An infusion monitoring and nursing compliance collaboration method, system, and media

Abstract

The application discloses an infusion monitoring and nursing compliance collaborative method and system and a medium, relates to the field of medical infusion safety and long-term nursing insurance nursing compliance, and solves the problems that the existing infusion monitoring only makes threshold alarm, lacks real-time re-reasoning of drug contraindications and integrates evidence chains of nursing compliance, and improves infusion safety and long-term nursing insurance fund supervision efficiency.

Description

Technical Field

[0001] This invention relates to the technical field of medical infusion safety and long-term care insurance compliance, specifically to a collaborative method, system, and medium for infusion monitoring and nursing compliance. This method monitors infusion and physiological parameters during infusion, performs real-time re-inference on medication contraindications, generates a verifiable chain of treatment time evidence, and maps infusion process data and nursing service data into a long-term care insurance nursing service log for compliance verification and fund review. Background Technology

[0002] In outpatient and inpatient intravenous infusion scenarios, medical institutions need to record the infusion rate and start and end times, and verify the rationality of medication and contraindications to reduce medical errors. Existing infusion monitoring equipment mostly only records the drip rate, provides remaining volume reminders, and matches the user's identity; some solutions issue threshold alarms when physiological parameters exceed limits.

[0003] The above scheme has three shortcomings: First, the alarm is based on a static threshold and does not link real-time physiological deviations with medical orders, test results and drug knowledge graphs to make a definitive re-inference on contraindications and drug interactions, making it difficult to identify risks in advance at the medication regimen level.

[0004] Secondly, the data in the infusion process is scattered and lacks an immutable chain of evidence that is indexed by the patient's identity and binds the person, medication, bed, and time, making it difficult to obtain evidence in medical disputes and nursing quality control.

[0005] Third, with the full implementation of the long-term care insurance system, the infusion and medication processes in home and institutional care need to be mapped into nursing service logs and subject to fund audits. The existing scheme does not integrate infusion monitoring with long-term care insurance compliance verification, making it difficult to meet the rigid requirements of fund intelligent auditing and settlement record keeping. Summary of the Invention

[0006] The technical problem this invention aims to solve is that existing infusion monitoring only performs threshold alarms and identity matching, lacking a collaborative mechanism to drive the deterministic reasoning of contraindications based on real-time physiological deviations and generate a legally verifiable evidence chain and a long-term care insurance compliance evidence chain. This results in delayed identification of infusion safety risks, difficulty in providing evidence, and inability to meet the review requirements of long-term care insurance funds.

[0007] To address the aforementioned technical problems, this invention provides a method, system, and medium for compliant collaboration in infusion monitoring and nursing care.

[0008] During the infusion process, the infusion parameters (drip rate, remaining volume, start and end time) and physiological parameters (body temperature, heart rate, blood oxygen saturation, respiratory rate, etc.) are linked to the same patient identity identifier through the identity verification module.

[0009] When any physiological parameter deviates from the baseline, the edge gateway's processor calls upon medical order information, test information, and drug knowledge graph to perform deterministic re-reasoning (including counterfactual reasoning) on ​​the contraindications, drug interactions, or incompatibilities associated with the deviation in the current medication regimen, thereby obtaining risk level data. Unlike the static incompatibilities check performed during prescription or dispensing, this invention uses the physiological deviation collected in real time during infusion as the reasoning trigger condition, continuously monitoring and re-reasoning during the medication implementation phase.

[0010] Using patient identification as an index, infusion parameters, physiological parameters, further reasoning conclusions, and timestamps are written into an immutable chain of treatment time evidence for use in medical disputes.

[0011] The infusion process data and nursing service data are mapped to structured fields of the long-term care insurance nursing service log, and compared item by item with the service plan or catalog to generate compliance judgment data and fund audit evidence chain.

[0012] The present invention also provides a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the above-described method for coordinated infusion monitoring and nursing compliance.

