A hospital outpatient and emergency resource allocation management method

Abstract

The present application provides a hospital outpatient and emergency resource allocation management method, which relates to the technical field of hospital resource allocation; the method comprises: real-time collection of a first information set of the emergency department, triggering resource allocation strategy optimization when patient flow data reaches a first threshold, and predicting emergency resource demand within a preset time period in the future; generating an initial population comprising N initial allocation schemes, each allocation scheme being encoded as a binary chromosome, the chromosome coding region being divided into an emergency resource segment and a general resource segment, and iteratively optimizing the initial population using a genetic algorithm to obtain an optimized resource allocation strategy; the method ensures that emergency resources are reserved while general outpatient resources are flexibly adjusted as needed to avoid resource conflicts through genetic lock protection of the emergency resource segment in the genetic algorithm, allowing genetic variation of clinics and equipment, and dynamically supplementing physicians. This method not only improves the dynamic response efficiency of emergency resources, but also has a priority guarantee mechanism.

Description

Technical Field

[0001] The present application relates to the technical field of hospital resource allocation, and in particular to a hospital outpatient and emergency resource allocation management method. Background Art

[0002] In the modern medical system, the hospital's outpatient and emergency services are important links in ensuring the health of patients, but the hospital's resources are limited. There is a certain sharing relationship between outpatient and emergency services in the use of resources. Therefore, a scientific resource allocation method for outpatient and emergency services is needed.

[0003] In existing hospital resource allocation technologies, the goal is often set to maximize patient satisfaction or minimize patient waiting time, while linear programming is performed considering constraints such as resource quantity and physician working hours. However, in actual hospital environments, patients' conditions are complex and diverse, and patient flow fluctuates greatly in different time periods. Physician work efficiency is also affected by multiple factors, and the relationship between these factors is often nonlinear. This results in traditional methods lacking the ability to respond to emergencies such as sudden increases in patient flow or the needs of special patients, and having poor adaptability. Summary of the Invention

[0004] In view of the above-mentioned defects or deficiencies in the prior art, this application aims to provide a hospital outpatient and emergency resource allocation management method to improve the adaptability of hospital resource allocation; the method comprises the following steps:

[0005] collecting a first set of information from the emergency department in real time, the first set of information including patient flow data, equipment usage status, and physician on-duty information;

[0006] When the patient flow data reaches a first threshold, triggering resource allocation strategy optimization; the resource allocation strategy optimization includes:

[0007] Based on the first information set, predict the demand for emergency resources within a preset time period in the future; the emergency resource demand includes the demand for the number of emergency rooms, the demand for emergency equipment, and the demand for emergency physicians;

[0008] Generate an initial population containing N initial allocation plans. Each allocation plan is encoded as a binary chromosome. The chromosome coding region is divided into an emergency resource segment and a general resource segment. The emergency resource segment contains gene sequences that characterize the number of emergency rooms, emergency equipment allocation, and emergency physician configuration. The resource allocation amount is generated based on the predicted emergency resource demand.

[0009] Iteratively optimizing the initial population using a genetic algorithm to obtain an optimized resource allocation strategy;

[0010] The iterative optimization of the initial population using a genetic algorithm to obtain an optimized resource allocation strategy specifically includes the following steps:

[0011] Calculating chromosome fitness using a fitness function, where the fitness function includes an emergency resource guarantee penalty item, and the penalty value of the emergency resource guarantee penalty item is related to the emergency resource demand prediction value and the actual emergency resource configuration value;

[0012] Select multiple chromosomes with higher fitness as parent chromosomes, perform a single-point crossover operation on the parent chromosomes, apply gene lock protection to the emergency resource segment during the crossover process, and only allow common resource segments to perform gene exchange to obtain daughter chromosomes;

[0013] Performing a mutation operation on the offspring chromosome, wherein only the number of emergency rooms and the emergency equipment allocation genes are allowed to be mutated in the emergency resource segment;

[0014] Detecting whether the number of emergency physicians in the offspring chromosome meets a dynamic minimum threshold; if not, selecting gene sites with emergency qualifications from the physician genes in the general resource segment for supplementation;

[0015] The parent and offspring populations are merged, and the N individuals with the highest fitness are retained to continue iterative optimization until the preset number of iterations is reached or the fitness value converges. The optimal chromosome is decoded to generate the optimized resource allocation strategy.

[0016] According to the technical solution provided by the present application, the emergency equipment demand is obtained by the proportion of critically ill patients; the method further includes determining a dynamic minimum threshold;

[0017] Determining the dynamic minimum threshold comprises the following steps:

[0018] Calculate the number of basic physicians required based on the proportion of critically ill patients in the emergency resource demand;

[0019] Add a% of the current number of patients stranded in the emergency department as the buffer value;

[0020] Combined with the real-time fatigue monitoring data of on-duty physicians, the number of basic physician requirements is weightedly modified to obtain a dynamic minimum threshold.

[0021] According to the technical solution provided by this application, before applying gene lock protection to the emergency resource segment during the crossover process, the following steps are included:

[0022] A protection flag is set for the gene bit of the emergency resource segment in the chromosome coding structure; the flag of the protection flag is 1;

[0023] The crossover process applies gene lock protection to the emergency resource segment, allowing only common resource segments to undergo gene exchange, specifically including the following steps:

[0024] Traversing the identifiers of all gene bits in the chromosome coding structure, locking the gene bit identified as 1 as the emergency resource segment gene;

[0025] Cross-operation is prohibited for the emergency resource segment gene.

[0026] According to the technical solution provided by this application, the execution of the emergency resource guarantee penalty item includes: when the cumulative penalty value exceeds the second threshold, starting the chromosome repair program, which performs the following operations:

[0027] Scanning chromosomes for loci where emergency resources are not met;

[0028] Obtain the number of physicians with emergency qualifications in the general resource segment to determine whether it is sufficient to fill the gene position where emergency resources do not meet the standards;

[0029] If so, call the physician genes with emergency qualifications in the general resource segment to fill the gene positions that do not meet the standards in the emergency resource segment; if not, reallocate equipment resources from high to low according to equipment utilization rate.

