Emergency waiting patient condition dynamic assessment method based on mobile terminal

Abstract

The application belongs to the technical field of medical information processing, and discloses an emergency patient condition dynamic evaluation method based on a mobile terminal. A patient identification, a priority weight and a condition baseline feature vector are encapsulated in a queuing two-dimensional code by acquiring patient triage results and baseline sign data. Queuing information and the baseline feature vector are obtained by scanning the code with a mobile terminal, and qualitative clinical observation information collected by nurses is converted into a quantitative offset value based on a clinical observation semantic knowledge base. The quantitative offset value and the baseline feature vector are subjected to offset amount operation, and a condition deterioration risk index is calculated in combination with queuing duration, triage level and queuing environment. The just-visit priority weight is dynamically adjusted according to the condition deterioration risk index, and a queuing sequence is reconstructed, which is reordered in an emergency information system and a change notification is pushed to related patients and clinics. The application provides high-quality technical support for emergency queuing management and intelligent triage.

Description

Technical Field

[0001] This invention relates to the field of medical information processing technology, and more specifically, to a method for dynamic assessment of the condition of emergency patients waiting for treatment based on mobile terminals. Background Technology

[0002] Existing emergency waiting management technologies face multi-dimensional technical bottlenecks in the dynamic assessment of complex conditions, particularly in the semantic ambiguity and quantification mismatch encountered when fusion of multi-source heterogeneous clinical data to process qualitative observation information. As the types of complaints from emergency patients increase, the granularity, subjectivity, and expression standards of different clinical observation dimensions vary significantly. Traditional fixed scoring scales or experience-based judgment strategies cannot adapt to the differences in observer capabilities and atypical presentations in actual diagnosis and treatment. In complex waiting scenarios, when nurses lack clinical experience or patients present with hidden symptoms, the system still mechanically assigns scores according to standardized triage scales, leading to significant deviations in the assessment of disease severity. To maintain basic triage efficiency, existing processes are forced to rely on a single objective indicator (usually vital signs data), artificially ignoring the complementary value of subjective clinical observations. This simplification strategy results in the omission of significant disease characteristics, especially key clinical manifestations such as ashen complexion (early signs of circulatory failure), confusion (signals of central nervous system damage), and abnormal respiratory rhythms (indicating metabolic disorders) that cannot be quantified and incorporated into the assessment system. Emergency physicians often point out that the triage system appears standardized but lacks individualized targeting.

[0003] Meanwhile, existing methods generally employ a single, static assessment framework, failing to consider both the initial baseline state and the cross-temporal information of disease evolution during the waiting period. Traditional triage processes perform an assessment only upon patient arrival and maintain a fixed queuing order, either focusing on initial symptoms while ignoring subsequent deterioration, or relying on patients actively seeking help and passively responding to changes. In particular, the risk of acute disease deterioration occurring within minutes and the cumulative risk of waiting time occurring within hours are intertwined in dynamic waiting queues. Existing rounds use a fixed-cycle, extensive observation approach, which, while controlling nurses' workload, misses critical time windows for observing disease turning points. Although some systems have introduced timed reassessment mechanisms, the lack of quantitative measurement tools for disease change trends means that reassessment results remain isolated snapshots of the state, failing to establish correlation analysis between pre- and post-assessments, making it difficult to identify progressive deterioration in a timely manner.

[0004] In view of this, the present invention proposes a method for dynamic assessment of the condition of emergency patients waiting for treatment based on mobile terminals to solve the above problems. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution: Mobile-based methods for dynamic assessment of the condition of emergency waiting patients include: The system obtains the triage results and baseline vital signs data of patients after the initial triage and generates a waiting slip based on them. The waiting slip is printed with a waiting QR code, which contains the patient's unique identifier, the initial triage timestamp, the current consultation priority weight, and the baseline feature vector of the patient's condition. Nurses periodically patrol the waiting area using mobile devices and obtain the corresponding patient's queuing information and baseline feature vector of the patient's condition by scanning the corresponding patient's waiting QR code; and simultaneously obtain the corresponding patient's complete baseline assessment record in the emergency information system. Nurses collect qualitative clinical observation information of target waiting patients through the clinical data collection interface on mobile terminals, and convert the targeted clinical observation information into quantitative offset values ​​through a preset clinical observation semantic knowledge base. The quantized offset value is offset by the corresponding dimensional component of the baseline feature vector of the patient's condition, and the patient's condition deterioration risk index is calculated by combining the patient's waiting time, initial triage level and waiting environment. The initial priority weight of the target patient is dynamically adjusted based on the risk index of disease deterioration, priority adjustment suggestions are generated, and it is determined whether to trigger queuing information reconstruction based on the suggestions; if triggered, the initial priority weight and disease baseline feature vector of the target waiting patient are updated. The updated initial patient priority weights are synchronized to the emergency information system. Based on the emergency information system, all patients waiting in the current queue are reordered, and a sequence change notification is sent to the relevant patients and clinics.

[0006] The technical effects and advantages of this invention, which uses a mobile terminal-based method for dynamic assessment of the condition of emergency waiting patients, are as follows: This invention scientifically integrates multi-dimensional heterogeneous clinical observation data to ensure that information from facial color, respiratory status, level of consciousness, pain manifestations, and limb movements truly contributes to the quantitative representation of disease characteristics. This significantly improves the accuracy of identifying changes in patient conditions during the waiting period and the precision of priority adjustment, enhancing the safety of emergency waiting management and the coordination of dynamic adjustments to the patient sequence. Furthermore, in practical applications, even in complex and ever-changing atypical disease evolution scenarios (such as insidious deterioration, interference from multiple underlying diseases, and unclear symptom expression), it maintains the integrity of cross-temporal disease characteristics from the initial baseline state to the entire waiting process, eliminating the need to rely entirely on objective vital sign measurements or nurses' personal experience, making priority adjustment decisions more intelligent and precise. In particular, the combination of dual-dimensional offset calculation and multi-factor risk fusion modeling overcomes the limitations of traditional single-session static triage, eliminating safety hazards such as missed diagnoses of deteriorating conditions during the waiting period and unreasonable rigid queuing orders, providing reliable technical support for the intelligent operation and maintenance of emergency departments. Attached Figure Description

[0007] Figure 1 This is a schematic diagram of the method for dynamic assessment of the condition of emergency patients waiting for treatment based on mobile terminals according to the present invention. Detailed Implementation

[0008] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0009] Please see Figure 1 As shown in this embodiment, the method for dynamic assessment of the condition of emergency waiting patients based on mobile terminals includes: The system acquires the triage results and baseline vital signs data of patients after initial triage, and generates a waiting slip based on this data. The waiting slip contains a QR code encapsulating the patient's unique identifier, the initial triage timestamp, the current priority weight, and the baseline feature vector of the patient's condition. The triage results include the patient's initial triage level, chief symptom description, and preliminary condition assessment, obtained through a systematic evaluation of the patient by the triage nurse or on-duty physician according to emergency triage standards. The patient's baseline vital signs data includes vital signs parameters such as body temperature, heart rate, respiratory rate, blood pressure, blood oxygen saturation, and pain score, collected by the triage nurse during the initial assessment using monitoring equipment. The waiting QR code encapsulates this information in a structured coding manner, providing each waiting patient with a digital medical record that can be quickly parsed by mobile devices, providing a data foundation for subsequent rounds and assessments.

[0010] Nurses periodically patrol the waiting area using mobile devices, scanning the QR codes corresponding to specific patients to obtain their queueing information and baseline feature vectors of their conditions. Simultaneously, they access the patient's complete baseline assessment record within the emergency information system. Periodic patrols are a standardized procedure for nurses to proactively observe and confirm the condition of patients in the waiting area. The patrol cycle is dynamically determined based on the number of patients in the waiting area and the overall severity of their conditions. The mobile device carries a dedicated patrol assessment application with a built-in camera scanning module. Nurses can retrieve structured data such as patient identity information, current queue position, initial triage level, and baseline feature vectors of their conditions by scanning the QR codes. Simultaneously, the mobile device communicates with the emergency information system in real-time via a wireless network, synchronously acquiring the patient's complete baseline assessment record. This record is a detailed assessment document entered into the emergency information system during the initial triage, containing a textual description of the patient's chief complaint, complete vital sign measurements, past medical history, medication history, allergy history, and the nurse's qualitative observation records. This synchronous acquisition is achieved through a real-time communication interface between the mobile device and the emergency information system, using the secure HTTPS protocol for data transmission. Complete baseline assessment records provide nurses with a panoramic view of the patient's condition, avoiding judgment bias caused by fragmented information. This is especially helpful for patients with underlying diseases or special physical conditions, as complete records help nurses understand the clinical significance of current observations.

