative diagnosis and treatment plan; This indicates the total number of examination items included in the list of examination items for this standardized treatment pathway template.

[0071] The third dimension is the treatment cycle comparison: the expected treatment cycle of each stage in the standardized treatment pathway template is compared with the suggested processing timeliness level of the corresponding item in the preliminary collaborative treatment plan, and the cycle matching deviation is calculated. If there is a difference between the number of days corresponding to the suggested processing timeliness level of the treatment suggestion item in the preliminary collaborative treatment plan and the expected number of days of the corresponding stage in the standardized treatment pathway template, the difference is recorded as the cycle deviation.

[0072] The comparison results of the three dimensions are summarized and labeled on each standardized treatment pathway template to obtain the pathway comparison results.

[0073] Step 5.3: Based on the path comparison results, for each standardized treatment path, and in conjunction with the specific risk factors marked in the multi-dimensional health status distribution characteristics of the patient, check one by one whether there are any contraindications that conflict with the patient's specific risk factors, and obtain the path safety verification results. Specifically, this includes: based on the obtained path comparison results, for each standardized treatment path, and in conjunction with the risk level information marked in the multi-dimensional health status distribution characteristics of the patient, check one by one whether there are any contraindications that conflict with the patient's specific risk factors.

[0074] From the obtained multi-dimensional health status distribution characteristics, the physiological system categories corresponding to all test indicators marked as high-risk for this patient are extracted. These physiological system categories have been determined through indicator category division, including five categories: cardiovascular system, endocrine and metabolic system, liver and kidney function system, hematologic system, and respiratory system. The physiological system categories to which all high-risk test indicators belong are summarized to form a set of specific risk factors for this patient.

[0075] For each standardized treatment pathway, a predefined list of contraindications is extracted from the pathway template. Each contraindication item in the list includes: the physiological system category involved, the drug component category or operation type applicable to the contraindication, the risk level threshold of the contraindication, and a description of the conditions triggering the contraindication. The list of contraindications for each standardized treatment pathway is cross-referenced with the patient's set of specific risk factors. The comparison rules are as follows: If a physiological system category in the patient's set of specific risk factors is the same as the physiological system category involved in an item in the contraindications list, then it is further determined whether the absolute value of the deviation of the high-risk detection indicator under that physiological system category is greater than or equal to the risk level threshold specified in the contraindication item; if so, the standardized treatment pathway is marked as having a contraindication conflict.

[0076] For standardized treatment pathways with contraindications, calculate the degree of contraindication conflict. : ; in, This indicates that the patient is number one. The weight coefficient of the physiological system category corresponding to each high-risk detection indicator that triggers a taboo conflict is determined by the relative importance of each physiological system category in clinical safety assessment, and is a preset constant. This indicates that the patient is number one. The absolute value of the deviation of a high-risk detection indicator that has triggered a taboo conflict; Indicates the first A high-risk detection indicator that triggers a taboo conflict is the risk level threshold specified in the taboo item.

[0077] If the degree of contraindication conflict of a certain standardized treatment pathway If the threshold for path safety is exceeded, the standardized treatment path is marked as failing safety verification and excluded from subsequent overall aggregation and accounting; if If the threshold for path safety is less than or equal to the preset threshold, the standardized treatment path is marked as having passed safety verification, and a contraindication / conflict level is added to the path. This serves as a security marker for reference in subsequent verification steps. This yields the path security verification result.

[0078] Step 5.4: Based on the path safety verification results, perform a comprehensive summary calculation on each standardized treatment path that has passed the safety verification according to three accounting dimensions: resource consumption, time period, and departmental collaboration requirements, to obtain the comprehensive summary calculation results. Specifically, based on the path safety verification results obtained in Step 5.3, perform a comprehensive summary calculation on each standardized treatment path marked as having passed the safety verification according to the three accounting dimensions: resource consumption, time period, and departmental collaboration requirements.

[0079] The first accounting dimension is resource consumption accounting: for each standardized treatment path, the resource consumption level corresponding to each stage of the path is extracted, and the resource consumption level of each stage is converted into a quantified resource consumption value. Quantified resource consumption value Calculate as follows: ; in, This indicates the first step of the standardized treatment pathway. The resource consumption levels for each stage of diagnosis and treatment are divided into five levels from low to high, with corresponding values ​​of one to five. Indicates the first The stage weight coefficient of each treatment stage is determined by the importance of that stage in the overall treatment pathway. It is a preset constant, and the sum of the stage weight coefficients of all stages is equal to one.

[0080] The second accounting dimension is time cycle accounting: for each standardized treatment pathway, the expected treatment cycle days for each stage of the pathway are extracted and added to the calculated cycle deviation to calculate the expected total treatment cycle for the pathway. : ; in: This indicates the first step of the standardized treatment pathway. The expected treatment cycle for each stage of diagnosis and treatment, in days; This represents the sum of the periodic deviations at each stage of the path.