[0013] This invention is not intended for the diagnosis or treatment of diseases, nor does it apply any therapeutic intervention to the human body; the risk level data and evidence chain output are used only for medical quality control, evidence collection and nursing compliance verification, and do not change the doctor's diagnosis or treatment conclusion. Beneficial effects

[0014] Compared with the prior art, the present invention has the following technical effects:

[0015] First, by triggering the determination of contraindications in the drug knowledge graph through real-time physiological offsets, compared with static threshold alarms, incompatibilities and drug interaction risks can be identified earlier at the medication regimen level, thereby reducing the identification lag and missed detection gaps caused by simple threshold alarms.

[0016] Second, by using patient identification as an index to generate an immutable chain of evidence for treatment time that is bound in four dimensions—person, medicine, bed, and time—this significantly improves the completeness and non-repudiation of evidence collection in medical disputes and nursing quality control compared to scattered records.

[0017] Third, by integrating infusion monitoring with the structured, compliance verification, and fund audit evidence chain of long-term care insurance nursing service logs, the workload of manual verification of nursing service compliance is reduced compared to isolated infusion records, thereby improving the efficiency of long-term care insurance fund supervision and settlement. Attached Figure Description

[0018] Figure 1 is a block diagram of the architecture of the infusion monitoring and nursing compliance collaborative system of the present invention; Figure 2 is a flowchart of the infusion monitoring and nursing compliance collaborative method of the present invention; Figure 3 is a flowchart of the long-term care insurance nursing service log mapping and fund audit evidence chain of the present invention; Figure 4 is a schematic diagram of the hardware structure of the bedside infusion monitoring terminal of the present invention.

[0019] In Figure 2, S1, S2, S3, and S4 represent the four steps of the method of the present invention, namely, data acquisition, physiological deviation detection and deterministic re-inference, generation of evidence chain for treatment time, and generation of evidence chain for long-term care insurance compliance. The four steps correspond to the following embodiments one to four. Detailed Implementation

[0020] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. The following embodiments are exemplary and are used to illustrate the implementation of the present invention, and are not intended to limit the scope of protection of the present invention.

[0021] The order in which the various steps of this invention are described is used to illustrate its implementation process. The steps may partially overlap or be executed in parallel in time, and their order does not constitute a limitation on this invention.

[0022] Example 1: This example illustrates the overall structure of the system of the present invention and the process of collecting infusion parameters and physiological parameters.

[0023] As shown in Figure 1, the system of the present invention includes a bedside infusion monitoring terminal (100), a physiological parameter acquisition module (200), an edge gateway (300), and an evidence chain and compliance module (400).

[0024] The bedside infusion monitoring terminal (100) is deployed at the patient's bed and is equipped with a drip rate sensor, a remaining volume sensor, a bubble sensor and a tubing pressure sensor. It is also equipped with an automatic clamping mechanism and an identity verification unit. The identity verification unit obtains the patient's identity and bed identification through radio frequency identification or barcode reading and generates infusion timing data with timestamps.

[0025] The physiological parameter acquisition module (200) is used to acquire at least one physiological parameter among body temperature, heart rate, blood oxygen saturation, and respiratory rate. Its implementation forms include, but are not limited to, wearable wristbands, bedside multi-parameter monitors, and central monitoring systems. The wearable wristband is an optional implementation that integrates infrared body temperature, photoplethysmography pulse wave, heart rate, blood oxygen, piezoelectric or inertial measurement units, and respiratory sensing, and can be expanded to include ECG patches and positioning units. The aforementioned implementation forms of the physiological parameter acquisition module (200) are all optional specific implementation forms of the physiological parameter acquisition of the present invention and fall within the protection scope of the physiological parameter acquisition module as defined by the claims of the present invention.

[0026] The edge gateway (300) is equipped with a processor, memory and communication interface. The communication interface supports HL7, FHIR and IEEE 11073 standard adaptation and is used to interface with hospital information system, laboratory information system, electronic medical record and drug knowledge graph. The processor of the edge gateway (300) executes the instructions in the memory to realize offset detection, deterministic re-inference and evidence chain writing.

[0027] Example 2: This example illustrates the deterministic re-inference process triggered by physiological deviation (corresponding to step S2 in Figure 2).

