Medical service item cost prediction method and device based on key factors

Abstract

The present invention discloses a method and device for predicting the cost of medical service items based on key factors. The method includes: processing the basic information of medical service items based on the canonical correlation analysis algorithm to obtain the key factors of medical service items; inputting the key factors of medical service items into a pre-trained deep neural network model to determine the attribution of medical service items; predicting the cost of medical service items in each affiliated department according to the attribution of medical service items; and predicting the cost of medical service items in the hospital based on the cost of medical service items in each affiliated department. Through the canonical correlation analysis algorithm and the deep neural network model, while greatly improving the processing speed of relevant information, it can simplify the process of predicting the cost of medical service items. By predicting the cost of medical service items in each affiliated department and then predicting the cost of medical service items in the hospital, using a step-by-step prediction method is conducive to improving the accuracy of the prediction results.

Description

Technical Field

[0001] The present invention relates to the technical field of project cost prediction, and in particular, to a method and device for predicting the cost of medical service projects based on key factors. Background Art

[0002] The cost prediction results of medical service projects can provide favorable data support and decision-making basis for the dynamic adjustment of medical service prices, the formulation of financial compensation policies for public hospitals, the reform of composite medical insurance payment methods, or the determination of insurance rates by medical insurance institutions.

[0003] In the process of predicting the cost of medical service projects, accurately predicting the cost of medical service projects in each department is the basis for predicting the cost of medical service projects in the entire hospital. At present, the main methods for predicting the cost of medical service projects in each department are the income coefficient method and the activity-based costing method. Under the income coefficient method, unreasonable pricing of medical service projects will lead to inaccurate calculation results of medical service project costs. The activity-based costing method takes the activity volume as the cost allocation basis and treats direct costs and indirect costs (including period costs) equally as the costs consumed by products for activities, broadening the scope of cost calculation, but the activity division is complex and the practical operation is difficult.

[0004] Currently, the accounting of medical service project costs is mainly carried out by professionals using the above methods for measurement and analysis, which has problems such as high prediction thresholds, cumbersome processes, being affected by human uncertain factors, and poor prediction accuracy. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art.

[0006] For this purpose, the present invention proposes a method and device for predicting the cost of medical service projects based on key factors, and solves the technical problems existing in the cost prediction of medical service projects in the prior art through the combined application of the canonical correlation analysis algorithm and the deep neural network model.

[0007] According to the first aspect of the present application, there is provided a method for predicting the cost of medical service projects, the method comprising:

[0008] Processing the basic information of the medical service project based on the canonical correlation analysis algorithm to obtain the key factors of the medical service project;

[0009] Inputting the key factors of the medical service project into a pre-trained deep neural network model to determine the attribution of the medical service project;

[0010] Predicting the cost of medical service projects in each attribution department according to the attribution of the medical service project;

[0011] Predict the medical service item costs of a hospital based on the medical service item costs of each of the affiliated departments.

[0012] In the above method, the step of processing the basic information of the medical service item based on the canonical correlation analysis algorithm to obtain the key factors of the medical service item includes: comparing the basic information of the medical service item with the relevant information in the standard library to screen out the key factors of the medical service item.

[0013] In the above method, before the step of inputting the key factors of the medical service item into a pre-trained deep neural network model to determine the affiliation of the medical service item, the method further includes: obtaining training samples; performing feature extraction on the training samples to obtain the sequence features corresponding to the training samples; and establishing and training a model according to the sequence features corresponding to the training samples to obtain the deep neural network model.

[0014] In the above method, the step of predicting the medical service item costs of each affiliated department according to the affiliation of the medical service item includes: de-duplicating the relevant data of the medical service item according to a preset de-duplication rule to avoid repeated prediction of the costs of the medical service item; and screening the relevant data of the medical service item after de-duplication according to a preset filtering rule to retain the data participating in the prediction of the medical service item costs.

[0015] In the above method, the step of predicting the medical service item costs of each affiliated department according to the affiliation of the medical service item further includes: checking whether the cost prediction of the medical service item of each affiliated department is abnormal according to a preset judgment rule; if so, giving an abnormal prompt and performing a retrospective inspection on the data of the cost prediction of the medical service item according to the abnormal prompt; if not, giving a normal prompt.

[0016] In the above method, the step of checking whether the cost prediction of the medical service item of each affiliated department is abnormal according to a preset judgment rule includes: checking whether the ratio of the income to the cost of the medical service item of each affiliated department is within the range of a first set threshold and a second set threshold, where the first set threshold is less than the second set threshold; if the ratio of the income to the cost of the medical service item of each affiliated department is greater than or equal to the first set threshold and less than or equal to the second set threshold, the cost prediction of the medical service item of each affiliated department is normal.