[0081] The third accounting dimension is the accounting of departmental collaboration needs: for each standardized treatment pathway, extract the list of collaborative department codes corresponding to each stage of the pathway, and count the number of dissimilar departments involved in the pathway. The complexity of departmental collaboration is calculated based on the frequency of participation of each collaborating department at each stage of the pathway. : ; in: This indicates the number of different departments involved in the standardized treatment pathway; This represents the average participation frequency of each collaborating department at each stage of the pathway, calculated by dividing the total number of times each collaborating department is cited at each stage by the number of dissimilar departments. Calculated.

[0082] After completing the independent accounting for the three dimensions, a total comprehensive accounting score is calculated for each standardized treatment path that passes the safety verification. : ; in: This represents the quantified value of resource consumption for this path. The larger the value, the more resources are consumed. Therefore, its reciprocal is used in the calculation, so that the lower the resource consumption, the greater the contribution of the score. This represents the estimated total processing time for the path; similarly, its reciprocal can be taken. This represents the complexity of interdepartmental collaboration along the path; similarly, its reciprocal can be taken. This indicates the degree of taboo conflict of the path, and serves as a penalty item in the calculation. The larger the value, the lower the score. This represents the accounting weighting coefficient for the resource consumption dimension; The accounting weight coefficient represents the time period dimension; This represents the weighting coefficient for the departmental collaboration dimension. This represents the weighting coefficient of the security penalty item. , , and The values ​​are all positive numbers between zero and one, and , and The sum of the three equals one. Independent settings.

[0083] All standardized treatment pathways that have passed safety verification will be aggregated and scored across the entire domain. The data are sorted in descending order, and detailed data and security labels for each path are appended to its three accounting dimensions. The results are then aggregated to obtain the overall accounting results.

[0084] In this embodiment of the invention, by accurately matching standardized treatment pathway templates in a biobank, authoritative and standardized clinical reference benchmarks are provided for preliminary collaborative treatment plans, effectively improving the scientific nature of treatment recommendations. Furthermore, the treatment stages, examination items, and expected cycles of the standard pathway are compared item by item with the preliminary plan, ensuring a high degree of consistency in the execution details of the intervention measures and avoiding overtreatment or omission of key examinations. Innovatively, a rigorous safety verification is conducted by incorporating individualized risk factors of each patient, accurately identifying and excluding potential contraindications, maximizing the clinical safety of personalized treatment plans, and reducing medical risks. Finally, compliant pathways are comprehensively calculated from three dimensions: resource consumption, time cycle, and departmental collaboration needs, achieving the final allocation and efficient scheduling of medical resources.

[0085] In a preferred embodiment of the present invention, step 6 above may include: Step 6.1: Based on the overall summary accounting results, extract the urgency score, resource requirement level, and expected processing cycle corresponding to each standardized treatment pathway in the accounting results to obtain the flow decision input parameters. Specifically, this includes: based on the obtained overall summary accounting results, processing the parameters according to... For each standardized treatment pathway arranged in descending order, three data points are extracted: urgency score, resource requirement level, and expected treatment cycle.

[0086] The urgency score is determined as follows: based on the expected treatment cycle length corresponding to the initial assessment stage marked by the standardized treatment pathway, and combined with the patient's risk level distribution on relevant detection indicators of the pathway, the urgency score is calculated. : ; in: This indicates the expected treatment cycle in days during the initial assessment phase of this standardized treatment pathway. This indicates the proportion of indicators marked as high-risk among the test results related to this pathway for patients. This represents the urgency weighting coefficient of the cycle; This represents the risk urgency weighting coefficient. and The sum of them equals one.

[0087] The resource demand level directly references the quantified resource consumption value of that path. ,according to The value range is mapped to one of three levels: low resource demand, medium resource demand, and high resource demand. The estimated processing cycle directly refers to the estimated total processing cycle of this path. .

[0088] The three data points corresponding to each standardized treatment pathway—urgency score, resource requirement level, and expected treatment cycle—are packaged into a single flow decision input record. The flow decision input records of all pathways in descending order from the overall aggregated accounting results are then summarized to obtain the flow decision input parameters.

[0089] Step 6.2: Based on the workflow decision input parameters, compare the urgency score with the preset urgency judgment threshold to determine the priority label of the work order and obtain the work order priority judgment result. Specifically, this includes: based on the obtained workflow decision input parameters, assigning an urgency score to each workflow decision input record. Compare with the preset urgency threshold.

[0090] The preset urgency determination threshold includes a first urgency threshold. Second emergency threshold ,satisfy Greater than The comparison rules are as follows: If Greater than or equal to If the priority label of the work order is set to first priority, then the priority label of the work order is set to first priority. Less than and greater than or equal to If so, the priority label of the work order is determined to be the second priority; if Less than If the priority label of the work order is determined to be the third priority, then the priority label of the work order will be determined after determining the priority label corresponding to each workflow decision input record. The overall summary accounting results will then be... The priority label corresponding to the standardized treatment path with the highest value is taken as the primary priority label for this work order. If multiple standardized treatment paths in the overall summary calculation result have urgency scores all reaching the first priority level, and these paths point to different treatment directions, then the primary priority label is taken. The priority label of the path with the highest value is used as the main priority label, and the remaining paths are noted as alternative directions in the flow instruction as associated paths, thus obtaining the work order priority determination result.