[0011] Nurses collect qualitative clinical observation information from target patients waiting for treatment using a mobile terminal's clinical data collection interface. This qualitative information is then converted into quantitative offset values ​​using a pre-defined clinical observation semantic knowledge base. Qualitative clinical observation information consists of non-quantitative descriptions of the patient's condition obtained by nurses through visual observation, brief inquiries, and clinical experience. It covers five dimensions: facial color, respiratory status, level of consciousness, pain presentation, and limb movement. The clinical data collection interface presents a structured term selection method. Nurses do not need to manually enter text; they only need to select the term that best matches the patient's current state from a pre-defined list of descriptive terms for each clinical observation dimension. The clinical observation semantic knowledge base pre-constructs a mapping relationship between descriptive terms and quantitative values ​​for each clinical observation dimension. A semantic quantification algorithm converts the qualitative descriptive terms selected by the nurses into clinically meaningful numerical scores, forming quantitative offset values, thus achieving a systematic transformation from subjective observation to objective quantification.

[0012] The quantitative offset value is offset by the corresponding dimensional components of the baseline feature vector of the patient's condition. This offset is then combined with the patient's waiting time, initial triage level, and waiting environment to calculate the patient's disease deterioration risk index. The offset calculation uses vectorized difference and distance metrics to perform multi-level quantitative analysis of the differences between the current observation state and the initial baseline state, as well as previous rounds of observation records, resulting in a disease change vector reflecting the direction and magnitude of disease changes. The disease deterioration risk index comprehensively considers multiple factors, including the intensity of disease changes, whether the waiting time exceeds a safe threshold, the crowding level of the waiting environment, and the burden on medical resources. It is calculated using a multi-factor risk fusion function, comprehensively reflecting the overall risk level of disease deterioration during the waiting period.

[0013] The system dynamically adjusts the initial priority weights of target patients based on their disease deterioration risk index, generating priority adjustment suggestions and determining whether queuing information reconstruction is triggered. If triggered, the initial priority weights and baseline disease feature vectors of the target waiting patients are updated. The dynamic adjustment process determines the incremental adjustment range of priority weights based on the risk range of the disease deterioration risk index. When the adjustment range exceeds a preset level change threshold, the system generates priority adjustment suggestions and presents them to nurses on mobile terminals. Nurses review and confirm these suggestions before executing the priority update. Priority adjustment suggestions include information such as the pre-adjustment level, post-adjustment level, dominant deterioration dimension, and suggested interventions, ensuring the adjustment process is transparent and traceable. The trigger condition for queuing information reconstruction is when the change in priority weights reaches a level sufficient to affect the queuing order. Once triggered, the system automatically updates the patient's priority weights and baseline disease feature vectors, providing an updated assessment benchmark for subsequent rounds.

[0014] The updated initial priority weights are synchronized to the emergency information system. Based on this, all waiting patients in the current queue are reordered, and sequence change notifications are sent to relevant patients and examination rooms. The synchronization process utilizes a secure communication channel between mobile terminals and the emergency information system. Upon receiving the updated priority weights, the emergency information system extracts the latest priority weights for all waiting patients, performs a global sorting based on these weights, and optimizes the sequence by considering waiting time and examination room resource allocation. After sorting, the system automatically identifies patients whose sequence positions have significantly changed and simultaneously sends sequence change notifications through the waiting area display screen, mobile terminal push notifications, and the examination room's call system. This ensures that patients, nurses, and attending physicians are promptly informed of the sorting adjustments, maintaining orderly operation of the waiting area.

[0015] In an embodiment of the present invention, the process of generating the waiting QR code includes: Upon initial triage, basic triage data is collected, and basic patient information and medical record summaries are extracted from the emergency information system. Basic triage data comprises structured assessment results generated during the initial triage process, including the initial triage level (Level 1: Critical, Level 2: Severe, Level 3: Emergency, Level 4: Non-Emergency), chief symptoms, onset time, pain score, and preliminary treatment recommendations. Basic information includes the patient's name, age, gender, and contact number. The medical record summary is extracted from the electronic medical record module of the emergency information system and contains key medical information such as past medical history, chronic disease diagnoses, long-term medication lists, and drug allergy records. This data collection provides a complete data source for the information encapsulation of the waiting area QR code.

[0016] An encrypted patient identification code is generated based on the patient's unique identifier. The unique patient identifier is a globally unique number assigned to each patient by the emergency information system, typically using a visitor registration number or appointment sequence number. The encryption process uses a symmetric encryption algorithm (such as AES-128) to encode the unique patient identifier, generating a fixed-length ciphertext string as the patient identification code. The purpose of encryption is to prevent unauthorized scanning of the waiting area QR code by others to directly obtain patient identity information, thus protecting patient privacy. Simultaneously, the encryption key is stored only in the emergency information system and authorized mobile terminals, ensuring that only authenticated nurses can decrypt and access the complete patient information after scanning the code.

[0017] The vital signs measured during initial triage are vectorized and encoded to generate a baseline feature vector for the patient's condition. Vectorization encoding is the process of converting various vital sign values ​​and initial clinical observation and assessment results into a standardized multidimensional numerical vector. The encoding process first normalizes the vital signs, mapping body temperature, heart rate, respiratory rate, systolic blood pressure, diastolic blood pressure, and blood oxygen saturation to the [0, 1] interval, based on the clinical normal range and extreme abnormal range of each parameter. Then, the assessment results of the nurse's observation of the patient's complexion, respiratory status, level of consciousness, pain manifestation, and limb movement during the initial triage are semantically quantified and converted into baseline quantified values ​​for the corresponding dimensions. Finally, the normalized vital sign values ​​and the baseline quantified values ​​of each clinical observation dimension are arranged in a fixed order and combined to form the baseline feature vector for the patient's condition. The baseline feature vector for the patient's condition fully records the patient's comprehensive condition at the time of initial triage and serves as a reference benchmark for offset calculations in subsequent rounds and assessments.

[0018] The chief complaints recorded by nurses during initial triage undergo text structuring, extracting symptom keywords and mapping them to symptom feature codes. Text structuring is the process of converting free-text descriptions of chief complaints into machine-readable coded forms. The process first involves word segmentation and keyword extraction of the chief complaint text, identifying core semantic elements such as symptom names, location descriptions, durations, and accompanying symptoms. Then, based on existing medical domain dictionaries, the extracted symptom keywords are mapped to standardized symptom codes. The symptom code dictionary is hierarchically organized according to human systems and symptom types, covering standardized expressions of common emergency symptoms. Finally, the symptom codes are combined to form a symptom feature code sequence. Symptom feature coding enables mobile terminals to quickly identify the patient's primary reason for seeking medical attention during subsequent scanning, assisting nurses in conducting targeted clinical observations during rounds and assessments.

[0019] Based on the patient's initial triage level, the corresponding initial priority weight value is retrieved from a pre-defined priority weight mapping table. This initial priority weight value is then normalized according to the overall distribution of the current waiting queue. The priority weight mapping table predefines the baseline weight value for each triage level; for example, Level 1 (critical) corresponds to a baseline weight of 100, Level 2 (severe) to 75, Level 3 (emergency) to 50, and Level 4 (non-emergency) to 25. Queue normalization is a process of fine-tuning the baseline weight value based on the distribution of patients at each level in the current waiting queue. When there are too many patients at a certain level, fine-grained weight differentiation is applied among patients of the same level based on their arrival order, ensuring fairness in the ordering of patients within the same level. The normalized initial priority weight value maintains the priority order between levels while also achieving a reasonable arrival order within the same level.

[0020] The patient identification code, initial triage timestamp, initial priority weight value, baseline feature vector of the patient's condition, and symptom feature code are encapsulated according to a predetermined data structure to generate a QR code data packet, with a reserved dynamic update field within the QR code data packet. The predetermined data structure uses a compact binary encoding format, comprising two parts: a fixed field area and a dynamic field area. The fixed field area stores information that remains unchanged during the waiting process, such as the patient identification code, initial triage timestamp, and symptom feature code. The dynamic field area stores information that may be updated after rounds of assessments, such as the current priority weight and baseline feature vector of the patient's condition. The reserved dynamic update field is used to record additional information such as the timestamp of the most recent round of assessment, the assessing nurse's identification, and the priority adjustment flag. A checksum is also embedded in the data packet to verify data integrity during mobile terminal scanning and parsing, preventing data parsing errors caused by physical damage to the QR code.

[0021] The QR code data package is encoded to generate a QR code, which is then printed on the patient's queuing pass. The queuing pass also includes the patient's name, estimated waiting time, and rounds assessment information. The QR code encoding uses the standard QR Code encoding method, selecting an appropriate error correction level (typically level M, with an error correction rate of approximately 15%) to balance data capacity and tolerance for physical damage. The printed queuing pass presents the patient's basic information and waiting instructions in a clear and legible manner. The rounds assessment information includes notices such as "Please keep the waiting QR code clearly visible," "Nurses will periodically scan the code to assess your condition," and "Please inform us immediately if you feel unwell," guiding the patient to cooperate with the rounds assessment process.