[0017] According to a second aspect of the present application, there is provided a cost prediction device, which includes:

[0018] A key factor acquisition module, which is used to process the basic information of medical service items based on the canonical correlation analysis algorithm to obtain the key factors of the medical service items;

[0019] An affiliated department determination module, which is used to input the key factors of the medical service items into a pre-trained deep neural network model to determine the affiliation of the medical service items;

[0020] An affiliated department cost prediction module, which is used to predict the medical service item costs of each affiliated department according to the affiliation of the medical service items;

[0021] A hospital cost prediction module, which is used to predict the medical service item costs of the hospital based on the medical service item costs of each affiliated department.

[0022] In the above device, the affiliated department cost prediction module includes:

[0023] A duplicate removal unit, which is used to remove duplicates from the relevant data of the medical service items according to a preset duplicate removal rule to avoid predicting the cost of the medical service items repeatedly;

[0024] A filtering unit, which is used to screen the relevant data of the medical service items after duplicate removal according to a preset filtering rule to retain the data participating in the prediction of the medical service item costs;

[0025] An inspection unit, which is used to check whether the prediction of the medical service item costs of each affiliated department is abnormal according to a preset judgment rule; if so, an abnormal prompt is given, and the data for predicting the medical service item costs is traced and inspected according to the abnormal prompt; if not, a normal prompt is given.

[0026] According to the third aspect of the present application, a terminal is provided. The terminal includes a memory and a processor. A computer program that can run on the processor is stored on the memory. When the processor runs the computer program, it executes the medical service item cost prediction method described in any one of the above.

[0027] According to the fourth aspect of the present application, a computer-readable storage medium is provided. The computer-readable storage medium includes a stored computer program. When the computer program is run by a processor, it controls the terminal where the storage medium is located to execute the medical service item cost prediction method described in any one of the above.

[0028] According to the technical solution provided by this application, it has at least the following beneficial effects: By processing the basic information of medical service items based on the canonical correlation analysis algorithm, the key factors corresponding to the medical service items can be quickly obtained. Compared with manual screening, processing the basic information based on the canonical correlation analysis algorithm greatly improves the processing speed of relevant information and is conducive to improving work efficiency. Inputting the key factors into a pre-trained deep neural network model to determine the attribution of the medical service items, the deep neural network model can introduce more key factor features on the input side, which is beneficial to simplifying the medical service item cost prediction process while making the attribution of the medical service items more accurate. Predicting the medical service item costs of each affiliated department according to the attribution of the medical service items, and then predicting the medical service item costs of the hospital based on the medical service item costs of each affiliated department. Adopting a step-by-step prediction method is conducive to improving the accuracy of the prediction results.

[0029] Other features and advantages of this application will be described in the subsequent specification, and part of them will become obvious from the specification or be understood by implementing this application. The purpose and other advantages of this application can be realized and obtained through the structure specifically pointed out in the specification, claims, and drawings. Brief Description of the Drawings

[0030] The above and / or additional aspects and advantages of the present invention will become obvious and easy to understand from the description of the embodiments in conjunction with the following drawings, where:

[0031] Figure 1 It is a flowchart of the medical service item cost prediction method provided by the embodiment of this application;

[0032] Figure 2 It is a partial list in the standard library provided by the embodiment of this application;

[0033] Figure 3 It is the network structure diagram of the deep neural network model provided by the embodiment of this application;

[0034] Figure 4 It is the algorithm framework diagram provided by the embodiment of this application;

[0035] Figure 5 It is the structural block diagram of the cost prediction device provided by the embodiment of this application. Detailed Description of the Embodiment

[0036] In order to make the purpose, technical solution, and advantages of this application clearer, the following further details this application in conjunction with the drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain this application and are not used to limit this application.

[0037] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, not to describe a specific order or sequence.

[0038] The solution shown in the embodiment of the present invention can be applied to various software development platform devices. For example, the solution can be applied to electronic devices based on personal computers or workstations running software development tools.

[0039] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0040] like Figure 1 As shown, an embodiment of the present application provides a method for predicting the cost of medical service items based on key factors, which can be performed by a cost prediction device, which can be implemented in software and / or hardware. The method for predicting the cost of medical service items includes: Step S101 to Step S104.

[0041] Step S101: Processing basic information of medical service items based on a typical association analysis algorithm to obtain key factors of the medical service items.

[0042] In this application, based on the typical association analysis algorithm, the basic information of the medical service items provided by the hospital is compared and screened with the relevant information in the standard library, so that the key factors corresponding to the medical service items can be quickly obtained. Compared with manual screening, the screening speed is greatly improved, which is conducive to improving work efficiency.

