Purchase task demonstration system and method suitable for hospital return-to-mouth management department

Abstract

The invention discloses a procurement task demonstration system and method suitable for a hospital return management department, and belongs to the technical field of intelligent medical treatment.The procurement task demonstration system and method comprises the steps that treatment path data of patients of the whole hospital in a target period is obtained, and a treatment path sequence set is constructed; a fixed connection mode between treatment units of different departments is identified through analysis; determining a medical material set associated with the fixed connection mode according to the standard department material relation mapping table; collaborative purchase prediction judgment is executed on each material set, the number of patients who will experience a related fixed connection mode is estimated by analyzing the treatment progress of the patients in the current period, and the collaborative prediction demand is calculated in combination with the standard unit dosage; and comparing the predicted demand quantity with the stock quantity and a safety threshold, generating a preliminary list, and performing credibility weighting by using a historical deviation record to form a final collaborative purchase prediction list. According to the invention, the procurement argument mode conversion from passive summarization to active prediction and from department isolation to hospital-wide collaboration is realized.

Description

Technical Field

[0001] This invention discloses a procurement task verification system and method applicable to hospital management departments, belonging to the field of smart healthcare technology. Background Technology

[0002] Currently, hospital management departments primarily rely on two traditional models when conducting medical supply procurement needs assessments. The first is trend extrapolation based on historical consumption data, which involves predicting future demand through simple time-series analysis of past departmental material requisition or consumption records. This method treats departments as independent consumption units, neglecting the dynamic and collaborative material needs arising from patient transfers between departments in modern multidisciplinary collaborative treatment models. Essentially, it is a static and lagging prediction, unable to respond to new treatment combinations or changes in treatment pathways, and prone to distorting the prediction of certain supplies.

[0003] Secondly, each clinical department submits its own procurement requests based on its experience, which are then compiled and reviewed by the relevant management department. While this approach reflects the immediate needs of each department to some extent, it lacks systematic coordination and forward-looking assessment at the hospital-wide level. Departments, acting solely from their own perspective, struggle to accurately predict the chain reaction of demand for supplies from downstream departments caused by changes in treatment volume in upstream departments. This easily leads to an imbalance in procurement plans, resulting in either stockpiling of some supplies or shortages of critical materials in collaborative treatment processes.

[0004] In summary, existing technologies lack a procurement justification method that can proactively identify cross-departmental treatment collaboration patterns based on the actual treatment process of all patients in the hospital, and accurately predict collaborative material needs accordingly. This results in insufficient scientific rigor, coordination, and foresight in hospital procurement plans, making it difficult to accurately guarantee the supply of materials in patient-centered, continuous medical activities, and also affecting the efficiency of hospital inventory fund utilization. Therefore, there is an urgent need for a new procurement task justification method that can be deeply integrated into the medical business process and achieve intelligent collaborative prediction of needs. Summary of the Invention

[0005] To achieve the above objectives, this application provides the following technical solution: A method for evaluating procurement tasks applicable to hospital management departments, characterized by comprising the following steps: S1, in response to the procurement justification instruction, obtains treatment pathway data for all patients in the hospital within the target justification period; S2, Based on the treatment path data, construct a set of treatment path sequences for the patient group; S3, parse the set of treatment path sequences, and identify a fixed connection pattern in all treatment path sequences that consists of at least two treatment units from different departments. In the course of historical patient treatment, after completing the treatment unit of the previous department, the patient has a predetermined probability of entering the next department for treatment unit connection. S4, Based on the hospital's pre-set standard departmental material relationship mapping table, determine the set of medical supplies associated with all departmental treatment units in the fixed connection mode; S5. Based on all treatment pathway sequences containing the fixed connection pattern associated with the material set, the estimated number of patients expected to generate demand for various materials in the material set through the fixed connection pattern within the target demonstration period is calculated. Based on the estimated number of patients and the standard unit usage in the standard department material relationship mapping table, the collaborative predicted demand for each type of material in the material set before the end of the target demonstration period is calculated. S6. Compare all the collaboratively predicted demand quantities with the current inventory and safety threshold of the materials in the hospital material management system, filter out the materials whose difference between the predicted demand quantity and the current inventory exceeds the safety threshold, and generate a preliminary collaborative procurement prediction list. S7. Retrieve the deviation records between the actual purchase quantity and the collaborative forecast demand quantity of each material in the preliminary collaborative procurement forecast list within the historical demonstration period of the predetermined quantity in the past. Based on the deviation records, perform credibility weighting on the collaborative forecast demand quantity of the current demonstration period to generate a final collaborative procurement forecast list with credibility weights.

[0006] Furthermore, S2 also includes: S21, standardize the treatment path data of each patient obtained by S1 to ensure that the identification of each department's treatment unit conforms to the hospital's unified coding standard; S22, sort the standardized departmental treatment unit sequence for each patient according to the recorded occurrence time or logical order to form the patient's initial treatment path sequence; S23 marks the status identifier for each departmental treatment unit node in the initial treatment path sequence for each patient. S24, gather the treatment path sequences of all patients' marked status identifiers to form the treatment path sequence set.

[0007] Furthermore, S3 also includes: S31, Traverse the set of treatment path sequences and extract all sub-sequences in each treatment path sequence that consist of at least two different departmental treatment unit identifiers; S32, establish a candidate connection pattern set. For each subsequence, convert it into a standard pattern expression, arrange its identifiers according to the order of appearance of the department treatment unit in the subsequence, and add the standard pattern expression as a candidate pattern to the candidate connection pattern set. S33, For each candidate pattern in the candidate connection pattern set, perform global frequency statistics in the treatment path sequence set; S34, set a first frequency threshold and a second frequency threshold. The first frequency threshold is used to determine whether the total frequency of the pattern is significant. The second frequency threshold is used to determine the frequency of the same pattern repeating in a single patient sequence. S35, for each candidate pattern in the candidate connection pattern set, determine whether its total frequency of the pattern is greater than the first frequency threshold, and whether the quotient obtained by dividing the total number of times the pattern appears by the total frequency of the pattern is less than the second frequency threshold. If both conditions are met, then the candidate pattern is determined to be a high-frequency and relatively discrete pattern. S36. For each patient treatment path sequence containing the candidate pattern, record the time interval or sequence interval between the treatment unit nodes of each department contained in the candidate pattern each time it appears in the sequence, and calculate the statistical dispersion of the interval between treatment unit nodes of the same sequence position in all instances of the candidate pattern. S37, Set a dispersion threshold, compare the calculated statistical dispersion with the dispersion threshold, and if the statistical dispersion is lower than the dispersion threshold, determine that the candidate pattern has stability in the time or order dimension. S38. Candidate patterns that simultaneously meet the judgment conditions of S35 and S37 are formally identified as the fixed connection patterns, and their standard pattern expression, total pattern frequency, and average occurrence interval dispersion are recorded as pattern feature information.

[0008] Furthermore, S5 also includes: S51, for the set of medical supplies currently being judged, determine the fixed connection patterns associated with them, and obtain a subset of all treatment path sequences associated with these fixed connection patterns; S52, for each patient treatment path sequence in the subset of treatment path sequences, analyze its current status and expected progress within the remaining time of the target demonstration period: S53, cumulatively all patients identified in S52 as included in the estimated number of patients, to obtain the estimated number of patients N for the current medical supplies set and its associated fixed connection mode; S54. Based on the standard departmental material relationship mapping table, obtain the standard unit dosage of each category of material in the current medical material set under each departmental treatment unit involved in its associated fixed connection mode. S55, For the i-th type of supplies in the current medical supplies set, calculate its collaborative predicted demand Di: Di ​​= N * Σ (standard unit usage under each related department treatment unit); S56, repeat S51 to S55 until the collaborative procurement forecast and determination for all medical supplies is completed.