[0028] In this embodiment, the edge gateway (300) performs deterministic re-inference, and its inference engine adopts the following hierarchical structure: the input layer receives physiological offset features, medical order information and test information; the knowledge graph layer stores contraindication rules, drug interaction rules and compatibility rules. The knowledge graph is constructed from historical medical orders, drug instructions and adverse reaction reports through entity extraction and relation extraction. The construction process includes corpus annotation, entity alignment and rule confidence assignment. The confidence threshold is set to 0.7 to 0.95 in one example. Those skilled in the art can adjust the value range according to the corpus size and accuracy requirements; the inference layer performs counterfactual inference on the current medication regimen, that is, under the condition that other medications except for the drug and its dosage associated with the offset remain unchanged, it calculates the quantitative estimate of the probability of adverse reaction occurring when the physiological offset exists, and outputs risk level data based on the quantitative estimate; the output layer sends the risk level data back to the responsible nurse terminal and writes it into the evidence chain.

[0029] The "deterministic re-reasoning" referred to in this invention refers to reasoning based on contraindication rules, drug interaction rules, and incompatibilities rules recorded in the drug knowledge graph, combined with counterfactual deduction, in which the reasoning link is interpretable and can be manually verified; its output includes quantitative estimates of the probability of adverse reactions and graded risk level data, which is different from black box probability prediction that relies solely on statistical learning models.

[0030] The counterfactual reasoning algorithm in the aforementioned deterministic re-reasoning includes the following steps: Step a, constructing a counterfactual scenario: keeping other medications unchanged except for the drugs and their dosages associated with the physiological shift, and only using the physiological shift as a conditional input; Step b, querying the knowledge edges and their weights from the associated drugs to each adverse reaction type in the drug knowledge graph, and superimposing a physiological shift correction factor, which is converted from the shift magnitude to the conditional probability increment according to a preset mapping relationship; Step c, calculating the quantitative estimate of the probability of adverse reaction occurrence according to P(adverse reaction) = f(graph path weight, physiological shift correction factor), where f is the path weighted sum or Bayesian network conditional probability calculation; Step d, mapping the quantitative estimate to risk level data according to a preset grading threshold: a quantitative estimate less than 0.2 is judged as low risk, 0.2 to 0.5 is judged as medium risk, and greater than 0.5 is judged as high risk.

[0031] As examples (not exhaustive), of the preset mapping relationship between the physiological offset magnitude and the conditional probability increment: a 10% decrease in respiratory rate from the baseline corresponds to a conditional probability increment of +0.08; a 15% decrease in respiratory rate from the baseline corresponds to +0.12 (i.e., the value used in Example 1); a 20% decrease in respiratory rate from the baseline corresponds to +0.16; early signs of chemotherapy drug extravasation correspond to +0.15 (i.e., the value used in Example 2 below). The above mappings are pre-configured adjustable parameters, not limited to the listed values, used to illustrate the specific source of the offset correction factor values, making the counterfactual reasoning loop reproducible.

[0032] In terms of hardware implementation, the counterfactual reasoning can be executed by a dedicated inference coprocessor, neural network processing unit, or general-purpose processor with instruction acceleration within the edge gateway (300); the hashing on-chain can be implemented by a security encryption module, a trusted platform module, or a blockchain node interface, and the blockchain node interface supports communication with judicial blockchains or notarized evidence storage platforms; the above-mentioned hardware carrier is the physical implementation of the software function on the device side and does not constitute a limitation on the scope of protection.

[0033] The following uses two sets of exemplary input values ​​to illustrate the execution process of steps a to d. The input values ​​and intermediate calculation results are exemplary values ​​set for the convenience of those skilled in the art, and are not clinical test data. Those skilled in the art can select other values ​​for equivalent implementation within the algorithm framework disclosed in this invention.