[0043] The basic information includes but is not limited to the project code, project name, project connotation and pricing unit of the medical service projects carried out by the hospital. It should be noted that the project connotation refers to the facilities (special environments, such as hyperbaric oxygen chambers, laminar flow purification rooms, etc.), equipment, and descriptions of the operation process, main steps and technical service content that must be used in the operation of the medical service price project. Its role is to help the pricing management department understand the facilities and equipment that must be used for the medical service and the main operation process, as well as the complexity and difficulty of the technical operation, which is conducive to accurate cost verification.

[0044] The standard library refers to the national medical service price item standard library, such as Figure 2The partial list shown above, the National Medical Service Price Item Specification Library stipulates relevant information such as the item code, item name, item connotation, pricing unit, human consumption and time consumption, low-value consumable consumption quota level, technical difficulty, and risk level of medical service item standards. Compare the basic information of the medical service items carried out by the hospital with the relevant information in the standard library. Specifically, the item connotation in the basic information can be compared with the item connotation of the medical service items in the standard library. If the similarity between the two is relatively high, it means that the medical service items carried out by the hospital can be fully matched with the medical service items in the standard library, and the key factors corresponding to the item connotation of the medical service items in the standard library can be directly used as the key factors of the medical service items carried out by the hospital. It should be noted that the key factors of medical service items include, but are not limited to, item code, item name, item connotation, pricing unit, human consumption and time consumption, low-value consumable consumption quota level, technical difficulty, and risk level. In this application, the attribution of medical service items can be determined by the item code, item name, item connotation, and pricing unit in the key factors.

[0045] There are various algorithms for mining information correlation relationships in the prior art, such as the Apriori algorithm, FreeSpan algorithm, prefixspan algorithm, and typical correlation analysis algorithm, etc. In this application, the typical correlation analysis algorithm (Canonical Correlation Analysis, abbreviated as CCA) is specifically selected to mine the correlation relationship between the basic information of medical service items and the relevant information in the standard library to obtain the key factors corresponding to the medical service items.

[0046] Specifically, the basic information of medical service items is represented by X, and the relevant information in the standard library is represented by Y. In this application, both the basic information X and the relevant information Y in the standard library can be used as multi-dimensional arrays. The correlation coefficient ρ(X, Y) of X and Y is calculated using the typical correlation analysis algorithm. According to the correlation coefficient ρ(X, Y), the correlation degree between the basic information X and the relevant information Y in the standard library can be judged, and then the key factors corresponding to the medical service items can be screened out.

[0047] In this application, the input quantities of this algorithm are X and Y, the dimensions of X and Y are both greater than 1, and the output quantity is the correlation coefficient ρ(X, Y). The calculation formula is as follows:

[0048]

[0049] In the formula, S XX represents the variance of X, S YY represents the variance of Y, S XYDenote the covariance of X and Y. Perform singular value decomposition on matrix M, and the largest singular value obtained is the correlation coefficient ρ(X, Y). The value range of the correlation coefficient ρ(X, Y) is [-1, 1]. The closer the absolute value of ρ(X, Y) is to 1, the higher the linear correlation between X and Y; the closer the absolute value of ρ(X, Y) is to 0, the lower the linear correlation between X and Y.

[0050] In this application, before processing the basic information of medical service items based on the canonical correlation analysis algorithm to obtain the key factors of medical service items, this algorithm can be used to dynamically collect basic information in real time from business systems such as hospital information systems, material management systems such as drug consumables, fixed asset management systems, and logistics management systems.

[0051] In this application, the canonical correlation analysis algorithm can be used to compare the basic information of medical service items with the relevant information in the standard library to screen out the key factors corresponding to the medical service items, and the basic information of the medical service items can also be transmitted to a third-party software through an interface for screening.

[0052] Step S102: Input the key factors of the medical service item into a pre-trained deep neural network model to determine the attribution of the medical service item.

[0053] It should be noted that there are a wide variety of medical service items. The implementation of some medical projects may span multiple departments, making it more difficult to predict the costs of medical projects in the entire hospital. Determining the accurate department attribution of the medical service item based on the key factors is beneficial to reducing the difficulty of cost prediction caused by medical projects spanning multiple departments, and thus beneficial to simplifying the process of cost prediction for medical projects in the entire hospital and improving the efficiency of cost prediction for medical projects.

[0054] In this application, determining the attribution of a medical service item refers to determining which department the medical service item belongs to. For example, it belongs to the prescribing department, the executing department, or the exclusive department.

[0055] In this application, the attribution rules of medical service items are set in the pre-trained deep neural network model, and the attribution rules include the attribution rules of the prescribing department, the attribution rules of the executing department, and the attribution rules of the exclusive department.