[0009] Furthermore, S52 also includes: S521, If ​​the current state of the patient sequence is that it has fully experienced all the treatment units of the fixed connection pattern it is associated with, and is not currently in any treatment unit of the pattern, then no expected demand based on the pattern will be generated in the current demonstration period, and the patient will not be included in the estimated number of patients. S522, if the current state of the patient sequence is that it is in the intermediate department treatment unit of a certain fixed connection pattern associated with it, then determine the probability that the patient is expected to complete the subsequent department treatment unit of the fixed connection pattern before the end of the target demonstration period; this probability is logically inferred based on the historical data of the average treatment course of this type of treatment in the hospital; if it is inferred that the probability of completion is high, then the patient is included in the estimated number of patients. S523, if the current state of the patient sequence is that the preceding part of the treatment unit of a certain fixed connection mode has been completed, but the subsequent key treatment unit of the mode has not yet been entered, then based on the patient's completed treatment progress, routine medical pathway planning and the remaining time of the target demonstration cycle, the probability of the patient entering and completing the fixed connection mode before the end of the cycle is comprehensively estimated. If the probability is estimated to be high, then the patient is included in the estimated number of patients.

[0010] Furthermore, in steps S522 and S523, the judgment that is presumed to be of high probability of completion or high probability of entering and completing is verified by means of the execution status of medical orders in the hospital information system, the timestamp data of the scheduled surgery or examination, and the department's bed availability and scheduling plan. When the auxiliary verification information supports a certainty exceeding a preset certainty threshold within the target validation period, the patient is included in the estimated number of patients.

[0011] Furthermore, S6 also includes: S61, for each item in the final collaborative procurement forecast list, obtain its current available inventory C and the pre-set inventory safety threshold S from the hospital's material management system in real time. S62, calculate the difference Δ = D - C between the collaborative forecast demand D and the current available inventory C for this material; S63, determine whether the difference Δ is greater than the safety threshold S; if Δ>S, determine that the material needs to be included in the procurement demonstration scope and kept in the list; otherwise, remove the material from the preliminary collaborative procurement forecast list.

[0012] Furthermore, the S7 also includes: S71, for each item in the preliminary collaborative procurement forecast list, retrieve the collaborative forecast demand generated by this method in each of the past M historical demonstration periods, and the actual procurement quantity of the item after the end of the corresponding period. S72, calculate the prediction deviation rate ε_i for each historical period i = |actual purchase quantity - collaboratively predicted demand quantity| / collaboratively predicted demand quantity; S73, Based on the prediction deviation rate of the past M periods, calculate the average prediction deviation rate E and the deviation stability coefficient σ of the material; the deviation stability coefficient reflects the degree of dispersion of the deviation rate in each historical period. S74, Based on the predicted average deviation rate E and the deviation stability coefficient σ, the confidence weight factor ω is determined through a preset weight mapping rule; S75, multiply the collaborative forecast demand D of the material in the current demonstration period by the credibility weight factor ω to obtain the weighted collaborative forecast demand D' = D * ω, and record the value of ω.

[0013] Furthermore, the method also includes a periodic calibration step, which is performed after each predetermined number of procurement justification cycles, including: S91, Collect the collaborative predicted demand for each material generated by this method in each of the most recently completed demonstration cycles of the predetermined number of times, as well as the actual consumption data of the materials in the corresponding cycles. S92, compare the predicted demand and actual consumption of the same material in different cycles to analyze the consistency of their changing trends and systematic deviations. S93, based on the analysis results of S92, propose revision suggestions for the relevant standard unit usage in the standard department material relationship mapping table, or propose adjustment suggestions for the first frequency threshold, second frequency threshold or dispersion threshold used to identify the fixed connection mode in S3, and submit the revision suggestions or adjustment suggestions for review. After the review is approved, they will be used to update the demonstration configuration for subsequent cycles.

[0014] According to a second aspect of the present invention, the present invention claims protection for a procurement task justification system applicable to hospital centralized management departments, comprising: One or more processors; A memory storing one or more programs, which, when executed by one or more processors, enable the processors to implement a procurement task justification method applicable to hospital management departments.

[0015] This invention discloses a procurement task verification system and method suitable for hospital centralized management departments, belonging to the field of smart healthcare technology. It acquires treatment path data of all patients in the hospital within a target period and constructs a set of treatment path sequences. Through analysis, it identifies fixed connection patterns between treatment units in different departments. Based on a standard departmental material relationship mapping table, it determines the set of medical supplies associated with the fixed connection patterns. For each material set, it performs collaborative procurement prediction and judgment, estimating the number of patients experiencing relevant fixed connection patterns by analyzing the treatment progress of patients in the current period, and calculating the collaborative predicted demand based on standard unit usage. The predicted demand is compared with inventory levels and safety thresholds to generate a preliminary list, and historical deviation records are used for credibility weighting to form the final collaborative procurement prediction list. This invention realizes a transformation in procurement verification mode from passive aggregation to proactive prediction, and from departmental isolation to hospital-wide collaboration. Attached Figure Description

[0016] Figure 1 A flowchart illustrating a procurement task justification method applicable to hospital centralized management departments, as claimed in this embodiment of the invention; Figure 2 The second flowchart of a procurement task verification method applicable to hospital centralized management departments, as claimed in an embodiment of the present invention; Figure 3 The third workflow diagram of a procurement task verification method applicable to hospital centralized management departments, as claimed in an embodiment of the present invention; Figure 4 The fourth workflow diagram of a procurement task verification method applicable to hospital centralized management departments, as claimed in this embodiment of the invention; Figure 5 The fifth flowchart of a procurement task verification method applicable to hospital centralized management departments, as claimed in this embodiment of the invention. Detailed Implementation

[0017] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0018] The terms "first," "second," and "third" in this application are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first," "second," or "third" may explicitly or implicitly include at least one of those features. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified. All directional indications in the embodiments of this application, such as up, down, left, right, front, back, etc., are only used to explain the relative positional relationships and movements between components in a specific orientation as shown in the accompanying drawings. If the specific orientation changes, the directional indications will change accordingly. Furthermore, the terms "including" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0019] References to embodiments herein mean that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0020] According to a first embodiment of the present invention, the present invention claims protection for a procurement task justification method applicable to hospital centralized management departments, referring to... Figure 1 The method includes the following steps: S1, in response to the procurement justification instruction, obtains treatment pathway data for all patients in the hospital within the target justification period; S2, Based on treatment pathway data, construct a set of treatment pathway sequences for the patient population; S3, analyze the set of treatment pathway sequences, and identify a fixed connection pattern in all treatment pathway sequences that consists of at least two treatment units from different departments. In the course of a patient's historical treatment, after completing the treatment unit of the previous department, the patient has a predetermined probability of entering the next department for treatment unit connection. S4, based on the hospital's pre-set standard departmental material relationship mapping table, determines the set of medical supplies associated with all departmental treatment units in the fixed connection mode; S5. Based on all treatment pathway sequences containing the fixed connection pattern associated with the material set, the estimated number of patients expected to generate demand for various materials in the material set through the fixed connection pattern within the target demonstration period is calculated. Based on the estimated number of patients and the standard unit usage in the standard department material relationship mapping table, the collaborative predicted demand for each type of material in the material set before the end of the target demonstration period is calculated. S6 compares all collaboratively predicted demand with the current inventory and safety threshold of materials in the hospital's material management system, filters out materials whose difference between predicted demand and current inventory exceeds the safety threshold, and generates a preliminary collaborative procurement forecast list. S7 retrieves the deviation records between the actual purchase quantity and the collaborative forecast demand quantity of each material in the preliminary collaborative procurement forecast list within the historical demonstration period of the predetermined quantity in the past. Based on the deviation records, the collaborative forecast demand quantity of the current demonstration period is weighted by credibility, and a final collaborative procurement forecast list with credibility weights is generated.

