A medicine consumable intelligent purchasing method and system

Abstract

The present disclosure relates to a kind of medicine consumable wisdom purchase method and system, method includes: obtaining hospital historical medicine purchase data, patient visit data and epidemiology data;Adopt medicine demand prediction model, according to the hospital historical medicine purchase data, patient visit data, epidemiology data, first medicine purchase data are determined to be predicted;Determine medicine purchase adjustment adaptability coefficient, according to the medicine purchase adjustment adaptability coefficient, the first medicine purchase data are adjusted, and second medicine purchase data are obtained.The technical scheme of the present application can make the system or business logic more suitable for actual scene changes by determining the medicine purchase adjustment adaptability coefficient.The coefficient follows dynamic change, reduces the error caused by fixed value, makes medicine purchase prediction more accurate, and improves data accuracy.

Description

Technical Field

[0001] This disclosure relates to the field of medical technology, and in particular to a smart procurement method and system for pharmaceuticals and consumables. Background Technology

[0002] In related technologies, hospital drug procurement mainly relies on pharmacy department heads to make purchases based on inventory and experience, lacking scientific estimation. This leads to a significant discrepancy between procurement demand and actual usage, resulting in drug stockpiling or shortages. This negatively impacts the patient's healthcare experience. Summary of the Invention

[0003] To overcome the problems existing in the related technologies, this disclosure provides a smart procurement method and system for pharmaceutical consumables to solve the above problems.

[0004] According to a first aspect of the present disclosure, a smart procurement method for pharmaceutical consumables is provided, comprising:

[0005] Obtain historical drug procurement data, patient visit data, and epidemiological data from hospitals;

[0006] Using a drug demand forecasting model, the first drug procurement data is predicted and determined based on the hospital's historical drug procurement data, patient visit data, and epidemiological data.

[0007] Determine the adaptability coefficient for drug procurement adjustments, specifically including:

[0008] The adaptability coefficient for drug procurement adjustments is determined using the following methods:

[0009] Determine the doctor coefficient, patient coefficient, and seasonal coefficient;

[0010] Based on external characteristics, the doctor coefficient, the patient coefficient, and the seasonal coefficient are nonlinearly refined and adjusted to obtain the drug procurement adjustment adaptability coefficient.

[0011] The drug procurement adjustment adaptability coefficient is obtained by jointly calculating the doctor coefficient, the patient coefficient, and the seasonal coefficient.

[0012] The drug procurement adjustment adaptability coefficient is determined based on the deviation between the predicted and actual values ​​of the hospital's historical drug procurement data.

[0013] Based on the drug procurement adjustment adaptability coefficient, the first drug procurement data is adjusted to obtain the second drug procurement data.

[0014] Secondly, this application proposes a smart procurement system for pharmaceuticals and consumables, comprising: a processor; and a memory for storing processor-executable instructions;

[0015] The processor is configured to execute the intelligent procurement method for pharmaceuticals and consumables described in any of the preceding embodiments.

[0016] The technical solutions provided by the embodiments of this disclosure may include the following beneficial effects:

[0017] Compared with existing technologies, the technical solution of this application, by determining the adaptability coefficient for drug procurement adjustments and adjusting the calculated drug procurement data, makes the system or business logic more closely aligned with actual scenario changes, resulting in more accurate drug procurement predictions and improved data precision. This invention proposes several calculation methods for the adaptability coefficient for drug procurement adjustments: nonlinear refinement adjustments to the doctor coefficient, patient coefficient, and seasonal coefficient based on external characteristics; joint calculation of the doctor coefficient, patient coefficient, and seasonal coefficient; and determination of the adaptability coefficient based on the deviation between predicted and actual values ​​of historical drug procurement data from hospitals. This multi-dimensional approach to closely approximating real-world drug procurement conditions makes the adaptability coefficient for drug procurement adjustments more accurate and better suited to actual scenarios.

[0018] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0019] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0020] Figure 1 This is a flowchart illustrating a smart procurement method for pharmaceutical consumables according to an exemplary embodiment;

[0021] Figure 2 This is a drug search page illustrated according to an exemplary embodiment;

[0022] Figure 3 This is an example of a timeliness setting page;

[0023] Figure 4 This is a statistics page illustrated according to an exemplary embodiment;

[0024] Figure 5 This is a structural block diagram of a smart procurement system for pharmaceutical consumables, according to an exemplary embodiment. Detailed Implementation

[0025] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0026] It should be noted that all actions involving the acquisition of signals, information, or data in this application are carried out in compliance with the relevant data protection laws and policies of the country where the application is located, and with the authorization granted by the owner of the relevant device.

[0027] Based on this, this application proposes a smart procurement method for pharmaceuticals and consumables, see appendix. Figure 1 The methods include:

[0028] In step S102, historical drug procurement data, patient visit data, and epidemiological data of the hospital are obtained.

[0029] In this embodiment, based on epidemiological and hospital visit characteristics, historical drug procurement data, patient visit data, and epidemiological data from hospitals are collected to analyze drug demand trends and influencing factors. Data sources are as follows: Existing epidemiological data: obtained through the rational drug use system. Drug usage during epidemics: obtained through the rational drug use system.

[0030] In step S104, a drug demand forecasting model is used to predict and determine the first drug procurement data based on the hospital's historical drug procurement data, patient visit data, and epidemiological data.

[0031] In this embodiment, regression analysis is used to establish a drug demand forecasting model, which can predict the demand for drugs, thereby optimizing procurement plans and improving inventory management efficiency. Utilizing an intelligent procurement system for accurate predictive procurement of drug demand avoids the pitfalls of traditional high-low shelf stocking, improving the accuracy and timeliness of procurement plans.

[0032] In some embodiments, by combining epidemic trend prediction (influenza as an example), by collecting and statistically analyzing data monitored in a region's influenza surveillance information system in previous years, it is possible to understand the peak period of influenza incidence in the city and whether there are temporal variation characteristics; by fitting and analyzing historical influenza data through an auto-regressive moving average model (ARIMA), it is possible to predict the influenza incidence trend in the future, providing a scientific basis for formulating drug procurement plans.

[0033] For example, from 2022 to 2024, a total of 1,205,887 cases of outpatient and emergency visits were collected in a certain region, of which 38,793 were influenza-like illness (ILI), with an ILI rate of 3.22%. Before August 2022, the influenza epidemic was relatively stable. The number of influenza cases gradually increased starting in August, peaking in November before gradually declining, but the number of cases in December was still higher than in the previous eight months. In 2023, the influenza peak occurred in July and August, with secondary peaks in August and November. In 2024, the number of influenza cases gradually increased starting in February, peaking in March, gradually declining starting in April, remaining stable from May to September, increasing again in October, and reaching a secondary peak in December.