[0030] According to the technical solution provided by this application, the weighted correction of the number of basic physician requirements to obtain a dynamic minimum threshold specifically includes the following steps:

[0031] According to the real-time fatigue monitoring data, the physician's fatigue level, current working hours, and historical consultation efficiency are obtained;

[0032] A three-dimensional fuzzy control rule table was constructed, with the input variables being physician fatigue level, continuous working hours, and historical reception efficiency;

[0033] Define an output correction factor μ∈[G, Q], and when the physician fatigue level reaches the third alert level, activate a mandatory correction instruction, which includes:

[0034] If the continuous working time is greater than or equal to the first preset time, the correction factor is lowered to G; if the historical reception efficiency is less than the product of the first coefficient and the average value of the emergency department, the correction factor is additionally deducted by the first deduction amount;

[0035] The dynamic minimum threshold is calculated based on the formula: dynamic minimum threshold = number of basic physician needs × (1 + a%) × μ.

[0036] According to the technical solution provided in this application, the mutation operation on the offspring chromosome specifically includes the following steps:

[0037] Establish {m×n} emergency equipment mutation probability matrix P=( ), It represents the probability of mutation of device type i in time slice j, which is given by the formula get;

[0038] Where: i represents the equipment type, j represents the time slice, m represents the number of equipment types, n represents the number of time slices, α and β represent dynamic weight coefficients, α + β = 1 and α = number of emergency waiting patients / department load limit, Indicates the equipment utilization rate, represents the forecast deviation rate;

[0039] When the device usage rate exceeds a third threshold, an anti-saturation mechanism is triggered. The anti-saturation mechanism automatically masks the mutation reduction operation of the device type and performs associated mutation on the device allocation genes of adjacent time slices.

[0040] According to the technical solution provided by this application, the protection flag can be dynamically updated, and the dynamic update includes:

[0041] After each iteration, the emergency resource allocation gap is detected, where the emergency resource allocation gap is the difference between the emergency resource demand forecast value and the actual emergency resource allocation value;

[0042] When the emergency resource allocation gap is greater than the fourth threshold, the gene lock protection range is expanded. The expanded gene lock protection range is to add an emergency backup resource segment at the end of the chromosome code, and the initial gene position flag is set to 1.

[0043] According to the technical solution provided in this application, after detecting the gap in emergency resource allocation, the following steps are also included:

[0044] When the emergency resource allocation gap is less than the fifth threshold for n consecutive iterations, the protection range is shrunk. The shrinking of the protection range is to release the identification protection of the emergency equipment allocation gene bit and convert the emergency backup resource segment into a common resource segment.

[0045] According to the technical solution provided in this application, the following steps are also included:

[0046] Retrieving a second information set from the outpatient department to construct an electronic patient file; the second information set includes the patient's historical medical data and nursing records, and the electronic patient file includes a medical timeline, a diagnosis report, a medication record, and a nursing tracking tag;

[0047] Obtaining, based on the patient electronic files, predicted target patients who will come for treatment within the future preset period, wherein the predicted target patients are patients with unclosed nursing tasks or high-risk complications;

[0048] The resource allocation strategy optimization further includes:

[0049] Applying a priority weight to the gene sequence of the associated physician corresponding to the predicted target patient, where the weight value is the product of the success rate of similar cases in history and the load pressure coefficient of the emergency department;

[0050] The fitness function is obtained by combining the priority weight and the emergency resource guarantee penalty item.

[0051] According to the technical solution provided by this application, it also includes establishing a two-way data interface between the patient side and the physician side;

[0052] The establishment of a two-way data interface between the patient and the physician comprises the following steps:

[0053] Establish a patient-side interface: Provide an electronic file query module that supports visual display of nursing records by timeline and pushes the associated physician's consultation time and emergency resource occupancy status as described in the resource allocation strategy in real time;

[0054] Establish a physician-side interface: After decoding the optimal chromosome to generate a resource allocation strategy, it automatically associates high-risk tags in the patient's electronic file within the current period and generates a dynamic reminder list. The dynamic reminder list includes unclosed nursing tasks that need to be handled first, similar case characteristics that match the current physician's historical consultation records, and equipment usage conflict warnings;

[0055] The dynamic prompt list is bound to the resource allocation plan in the resource allocation strategy. When equipment conflict or physician overload is detected, a secondary resource negotiation mechanism is triggered to select substitute resources from the idle gene bits in the general resource segment in descending order of fitness scores.

[0056] Compared with the existing technology, the beneficial effects of the present application are: the real-time data-driven dynamic trigger mechanism proposed in the present application enables the hospital to timely capture changes in the resource demand of the emergency department, especially when an emergency causes a surge in patient flow, and initiates the optimization of the resource allocation strategy, thereby improving the response capability to emergencies; the genetic algorithm is used to iteratively optimize the initial allocation plan to better adapt to the nonlinear relationship in the hospital environment. The emergency resource guarantee penalty term in the fitness function takes into account the relationship between the predicted value of emergency resource demand and the actual configuration value. This nonlinear penalty mechanism can more accurately reflect the rationality of resource allocation. It also ensures the relative stability of emergency resources through the emergency resource segment gene lock protection in the genetic algorithm, allows the clinic and equipment gene mutation, and dynamically supplements physicians. It provides a certain flexibility for the optimization of resource allocation plans, and can better coordinate the resource usage of outpatient and emergency departments. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 A flowchart of the steps of the hospital outpatient and emergency resource allocation management method provided for this application. DETAILED DESCRIPTION

[0058] The present application will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the relevant invention and are not intended to limit the invention. It should also be noted that, for ease of description, only portions relevant to the invention are shown in the accompanying drawings.