[0022] When a patient's priority weight is updated after a round of assessments, a QR code update request is sent to the emergency information system via a mobile terminal. This request includes the patient's identification code and the updated priority weight value. After verifying the validity of the update request, the emergency information system generates a new QR code data packet and returns it to the mobile terminal. The mobile terminal updates the patient's QR code data in its local cache and displays the updated information synchronously upon the next scan. The validity verification of the update request includes: verifying whether the request sender is an authorized nurse account, verifying the existence of the patient's identification code in the current waiting queue, and verifying the consistency between the timestamp of the update request and the time of the most recent round of assessments. Since the QR code on the paper queuing slip cannot be physically rewritten, an "electronic overlay" mechanism is used: after scanning the paper QR code, the mobile terminal first reads the patient's identification code, then queries the latest data version of the patient from the local cache or the emergency information system. The latest version data is used as the basis for subsequent assessments, ensuring that the old data on the paper slip does not affect the accuracy of the assessment.

[0023] In an embodiment of the present invention, the process of obtaining the quantization offset value includes: A pre-constructed semantic knowledge base for clinical observation was developed, comprising baseline status terms and degree-of-change terms for five clinical observation dimensions: facial color, respiratory status, level of consciousness, pain presentation, and limb movement. This semantic knowledge base is a structured medical terminology-numerical mapping system, constructed based on emergency medicine expert consensus and clinical observation assessment standards. The facial complexion observation dimension includes terms such as rosy complexion (baseline state) and variations in color, including pale, ashen, cyanotic, and bluish-purple complexion; the respiratory status dimension includes terms such as stable breathing (baseline state) and variations in color, including slightly rapid breathing, rapid breathing, severe dyspnea, and weak and irregular breathing; the consciousness status dimension includes terms such as clear consciousness (baseline state) and variations in color, including drowsiness, confusion, stupor, coma, restlessness, agitation, and delirium; the pain manifestation dimension includes terms such as no pain (baseline state) and variations in color, including mild pain, moderate pain, severe pain, and extreme pain; the limb movement dimension includes terms such as free movement (baseline state) and variations in color, including slightly limited movement, significantly limited movement, limb weakness, and limb paralysis. The terminology within each clinical observation dimension is organized in ascending order of clinical severity, providing a structured basis for subsequent semantic distance calibration.

[0024] A zero-benchmark quantification value was assigned to the baseline status term, and semantic distance was simultaneously calibrated for each degree of change term according to the direction and severity of disease progression, resulting in a semantic distance benchmark value. The zero-benchmark quantification value indicates that the baseline status term represents the normal reference state for this clinical observation dimension, and the quantification values ​​of all degree of change terms are calibrated starting from this value. Semantic distance calibration, through a systematic process of rank assignment, spacing verification, and cumulative reconstruction, transforms qualitative descriptions of degree of change into quantitative numerical representations. The calibrated semantic distance benchmark value reflects the clinical severity of each degree of change term's deviation from the baseline state. The semantic distance benchmark value is normalized to the interval [0, 10], where 0 corresponds to the baseline normal state and 10 corresponds to the most severe deterioration state in this clinical observation dimension.

[0025] The semantic distance benchmark is adjusted based on patient age and underlying disease information to obtain the specific semantic quantification coefficient for each patient. The specific semantic quantification coefficient is a quantification result after personalized adjustment of the general semantic distance benchmark for individual patient characteristics, reflecting the differences in the clinical significance of the same symptoms in different patient groups. The adjustment process first determines the age sensitivity coefficient based on patient age. The age sensitivity coefficient for elderly patients (over 65 years old) and children (under 14 years old) is set between 1.2 and 1.5 (the specific value is determined according to the age group subdivision), reflecting the physiological vulnerability of these two age groups to changes in disease condition; the age sensitivity coefficient for adult patients (14 to 65 years old) is set to 1.0. Then, a disease correction coefficient is determined based on the patient's underlying disease information. For example, the disease correction coefficient for the facial color observation dimension and respiratory status dimension is increased to 1.2 for patients with cardiovascular disease, and the disease correction coefficient for the respiratory status dimension is increased to 1.3 for patients with chronic respiratory diseases. The formula for calculating the specific semantic quantification coefficient is: ,in, For specific semantic quantization coefficients, As the semantic distance benchmark, Age sensitivity coefficient represents the correction coefficient for the k-th underlying disease. The corrected specific semantic quantification coefficient enables the same descriptive term to produce differentiated quantification results in different patients, improving the clinical accuracy of the assessment.

[0026] The specific semantic quantification coefficients corresponding to the qualitative descriptive terms selected by the nurse based on the clinical data collection interface are obtained, and then the difference is calculated between these coefficients and the corresponding dimensions in the patient's baseline feature vector to obtain the quantification offset value for that dimension. The difference calculation is the core computational step for quantifying the degree of change between the current clinical observation state and the initial baseline state. The calculation process first determines the term of change selected by the nurse in a certain clinical observation dimension, and obtains the specific semantic quantification coefficient of this term after patient-specific correction as the current observation quantification value; then, the baseline component value corresponding to that dimension is extracted from the baseline feature vector; finally, the difference between the two is calculated to obtain the quantification offset value for that dimension; a positive quantification offset value indicates that the condition in that dimension has worsened compared to the initial triage, a quantification offset value of zero indicates no change in that dimension, and a negative quantification offset value indicates that the condition has improved. The quantification offset values ​​of each clinical observation dimension collectively constitute the offset information for the current rounds assessment.

[0027] The process involves acquiring the quantitative offset values ​​corresponding to different clinical observation dimensions within the same observation and performing cross-dimensional consistency checks. Based on the consistency check results, it's determined whether to provide feedback to nurses with conflicting information and request confirmation or correction. Cross-dimensional consistency checks are a quality control mechanism for detecting the logical rationality of observation data, identifying inconsistencies based on clinical experience rules and physiological correlation patterns. The check process first extracts the quantitative offset values ​​of all dimensions within the same observation to construct the current observation vector. Then, it applies preset consistency rules for checking. These rules include: dyspnea (respiratory state offset value > 5) is usually accompanied by abnormal facial color (facial color observation offset value > 3) or altered consciousness (consciousness state offset value > 2). If respiration is severely abnormal but facial color and consciousness are normal, it is marked as suspicious inconsistency; severe pain (pain manifestation offset value > 7) is usually accompanied by pallor, sweating, or rapid breathing. If pain is severe but other dimensions are completely normal, it is marked as suspicious inconsistency; the combination of altered consciousness (consciousness state offset value > 6) and completely normal limb movement is also a low-probability situation. The verification algorithm calculates the correlation matrix of the offset values ​​of each dimension and identifies combinations that deviate significantly from the expected physiological correlation pattern. When an inconsistency is detected, a contradictory prompt message pops up on the mobile terminal, highlighting the suspicious dimension combination and providing two processing options: the nurse confirms that the observation is correct (such as the existence of atypical manifestations), records the confirmation mark, and continues processing; or the nurse re-examines and corrects the selected terms, and then returns to the data collection interface to allow modification. This interactive verification mechanism effectively reduces human input errors and observation omissions, improving data quality and assessment reliability.

[0028] Once the consistency check passes, a quantitative offset value is formed based on the combination of quantitative offset values ​​corresponding to different clinical observation dimensions. The quantitative offset value is the final output result of integrating the quantitative offset values ​​of the five clinical observation dimensions into a unified scoring structure, represented as a five-dimensional vector. Each component corresponds to the quantitative offset value of the dimensions of facial color observation, respiratory status, consciousness status, pain manifestation, and limb movement. The quantitative offset value completely records the direction and magnitude of the patient's deviation from the baseline state in each clinical observation dimension during this round of assessment, providing multi-dimensional quantitative input for subsequent offset calculation and the calculation of the disease deterioration risk index.

[0029] In embodiments of the present invention, the process of semantic distance calibration includes: For each clinical observation dimension, the baseline state terminology under that dimension is set as the semantic origin. All terms related to the degree of change under that dimension are arranged according to the progressive direction of clinical deterioration, constructing a single-dimensional semantically ordered terminology sequence for that dimension. The semantic origin is the zero-reference point for semantic distance quantification, corresponding to the normal or baseline state under that dimension. The arrangement of the degree of change terms strictly follows the progressive relationship of clinical severity, from slight changes to severe deterioration, forming a monotonically increasing ordered sequence. For example, the single-dimensional semantically ordered terminology sequence for the facial complexion observation dimension is: rosy complexion (semantic origin) – pale complexion – ashen complexion – cyanotic complexion – cyanotic complexion; the sequence for the pain manifestation dimension is: no pain (semantic origin) – mild pain – moderate pain – severe pain – extreme pain. The construction of the ordered sequence provides a definite arrangement basis for subsequent hierarchical scoring and distance analysis.