[0021] In this embodiment, an electronic procurement justification instruction from the relevant management department is received and parsed. This instruction contains at least a clearly defined target justification period. In response to this instruction, the hospital automatically queries and extracts anonymized treatment pathway data from its core business database for all inpatients and patients planned to be admitted within the target justification period, from the start of the target justification period to the current time. The treatment pathway data is organized by patient, and each data entry contains an irreversibly anonymized unique patient identifier, a list of all departmental treatment units that the patient has completed within the time period and their implementation times, and a unique code for the departmental treatment unit that the patient is currently registered to receive treatment at the current time. Based on the raw data obtained from S1, a structured set of treatment path sequences is constructed for all patients. Specifically, an independent treatment path sequence object is created for each patient. This sequence object arranges the codes of each departmental treatment unit that the patient has completed in strict chronological order of the treatment actions. Then, at the end of this time sequence, the code of the departmental treatment unit that the patient is currently undergoing is appended, and the last unit is marked as in progress. The treatment path sequence objects of all individual patients together constitute the set of treatment path sequences, which serves as the unified data foundation for subsequent analysis. Deep pattern mining is performed on the treatment path sequence set constructed by S2 to identify recurring combinations of departmental treatment units with temporal regularity. Each patient treatment path sequence in the set is traversed one by one, searching for all continuous subsequences of departmental treatment units with a length greater than or equal to 2. For each such subsequence, it is abstracted into a candidate connection pattern, which records the sequential relationship of the departmental treatment unit codes within the subsequence. Subsequently, in all patient treatment path sequences, the total number of patients and the total number of occurrences of each candidate connection pattern are counted. Based on preset criteria for the significance and dispersion of occurrence frequency, patterns that are both ubiquitous and non-personally frequently repeated are selected from numerous candidate patterns. For the selected patterns, the stability of the time interval or sequence interval between units when they occur in different patient sequences is further analyzed. Finally, patterns that simultaneously meet the three conditions of high frequency, non-personally repeated, and temporally stable are formally identified as fixed connection patterns, and their complete unit sequence, occurrence intensity, and stability metric are recorded. Access the hospital's pre-maintained and approved knowledge base of material consumption standards, which exists in the form of a standard departmental material relationship mapping table. For each fixed connection pattern identified by S3, based on the code of each departmental treatment unit included in the pattern, retrieve from the mapping table each type of medical material that is routinely consumed to complete the standard diagnosis and treatment activities of that unit, including its unique material code, standard specification, and the theoretical consumption quantity corresponding to each patient unit served, such as a single treatment or a single day of hospitalization, i.e., the standard unit usage. All related materials belonging to the same fixed connection pattern but possibly scattered under different departmental treatment units are aggregated into a medical material set strongly related to that pattern. For each medical supply set generated by S4, a core collaborative procurement prediction and judgment operation is performed. This operation first locates all patient sequences in the treatment path sequence set that contain fixed connection patterns associated with that supply set. Then, it analyzes the expected treatment progress of these patients in the remaining time of the current target demonstration period: for patients who have not yet started the pattern, have started but not completed it, or are in the middle of the pattern, it logically infers the probability that they will complete all or part of the subsequent treatment units covered by the pattern before the end of the target demonstration period, based on their completed treatment content, the standard treatment cycle length of the department, and the appointment scheduling information in the hospital information. Patients inferred to have a high probability of completion are included in the statistical scope, and the estimated number of patients for that supply set is calculated. Finally, based on the standard unit usage of each supply in the supply set under its associated treatment units in each department as defined in the standard department supply relationship mapping table, it is multiplied by the estimated number of patients, and the potential repeated consumption caused by the same patient experiencing different units in the pattern is accumulated, thereby calculating the collaborative predicted demand of the supply before the end of the target demonstration period. The collaborative forecast demand list of all materials calculated by S5 is compared one by one with the real-time current available inventory of the corresponding materials recorded in the hospital's material management system and the pre-set inventory safety threshold for each type of material. For each material, the difference between its collaborative forecast demand and current available inventory is calculated and compared with the inventory safety threshold. Only when the difference is greater than the inventory safety threshold is it determined that there is a substantial inventory gap based on the collaborative forecast for that material, and it is retained. All retained materials and their related information constitute a preliminary collaborative procurement forecast list. To enhance the reliability of forecasts, a historical deviation calibration mechanism is introduced. From the historical data database, the collaborative forecast demand and actual procurement volume of each item in the preliminary collaborative procurement forecast list, generated by this method, are retrieved over several complete demonstration periods. The forecast deviation ratio for each historical period is calculated, and the average level and fluctuation of these deviation ratios are analyzed. Based on this historical deviation analysis, a reliability weighting factor is assigned to the collaborative forecast demand for the current period. This factor reflects the accuracy and stability of historical forecasts; the higher the weighting factor, the higher the historical reference reliability of the current forecast value. This weighting factor is used to weight and correct the current collaborative forecast demand, generating a weighted collaborative forecast demand. The list is then updated accordingly, forming a final collaborative procurement forecast list with reliability weights.

[0022] Furthermore, referring to Figure 2 In S2, it also includes: S21, standardize the treatment path data of each patient obtained by S1 to ensure that the identification of each department's treatment unit conforms to the hospital's unified coding standard; S22, sort the standardized departmental treatment unit sequence for each patient according to the recorded occurrence time or logical order to form the patient's initial treatment path sequence; S23 marks the status identifier for each departmental treatment unit node in the initial treatment path sequence for each patient. S24, gather the treatment path sequences of all patients' marked status identifiers to form the treatment path sequence set.

[0023] In this embodiment, data standardization and cleaning are performed on each original patient treatment path data acquired in S1. This operation first verifies and standardizes the format of all departmental treatment unit identifiers, ensuring they fully comply with the latest coding standards published by the hospital's master data management system. For data with missing codes, incorrect formats, or using outdated codes, automatic conversion or anomaly marking is performed according to preset coding mapping rules. Simultaneously, treatment time point information is standardized into an internal timestamp format, ensuring logical consistency in the time sequence.

[0024] For each patient whose data has been standardized, all their departmental treatment units are arranged in strict ascending chronological order according to their corresponding standardized treatment time points. For records of multiple different treatments occurring on the same day with identical time points, a secondary sorting is performed based on the usual logical sequence of diagnostic and treatment activities, such as diagnosis before treatment, and preoperative preparation before surgery. This generates an ordered list of departmental treatment unit codes arranged chronologically for each patient, known as the patient's initial treatment path sequence.