[0056] For example, medical service items in the comprehensive medical service category, including medical service items such as examination, bed, nursing, injection, debridement, dressing change, etc., belong to the prescribing department; medical service items in the medical technology diagnosis and treatment category, including medical imaging, X-ray contrast, magnetic resonance, CT, ultrasound, nuclear medicine, laboratory tests, pathology, etc., belong to the performing department; medical service items in the clinical diagnosis and treatment category and traditional Chinese medicine diagnosis and treatment category belong to the performing department; medical service items that can only be carried out in a single department, such as a hospital where only a single department conducts magnetic resonance imaging projects, then such projects belong to this single exclusive department. The coding of these different types of medical service items has general rules.

[0057] It should be noted that a deep neural network model (Deep Neural Networks, abbreviated as DNN) can be understood as a neural network with many hidden layers. Divided according to the positions of different layers in the deep neural network model, as Figure 3 shown, the neural network layers inside the deep neural network model can be divided into three categories: the input layer, the hidden layers, and the output layer. Generally speaking, the first layer is the input layer, the last layer is the output layer, and the middle layers are all hidden layers. The layers are fully connected to each other. That is, any neuron in the i-th layer must be connected to any neuron in the i + 1-th layer. In this application, the input of the deep neural network model is the item code, item name, item connotation, and pricing unit among the key factors of medical service items, and the output is the affiliated department of the medical service item. By adopting the network structure of the deep neural network model, more key factor features of medical service items can be introduced on the input side, making the attribution of medical service items more accurate.

[0058] In this application, after determining the affiliated department of the medical service item according to the item code, item name, item connotation, and pricing unit among the key factors, then determine the personnel expenses for performing the medical project according to the labor consumption, time consumption, risk level, and technical difficulty among the key factors, and determine the non-separately charged health material costs, fixed asset depreciation expenses, intangible asset amortization expenses, and other operating expenses corresponding to the medical project according to the workload, project time consumption, and low-value consumable consumption quota grade among the key factors, and determine the medical risk fund to be accrued according to the key factor of the risk level.

[0059] It should be noted that the above-mentioned expenses can be obtained through different deep neural network models, and different output quantities can be obtained by adjusting the input quantity of the deep neural network model and related model training parameters.

[0060] Step S103: Predict the medical service item costs of each affiliated department according to the attribution of the medical service items.

[0061] Specifically, after determining the affiliated department of the medical service item based on step S102, it is also necessary to deduplicate the relevant data of the medical service item according to the preset deduplication rules through the Simhash algorithm to avoid the cost of predicting the medical service item repeatedly. Moreover, the relevant data of the medical service item after deduplication is screened according to the preset filtering rules to filter out the data that does not participate in the cost prediction of the medical service item and retain the data that participates in the cost prediction of the medical service item.

[0062] It should be noted that the preset deduplication rules include but are not limited to the following two situations: One is to delete the item data of the execution department with repeated affiliation in the medical service items affiliated by the issuing department. That is, if a certain medical service item is affiliated to the issuing department, the item data that is repeatedly affiliated to other execution departments in this medical service item can be automatically judged and excluded according to the preset deduplication rules; The other is to delete the item data of the issuing department with repeated affiliation in the medical service items affiliated by the execution department. That is, if a certain medical service item is affiliated to the execution department, the item data that is repeatedly affiliated to other issuing departments in this medical service item can be automatically judged and excluded according to the preset deduplication rules.

[0063] The preset filtering rules include but are not limited to the following three situations: The first is to filter out the data with zero workload of the medical service item. For example, the workload of a single medical service item is zero, or the aggregated workload of the medical service item is zero; The second is to filter out the data of the refunded medical service item. For example, the refunded medical service item in the hospital information system, that is, the medical service item that has not been actually executed, does not participate in the cost prediction of the medical service item, and the relevant data of this type of medical service item is excluded; The third is to filter out the data that does not belong to the issuing department, execution department or exclusive department in the medical service item. For example, the medical service item that does not belong to the issuing department, execution department or exclusive department does not participate in the cost prediction of the medical service item, and this type of data is excluded.

[0064] In actual application, the preset deduplication rules and filtering rules can be adjusted according to actual needs and are not specifically limited in this application.

[0065] In this application, the Simhash algorithm is used to achieve the deduplication and filtering of relevant data. The comparison speed of the Simhash algorithm is relatively fast, and in the tasks of deduplicating and filtering massive texts, it can greatly improve the speed of deduplication and filtering.

