Medical Consumables Anomaly Analysis System

Abstract

A medical institution consumable anomaly analysis system includes a preprocessing module, a consumable and physician cost prediction module, and an abnormal physician behavior detection module. The preprocessing module preprocesses the raw data and converts it into a feature set; the consumable and physician cost prediction module predicts the consumable costs for multiple physicians based on the feature set and establishes a consumable cost baseline accordingly; the abnormal physician behavior detection module generates an anomaly warning message when the deviation between the real-time consumable cost data and the consumable cost baseline exceeds a predetermined threshold.

Claims

1. A medical institution consumable anomaly analysis system, comprising: a preprocessing module for connecting to at least one data source and receiving a plurality of raw data from the at least one data source; the preprocessing module preprocessing the raw data and converting it into a feature set, wherein the feature set corresponds to a plurality of clinical surgical operation data of a plurality of physicians; the preprocessing module outputting the feature set; and a consumable and physician cost prediction module connected to the preprocessing module and receiving the feature set, wherein the consumable and physician cost prediction module predicts individual consumable costs for a plurality of physicians based on the feature set, the individual consumable costs for a plurality of physicians corresponding to the physicians, and establishes at least one consumable cost baseline based on the individual consumable costs for a plurality of physicians; An abnormal physician behavior detection module is connected to the consumables and physician cost prediction module and the at least one data source. The abnormal physician behavior detection module receives the at least one consumable cost baseline from the consumables and physician cost prediction module and receives real-time consumable cost data from the at least one data source. The real-time consumable cost data corresponds to a physician. The abnormal physician behavior detection module generates an abnormal warning message when a deviation between the real-time consumable cost data and the at least one consumable cost baseline exceeds a predetermined threshold. A report generation module is connected to the abnormal physician behavior detection module. The report generation module generates a management report based on the abnormal warning message. The abnormal physician behavior detection module receives the predicted costs of individual consumables from the consumables and physician cost prediction module. When the deviation between the real-time consumable cost data and the at least one consumable cost baseline exceeds the predetermined threshold, the abnormal physician behavior detection module adds the predicted cost of individual consumables corresponding to the real-time consumable cost data and the deviation to the abnormal warning message. The report generation module receives the abnormal warning message containing the predicted cost of individual consumables and the deviation, and the generated management report includes the predicted cost of individual consumables and the deviation.

2. The medical institution consumable anomaly analysis system as described in claim 1, wherein the consumable and physician cost prediction module converts the feature set into a high-dimensional feature set for each physician, the high-dimensional feature set including surgical operation feature data, consumable usage feature data, technician cost feature data, patient feature data and physician attribute feature data, the consumable and physician cost prediction module inputs the high-dimensional feature set of each physician into a prediction model to predict the individual consumable cost forecast value for each physician.

3. The medical institution consumable anomaly analysis system as described in claim 1, wherein the real-time consumable cost data and the consumable cost baseline have multiple dimensions of deviation, including a consumable type deviation, a consumable quantity deviation and a technician cost deviation, and the abnormal physician behavior detection module calculates the deviation based on these dimensions of deviation.

4. The medical institution consumable anomaly analysis system as described in claim 1, wherein the report generation module has a comparison unit, a data interpretation rule base, and a language model. The data interpretation rule base has a plurality of data conditions and a plurality of statement templates corresponding to these data conditions. The comparison unit compares the anomaly warning message with the data interpretation rule base to select one of the statement templates corresponding to the anomaly warning message and inputs the selected statement template into the language model. The language model generates the management report based on the selected statement template and the anomaly warning message.