[0030] Based on pre-defined clinical severity evaluation criteria, initial severity level scores are assigned to each term in a single-dimensional semantically ordered sequence, resulting in an initial severity level value for each term. The clinical severity evaluation criteria are developed based on expert consensus on emergency triage and clinical nursing assessment scales, reflecting the severity level of each descriptive term in clinical practice. The scoring process is determined by an emergency medicine expert team through multiple rounds of review using the Delphi method, ensuring that the level values ​​for each term have a basis in clinical consensus. For example, in the facial color observation dimension, pallor is assigned an initial severity level value of 2, ashen complexion is assigned 3, cyanosis is assigned 6, and cyanosis is assigned 9. The differences in level values ​​reflect the non-uniform distribution of different facial color abnormalities in clinical severity. The initial severity level values ​​serve as the raw input for subsequent semantic distance calculations.

[0031] For a single-dimensional semantically ordered sequence of terms, the difference between the initial severity levels of two adjacent terms is calculated to obtain the initial semantic distance between adjacent terms; then, a non-uniformity check is performed on the initial semantic distance. The initial semantic distance between adjacent terms reflects the semantic span between adjacent severity levels in the sequence. The semantic distance between different pairs of adjacent terms is usually not uniform to conform to clinical reality—the leap from mild to moderate symptoms is clinically significantly different from the leap from severe to extremely severe symptoms; the purpose of the non-uniformity check is to identify and verify the rationality of this distance distribution. The verification process first calculates the mean and standard deviation of the initial semantic distance between all adjacent terms, and then calculates the coefficient of variation (the ratio of the standard deviation to the mean). When the coefficient of variation is less than the preset threshold, it indicates that the distance distribution is too uniform and may not fully reflect the non-linear characteristics of clinical severity. In this case, clinical experts need to review and adjust the scoring. When the coefficient of variation is greater than the preset upper threshold, it indicates that the distance distribution is too disparate and there may be scoring deviations for individual terms. In this case, review and adjustment are also required. The standard for passing the verification is that the coefficient of variation is within a reasonable range, confirming that the semantic distance distribution between adjacent terms is consistent with clinical understanding.

[0032] Based on the semantic distance between adjacent terms after non-uniformity verification, the initial severity level values ​​of each term with varying degrees of change are progressively reconstructed by accumulating the values ​​from the semantic origin. This results in the cumulative reconstructed semantic distance value for each term with varying degrees of change. Progressive cumulative reconstruction transforms discrete level scores into continuous cumulative distance measures. The reconstruction process starts from the semantic origin (baseline term, with a distance value of 0). The cumulative reconstructed semantic distance value of the first term with varying degrees of change is equal to its semantic distance from the semantic origin, and the cumulative reconstructed semantic distance value of the a-th term with varying degrees of change is equal to the progressive summation of the semantic distances between the first a adjacent terms. The cumulative reconstruction preserves information about the non-uniform distances between adjacent terms, ensuring that the final distance value accurately reflects the semantic gap between each term and the baseline state. The cumulative reconstructed semantic distance value is then normalized to the [0, 1] interval. This normalization ensures that the semantic distance benchmark values ​​for different clinical observation dimensions have a unified range, facilitating subsequent cross-dimensional comparisons and calculations.

[0033] To address the situation where the direction of disease deterioration is not unique within the same clinical observation dimension, separate ordered semantic term sequences corresponding to each deterioration direction are constructed. Initial grading, non-uniformity verification, and cumulative reconstruction are then performed independently on each ordered semantic term sequence. A typical case of non-unique deterioration direction occurs in the consciousness state dimension. Consciousness abnormalities include both inhibitory (deterioration from wakefulness to drowsiness, confusion, stupor, and coma) and excitatory (deterioration from wakefulness to agitation, restlessness, and delirium). Both directions represent deterioration of consciousness, but the pathological mechanisms and clinical manifestations differ. For such cases, using the common baseline state term (awake consciousness) as the semantic origin, separate ordered semantic term sequences for inhibitory and excitatory directions are constructed. Each branch sequence is independently graded and distance-calibrated. When a nurse selects a descriptive term belonging to a branch sequence, the semantic distance benchmark value corresponding to that branch sequence is used for quantification. Simultaneously, directional identification information is added to the semantic distance benchmark value of the branch sequence so that subsequent offset calculations can identify the deterioration direction, providing support for the directional analysis of disease changes.

[0034] In an embodiment of the present invention, the process of performing offset calculation includes: The quantized offset values ​​of each clinical observation dimension are organized into a vector form to form the instantaneous offset vector at the current observation moment. The instantaneous offset vector is a five-dimensional numerical vector formed by arranging the quantized offset values ​​of each clinical observation dimension in this round of assessment according to a uniform dimensional order. The instantaneous offset vector completely records the patient's deviation from the baseline state in each clinical observation dimension at the current observation moment, and serves as the basic input object for subsequent distance measurements and vector operations. The component values ​​of the vector are derived from the quantized offset values ​​of each dimension, and have undergone specific semantic quantization coefficient correction and cross-dimensional consistency verification.

[0035] The initial baseline vector is constructed by extracting the clinical observation dimensions of the patient at the initial triage time from the complete baseline assessment record. The initial baseline vector reflects the patient's initial condition upon entering the waiting area. For patients with abnormal symptoms at the initial triage, the corresponding dimension component values ​​in their baseline vector are non-zero. For example, for a patient presenting with abdominal pain as the chief complaint, the baseline component values ​​for the pain dimension are determined based on the pain assessment results at the initial triage. When the target waiting patient has multiple rounds of assessment records, the quantified offset values ​​of each dimension from the most recent round are extracted to construct the previous observation vector. The previous observation vector records the offset state at the time of the previous assessment and serves as a benchmark for assessing short-term trends in the patient's condition. When the target patient is receiving their first assessment, the previous observation vector is defaulted to the zero vector of the corresponding dimension in the initial baseline vector.

[0036] The Euclidean distance between the instantaneous offset vector and the initial baseline vector is calculated as the cumulative change in patient condition; the Euclidean distance between the instantaneous offset vector and the previous observation vector is calculated as the interval change in patient condition. The cumulative change in patient condition measures the overall change in the patient's condition from the initial triage to the current observation time, and is calculated using the following formula: ;in, To accumulate the range of changes in the condition, Let i be the i-th dimension component of the instantaneous offset vector. is the i-th dimension component of the initial baseline vector.

[0037] The interval range of disease change measures the rate of recent change in the patient's condition from the last round of observation to the current observation time. The calculation formula is as follows: ;in, This indicates the range of changes in the severity of the illness within a given interval. is the i-th dimension component of the previously observed vector. The cumulative disease change amplitude and the interval disease change amplitude reflect the changing characteristics of the disease across different time spans. The former reveals the overall deterioration trend, while the latter reveals the dynamic characteristics of recent acceleration or deceleration in deterioration. The synergistic analysis of the two can comprehensively describe the evolution of the disease.

[0038] The cumulative and interval-level changes in disease severity were normalized, and then differential weighting coefficients were assigned based on these normalizations. Normalization maps the two types of changes to a comparable scale using a maximum normalization method, with the maximum theoretical Euclidean distance corresponding to the maximum offset of each pre-defined clinical observation dimension as the normalization denominator. The differential weighting coefficients are dynamically determined based on the relative magnitudes of the two types of changes: when the normalized value of the interval-level change is significantly greater than that of the cumulative change, it indicates a recent rapid change in the disease, and the weighting coefficient of the interval change is appropriately increased (e.g., interval weight 0.6, cumulative weight 0.4); when the normalized value of the cumulative change is significantly greater than that of the interval change, it indicates a continued slow deterioration of the disease with little recent change, and the weighting coefficient of the cumulative change is appropriately increased (e.g., cumulative weight 0.6, interval weight 0.4); when the two are close in magnitude, a balanced weight (e.g., 0.5 for each) is used. This differential weighting allows the synthesized results to adaptively reflect the clinical importance of different disease severity patterns.

[0039] Based on the differentiated weighting coefficient allocation results, the cumulative disease change amplitude and the interval disease change amplitude are vector-synthesized to generate a disease change vector. The vector synthesis process linearly weights the two types of normalized change amplitudes using weighting coefficients to obtain the synthesized scalar intensity value of the disease change. On this basis, the disease change vector is constructed by combining the directional information of the instantaneous offset vector, with the direction determined by the difference vector direction between the instantaneous offset vector and the initial baseline vector. The disease change vector is then decomposed directionally to identify the contribution of each clinical observation dimension to the overall disease change. The directional decomposition process calculates the ratio of the component of the disease change vector in each dimension to the total vector magnitude, i.e., the directional contribution ratio of each dimension. The clinical observation dimension with the largest contribution is marked as the dominant deterioration dimension, which represents the most significant abnormal direction in the patient's current disease change, providing a directional reference for subsequent risk assessment and intervention recommendations.