[0066] Furthermore, the medical service item costs of each department include direct costs and indirect costs. In this application, direct costs refer to various expenses that can be directly included in or calculated by a certain method and then directly included in the department costs when the department conducts medical service activities. For example, the personnel expenses of each department, the cost of non-separately charged medical materials, the depreciation of fixed assets, the amortization of intangible assets, the provision for medical risk funds, and other operating expenses. The relevant data of direct costs can be applied to the above deduplication rules and filtering rules. Indirect costs refer to various costs shared by clinical and medical technology departments that carry out medical service items from management and auxiliary departments.

[0067] In this application, after deduplicating and filtering and screening the data of the direct costs of medical service items whose attribution departments are determined, and then adding the corresponding indirect costs, the cost prediction of medical service items can be carried out for each attribution department.

[0068] In this application, after predicting the costs of medical service items for each attribution department, it is also possible to check whether the cost prediction of medical service items for each attribution department is abnormal based on the Simhash algorithm according to the preset judgment rules; if so, an abnormal prompt is given, and the data of the cost prediction of medical service items is traced and audited according to this abnormal prompt; if not, a normal prompt is given.

[0069] Furthermore, the preset judgment rules include: whether the ratio of the income of medical service items of each attribution department to the cost of medical service items of each attribution department is within the range of the first set threshold and the second set threshold, where the first set threshold is less than the second set threshold.

[0070] If the ratio of the income of medical service items of each attribution department to the cost of medical service items of each attribution department is less than the first set threshold or greater than the second set threshold, it means that the cost prediction of medical service items of each department is abnormal; if the ratio of the income of medical service items of each attribution department to the cost of medical service items of each attribution department is greater than or equal to the first set threshold and less than or equal to the second set threshold, it means that the cost prediction of medical service items of each department is normal.

[0071] In actual application, the numerical values of the first set threshold and the second set threshold can be adjusted according to the actual situation, and no specific limitation is made in this application.

[0072] Step S104: Predict the cost of medical service items of the hospital based on the cost of medical service items of each attribution department.

[0073] Specifically, based on the cost of medical service items of each attribution department obtained in step S103, the cost of medical service items of each attribution department is weighted and summed using the summation algorithm to obtain the cost of medical service items of the hospital.

[0074] Combined with Figure 4 Shown in the algorithm framework diagram, for the medical service item cost prediction method in this application, first, based on the canonical correlation analysis algorithm, the correlation relationship between the basic information of the medical service item and the relevant information in the standard library is mined, and the basic information is compared with the relevant information in the standard library to quickly obtain the key factors corresponding to the medical service item. Compared with manual screening, processing the basic information based on the canonical correlation analysis algorithm greatly improves the processing speed of relevant information and is beneficial to improving work efficiency. Secondly, the key factors are input into a pre-trained deep neural network model to determine the attribution of the medical service item. The deep neural network model can introduce more key factor features on the input side, which is beneficial to simplifying the medical service item cost prediction process while making the attribution of the medical service item more accurate. Thirdly, after determining the affiliated department of the medical service item, the Simhash algorithm is used to quickly deduplicate and filter the relevant data of the medical service item, which can effectively retain the data participating in the medical service item cost prediction while avoiding repeated prediction of the medical service item cost to ensure the accuracy of the medical service item cost prediction for each affiliated department. Finally, the summation algorithm is used to predict the medical service item cost of the hospital based on the medical service item cost predictions of each affiliated department, adopting a step-by-step prediction method to improve the accuracy of the prediction result.

[0075] In some embodiments, before step S102, the method further includes: steps S1021 to S1023 (not shown in the figure).

[0076] Step S1021: Obtain training samples.

[0077] Before predicting the medical service item cost, in order to build a deep neural network model, a large number of training samples are required as the training basis of the deep neural network model. Only by continuously correcting through obtaining a large number of training samples can a relatively accurate deep neural network model be obtained, and then the attribution of the medical service item can be determined more accurately.

[0078] In this embodiment, the training samples are stored locally in advance and can be obtained from the local medical service item file.

[0079] As mentioned above, the relevant information of the medical service item is stored in the medical service item file, and the relevant information includes but is not limited to the basic information and key factors of the medical service item.

[0080] Step S1022: Extract features from the training samples to obtain the sequence features corresponding to the training samples.

[0081] The obtained training samples are massive. Correspondingly, the sequence features correspond to the training samples, that is, after feature extraction, the sequence features corresponding to each group of training samples are obtained.

[0082] Step S1023: Establish and train a model based on the sequence features corresponding to the training samples to obtain a deep neural network model.

[0083] Specifically, using the sequence features corresponding to the training samples as input for model establishment and training, a deep neural network model reflecting the feature distribution of medical service items can be obtained.