[0034] In some embodiments, the aforementioned drug demand forecasting model can be established using an Autoregressive Integrated Moving Average (ARIMA) model combined with a Long Short-Term Memory (LSTM) network. The ARIMA model is used to predict the linear characteristics of the drug consumption sequence, generating corresponding residual sequences. The nonlinear characteristics of the original sequence are embedded in these residual sequences. The powerful nonlinear data processing capabilities of the LSTM model are used to correct these residual data sequences. The linear prediction results from the ARIMA model are combined with the residual prediction values ​​from the LSTM model to obtain the predicted drug consumption value. Step S106, determining the drug procurement adjustment adaptability coefficient, specifically includes:

[0035] The following method is used to determine the adaptability coefficient for drug procurement adjustments.

[0036] Determine the doctor coefficient, patient coefficient, and seasonal coefficient.

[0037] Among them, the seasonal coefficient is a coefficient set according to the changes in the demand for medicines in the market in different seasons, and is used to reflect the impact of seasonal factors on the amount of medicines purchased.

[0038] The physician coefficient, determined by physician-related factors (such as physicians' prescribing habits, number of patients treated, and medication preferences), is used to quantify the impact of physician factors on drug procurement volume.

[0039] The patient coefficient, determined by patient-related factors (such as the total number of patients, case structure, severity, and duration of medication), is used to quantify the impact of patient factors on drug procurement volume.

[0040] The first approach involves nonlinearly refining the doctor coefficient, patient coefficient, and seasonal coefficient based on external characteristics to obtain the drug procurement adjustment adaptability coefficient.

[0041] Alternatively, the second method can be used to jointly calculate the doctor coefficient, patient coefficient, and seasonal coefficient to obtain the drug procurement adjustment adaptability coefficient.

[0042] Alternatively, a third approach is to determine the drug procurement adjustment adaptability coefficient based on the deviation between the predicted and actual values ​​of drug procurement for each set of historical data.

[0043] The three methods described above can be used individually or in combination to determine the adaptability coefficient for drug procurement adjustments. For example, the first method can be used first, followed by the third method to obtain the adaptability coefficient. Alternatively, the second method can be used first, followed by the third method. Or, the first method can be used first, followed by the second method. Or, all three methods can be used to obtain the adaptability coefficient for drug procurement adjustments.

[0044] In step S108, the first drug procurement data is adjusted according to the drug procurement adjustment adaptability coefficient to obtain the second drug procurement data.

[0045] The technical solution of this application, by determining the adaptability coefficient for drug procurement adjustments and adjusting the calculated drug procurement data, enables the system or business logic to better align with actual scenario changes. This leads to more accurate drug procurement forecasts and improved data precision.

[0046] This invention proposes several calculation methods for the adaptability coefficient of drug procurement adjustments: First, nonlinear refinement adjustments are made to the doctor coefficient, patient coefficient, and seasonal coefficient based on external characteristics to obtain the adaptability coefficient. Second, a joint calculation of the doctor coefficient, patient coefficient, and seasonal coefficient yields the adaptability coefficient. Third, the adaptability coefficient is determined based on the deviation between predicted and actual values ​​of historical drug procurement data from the hospital. Instead of using fixed doctor, patient, and seasonal coefficients, adaptive adjustments are made, with the coefficients changing dynamically, reducing errors caused by fixed values. This allows for rapid response to external changes, avoiding the "rigidity" problem of fixed coefficients when scenarios change, and enhancing business flexibility. Real-time adaptive coefficients enable more accurate matching of drug procurement volumes, reducing resource waste or supply shortages, and optimizing resource allocation. Analysis results based on dynamic coefficients more accurately reflect business patterns, providing a more reliable basis for management decisions and improving the scientific nature of decision-making.

[0047] In some embodiments, the determination of the doctor coefficient, patient coefficient, and seasonal coefficient can be achieved in the following ways.

[0048] In the first implementation, the predetermined doctor coefficient, patient coefficient, and seasonal coefficient mentioned above can be obtained from historical data. For example, the doctor coefficient, patient coefficient, and seasonal coefficient calculated last year can be used as the predetermined doctor coefficient, patient coefficient, and seasonal coefficient for this year.

[0049] In the second implementation, the predetermined doctor coefficient, patient coefficient, and seasonal coefficient mentioned above can also be set randomly.

[0050] In the third implementation method, the aforementioned predetermined doctor coefficient, patient coefficient, and seasonal coefficient can also be calculated using a predetermined calculation formula:

[0051] Taking the doctor coefficient as an example, the doctor coefficient reflects the impact of doctor-related factors on drug procurement. The calculation steps for the doctor coefficient are as follows:

[0052] The physician coefficient can be simplified into the combined rate of change of three core physician factors: physician size, number of consultations, and medication habits, relative to the baseline period.

[0053] 1. Determine "baseline period physician coefficient = 1.0" (anchoring the conventional relationship between "physician-procurement").

[0054] Select "previous year" as the benchmark:

[0055] Doctor staff size: 50 doctors in total (permanent);

[0056] Patient volume: Doctors throughout the hospital saw a total of 30,000 patients during the quarter;

[0057] Medication habits: The total drug procurement amount corresponding to the number of outpatient visits is 3 million yuan (that is, 100 yuan of drug procurement per outpatient visit per person).

[0058] At this point, it is clear that the "overall status of doctors" for these 50 doctors, 30,000 patient visits, and 3 million in procurement amount corresponds to a coefficient of 1.0 (meaning that "under the baseline condition, the impact of doctors on procurement is 1 times the baseline amount").

[0059] 2. Calculate the current "changes in various factors related to the doctor":

[0060] Statistics for this year:

[0061] Number of doctors: 55 (5 more than the baseline, change rate = (55-50)÷50 = 10%)

[0062] Number of outpatient visits: 36,000 (6,000 more than the baseline, change rate = (36,000 - 3) ÷ 3 = 20%)

[0063] Medication habits: The total drug procurement amount for this quarter was 3.96 million yuan (110 yuan per visit per person, 10% more than the benchmark of 100 yuan, change rate = 10%).