[0034] Example 1 (Early identification of respiratory depression in patient-controlled analgesia scenarios using pain pumps): A 15% decrease in the patient's respiratory rate from baseline constitutes a physiological shift. The associated medication is an opioid. The graph path weight corresponds to a baseline probability of respiratory depression of 0.20. The physiological shift correction factor is increased by 0.12. The quantitative estimate calculated in step c is 0.32. According to step d, it is mapped to a risk level of "medium". Based on this, the system pushes an early warning of respiratory depression to the responsible nurse's terminal.

[0035] Example 2 (Vascular extravasation monitoring in chemotherapy drug infusion scenarios): Early signs of drug extravasation (local swelling and temperature changes) are detected, constituting a physiological deviation. The associated drug is a vesicant chemotherapy drug. The map path weight corresponds to a baseline probability of extravasation tissue damage of 0.25. The physiological deviation correction factor is increased by 0.15. The quantitative estimate is calculated to be 0.40 in step c. According to step d, it is mapped to a risk level of "medium". Based on this, the system pushes an extravasation treatment alarm and generates an evidence chain record.

[0036] Example 3 This example illustrates the process of generating and judicially verifying the chain of evidence for treatment time (corresponding to step S3 in Figure 2).

[0037] In this embodiment, the edge gateway (300) calculates the hash value of the infusion parameters, physiological parameters, re-inference conclusions and timestamps, writes the hash value into the judicial blockchain or notarized evidence storage platform and retains the original text at the hospital. The treatment time evidence chain is bound in four dimensions with patient identity, drug identity, bed identity and timestamp, so that judicial or administrative agencies can verify the originality.

[0038] Example 4 This example illustrates the process of mapping long-term care insurance nursing service logs and generating evidence chains for fund audit (corresponding to step S4 in Figure 2 and Figure 3).

[0039] In this invention, the infusion process data includes infusion parameters, physiological parameters, and re-inference conclusions, as well as their respective timestamps; the nursing service data includes nursing service items, personnel performing the services, and execution records.

[0040] In this embodiment, the evidence chain and compliance module (400) maps the infusion process data and nursing service data into structured fields of the long-term care insurance nursing service log. The fields include service item code, execution start and end time, execution duration, execution personnel identification and physiological monitoring items. The data is compared item by item with the long-term care insurance service item catalog and service plan to obtain compliance judgment data. Based on the compliance judgment data, a fund audit evidence chain is generated for intelligent auditing and settlement voucher generation of the long-term care insurance fund.

[0041] The metering unit and billing unit are generated based on the aforementioned structured fields (service item code, execution start and end time, execution duration) according to the billing rules of the long-term care insurance smart care equipment or assistive device payment catalog. There is a one-to-one data mapping relationship between the structured fields and the metering and billing.

[0042] Example 5: This example illustrates the lower-level application scenarios of the present invention.

[0043] As a secondary implementation, the present invention is applicable to early identification of respiratory depression in patient-controlled analgesia scenarios using pain pumps, monitoring of extravasation of chemotherapy drugs, remote drug administration monitoring in home-based continuing care at internet hospitals, and remote alarms for preventing extubation and orthostatic hypotension in home-based infusion scenarios for disabled or semi-disabled patients.

[0044] In implementations involving speed adjustment, the system only issues speed adjustment suggestions or alarms to the responsible nurse and does not automatically perform therapeutic speed adjustments; therefore, this invention does not constitute a treatment method. The home-adaptive unit's identification of the risk of orthostatic hypotension based on physiological parameters is an algorithmic warning and does not constitute a disease diagnosis method, nor does it replace physician diagnosis. Data compliance and ethical design