[0040] In an embodiment of the present invention, the process of obtaining the disease progression risk index includes: The baseline value of the intensity of disease change is calculated based on the vector of disease change, and then adjusted for dimensional sensitivity according to the clinical urgency of the dominant deterioration dimension. The baseline value of the intensity of disease change is taken as the magnitude of the disease change vector, directly reflecting the overall magnitude of the disease change. Dimensional sensitivity adjustment is a process of correcting the intensity of disease change based on the clinical urgency weights of different clinical observation dimensions in an emergency setting, based on the objective law that there are differences in the degree of clinical urgency of abnormalities in different dimensions in emergency medicine. The clinical urgency weights of each clinical observation dimension are preset as follows: the urgency weight of the consciousness dimension is preset to the highest, because a rapid deterioration of consciousness usually indicates life-threatening emergencies such as intracranial lesions, severe metabolic disorders, or circulatory failure; the urgency weight of the respiratory status dimension is set to the second highest, because a rapid deterioration of respiratory function directly threatens oxygenation and ventilation; the urgency weight of the facial color observation dimension is set to the third highest, because changes in facial color are an external reflection of circulatory status and oxygenation level; the urgency weight of the pain manifestation dimension is set to the fourth highest, serving as the baseline weight; and the urgency weight of the limb movement dimension is set to the lowest, because although changes in limb movement may indicate neurological lesions, the degree of urgency is relatively low. The formula for calculating dimensional sensitivity adjustment is: ;in, To adjust the intensity of the changes in the condition. The clinical urgency weights corresponding to the dominant deterioration dimension. This represents the magnitude of the disease change vector. The adjusted intensity of disease change more accurately reflects the clinical risk level of the current disease change.

[0041] The system obtains the waiting time of target patients and sets a safe threshold for waiting time based on the patient's initial triage level. When the waiting time exceeds the safe threshold, an overtime waiting risk factor is calculated. The waiting time is the difference between the current time and the initial triage timestamp, reflecting the total time the patient has waited in the waiting area. The safe threshold is set according to the clinical timeliness requirements of different triage levels; for example, the safe threshold for Level 1 critically ill patients is 0 minutes (they should be seen immediately and should not be waiting under normal circumstances); the safe threshold for Level 4 non-emergency patients is 120 minutes. When the actual waiting time exceeds the corresponding level's safe threshold, it indicates that the patient has exceeded the recommended waiting time, and the possibility of their condition worsening increases with time. The formula for calculating the overtime waiting risk factor is: ;in, For the risk factor of waiting overtime, This refers to the actual waiting time. As a safety threshold, This is a time sensitivity coefficient (valued according to triage level, e.g., 1.5 for level 2, 1.0 for level 3); when the waiting time does not exceed the safety threshold, the overtime waiting risk factor is set to 0. The overtime waiting risk factor quantifies the additional risk increment caused by excessive waiting time, reflecting the important impact of time factors on the assessment of disease deterioration.

[0042] The environmental load parameters of the current waiting area are obtained, and the environmental stress coefficient is calculated based on these parameters. The environmental load parameters include the total number of waiting patients, the number of patients at the same level of waiting, and the doctor's consultation rate. The total number of waiting patients reflects the overall congestion level of the waiting area and is obtained from real-time queuing data in the emergency information system; the number of patients at the same level of waiting reflects the scale of patients competing for treatment with the target patient; the doctor's consultation rate is the number of patients seen in each consultation room per unit time, calculated in real-time based on recent consultation records. The formula for calculating the environmental stress coefficient is: ;in, Environmental stress coefficient, This represents the total number of patients currently waiting for treatment. The standard carrying capacity of the waiting area The number of patients waiting for the same level of treatment. This serves as a reference benchmark for the number of patients waiting at the same level. Based on the current rate of doctor consultations, This serves as a reference value for the standard patient reception rate. A higher environmental stress coefficient indicates higher stress in the waiting environment, a lower likelihood of patients receiving timely treatment, and indirectly increases the risk of disease deterioration. The calculated environmental stress coefficient is truncated to an upper limit, restricting it to a preset maximum value (e.g., a maximum of 2.0) to prevent extreme environmental data from causing excessive bias in risk assessment.

[0043] The baseline risk value corresponding to the patient's initial triage level is extracted. The baseline value of the intensity of disease change, the risk factor of prolonged waiting time, and the environmental stress coefficient are used as risk increments. A multifactor risk fusion function is used to calculate the patient's disease deterioration risk index. The baseline risk value is a preset risk starting level based on the initial triage level, reflecting the difference in baseline risk among patients at different triage levels: for example, the baseline risk value for a Level 1 critically ill patient is 0.8. The multifactor risk fusion function comprehensively calculates the baseline risk value and the three risk increments using the following formula: Where R represents the risk index of disease progression. As the benchmark risk value, The intensity of disease changes after dimensional sensitivity adjustment. For the risk factor of waiting overtime, Environmental stress coefficient, , and The fusion weight coefficient is the weighting factor for each risk increment. The setting of the fusion weight coefficient reflects the relative importance of each factor in the comprehensive risk assessment. The typical configuration is: intensity of disease change > risk factor of waiting time > environmental stress coefficient. The risk index obtained by fusion calculation is truncated in the range of [0, 1]. Values ​​outside the range are truncated to 0 or 1 respectively.

[0044] The calculated risk index for disease deterioration is dynamically segmented into low-risk, medium-risk, high-risk, and extremely high-risk intervals. Dynamic threshold segmentation is a classification process that maps a continuous risk index to discrete risk levels. The low-risk interval indicates that the patient's condition is basically stable and requires no emergency intervention; the medium-risk interval indicates that the patient's condition has changed somewhat and requires attention but is not currently in imminent danger; the high-risk interval indicates that the patient's condition has significantly deteriorated and requires immediate medical attention or temporary treatment; the extremely high-risk interval indicates that the patient's condition has deteriorated rapidly and is life-threatening, requiring immediate medical attention or the initiation of emergency treatment procedures. Different priority adjustment strategies are set for different risk ranges: low-risk ranges do not trigger priority adjustments and maintain the existing queuing order; medium-risk ranges trigger small priority weight increments (e.g., increments of 5 to 15) to improve the consultation order but do not change the triage level; high-risk ranges trigger medium-sized priority weight increments (e.g., increments of 15 to 30) and generate triage level upgrade suggestions; very high-risk ranges trigger large priority weight increments (e.g., increments of 30 to 50) and automatically trigger doctor consultation requests and emergency treatment notifications.

[0045] In an embodiment of the present invention, the dynamic adjustment process of the initial medical visit priority weight includes: The patient's current priority weight is extracted as the baseline weight before adjustment. The baseline weight before adjustment is the patient's latest priority weight value before this round of evaluation. For patients receiving their first round of evaluation, the baseline weight before adjustment is equal to the initial priority weight value; for patients with multiple rounds of evaluation, the baseline weight before adjustment is the weight value after the last adjustment. The baseline weight is extracted from the dynamic field area of ​​the waiting QR code or the patient record in the emergency information system.

[0046] Based on the risk range of the disease progression risk index, the incremental adjustment range of the priority weight is determined, and the current medical treatment priority weight is adjusted accordingly. The determination of the incremental adjustment range is based on the differentiated priority adjustment strategy corresponding to each risk range. Within the incremental range of each risk range, the precise incremental value is determined by linear interpolation based on the specific position of the risk index within that range.

[0047] The system calculates the difference between the adjusted priority weight and the baseline weight before adjustment. When the absolute value of the difference exceeds a preset level change threshold, a priority adjustment suggestion is generated and presented on the mobile terminal. The level change threshold is a critical value for determining whether a priority change is sufficient to cause a jump in triage level, and is usually set to half the difference between the baseline weights of two adjacent triage levels. The priority adjustment suggestion includes four core pieces of information: the level before adjustment, the level after adjustment, the dominant deterioration dimension, and the suggested intervention measures. The level before adjustment is the patient's current triage level, and the level after adjustment is determined based on the level range that the adjusted priority weight falls into. The dominant deterioration dimension indicates the most important clinical observation dimension that caused the change in condition. The suggested intervention measures are generated by matching the dominant deterioration dimension and risk interval from a preset intervention measure knowledge base. For example, when the dominant deterioration dimension is respiratory status and the risk interval is high risk, the suggested intervention measures may include "immediately retest blood oxygen saturation and respiratory rate, and provide oxygen support if necessary." The priority adjustment suggestion is presented as a structured card in a pop-up window on the mobile terminal screen, and nurses can choose from three operation options: "Confirm Adjustment," "Postpone Adjustment," or "Confirm After Modification."

