An artificial intelligence-based tumor nursing service management system

Abstract

The present invention relates to the field of tumor nursing technology, specifically an artificial intelligence-based tumor nursing service management system, the system includes a medication detection module, a combined medication module, a vital sign recognition module, an action push module and a resource scheduling module. Based on the daily medication data of tumor patients, the system analyzes the drug name, dosage and medication time, and compares the actual and standard dosage differences. The present invention automatically analyzes the patient's continuous medication pattern, vital sign changes and real-time laboratory monitoring data, integrates multi-source health parameters to complete dynamic aggregation and linkage judgment, realizes high-frequency identification of risk factors and monitoring of sensitive vital sign fluctuations, timely generates intervention measure push information, optimizes the nursing task response process, and uses multi-dimensional data fusion and intelligent task sorting to achieve accurate matching of patient risk management and nursing resource scheduling, thereby promoting the data closed loop and intervention closed loop of the entire tumor nursing process and enhancing the initiative and adaptability of nursing decision-making.

Description

Technical Field

[0001] The present invention relates to the technical field of tumor nursing, and in particular to a tumor nursing service management system based on artificial intelligence. Background Art

[0002] The field of oncology nursing involves health management, risk assessment, nursing intervention, symptom monitoring, and follow-up management information integration for cancer patients throughout the entire process of diagnosis, treatment, and rehabilitation. It covers the multi-dimensional nursing needs of cancer patients, including physiological, psychological, social, and behavioral needs, and is an important foundation for achieving health management and personalized nursing services for cancer patients throughout their life cycle. Among them, traditional oncology nursing service management systems refer to nursing management activities such as collecting patient information, tracking disease conditions, recording risk assessment data, analyzing nursing plan recommendations, and arranging follow-up plans through manual entry. Common methods include paper medical records, manual registration, spreadsheet information entry, rule-based data query and early warning tools, etc. These methods are mostly dominated by manual experience and rely mainly on nursing staff to regularly evaluate and address patients' health and nursing needs.

[0003] Traditional technologies mostly use manual collection and decentralized registration of information, which makes it difficult to support real-time linkage of health data and full-process risk management. There is a data gap between nursing risk indicators and patient status monitoring links, which can easily lead to missing key information or delayed abnormal responses during the stage of disease changes. When patients' vital signs fluctuate frequently or medication regimens are adjusted, traditional methods are difficult to quickly coordinate nursing tasks and manpower arrangements, affecting the intervention timeliness and service continuity of high-risk patients. Summary of the Invention

[0004] The purpose of the present invention is to solve the shortcomings of the existing technology and propose an artificial intelligence-based tumor nursing service management system.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: an artificial intelligence-based tumor nursing service management system, the system comprising:

[0006] The medication detection module analyzes the drug name, dosage, and time of administration based on the daily medication data of cancer patients, compares the difference between the actual and standard dosages, calculates the difference between the medication interval and the drug metabolism interval, determines the risk characteristics, and obtains the deviation medication characteristics;

[0007] Based on the offset medication characteristics, the combined medication module compares the order of drug combinations and the adverse reaction time, screens for combinations with overlapping time, analyzes the reference range of laboratory tests, calculates the risk groups, and obtains the drug risk superposition amount;

[0008] The physical sign recognition module compares the patient care stage with the risk information based on the drug risk superposition, collects daily physiological data, determines the variation range of physical sign parameters and stage standards, and analyzes the cumulative fluctuations to obtain the abnormal physical sign concentration rate;

[0009] The action push module screens the signs with high fluctuation frequency based on the abnormal concentration rate of the signs, analyzes the combined performance of the signs with blood oxygen and respiratory rate, determines whether it is a respiratory intervention event, identifies the corresponding nursing action, and obtains the intervention action instruction sequence number.

[0010] The present invention has the following improvements: the deviation medication characteristics include dosage deviation type, medication interval category, and drug warning mark; the drug risk superposition amount includes combined risk level, adverse reaction mark, and laboratory abnormality item; the abnormal physical sign concentration rate includes sensitive physical sign grouping, stage abnormal performance, and continuous fluctuation record; the intervention action instruction sequence number includes task action number, intervention category label, and nursing response method.

[0011] The present invention is improved in that the medication detection module includes:

[0012] The dose difference determination submodule compares the actual dosage with the standard dosage of each drug based on the daily medication data of cancer patients, determines whether the dosage difference of each drug exceeds the safety reference range, screens drug combinations with abnormal dosages, and obtains the total amount of abnormal dosage differences;

[0013] The interval comparison submodule calculates the interval between two consecutive doses of the same drug based on the total amount of dosage difference abnormalities, compares the actual interval with the recommended interval of the drug, determines whether there is a situation that exceeds the reference range, and counts the cumulative situation of drug interval abnormalities to obtain the interval time deviation frequency;

[0014] The offset feature generation submodule extracts the corresponding drug name based on the interval time offset frequency, combines and associates it with the abnormal indicators, obtains the combined abnormal response intensity, extracts the drug combination with high frequency and concentrated numerical amplitude, and obtains the offset medication feature.

[0015] The present invention is improved in that the combined medication module comprises:

[0016] The time series comparison submodule analyzes the drug names and medication time series based on the offset medication characteristics, compares the actual medication sequence of each drug combination with the time sequence of adverse reactions, identifies the combinations of actual medication and adverse reactions that are temporally correlated, determines whether there is a pattern in the combination, and obtains a time series coincidence group sequence;

[0017] The risk linkage submodule analyzes the laboratory test results of the combination within the same medication cycle based on the time-series coincidence group sequence, compares the synchronization of the medication time of the drug combination and the test abnormality, and counts the coincidence frequency within the same cycle to obtain the frequency of risk coupling items;

[0018] The abnormal co-occurrence submodule calculates the abnormal co-occurrence amplitude according to the frequency of the risk coupling items, sorts the combinations according to the abnormal co-occurrence amplitude, screens the combination with the best co-occurrence degree, and obtains the drug risk superposition amount.

[0019] The present invention is improved in that the vital sign recognition module includes:

[0020] The stage benchmark comparison submodule analyzes the physiological data of the cancer patient at the current care stage based on the drug risk superposition, compares the difference between each physical sign and the standard data of the corresponding stage, determines the physical sign that causes the fluctuation, and obtains the physical sign stage deviation amplitude;

[0021] The continuous fluctuation measurement submodule calculates the daily change trend of the vital sign during the continuous monitoring process based on the deviation amplitude of the vital sign stage, compares the vital sign data of each day with the data performance of the previous cycle, analyzes the continuous fluctuation of the vital sign over time, optimizes the distribution of the fluctuation data, and obtains the intensity level of the vital sign fluctuation;

[0022] The abnormality rate calculation submodule analyzes the physical sign items showing abnormal trends based on the level of physical sign fluctuation intensity, screens the physical sign manifestations that show abnormal changes during the continuous monitoring period, determines the concentration of abnormal distribution and stage benchmark manifestations, and statistically analyzes the distribution of abnormal data to obtain the physical sign abnormality concentration rate.