[0045] The data processing of this invention follows the principles of legality, legitimacy, and necessity. The legal basis for processing patient physiological and infusion data is the statutory duty required for the medical institution's professional activities and long-term care insurance management, and the authorization and consent of the patient or guardian are obtained. Data collection is limited to the minimum scope necessary for infusion safety and nursing compliance. Physiological parameters and identity information are de-identified. The original data does not leave the terminal or domain. Reasoning and hashing are only completed within the edge gateway (300). Only necessary summaries are retained at the hospital and on the chain. The reasoning and review rules are not based on discriminatory labels such as race or region, do not output diagnostic conclusions or treatment plans, and do not change the doctor's diagnosis or treatment judgment. The de-identification process involves replacing directly identifiable natural person identity information with a pseudonym for the patient. The patient identity identifier is the pseudonym after the de-identification process, which remains consistent within the same medical institution and can be restored to the real identity through authorized mapping. Therefore, it does not affect the association of the aforementioned parameters, the writing and storage indexed by the patient identity identifier, nor does it affect the verification of the originality of the treatment time evidence chain by judicial or administrative agencies. The de-identification process does not change the data's associativity, but only removes information that can directly identify natural persons.

Claims

1. A method for compliant collaboration between infusion monitoring and nursing care, characterized in that, include: During the infusion process, infusion parameters are collected through the bedside infusion monitoring terminal (100) and physiological parameters are collected through the physiological parameter acquisition module (200). The infusion parameters include drip rate, remaining drug volume, and infusion start and end time. The physiological parameters include at least one of body temperature, heart rate, blood oxygen saturation, and respiratory rate. The infusion parameters and physiological parameters are all associated with the same patient identity and bed position through the identity verification module. Subsequently, when any physiological parameter deviates from its baseline value, the processor of the edge gateway (300) calls the medical order information, test information, and drug knowledge graph to perform deterministic re-reasoning on the contraindications, drug interactions, or incompatibilities associated with the deviation in the current medication regimen, and obtains a re-reasoning conclusion in the form of risk level data; then, using the patient's identity identifier as an index, the infusion parameters, the physiological parameters, the re-reasoning conclusion, and their respective timestamps are written into an immutable treatment time evidence chain, which is used for evidence in medical disputes.

2. A collaborative system for infusion monitoring and nursing compliance, characterized in that, include: A bedside infusion monitoring terminal (100) is configured to collect infusion parameters and obtain patient identification and bed identification. The infusion parameters include drip rate, remaining drug volume, and infusion start and end time. The physiological parameter acquisition module (200) is configured to acquire at least one physiological parameter among body temperature, heart rate, blood oxygen saturation, and respiratory rate. An edge gateway (300) includes a processor and a memory. The memory stores computer instructions. When the processor executes the instructions, it performs the following: when any physiological parameter deviates from its baseline value, it calls up medical order information, test information, and a drug knowledge graph to perform deterministic re-reasoning on the contraindications, drug interactions, or incompatibilities of the drugs associated with the deviation in the current medication regimen, and obtains a re-reasoning conclusion in the form of risk level data; and writes the infusion parameters, the physiological parameters, the re-reasoning conclusion, and the corresponding timestamp into an immutable treatment time evidence chain using the patient's identity identifier as an index. The evidence chain and compliance module (400) is configured to store and maintain the treatment time evidence chain using the patient identity identifier as an index; wherein the infusion parameters, the physiological parameters and the re-inference conclusion are all associated with the same patient identity identifier.

3. A computer-readable storage medium storing computer instructions thereon, characterized in that, When the computer instructions are executed by the processor, the following is achieved: during the infusion process, infusion parameters are collected through the bedside infusion monitoring terminal (100), and physiological parameters are collected through the physiological parameter acquisition module (200). The infusion parameters include drip rate, remaining drug volume, and infusion start and end time. The physiological parameters include at least one of body temperature, heart rate, blood oxygen saturation, and respiratory rate. The infusion parameters and physiological parameters are both associated with the same patient identity identifier and bed identifier through the identity verification module. Subsequently, when any physiological parameter deviates from its baseline value, the processor of the edge gateway (300) calls the medical order information, test information, and drug knowledge graph to perform deterministic re-reasoning on the contraindications, drug interactions, or incompatibilities associated with the deviation in the current medication regimen, and obtains a re-reasoning conclusion in the form of risk level data; then, using the patient's identity identifier as an index, the infusion parameters, the physiological parameters, the re-reasoning conclusion, and their respective timestamps are written into an immutable treatment time evidence chain, which is used for evidence in medical disputes.