[0064] Assign weights to the three factors (e.g., size accounts for 20%, treatment volume accounts for 50%, and medication habits account for 30%), and calculate the weighted change:

[0065] Scale contribution: 10% × 20% = 2%;

[0066] Contribution to patient volume: 20% × 50% = 10%;

[0067] Contribution of medication habits: 10% × 30% = 3%;

[0068] Total weighted rate of change = 2% + 10% + 3% = 15%.

[0069] 3. Calculate the current coefficient of all doctors in the hospital.

[0070] Current doctor coefficient = baseline value 1.0 × (1 + total weighted rate of change) = 1.0 × 1.15 = 1.15.

[0071] In some embodiments, the steps for determining the patient coefficient are as follows (for reference in drug procurement).

[0072] 1. Set “Baseline patient coefficient = 1.0” (anchoring baseline status).

[0073] Select the previous year's "quarter with stable operations" as the baseline period and collect key patient data (this data directly affects medication demand):

[0074] Total number of outpatient and inpatient patients in the baseline quarter: 10,000;

[0075] Patient composition: 30% are patients with chronic diseases (requiring long-term medication), 50% are patients with acute diseases (requiring short-term symptomatic medication), and 20% are patients with other conditions (low medication needs).

[0076] Average drug consumption per person: 150 yuan (i.e., total drug consumption in the baseline period: 10,000 × 150 = 1.5 million yuan).

[0077] The coefficient corresponding to the number of patients, their structure, and medication needs in this baseline period is set to 1.0. This serves as the "reference origin" for subsequent calculations.

[0078] 2. Compile statistics on the "change data" of current patients;

[0079] Taking the current quarter as an example, let's analyze the data on the same dimension:

[0080] Current total number of patients: 12,000 (2,000 more than the baseline period, change rate = (12,000 - 10,000) ÷ 10,000 = 20%).

[0081] Patient composition: 40% of patients have chronic diseases (an increase of 10 percentage points from the baseline period, with a change rate of (40%-30%) ÷ 30% ≈ 33.3%), 45% have acute diseases (a slight decrease), and 15% have other diseases (small change).

[0082] Average drug consumption per person: 165 yuan (15 yuan more than the baseline period, change rate = (165-150)÷150=10%).

[0083] 3. Calculate the overall rate of change based on the "influence weight".

[0084] The number, structure, and average medication usage of patients have different impacts on drug procurement, and weights need to be assigned (which can be adjusted according to the actual situation of the hospital, for example: number accounts for 40%, structure accounts for 30%, and average medication usage per patient accounts for 30%):

[0085] Change in the contribution of patient numbers: 20% × 40% = 8%;

[0086] Changes in patient demographics (increased proportion of chronic diseases, higher demand for medication, calculated positively): 33.3% × 30% ≈ 10%.

[0087] Changes in per capita medication consumption: 10% × 30% = 3%;

[0088] The overall change rate = 8% + 10% + 3% = 21%.

[0089] 4. Derive the current "patient coefficient".

[0090] Current patient coefficient = baseline value 1.0 × (1 + total comprehensive change rate) = 1.0 × (1 + 21%) = 1.21.

[0091] The practical significance of the patient coefficient: For example, if the baseline period requires the purchase of 1.5 million yuan worth of drugs based on patient needs, and the current patient coefficient is 1.21, it means that "patient-related factors (more patients, a higher proportion of chronic disease patients, and a slight increase in average drug use per person) have increased drug demand by 21%". Therefore, the "patient-related reference amount" for drug procurement in the current quarter can be initially planned at 1.5 million × 1.21 = 1.815 million yuan, which is more in line with the actual drug needs of patients.

[0092] The patient coefficient is the "influence multiple" of the current patient status relative to the baseline status. The core is to find the "patient dimensions that affect medication / resources" (number, structure, condition, etc.) and then quantify the changes.

[0093] In the context of drug purchasing, the core of calculating the seasonal coefficient is to "quantify the fluctuation range of drug demand in different seasons relative to a baseline state." It is typically determined by "historical data from the same period" and using the "ratio of current seasonal demand to baseline seasonal demand." The specific steps can be found in the following example:

[0094] Example: Calculation of "Seasonal Coefficient of Cold and Flu Medications" at a hospital.

[0095] 1. Determine the “baseline season” and “baseline demand value”.

[0096] Select a season with stable demand and no significant seasonal fluctuations as the benchmark (such as autumn, when the incidence of colds is relatively balanced), and statistically analyze the drug procurement demand in the benchmark season:

[0097] Assuming the average monthly purchase volume of cold medicines in the past three autumns (a total of three autumns) is 1000 boxes, 1100 boxes, and 900 boxes respectively, the average value is taken as the baseline demand:

[0098] The baseline monthly average purchase volume = (1000 + 1100 + 900) ÷ 3 = 1000 boxes.

[0099] At this point, the seasonal coefficient for the baseline season (autumn) is set to 1.0.

[0100] 2. Compile statistics on "actual demand in the target season".

[0101] Taking "winter" as an example (when colds are prevalent and demand increases), we also compiled the average monthly purchase volume for the past three winters:

[0102] Assuming the average monthly purchase volume for the past three winters was 2200 boxes, 2400 boxes, and 1900 boxes respectively, take the average:

[0103] Average monthly purchase volume in winter = (2200 + 2400 + 1900) ÷ 3 = 2167 boxes (rounded to the nearest integer).

[0104] 3. Calculate the seasonal coefficient for the target season.

[0105] Seasonal coefficient = average demand in the target season ÷ average demand in the baseline season.

[0106] Winter seasonal coefficient = 2167 ÷ 1000 ≈ 2.17.

[0107] 4. The same calculation applies to other seasons.

[0108] For example, during the summer (when the incidence of colds is low), the average monthly purchase volume over the past three years has been 500 boxes.

[0109] Summer seasonal coefficient = 500 ÷ 1000 = 0.5.

[0110] The calculation of the seasonal coefficient is essentially "taking the base season as '1' and dividing the actual demand of the target season (usually taking the historical average for the same period to reduce random fluctuations) by the base value". A result greater than 1 indicates that the demand in that season is higher than the base, and a result less than 1 indicates that the demand is lower than the base.

[0111] In some embodiments, the joint calculation of the doctor coefficient, the patient coefficient, and the seasonal coefficient includes:

[0112] The interaction coefficient weights are determined based on any combination of the doctor coefficient, the patient coefficient, and the seasonal coefficient.

[0113] The interaction coefficient weights are used to measure the degree of influence of the interaction coefficient on the interaction coefficient when the three factors of doctors, patients and seasons interact.