[0023] The present invention is improved in that the action push module includes:

[0024] The fluctuation feature screening submodule analyzes the continuous fluctuation records based on the abnormal concentration rate of the physical signs, compares the fluctuation frequency of each physical sign in the current nursing stage, screens the key physical signs, and determines the prominence of the key physical signs in the stage to obtain the key physical sign identification index;

[0025] The combined abnormality judgment submodule compares the daily fluctuations of blood oxygen saturation and respiratory rate based on the key vital sign identification indicators, obtains the intensity level of combined respiratory abnormality, compares it with the intervention initiation standard, determines whether the intervention conditions are met, and determines the respiratory system intervention event;

[0026] The nursing action matching submodule screens the event types that meet the intervention conditions based on the respiratory system intervention events, compares the current events with the matching items in the nursing action list, determines the category and response method of the matching actions, optimizes the nursing resource allocation structure, and obtains the intervention action instruction sequence number.

[0027] The present invention is improved in that the system further comprises:

[0028] The resource scheduling module analyzes the remaining working hours, task progress and skill tags of the nursing staff based on the intervention action instruction sequence number, selects and matches the nursing staff, and sorts and allocates priorities based on the task progress and historical efficiency to obtain the nursing task priority sequence;

[0029] The nursing task priority sequence includes job assignment order, resource allocation number, and scheduling priority label.

[0030] The present invention is improved in that the resource scheduling module includes:

[0031] The work time analysis submodule analyzes the remaining work time of the nursing staff based on the intervention action instruction sequence number, and screens the staff who can complete the nursing task and have matching skills based on their actual participation time and skill tags. It also compares the current work time schedule of each staff member with the time requirement of the upcoming task to generate the amount of available nursing resources.

[0032] The task progress judgment submodule analyzes the completion status of the nursing staff's current task based on the available nursing resources, compares the planned task progress, determines whether the actual progress is within the required range, and generates a scheduling adaptation deviation rate;

[0033] The priority sorting submodule analyzes the historical task completion performance of personnel in the nursing response records based on the scheduling adaptation deviation rate, compares their remaining working hours with their historical task participation, determines the task response order of each nursing staff in the actual nursing tasks, and obtains the nursing task priority sequence.

[0034] Compared with the prior art, the advantages and positive effects of the present invention are:

[0035] In the present invention, by automatically analyzing the patient's continuous medication pattern, changes in vital signs and real-time laboratory monitoring data, multi-source health parameters are integrated to complete dynamic collection and linkage judgment, realize high-frequency identification of risk factors and monitoring of sensitive vital sign fluctuations, timely generate intervention measures push information, optimize the nursing task response process, and use multi-dimensional data fusion and intelligent task sorting to achieve accurate matching of patient risk management and nursing resource scheduling, thereby promoting the data closed loop and intervention closed loop of the entire tumor nursing process and enhancing the initiative and adaptability of nursing decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 is a system flow chart of the present invention;

[0037] Figure 2 This is a flow chart of the medication detection module in the present invention;

[0038] Figure 3 This is a flow chart of the combined medication module of the present invention;

[0039] Figure 4 This is a flow chart of the physical sign recognition module in the present invention;

[0040] Figure 5 This is a flow chart of the action push module in the present invention;

[0041] Figure 6 This is a flow chart of the resource scheduling module in the present invention. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0043] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings and are only for the convenience of describing the present invention and simplifying the description. They do not indicate or imply that the devices or elements referred to must have a specific direction, be constructed and operate in a specific direction, and therefore should not be understood as limiting the present invention. In addition, in the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0044] Example:

[0045] See also Figure 1 The present invention provides a technical solution: an artificial intelligence-based tumor nursing service management system comprising:

[0046] The medication detection module analyzes the drug name, dosage, and time of administration based on daily medication data of cancer patients. It compares the actual dosage with the standard dosage, calculates the difference between the time between two consecutive medication administrations and the drug metabolism time, and determines whether any difference exceeds the drug safety reference range. By correlating the drug name, dosage difference, and time difference, it obtains the deviation medication characteristics.

[0047] The combined medication module compares the order of drug combination use and the time of adverse reaction onset based on the characteristics of offset medication use, screens for drug combinations with overlapping time periods, analyzes the reference ranges of laboratory test items on the day, and counts the simultaneous occurrence of combined medication combinations and out-of-range test items to obtain the drug risk superposition amount;

[0048] The vital sign recognition module compares the patient's current care stage with the risk data based on the drug risk superposition, collects daily physiological data during the patient's care stage, determines the magnitude of change between each vital sign parameter and the stage reference standard, analyzes the cumulative magnitude of continuous fluctuations of each vital sign, and counts the cumulative number of fluctuation changes to obtain the abnormal vital sign concentration rate;

[0049] The action push module selects key physical sign items with high fluctuation frequency based on the abnormal concentration rate of physical signs, analyzes the combined performance of the physical sign in the blood oxygen and respiratory rate of the day, determines whether the changes in blood oxygen and respiratory rate are abnormal at the same time, determines the respiratory system intervention event, and screens the nursing action corresponding to the event to obtain the intervention action instruction sequence number;

[0050] The resource scheduling module analyzes the remaining working hours, task progress and skill tags of nursing staff based on the intervention action instruction sequence number, screens nursing staff with matching skill tags, determines whether the current task progress of the staff meets the scheduling requirements, compares the historical completion efficiency in the nursing response records, and prioritizes the screened nursing staff to obtain the nursing task priority sequence.

[0051] The characteristics of medication deviation include dosage deviation type, medication interval category, and drug warning label; the drug risk superposition includes joint risk level, adverse reaction marker, and laboratory abnormality items; the abnormal physical sign concentration rate includes sensitive physical sign grouping, stage abnormal manifestation, and continuous fluctuation record; the intervention action instruction sequence includes task action number, intervention category label, and nursing response method; the nursing task priority sequence includes job assignment order, resource allocation number, and scheduling priority label.