[0059] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0060] Example 1

[0061] As mentioned in the background technology, in order to solve the problems in the existing technology, this application proposes a hospital outpatient and emergency resource allocation management method, such as Figure 1 As shown, the following steps are included:

[0062] S1. Real-time collection of a first information set of the emergency department, wherein the first information set includes patient flow data, equipment usage status, and physician on-duty information;

[0063] Specifically, the first set of information is collected in real time through the hospital information system (HIS) and IoT devices. This includes patient flow data, including the number of emergency department waiting rooms, patient classification (mild / severe), and the time interval between patient arrivals. Equipment usage status includes the occupancy status and usage duration of equipment such as ventilators and ECG monitors.

[0064] Physician availability information includes the physician resource allocation table, current work status (idle / busy), and qualification certification (whether or not they are qualified for emergency care). For patient flow data, crowd counters can be installed at the entrance to the emergency department and in each waiting area to count the number of patients entering and waiting in real time. For equipment usage, status monitoring sensors are installed on each emergency device to transmit real-time status information such as device on / off, fault status, etc. to the system. Physician availability information is obtained through the hospital's attendance system and physician location system to ensure accurate information about each physician's availability and location. This data is preprocessed by edge computing nodes to filter out noise, standardize the format, and then be transmitted to a central database.

[0065] S2. When the patient flow data reaches a first threshold, triggering resource allocation strategy optimization; the resource allocation strategy optimization includes:

[0066] S21. Based on the first information set, predict the demand for emergency resources within a preset time period in the future; the emergency resource demand includes the demand for the number of emergency rooms, emergency equipment, and emergency physicians;

[0067] Specifically, when patient traffic reaches 80% of the department's capacity for three consecutive hours, resource allocation strategy optimization is triggered, sending a signal to the optimization module. Based on historical patient traffic data, a time series model is used to predict the number of clinics required. Equipment requirements are calculated based on the proportion of critically ill patients (e.g., using the disease classification labels in the electronic medical record). For example, the number of ventilators required = the number of critically ill patients × 0.8 (empirical coefficient). Emergency physician requirements are calculated based on patient classification weights, for example: .

[0068] S22. Generate an initial population containing N initial allocation plans, where each allocation plan is encoded as a binary chromosome, and the chromosome encoding region is divided into an emergency resource segment and a general resource segment. The emergency resource segment includes a gene sequence representing the number of emergency rooms, emergency equipment allocation, and emergency physician configuration, and its resource allocation amount is generated based on the predicted emergency resource demand;

[0069] Specifically, binary encoding is used to represent the allocation of clinics, equipment, and physicians. Clinic Gene Bits: 4 binary bits directly represent the number of clinics allocated to the emergency department (0-15). Equipment Gene Bits: Each bit corresponds to a type of equipment. Physician Gene Bits: Each bit corresponds to the resource allocation status of a physician. N chromosomes are randomly generated (e.g., N = 100). The initial allocation plan ensures that the allocation of emergency department resources meets 80% of the predicted demand.

[0070] For example, the predicted emergency resource demand is as follows: Emergency Rooms: Six rooms are predicted to be needed; Emergency Equipment: Assuming there are four types of equipment: ECG monitors, defibrillators, ventilators, and gastric lavage machines, the predicted demand is 3, 2, 1, and 1, respectively. Emergency Physicians: Five physicians are predicted to be needed. Chromosome Structure: Emergency Resource Segment: Clinic Room Gene Bits: 4 binary bits directly represent the number of clinics allocated to the emergency department (0-15). Equipment Gene Bits: Assuming 4 bits correspond to each of the four types of equipment (ECG monitors are bits 1-1, defibrillators are bits 2-2, ventilators are bits 3-3, and gastric lavage machines are bits 4-4). Physician Gene Bits: Assuming there are 8 physicians, 8 binary bits represent the resource allocation status of each physician. The General Resource Segment: The structure is the same as the Emergency Segment. According to the requirements, the initial allocation plan ensures that the emergency resource segment meets 80% of the forecasted demand. Therefore, in the initial allocation plan (all allocations are rounded up), there are 5 emergency rooms (the dynamic minimum threshold for emergency rooms), 3 ECG monitors, 2 defibrillators, 1 ventilator, 1 gastric lavage machine (the dynamic minimum threshold for emergency equipment), and 4 emergency physicians (the dynamic minimum threshold for emergency physicians). The emergency resource segment is coded as follows: Clinic Gene Bit: A 4-bit binary representation of the number of clinics allocated to the emergency department (value range 0-15). For example, the Clinic Gene Bit: indicates that the number of clinics allocated to the emergency department is 5, encoded as 0101 (binary 0101 corresponds to the decimal value 5), which represents the number of clinics. The Equipment Gene Bit: A 4-bit binary representation of the occupancy status of the four types of equipment. A code of 1111 indicates that the ECG monitor, defibrillator, ventilator, and gastric lavage machine are all occupied, meeting the equipment requirement calculated above. Physician gene bits: 8 binary bits indicate that 4 of the 8 physicians have resources allocated. For example, the code 11001000 indicates that physicians 1, 2, and 5 have resources allocated. The complete code for the emergency resource segment is 0101111111001000. General resource segment code: The general resource segment is randomly generated. Assume the randomly generated code is: clinic gene bit: 0011 (indicating 3 clinics allocated to the outpatient department), equipment gene bit: 0101, physician gene bit: 00110100. The complete code for the general resource segment is: 0011010100110100. The complete chromosome is the combination of the emergency resource segment and the general resource segment, resulting in the following chromosome: 01011111110010000011010100110100. Following the above method, 100 such chromosomes are randomly generated, forming an initial population containing 100 initial allocation plans.