[0084] The pre - construction of the deep neural network model provides the support of mathematical algorithms for the subsequent attribution of medical service items, which is beneficial to improving the accuracy of medical service item attribution.

[0085] In addition, the construction of the deep neural network model is based on a large number of real training samples, and real training samples are the premise for determining the attribution of medical service items.

[0086] The following is the device embodiment of the present application, which can be used to execute the method embodiment of the present application. For details not disclosed in the device embodiment of the present application, please refer to the method embodiment of the present invention.

[0087] As Figure 5 shown, an embodiment of the present application provides a cost prediction device, which includes a key factor acquisition module 41, an attribution department determination module 42, an attribution department cost prediction module 43, and a hospital cost prediction module 44.

[0088] Among them,

[0089] The key factor acquisition module 41 is used to process the basic information of medical service items based on the canonical correlation analysis algorithm to obtain the key factors of medical service items;

[0090] The attribution department determination module 42 is used to input the key factors of medical service items into a pre - trained deep neural network model to determine the attribution of medical service items;

[0091] The attribution department cost prediction module 43 is used to predict the costs of medical service items of each attribution department according to the attribution of medical service items;

[0092] The hospital cost prediction module 44 is used to predict the costs of medical service items of the hospital based on the costs of medical service items of each attribution department.

[0093] In some embodiments, the key factor acquisition module 41 further includes a comparison unit (not shown in the figure), and the comparison unit is configured to compare the basic information of the medical service item with the relevant information in the standard library to screen out the key factors of the medical service item.

[0094] In some embodiments, the above device further includes a sample acquisition module, a feature extraction module, and a model creation module (not shown in the figure).

[0095] Among them,

[0096] The sample acquisition module is configured to acquire training samples;

[0097] The feature extraction module is configured to perform feature extraction on the training samples to obtain sequence features corresponding to the training samples;

[0098] The model creation module establishes and trains a model according to the sequence features corresponding to the training samples to obtain a deep neural network model.

[0099] In some embodiments, the cost prediction module 43 for the affiliated department includes a duplicate removal unit, a filtering unit, and an inspection unit (not shown in the figure).

[0100] Among them,

[0101] The duplicate removal unit is configured to remove duplicates from the relevant data of the medical service item according to a preset duplicate removal rule to avoid predicting the cost of the medical service item repeatedly;

[0102] The filtering unit is configured to screen the relevant data of the medical service item after duplicate removal according to a preset filtering rule to retain the data participating in the cost prediction of the medical service item;

[0103] The inspection unit is configured to check whether the cost prediction of the medical service item for each affiliated department is abnormal according to a preset judgment rule; if so, an abnormal prompt is given, and the data for the cost prediction of the medical service item is traced and inspected according to the abnormal prompt; if not, a normal prompt is given.

[0104] In some embodiments, the inspection unit is further configured to check whether the ratio of the income to the cost of the medical service item for each affiliated department is within the range of a first set threshold and a second set threshold, where the first set threshold is less than the second set threshold; if the ratio of the income to the cost of the medical service item for each of the affiliated departments is greater than or equal to the first set threshold and less than or equal to the second set threshold, the cost prediction of the medical service item for each affiliated department is normal.

[0105] In some embodiments, the cost prediction module 43 for the affiliated department further includes a duplicate removal rule setting unit, a filtering rule setting unit, and a judgment rule setting unit (not shown in the figure).

[0106] Among them,

[0107] a duplicate removal rule setting unit, configured to set duplicate removal rules for relevant data of medical service items;

[0108] a filtering rule setting unit, configured to set filtering rules for relevant data of medical service items after duplicate removal;

[0109] a judgment rule setting unit, configured to set judgment rules for whether the cost prediction of medical service items in each affiliated department is abnormal.

[0110] In some embodiments, the affiliated department cost prediction module 43 further includes a threshold setting unit (not shown in the figure), and the threshold setting unit is configured to set a first set threshold and a second set threshold according to the actual situation.

[0111] An embodiment of the present application further provides a terminal, including a memory and a processor. A computer program that can run on the processor is stored on the memory. When the processor runs the computer program, it executes the above-mentioned Figure 1 medical service item cost prediction method based on key factors as shown.

[0112] Specifically, the processor may be a CPU, a general-purpose processor, a DSP, an ASIC, an FPGA or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It can implement or execute various exemplary logic blocks, modules and circuits described in combination with the disclosure of the present application. The processor may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0113] Specifically, the processor is connected to the memory through a bus, and the bus may include a path for transmitting information. The bus may be a PCI bus or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc.

[0114] The memory may be a ROM or other type of static storage device that can store static information and instructions, a RAM or other type of dynamic storage device that can store information and instructions, or an EEPROM, a CD-ROM or other optical disc storage, optical disc storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto.