[0052] In Module 1, the difference between the actual and standard drug doses refers to the difference between the actual dose of a drug taken by a patient and the standard dose of the drug recommended in the pharmacopoeia, clinical pathway or drug instructions; the interval difference refers to the difference between the actual interval between two doses of the same drug when the patient takes the same drug twice in a row and the interval recommended for the normal pharmacokinetics (metabolic cycle) of the drug; the drug safety reference range refers to the reasonable range of doses and medication intervals set for each drug in clinical applications to ensure efficacy and safety. This range is usually derived from the pharmacopoeia or authoritative clinical guidelines.

[0053] In Module 2, the order of use refers to the actual chronological order of the use of multiple drugs by the patient, which is compared with the chronological order of the adverse reactions; overlapping drug combinations refer to the overlap of the time points of the two or more drugs used by the patient in combination with the occurrence of adverse reactions within a specific time period, and such drug combinations are screened out and focused on; the reference range of the test items refers to the normal range or standard value of each indicator in laboratory tests (such as liver function, kidney function, etc.), and the range is used to determine whether the test results are abnormal; simultaneous occurrence refers to the discovery during the analysis process that the combination of drugs and abnormal laboratory test results are recorded on the same day or period, indicating that the two are related.

[0054] In Module 3, the nursing stage refers to the specific stage of the nursing process for cancer patients, such as before surgery, after surgery, during chemotherapy or the recovery stage; risk data refers to the drug risk superposition results generated in the aforementioned combination medication module, which is used to remind patients of the drug-related nursing risks at the current stage; physiological data refers to the vital signs collected daily by patients, including blood pressure, heart rate, body temperature, blood oxygen, respiratory rate and other parameters directly related to the patient's health status; the stage reference standard refers to the normal range of various physiological data specified in the clinical nursing guidelines for each nursing stage; the amplitude of change refers to the absolute difference in the increase or decrease of a physiological parameter actually monitored during continuous monitoring; the cumulative amplitude refers to the cumulative sum of all single change amplitudes of a physiological parameter over a period of time, which is used to reflect the degree of fluctuation of the sign.

[0055] In Module 4, key vital signs refer to the physiological parameters that are found to fluctuate most frequently and have the greatest impact on the patient's current health risks after analysis, such as blood pressure, heart rate, blood oxygen, etc.; the combined performance of blood oxygen and respiratory rate refers to the joint changes in the two vital sign parameters of blood oxygen saturation and respiratory rate on the same day or in the same period; respiratory system intervention events refer to situations where respiratory system-related nursing intervention is required for patients based on the combined abnormalities of blood oxygen and respiratory rate; corresponding nursing actions refer to specific nursing measures matched according to the system's judgment results, such as oxygen inhalation, respiratory monitoring, doctor notification and other operation sequences.

[0056] In Module 5, matching skill tags refer to the professional capabilities or qualifications (such as chemotherapy nursing, respiratory nursing, etc.) recorded in the nursing staff's files, which correspond to the capabilities required for the current nursing task; scheduling requirements refer to the job allocation requirements of the current nursing task for personnel, such as professional skill requirements, allocable working hours, task progress, etc.; nursing response records refer to historical information such as the response speed and completion status of nursing staff who have received and completed nursing tasks, which are automatically recorded by the system; historical completion efficiency refers to the average response time or average completion quality of nursing staff who have completed nursing tasks of the same type or difficulty in the past, which is used to support personnel priority sorting.

[0057] See also Figure 2 , the medication detection module includes:

[0058] The dose difference determination submodule compares the actual dosage with the standard dosage of each drug based on the daily medication data of cancer patients, determines whether the dosage difference of each drug exceeds the safety reference range, screens drug combinations with abnormal dosages, and obtains the total amount of abnormal dosage differences;

[0059] The medication records submitted by patients every day are automatically extracted, and the drug name, actual dosage and corresponding medication time of each drug are extracted respectively. The drug name is used as the search key value, and the recommended dosage range of the drug registered in the pharmacopoeia, instructions or tumor treatment pathway is called in the preset database, and the upper and lower limits of its safe use are clearly defined as the dosage reference boundaries. Then, the actual dosage value taken by the patient and the called recommended dosage are directly calculated by difference. If the result is within the safe use range of the recommended dosage, the drug is marked as normal dosage. If it exceeds the boundary on either side of the upper and lower limits, it is marked as abnormal dosage. All drugs are judged for differences item by item and the corresponding identifications are stored. Then, all drugs are screened according to their dosage identifications, and all drugs judged to be of abnormal dosages are extracted. A set of abnormal drug combinations is constructed, and the drugs in the set are counted to form the statistical results of abnormal drugs as the output item. For example, a patient's daily medication record contains five drugs. The actual dosages of two of the drugs are twice and half of the recommended dosages, respectively. After comparison, it is judged that they exceed the recommended upper limit and are lower than the recommended lower limit. In this case, the two drugs are included in the abnormal combination, and the number of drugs in the abnormal combination is counted as 2 to determine the total amount of abnormal daily dosage differences.

[0060] The interval comparison submodule calculates the interval between two consecutive doses of the same drug based on the total amount of dose difference abnormalities, compares the actual interval with the recommended interval of the drug, determines whether it exceeds the reference range, and counts the cumulative abnormalities of the drug interval to obtain the frequency of interval time deviation;

[0061] For each drug, all relevant medication time data from the patient's historical records are retrieved. The two medication times for each drug within a consecutive number of days are arranged in order, and the medication time interval is calculated in chronological order. The medication time interval is then compared with the medication time interval setting range recommended in the authoritative data for the drug. Each medication time interval is determined to see if it deviates from the recommended interval range. If the interval is shorter than the recommended lower limit or longer than the recommended upper limit, the medication time interval is marked as abnormal. The medication time interval is continuously compared across multiple days of medication records for this drug, and the number of intervals determined to be abnormal is recorded. The number is then accumulated to obtain the interval deviation frequency. For example, for a certain drug, if a patient takes medication at 8:00 AM and 10:00 PM in one day, and the recommended interval for this drug is 12 hours, if the interval exceeds the recommended upper limit, the medication is marked as a deviation. For example, if the medication has similar interval behavior of exceeding the upper limit or falling short of the lower limit on three days out of five consecutive days, the cumulative interval deviation frequency for this drug in this period is recorded as 3. Finally, the interval deviation frequency of all drugs with abnormal dosages is calculated separately.