[0091] Step 6.3: Based on the work order priority determination result and resource requirement level, match the corresponding flow path template in the preset work order flow rule table to obtain the initial flow path. Specifically, this includes: matching the corresponding flow path template in the preset work order flow rule table based on the obtained work order priority determination result and the resource requirement level of each flow decision input record. The preset work order flow rule table uses priority label as the first-level index key and resource requirement level as the second-level index key, with each composite index key associated with one flow path template. Each flow path template defines an ordered sequence of department nodes. Each department node in the sequence includes a department code, the department's position in the flow sequence, the upper limit of the expected dwell time of the department node, and the mandatory or optional attribute marker of the department node.

[0092] The matching process is as follows: Using the determined primary priority tag as the first matching key, ... The resource requirement level corresponding to the highest-valued path is used as the second matching key. Two-level exact matching is performed in the work order flow rule table. If the exact match is successful, the corresponding flow path template is directly output as the initial flow path. If the exact match fails, the main priority label remains unchanged, and the resource requirement level is increased or decreased by one level before rematching until all resource requirement levels are matched or exhausted. If no flow path template can be matched under any priority label, the default flow path template corresponding to the main priority label is retrieved from the work order flow rule table. This default flow path template does not distinguish between resource requirement levels and is applicable to general work order types under this priority, thus obtaining the initial flow path.

[0093] Step 6.4: Based on the initial workflow path, and combined with the real-time service load status of each department and the availability of doctors on duty, perform availability checks and adaptive adjustments on each department node in the initial workflow path one by one to obtain the final medical and health service work order workflow instruction. Specifically, this includes: based on the obtained initial workflow path, and combined with the real-time service load status of each department and the availability of doctors on duty in the current system, performing availability checks and adaptive adjustments on each department node in the initial workflow path one by one.

[0094] The real-time service load status is collected by the system from the queues of each department at preset time intervals, including two indicators: the number of currently queued work orders in each department and the current average work order processing time. The available doctor on-duty information is read in real time by the system from the doctor scheduling management module, including the number of doctors currently on duty in each department and the employee number of each doctor on duty.

[0095] For each department node in the initial workflow sequence, perform the following availability check in sequence according to its order in the sequence: obtain the current number of queued work orders corresponding to the department code of that department node. and the number of doctors currently on duty Calculate the real-time load index of the department. : ; in, This indicates the number of work orders currently in the queue corresponding to this department code; This indicates the number of doctors currently on duty corresponding to the department code.

[0096] Real-time load index Compared with the preset departmental load saturation threshold Perform a comparison. If... Less than or equal to If so, the department node passes the availability check and maintains its original position in the workflow path. Greater than If the node in that department fails the availability check, an adaptive adjustment will be triggered.

[0097] The adaptive adjustment method is as follows: Using the department code of the department node that failed the availability check as a condition, search the doctor resource information pool for an alternative department with the same specialty as that department code; if an alternative department exists and the real-time load index of the alternative department is... Less than or equal to If a suitable alternative department is found, the original department code of that node in the initial workflow path is replaced with the department code of the alternative department. If no suitable alternative department exists, the original department node is retained, but its mandatory or optional attribute is changed from mandatory to optional, and a buffer waiting node is inserted after that node. The buffer waiting node is not bound to a specific department; only a buffer waiting time limit is set. After completing availability verification and adaptive adjustments for all department nodes, an adjusted department node sequence is obtained. This adjusted department node sequence is then assembled with the determined work order primary priority label to obtain the final medical and health service work order workflow instruction.

[0098] In this embodiment of the invention, by extracting key indicators such as the urgency, resource requirements, and expected cycle of standardized treatment pathways, the decision-making elements for work order flow are fully quantified, laying a data foundation for precise scheduling. Furthermore, the urgency level is compared with preset thresholds to establish work order priority labels, effectively ensuring priority response to high-risk and urgent medical needs and avoiding delays in critical cases. Combining priority and resource requirement matching flow rule tables, initial flow paths are automatically generated, significantly improving the standardization and processing efficiency of path planning. Finally, innovatively, the initial paths are dynamically verified and adaptively adjusted based on the real-time service load of each department and the on-duty status of doctors, overcoming resource mismatch and node congestion problems caused by static planning, ensuring the high feasibility and accurate implementation of final flow instructions in complex medical environments.

[0099] In a preferred embodiment of the present invention, step 7 above may include: Step 7.1: Based on the final medical and health service work order flow instruction, extract the target department code and the required physician specialty identifier specified in the instruction to obtain the allocation target parameters. Specifically, this includes: parsing the department node sequence in the final medical and health service work order flow instruction and extracting the information of the first department node in the sequence. If the first department node is a node bound to a specific department code, extract the target department code from that node. Based on the target department code, find the treatment stage corresponding to the department node from the treatment stage division sequence of the standardized treatment path template corresponding to the work order, and extract the physician specialty identifier required for that treatment stage. If the first department node is a buffer waiting node, skip that node and extract the next node bound to a specific department code in the sequence as the target department node. This yields the allocation target parameters containing the target department code and the required physician specialty identifier.