5. A medical institution consumable anomaly analysis system, comprising: a preprocessing module for connecting to at least one data source and receiving a plurality of raw data from the at least one data source; the preprocessing module preprocessing the raw data and converting it into a feature set, wherein the feature set corresponds to a plurality of clinical surgical operation data of a plurality of physicians; the preprocessing module outputting the feature set; and a consumable and physician cost prediction module connected to the preprocessing module and receiving the feature set, wherein the consumable and physician cost prediction module predicts individual consumable costs for a plurality of physicians based on the feature set, the individual consumable costs for a plurality of physicians corresponding to the physicians, and establishes at least one consumable cost baseline based on the individual consumable costs for a plurality of physicians. An abnormal physician behavior detection module is connected to the consumables and physician cost prediction module and the at least one data source. The abnormal physician behavior detection module receives the at least one consumable cost baseline from the consumables and physician cost prediction module and receives real-time consumable cost data from the at least one data source, the real-time consumable cost data corresponding to a physician. The abnormal physician behavior detection module generates an abnormal warning message when a deviation between the real-time consumable cost data and the at least one consumable cost baseline exceeds a predetermined threshold. A report generation module is connected to the abnormal physician behavior detection module, and the report generation module generates a management report based on the abnormal warning message. The physicians have multiple levels; the number of at least one consumable cost baselines is multiple and each corresponds to one of the levels. The consumables and physician cost prediction module establishes consumable cost baselines based on the physicians' levels and their individual consumable cost predictions. The abnormal physician behavior detection module receives these consumable cost baselines from the consumables and physician cost prediction module and selects a corresponding consumable cost baseline based on the physician's level corresponding to the real-time consumable cost data.

6. A medical institution consumable anomaly analysis system, comprising: a preprocessing module for connecting to at least one data source and receiving a plurality of raw data from the at least one data source; the preprocessing module preprocessing the raw data and converting it into a feature set, wherein the feature set corresponds to a plurality of clinical surgical operation data of a plurality of physicians; the preprocessing module outputting the feature set; and a consumable and physician cost prediction module connected to the preprocessing module and receiving the feature set, wherein the consumable and physician cost prediction module predicts individual consumable costs for a plurality of physicians based on the feature set, the individual consumable costs for a plurality of physicians corresponding to the physicians, and establishes at least one consumable cost baseline based on the individual consumable costs for a plurality of physicians; An abnormal physician behavior detection module connects the consumables and physician cost prediction module to the at least one data source. The abnormal physician behavior detection module receives the at least one consumable cost baseline from the consumables and physician cost prediction module and receives real-time consumable cost data from the at least one data source, the real-time consumable cost data corresponding to a physician. The abnormal physician behavior detection module generates an abnormal warning message when a deviation between the real-time consumable cost data and the at least one consumable cost baseline exceeds a predetermined threshold. The real-time consumable cost data and the consumable cost baseline have multiple dimensions of deviation, including a consumable type deviation, a consumable quantity deviation, and a technician cost deviation. The abnormal physician behavior detection module calculates the deviation based on these dimensions of deviation. The medical institution consumable anomaly analysis system includes a report generation module connected to the abnormal physician behavior detection module. The report generation module generates a management report based on the anomaly warning message. The abnormal physician behavior detection module includes an interpretability analysis unit that determines the contribution of each dimension deviation, generates a contribution list, and attaches the contribution list to the anomaly warning message. The management report generated by the report generation module includes the contribution list.

7. The medical institution consumable anomaly analysis system as described in claim 6 includes a clinical terminal device connected to the abnormal physician behavior detection module and the at least one data source. The clinical terminal device transmits consumable demand data to the at least one data source, and the at least one data source converts the consumable demand data into real-time consumable cost data. The abnormal physician behavior detection module can output the abnormal warning message to the clinical terminal device.

8. The medical institution consumable anomaly analysis system as described in claim 7, wherein the clinical terminal device transmits new consumable demand data to the at least one data source, the at least one data source converts the new consumable demand data into new real-time consumable cost data and transmits it to the abnormal physician behavior detection module; the abnormal physician behavior detection module compares whether the deviation between the new real-time consumable cost data and the at least one consumable cost baseline exceeds the predetermined threshold, and if the deviation still exceeds the predetermined threshold, the abnormal warning message is generated again.

Images