[0114] The interaction coefficient is used to quantify the combined impact of the interaction between doctors, patients, and seasonal factors on drug procurement demand.

[0115] In some embodiments, any combination of the doctor coefficient, the patient coefficient, and the seasonal coefficient can be: there can be only the doctor coefficient and the patient coefficient.

[0116] In some embodiments, only doctor coefficients and seasonal coefficients may be used. In some embodiments, only patient coefficients and seasonal coefficients may be used. In some embodiments, doctor coefficients, patient coefficients, and seasonal coefficients may be included.

[0117] The interaction coefficient is determined based on the interaction coefficient weights and any combination of the doctor coefficient, the patient coefficient, and the seasonal coefficient.

[0118] A joint coefficient is determined based on any combination of the doctor coefficient, patient coefficient, and seasonal coefficient, and the interaction coefficient, and the joint coefficient serves as the adaptability coefficient for adjusting drug procurement.

[0119] In this embodiment, if three coefficients are used, and the three factors are strongly correlated, such as seasonal high incidence + patient urgency + doctor's prescription tendency, the impact of the three factors combined is doubled, not simply added together. The interaction coefficient weight reflects the "amplification effect". The three factors are strongly correlated, and the weight can be set to 1.3. A value higher than 1 means that the impact after linkage is stronger than the sum of the individual factors. If there are only two factors (such as only doctor and patient, without the strong background of season), the weight can be set to 1.1.

[0120] The interaction coefficient is determined based on the interaction coefficient weights mentioned above. Specifically, in one embodiment, the interaction coefficient is equal to the product of the mean of the doctor coefficient, the patient coefficient, and the seasonal coefficient, and the interaction coefficient weights mentioned above.

[0121] For example, the doctor coefficient is 1.3, the patient coefficient is 1.2, and the seasonal coefficient is 1.4; the mean is (1.3 + 1.2 + 1.4) ÷ 3 = 1.3, and the mean multiplied by the interaction coefficient weight of 1.3 gives the interaction coefficient = 1.69.

[0122] In some embodiments, the joint coefficient can be set to be equal to the product of the doctor coefficient, patient coefficient, seasonal coefficient, and interaction coefficient.

[0123] Based on the example, 1.3×1.2×1.4×1.69≈3.69, therefore, the joint coefficient is approximately 3.69.

[0124] In some embodiments, taking "purchasing rehydration salts for children with diarrhea" as an example, only the patient coefficient and seasonal coefficient are considered, without taking into account the doctor coefficient, and the calculation is performed according to the logic of "interaction coefficient weight → interaction coefficient → joint coefficient":

[0125] Basic settings (two coefficients):

[0126] Patient coefficient (P): Among recent children with diarrhea, the proportion of expedited patients who needed to get their medication on the same day for acute diarrhea was 50% (higher than usual), and the adjusted P=1.4 (the normal value is 1.0).

[0127] Seasonal coefficient (S): It is currently summer (high temperatures make it easy for bacteria to grow, and it is the peak season for childhood diarrhea), so the seasonal coefficient S = 1.5 (the normal value is 1.0).

[0128] Step 1: Determine the "interaction coefficient weights".

[0129] Because it is a combination of patient and season factors, and summer is a season when the number of diarrhea patients tends to increase, the season and patient needs are directly related and have a strong correlation, so the weight should be greater than 1.

[0130] The interaction coefficient weight is set to 1.2, which is higher than the weight of a single factor but slightly lower than the weight of a strong three-factor association, which is reasonable.

[0131] If it's just a "weak correlation," such as when the season and patient needs are unrelated, then the weight should be set to 1.0.

[0132] Step 2: Calculate the "interaction coefficient". Interaction coefficient = (patient coefficient + seasonal coefficient) ÷ 2 × interaction coefficient weight.

[0133] Substituting the data: the actual mean of the two coefficients = (1.4 + 1.5) ÷ 2 = 1.45; multiplied by the weight 1.2, the interaction coefficient = 1.74.

[0134] Step 3: Calculate the "joint coefficient".

[0135] Joint coefficient = (patient coefficient × seasonal coefficient) × interaction coefficient.

[0136] Substituting the data: Patients × Seasonal coefficient = 1.4 × 1.5 = 2.1; 2.1 × 1.74 = 3.654, Combined coefficient ≈ 3.65.

[0137] Practical significance: If the "standard basic purchase quantity of rehydration salts is 200 boxes," then with a combination coefficient of 3.65, the adjusted purchase quantity would be approximately 200 × 3.65 = 730 boxes. This aligns with reality, as there are already many children with diarrhea in the summer, and half of these patients require urgent medication. The combination of these two factors "drives up demand," naturally resulting in a higher combination coefficient than when only one factor is present, thus increasing the purchase quantity.

[0138] In some embodiments, determining the drug procurement adjustment adaptability coefficient based on the deviation between the predicted and actual values ​​of historical drug procurement data from the hospital specifically includes:

[0139] The hospital's historical drug procurement data is grouped according to preset dimensions to obtain multiple groups of historical drug procurement data.

[0140] In this embodiment, the hospital's historical drug procurement data is grouped according to different dimensions. Dimensions can be, for example, time-based dimensions, such as a period, where the period can be months. The hospital's historical drug procurement data is divided into multiple groups based on months. Dimensions can also be doctor-related factors, such as the doctor's professional title; based on different professional titles, the historical drug procurement data is divided into multiple groups.

[0141] Determine the predicted and actual values ​​of historical drug procurement data for each group of hospitals.

[0142] The correction coefficient for each group is determined based on the predicted and actual values ​​of the historical drug procurement data of each hospital.

[0143] The appropriateness coefficient for adjusting the procurement of the aforementioned drugs is determined based on the correction coefficients for each group mentioned above.

[0144] In this embodiment, based on a time period, taking a cough syrup from the respiratory department as an example, the drug, patient (adult), and prescription scenario (outpatient) are fixed. The prescriptions are divided into groups based on the "last 5 months (January-May)" period, with each group corresponding to one purchase in that month.

[0145] Step 1: Divide the data into groups.

[0146] Group 1: Purchases in January (1 time);

[0147] Group 2: Purchases in February (1 time);

[0148] Group 3: Purchases in March (1 time);

[0149] Group 4: Purchases in April (1 time);

[0150] Group 5: Procurement in May (1 time).

[0151] Step 2: Calculate the correction factor for each group.