[0100] Step 7.2: Based on the assigned target parameters, perform a first-level screening in the preset doctor resource information pool according to the target department code, and perform a second-level screening in the resulting set of doctors in the department according to the required doctor specialty direction identifier, to obtain a candidate doctor set. Specifically, this includes performing two-level screening in the preset doctor resource information pool according to the obtained assigned target parameters. The preset doctor resource information pool uses the department code as the first-level classification field. Under each department code, it stores the personal files of all registered doctors in that department. Each doctor's personal file includes the doctor's employee number, doctor's name identifier, department code, department group identifier, list of specialty direction identifiers, list of areas of expertise tags, current on-duty status identifier, number of currently processed work orders, and average work order processing time.

[0101] The first level of screening is department screening: using the obtained target department code as the screening condition, all doctors whose department codes match the target department code are selected from the doctor resource information pool, thus obtaining the department doctor set.

[0102] The second level of screening is professional direction screening: In the set of doctors in the department, the desired doctor's professional direction identifier is used as the screening condition to select doctors whose professional direction identifier is included in the list of professional direction identifiers, thus obtaining a set of candidate doctors.

[0103] If the candidate doctor set is empty, the second-level screening criteria are relaxed, and instead, doctors whose professional direction identifiers in the list contain similar professional direction identifiers that have a preset relationship with the professional direction identifiers of the required doctor are screened, and the candidate doctor set is reconstructed.

[0104] Step 7.3: Based on the candidate doctor set, obtain the current on-duty status identifier, the number of currently processed work orders, and the field of expertise tag for each candidate doctor. Calculate the matching score between each candidate doctor and the current work order requirements to obtain the doctor matching ranking result. Specifically, based on the candidate doctor set obtained in Step 7.2, for each candidate doctor in the set, obtain three parts of data: their current on-duty status identifier, the number of currently processed work orders, and the field of expertise tag.

[0105] First, the on-duty status of each doctor in the candidate doctor set is checked: if a doctor's current on-duty status is marked as absent or on leave, that doctor is removed from the candidate doctor set. Subsequent matching degree calculations are only performed on candidate doctors whose current on-duty status is marked as present.

[0106] For each candidate doctor who passes the on-duty status verification, a matching score between the doctor and the current work order requirements is calculated based on the following three dimensions. : ; in: The component representing the doctor's current workload matching degree is calculated as follows: The number of work orders currently in progress for that doctor is obtained. With respect to the preset reasonable workload limit for doctors Comparison, Values and The normalized result of the difference The smaller the value, the higher the value of the component, indicating that the doctor currently has more capacity to accept new work orders; The matching component representing the doctor's area of ​​expertise is calculated as follows: The disease category tags involved in the current work order within the determined alternative diagnostic directions are compared with each tag in the doctor's area of ​​expertise tag list using Jaccard similarity calculation. The Jaccard similarity coefficient is used as the... The higher the similarity value, the better the doctor's professional background matches the current work order's diagnostic and treatment needs; The expected response time matching component for doctors is calculated as follows: based on the doctor's average work order processing time. For parameters, Values The reciprocal after normalization The shorter the length, the higher the value of the component. This represents the weighting coefficient for the physician workload dimension; This represents the weighting coefficient for the area of ​​expertise. This represents the weighting coefficient for the response time dimension. , and The values ​​are all positive numbers between zero and one, and the sum of the three values ​​equals one. A matching score is calculated for each currently employed candidate doctor in the candidate doctor set. Then, press The doctors are sorted in descending order of matching degree to obtain the ranking results.

[0107] Step 7.4: Based on the doctor matching score ranking results, select target doctors from high to low matching scores, and assemble work order dispatch parameters based on the department group identifier, doctor's employee number, and expected first response time of the target doctor. Specifically, this includes: assembling the work order dispatch parameters based on the obtained doctor matching score ranking results. The candidate doctor ranked first is selected as the target doctor for allocation, in descending order of priority.

[0108] Extract the department group identifier and doctor's employee number from the doctor's personal file. Calculate the estimated first response time based on the doctor's ranking in the current work order queue and the average work order processing time. : ; in: This indicates the number of pending work orders preceding the current work order's position in the doctor's work order queue, i.e., the number of work orders currently being processed for that doctor. ; This indicates the doctor's average work order processing time, in minutes. The target is assigned the doctor's employee ID, department group identifier, and estimated first response time. The target department code and the identifier of the first department node in the obtained flow node sequence are encapsulated in a structured manner to assemble the work order dispatch parameters.