[0062] The offset feature generation submodule extracts the corresponding drug name based on the interval time offset frequency and combines it with the abnormal indicators using the formula:

[0063] ;

[0064] Get the combined abnormal response strength , extract the drug combinations with high frequency and concentrated numerical amplitude, and obtain the offset medication characteristics, among which, Indicates the The actual dosage of the drug, Indicates the The standard dose of a drug, Indicates the The actual time interval between two consecutive doses of the drug, Indicates the The recommended dosing interval for each drug, Indicates the The status flag for both the dose difference and the interval difference of the drug is abnormal, which can be 0 or 1. It represents the abnormal item collaborative amplification coefficient, which is used to adjust the impact of the simultaneous abnormal state. Indicates the number of drug types, Refers to the synergistic abnormality adjustment item, which gives higher / lower weight to the risk when both the dose and interval are abnormal. It means that the drug has both abnormal dosage and abnormal interval on the same day, that is, "combined abnormality". The value is 0 or 1. 0 means that the drug's dosage difference and interval difference are not abnormal at the same time on this day; 1 means that the drug's dosage difference and interval difference are abnormal on this day, that is, "abnormal superposition".

[0065] Combined abnormal response intensity refers to the difference between the actual dose and the standard dose of each drug, as well as the difference between the actual interval between two consecutive doses and the recommended interval, when analyzing the use of multiple medications in cancer patients. Combined with the comprehensive situation of abnormalities in both dose and interval, the indicators are uniformly converted into an aggregated indicator that can measure the strength of the abnormality. It measures the comprehensive strength and concentrated manifestation of the two types of abnormal phenomena, "dose difference" and "drug interval difference", in all medication items of a patient within a certain period. The larger the value, the more obvious the abnormal manifestation or the more drugs involved, and the higher the risk.

[0066] Call the names of drugs that have been identified as having dose differences and interval deviations, and establish corresponding relationships between them and the corresponding actual doses, standard doses, actual intervals, recommended intervals, and combined abnormal states. Analyze the structural distribution of each parameter in terms of numerical performance, and use standardization and normalization operations to uniformly calculate and process the values ​​under different dimensions and then enter them into the calculation formula. For example, the actual dose of cisplatin is 90 mg, the standard dose is 100 mg, and the normalized difference is , the actual interval is 18 hours, the recommended interval is 24 hours, and the normalized interval difference is , the joint abnormal state is 1, then its participation calculation is:

[0067] ;

[0068] The actual dose of gemcitabine is 1200 mg, the standard dose is 1000 mg, and the normalized difference is , the actual interval is 23 hours, the recommended interval is 24 hours, and the normalized difference is , the combined abnormal state is 1, Set it to 2 and substitute it into the calculation:

[0069] ;

[0070] The dose of paclitaxel is 160 mg, the standard dose is 175 mg, and the normalized difference is , the interval is 19 hours, the recommended interval is 24 hours, the normalized difference is , the joint abnormal state is 1, and the calculation is:

[0071] ;

[0072] The three results are 、 、 , and sum to get:

[0073] ;

[0074] This numerical result represents the concentration of response density formed by multiple drug deviation indicators under the dual abnormal conditions of dose and time interval. The higher the value, the stronger the deviation trend. Based on this, drug combinations with outstanding response density performance are extracted, and the ranking of drugs in risk distribution is established to form the deviation medication characteristics. This formula realizes the quantitative description of risk-concentrated drug combinations through the coordinated consideration of dose and interval differences and the participation of abnormal identification in the calculation, thus building an evaluation basis in the drug screening link.

[0075] See also Figure 3 , the combination medication module includes:

[0076] The time series comparison submodule analyzes the drug names and medication time series based on the offset medication characteristics, compares the actual medication sequence of each drug combination with the time sequence of adverse reactions, identifies the combinations of actual medication and adverse reactions that are temporally correlated, determines whether there is a pattern in the combination, and obtains a sequence of temporal coincidence groups;

[0077] First, retrieve all drug items that have been marked as having dosage deviation, abnormal medication interval or drug warning signs, and match their daily medication records one by one, extract the medication time of each drug, and construct a data table structure with the drug name as the index field and the daily medication time as the sequence. Then, all drug combinations involved are processed in pairs according to the patient's actual medication behavior, and the medication time series of each group of drug combinations are cross-sorted to determine the actual medication time sequence of each drug. Then, the time of occurrence of adverse reactions on that day and the next three days is extracted from the patient's electronic medical record, and the first recorded time of the adverse reaction is compared with the actual medication time of each drug in the drug combination. A comparison is performed to determine whether the adverse reaction occurs within 48 hours after all drugs are taken. If this condition is met, the combination is marked as having an associated relationship. For this type of marked combination, whether the patient repeatedly experiences this time sequence association relationship in multiple consecutive cycles is accumulated to determine whether there is a repetitive regularity. For example, in four chemotherapy cycles, the patient experiences nausea and vomiting within one day after taking fluorouracil and oxaliplatin. The time sequence relationship between the drug combination and the adverse reaction is judged to be associated and regular, and recorded as a stable overlap group. Finally, all drug combinations that meet the time sequence association and regular repetition are output as a time sequence overlap group sequence.

[0078] The risk linkage submodule analyzes the laboratory test results of the drug combination within the same medication cycle based on the time-series coincidence group sequence, compares the synchronization of the medication time and the test abnormalities of the drug combination, and counts the coincidence frequency within the same cycle to obtain the frequency of risk coupling items;

[0079] First, retrieve the medication time corresponding to each pair of drug combinations, and synchronously retrieve the laboratory test records of the same patient during the medication cycle in the database, retrieve the data values ​​and test times of each key test item respectively, and then set the cycle boundary according to the medication time window of the drug combination, and define the medication cycle window from the start day to the seventh day of the combination medication. Then, filter all test records whose test time falls within the cycle window from the laboratory records, and further judge whether the test value exceeds the corresponding reference value range. For example, the normal range of serum creatinine is set to 0.6 to 1.2 mg / dL. If the patient's test result is 1.6 mg / dL, the test item is marked as an abnormal item, and the time point when the abnormality occurred is recorded. Compare with the medication time of the corresponding drug combination to determine whether the test abnormality occurs within 48 hours after taking the drug combination. If the occurrence time overlaps with the medication time by no more than 48 hours, it is regarded as a synchronous event and recorded as an overlap. The number of overlaps is accumulated for all test records that meet the synchronization conditions, and the synchronous overlaps of the same drug combination for all patients in different cycles are counted. Finally, the number of times the drug combination and laboratory abnormalities appear together in the medication cycle is formed as the frequency output of the risk coupling item of the combination. For example, a patient used the same combination of drugs in three cycles, and an abnormal record of increased ALT appeared each time, and all occurred within 24 hours after taking the medicine. The risk coupling item frequency is recorded as 3.