[0115] Optionally, the memory is used to store the code of the computer program for executing the solution of the present application, and is controlled by the processor to execute. The processor is used to execute the application program code stored in the memory to implementFigure 5 The operations of the cost prediction device provided by the illustrated embodiment.

[0116] An embodiment of the present application further provides a computer-readable storage medium. The computer-readable storage medium includes a stored computer program. When the computer program is run by a processor, it controls the terminal where the storage medium is located to execute the above-mentioned Figure 1 medical service item cost prediction method based on key factors as illustrated.

[0117] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0118] Those of ordinary skill in the art can understand that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and their appropriate combinations. Some physical components or all physical components can be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include a computer storage medium (or non-transitory medium) and a communication medium (or transitory medium). As is well known to those of ordinary skill in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable instructions, data structures, program modules, or other data. Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassette, tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, as is well known to those of ordinary skill in the art, a communication medium generally includes computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transmission mechanism, and can include any information delivery medium.

[0119] The above is a specific description of the preferred embodiment of the present application. However, the present application is not limited to the above-mentioned embodiments. Those skilled in the art can make various equivalent deformations or substitutions without departing from the spirit of the present application. These equivalent deformations or substitutions are all included within the scope defined by the claims of the present application.

Claims

1. A method for predicting the cost of medical service items based on key factors, characterized in that, it includes: Processing the basic information of medical service items based on the canonical correlation analysis algorithm to obtain the key factors of the medical service items; Using the canonical correlation analysis algorithm to compare the basic information of the medical service items with the relevant information in the standard library to screen out the key factors of the medical service items; the basic information of the medical service items is represented by X, and the relevant information in the standard library is represented by Y. Both the basic information X and the relevant information Y in the standard library can be used as multi-dimensional arrays. Using the canonical correlation analysis algorithm to calculate the correlation coefficient ρ(X, Y) between X and Y, and based on the correlation coefficient ρ(X, Y), the correlation degree between the basic information X and the relevant information Y in the standard library can be judged, and then the key factors corresponding to the medical service items can be screened out. The calculation formula is as follows: ； Wherein, the input quantities are X and Y, the dimensions of X and Y are both greater than 1, and the output quantity is the correlation coefficient ρ(X, Y); S XX represents the variance of X, S YY represents the variance of Y, S XY represents the covariance of X and Y. The matrix M is subjected to singular value decomposition, and the largest singular value obtained is the correlation coefficient ρ(X, Y). The value range of the correlation coefficient ρ(X, Y) is [-1, 1]. The closer the absolute value of ρ(X, Y) is to 1, the higher the linear correlation between X and Y; the closer the absolute value of ρ(X, Y) is to 0, the lower the linear correlation between X and Y; The key factors of the medical service items include but are not limited to item code, item name, item connotation, pricing unit, manpower consumption and time consumption, low-value consumable consumption quota level, technical difficulty and risk level; Inputting the key factors of the medical service items into a pre-trained deep neural network model to determine the attribution of the medical service items; after determining the attribution department of the medical service items according to the item code, item name, item connotation and pricing unit in the key factors, then determine the personnel expenses for executing the medical project according to the manpower consumption and time consumption, risk level and technical difficulty in the key factors, and determine the non-separately charged health material costs, fixed asset depreciation costs, intangible asset amortization costs and other operating costs corresponding to the medical project according to the workload, project time consumption and low-value consumable consumption quota level in the key factors, and determine the provision of medical risk funds according to the key factor of risk level; Predicting the costs of medical service items of each attribution department according to the attribution of the medical service items; Removing duplicates from the relevant data of the medical service items according to the preset deduplication rules to avoid repeated prediction of the costs of the medical service items; screening the relevant data of the medical service items after deduplication according to the preset filtering rules to retain the data participating in the cost prediction of the medical service items; checking whether the cost prediction of the medical service items of each attribution department is abnormal according to the preset judgment rules; if so, giving an abnormal prompt, and performing retrospective inspection on the data of the cost prediction of the medical service items according to the abnormal prompt; if not, giving a normal prompt; Predicting the costs of the hospital's medical service items based on the cost predictions of the medical service items of each attribution department.

2. The method for predicting the cost of medical service items according to claim 1, characterized in that, Before the step of inputting the key factors of the medical service items into a pre-trained deep neural network model to determine the attribution of the medical service items, the method further includes: Obtaining training samples; Performing feature extraction on the training samples to obtain the sequence features corresponding to the training samples; Establishing and training a model according to the sequence features corresponding to the training samples to obtain the deep neural network model.