[0109] In this embodiment of the invention, by accurately extracting the target department and professional direction identifiers, the clear quantification of work order allocation requirements is achieved, laying the foundation for subsequent precise scheduling. Furthermore, a dual progressive screening mechanism based on department and professional direction is adopted to quickly narrow down the candidate range, significantly improving resource filtering efficiency and professional matching rate. Breaking through the traditional static allocation model, the invention comprehensively considers the doctor's real-time on-duty status, current workload of work orders, and expertise tags, dynamically calculating multi-dimensional matching degree. This effectively avoids processing delays caused by doctor overload or mismatch in professional subdivision, ensuring the global optimal allocation of medical resources. Finally, the target doctor is selected based on the matching degree, and complete dispatch parameters, including the expected response time, are assembled, achieving seamless connection between work orders and optimal medical resources.

[0110] In a preferred embodiment of the present invention, step 8 above may include: Step 8.1: Based on the target assigned doctor's ID and department group identifier in the work order dispatch parameters, overwrite the current handler field and current handling department field of the medical and health service work order with the target assigned doctor's ID and department group identifier, respectively, to obtain the handler update result. Specifically, this includes: locating the stored record of the medical and health service work order in the work order data table based on the target assigned doctor's ID and department group identifier contained in the obtained work order dispatch parameters; overwriting the value of the current handler field in the stored record with the target assigned doctor's ID; and overwriting the value of the current handling department field with the department group identifier. The overwrite operation adopts a transactional write method, that is, before overwriting, a snapshot of the values ​​before the overwrite operation is recorded. This snapshot includes the original values ​​of the current handler field and the current handling department field before the overwrite. The snapshot of the values ​​before the overwrite is saved as part of the workflow log record to ensure the integrity and traceability of the workflow. After completing the overwrite operation, the handler update result is obtained.

[0111] Step 8.2: Based on the update result from the processor, write the target node identifier in the work order dispatch parameters into the work order's flow node log table. Simultaneously record the trigger timestamp, source node identifier, and target node identifier for this flow in the log table to obtain the flow node update result. Specifically, based on the obtained processor update result, use the target department code in the obtained work order dispatch parameters as the target node identifier for this flow, along with the source node identifier of the work order before this flow, and the obtained current system time as the trigger timestamp for this flow. Write all of these into the flow node log table associated with the work order. The flow node log table uses the work order number as the primary index key. Each flow node log record includes a flow log sequence number, source node identifier, target node identifier, flow trigger timestamp, flow completion timestamp, and a snapshot of the record's value before overwriting. The flow log sequence number is incremented from the beginning based on the number of flows since the work order was created. The workflow completion timestamp field is temporarily set to empty during this write. It will be automatically filled back by the system after the work order completes the current node processing, thus obtaining the workflow node update result.

[0112] Step 8.3: Based on the update result of the workflow node, switch the processing status marker of the medical and health service work order from the pending allocation status to the dispatched pending processing status, and write the status change timestamp to obtain the processing status update result. Specifically, this includes: based on the obtained workflow node update result, switching the processing status marker of the medical and health service work order from the current pending allocation status to the dispatched pending processing status. The processing status marker is the enumeration value stored in the processing status field of the work order data table. The status switching operation also adopts a transactional write method: before writing the new status value, first read the original value of the current processing status field and record it in the status change log; then, overwrite the value of the processing status field with the enumeration value corresponding to the dispatched pending processing status; use the current system time of the status change operation as the timestamp of this status change and write it into the status change timestamp field associated with the work order. After the status switch is completed, the system triggers a new work order arrival notification to the doctor's workstation terminal corresponding to the target assigned doctor. The notification includes the work order number, a summary of the patient's basic information, the work order priority tag, and a summary of the standardized treatment path with the highest global aggregate score determined by the work order, thereby obtaining the processing status update result.

[0113] Step 8.4: Based on the processing status update result, archive all processing trajectory data of the work order since its creation to the processed work order record database in chronological order. The all processing trajectory data includes records of changes to each flow node, information of the processor at each stage, and records of each status change, to obtain the work order record of completed processing. Specifically, based on the obtained processing status update result, trigger the work order archiving process. The archiving process uses the unique work order number of the work order as the search condition to extract all records of the work order since its creation from the three data tables: the work order data table, the flow node log table, and the status change log table.

[0114] Extract the work order creation timestamp, patient unique identifier, work order source channel identifier, and work order type identifier from the work order data table. Extract all workflow node change records corresponding to the sequence numbers of all workflow logs for this work order from the workflow node log table. Each workflow node change record includes the source node identifier, target node identifier, workflow trigger timestamp, workflow completion timestamp, and the corresponding snapshot of the value taken before overwriting. Extract all status change records for this work order from the status change log. Each status change record includes the status value before the change, the status value after the change, the status change trigger timestamp, and the status change operation source identifier. The status change operation source identifier is used to distinguish whether the change was automatically triggered by the system or manually triggered. Arrange all the above extracted processing trajectory data in chronological order, starting from the work order creation timestamp and ending with the most recent status change timestamp, and assemble them into a complete work order lifecycle processing trajectory record. The complete lifecycle processing trajectory of the work order is inserted into the processed work order record library, and an archive category label is added to the archived record according to the main priority label of the work order. The archive category label is used to quickly retrieve processed work orders by priority category in subsequent statistical analysis.