[0048] After the nurse confirms the priority adjustment suggestion, the adjusted priority weight is written into the waiting area QR code, and the patient's priority weight is simultaneously updated in the emergency information system. The timestamp of this adjustment, the basis for the adjustment, and the identifier of the operating nurse are also recorded, forming a priority adjustment audit log. The operation of writing to the waiting area QR code is achieved by updating the patient's QR code data packet in the emergency information system and the local cache of the mobile terminal, ensuring that the latest priority weight information can be obtained the next time the code is scanned. The audit log fully records all contextual information of each priority adjustment operation, including the adjustment time, the weight values ​​and levels before and after the adjustment, the risk index value of disease deterioration that triggered the adjustment, the dominant deterioration dimension, the quantitative offset values ​​of each clinical observation dimension, and the employee number and name of the operating nurse, providing complete data support for subsequent quality management reviews and medical dispute tracing.

[0049] A rationality check is performed on any shifts in the adjusted priority weighting. This check includes verifying whether the adjusted priority weighting matches the patient's vital sign measurements. If a mismatch exists, the nurse is required to measure the current vital signs, and the priority weighting is readjusted based on the updated data. This rationality check is a safety mechanism to objectively verify the priority adjustment results, preventing inconsistencies between priority adjustments based solely on qualitative observations and the patient's actual physiological state. The check first determines the corresponding vital sign reference range based on the adjusted triage level, then compares the patient's most recent vital sign measurement with this reference range. If the vital sign measurements are entirely within the normal range but the adjusted level is Level II critical, a mismatch is identified. In this case, the mobile terminal prompts the nurse: "The current priority has been adjusted to Level II critical, but the most recent vital sign record shows no abnormalities. Please measure the current vital signs to confirm." Nurses use bedside monitoring equipment or portable monitoring devices to perform real-time vital sign measurements, and the results are transmitted to the mobile terminal via Bluetooth or manual input. Based on the updated vital signs data, the system reassesses the rationality of the priority adjustment: if the newly measured vital signs are indeed abnormal, the priority adjustment is confirmed to be effective; if the newly measured vital signs are still within the normal range, the system suggests reducing the adjustment range or maintaining the original level, and records the situation in the audit log for review.

[0050] In an embodiment of the present invention, the process of reordering all waiting patients in the current queue includes: When the emergency information system receives updated patient priority weights, it extracts the priority weights of all currently waiting patients and constructs a waiting queue priority vector. The waiting queue priority vector is a one-dimensional numerical vector formed by arranging the latest priority weights of all waiting patients in order of their patient numbers. The length of the vector is equal to the total number of patients in the current waiting queue. The extraction process iterates through all patient records in the emergency information system marked "waiting," reading the latest priority weight value for each patient, including the dynamically adjusted weight value after rounds of evaluation and the initial weight value before adjustment.

[0051] The priority vectors of the waiting queue are sorted in descending order to obtain a theoretical patient appointment sequence based on priority weights. This descending order places patients with the highest priority weights at the beginning of the sequence and patients with the lowest weights at the end, establishing a basic "higher priority, first-come" ranking principle. For patients with the same priority weights, the initial triage time stamp is used as a secondary ranking criterion, placing patients with earlier triage times at the beginning, reflecting the fairness principle of "first-come, first-served" for patients of the same priority. The theoretical patient appointment sequence is the optimal ranking result calculated based on the current weight values ​​of all patients, serving as the initial scheme for subsequent sequence adjustments.

[0052] Based on the theoretical patient access sequence, patients whose waiting time exceeds a preset safety threshold but have a low priority weight are identified, and their positions within the theoretical access sequence are adaptively adjusted. This adaptive adjustment introduces a fairness compensation mechanism on top of strict priority ranking, aiming to prevent low-priority patients from experiencing prolonged waiting times due to the continuous influx of high-priority patients. The adjustment process first iterates through all patients in the theoretical access sequence, marking the set of patients whose waiting time exceeds the safety threshold corresponding to their triage level (e.g., a level 4 non-emergency patient waiting more than 120 minutes). Then, a timeout coefficient is calculated for these patients, equal to the ratio of actual waiting time to the safety threshold. When the timeout coefficient exceeds a preset adaptive adjustment trigger threshold (e.g., 1.5 times), the patient's position in the theoretical access sequence is moved forward, with the adjustment magnitude positively correlated with the timeout coefficient but limited to not exceeding the position of adjacent higher-priority patients. The adaptively adjusted access sequence maintains the core priority logic while ensuring fairness in waiting times and avoiding extreme waiting situations.

[0053] After adjusting the patient flow sequence, clinic resource matching is performed. Based on the specialty and current workload of each clinic, patients are assigned to the most suitable clinic queue. Clinic resource matching is the process of subdividing a single global patient flow sequence into clinic-specific sub-sequences. The matching process first retrieves a list of currently open clinics, specialty labels for each clinic (e.g., internal medicine, surgery, pediatrics, orthopedics), and the current number of patients in the queue from the emergency information system. Then, based on the patient's chief complaint and symptom characteristic codes, the most suitable specialty specialty is matched. When multiple suitable clinics exist, the clinic with the fewest patients in the queue is selected to achieve load balancing. After matching, each patient is assigned to a corresponding clinic queue, maintaining their relative position within the global sequence within the clinic queue.

[0054] The system calculates the change in sequence position for each patient before and after sequence rearrangement, identifying patients with significantly advanced or significantly delayed sequence positions. The change in sequence position is the difference between the rearranged and unrearranged position numbers; a negative value indicates an earlier arrival (earlier consultation), and a positive value indicates a later arrival (later consultation). A significant change in sequence position is defined as an absolute value greater than a preset significance threshold (e.g., 3 positions). For patients with significantly advanced sequence positions, a priority consultation notification is sent via push notification to their mobile device and to the waiting area's large screen. The notification includes the message, "Your consultation order has been adjusted to position X based on your condition assessment. Please pay attention to the call information." For patients with significantly delayed sequence positions, an explanation of the sequence adjustment and the estimated delay is sent, including the message, "Due to a patient in the waiting area whose condition has changed and requires priority consultation, your estimated consultation time has been delayed by approximately X minutes. Thank you for your understanding." The method of sending notifications can be flexibly chosen based on the patient's receiving conditions; it can be publicly displayed on the waiting area's large screen or sent via SMS to the patient's registered mobile phone number.

[0055] The patient appointment sequence is updated synchronously in the consultation room, and patients are called according to the new sequence in the queuing system. Simultaneously, the triggering reason, adjustment range, and number of affected patients are recorded in the emergency information system, forming a sequence adjustment log. Updates in the consultation room are achieved by pushing the latest patient appointment sequence data to each consultation room workstation in real time through the emergency information system. Attending physicians can view the updated list of waiting patients and their priority markings on the consultation room workstation interface. The queuing system automatically adjusts the calling order according to the new appointment sequence. When a sequence reordering occurs, if the currently called patient has not arrived, the system automatically jumps to the next patient and delays the recall. The sequence adjustment log records complete information for each sequence reordering operation, including the patient number that triggered the reordering, the triggering reason (the risk index of disease deterioration reaching the high-risk or extremely high-risk range), a global sequence snapshot before and after the adjustment, the number of affected patients, and the change in each patient's position, providing a data foundation for emergency department operation management and quality assessment.

[0056] In embodiments of the present invention, the method further includes: A patrol task management module is set up on the mobile terminal. This module generates a patrol priority list based on the initial triage level and waiting time of waiting patients, and pushes a list of target waiting patients requiring priority patrol to the nurses. The generation of the patrol priority list comprehensively considers the urgency of the patient's triage level and the cumulative risk of waiting time: patients with higher triage levels (such as level 2 critical illness) are patrolled before those with lower triage levels; among patients of the same level, those with longer waiting times are patrolled before those with shorter waiting times. The target waiting patient list is arranged by priority and color-coded to indicate urgency (e.g., red indicates level 2 critical illness patients who have exceeded the waiting time, yellow indicates level 3 emergency patients who are close to exceeding the waiting time), helping nurses quickly identify those who need priority patrol. The patrol task management module also automatically updates the list status based on the nurses' patrol completion records; patients who have completed patrols are marked as "assessed" with the assessment time noted, while patients who have not been patrolled remain in the "pending assessment" status.

[0057] After the nurse scans the waiting QR code of the target patient, the initial triage information and summaries of previous rounds of assessments are displayed at the top of the clinical data collection interface. The initial triage information is presented in a concise, structured card format, including key information such as triage level, chief complaint, onset time, initial vital signs summary, and allergy information, allowing nurses to quickly understand the patient's basic condition without having to read the entire medical record. The summaries of previous rounds of assessments are displayed in a timeline format, with each round record including the round time, selected terms for each clinical observation dimension, quantitative offset value, risk index of disease deterioration, and whether priority adjustments were made. This allows nurses to clearly grasp the trajectory of the patient's condition and provides historical reference for current observations.