[0025] Each departmental treatment unit node in the initial treatment pathway sequence is labeled with a status. The labeling is based on the relationship between the treatment time point recorded in each treatment unit and the current time point, and whether the record is marked as completed. Specifically, unit nodes whose treatment time points in the sequence are earlier than or equal to the current time point and whose recorded status is completed are marked as completed. For nodes whose last time point in the sequence is not later than the current time, and whose patient is still in the treatment unit's consultation status according to real-time business feedback (such as hospitalization or surgery), the node is specially marked as in progress. This status labeling is crucial for distinguishing between historical treatment and anticipated future treatment.

[0026] A global collection container is created to hold the treatment pathway sequences of all patients. Each patient's treatment pathway sequence, marked as completed (output from S23), is added to this collection container as an independent data element. This container constitutes a complete set of treatment pathway sequences, where each element has a unified format: a time-sorted list of departmental unit codes, and a clear status identifier for each node, including completed or in progress. This provides a structured and status-rich data foundation for subsequent pattern mining and progress analysis.

[0027] Furthermore, S3 also includes: S31, Traverse the set of treatment path sequences and extract all sub-sequences in each treatment path sequence that consist of at least two different departmental treatment unit identifiers; S32, establish a candidate connection pattern set. For each subsequence, convert it into a standard pattern expression, arrange its identifiers according to the order of appearance of the department treatment unit in the subsequence, and add the standard pattern expression as a candidate pattern to the candidate connection pattern set. S33, For each candidate pattern in the candidate connection pattern set, perform global frequency statistics in the treatment path sequence set; S34, set a first frequency threshold and a second frequency threshold. The first frequency threshold is used to determine whether the total frequency of the pattern is significant. The second frequency threshold is used to determine the frequency of the same pattern repeating in a single patient sequence. S35, for each candidate pattern in the candidate connection pattern set, determine whether its total frequency of the pattern is greater than the first frequency threshold, and whether the quotient obtained by dividing the total number of times the pattern appears by the total frequency of the pattern is less than the second frequency threshold. If both conditions are met, then the candidate pattern is determined to be a high-frequency and relatively discrete pattern. S36. For each patient treatment path sequence containing the candidate pattern, record the time interval or sequence interval between the treatment unit nodes of each department contained in the candidate pattern each time it appears in the sequence, and calculate the statistical dispersion of the interval between treatment unit nodes of the same sequence position in all instances of the candidate pattern. S37, Set a dispersion threshold, compare the calculated statistical dispersion with the dispersion threshold, and if the statistical dispersion is lower than the dispersion threshold, determine that the candidate pattern has stability in the time or order dimension. S38. Candidate patterns that simultaneously meet the judgment conditions of S35 and S37 are formally identified as the fixed connection patterns, and their standard pattern expression, total pattern frequency, and average occurrence interval dispersion are recorded as pattern feature information.

[0028] In this embodiment, a traversal process is initiated to sequentially read each patient treatment path sequence in the treatment path sequence set. For each sequence, starting from its starting position, a sliding window technique is used to extract all possible consecutive subsequences. The length of these subsequences starts from 2 and extends up to the total number of completed state nodes in the sequence. For example, for a sequence with 5 completed units, all consecutive subsequences of length 2, 3, 4, and 5 will be extracted. Each extracted subsequence retains the original order of its internal departmental treatment unit codes.

[0029] Initialize an empty set of candidate connection patterns. For each continuous subsequence extracted in S31, convert it into a standardized string expression. This expression is formed by sequentially concatenating the codes of the departmental treatment units within the subsequence, separated by a specific delimiter such as -> to clearly express their sequential relationship. For example, the subsequence [Department A Surgery, Department B Resuscitation] is converted into the expression Department A Surgery -> Department B Resuscitation. Check if the exact same expression already exists in the candidate connection pattern set. If not, add this expression as a new candidate connection pattern to the set; if it already exists, do not add it again. This step aims to collect all unique combinations of departmental treatment unit sequences that have appeared in the data.

[0030] A full data scan is performed on each candidate pattern in the candidate connection pattern set to calculate its key statistical indicators. The entire treatment pathway sequence set is then re-traversed, and for each patient sequence, the candidate pattern is checked to see if it appears as a consecutive subsequence. Two core indicators are calculated: the total frequency of the pattern (how many different patient sequences contain the pattern regardless of how many times it appears); and the total number of times the pattern appears (the cumulative number of times the pattern appears as a consecutive subsequence across all patient sequences, possibly appearing multiple times within the same patient sequence).

[0031] Administrators or the knowledge base predefine quantitative thresholds for filtering. A first frequency threshold, such as total frequency > 5% of total patients, is used to determine whether a pattern is prevalent in the patient population, rather than appearing sporadically. A second frequency threshold, such as average occurrences / total frequency < 2, is used to determine whether the pattern is scattered across different patients, rather than concentrated in a few patients with high-frequency repetition; the latter may represent complex conditions in individual patients rather than a general process.

[0032] For each candidate pattern, a first-level screening logic is applied. It calculates the quotient of the total number of occurrences of the candidate pattern divided by the total frequency of the pattern itself. This quotient reflects the average number of occurrences of the pattern in a single patient. Then, it determines whether the total frequency of the candidate pattern is greater than a first frequency threshold and whether the calculated quotient is less than a second frequency threshold. Only candidate patterns that simultaneously meet both conditions are considered to be widely present in the patient population, exhibiting high frequency and relatively discrete distribution, and are therefore considered non-individually high-frequency repetitions. This allows them to pass this screening stage and proceed to a more in-depth temporal stability analysis phase.

[0033] For candidate patterns selected through S35 screening, a detailed temporal stability analysis was performed. For each such pattern, all patient sequences containing that pattern were scanned again. For each specific occurrence of the pattern in each sequence, the position index or corresponding timestamp of each departmental treatment unit node within it was precisely recorded in the original treatment path sequence. Then, the statistical distribution of the interval position difference or time difference from the first unit to the second unit was calculated across all occurrences of the pattern, such as calculating the standard deviation or coefficient of variation. Similarly, the interval distribution from the second unit to the third unit was calculated for patterns with a length greater than 2, and so on. The purpose of this step is to quantify the tightness and regularity of the temporal or sequential connections between the steps within the pattern.

[0034] Temporal stability is assessed based on another predefined dispersion threshold, such as a coefficient of variation of the time interval being less than 0.5. The statistical dispersion of each inter-unit interval calculated in S36 is compared to the dispersion threshold. If the statistical dispersion of all inter-unit intervals of a candidate pattern is below the dispersion threshold, the pattern is deemed to have high stability and predictability in the temporal or sequential dimension, indicating a regular medical pathway component.

[0035] Candidate patterns that successfully passed the rigorous dual screening of S35 high-frequency and discrete patterns and S37 time-series stability were officially recognized as fixed connection patterns. A complete profile was created for each recognized fixed connection pattern, recording its standard pattern expression, total pattern frequency, total number of occurrences, average value and dispersion of the intervals between units, and other characteristic information. These profiles constitute the core knowledge basis for subsequent collaborative demand forecasting.