[0080] The abnormal co-occurrence submodule uses the formula according to the frequency of risk coupling items:

[0081] ;

[0082] Calculate the abnormal co-occurrence amplitude , sorting by the abnormal co-occurrence amplitude of each combination, screening the combination with the best co-occurrence degree, and obtaining the drug risk superposition amount, among which, Indicates the total number of drugs in the drug combination, Indicates the The time to take the medicine, Indicates the The time of occurrence of adverse reactions corresponding to the drugs, Indicates the The number of abnormal test items in the laboratory tests of the drug on the same day, Indicates the The dose deviation of the drug Represents the total number of abnormal laboratory test items, Indicates the The co-occurrence frequency of abnormal laboratory tests, It is the serial number of the abnormal item detected by the laboratory. It is the serial number in the drug combination.

[0083] The abnormal co-occurrence amplitude refers to a comprehensive quantitative indicator used to measure the degree of synchronous co-occurrence between medication behavior, adverse reaction events and laboratory abnormality tests of a drug combination within a specific medication cycle. It can be used as a numerical basis to measure the drug risk superposition generated by the drug combination in actual application, to rank and screen the risk co-occurrence degrees of different drug combinations, and to identify drug combinations with potential high risks, providing data support for subsequent risk intervention and nursing decisions.

[0084] It is the absolute value of the difference between the two (square root), which is used to measure the time distance between the administration of each drug and the occurrence of adverse reactions. The smaller the value, the closer the adverse reaction is to the time of drug administration (i.e., the linkage / risk is greater); the larger the value, the farther the two are apart (relatively lower risk); The "abnormality intensity" is calculated by adding laboratory test abnormalities and dosage abnormalities together and taking the square root to avoid the impact of any single abnormality. A larger value indicates a more significant abnormality (laboratory / dosage) for that particular drug on that day. The absolute value of the temporal distance is combined with the abnormality intensity to reflect the "comprehensive intensity of the co-occurrence of abnormalities." If both are high, the abnormality risk is amplified; if either is low, the overall risk impact is limited.

[0085] Reflects the overall co-occurrence level of abnormalities detected during the medication cycle. The more abnormalities there are in the entire combination, the larger the denominator, and the lower the risk score. It is to ensure that there is a "scale correction" when calculating different drug groups. The more drugs there are, the larger the denominator will be, and the abnormal impact of a single drug will not be infinitely magnified.

[0086] The time to take each drug in the drug combination , adverse reaction time , number of laboratory abnormalities , dose offset amplitude and anomaly detection co-occurrence frequency Perform joint calculations and set the normalized parameters as follows:

[0087] 、 、 ;

[0088] 、 、 ;

[0089] 、 、 ;

[0090] 、 、 ;

[0091] 、 .

[0092] Substitute the normalized parameters into the formula and calculate the numerator first:

[0093] Group 1:

[0094] ;

[0095] Group 2:

[0096] ;

[0097] Group 3:

[0098] ;

[0099] The sum of the numerators is:

[0100] ;

[0101] Then calculate the denominator, the number of drugs , the sum of co-occurrence frequencies is , so the denominator is:

[0102] ;

[0103] It turns out that:

[0104] ;

[0105] The obtained abnormal co-occurrence amplitude , as an aggregate metric, is used to quantify the contribution of drug combinations to the potential risk of patients throughout the care cycle, based on the corresponding The results are sorted to screen out the combination with the best co-occurrence degree (i.e., the highest risk aggregation degree). Therefore, the abnormal co-occurrence amplitude not only reflects the linkage risk level of the current combination, but also serves as the basis for judging and generating the drug risk superposition amount.

[0106] See also Figure 4 , the vital signs recognition module includes:

[0107] The stage benchmark comparison submodule analyzes the physiological data of cancer patients at the current stage of care based on the drug risk superposition, compares the difference between each physical sign and the standard data of the corresponding stage, determines the key physical signs that cause fluctuations, and obtains the physical sign stage deviation amplitude;

[0108] First, retrieve the patient's current nursing stage information, extract the corresponding reference standard value set according to the stage label, which contains the normal fluctuation range of physiological parameters such as blood pressure, heart rate, body temperature, blood oxygen saturation, and respiratory rate. Then call the physiological data records reported by the patient every day during this stage, compare the actual value of each sign with the stage standard value set item by item, perform difference calculation on each sign parameter, and make bilateral judgments with the upper and lower limit thresholds set for the sign in the reference standard. If any difference exceeds the upper and lower limit intervals, the sign parameter is marked as an offset item, and all sign offset items are aggregated to form an abnormal set, and then the abnormal set is analyzed. The various physical signs are sorted from large to small according to the absolute value of the difference to determine the key signs that cause the deviation. For example, the heart rate of a patient in the recovery stage after chemotherapy is measured to be 110 beats / minute, while the upper limit of the standard for this stage is 100 beats / minute. The difference is 10 beats / minute, which is marked as a deviation. Another example is that the body temperature is measured to be 38.3℃, which exceeds the upper limit of the reference value of 37.5℃ for this stage. The difference is 0.8℃, which is also included in the deviation set. The degree of deviation of each physical sign in the same stage is comprehensively considered, and the weight of its impact on health risks is combined to sort them. Among them, heart rate and body temperature fluctuations rank in the top two, and they are identified as key fluctuation signs. Finally, the deviation amplitude of its physical sign stage is recorded with the maximum deviation amplitude.

[0109] The continuous fluctuation measurement submodule calculates the daily change trend of the sign during continuous monitoring based on the deviation amplitude of the sign stage, compares the daily sign data with the data performance of the previous cycle, analyzes the continuous fluctuation of the sign over time, optimizes the distribution of fluctuation data, and obtains the intensity level of the sign fluctuation;

[0110] Read the historical monitoring data of patients who have been identified as key fluctuating vital signs, construct a time series for each vital sign parameter in the order of daily reporting, calculate the absolute difference between the two adjacent days on each day as the daily fluctuation value, and record the daily mean sequence of the vital sign in the past complete monitoring cycle. Then perform a difference analysis operation on the vital sign value of the current day and the vital sign value of the corresponding day of the previous cycle to determine whether the vital sign data is continuously higher or lower than the similar data of the previous cycle. If the direction of the vital sign difference is consistent for more than three consecutive days, it is marked as a trend fluctuation event. The frequency and fluctuation direction of the trend fluctuation are counted, and then the data of all fluctuation events are sorted into fluctuations. The dynamic values ​​are grouped and the frequency of daily fluctuation values ​​falling into each grouping interval is distributed and statistically analyzed, and a hierarchical judgment is made on the overall fluctuation trend according to the frequency weight. For example, for the patient's respiratory rate data, the data for 5 consecutive days are 20, 22, 24, 23, and 25 times / minute, and the data on the same date in the previous cycle are 18, 20, 19, 20, and 21 times / minute. If the respiratory rate is found to rise continuously and the daily difference is greater than 2, it is judged to be a continuous fluctuation. Subsequently, the fluctuation values ​​are divided into 0-1, 1-3, and 3-5 intervals. It is found that more than 70% of the daily fluctuation values ​​fall into the 3-5 interval. Based on this, the sign is evaluated as the fluctuation level with the highest intensity level.