[0058] After nurses complete the selection of qualitative descriptive terms, the mobile terminal calculates and displays the quantitative offset value and the vector of disease condition changes in real time, presenting the changing trends of each clinical observation dimension in the form of visual charts to help nurses intuitively judge the degree of disease change. The visual charts include two forms: first, a radar chart, which uses the five clinical observation dimensions as axes to simultaneously display the distribution of quantitative values ​​for the initial baseline state, previous observation state, and current observation state, allowing nurses to clearly identify the direction and magnitude of changes in each dimension; second, a trend line chart, with time as the horizontal axis and quantitative offset value as the vertical axis, displaying the curve of offset value changes from the initial triage to the current rounds of assessments, enabling nurses to identify whether the disease is continuously deteriorating, fluctuating, or trending towards stability. The dominant deterioration dimension in the visual charts is highlighted with striking colors and labels, and the disease deterioration risk index is simultaneously displayed with numbers and color codes (green for low risk, yellow for medium risk, orange for high risk, and red for extremely high risk), assisting nurses in making rapid clinical judgments and decisions.

[0059] When the risk index of a patient's condition worsening reaches the high-risk or extremely high-risk range, the mobile terminal triggers a doctor's consultation suggestion and adds a consultation prompt to the priority adjustment suggestion. The triggering of the doctor's consultation suggestion is an emergency contact mechanism automatically initiated by the system based on the risk level, aiming to obtain timely professional judgment and treatment decisions from doctors in cases where the patient's condition may deteriorate rapidly. Nurses can initiate a doctor's consultation request with a single click via their mobile terminals, pushing information about the patient's condition changes and current location to the on-duty doctor. The consultation request push includes: patient's name and identification code, initial triage level and current risk level, dominant deterioration dimension and specific manifestations, latest quantitative offset values ​​for each clinical observation dimension, a trend chart of condition changes, and the patient's current location information in the waiting area. The push is sent to the on-duty doctor's work terminal or mobile device through the emergency information system's instant messaging module, with a confirmation mechanism set up. If the on-duty doctor does not confirm receipt within a preset time, the system automatically escalates the push to the second-line on-duty doctor or the head of the emergency department.

[0060] A dynamic assessment and monitoring dashboard is simultaneously installed on the computer terminals at the triage station, displaying the latest round of assessment times, disease deterioration risk indices, and priority adjustment records for each waiting patient in real time. The dashboard uses a combination of tables and graphs, displaying key status information for all waiting patients in a patient list format, arranged from highest to lowest disease deterioration risk index, with risk levels clearly marked by color coding. The dashboard also includes a statistical overview area displaying macro-level operational indicators such as the total number of patients currently waiting, distribution across triage levels, average waiting time, and the number of patients exceeding their waiting time, helping triage nurses comprehensively grasp the overall situation in the waiting area. For high-risk patients who have not undergone rounds and assessments within the preset time, the monitoring dashboard generates a round reminder via flashing markers and sound alerts, which is then pushed to the mobile terminal. The preset time is set differently based on the triage level: the maximum rounds and assessment interval is 15 minutes for level 2 critically ill patients, 30 minutes for level 3 emergency patients, and 60 minutes for level 4 non-emergency patients. When the time elapsed since a patient's last assessment reaches the upper limit of the corresponding interval, the system automatically highlights the patient's record in red on the monitoring dashboard and simultaneously pushes an inspection reminder notification to the mobile terminal of the nurse responsible for that area. The notification includes the patient's name, location, initial triage level, time elapsed since the last assessment, and current risk index of deterioration of condition, ensuring that high-risk patients receive continuous attention and timely assessment.

[0061] This invention achieves systematic and dynamic assessment of the condition of patients waiting in the emergency department and intelligent management of their priority by encapsulating and dynamically updating structured data of waiting QR codes, driving qualitative-quantitative conversion with a clinical observation semantic knowledge base, performing vectorized offset operations on quantified offset values, calculating a disease deterioration risk index based on multi-factor fusion, and implementing a differentiated priority adjustment and queuing sequence reconstruction mechanism. The semantic knowledge base quantification method of this invention transforms nurses' subjective clinical observations into calculable and comparable numerical scores, overcoming the problems of traditional qualitative assessments being difficult to objectively measure and lacking consistency across assessors. The offset calculation method based on vector space comprehensively describes the evolution characteristics of the disease from two time dimensions: cumulative change and interval change, and provides directional guidance for clinical intervention by identifying the dominant deterioration dimension. The multi-factor risk fusion mechanism comprehensively considers multiple factors such as the intensity of disease changes, waiting time, and environmental load, avoiding the one-sidedness of single-factor judgment. The closed-loop feedback priority dynamic adjustment and queue reconstruction mechanism ensure that the order of waiting patients always matches their real-time disease status, effectively reducing the risk of delayed treatment due to the failure to detect changes in the disease during the waiting period, and providing an intelligent technical solution for emergency waiting management. It should be noted that the parameter values ​​involved in this application are only examples for reference.

[0062] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0063] All formulas in this manual are dimensionless and calculated numerically. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0064] The parameter values ​​involved in the embodiments of this application are only illustrative examples and are not intended to limit this application. In practical applications, they can be adaptively adjusted and selected according to specific scenarios. Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for dynamic assessment of the condition of emergency waiting patients based on mobile terminals, characterized in that, include: The system obtains the triage results and baseline vital signs data of patients after the initial triage and generates a waiting slip based on them. The waiting slip is printed with a waiting QR code, which contains the patient's unique identifier, the initial triage timestamp, the current consultation priority weight, and the baseline feature vector of the patient's condition. Nurses periodically patrol the waiting area using mobile devices and obtain the corresponding patient's queuing information and baseline feature vector of the patient's condition by scanning the corresponding patient's waiting QR code; and simultaneously obtain the corresponding patient's complete baseline assessment record in the emergency information system. Nurses collect qualitative clinical observation information of target waiting patients through the clinical data collection interface on mobile terminals, and convert the targeted clinical observation information into quantitative offset values ​​through a preset clinical observation semantic knowledge base. The quantized offset value is offset by the corresponding dimensional component of the baseline feature vector of the patient's condition, and the patient's condition deterioration risk index is calculated by combining the patient's waiting time, initial triage level and waiting environment. The initial priority weight of the target patient is dynamically adjusted based on the risk index of disease deterioration, priority adjustment suggestions are generated, and it is determined whether to trigger the reconstruction of queuing information based on these suggestions. If triggered, update the initial consultation priority weight and the baseline feature vector of the patient's condition. The updated initial patient priority weights are synchronized to the emergency information system. Based on the emergency information system, all patients waiting in the current queue are reordered, and a sequence change notification is sent to the relevant patients and clinics.

2. The method for dynamic assessment of the condition of emergency waiting patients based on mobile terminals according to claim 1, characterized in that, The process of obtaining the quantization offset value includes: A clinical observation semantic knowledge base is pre-constructed, which includes baseline status terms and degree of change terms under five clinical observation dimensions: facial color observation, respiratory status, consciousness status, pain manifestation, and limb movement. The baseline status terms are assigned a zero baseline quantization value, and the semantic distance of each term with a degree of change is simultaneously calibrated according to the direction and degree of disease deterioration to obtain the semantic distance baseline value; The semantic distance baseline value is corrected based on the patient's age and underlying disease information to obtain the specific semantic quantification coefficient corresponding to the patient. Obtain the specific semantic quantization coefficients corresponding to the qualitative descriptive terms selected by nurses based on the clinical data collection interface, and perform a difference operation between them and the corresponding dimension in the patient's corresponding disease baseline feature vector to obtain the quantization offset value under the corresponding dimension. Obtain the quantitative offset values ​​corresponding to different clinical observation dimensions involved in the same observation, and perform cross-dimensional consistency verification. Based on the consistency verification results, choose whether to provide feedback on contradictory information to the nurses and request confirmation or correction. Once the consistency check is passed, a quantitative offset value is formed based on the combination of quantitative offset values ​​corresponding to different clinical observation dimensions.

3. The method for dynamic assessment of the condition of emergency waiting patients based on mobile terminals according to claim 2, characterized in that, The process of semantic distance calibration includes: For each clinical observation dimension, the baseline state term under that dimension is set as the semantic origin, and all the degree of change terms under that dimension are arranged according to the progressive direction of clinical disease deterioration to construct a single-dimensional semantically ordered term sequence corresponding to that dimension. Based on the preset clinical severity evaluation criteria, the initial severity level score is assigned to each word in the single-dimensional semantic ordered word sequence to obtain the initial severity level value corresponding to each word. For two adjacent terms with varying degrees of change in a single-dimensional semantically ordered term sequence, the difference between their initial severity levels is calculated to obtain the initial semantic distance between adjacent terms; the initial semantic distance is then checked for non-uniformity. Based on the semantic distance between adjacent terms after non-uniformity verification, the initial severity level values ​​of each term with varying degree of change are progressively accumulated and reconstructed, starting from the semantic origin, to obtain the accumulated reconstructed semantic distance value corresponding to each term with varying degree of change. The accumulated reconstructed semantic distance value is then subjected to interval normalization to obtain the semantic distance benchmark value corresponding to each term with varying degree of change. To address the situation where the direction of disease deterioration is not unique under the same clinical observation dimension, we construct a branch semantic ordered word sequence corresponding to each deterioration direction, and independently perform initial level assignment, non-uniformity verification, and cumulative reconstruction process on each branch semantic ordered word sequence.