[0036] Furthermore, referring to Figure 3 S5 also includes: S51, for the set of medical supplies currently being judged, determine the fixed connection patterns associated with them, and obtain a subset of all treatment path sequences associated with these fixed connection patterns; S52, for each patient treatment path sequence in the subset of treatment path sequences, analyze its current status and expected progress within the remaining time of the target demonstration period: S53, cumulatively all patients identified in S52 as included in the estimated number of patients, to obtain the estimated number of patients N for the current medical supplies set and its associated fixed connection mode; S54. Based on the standard departmental material relationship mapping table, obtain the standard unit dosage of each category of material in the current medical material set under each departmental treatment unit involved in its associated fixed connection mode. S55, For the i-th type of supplies in the current medical supplies set, calculate its collaborative predicted demand Di: Di ​​= N * Σ (standard unit usage under each related department treatment unit); S56, repeat S51 to S55 until the collaborative procurement forecast and determination for all medical supplies is completed.

[0037] In this embodiment, a specific set of medical supplies is processed first. First, the unique identifiers of one or more associated fixed connection patterns are read from the metadata of this set of supplies. Then, a filtering query is performed on the set of treatment pathway sequences to identify all patient records whose treatment pathway sequences contain any of the aforementioned associated fixed connection patterns. These filtered patient sequences constitute the target patient subset for this prediction.

[0038] For each patient in the target patient subset, an individualized treatment progress analysis and future expectation projection are performed. This is the core of the prediction and judgment, employing a case-by-case processing logic; After completing the individual analysis of all patients in the target patient subset, the counts of all patients marked as included in the estimated patient count in S52 are summed. The resulting value N is the estimated number of patients expected to generate material consumption needs within the target validation period, based on the current medical supply set and its associated fixed connection pattern.

[0039] Based on the standard departmental material relationship mapping table, perform a precise material consumption standard query. For each material in the current medical material set, identified by its unique material code, query the defined standard unit usage under each specific departmental treatment unit covered by the associated fixed connection mode. For example, for the mode A surgery -> B resuscitation, query the standard usage of a certain hemostatic material in a single surgery under the A surgery unit, and the standard daily usage under the B resuscitation unit.

[0040] For the i-th type of medical supplies in the set, calculate its collaborative predicted demand Di. The logic of the calculation formula is: Di = estimated number of patients N × sum of standard unit usage of this supply in all relevant departmental treatment units of the associated fixed connection mode. In specific calculation, the standard unit usage of this supply in each unit of the mode, as queried in S54, is summed to obtain the average total consumption of this supply for a single patient experiencing the complete mode. Then, this total consumption for a single patient is multiplied by the estimated number of patients N. If a supply is consumed in different units of the mode, its usage is accumulated during the summation. This step combines the prediction of patient flow with the standard supply consumption of a single treatment, transforming it into a specific supply demand prediction.

[0041] S51 to S55 are executed in a loop. After each set of medical supplies is determined, its predicted demand is stored. Then, the next set of medical supplies to be processed is automatically selected, and the entire process is repeated. The overall determination task of S5 is completed only after all medical supply sets generated by S4 have been processed.

[0042] Furthermore, referring to Figure 4 S52 also includes: S521, If ​​the current state of the patient sequence is that it has fully experienced all the treatment units of the fixed connection pattern it is associated with, and is not currently in any treatment unit of the pattern, then no expected demand based on the pattern will be generated in the current demonstration period, and the patient will not be included in the estimated number of patients. S522, if the current state of the patient sequence is that it is in the intermediate department treatment unit of a certain fixed connection pattern associated with it, then determine the probability that the patient is expected to complete the subsequent department treatment unit of the fixed connection pattern before the end of the target demonstration period; this probability is logically inferred based on the historical data of the average treatment course of this type of treatment in the hospital; if it is inferred that the probability of completion is high, then the patient is included in the estimated number of patients. S523, if the current state of the patient sequence is that the preceding part of the treatment unit of a certain fixed connection mode has been completed, but the subsequent key treatment unit of the mode has not yet been entered, then based on the patient's completed treatment progress, routine medical pathway planning and the remaining time of the target demonstration cycle, the probability of the patient entering and completing the fixed connection mode before the end of the cycle is comprehensively estimated. If the probability is estimated to be high, then the patient is included in the estimated number of patients.

[0043] In this embodiment, individualized treatment progress analysis and future prediction are performed for each patient in the target patient subset. This is the core of the prediction and judgment, employing a case-by-case processing logic: Scenario 1: Completed Pattern: Examine the patient's treatment pathway sequence. If the sequence shows that all departmental treatment units involved in the associated fixed connection pattern have been marked as completed, and the patient's current treatment unit does not belong to any part of that pattern, then it is determined that the patient's treatment needs based on this pattern have been met in the past. During this target validation period, the patient will not have any new material needs due to this pattern. Therefore, this patient is not included in the estimated patient count for the current period.

[0044] Scenario Two: In-Progress Mode. Examine the patient's treatment pathway sequence. If the sequence shows that the patient is currently in a specific departmental treatment unit within a fixed connection mode (i.e., the unit is marked as in progress) and has not yet completed all subsequent units of the mode, it is necessary to infer the probability that the patient will complete the remaining treatment units of the mode before the end of the target justification period. The inference is based on: the standard treatment duration of the remaining units of the mode derived from the medical pathway knowledge base, the patient's length of hospital stay, and the scheduled surgeries or examinations in the hospital information. For example, if the patient is currently in the in-progression unit of Department A surgery -> Department B resuscitation mode, and the standard period for Department B resuscitation is 3 days, with 5 days remaining in the target justification period and no scheduling conflicts, then logically, it is presumed that completion is highly probable. Patients meeting this condition are expected to receive subsequent treatment within the period and are therefore included in the estimated patient count.

[0045] Scenario 3: Pending Start Mode: Examine the patient's treatment pathway sequence. If the sequence shows that the patient has completed some pre-processing units of the associated fixed-connection mode and marked them as completed, but has not yet started the key subsequent units of the mode—these units are neither completed nor in progress—a comprehensive assessment of the patient's likelihood of initiating and completing the mode within the cycle is required. Assessment factors include: the patient's current condition based on diagnostic grouping, whether the routine initiation conditions for the mode are met, the department's appointment queuing situation, and the remaining time of the target justification cycle. Inferences are made based on pre-defined rule logic. For example, if a patient has completed the pre-operative examination mode pre-processing and is waiting for the elective surgery mode core, and if the operating room schedule shows an opening within the cycle, it is presumed that the surgery will be performed within the cycle. Patients meeting this scenario are also included in the estimated patient count.

[0046] Furthermore, in steps S522 and S523, the judgment that is presumed to be of high probability of completion or high probability of entering and completing is verified by means of the execution status of medical orders in the hospital information system, the timestamp data of the scheduled surgery or examination, and the department's bed availability and scheduling plan. When the auxiliary verification information supports a certainty exceeding a preset certainty threshold within the target validation period, the patient is included in the estimated number of patients.

[0047] In this embodiment, the judgments in S522 and S523 that presume a high probability of completion or entry and completion are enhanced by the rigorousness of logical reasoning through access to auxiliary information flow from the hospital's real-time operations. During the presumption process, instead of relying solely on historical average durations, the system actively queries authoritative electronic information records directly related to the patient for cross-validation. These records include: electronic medical orders issued by doctors and their current execution status, precise timestamps of scheduled operating rooms or examination departments, and bed occupancy and turnover schedules published by relevant clinical departments. A certainty threshold is set. Only when the strength of positive evidence extracted from these auxiliary information sources supporting the patient's completion of the relevant treatment unit within the target validation period—such as a clear appointment time within the period or a scheduled medical order exceeding the preset certainty threshold—is the patient's status ultimately presumed as high probability, and they are officially included in the estimated patient count. This multi-source information cross-validation mechanism aims to reduce misjudgments based solely on statistical averages, making patient progress predictions closer to actual medical arrangements.