[0111] The abnormality rate calculation submodule analyzes the physical sign items showing abnormal trends based on the intensity level of physical sign fluctuations, screens the physical sign manifestations that show abnormal changes during the continuous monitoring period, determines the concentration of abnormal distribution and stage baseline manifestations, and calculates the distribution of abnormal data to obtain the physical sign abnormality concentration rate;

[0112] All physical sign items marked as high-intensity fluctuations are reviewed item by item, and the daily data records are used to determine whether there are abnormal changes in the physical sign. The logic for judging abnormal changes is that the actual measured value of the physical sign exceeds the upper or lower reference limit of the stage, or the fluctuation direction is consistent for three consecutive days and the daily change amplitude exceeds the preset single-day deviation threshold. All records of each physical sign item in a complete monitoring cycle are judged item by item and the number of abnormalities is accumulated. The data corresponding to the date of the abnormal occurrence is counted to form an abnormal record sequence. The deviation direction and absolute difference of each data in the abnormal record are then compared with the stage standard deviation to determine whether such abnormalities are statistically clustered on a specific day. The abnormal sign is concentrated in a specific parameter interval or a specific parameter interval. For example, a patient's temperature exceeds 37.8°C for 5 days within 8 days, and 4 of them occur from the 2nd to the 5th day. The abnormal record is determined to be concentrated in time. Subsequently, the temperature deviation value is distributed and analyzed. It is found that the deviation values ​​in the 4 abnormalities are all above 1.0°C, and more than twice the average fluctuation value of 0.5°C in this stage. Based on this, the abnormal sign is recorded as having a strong centralized feature, and the abnormal records are accumulated and counted. The abnormal concentration rate of the sign is output as the percentage of clustered abnormalities in the total abnormal records. For example, there are 5 abnormal temperature records in total, of which 4 are centralized abnormal records, and the calculated result is 80%.

[0113] See also Figure 5 , the action push module includes:

[0114] The fluctuation feature screening submodule analyzes continuous fluctuation records based on the abnormal concentration rate of physical signs, compares the fluctuation frequency of each physical sign in the current nursing stage, screens key physical signs, and determines the prominence of key physical signs in the stage to obtain key physical sign identification indicators;

[0115] Read all the vital sign data items marked as abnormal, sort out their continuous fluctuation records in chronological order, count the number of daily fluctuations of each vital sign in the current nursing stage, form a corresponding relationship table between vital signs and fluctuation frequencies, and then compare the fluctuation frequency of each vital sign item in the table with the fluctuation threshold set in the nursing stage. If the fluctuation frequency of a certain vital sign in the cycle is greater than the set threshold, the vital sign is marked as a frequency-prominent item. For example, when the nursing stage is the recovery period after chemotherapy, the temperature fluctuation threshold is set to 3 times / week. If the patient's temperature record exceeds the standard value more than 3 times, it is regarded as a high-frequency fluctuation sign, and this sign is screened into the candidate set. Then the vital signs in the candidate set are normalized and each item is calculated. The fluctuation amplitude and abnormal concentration rate of the physical signs are compared with the reference value of the corresponding nursing stage. If the ratio exceeds the preset strong deviation coefficient (such as set to 2.0), the physical sign will be further marked as a key physical sign. For example, among the three items of heart rate, respiratory rate, and body temperature, the reference range of heart rate in this stage is 60-100 times / minute. The patient exceeds 120 times / minute five times in a week. It is judged to be a high-frequency and high-amplitude fluctuation. The difference ratio is (120-100) / (100-60)=0.5, which meets the strong deviation standard. The body temperature only slightly exceeds the reference upper limit, and the difference ratio is less than 0.2. It is excluded from the key signs, and the output is the key sign identification index consisting of key signs with high fluctuation frequency and large deviation amplitude.

[0116] The joint abnormality judgment submodule compares the daily fluctuations of blood oxygen saturation and respiratory rate based on key physical sign identification indicators, using the formula:

[0117] ;

[0118] Obtain respiratory abnormality intensity level , compared with the intervention initiation criteria to determine whether the intervention conditions are met and determine the respiratory system intervention event, among which, Represents the fluctuation range of blood oxygen saturation on the same day, Represents the fluctuation range of respiratory rate on the day, Represents the base of the systolic blood pressure fluctuation range on that day, Indicates the number of abnormal physical signs on that day. Indicates the Fluctuation data of respiratory rate, Indicates the Fluctuation data of blood oxygen saturation, Indicates the total number of times blood oxygen and respiratory rate are collected together on that day;

[0119] The intensity level of combined respiratory abnormality refers to a numerical grading indicator based on the fluctuation amplitude of the patient's blood oxygen saturation and respiratory rate on the same day, through a comprehensive analysis of the synchronous changes of the two during the same nursing cycle, as well as factors such as systolic blood pressure fluctuations and the number of abnormal physical signs. This indicator is used to measure the degree of abnormality in the combined fluctuations of blood oxygen and respiratory rate, and reflects the deviation level of the patient's current respiratory system status from the normal standard. The higher the value, the more prominent the risk of synchronous abnormal fluctuations in blood oxygen and respiratory rate. The grade score is calculated first and then compared with the standard threshold range to decide whether to intervene.

[0120] Compare the fluctuations of blood oxygen saturation and respiratory rate on the same day. The monitoring time points are set as 08:00, 12:00, 16:00, and 20:00. The fluctuation amplitudes of blood oxygen saturation recorded at the four time points are 3.2%, 2.5%, 4.1%, and 3.7%, respectively. The fluctuation amplitudes of respiratory rate are 2.8, 3.0, 3.5, and 2.9 times / min, respectively. The base of systolic blood pressure fluctuation is 6.1, 5.8, 6.5, and 6.0 mmHg. The number of abnormal physical signs on the day is 2, 3, 2, and 4, respectively. The pairs of respiratory rate fluctuation data and blood oxygen saturation fluctuation data recorded during the sampling period are (0.6, 0.5) and (0.7, 0.6), respectively. The data of the 16:00 period are used as the core sample. After dimensionless processing of all data, the normalized parameter values ​​are:

[0121] 、 、 、 、 、 、 、 , substitute the above parameters into the formula to calculate the numerator:

[0122] ;

[0123] Denominator:

[0124] ;

[0125] Comprehensive calculation:

[0126] ;

[0127] The results show that at 16:00 on the day of sampling, the respiratory combined abnormality intensity level reflected by the synchronous fluctuations of respiratory rate and blood oxygen saturation was 1.10. If this level value is above the upper limit of the intervention start level range in the judgment criteria, it is considered to trigger the intervention condition, and the corresponding time point of the data will be marked as the starting node of the respiratory system intervention event. The level will subsequently be used to match the task category that requires immediate response in the intervention action list, so as to complete the process of the next link in the nursing scheduling task chain. The formula effectively integrates the fluctuation range, frequency and synchronization of physical signs by introducing multiple normalized physiological data differences and combined calculation methods, making the abnormal intensity level more accurate in assessment and more valuable in identification.