[0071] S23, iteratively optimizing the initial population using a genetic algorithm to obtain an optimized resource allocation strategy;

[0072] The iterative optimization of the initial population using a genetic algorithm to obtain an optimized resource allocation strategy specifically includes the following steps:

[0073] S231. Calculating chromosome fitness using a fitness function, where the fitness function includes an emergency resource guarantee penalty item, wherein the penalty value of the emergency resource guarantee penalty item is related to the emergency resource demand prediction value and the actual emergency resource configuration value;

[0074] Specifically, assume the predicted demand is: 6 consulting rooms, 3 ECG monitors, and 5 doctors. The emergency section of chromosome 1 is configured as: 5 consulting rooms, 3 ECG monitors, and 4 doctors. The penalty calculation is:

[0075] Clinic punishment: ;Physician punishment: ; Total penalty = 0.5 + 0.3 = 0.8; Fitness = 1 / (1+0.8)≈0.555; Calculate the fitness of 100 chromosomes in this way;

[0076] S232. Select multiple chromosomes with higher fitness as parent chromosomes, perform a single-point crossover operation on the parent chromosomes, apply gene lock protection to the emergency resource segments during the crossover process, and only allow common resource segments to perform gene exchange, so as to obtain daughter chromosomes;

[0077] Furthermore, before applying gene lock protection to the emergency resource segment during the crossover process, the following steps are included:

[0078] A protection flag is set for the gene bit of the emergency resource segment in the chromosome coding structure; the flag of the protection flag is 1;

[0079] Specifically, within the chromosome encoding structure, protection flags are set for the gene bits representing the allocation status of each resource within the emergency resource segment. These include: the clinic gene bit (4 bits) representing the number of emergency clinics, the equipment gene bit (one bit corresponding to one piece of equipment), and the physician gene bit (one bit corresponding to one physician) representing the allocation status of emergency physicians. The protection flag is set to 1.

[0080] The crossover process applies gene lock protection to the emergency resource segment, allowing only common resource segments to undergo gene exchange, specifically including the following steps:

[0081] Traversing the identifiers of all gene bits in the chromosome coding structure, locking the gene bit identified as 1 as the emergency resource segment gene;

[0082] Cross-operation is prohibited for the emergency resource segment gene.

[0083] Specifically, after the parent chromosome is selected using the tournament selection method, a single-point crossover operation is performed on the parent chromosome:

[0084] For example, the common segment of chromosome 1 is: 0011010100110100

[0085] Common segment of chromosome 2: 0101010001101000

[0086] The intersection point is after the 5th position in the general segment (i.e. the 21st position after the end of the emergency segment):

[0087] Chromosome 1 common segment decomposition: first 5 bits "00110" + last 11 bits "10100110100"

[0088] Chromosome 2 common segment decomposition: first 5 bits "01010" + last 11 bits "10001101000"

[0089] After crossing, it generates:

[0090] Child 1 normal segment: "01010" + "10100110100" = 0101010100101100

[0091] Child 2 normal segment: "00110" + "10001101000" = 0011010001101000

[0092] Final daughter chromosomes:

[0093] Offspring 1: Chromosome 1 emergency segment (0101111111001000) + new common segment (0101010100101100)

[0094] Offspring 2: chromosome 2 emergency segment (0111111111110000) + new common segment (0011010001101000).

[0095] S233. Perform a mutation operation on the offspring chromosome, wherein only the number of emergency rooms and the emergency equipment allocation genes are allowed to be mutated in the emergency resource segment;

[0096] Specifically, the sentence "Only the number of emergency clinics and emergency equipment allocation genes are allowed to mutate in the emergency resource segment" means "Only the gene bits representing the number of emergency clinics and the gene bits representing the emergency equipment allocation status are allowed to mutate in the emergency resource segment". The mutation range is: Emergency resource segment: only the number of clinics gene and equipment allocation gene are allowed to mutate. General resource segment: All genes can mutate, mutation rule: randomly flip gene bits with a probability of 2% (mutation between 0 and 1). After the clinic gene mutates, the rationality needs to be verified (such as 3-bit binary represents that the number of clinics does not exceed 8). For example, the original code of the chromosome 1 emergency segment: 110111111001000 may mutate to: 111111111001000 (clinic gene changes from 110→111, and the number of clinics changes from 5→7).

[0097] S234: Detect whether the number of emergency physicians in the offspring chromosome meets a dynamic minimum threshold; if not, select gene sites with emergency qualifications from the physician genes in the general resource segment for supplementation;

[0098] Specifically, detecting whether the number of emergency physicians in the offspring chromosomes meets the dynamic minimum threshold means "detecting whether the number of emergency physicians calculated from the gene positions representing the configuration status of emergency physicians in the offspring chromosomes meets the dynamic minimum threshold."

[0099] Furthermore, the execution of the emergency resource guarantee penalty item includes: when the cumulative penalty value exceeds a second threshold, starting a chromosome repair program, which performs the following operations:

[0100] Scanning chromosomes for loci where emergency resources are not met;

[0101] Specifically, the gene positions in the chromosomes where emergency resources do not meet the standards are scanned; for clinic resources, it means that the value represented by the gene position of the number of clinics is less than the required value; for equipment and physician resources, it means that the occupied quantity or status represented by the corresponding gene position does not meet the requirements.

[0102] Obtain the number of physicians with emergency qualifications in the general resource segment to determine whether it is sufficient to fill the gene position where emergency resources do not meet the standards;

[0103] If so, call the physician genes with emergency qualifications in the general resource segment to fill the gene positions that do not meet the standards in the emergency resource segment; if not, reallocate equipment resources from high to low according to equipment utilization rate.

[0104] Specifically, the second threshold is 100. If the number of emergency physicians in the offspring chromosome is less than the dynamic minimum threshold for emergency physicians, an unoccupied physician with emergency qualifications is selected from the general segment. The corresponding gene bit in the general segment is set from 0 to 1, and the corresponding bit in the emergency segment is set from 0 to 1.

[0105] S235. Merge the parent and offspring populations, retain the N individuals with the highest fitness, and continue iterative optimization until the preset number of iterations is reached or the fitness value converges, decode the optimal chromosome and generate the optimized resource allocation strategy.

[0106] Specifically, the parent generation (50) and the offspring generation (50) are merged into 100 chromosomes, which are sorted by fitness, and the top 100 individuals are retained for the next generation. When the maximum number of iterations is reached (e.g., 200 generations) or the fitness value changes by <1% for 10 consecutive generations, the chromosome with the highest fitness is selected as the optimal chromosome, and the optimal chromosome is decoded to obtain a resource allocation plan for emergency and general outpatient departments. This solution can quickly generate a resource allocation strategy that balances efficiency and fairness when emergency resources are tight, effectively improving the responsiveness of emergency departments, pre-deploying resources for impending surges in emergency visits, and overcoming situations where emergency resources are insufficient or outpatient resources are overly squeezed. This solution is more suitable for medium- and short-term resource scheduling (e.g., triggering optimization every 2 hours).