[0152] The correction factor for each group is obtained by dividing the actual purchase volume of the month by the predicted purchase volume of the month.

[0153] See Table 1:

[0154] Table 1 Monthly Procurement Statistics

[0155]

[0156] In one embodiment, the drug procurement adjustment adaptability coefficient is determined based on each set of correction coefficients. The average value of each set of correction coefficients can be calculated and used as the drug procurement adjustment adaptability coefficient.

[0157] In one embodiment, the proportion of each data set to the total data can be calculated, and the product of the proportion of each data set and the correction coefficient can be calculated. The sum of the products of the proportion of each data set and the correction coefficient is then obtained to obtain the drug procurement adjustment adaptability coefficient.

[0158] For example, in the above embodiment, the purchase volume of each group is calculated according to its weight in the total purchase volume (total purchase volume = 180 + 168 + 130 + 96 + 88 = 662 bottles):

[0159] Group 1 weight: 180 ÷ 662 ≈ 27%;

[0160] Group 2 weight: 168 ÷ 662 ≈ 25%;

[0161] Group 3 weight: 130 ÷ 662 ≈ 20%;

[0162] Group 4 weight: 96 ÷ 662 ≈ 14%;

[0163] Group 5 weight: 88 ÷ 662 ≈ 14%;

[0164] The drug procurement adjustment adaptability coefficient = (1.2×27%) + (1.2×25%) + (1.0×20%) + (0.8×14%) + (0.8×14%) ≈ 1.04.

[0165] January and February are flu season, and the actual usage was 20% higher than predicted (correction factor 1.2); the usage was lower in April and May, 20% lower than predicted (correction factor 0.8), which corresponds to the actual medication usage pattern.

[0166] In some embodiments, taking a certain antihypertensive drug as an example, with fixed drugs, time (within the past month), and patient types, patients are divided into 3 groups based solely on their doctor's professional title, with each group corresponding to one procurement:

[0167] Step 1: Divide the data into groups.

[0168] Group 1: Procurement related to prescriptions issued by chief physicians;

[0169] Group 2: Procurement related to prescriptions issued by attending physicians;

[0170] Group 3: Procurement of prescriptions associated with resident physicians.

[0171] Step 2: Calculate the correction factor for each group.

[0172] For each group, the correction factor is calculated by dividing the actual purchase quantity by the predicted purchase quantity.

[0173] See Table 2:

[0174] Table 2. Procurement Statistics for Each Group of Doctors

[0175]

[0176] Step 3: Calculate the "overall adjustment coefficient".

[0177] Calculate based on the weight of each group's purchase volume relative to the total purchase volume (assuming total purchase volume = 102 + 165 + 72 = 339 boxes):

[0178] Group 1 weight: 102 ÷ 339 ≈ 30%;

[0179] Group 2 weight: 165 ÷ 339 ≈ 49%;

[0180] Group 3 weight: 72 ÷ 339 ≈ 21%;

[0181] The adaptability coefficient for drug procurement adjustments = (1.02 × 30%) + (1.1 × 49%) + (0.9 × 21%) ≈ 1.03.

[0182] In some embodiments, the refinement adjustment of the predetermined seasonal coefficients based on external characteristics includes:

[0183] Based on the aforementioned seasonal coefficients, the adjusted seasonal coefficients are determined by adjusting the weights according to time and seasonal coefficients.

[0184] The seasonal coefficient adjustment weights are determined based on holidays or peak flu seasons.

[0185] In this embodiment, taking "traditional Chinese medicine for respiratory diseases in hospitals" as an example, the basic coefficients for the four seasons are first determined, and then the seasonal coefficients are adjusted based on whether "the whole season includes holidays / flu peaks", and finally the "adjusted complete seasonal coefficients" are calculated.

[0186] Step 1: Determine the seasonal coefficient as the basic coefficient. The following steps can be used.

[0187] Based on routine data from the past three years (excluding special holidays and large-scale influenza outbreaks), calculate the basic coefficient for each season (based on the average annual usage, coefficient = actual seasonal usage ÷ average annual usage):

[0188] Spring (March-May): Basic coefficient = 1.1 (Spring pollen season, respiratory discomfort is slightly more common);

[0189] Summer (June-August): Basic coefficient = 0.8 (lowest dosage, fewer respiratory diseases);

[0190] Autumn (September-November): Basic coefficient = 1.3 (due to cross-infection after the start of school, the dosage increases).

[0191] Winter (December-February): Basic coefficient = 1.6 (highest usage due to low temperature and enclosed indoor environment).

[0192] Step 2: Determine the seasonal coefficient adjustment weights.

[0193] A weight is assigned to "whether the entire season contains key external features" (the weight is a correction ratio to the "entire season coefficient"):

[0194] If the season includes "major holidays" (such as the Spring Festival in winter and the National Day in autumn, and the duration of the holidays accounts for more than 1 / 3 of the season): weight = +0.15 (due to the movement / gathering of people, the usage is 15% more than in the regular season).

[0195] If the season falls within the "influenza peak season as defined by the CDC" (e.g., winter and autumn are both designated as peak seasons): weight = +0.2 (dosage is 20% more than in the regular season).

[0196] If both include: weight superposition (but not exceeding 0.3, to avoid over-adjustment).

[0197] Step 3: Refine and adjust the "final seasonal coefficient" according to the season.

[0198] 1. Winter (December-February);

[0199] Basic characteristics: It is itself a "high season for influenza" (meets the weighting of +0.2), and it includes the Spring Festival (lasting about 20 days, accounting for 1 / 4 of the winter season, meets the "major holiday" weighting of +0.15).

[0200] Adjusting the weight: Since the two are combined, the upper limit of 0.3 is taken (to avoid excessively high coefficients leading to inventory backlog).

[0201] The adjusted seasonal coefficient = basic coefficient × (1 + adjustment weight) = 1.6 × (1 + 0.3) = 2.08.

[0202] 2. Autumn (September-November);

[0203] Basic characteristics: It is the "flu season" (weight +0.2) and includes the National Day holiday (7 days, accounting for about 1 / 10 of the autumn season, less than 1 / 3, and does not meet the weight condition of "major holidays").

[0204] Adjust the weights: only use the peak influenza weight + 0.2;

[0205] The adjusted seasonal coefficient is 1.3 × (1 + 0.2) = 1.56.

[0206] 3. Spring (March-May);

[0207] Basic characteristics: Not during peak flu season, only includes Qingming / May Day (single holiday ≤ 3 days, low proportion), no external characteristics meeting the criteria.