[0048] Furthermore, S6 also includes: S61, for each item in the final collaborative procurement forecast list, obtain its current available inventory C and the pre-set inventory safety threshold S from the hospital's material management system in real time. S62, calculate the difference Δ = D - C between the collaborative forecast demand D and the current available inventory C for this material; S63, determine whether the difference Δ is greater than the safety threshold S; if Δ>S, determine that the material needs to be included in the procurement demonstration scope and kept in the list; otherwise, remove the material from the preliminary collaborative procurement forecast list.

[0049] In this embodiment, a temporary comparison task is established for each item to be included in the preliminary collaborative procurement forecast list. A query request is initiated to the hospital's supplies management system in real-time or near real-time via a pre-configured data interface to obtain the current available inventory level C of that item. This inventory level typically refers to the actual quantity of good-quality inventory that is available in the warehouse and has not been pre-booked or locked. Simultaneously, the pre-set inventory safety threshold S for this type of item is read from the local supplies master data configuration table. This threshold is a minimum buffer inventory level set based on factors such as historical consumption fluctuations, procurement lead times, and supplier reliability.

[0050] For each item, perform a simple arithmetic operation: subtract the current available inventory C obtained from materials management from the collaborative forecast demand D calculated using S5, to obtain a theoretical immediate inventory gap Δ, i.e., Δ = D - C. This value represents the net gap between forecast demand and current inventory, without considering safety buffers.

[0051] The calculated theoretical immediate demand gap Δ is logically compared with the inventory safety threshold S. The judgment logic is as follows: if Δ > S, it means that even considering the reserved safety buffer inventory S, the predicted demand gap Δ will still consume this buffer and result in an additional shortage, indicating an urgent need for procurement. Therefore, the item is determined to need procurement and remains on the list. Conversely, if Δ ≤ S, it means that the predicted demand gap is within the range that the safety buffer inventory can cover, and there is no urgent risk of inventory depletion. Therefore, the item is removed from the current preliminary collaborative procurement forecast list. After this step, all items on the list are items whose predicted demand clearly exceeds the current inventory plus the safety buffer, allowing the focus of procurement justification to be concentrated on the real risk points.

[0052] Furthermore, referring to Figure 5 The S7 also includes: S71, for each item in the preliminary collaborative procurement forecast list, retrieve the collaborative forecast demand generated by this method in each of the past M historical demonstration periods, and the actual procurement quantity of the item after the end of the corresponding period. S72, calculate the prediction deviation rate ε_i for each historical period i = |actual purchase quantity - collaboratively predicted demand quantity| / collaboratively predicted demand quantity; S73, Based on the prediction deviation rate of the past M periods, calculate the average prediction deviation rate E and the deviation stability coefficient σ of the material; the deviation stability coefficient reflects the degree of dispersion of the deviation rate in each historical period. S74, Based on the predicted average deviation rate E and the deviation stability coefficient σ, the confidence weight factor ω is determined through a preset weight mapping rule; S75, multiply the collaborative forecast demand D of the material in the current demonstration period by the credibility weight factor ω to obtain the weighted collaborative forecast demand D' = D * ω, and record the value of ω.

[0053] In this embodiment, a historical data analysis routine is initiated for each item on the preliminary collaborative procurement forecast list. The historical verification database is accessed, and a search is performed according to the item code. Two sets of key data for the item over the past M consecutive historical verification cycles (e.g., M=6) are extracted: first, the original value of the collaborative forecast demand generated by method S5 in each cycle; second, the actual purchase quantity in the purchase order actually initiated by the hospital for the item after each cycle. The actual purchase quantity is considered an approximate reflection of the actual execution and confirmation of demand.

[0054] For each historical period i (i=1 to M), a forecast deviation rate ε_i is calculated. It is calculated as follows: take the absolute value of the difference between the actual purchase quantity and the collaboratively forecasted demand quantity within that period, and then divide it by the collaboratively forecasted demand quantity. That is, ε_i = |actual purchase quantity - collaboratively forecasted demand quantity| / collaboratively forecasted demand quantity. This ratio reflects the relative degree to which the forecast value deviates from the actual purchase quantity within that period.

[0055] Based on the prediction deviation rate sequence [ε_1, ε_2, …, ε_M] over the past M periods, two key statistical indicators are calculated. The first is the average prediction deviation rate E, which is the arithmetic mean of all ε_i, representing the overall average level of prediction deviation for this material. The second is the deviation stability coefficient σ, which represents the dispersion of these M deviation rates and can be obtained by calculating their standard deviation or coefficient of variation. A smaller σ value indicates smaller fluctuations in historical prediction deviations and higher stability; conversely, a larger σ value indicates larger fluctuations and lower stability.

[0056] Based on the calculated E and σ, the credibility weight factor ω is determined by querying a predefined credibility weight mapping table. This mapping table is set by management based on experience and risk preference. Its core mapping principle is: the lower the average prediction deviation rate E and the smaller the deviation stability coefficient σ, the higher the credibility weight factor ω is assigned. For example, materials with very low E and very small σ may receive ω=1.0 (fully credible) or higher; materials with moderate E but large σ may receive ω=0.8 (relatively credible), but fluctuations should be noted; materials with very high E regardless of σ may receive a lower ω value, such as 0.6, indicating poor historical prediction accuracy, and the current prediction should be treated with caution.

[0057] The original collaborative forecast demand D for the current evaluation period is revised using weighting factors. The revised calculation is: Weighted collaborative forecast demand D' = D × ω. Simultaneously, the weighting factor ω used in the calculation is recorded and associated with D'. The forecast demand for each item in the list is replaced by its corresponding D', with an accompanying ω value as a credibility annotation. Thus, the original preliminary collaborative procurement forecast list is updated into a final collaborative procurement forecast list with credibility weights. This list not only provides demand forecasts but also historical reliability metrics for each forecast value, offering managers a richer dimension for decision-making.

[0058] Furthermore, the method also includes a periodic calibration step, which is performed after each predetermined number of procurement justification cycles, including: S91, Collect the collaborative predicted demand for each material generated by this method in each of the most recently completed demonstration cycles of the predetermined number of times, as well as the actual consumption data of the materials in the corresponding cycles. S92, compare the predicted demand and actual consumption of the same material in different cycles to analyze the consistency of their changing trends and systematic deviations. S93, based on the analysis results of S92, propose revision suggestions for the relevant standard unit usage in the standard department material relationship mapping table, or propose adjustment suggestions for the first frequency threshold, second frequency threshold or dispersion threshold used to identify the fixed connection mode in S3, and submit the revision suggestions or adjustment suggestions for review. After the review is approved, they will be used to update the demonstration configuration for subsequent cycles.

[0059] In this embodiment, after the calibration cycle is initiated, all generated demonstration record data are collected within the most recently completed demonstration cycle, such as 12 cycles. The focus is on collecting two key data sequences for each predicted material in each demonstration cycle: first, the collaboratively predicted demand or weighted demand generated by method S5 / S7; and second, the actual consumption of materials reflected in the hospital's material management consumption and outbound records or financial settlement records after the demonstration cycle ends. This is a posterior indicator that more directly reflects the true clinical demand than the actual procurement quantity.