[0107] In a preferred embodiment, the emergency equipment demand is obtained by the proportion of critically ill patients; the method further comprises determining a dynamic minimum threshold;

[0108] Determining the dynamic minimum threshold comprises the following steps:

[0109] Calculate the number of basic physicians required based on the proportion of critically ill patients in the emergency resource demand;

[0110] Add a% of the current number of patients stranded in the emergency department as the buffer value;

[0111] Combined with the real-time fatigue monitoring data of on-duty physicians, the number of basic physician requirements is weightedly modified to obtain a dynamic minimum threshold.

[0112] Specifically, the number of required primary care physicians is calculated based on the proportion of critically ill patients in the predicted emergency resource demand. For example, if one physician is required for every five critically ill patients and 20 critically ill patients are predicted within the next 24 hours, the number of required primary care physicians is 20 ÷ 5 = 4. A buffer value is added: Assuming the current number of patients in the emergency department is 30 and a% is 20%, the buffer value is 30 × 20% = 6. Based on the standard of one physician for every five patients, the number of physicians corresponding to the buffer value is 6 ÷ 5 = 1.2, rounded up to 2. A weighted correction is applied to obtain a dynamic minimum threshold: a weighted correction is performed based on real-time fatigue monitoring data of on-duty physicians. Real-time fatigue monitoring data can be obtained through wearable devices, such as smart bracelets, which monitor the physician's heart rate, step count, and other data. Physician fatigue is then calculated based on a pre-set algorithm. Assuming that the number of basic physicians required is 4, plus 2 physicians corresponding to the buffer value, the total number is 6. Based on the fatigue monitoring data, the correction factor μ = 0.8 is determined, then the dynamic minimum threshold = 6 × 0.8 = 4.8, rounded up to 5.

[0113] Furthermore, performing weighted correction on the number of basic physician requirements to obtain a dynamic minimum threshold specifically includes the following steps:

[0114] According to the real-time fatigue monitoring data, the physician's fatigue level, current working hours, and historical consultation efficiency are obtained;

[0115] Specifically, real-time fatigue monitoring data is obtained through smart bracelets and hospital information systems.

[0116] A three-dimensional fuzzy control rule table was constructed, with the input variables being physician fatigue level, continuous working hours, and historical reception efficiency;

[0117] For example, the physician fatigue level is level 2, the current working hours are 8 hours, and the historical reception efficiency is 3 patients per hour. By querying the fuzzy control rule table, the correction factor μ = 0.9 is obtained;

[0118] Define an output correction factor μ∈[G, Q], and when the physician fatigue level reaches the third alert level, activate a mandatory correction instruction, which includes:

[0119] If the continuous working time is greater than or equal to the first preset time, the correction factor is lowered to G; if the historical reception efficiency is less than the product of the first coefficient and the average value of the emergency department, the correction factor is additionally deducted by the first deduction amount;

[0120] For example, when a physician's fatigue level reaches level three, a mandatory correction is activated. If the continuous working time is greater than or equal to a first preset duration (e.g., 10 hours), the correction factor is lowered to G (the lower limit of the correction factor, e.g., 0.7). If the historical patient reception efficiency is less than the product of the first coefficient (e.g., 0.8) and the emergency department average (e.g., 4 patients per hour) (i.e., 0.8 × 4 = 3.2), the correction factor is further deducted by the first deduction amount (e.g., 0.1).

[0121] The dynamic minimum threshold is calculated based on the formula: dynamic minimum threshold = number of basic physician needs × (1 + a%) × μ.

[0122] For example, assuming that the number of basic physicians required is 5, a% is 20%, and the correction factor μ = 0.9, then the dynamic minimum threshold = 5×(1 + 20%)×0.9 = 5.4, rounded up to 6.

[0123] This implementation fully details the impact of individual physician status on the demand threshold. By factoring in individual physician status differences and introducing fuzzy control rules, the dynamic minimum threshold is dynamically adjusted. When physician fatigue levels are high, continuous work hours are long, and historical patient efficiency is low, the fuzzy control rule table can appropriately adjust the correction factor downward based on pre-set rules, thereby increasing the dynamic minimum threshold and ensuring sufficient physician resources to support emergency medical services.

[0124] In a preferred embodiment, performing a mutation operation on the offspring chromosome specifically comprises the following steps:

[0125] Establish {m×n} emergency equipment mutation probability matrix P=( ), It represents the probability of mutation of device type i in time slice j, which is given by the formula get;

[0126] Where: i represents the equipment type, j represents the time slice, m represents the number of equipment types, n represents the number of time slices, α and β represent dynamic weight coefficients, α + β = 1 and α = number of emergency waiting patients / department load limit, Indicates the equipment utilization rate, represents the forecast deviation rate;

[0127] Specifically, the prediction deviation rate is an estimate of the possible error in the prediction results due to the incompleteness, uncertainty and limitations of the prediction model itself when predicting the demand for emergency equipment in the future preset period. The matrix P = ( ) is determined by the number of device types m and the number of time slices n. For example, when there are three types of devices (such as ventilators, ECG monitors, and defibrillators) and time is divided into four time slices (for example, one time slice every 6 hours, and a total of 4 time slices per day), we will get a Each element in the matrix Represents the probability that device type i will mutate within time slice j.

[0128] When the device usage rate exceeds a third threshold, an anti-saturation mechanism is triggered. The anti-saturation mechanism automatically masks the mutation reduction operation of the device type and performs associated mutation on the device allocation genes of adjacent time slices.

[0129] Specifically, the anti-saturation mechanism is designed to prevent unreasonable device variability during high utilization. For example, if the utilization of a device (such as a ventilator) exceeds 80% during the second time slice, the device's variability reduction operation will be prohibited during the second time slice. This prevents the device from experiencing a reduction in allocated capacity during this busy time slice, ensuring the normal use of emergency equipment.