4. The method according to claim 1, characterized in that, The deterministic re-reasoning specifically includes: constructing counterfactual reasoning for the current medication regimen based on the contraindication rules, drug interaction rules, and incompatibilities rules recorded in the drug knowledge graph. That is, while keeping other medications unchanged except for the drugs and their dosages associated with the deviation, the quantitative estimate of the probability of adverse reactions occurring when the physiological deviation exists is calculated, and risk level data is output based on the quantitative estimate.

5. The method according to claim 1, characterized in that, The process of writing an immutable chain of evidence for treatment time specifically includes: calculating a hash value for the infusion parameters, physiological parameters, re-inference conclusions, and timestamps using a preset hash algorithm; writing the hash value into a judicial blockchain or notarized evidence storage platform; and storing the original text at the hospital. The chain of evidence for treatment time is bound together with the patient's identity identifier, drug identifier, bed identifier, and timestamp in four dimensions for verification of originality by judicial or administrative agencies.

6. The system according to claim 2, characterized in that, The edge gateway (300) is also configured to: when the same patient identity corresponds to multiple parallel infusions, perform incompatibility and timing conflict detection across multiple drug solutions, identify physicochemical incompatibility and drug administration timing conflicts between multiple drugs within the same time window, and write the conflict results into the treatment time evidence chain.

7. The system according to claim 2, characterized in that, It also includes a home adaptation unit configured to: detect the anti-removal status of unplanned tubing removal for disabled or semi-disabled patients receiving home intravenous infusions, monitor changes in body position to identify the risk of orthostatic hypotension, and send a remote alarm to the caregiver or responsible nurse terminal when an abnormality is detected; the home adaptation unit communicates with the nurse station or long-term care insurance agency platform through the edge gateway (300).

8. The method according to claim 1, characterized in that, The physiological parameter acquisition module (200) includes at least one of a wearable wristband, a bedside multi-parameter monitor, and a central monitoring system; the edge gateway (300) performs multimodal temporal alignment of physiological parameters and infusion parameters from multiple sources according to timestamps, and then feeds in the drug knowledge graph to perform the deterministic re-inference. The multimodal temporal alignment is completed on the edge side and does not change the diagnosis or treatment conclusion.

9. The method, system, or medium according to claim 1, 2, or 3, characterized in that: The method further includes: mapping the infusion parameters, the physiological parameters and the re-inference conclusion to structured fields of the long-term care insurance nursing service log, wherein the structured fields include service item code, execution start and end time, execution duration, execution personnel identification and physiological monitoring items, and comparing the structured fields with the pre-configured long-term care insurance service item catalog and service plan item by item to obtain compliance judgment data and generate a fund audit evidence chain; The system is further configured to: map the infusion parameters, the physiological parameters, and the re-inference conclusions to the structured fields, and compare the structured fields with the long-term care insurance service item catalog and service plan item by item to obtain the compliance judgment data and generate the fund audit evidence chain; The computer instructions in the medium also cause the processor to execute: mapping the infusion parameters, the physiological parameters, and the re-inference conclusion to the structured fields, and comparing the structured fields with the long-term care insurance service item catalog and service plan item by item to obtain the compliance determination data and generate the fund audit evidence chain; The aforementioned fund audit evidence chain is used for intelligent auditing and settlement voucher generation of long-term care insurance funds.

10. The method, system, or medium according to claim 1, 2, or 3, characterized in that: The method further includes: generating a metering unit and a billing unit from the infusion parameters, the physiological parameters, the re-inference conclusions, and the nursing service data according to the payment catalog of long-term care insurance smart care devices or assistive devices; The system is also configured to generate the metering unit and the billing unit based on the infusion parameters, the physiological parameters, the re-inference conclusions, and the nursing service data, according to the payment catalog. The computer instructions in the medium also cause the processor to execute: generating the metering unit and the billing unit based on the infusion parameters, the physiological parameters, the re-inference conclusions, and the nursing service data, according to the payment catalog; The metering unit and billing unit are used to submit payment applications to the long-term care insurance agency.

Images