[0208] Adjust weight: 0 (no additional adjustment required);

[0209] The adjusted seasonal coefficient = the basic coefficient = 1.1.

[0210] 4. Summer (June-August);

[0211] Basic characteristics: Not during peak flu season, only includes the Dragon Boat Festival (1 day), no matching external characteristics.

[0212] Adjusted weight: 0;

[0213] The adjusted seasonal coefficient = the basic coefficient = 0.8.

[0214] In summary, after the above adjustments, each season has a "final coefficient": Winter 2.08, Autumn 1.56, Spring 1.1, and Summer 0.8. For example, when purchasing in winter, stocking up directly based on "annual average usage × 2.08" takes into account both the basic seasonal patterns and the impact of "flu peak + Spring Festival," making it more in line with actual needs than simply using the basic coefficient of 1.6.

[0215] In some embodiments, the refinement adjustment of the predetermined patient coefficients based on external characteristics includes:

[0216] The external characteristic is the level of urgency. When the level of urgency is normal, the patient coefficient is the baseline value. The baseline value is the value set corresponding to the normal level of urgency.

[0217] In some embodiments, the baseline value of the patient coefficient can be directly used as the patient coefficient.

[0218] In some embodiments, the patient coefficient baseline value can be multiplied by a predetermined weight to obtain the patient coefficient.

[0219] In cases where the urgency level is expedited, the patient coefficient is determined based on the aforementioned patient coefficient baseline value and the first patient coefficient weight value.

[0220] The patient coefficient is obtained by multiplying the above-mentioned patient coefficient baseline value with the above-mentioned first weight value.

[0221] In cases where the urgency level is sudden, the patient coefficient is determined based on the aforementioned patient coefficient baseline value and the second patient coefficient weight value.

[0222] The patient coefficient is obtained by multiplying the baseline value of the above-mentioned patient coefficient with the weight value of the above-mentioned second patient coefficient.

[0223] In this embodiment, the patient coefficient is adjusted for different levels of urgency (assuming a baseline patient coefficient of 1.0, the adjustment factor is only an example).

[0224] Urgency level "normal" (non-urgent, stable demand): The coefficient is not adjusted, that is, the adjusted coefficient = the predetermined value of the first patient's coefficient × 1.0.

[0225] Example: A community hospital regularly procures standard antihypertensive drugs for patients with stable conditions and fixed medication needs, with no urgent need for replenishment. In this case, the procurement quantity can be calculated based on the baseline patient coefficient, without the need to increase the quantity or prioritize the medication based on "urgency."

[0226] Urgency level "Extreme" (urgent need, requiring a shorter procurement cycle): The coefficient is usually adjusted upwards, for example, the adjusted coefficient = the predetermined value of the second patient's coefficient × 1.5. Here, 1.5 is the weighting factor.

[0227] Example: A department receives a batch of postoperative patients who need a replenishment of a specific antibiotic within 36 hours (the original stock is only enough for 12 hours), which is considered an "urgent" need. In this case, by adjusting the coefficient (for example, from the basic 1.0 to 1.5), the model will appropriately increase the procurement quantity when predicting the procurement volume (to avoid stockouts) and at the same time improve the procurement priority of this drug.

[0228] The urgency level is "sudden" (urgent and potentially extraordinary demand, such as a public health emergency): the coefficient is significantly increased, for example, the adjusted coefficient = the predetermined value of the third patient coefficient × 2.0. Here, 2 is the second weight.

[0229] Example: A sudden influenza outbreak occurs in a certain area, leading to a surge in patients at fever clinics. The demand for antipyretics and antiviral drugs is "sudden" and far exceeds the usual level. At this time, the patient coefficient is adjusted to 2.5 times the base value. Based on this, the model will significantly increase the procurement forecast (for example, from the usual 500 boxes to 1200 boxes), and trigger the "emergency procurement channel" to prioritize the supply of these drugs and avoid inventory shortages.

[0230] In some embodiments, the refinement adjustment of the predetermined physician coefficients based on external characteristics includes:

[0231] The external feature is that it is segmented into weekdays and holidays;

[0232] On weekdays, the doctor's coefficient is the base value;

[0233] The aforementioned baseline values ​​are the benchmark values ​​for doctor coefficients set in a weekday scenario. They provide a unified and standardized reference scale for the calculation of doctor coefficients and serve as the initial basis for calculating doctor coefficients (by multiplying the baseline value by the fluctuation coefficient) in special scenarios such as holidays.

[0234] During holidays, the doctor coefficient is the product of the fluctuation coefficient and the base value.

[0235] In this embodiment, taking the prescription coefficient of a cardiologist for a certain lipid-lowering drug as an example, the doctor coefficient is essentially "the ratio of the doctor's prescription volume to the department's average prescription volume," used to predict medication needs. Adjustments are made in conjunction with the external feature of "weekdays / holidays," as follows:

[0236] Step 1: Determine the baseline value for the doctor's coefficient.

[0237] First, calculate the baseline values ​​for three cardiologists (based on prescription data from the past three working days, excluding holiday interference; baseline value = doctor's average daily prescription volume ÷ department's average daily prescription volume):

[0238] Dr. Zhang (Chief Physician, stable prescriptions on weekdays): Baseline value = 1.2 (daily prescription volume is 20% higher than the department average).

[0239] Dr. Li (Attending Physician, Weekday Prescription Routine): Baseline value = 1.0 (Daily average prescription volume is equal to the department average);

[0240] Dr. Wang (resident physician, weekday prescriptions tend to be conservative): Baseline value = 0.8 (daily prescription volume is 20% lower than the department average).

[0241] Step 2: Determine the "holiday fluctuation coefficient".

[0242] Based on historical data, the number of cardiology outpatient visits and prescription habits fluctuate regularly during holidays (such as weekends, National Day, and Spring Festival). A "fluctuation coefficient" (reflecting the proportion of prescriptions on holidays relative to weekdays) is established:

[0243] On regular holidays (such as weekends, outpatient visits are mainly for follow-up patients): fluctuation coefficient = 0.7 (the number of prescriptions is 70% of that on weekdays, due to fewer initial visits and fewer prescription adjustments).

[0244] During long holidays (such as the 7-day National Day and the 7-day Spring Festival, some patients may seek medical treatment earlier or later): fluctuation coefficient = 0.4 (the number of prescriptions is 40% of that on weekdays, because outpatient clinics are open for fewer days and patients prepare their medications in advance).