[0130] At the same time, linked mutation is performed on the device allocation genes for adjacent time slices (the first and third time slices). This takes into account the temporal correlation of device usage. High utilization in one time slice may indicate high demand in the adjacent time slice. Linked mutation allows for more balanced device allocation over time, avoiding the irrational allocation of device resources that would otherwise occur if mutations were performed on each time slice in isolation. This implementation not only dynamically adjusts the mutation probability based on accuracy but also triggers an anti-saturation mechanism for frequently used devices to prevent frequently used devices from being under-allocated.

[0131] In a preferred embodiment, the protection flag can be dynamically updated, and the dynamic update includes:

[0132] After each iteration, the emergency resource allocation gap is detected, where the emergency resource allocation gap is the difference between the emergency resource demand forecast value and the actual emergency resource allocation value;

[0133] When the emergency resource allocation gap is greater than the fourth threshold, the gene lock protection range is expanded. The expanded gene lock protection range is to add an emergency backup resource segment at the end of the chromosome code, and the initial gene position flag is set to 1.

[0134] Furthermore, after detecting the gap in emergency resource allocation, the method further includes the following steps:

[0135] When the emergency resource allocation gap is less than the fifth threshold for n consecutive iterations, the protection range is shrunk. The shrinking of the protection range is to release the identification protection of the emergency equipment allocation gene bit and convert the emergency backup resource segment into a common resource segment.

[0136] Specifically, dynamic adjustments to the gene lock's protection range prevent excessive fluctuations in emergency resources during the iteration process. When resource gaps are large, protection is strengthened to prevent further resource loss; while when resource gaps are small, the protection range contraction mechanism can appropriately relax restrictions, allowing for more flexible resource allocation. This balanced adjustment approach helps maintain relative stability in emergency resource allocation and ensures the continued operation of emergency departments.

[0137] In a preferred embodiment, the following steps are also included:

[0138] Retrieving a second information set from the outpatient department to construct an electronic patient file; the second information set includes the patient's historical medical data and nursing records, and the electronic patient file includes a medical timeline, a diagnosis report, a medication record, and a nursing tracking tag;

[0139] Specifically, historical patient data: Patient diagnosis records, medication records, and surgical records are extracted from the hospital information system (HIS) in a structured database table (such as MySQL). Nursing records: Real-time nursing data, including wound care time, infusion completion status, and vital sign monitoring records, is collected through mobile nursing devices (PDAs) in a JSON stream format. Unstructured nursing records (such as handwritten notes by nurses) are subjected to optical character recognition (OCR) and converted into structured data. Timestamps are used to align data from different sources to ensure timeline consistency.

[0140] Based on a cloud-based database, we build electronic patient records. The cloud-based database is constructed as follows: Patient unique ID: Using the medical insurance number as the primary key, it links emergency and outpatient data. Visit timeline: This stores all patient visits in chronological order (e.g., "2023-10-01 09:00 Emergency Department admission"). Diagnostic reports: This stores CT and MRI reports and physicians' handwritten diagnoses in PDF format. Medication records: This stores drug name, dosage, and time of use in a JSON format (e.g., {"drug":"aspirin", "dose":"100mg", "time":"2023-10-01 10:00"}). Nursing tracking tags: This automatically generates tags using natural language processing (NLP). For example, unclosed nursing tasks: This extracts uncompleted nursing items from nursing records (e.g., "postoperative drainage tube not removed"). High-risk complication tags: This matches risk levels based on diagnostic keywords (e.g., "diabetic foot" and "deep vein thrombosis"). Data synchronization mechanism: Use message queues (such as Kafka) to synchronize emergency and outpatient data to the cloud in real time, ensuring that the delay in electronic file updates is less than 1 second.

[0141] Obtaining, based on the patient electronic files, predicted target patients who will come for treatment within the future preset period, wherein the predicted target patients are patients with unclosed nursing tasks or high-risk complications;

[0142] Specifically, by comparing the expected completion time of the task with the current time, we determine whether the task is in an open state. For example, if a nursing task is expected to be completed within a week, but five days have passed and it has not yet been marked as completed, then the task is an open nursing task within the preset future time period. Using machine learning algorithms (such as logistic regression, decision trees, random forests, etc.) or statistical methods, we construct a complication risk prediction model based on historical patient data. The extracted risk factors are used as input variables, and the occurrence of complications is used as the output variable to train and optimize the model. The model calculates the probability of each patient developing a high-risk complication within the preset future time period, thereby predicting the target patient.

[0143] The resource allocation strategy optimization further includes:

[0144] Applying a priority weight to the gene sequence of the associated physician corresponding to the predicted target patient, where the weight value is the product of the success rate of similar cases in history and the load pressure coefficient of the emergency department;

[0145] The fitness function is obtained by combining the priority weight and the emergency resource guarantee penalty item.

[0146] Specifically, when resource allocation strategy optimization is triggered, the care tags of patients visiting during the current period are retrieved from electronic records. Priority weight calculation: Historical success rate for similar cases: The physician's success rate for similar cases over the past year is extracted from the physician performance database (e.g., success rate = number of patients cured / total number of patients seen). Emergency department load pressure coefficient: Calculated in real time as the current number of patients waiting / department capacity limit, with a value range of [0, 1]. Weight formula: Weight = success rate × load pressure coefficient. For example, if a physician has an 80% success rate for similar cases and the current load pressure coefficient is 0.6, then the weight = 0.8 × 0.6 = 0.48. Fitness function update: The original fitness function (including the emergency resource penalty term) is expanded to: Fitness value = base fitness + ∑(priority weight) - penalty value.

[0147] This embodiment is based on the dynamic optimization of the priority of the patient's electronic file, combined with the outpatient treatment data, to establish a patient file, and predicts based on the patient file whether patients with unclosed nursing tasks or high-risk complications will come for treatment in the future period. If they come for treatment, the physicians related to their conditions will be given a high weight to meet the patients' personalized needs.