[0060] A thorough comparative analysis is conducted on a material-by-material basis. This involves a horizontal comparison of the predicted demand series and the actual consumption series for the same material over the past 12 periods. The analysis focuses on two aspects: first, trend consistency analysis, observing whether the changing trends of predicted demand, such as monthly increases, decreases, or seasonal fluctuations, largely match the changing trends of actual consumption; second, deviation analysis, calculating whether there is a persistent overestimation or underestimation of the predicted value relative to the actual consumption value throughout the entire calibration period, i.e., whether the deviation rate E significantly deviates from 0.

[0061] Based on the quantitative and qualitative analysis results of S92, specific calibration recommendation reports are generated. These recommendations may involve revisions to two core knowledge bases: First, if the predictions for certain supplies consistently deviate, and this deviation can be traced back to consumption differences within a specific departmental treatment unit, a suggestion is made to revise the standard unit usage value of that supply in the standard departmental supply relationship mapping table for that unit, accompanied by historical data as evidence. Second, if the identification results for certain fixed connection patterns are found to be unstable or their predictive contribution low, a suggestion may be made to adjust the threshold parameters used for pattern identification in S3, such as the first frequency threshold, the second frequency threshold, or the dispersion threshold, to make pattern recognition more consistent with actual, stable medical collaboration relationships. These calibration recommendation reports do not take effect automatically but are submitted as important knowledge assets to the hospital's relevant management department, clinical expert committee, and supplies management department for joint review. After approval, authorized administrators update the corresponding configuration parameters or knowledge base content, ensuring that the next stage of procurement task justification is based on a more accurate and realistic foundation.

[0062] According to a second embodiment of the present invention, the present invention claims protection for a procurement task justification system applicable to hospital centralized management departments, comprising: One or more processors; A memory storing one or more programs, which, when executed by one or more processors, enable the processors to implement a procurement task justification method applicable to hospital management departments.

[0063] The following is a specific example: On the 25th of a certain month, the hospital's procurement task verification system received a procurement verification instruction from the relevant management department for the target verification cycle of the following month. The system responded immediately, retrieving all treatment records of inpatients from the 1st to the 25th of that month from the hospital's core database, and integrating these records with the pre-hospitalization plans for the following month. The data was anonymized, and each record included an anonymous patient ID, the departmental treatment units the patient had completed (e.g., Orthopedics Outpatient - Initial Consultation, Radiology - Knee MRI, Anesthesiology - Preoperative Assessment, Orthopedics - Knee Replacement Surgery List and Dates), and the patient's current departmental unit (e.g., Orthopedics Ward - Postoperative Observation).

[0064] Using this data, the system constructs a chronologically ordered treatment pathway sequence for each patient, marking the currently ongoing unit at the end of the sequence. All patient sequences constitute the initial analysis set. The system performs in-depth analysis of this set, automatically identifying frequently occurring and stable combinations of departmental unit sequences. For example, the system discovers that after the orthopedics-knee replacement surgery unit, there is a very high probability, and usually within a specific time window, that it connects to the rehabilitation department-early postoperative rehabilitation unit. This combination is identified as a fixed connection pattern called "post-knee replacement rehabilitation." The system then queries the hospital's standard knowledge base, namely the standard departmental material relationship mapping table, to determine the material sets associated with this pattern, such as specific artificial joints, bone cement, and special sutures used in surgery, as well as pressure cryotherapy devices and specific models of rehabilitation braces used in the rehabilitation phase.

[0065] Subsequently, the system performs collaborative prediction on this set of resources. It first identifies patient sequences across all treatment pathways that include post-knee replacement rehabilitation. Then, it analyzes the treatment progress of these patients within the target validation period for the following month: for patients who have completed surgery but are still hospitalized and are expected to be transferred to the rehabilitation department next month, and for patients who have scheduled surgery next month and will inevitably enter the rehabilitation department post-surgery, the system logically infers the probability of them completing the full surgery + rehabilitation program or a key component next month based on their current status, standard treatment cycle, and existing scheduling information, and includes high-probability patients in the estimated number. Assuming the estimated number of patients is N, the system then calculates the collaborative predicted demand Di for each resource based on the single-unit usage of materials such as artificial joints in the surgical unit, and the average daily usage and average number of rehabilitation days of the pressurized cryotherapy device in the rehabilitation unit, according to the standard mapping table.

[0066] The system then compares these projected demand quantities with the real-time inventory levels of the corresponding materials in the materials management system, as well as the set safety stock thresholds. For materials whose projected demand minus the current inventory still exceeds the safety threshold, the system includes them in the preliminary procurement forecast list. The system further retrieves historical forecast and final actual procurement data for these materials over the past six months, calculates patterns of historical deviations, and assigns a reliability weight to the current forecast value: a weight close to 1 is assigned for small and stable historical deviations, while a weight is lowered for large historical fluctuations. Finally, the system generates a weighted final procurement forecast list and automatically compiles a detailed justification report. The report not only lists the recommended procurement materials and their weighted projected quantities but also elaborates on the basis for each recommendation, including the fixed connection mode relied upon, the calculation process for the estimated number of patients, the source of the standard unit dosage used, and the reasons for the weighting of historical deviations. This report is automatically pushed to the procurement management department's work platform through the hospital's OA system.

[0067] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between apparatuses or units, and may be electrical, mechanical, or other forms.

[0068] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated units described above can be implemented in hardware or as software functional units. The above are merely embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made based on the description and drawings of this application, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

[0069] The specific embodiments of the invention have been described in detail above, but they are only examples, and this application is not limited to the specific embodiments described above. For those skilled in the art, any equivalent modifications or substitutions to the invention are also within the scope of this application. Therefore, all equivalent changes, modifications, and improvements made without departing from the spirit and principles of this application should be covered within the scope of this application.

Claims

1. A method for justifying procurement tasks applicable to hospital management departments, characterized in that, The method includes the following steps: S1, in response to the procurement justification instruction, obtains treatment pathway data for all patients in the hospital within the target justification period; S2, Based on the treatment path data, construct a set of treatment path sequences for the patient group; S3, parse the set of treatment path sequences, and identify a fixed connection pattern in all treatment path sequences that consists of at least two treatment units from different departments. In the course of historical patient treatment, after completing the treatment unit of the previous department, the patient has a predetermined probability of entering the next department for treatment unit connection. S4, Based on the hospital's pre-set standard departmental material relationship mapping table, determine the set of medical supplies associated with all departmental treatment units in the fixed connection mode; S5. Based on all treatment pathway sequences containing the fixed connection pattern associated with the material set, the estimated number of patients expected to generate demand for various materials in the material set through the fixed connection pattern within the target demonstration period is calculated. Based on the estimated number of patients and the standard unit usage in the standard department material relationship mapping table, the collaborative predicted demand for each type of material in the material set before the end of the target demonstration period is calculated. S6. Compare all the collaboratively predicted demand quantities with the current inventory and safety threshold of the materials in the hospital material management system, filter out the materials whose difference between the predicted demand quantity and the current inventory exceeds the safety threshold, and generate a preliminary collaborative procurement prediction list. S7. Retrieve the deviation records between the actual purchase quantity and the collaborative forecast demand quantity of each material in the preliminary collaborative procurement forecast list within the historical demonstration period of the predetermined quantity in the past. Based on the deviation records, perform credibility weighting on the collaborative forecast demand quantity of the current demonstration period to generate a final collaborative procurement forecast list with credibility weights.