[0245] Step 3: Refine and adjust the "doctor coefficient" according to "weekdays / holidays".

[0246] 1. Weekday scenario (basic situation).

[0247] Use the base value directly, no adjustment needed:

[0248] Dr. Zhang: Doctor coefficient = 1.2 (Based on this prediction, his medication needs on weekdays are 20% higher than the department average).

[0249] Dr. Li: Doctor coefficient = 1.0 (consistent with the department's average demand).

[0250] Dr. Wang: Doctor coefficient = 0.8 (demand is 20% lower than the department average).

[0251] 2. Regular public holidays (such as Saturdays and Sundays);

[0252] Adjusted by "volatility coefficient × base value":

[0253] Dr. Zhang: The adjusted doctor coefficient = 0.7 × 1.2 = 0.84 (Although he usually prescribes many prescriptions, his outpatient volume is low on holidays, and the demand drops to 70% of the weekday volume, corresponding to a coefficient of 0.84).

[0254] Dr. Li: The adjusted doctor coefficient = 0.7 × 1.0 = 0.7 (the demand is 70% of the weekdays).

[0255] Dr. Wang: The adjusted doctor coefficient = 0.7 × 0.8 = 0.56 (demand further reduced).

[0256] 3. Long holidays (such as the 7-day National Day holiday);

[0257] Adjusted by "volatility coefficient × base value":

[0258] Dr. Zhang: The adjusted doctor coefficient = 0.4 × 1.2 = 0.48 (fewer outpatient visits during long holidays, his demand is only 40% of that on weekdays).

[0259] Dr. Li: The adjusted doctor coefficient = 0.4 × 1.0 = 0.4 (the demand is 40% of the weekdays).

[0260] Dr. Wang: The adjusted doctor coefficient = 0.4 × 0.8 = 0.32 (lowest demand).

[0261] In practical applications, such as a hospital pharmacy needing to predict "Dr. Zhang's dosage of lipid-lowering drugs during the National Day holiday":

[0262] If the department needs to prepare 100 boxes per weekday, Dr. Zhang's weekday demand = 100 × 1.2 = 120 boxes;

[0263] During the National Day holiday, based on the adjusted doctor coefficient of 0.48, his demand is 100 × 0.48 = 48 boxes. Preparing stock based on this number will avoid overestimating (preparing 120 boxes based on the base value of 1.2 would lead to stockpiling) and underestimating (preparing 40 boxes based on the uniform fluctuation coefficient of 0.4 might not be enough for his patients), thus more accurately matching the actual prescription needs.

[0264] In some embodiments, the refinement adjustment of the predetermined physician coefficients based on external characteristics includes:

[0265] External characteristics include: drug classification and physician scheduling.

[0266] The doctor coefficient is further refined and adjusted based on drug classification and doctor scheduling.

[0267] In this embodiment, the drug procurement plan was not linked to the doctor (expert) scheduling system, resulting in a shortage or backlog of expert medications (especially high-value drugs). 1. Classify the drugs.

[0268] (1) Category A (Schedule-Sensitive): High-value drugs (such as targeted cancer drugs) that are strongly correlated with expert scheduling. Monitor inventory daily and set dynamic safety stock (e.g., replenish stock to 120% of usage 3 days before the scheduled shift).

[0269] (2) Category B (Department-wide): Drugs shared by multiple departments (such as antibiotics), replenished weekly.

[0270] (3) Category C (low turnover essential drugs): regular reserves, replenished monthly.

[0271] In this embodiment, the doctor coefficient can be dynamically adjusted by combining expert scheduling, and at the same time, drug classification and inventory strategies can be linked to make the doctor coefficient more in line with the actual business scenario, such as the fluctuation of drug demand caused by expert scheduling, thereby optimizing inventory efficiency.

[0272] Step 1: Identify the core correlation dimensions of the "doctor coefficient".

[0273] First, define the "doctor coefficient" as "the strength of the correlation between a certain type of drug and a specific specialist / department's drug use." A higher coefficient indicates that the use of that drug is more significantly affected by the specialist's scheduling. Combining this with the ABC classification above, the focus should be on dynamically adjusting category A (scheduling-sensitive), while categories B and C can be simplified.

[0274] Category A (e.g., targeted cancer drugs): The coefficient needs to be directly linked to the corresponding expert's schedule, because the expert's consultation time directly determines the short-term demand for this type of drug (for example, if the expert is seeing patients for 3 days this week, the usage of this type of drug may double compared to last week).

[0275] Category B (Department-wide): The coefficient can be linked to the "overall outpatient volume of the department" (rather than a single expert). Because it is shared by multiple departments, the demand fluctuations are more stable, and replenishment can be done weekly.

[0276] Category C (Low-Cycle Transition): The coefficients are basically stable (demand is infrequent and fixed), requiring no frequent adjustments; minor adjustments can be made when replenishing inventory monthly.

[0277] Step 2: Dynamic adjustment logic for physician coefficients of Class A drugs.

[0278] Category A is scheduling-sensitive; the key is to ensure that the coefficients "respond in advance" according to expert scheduling, thereby linking to dynamic safety stock. This can be broken down into three steps:

[0279] 1. First, associate "relevant experts" and basic coefficients with Class A drugs.

[0280] First, identify the core expert corresponding to each Class A drug (e.g., "Osimertinib" is associated with Director Li of the Oncology Department), and set a base coefficient (e.g., 1.0, representing "the strength of routine association when there is no scheduling").

[0281] The coefficient can be initially determined by combining historical data: for example, in the past 6 months, the amount of osimertinib used on the days when Director Li was on duty was 2.5 times that on the days when he was not on duty, then the base coefficient can be set to 2.5 (meaning "when the expert is on duty, the correlation between the drug and him is 2.5 times that of the norm").

[0282] 2. Adjust the coefficient according to the scheduling plan.

[0283] Three days before the scheduled appointment: When the system confirms the expert's appointment (e.g., Director Li will be seeing patients next Wednesday), the coefficient of the corresponding medicine will be increased to the "peak coefficient" (e.g., the base coefficient 2.5 → increased to 3.0) from that day onwards. At the same time, a replenishment instruction ("replenish to 100% quantity") will be triggered, because sufficient inventory needs to be prepared in advance to meet the concentrated demand on the day of the appointment.

[0284] On the day of scheduling and the following day: maintain the coefficient at its peak (3.0), monitor the actual usage, and if the usage exceeds expectations, temporarily increase it by 0.2-0.3 (for example, if the actual usage is 10% more than the estimate, adjust the coefficient to 3.2) to avoid stockouts.