[0115] In this embodiment of the invention, by accurately overwriting the target doctor and department identifiers, the automatic locking and seamless handover of the responsible party for the work order is achieved, effectively eliminating errors and delays caused by manual allocation. Furthermore, the flow node identifiers and multi-dimensional timestamps are synchronously written into the log table, constructing a transparent and traceable flow tracking system that ensures that the trajectory during multi-department collaboration is auditable and anomalies are locable. The work order status is automatically driven to switch to dispatched and pending, realizing refined state machine management of the work order lifecycle and providing an objective data benchmark for service timeliness assessment. Finally, the complete trajectory data, including all node changes, handlers, and status changes, is archived in deep chronological order, forming a structured digital archive of work orders.

[0116] like Figure 2 As shown, embodiments of the present invention also provide a collaborative processing system for a medical and health service work order knowledge base, including: The work order receiving and parsing module is used to receive medical and health service work orders; parse the patient's chief complaint and symptoms and signs in the medical and health service work orders to obtain the parsing results; and obtain structured work order feature data based on the parsing results. The knowledge base retrieval and matching module is used to retrieve disease localization detection data from a preset knowledge base based on the symptoms and signs in the structured work order feature data to obtain retrieval results; and to obtain a set of matching disease features based on the retrieval results. The health trend analysis module is used to extract historical health assessment results from the health trend map based on the matched disease feature set; and to obtain multi-dimensional health status distribution characteristics based on the extraction results. The collaborative diagnosis and treatment derivation module is used to deduce the correlation between corresponding preventive health care measures and preliminary diagnosis and treatment suggestions segment by segment based on the multi-dimensional health status distribution characteristics, so as to obtain the derivation results; based on the derivation results, a preliminary collaborative diagnosis and treatment plan is obtained. The global summary and accounting module is used to perform global summary and accounting based on the preliminary collaborative diagnosis and treatment plan and the standard treatment pathway in the biobank to obtain the accounting results; The workflow instruction generation module is used to obtain the final medical and health service work order workflow instructions based on the calculation results; The doctor and department matching module is used to match the corresponding doctor and department group according to the final medical and health service work order flow instructions to obtain the matching result; and to obtain the work order dispatch parameters based on the matching result. The work order status archiving module is used to update the processing status and flow nodes of medical and health service work orders according to the work order dispatch parameters to obtain the update results; based on the update results, the work order records that have been processed are obtained.

[0117] It should be noted that this system is a system corresponding to the above method. All implementation methods in the above method embodiments are applicable to this embodiment and can achieve the same technical effect.

[0118] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A collaborative processing method for a medical and health service work order knowledge base, characterized in that, The method includes: Step 1: Receive medical and health service work orders; parse the patient's chief complaint and symptoms in the medical and health service work orders to obtain the parsing results; based on the parsing results, obtain structured work order feature data; Step 2: Based on the symptoms and signs in the structured work order feature data, retrieve the disease localization detection data from the preset knowledge base to obtain the retrieval results; based on the retrieval results, obtain the matching disease feature set; Step 3: Extract historical health assessment results from the health trend map based on the matched disease feature set; obtain multi-dimensional health status distribution features based on the extraction results; Step 4: Based on the multi-dimensional distribution characteristics of health status, deduce the correlation between corresponding preventive healthcare measures and preliminary treatment recommendations segment by segment to obtain the deduction results; based on the deduction results, obtain a preliminary collaborative treatment plan. Step 5: Based on the preliminary collaborative treatment plan and the standard treatment pathways in the biobank, perform a comprehensive summary and calculation to obtain the calculation results; Step 6: Based on the calculation results, obtain the final medical and health service work order transfer instructions; Step 7: Based on the final medical and health service work order flow instructions, match the corresponding assigned doctor and department group to obtain the allocation result; based on the matching result, obtain the work order dispatch parameters; Step 8: Update the processing status and flow nodes of the medical and health service work orders according to the work order dispatch parameters to obtain the update results; based on the update results, obtain the work order records that have been processed.

2. The collaborative processing method for a medical and health service work order knowledge base according to claim 1, characterized in that, Step 1: Receive a medical and health service work order; The chief complaint and symptoms / signs of patients in medical and health service work orders are analyzed to obtain the analysis results; based on the analysis results, structured work order feature data is obtained, including: Based on the received medical and health service work orders, extract the patient's chief complaint text field, symptom description text field, and vital sign index value field from the work order form to obtain the original field extraction results; Based on the original field extraction results, the patient's chief complaint text field and the symptom description text field are processed by medical terminology segmentation and medical entity recognition to obtain preliminary parsing results containing medical entity labels and entity boundaries. Based on the preliminary analysis results, each identified medical entity is sequentially mapped to the corresponding standard term in the pre-set standard medical terminology glossary to obtain the terminology standardization processing results; Based on the terminology standardization process, the standardized chief complaint entity, symptom entity, and sign values ​​are assembled into key-value pair feature data according to a preset structured template. Each feature item is labeled with a data type label to obtain structured work order feature data.