4. The method for dynamic assessment of the condition of emergency waiting patients based on mobile terminals according to claim 1, characterized in that, The process of offset calculation includes: The quantitative offset values ​​of each clinical observation dimension are organized into a vector form to form the instantaneous offset vector at the current observation time; The patient's clinical observation dimensions were measured at the initial triage time from the complete baseline assessment record to construct the initial baseline vector; and when there were multiple rounds of inspection records for the target waiting patient, the quantitative offset values ​​of each dimension from the most recent inspection record were extracted to construct the previous observation vector. Calculate the Euclidean distance between the instantaneous offset vector and the initial baseline vector as the cumulative change in the patient's condition; calculate the Euclidean distance between the instantaneous offset vector and the previous observation vector as the interval change in the patient's condition. The cumulative and interval variations in disease severity were normalized, and differential weighting coefficients were assigned based on these normalizations. Based on the results of the differential weighting coefficient allocation, the cumulative disease change amplitude and the interval disease change amplitude are vectorized to generate a disease change vector; and the disease change vector is decomposed in direction to identify the contribution of each clinical observation dimension to the overall disease change, and the clinical observation dimension with the largest contribution is marked as the dominant deterioration dimension.

5. The method for dynamic assessment of the condition of emergency waiting patients based on mobile terminals according to claim 1, characterized in that, The process of obtaining the risk index for disease progression includes: The baseline value of the intensity of disease change is calculated based on the disease change vector; and the baseline value of the intensity of disease change is adjusted for dimensional sensitivity according to the clinical urgency of the dominant deterioration dimension. The system obtains the waiting time of target patients and sets a safe threshold for the waiting time based on the patient's initial triage level; when the waiting time exceeds the safe threshold, the system calculates the overtime waiting risk factor. Obtain the environmental load parameters of the current waiting area and calculate the environmental stress coefficient based on the environmental load parameters; the environmental load parameters include the total number of waiting patients, the number of waiting patients of the same level, and the doctor's consultation rate; The baseline risk value corresponding to the patient's initial triage level was extracted, and the baseline value of the intensity of change in condition, the risk factor of waiting time, and the environmental stress coefficient were used as risk increments. The patient's condition deterioration risk index was calculated through a multi-factor risk fusion function. The calculated risk index of disease deterioration is dynamically segmented into low-risk, medium-risk, high-risk, and extremely high-risk ranges; and differentiated priority adjustment strategies are set for different risk ranges.

6. The method for dynamic assessment of the condition of emergency waiting patients based on mobile terminals according to claim 1, characterized in that, The dynamic adjustment process of initial medical visit priority weights includes: Extract the patient's current medical priority weight as the baseline weight before adjustment; Based on the risk range of the disease deterioration risk index, determine the incremental adjustment range of the priority weight, and adjust the current medical treatment priority weight accordingly. The difference between the adjusted priority weight and the baseline weight before adjustment is calculated. When the absolute value of the difference exceeds the preset level change threshold, a priority adjustment suggestion is generated and presented on the mobile terminal. The priority adjustment suggestion includes the level before adjustment, the level after adjustment, the dominant deterioration dimension, and the suggested intervention measures. After the nurse confirms the priority adjustment suggestion, the adjusted priority weight is written into the waiting QR code and the patient's priority weight is updated synchronously in the emergency information system; at the same time, the timestamp of this adjustment, the basis for the adjustment and the identification of the operating nurse are recorded to form a priority adjustment audit log.

7. The method for dynamic assessment of the condition of emergency waiting patients based on mobile terminals according to claim 6, characterized in that, The dynamic adjustment process for initial medical visit priority weights also includes: The rationality of the adjusted priority weights for medical visits is verified. The rationality verification includes checking whether the adjusted priority weights for medical visits match the patient's vital sign measurements. If there is a mismatch, the nurse is required to measure the current vital signs and readjust the priority weights for medical visits based on the updated vital sign data.

8. The method for dynamic assessment of the condition of emergency waiting patients based on mobile terminals according to claim 1, characterized in that, The process of generating the waiting area QR code includes: When a patient completes the initial triage, the patient's basic triage data is collected, and the patient's basic information and medical record summary are extracted from the emergency information system. An encrypted patient identification code is generated based on the patient's unique identifier; the vital signs measured during the initial triage are vectorized and encoded to generate a baseline feature vector of the patient's condition; and the chief complaints recorded by the nurse during the initial triage are processed by text structuring, extracting symptom keywords and mapping them to symptom feature codes. Based on the patient's initial triage level, the corresponding initial priority weight value is retrieved from the preset priority weight mapping table; and the initial priority weight value is normalized according to the overall distribution of the current waiting queue. The patient identification code, initial triage timestamp, initial priority weight value, disease baseline feature vector, and symptom feature code are encapsulated according to a predetermined data structure to generate a QR code data packet; and a dynamic update field is reserved in the QR code data packet. The QR code data packet is encoded into a QR code and printed on the patient's queuing pass; the patient's name, estimated waiting time, and round-trip assessment prompts are also printed on the queuing pass. When a patient's priority weight is updated after rounds and assessments, the patient sends a QR code update request to the emergency information system via a mobile terminal. The QR code update request includes the patient's identification code and the updated priority weight value. After verifying the validity of the update request, the emergency information system generates a new QR code data packet and returns it to the mobile terminal. The mobile terminal updates the patient's QR code data in its local cache and displays the updated information synchronously the next time the code is scanned.

9. The method for dynamic assessment of the condition of emergency waiting patients based on mobile terminals according to claim 1, characterized in that, The process of reordering all waiting patients in the current queue includes: When the emergency information system receives patient information with updated priority weights, it extracts the priority weights of all waiting patients and constructs a priority vector for the waiting queue. The priority vectors of the waiting queue are sorted in descending order to obtain the theoretical medical treatment sequence based on priority weights; Based on the theoretical medical treatment sequence, patients whose waiting time exceeds a preset safety threshold but have a low priority weight are identified, and their sequence position within the theoretical medical treatment sequence is adaptively adjusted. After adjusting the order of patients, the clinic resources are matched, and patients are assigned to the most suitable clinic queue based on the specialty and current workload of each clinic. Calculate the change in sequence position for each patient before and after sequence rearrangement, and identify patients whose sequence position is significantly advanced and those whose sequence position is significantly delayed; for patients whose sequence position is significantly advanced, send priority treatment notifications via mobile terminals and large display screens; for patients whose sequence position is significantly delayed, send sequence adjustment instructions and expected delay durations. The appointment sequence is updated synchronously in the consultation room, and patients are called according to the new appointment sequence in the queuing system. At the same time, the triggering reason, adjustment range and number of affected patients are recorded in the emergency information system to form a sequence adjustment log.

10. The method for dynamic assessment of the condition of emergency waiting patients based on mobile terminals according to claim 1, characterized in that, Also includes: A patrol task management module is set up on the mobile terminal. The patrol task management module generates a patrol priority list based on the initial triage level and waiting time of the waiting patients, and pushes a list of target waiting patients who need to be patrolled to the nurses. After the nurse scans the waiting QR code of the target patient, the initial triage information and summary of previous rounds of assessments of the corresponding patient are displayed at the top of the clinical data collection interface. After the nurse completes the selection of qualitative descriptive terms, the mobile terminal calculates and displays the quantitative offset value and the vector of changes in the condition in real time, presenting the changing trends of each clinical observation dimension in the form of visual charts, to help the nurse intuitively judge the degree of change in the condition. When the risk index of the condition worsening reaches the high-risk or extremely high-risk range, the mobile terminal triggers a doctor's consultation suggestion and adds a consultation prompt to the priority adjustment suggestion. Nurses can initiate a doctor consultation request with one click via mobile terminal, and push information about changes in the patient's condition and current location to the on-duty doctor; Simultaneously, a dynamic assessment and monitoring dashboard for patients' conditions is set up on the computer terminal equipment at the triage station, displaying the latest inspection and assessment time, the risk index of disease deterioration, and priority adjustment records for each waiting patient in real time; for high-risk patients who have not been inspected and assessed for more than the preset time, an inspection reminder is generated and pushed to the mobile terminal.

Images