3. The medical service item cost prediction method according to claim 1, characterized in that, the step of checking whether the medical service item cost prediction of each of the affiliated departments is abnormal according to a preset judgment rule includes: checking whether the ratio of the income to the cost of the medical service items of each of the affiliated departments is within the range of a first set threshold and a second set threshold, wherein the first set threshold is less than the second set threshold; if the ratio of the income to the cost of the medical service items of each of the affiliated departments is greater than or equal to the first set threshold and less than or equal to the second set threshold, then the medical service item cost prediction of each of the affiliated departments is normal.

4. A cost prediction device, characterized in that, comprising: a key factor acquisition module, configured to process the basic information of the medical service item based on a canonical correlation analysis algorithm to obtain the key factors of the medical service item; using the canonical correlation analysis algorithm to compare the basic information of the medical service item with the relevant information in the standard library to screen out the key factors of the medical service item; the basic information of the medical service item is represented by X, and the relevant information in the standard library is represented by Y. Both the basic information X and the relevant information Y in the standard library can be used as multi-dimensional arrays. Using the canonical correlation analysis algorithm to calculate the correlation coefficient ρ(X, Y) between X and Y, and according to the correlation coefficient ρ(X, Y), the correlation degree between the basic information X and the relevant information Y in the standard library can be judged, and then the key factors corresponding to the medical service item can be screened out; the calculation formula is as follows: ； Wherein, the input quantities are X and Y, the dimensions of X and Y are both greater than 1, and the output quantity is the correlation coefficient ρ(X, Y); S XX represents the variance of X, S YY represents the variance of Y, S XY represents the covariance of X and Y. Perform singular value decomposition on the matrix M, and the largest singular value obtained is the correlation coefficient ρ(X, Y). The value range of the correlation coefficient ρ(X, Y) is [-1, 1]. The closer the absolute value of ρ(X, Y) is to 1, the higher the linear correlation between X and Y; the closer the absolute value of ρ(X, Y) is to 0, the lower the linear correlation between X and Y; the key factors of the medical service item include, but are not limited to, item code, item name, item connotation, pricing unit, man-hour consumption and time-consuming, low-value consumable consumption quota level, technical difficulty and risk level; an affiliated department determination module, configured to input the key factors of the medical service item into a pre-trained deep neural network model to determine the affiliation of the medical service item; after determining the affiliated department of the medical service item according to the item code, item name, item connotation and pricing unit in the key factors, then determine the personnel expenses for the personnel performing the medical project according to the man-hour consumption and time-consuming, risk level and technical difficulty in the key factors, and determine the non-separately charged health material costs, fixed asset depreciation costs, intangible asset amortization costs and other operating costs corresponding to the medical project according to the workload, project time-consuming and low-value consumable consumption quota level in the key factors, and determine the provision for medical risk funds according to the key factor of the risk level; The affiliated department cost prediction module is used to predict the medical service item costs of each affiliated department according to the affiliation of the medical service items; de-duplicate the relevant data of the medical service items according to the preset de-duplication rules to avoid duplicate prediction of the costs of the medical service items; screen the relevant data of the medical service items after de-duplication according to the preset filtering rules to retain the data participating in the cost prediction of the medical service items; check whether the cost prediction of the medical service items of each affiliated department is abnormal according to the preset judgment rules; if so, give an abnormal prompt and conduct a retrospective inspection of the data for the cost prediction of the medical service items according to the abnormal prompt; if not, give a normal prompt. The hospital cost prediction module is used to predict the medical service item costs of the hospital based on the cost predictions of the medical service items of each affiliated department.

5. The cost prediction device according to claim 4, wherein, the affiliated department cost prediction module includes: a de-duplication unit, configured to de-duplicate the relevant data of the medical service items according to the preset de-duplication rules to avoid duplicate prediction of the costs of the medical service items; a filtering unit, configured to screen the relevant data of the medical service items after de-duplication according to the preset filtering rules to retain the data participating in the cost prediction of the medical service items; an inspection unit, configured to check whether the cost prediction of the medical service items of each affiliated department is abnormal according to the preset judgment rules; if so, give an abnormal prompt and conduct a retrospective inspection of the data for the cost prediction of the medical service items according to the abnormal prompt; if not, give a normal prompt.

6. A terminal, comprising a memory and a processor, and a computer program is stored on the memory and can run on the processor, wherein, when the processor runs the computer program, it executes the medical service item cost prediction method according to any one of claims 1 to 3.

7. A computer-readable storage medium, wherein, the computer-readable storage medium includes a stored computer program, wherein when the computer program is run by a processor, it controls the terminal where the storage medium is located to execute the medical service item cost prediction method according to any one of claims 1 to 3.

Images