2. The method according to claim 1, characterized in that, S2 also includes: S21, standardize the treatment path data of each patient obtained by S1 to ensure that the identification of each department's treatment unit conforms to the hospital's unified coding standard; S22, sort the standardized departmental treatment unit sequence for each patient according to the recorded occurrence time or logical order to form the patient's initial treatment path sequence; S23 marks the status identifier for each departmental treatment unit node in the initial treatment path sequence for each patient. S24, gather the treatment path sequences of all patients' marked status identifiers to form the treatment path sequence set.

3. The method according to claim 2, characterized in that, S3 also includes: S31, Traverse the set of treatment path sequences and extract all sub-sequences in each treatment path sequence that consist of at least two different departmental treatment unit identifiers; S32, establish a candidate connection pattern set. For each subsequence, convert it into a standard pattern expression, arrange its identifiers according to the order of appearance of the department treatment unit in the subsequence, and add the standard pattern expression as a candidate pattern to the candidate connection pattern set. S33, For each candidate pattern in the candidate connection pattern set, perform global frequency statistics in the treatment path sequence set; S34, set a first frequency threshold and a second frequency threshold. The first frequency threshold is used to determine whether the total frequency of the pattern is significant. The second frequency threshold is used to determine the frequency of the same pattern repeating in a single patient sequence. S35, for each candidate pattern in the candidate connection pattern set, determine whether its total frequency of the pattern is greater than the first frequency threshold, and whether the quotient obtained by dividing the total number of times the pattern appears by the total frequency of the pattern is less than the second frequency threshold. If both conditions are met, then the candidate pattern is determined to be a high-frequency and relatively discrete pattern. S36. For each patient treatment path sequence containing the candidate pattern, record the time interval or sequence interval between the treatment unit nodes of each department contained in the candidate pattern each time it appears in the sequence, and calculate the statistical dispersion of the interval between treatment unit nodes of the same sequence position in all instances of the candidate pattern. S37, Set a dispersion threshold, compare the calculated statistical dispersion with the dispersion threshold, and if the statistical dispersion is lower than the dispersion threshold, determine that the candidate pattern has stability in the time or order dimension. S38. Candidate patterns that simultaneously meet the judgment conditions of S35 and S37 are formally identified as the fixed connection patterns, and their standard pattern expression, total pattern frequency, and average occurrence interval dispersion are recorded as pattern feature information.

4. The method according to claim 1, characterized in that, S5 also includes: S51, for the set of medical supplies currently being judged, determine the fixed connection patterns associated with them, and obtain a subset of all treatment path sequences associated with these fixed connection patterns; S52, for each patient treatment path sequence in the subset of treatment path sequences, analyze its current status and expected progress within the remaining time of the target demonstration period: S53, cumulatively all patients identified in S52 as included in the estimated number of patients, to obtain the estimated number of patients N for the current medical supplies set and its associated fixed connection mode; S54. Based on the standard departmental material relationship mapping table, obtain the standard unit dosage of each category of material in the current medical material set under each departmental treatment unit involved in its associated fixed connection mode. S55, For the i-th type of supplies in the current medical supplies set, calculate its collaborative predicted demand Di: Di ​​= N * Σ (standard unit usage under each related department treatment unit); S56, repeat S51 to S55 until the collaborative procurement forecast and determination for all medical supplies is completed.

5. The method according to claim 1, characterized in that, S52 also includes: S521, If ​​the current state of the patient sequence is that it has fully experienced all the treatment units of the fixed connection pattern it is associated with, and is not currently in any treatment unit of the pattern, then no expected demand based on the pattern will be generated in the current demonstration period, and the patient will not be included in the estimated number of patients. S522, if the current state of the patient sequence is that it is in the intermediate department treatment unit of a certain fixed connection pattern associated with it, then determine the probability that the patient is expected to complete the subsequent department treatment unit of the fixed connection pattern before the end of the target demonstration period; this probability is logically inferred based on the historical data of the average treatment course of this type of treatment in the hospital; if it is inferred that the probability of completion is high, then the patient is included in the estimated number of patients. S523, if the current state of the patient sequence is that the preceding part of the treatment unit of a certain fixed connection mode has been completed, but the subsequent key treatment unit of the mode has not yet been entered, then based on the patient's completed treatment progress, routine medical pathway planning and the remaining time of the target demonstration cycle, the probability of the patient entering and completing the fixed connection mode before the end of the cycle is comprehensively estimated. If the probability is estimated to be high, then the patient is included in the estimated number of patients.

6. The method according to claim 5, characterized in that, In steps S522 and S523, the judgment that is presumed to be of high probability of completion or high probability of entry and completion is verified by means of the execution status of medical orders in the hospital information system, the timestamp data of the scheduled surgery or examination, and the department's bed availability and scheduling plan. When the auxiliary verification information supports a certainty exceeding a preset certainty threshold within the target validation period, the patient is included in the estimated number of patients.

7. The method according to claim 1, characterized in that, S6 also includes: S61, for each item in the final collaborative procurement forecast list, obtain its current available inventory C and the pre-set inventory safety threshold S from the hospital's material management system in real time. S62, calculate the difference Δ = D - C between the collaborative forecast demand D and the current available inventory C for this material; S63, determine whether the difference Δ is greater than the safety threshold S; if Δ > S, determine that the material needs to be included in the procurement demonstration scope and kept in the list; otherwise, remove the material from the preliminary collaborative procurement forecast list.

8. The method according to claim 1, characterized in that, S7 also includes: S71, for each item in the preliminary collaborative procurement forecast list, retrieve the collaborative forecast demand generated by this method in each of the past M historical demonstration periods, and the actual procurement quantity of the item after the end of the corresponding period. S72, calculate the prediction deviation rate ε_i for each historical period i = |actual purchase quantity - collaboratively predicted demand quantity| / collaboratively predicted demand quantity; S73, Based on the prediction deviation rate of the past M periods, calculate the average prediction deviation rate E and the deviation stability coefficient σ of the material; the deviation stability coefficient reflects the degree of dispersion of the deviation rate in each historical period. S74, Based on the predicted average deviation rate E and the deviation stability coefficient σ, the confidence weight factor ω is determined through a preset weight mapping rule; S75, multiply the collaborative forecast demand D of the material in the current demonstration period by the credibility weight factor ω to obtain the weighted collaborative forecast demand D' = D * ω, and record the value of ω.

9. The method according to claim 1, characterized in that, The method also includes a periodic calibration step, which is performed after each predetermined number of procurement validation cycles, including: S91, Collect the collaborative predicted demand for each material generated by this method in each of the most recently completed demonstration cycles of the predetermined number of times, as well as the actual consumption data of the materials in the corresponding cycles. S92, compare the predicted demand and actual consumption of the same material in different cycles to analyze the consistency of their changing trends and systematic deviations. S93, based on the analysis results of S92, propose revision suggestions for the relevant standard unit usage in the standard department material relationship mapping table, or propose adjustment suggestions for the first frequency threshold, second frequency threshold or dispersion threshold used to identify the fixed connection mode in S3, and submit the revision suggestions or adjustment suggestions for review. After the review is approved, they will be used to update the demonstration configuration for subsequent cycles.

10. A procurement task justification system applicable to hospital centralized management departments, characterized in that, include: One or more processors; A memory having stored one or more programs that, when executed by one or more processors, cause the one or more processors to implement a procurement task justification method applicable to hospital centralized management departments according to any one of claims 1 to 9.

Images