3. The collaborative processing method for a medical and health service work order knowledge base according to claim 2, characterized in that, Step 2: Based on the symptoms and signs in the structured work order feature data, retrieve disease localization detection data from the preset knowledge base to obtain retrieval results; based on the retrieval results, obtain the matching disease feature set, including: By extracting the labeled symptom entities and signs from the structured work order feature data as retrieval query conditions, a symptom and sign query vector is formed. Based on the symptom and sign query vector, the disease localization detection data is searched in a preset knowledge base at multiple levels. The multi-level search sequentially executes three search levels: exact matching search, fuzzy matching search, and semantic association search, to obtain the disease entries hit at each search level and the matching degree score corresponding to each entry. Based on the disease entries hit at each retrieval level and the matching score of each entry, the disease entries returned by different retrieval levels are deduplicated and merged, and then sorted in descending order of matching score to obtain a sorted list of candidate diseases. Based on the sorted list of candidate diseases, select the candidate diseases whose matching scores rank within a preset threshold range, and extract the clinical feature description information corresponding to each candidate disease to obtain the matching disease feature set.

4. The collaborative processing method for a medical and health service work order knowledge base according to claim 3, characterized in that, Step 3: Based on the matched disease feature set, extract historical health assessment results from the health trend map; based on the extraction results, obtain multi-dimensional health status distribution features, including: Based on the matched set of disease features, and using the patient's unique identifier as an index, the historical health assessment records of the patient are queried in the health trend map to obtain the historical assessment query results. Based on the historical assessment query results, the values ​​of the detection indicators in each assessment record are arranged in ascending order by assessment timestamp. The direction and magnitude of change of the same detection indicator between two adjacent assessments are calculated one by one to obtain the health trend characteristics in the time dimension. Based on the health trend characteristics over time, the current values ​​of each test indicator of the patient are compared with the baseline reference values ​​of the same age group stored in the health trend map, and the deviation of each indicator is calculated to obtain the deviation analysis results. Based on the deviation analysis results, and combined with the typical abnormal patterns of indicators for each candidate disease in the matched disease feature set, a three-dimensional cross-summary was performed according to the indicator category dimension, time change dimension, and risk level dimension to obtain multi-dimensional health status distribution characteristics.

5. The collaborative processing method for a medical and health service work order knowledge base according to claim 4, characterized in that, Step 4: Based on the multi-dimensional distribution characteristics of health status, deduce the correlation between corresponding preventive healthcare measures and preliminary treatment recommendations segment by segment to obtain the derivation results; Based on the derivation results, a preliminary collaborative treatment plan was obtained, including: Each abnormal indicator in the multidimensional health status distribution features is mapped to a pre-set preventive healthcare measures knowledge base, and the corresponding intervention measures are searched to obtain a set of candidate preventive healthcare measures for each abnormal indicator. Based on the set of candidate preventive healthcare measures corresponding to each abnormal indicator, the causal chain between each preventive healthcare measure and the abnormal indicator it targets is deduced segment by segment in descending order of risk level, resulting in a segmented deduction result with causal annotation. Based on the step-by-step derivation results with causal annotations, for each candidate disease in the matched disease feature set, the confidence score is calculated by combining the symptom and sign matching degree and the deviation degree of historical health trend. The candidate diagnostic directions are selected from high to low according to the confidence score. For each candidate diagnostic direction, a corresponding preliminary diagnosis and treatment suggestion item is generated, resulting in a preliminary diagnosis and treatment suggestion list with priority sorting. Based on the set of candidate preventive healthcare measures and the list of preliminary treatment recommendations with priority, the measures are merged and organized according to the hierarchical structure of preventive measures taking precedence over treatment recommendations. Redundancy is then removed from the merged items to obtain a preliminary collaborative treatment plan.

6. The collaborative processing method for a medical and health service work order knowledge base according to claim 5, characterized in that, Step 5: Based on the preliminary collaborative treatment plan and the standard treatment pathways in the biobank, perform a comprehensive summary and calculation to obtain the calculation results, including: By extracting the disease codes and treatment direction identifiers corresponding to each treatment suggestion item in the preliminary collaborative treatment plan, and using the disease codes and treatment direction identifiers as joint search conditions, the standardized treatment path templates that match are queried in the biobank to obtain the standardized treatment path matching results. Based on the standardized treatment pathway matching results, the predefined treatment stages, required examination items for each stage, and expected treatment cycles for each stage are extracted from each standardized treatment pathway template. These are then compared item by item at the stage level with the corresponding items in the preliminary collaborative treatment plan to obtain the pathway comparison results. Based on the path comparison results, for each standardized treatment path, combined with the special risk factors marked in the multi-dimensional health status distribution characteristics of the patient, the standardized treatment path is checked one by one to see if there are any contraindications that conflict with the patient's special risk factors, and the path safety verification results are obtained. Based on the path safety verification results, each standardized treatment path that passed the safety verification was aggregated and calculated across the entire domain according to three accounting dimensions: resource consumption, time cycle, and departmental collaboration needs, to obtain the aggregated accounting results.