[0148] In a preferred embodiment, it also includes establishing a two-way data interface between the patient side and the physician side;

[0149] The establishment of a two-way data interface between the patient and the physician comprises the following steps:

[0150] Establish a patient-side interface: Provide an electronic file query module that supports visual display of nursing records by timeline and pushes the associated physician's consultation time and emergency resource occupancy status as described in the resource allocation strategy in real time;

[0151] Specifically, timeline visualization uses the ECharts library to generate interactive timelines, supporting clickable events for detailed information. Nursing record display displays data in a collapsed format by category (e.g., infusion, wound care), with key events (e.g., "postoperative infection alert") highlighted. Real-time push notifications push resource allocation policy updates to patients' mobile devices via the WebSocket protocol.

[0152] Establish a physician-side interface: After decoding the optimal chromosome to generate a resource allocation strategy, it automatically associates high-risk tags in the patient's electronic file within the current period and generates a dynamic reminder list. The dynamic reminder list includes unclosed nursing tasks that need to be handled first, similar case characteristics that match the current physician's historical consultation records, and equipment usage conflict warnings;

[0153] Specifically, high-risk label association: After decoding the resource allocation strategy, a reminder list is generated based on the label matching rules in the patient's electronic file (such as "diabetic foot → high risk endocrinology department"). List content: Unclosed nursing tasks: Uncompleted items are extracted from the nursing record (such as "Patient A did not complete the postoperative dressing change").

[0154] Similar Case Characteristics: Based on the physician's historical consultation data, the current patient's similarity with historical cases is matched (e.g., "85% similarity to 2022 case #123"). Equipment Conflict Alert: Detects equipment allocation conflicts in resource allocation strategies (e.g., "The ventilator is overloaded at 2:00 PM"). Checklist Binding and Display: Embed dynamic reminder lists into the physician's workstation interface, sorted by urgency.

[0155] The dynamic prompt list is bound to the resource allocation plan in the resource allocation strategy. When equipment conflict or physician overload is detected, a secondary resource negotiation mechanism is triggered to select substitute resources from the idle gene bits in the general resource segment in descending order of fitness scores.

[0156] Specifically, conflict detection: Equipment conflict: Checks whether the equipment allocation gene bit exceeds the actual number of devices (e.g., the ventilator gene bit is "1111" but there are only 3 in reality). Physician overload: Determines whether the physician's continuous working hours exceed 8 hours. Backup resource selection: Fitness score calculation: For idle gene bits in the general resource segment (e.g., unoccupied physicians or equipment), calculate their fitness with current demand: fitness score = resource idle rate × historical utilization rate; backup call rule: Select resources in descending order of fitness score (e.g., the backup ventilator with the highest score is called first). Update the chromosome gene bit to mark the backup resource as occupied in the emergency segment (the gene bit changes from "0" to "1").

[0157] For example, when patient Zhang's electronic file displayed the label "Postoperative Observation," the system automatically assigned him a primary physician with similar surgical experience, improving consultation efficiency by 30%. A dynamic reminder list also appears: Dr. Li's workstation receives a reminder: "Patient Wang's ventilator is not closed. Nursing task: Postoperative drainage tube maintenance; Equipment conflict: 2 ventilators overloaded at 2:00 PM." This triggers the system to automatically dispatch backup ventilators from the general outpatient clinic.

[0158] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. The above is only the preferred implementation method of this application. It should be pointed out that due to the limitations of textual expression, there are objectively infinite specific structures. For ordinary technicians in this technical field, without departing from the principles of the present invention, they can also make several improvements, modifications or changes, and can also combine the above technical features in an appropriate manner; these improvements, modifications, changes or combinations, or the direct application of the inventive concept and technical solution to other occasions without improvement, should be regarded as the scope of protection of this application.

Claims

1. A hospital outpatient and emergency resource allocation management method, characterized in that: The following steps are involved: collecting a first set of information from the emergency department in real time, the first set of information including patient flow data, equipment usage status, and physician on-duty information; When the patient flow data reaches a first threshold, triggering resource allocation strategy optimization; The resource allocation strategy optimization includes: Based on the first information set, predict the demand for emergency resources within a preset time period in the future; the emergency resource demand includes the demand for the number of emergency rooms, the demand for emergency equipment, and the demand for emergency physicians; Generate an initial population containing N initial allocation plans. Each allocation plan is encoded as a binary chromosome. The chromosome coding region is divided into an emergency resource segment and a general resource segment. The emergency resource segment contains gene sequences that characterize the number of emergency rooms, emergency equipment allocation, and emergency physician configuration. The resource allocation amount is generated based on the predicted emergency resource demand. Iteratively optimizing the initial population using a genetic algorithm to obtain an optimized resource allocation strategy; The iterative optimization of the initial population using a genetic algorithm to obtain an optimized resource allocation strategy specifically includes the following steps: Calculating chromosome fitness using a fitness function, where the fitness function includes an emergency resource guarantee penalty item, and the penalty value of the emergency resource guarantee penalty item is related to the emergency resource demand prediction value and the actual emergency resource configuration value; Select multiple chromosomes with higher fitness as parent chromosomes, perform a single-point crossover operation on the parent chromosomes, apply gene lock protection to the emergency resource segment during the crossover process, and only allow common resource segments to undergo gene exchange to obtain daughter chromosomes; Performing a mutation operation on the offspring chromosome, wherein only the number of emergency rooms and the emergency equipment allocation genes are allowed to be mutated in the emergency resource segment; Detecting whether the number of emergency physicians in the offspring chromosome meets a dynamic minimum threshold; if not, selecting gene sites with emergency qualifications from the physician genes in the general resource segment for supplementation; The parent and offspring populations are merged, and the N individuals with the highest fitness are retained to continue iterative optimization until the preset number of iterations is reached or the fitness value converges. The optimal chromosome is decoded to generate the optimized resource allocation strategy.