[0128] The nursing action matching submodule is based on respiratory system intervention events, screens event types that meet the intervention conditions, compares the current event with the matching items in the nursing action list, determines the category and response method of the matching action, optimizes the nursing resource allocation structure, and obtains the intervention action instruction sequence number.

[0129] Extract the event content that triggers the intervention and read the event time, physiological parameter type, parameter abnormality and parameter abnormality amplitude. Then read the set nursing action list from the database, expand and match the matching condition fields of each nursing action in the list, and compare the parameter name recorded in the event with the trigger parameter type set in the nursing action one by one to see if it is consistent. Determine the field matching items. If they are consistent, enter the next matching stage. Match the abnormal performance recorded in the event with the abnormal description set in the nursing action. If a match exists, the nursing action is determined to be a candidate. Then, the abnormal amplitude is compared with the response level set for the nursing action. For example, the blood oxygen saturation recorded in the event is lower than 88%, and the oxygen intervention trigger threshold set in the nursing action list is 90%, then the event meets the response conditions of the action, and the nursing action is marked as an executable action. If multiple nursing actions meet the matching conditions, the matching actions are sorted according to the nursing resource usage priority. For example, if the oxygen intervention priority is set to 1 and the doctor notification priority is set to 2, the one with the lower priority number will be executed first. Finally, the action number, intervention category and response method in the matching action are combined and packaged to generate the intervention action instruction sequence number of the event, which is output as an operation instruction. For example, number A301 represents oxygen treatment, the category is "respiratory support", and the response method is "nurse on-site execution", and the intervention action instruction sequence number is obtained.

[0130] See also Figure 6 , the resource scheduling module includes:

[0131] The work time analysis submodule analyzes the remaining work time of nursing staff based on the intervention action instruction sequence number. It then combines their actual task participation time and skill tags to screen staff who can complete nursing tasks and have matching skills. It also compares each staff member's current work time schedule with the time requirements of the upcoming tasks to generate the available nursing resources.

[0132] The execution time required for each intervention action is retrieved as the task time requirement. At the same time, the current remaining working hours data of all on-duty nurses and the scheduled task duration of the day are retrieved to build a working hour usage table for each nurse. The nurse's skill label information is then retrieved to perform a one-to-one matching judgment with the skill requirements required in the intervention action. If a nurse's skill label is marked with the ability required for the intervention, such as "respiratory intervention" or "chemotherapy nursing", the person is judged as a skill matching person. Subsequently, a numerical comparison operation is performed on the matching person's current remaining working hours and the required task duration. If the remaining working hours are greater than or equal to the required task duration, the person is qualified. If the task time requirement is equal to the task time requirement, the person is marked as "assignable", otherwise it is "unassignable". For example, the task time required for the intervention instruction is 1.5 hours, and nurse Zhang has 2 hours of remaining working hours on that day and her skill tag contains "respiratory care", Zhang is judged to be an assignable candidate. On the contrary, Li has only 0.5 hours of remaining working hours or her skill tag does not contain relevant items, so she is excluded. All nursing staff who meet the task time requirement and have matching skills are included in the resource pool. The number of nursing staff in the resource pool and the sum of the time length available for task assignment for each person are counted, and the output is the available nursing resources for the current task.

[0133] The task progress judgment submodule analyzes the completion status of the nursing staff's current tasks based on the available nursing resources, compares the planned task progress, determines whether the actual progress is within the required range, and generates a scheduling adaptation deviation rate;

[0134] Based on the nurse's identity information, the list of ongoing tasks and the planned start and end times of each task are retrieved. The actual operation records for each task are then read, and the completion time and completion progress percentage of each task are extracted. The actual completion status is then compared with the planned progress. If the actual progress value is less than the planned progress value, it is recorded as progress lagging; otherwise, progress is normal or ahead of schedule. The number of tasks with lags and the average lag time for each nurse's current tasks are counted. Combined with the idle time window in the available nursing resources, the scheduling capacity of the nurse is determined to be sufficient to accept new tasks. For example, if nurse A's planned task completion rate should be 70%, but the actual record shows only 50% completion, a 20% lag. Her available nursing time is 1 hour, and the average lag time for the current task completion is 2 hours, her scheduling load is considered too heavy and she is not suitable for the current task allocation. Based on this, a scheduling adaptation record is established for each nurse. A numerical conversion is performed based on the lag degree and resource matching degree. A lag deviation greater than 20% is considered high, and less than 5% is considered low. This forms the scheduling adaptation deviation rate for each nurse.

[0135] The priority sorting submodule analyzes the historical task completion performance of personnel in the nursing response records based on the scheduling adaptation deviation rate, compares their remaining working hours with their historical task participation, determines the task response order of each nursing staff in the actual nursing tasks, and obtains the nursing task priority sequence;

[0136] The nursing response record database is called to extract the historical task completion status of all currently assignable personnel. The response start time and task completion time of their past similar tasks are calculated respectively, and compared item by item with the standard response time and completion deadline set for the task. If the response time in multiple historical tasks exceeds the set threshold, it is recorded as weak response capability. If the response time falls within the standard response time, it is recorded as stable response. Then, the current remaining working hours of each person and the total participation frequency of historical tasks are extracted. Combined with their average historical task completion time, the workload distribution density of the personnel in nursing tasks is calculated. For example, if nurse B has participated in tasks 40 times in the past, with an average response time of 10 minutes, an average task time of 40 minutes, and a current remaining working time of 3 hours, his remaining task capacity is assessed as 3 hours, which can cover 4 to 5 routine tasks. Compared with similar nurses with only 1 hour of remaining working time, his priority ranking is higher. Based on this, a priority ranking list is established, and those with low scheduling adaptation deviation rate, good historical response record, and sufficient current remaining working time are listed as high-priority personnel. The priority sequence of nursing tasks arranged in order of priority is output.