7. The collaborative processing method for a medical and health service work order knowledge base according to claim 6, characterized in that, Step 6: Based on the calculation results, obtain the final medical and health service work order transfer instructions, including: Based on the overall summary and accounting results, the urgency score, resource demand level, and expected processing cycle corresponding to each standardized treatment path in the accounting results are extracted to obtain the input parameters for the transfer decision. Based on the input parameters of the workflow decision, the urgency score is compared with the preset urgency judgment threshold to determine the priority label of the work order and obtain the work order priority judgment result. Based on the work order priority determination result and resource demand level, the corresponding flow path template is matched in the preset work order flow rule table to obtain the initial flow path; Based on the initial workflow path, and combined with the real-time service load status and available doctor availability information of each department, the availability of each department node in the initial workflow path is checked and adaptively adjusted one by one to obtain the final medical and health service work order workflow instruction.

8. The collaborative processing method for a medical and health service work order knowledge base according to claim 7, characterized in that, Step 7: Based on the final medical and health service work order flow instructions, match the corresponding assigned doctor and department group to obtain the allocation result; Based on the matching results, the work order dispatch parameters are obtained, including: Based on the final medical and health service work order flow instructions, extract the target department code and the required doctor's professional direction identifier specified in the instructions to obtain the assigned target parameters; Based on the assigned target parameters, a first-level screening is performed in the preset doctor resource information pool according to the target department code, and a second-level screening is performed in the selected department doctor set according to the required doctor professional direction identifier to obtain a candidate doctor set; Based on the candidate doctor set, obtain the current on-duty status identifier, the number of work orders currently being processed, and the field of expertise tag for each candidate doctor. Calculate the matching score between each candidate doctor and the current work order requirements to obtain the doctor matching score ranking result. Based on the doctor matching score ranking results, target doctors are selected from high to low matching score and assigned to them. The work order dispatch parameters are then assembled based on the department group identifier, doctor's employee number, and expected first response time of the target assigned doctor.

9. The collaborative processing method for a medical and health service work order knowledge base according to claim 8, characterized in that, Step 8: Update the processing status and workflow nodes of the medical and health service work orders according to the work order dispatch parameters to obtain the update results; based on the update results, obtain the work order records that have been completed, including: Based on the target doctor's employee number and department group identifier in the work order dispatch parameters, the current handler field and current handling department field of the medical and health service work order are overwritten with the target doctor's employee number and department group identifier, respectively, to obtain the handler update result; Based on the update result from the processor, the target node identifier in the work order dispatch parameters is written into the work order's flow node log table, and the trigger timestamp, source node identifier, and target node identifier of this flow are synchronously recorded in the log table to obtain the flow node update result; Based on the update results of the transfer nodes, the processing status of the medical and health service work order is changed from pending assignment to dispatched pending processing, and a status change timestamp is written to obtain the processing status update result. Based on the processing status update results, all processing trajectory data of the work order since its creation is archived in chronological order to the processed work order record database. The complete processing trajectory data includes records of changes to each flow node, information of the processor at each stage, and records of each status change, thus obtaining the work order record of completed processing.

10. A collaborative processing system for a medical and health service work order knowledge base, wherein the system implements the method as described in any one of claims 1 to 9, characterized in that, include: The work order receiving and parsing module is used to receive medical and health service work orders; The chief complaint and symptoms of patients in medical and health service work orders are analyzed to obtain the analysis results; based on the analysis results, structured work order feature data is obtained. The knowledge base retrieval and matching module is used to retrieve disease localization detection data from a preset knowledge base based on the symptoms and signs in the structured work order feature data to obtain retrieval results; and to obtain a set of matching disease features based on the retrieval results. The health trend analysis module is used to extract historical health assessment results from the health trend map based on the matched disease feature set; and to obtain multi-dimensional health status distribution characteristics based on the extraction results. The collaborative diagnosis and treatment derivation module is used to deduce the correlation between corresponding preventive health care measures and preliminary diagnosis and treatment suggestions segment by segment based on the multi-dimensional health status distribution characteristics, so as to obtain the derivation results; based on the derivation results, a preliminary collaborative diagnosis and treatment plan is obtained. The global summary and accounting module is used to perform global summary and accounting based on the preliminary collaborative diagnosis and treatment plan and the standard treatment pathway in the biobank to obtain the accounting results; The workflow instruction generation module is used to obtain the final medical and health service work order workflow instructions based on the calculation results; The doctor and department matching module is used to match the corresponding doctor and department group according to the final medical and health service work order flow instructions to obtain the matching result; and to obtain the work order dispatch parameters based on the matching result. The work order status archiving module is used to update the processing status and flow nodes of medical and health service work orders according to the work order dispatch parameters to obtain the update results; based on the update results, the work order records that have been processed are obtained.

Images