2. The hospital outpatient and emergency resource allocation management method according to claim 1, characterized in that: The emergency equipment demand is obtained from the proportion of critically ill patients; the method also includes determining a dynamic minimum threshold; Determining the dynamic minimum threshold comprises the following steps: Calculate the number of basic physicians required based on the proportion of critically ill patients in the emergency resource demand; Add a% of the current number of patients stranded in the emergency department as the buffer value; Combined with the real-time fatigue monitoring data of on-duty physicians, the number of basic physician requirements is weightedly modified to obtain a dynamic minimum threshold.

3. The hospital outpatient and emergency resource allocation management method according to claim 1, characterized in that: Before applying gene lock protection to the emergency resource segment during the crossover process, the following steps are included: A protection flag is set for the gene bit of the emergency resource segment in the chromosome coding structure; the flag of the protection flag is 1; The crossover process applies gene lock protection to the emergency resource segment, allowing only common resource segments to undergo gene exchange, specifically including the following steps: Traversing the identifiers of all gene bits in the chromosome coding structure, locking the gene bit identified as 1 as the emergency resource segment gene; Cross-operation is prohibited for the emergency resource segment gene.

4. The hospital outpatient and emergency resource allocation management method according to claim 1, characterized in that: The execution of the emergency resource guarantee penalty item includes: when the cumulative penalty value exceeds the second threshold, starting a chromosome repair program, which performs the following operations: Scanning chromosomes for loci where emergency resources are not met; Obtain the number of physicians with emergency qualifications in the general resource segment to determine whether it is sufficient to fill the gene position where emergency resources do not meet the standards; If so, call the gene of the physician with emergency qualification in the general resource segment to fill the gene position that does not meet the standard in the emergency resource segment; if not, reallocate the equipment resources from high to low according to the equipment utilization rate.

5. The hospital outpatient and emergency resource allocation management method according to claim 2, characterized in that: The weighted correction of the number of basic physician needs to obtain a dynamic minimum threshold specifically includes the following steps: According to the real-time fatigue monitoring data, the physician's fatigue level, current working hours, and historical consultation efficiency are obtained; A three-dimensional fuzzy control rule table was constructed, with the input variables being physician fatigue level, continuous working hours, and historical reception efficiency; Define an output correction factor μ∈[G, Q], and when the physician fatigue level reaches the third alert level, activate a mandatory correction instruction, which includes: If the continuous working time is greater than or equal to the first preset time, the correction factor is lowered to G; if the historical reception efficiency is less than the product of the first coefficient and the average value of the emergency department, the correction factor is additionally deducted by the first deduction amount; The dynamic minimum threshold is calculated based on the formula: dynamic minimum threshold = number of basic physician needs × (1 + a%) × μ.

6. The hospital outpatient and emergency resource allocation management method according to claim 3, characterized in that: The mutation operation on the offspring chromosome specifically comprises the following steps: Establish {m×n} emergency equipment mutation probability matrix P=( ), It represents the probability of mutation of device type i in time slice j, which is given by the formula get; Where: i represents the equipment type, j represents the time slice, m represents the number of equipment types, n represents the number of time slices, α and β represent dynamic weight coefficients, α + β = 1 and α = number of emergency waiting patients / department load limit, Indicates the equipment utilization rate, represents the forecast deviation rate; When the device usage rate exceeds a third threshold, an anti-saturation mechanism is triggered. The anti-saturation mechanism automatically masks the mutation reduction operation of the device type and performs associated mutation on the device allocation genes of adjacent time slices.

7. The hospital outpatient and emergency resource allocation management method according to claim 3, characterized in that: The protection flag can be dynamically updated, and the dynamic update includes: After each iteration, the emergency resource allocation gap is detected, where the emergency resource allocation gap is the difference between the emergency resource demand forecast value and the actual emergency resource allocation value; When the emergency resource allocation gap is greater than the fourth threshold, the gene lock protection range is expanded. The expanded gene lock protection range is to add an emergency backup resource segment at the end of the chromosome code, and the initial gene position flag is set to 1.

8. The hospital outpatient and emergency resource allocation management method according to claim 7, characterized in that: After detecting the gap in emergency resource allocation, the following steps are also included: When the emergency resource allocation gap is less than the fifth threshold for n consecutive iterations, the protection range is shrunk. The shrinking of the protection range is to release the identification protection of the emergency equipment allocation gene bit and convert the emergency backup resource segment into a common resource segment.

9. The hospital outpatient and emergency resource allocation management method according to claim 1, characterized in that: The following steps are also included: Retrieving a second information set from the outpatient department to construct an electronic patient file; the second information set includes the patient's historical medical data and nursing records, and the electronic patient file includes a medical timeline, a diagnosis report, a medication record, and a nursing tracking tag; Obtaining, based on the patient electronic files, predicted target patients who will come for treatment within the future preset period, wherein the predicted target patients are patients with unclosed nursing tasks or high-risk complications; The resource allocation strategy optimization further includes: Applying a priority weight to the gene sequence of the associated physician corresponding to the predicted target patient, where the weight value is the product of the success rate of similar cases in history and the load pressure coefficient of the emergency department; The fitness function is obtained by combining the priority weight and the emergency resource guarantee penalty item.

10. The hospital outpatient and emergency resource allocation management method according to claim 9, characterized in that: It also includes establishing a two-way data interface between the patient and physician sides; The establishment of a two-way data interface between the patient and the physician comprises the following steps: Establish a patient-side interface: Provide an electronic file query module that supports visual display of nursing records by timeline and pushes the associated physician's consultation time and emergency resource occupancy status as described in the resource allocation strategy in real time; Establish a physician-side interface: After decoding the optimal chromosome to generate a resource allocation strategy, it automatically associates high-risk tags in the patient's electronic file within the current period and generates a dynamic reminder list. The dynamic reminder list includes unclosed nursing tasks that need to be handled first, similar case characteristics that match the current physician's historical consultation records, and equipment usage conflict warnings; The dynamic prompt list is bound to the resource allocation plan in the resource allocation strategy. When equipment conflict or physician overload is detected, a secondary resource negotiation mechanism is triggered to select substitute resources from the idle gene bits in the general resource segment in descending order of fitness scores.

Images