[0137] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. An artificial intelligence-based tumor nursing service management system, characterized in that: The system comprises: The medication detection module analyzes the drug name, dosage, and time of administration based on the daily medication data of cancer patients, compares the difference between the actual and standard dosages, calculates the difference between the medication interval and the drug metabolism interval, determines the risk characteristics, and obtains the deviation medication characteristics; Based on the offset medication characteristics, the combined medication module compares the order of drug combinations and the adverse reaction time, screens drug combinations with overlapping time, analyzes the reference range of laboratory tests, calculates the risk groups, and obtains the drug risk superposition amount; The physical sign recognition module compares the patient care stage with the risk information based on the drug risk superposition, collects daily physiological data, determines the variation range of physical sign parameters and stage standards, and analyzes the cumulative fluctuations to obtain the abnormal physical sign concentration rate; The action push module screens the signs with high fluctuation frequency based on the abnormal concentration rate of the signs, analyzes the combined performance of the signs with blood oxygen and respiratory rate, determines whether it is a respiratory intervention event, identifies the corresponding nursing action, and obtains the intervention action instruction sequence number.

2. The artificial intelligence-based tumor nursing service management system according to claim 1, characterized in that: The deviation medication characteristics include dosage deviation type, medication interval category, and drug warning mark; the drug risk superposition amount includes combined risk level, adverse reaction marker, and laboratory abnormality items; the abnormal physical sign concentration rate includes sensitive physical sign grouping, stage abnormal performance, and continuous fluctuation record; the intervention action instruction sequence number includes task action number, intervention category label, and nursing response method.

3. The artificial intelligence-based tumor nursing service management system according to claim 1, characterized in that: The medication detection module includes: The dose difference determination submodule compares the actual dosage with the standard dosage of each drug based on the daily medication data of cancer patients, determines whether the dosage difference of each drug exceeds the safety reference range, screens drug combinations with abnormal dosages, and obtains the total amount of abnormal dosage differences; The interval comparison submodule calculates the interval between two consecutive doses of the same drug based on the total amount of dosage difference abnormalities, compares the actual interval with the recommended interval of the drug, determines whether there is a situation that exceeds the reference range, and counts the cumulative situation of drug interval abnormalities to obtain the interval time deviation frequency; The offset feature generation submodule extracts the corresponding drug name based on the interval time offset frequency, combines and associates it with the abnormal indicators, obtains the combined abnormal response intensity, extracts the drug combination with high frequency and concentrated numerical amplitude, and obtains the offset medication feature.

4. The artificial intelligence-based tumor nursing service management system according to claim 1, characterized in that: The combined medication module includes: The time series comparison submodule analyzes the drug names and medication time series based on the offset medication characteristics, compares the actual medication sequence of each drug combination with the time sequence of adverse reactions, identifies drug combinations with temporal correlation between actual medication and adverse reactions, determines whether there is a pattern in the drug combination, and obtains a time series coincidence group sequence; The risk linkage submodule analyzes the laboratory test results of the drug combination in the same medication cycle based on the time-series coincidence group sequence, compares the synchronization of the medication time of the drug combination and the abnormality of the test, and counts the coincidence frequency in the same cycle to obtain the frequency of risk coupling items; The abnormal co-occurrence submodule calculates the abnormal co-occurrence amplitude according to the frequency of the risk coupling items, sorts the drug combinations according to the abnormal co-occurrence amplitude, screens the drug combination with the best co-occurrence degree, and obtains the drug risk superposition amount.

5. The artificial intelligence-based tumor nursing service management system according to claim 1, characterized in that: The vital signs recognition module includes: The stage benchmark comparison submodule analyzes the physiological data of the cancer patient at the current care stage based on the drug risk superposition, compares the difference between each physical sign and the standard data of the corresponding stage, determines the physical sign that causes the fluctuation, and obtains the physical sign stage deviation amplitude; The continuous fluctuation measurement submodule calculates the daily change trend of the vital sign during the continuous monitoring process based on the deviation amplitude of the vital sign stage, compares the vital sign data of each day with the data performance of the previous cycle, analyzes the continuous fluctuation of the vital sign over time, optimizes the distribution of the fluctuation data, and obtains the intensity level of the vital sign fluctuation; The abnormality rate calculation submodule analyzes the physical sign items showing abnormal trends based on the level of physical sign fluctuation intensity, screens the physical sign manifestations that show abnormal changes during the continuous monitoring period, determines the concentration of abnormal distribution and stage benchmark manifestations, and statistically analyzes the distribution of abnormal data to obtain the physical sign abnormality concentration rate.

6. The artificial intelligence-based tumor nursing service management system according to claim 1, characterized in that: The action push module includes: The fluctuation feature screening submodule analyzes the continuous fluctuation records based on the abnormal concentration rate of the physical signs, compares the fluctuation frequency of each physical sign in the current nursing stage, screens the key physical signs, and determines the prominence of the key physical signs in the stage to obtain the key physical sign identification index; The combined abnormality judgment submodule compares the daily fluctuations of blood oxygen saturation and respiratory rate based on the key vital sign identification indicators, obtains the intensity level of combined respiratory abnormality, compares it with the intervention initiation standard, determines whether the intervention conditions are met, and determines the respiratory system intervention event; The nursing action matching submodule screens the event types that meet the intervention conditions based on the respiratory system intervention events, compares the current events with the matching items in the nursing action list, determines the category and response method of the matching actions, optimizes the nursing resource allocation structure, and obtains the intervention action instruction sequence number.

7. The artificial intelligence-based tumor nursing service management system according to claim 1, characterized in that: The system further comprises: The resource scheduling module analyzes the remaining working hours, task progress and skill tags of the nursing staff based on the intervention action instruction sequence number, selects and matches the nursing staff, and sorts and allocates priorities based on the task progress and historical efficiency to obtain the nursing task priority sequence; The nursing task priority sequence includes job assignment order, resource allocation number, and scheduling priority label.

8. The artificial intelligence-based tumor nursing service management system according to claim 7, characterized in that: The resource scheduling module includes: The work time analysis submodule analyzes the remaining work time of the nursing staff based on the intervention action instruction sequence number, and screens the staff who can complete the nursing task and have matching skills based on their actual participation time and skill tags. It also compares the current work time schedule of each staff member with the time requirement of the upcoming task to generate the amount of available nursing resources. The task progress judgment submodule analyzes the completion status of the nursing staff's current task based on the available nursing resources, compares the planned task progress, determines whether the actual progress is within the required range, and generates a scheduling adaptation deviation rate; The priority sorting submodule analyzes the historical task completion performance of personnel in the nursing response records based on the scheduling adaptation deviation rate, compares their remaining working hours with their historical task participation, determines the task response order of each nursing staff in the actual nursing tasks, and obtains the nursing task priority sequence.

Images