[0285] After the shift schedule is completed: the coefficient drops back to the base coefficient (2.5). If there is remaining inventory, the safety stock can be adjusted back to the normal value (e.g., from "100% quantity" to "70% quantity") to reduce backlog.

[0286] 3. Combine with the expert consultation frequency correction coefficient (long-term dimension).

[0287] If a specialist's clinic frequency increases this month compared to last month (e.g., from 4 times to 6 times), the base coefficient for the corresponding medicine can be adjusted upwards (e.g., from 2.5 to 2.8), because the stronger association between the medicine and the specialist increases with more clinic visits over a long period. If a specialist reduces clinic visits due to vacation (e.g., only 2 times this month), the base coefficient should be adjusted downwards (e.g., from 2.5 to 2.0) to avoid overstocking.

[0288] Step 3: Coefficient adjustment for Category B and Category C drugs (simplified version).

[0289] Category B (Department-wide): The doctor coefficient is linked to the "total number of outpatient visits per week in the department". For example, if there are 10 doctors making outpatient visits in the internal medicine department this week (8 last week), the doctor coefficient for Category B drugs (such as conventional antihypertensive drugs) will be increased from the base value of 1.0 to 1.2 (corrected according to the "increase in the number of outpatient visits per week"). The purchase quantity will then be adjusted in conjunction with the weekly replenishment plan (for example, if 50 boxes were replenished last week, 60 boxes will be replenished this week).

[0290] Category C (Low-cycle transition): The doctor coefficient is basically fixed and is only adjusted during quarterly inventory checks. For example, if the usage of a certain Category C drug (such as a drug for rare diseases) has decreased by 20% in the past 3 months, the coefficient will be lowered from 1.0 to 0.8. When replenishing stock monthly, the purchase volume will be reduced (for example, from 10 boxes per month to 8 boxes per month) to avoid inventory backlog.

[0291] Step 4: Implement the "Doctor Coefficient - Inventory" linkage rule.

[0292] Finally, the corrected coefficients must be linked to inventory operations to form a closed loop:

[0293] When the coefficient for Class A drugs is ≥3.0 (3 days before the shift): "Emergency replenishment to 100% inventory" will be automatically triggered;

[0294] When the coefficient for Category B drugs fluctuates by ±0.2 (weekly outpatient volume change): the replenishment quantity for this week will be automatically adjusted (increased or decreased proportionally to the coefficient fluctuation).

[0295] When the coefficient for Class C drugs is ≤0.8 (long-term demand decline): Automatic reminder to "reduce monthly replenishment quantity" and extend the next replenishment cycle (e.g., from once a month → once every 6 weeks).

[0296] In this way, the doctor coefficient is not a fixed value, but can change dynamically with the scheduling and frequency of visits. At the same time, it adapts to the characteristics of the three categories of drugs: A, B, and C. Category A accurately responds to the expert scheduling, Category B meets the routine needs of the department, and Category C reduces ineffective adjustments. This ensures both inventory efficiency and conforms to the logic of related business scenarios.

[0297] In some embodiments, the method further includes: optimizing the procurement cycle in conjunction with the work schedule.

[0298] (1) Procurement on demand: For departments with intensive expert schedules (such as expert outpatient weeks), shorten the procurement cycle to 3-5 days to reduce inventory backlog. For example, during weeks when ophthalmologists have concentrated outpatient services, increase the frequency of delivery of artificial tears and anti-glaucoma drugs.

[0299] (2) Procurement in batches: For experts who are taking long-term leave of absence (such as academic leave or further study), the regular procurement of their commonly used high-value drugs (such as monoclonal antibody drugs) is suspended and replaced with temporary allocation. In some embodiments, the method further includes: dynamically adjusting the drug catalog.

[0300] (1) Establish a “schedule-related drug catalog” to distinguish between regular drugs and drugs that depend on the schedule. For example, during the period when a neurology specialist is not seeing patients, his special drugs (such as new antiepileptic drugs) can be temporarily removed from the procurement catalog to reduce the capital occupation.

[0301] (2) With authorization from the hospital's pharmacy committee, the pharmacy department is permitted to adjust the procurement ratio of non-essential drugs according to the scheduling plan. For example, on expert surgery days, additional high-priced intraoperative contrast agents may be procured temporarily. In some embodiments, the method further includes: integrating information systems.

[0302] Specifically, data interfaces between scheduling systems (such as HRP), drug closed-loop management systems, and procurement platforms are integrated to enable real-time synchronization of scheduling changes to the procurement module. For example, hospitals have developed a "scheduling-procurement linkage dashboard" that visually displays a heatmap of drug demand corresponding to the scheduling for the next 7 days.

[0303] In some embodiments, the method further includes: modeling the association between scheduling data and medication demand.

[0304] (1) Analysis of expert scheduling and departmental medication characteristics.

[0305] Historical data mining: By extracting expert scheduling records and corresponding departmental drug prescription data from the Hospital Information System (HIS), a "scheduling-medication" correlation model is established. For example, on a cardiovascular specialist's consultation day, the consumption of anticoagulants (such as warfarin) and antihypertensive drugs (such as amlodipine) may increase by 20%-30%.

[0306] Medication usage habit labeling: Categorizing experts' prescribing habits (such as preference for innovative / generic drugs, combination therapy preferences) to create personalized procurement references. For example, if an oncology expert prefers a certain brand of targeted drug, the procurement cycle for that drug can be adjusted accordingly.

[0307] In some embodiments, dynamic demand forecasting of medicines can also be performed based on the scheduling situation.

[0308] (1) Predict peak drug consumption based on scheduling cycles (such as outpatient days, surgery days, and academic activity periods). For example, on days when surgical experts have a concentrated number of surgeries, sufficient anesthetic drugs (such as propofol) and hemostatic materials (such as gelatin sponges) need to be prepared in advance.

[0309] (2) Utilize machine learning algorithms (such as random forests and time series models) to predict the impact of scheduling changes on drug demand. For example, a tertiary hospital predicted through a model that drug demand would decrease by 15% during expert consultations outside the hospital, and reduced the corresponding procurement volume. In some embodiments, the method further includes: using an intelligent early warning system for early warning.

[0310] An inventory management system integrated with a scheduling calendar automatically triggers replenishment reminders. For example, after specialist outpatient schedules are entered, the system automatically calculates the required amount of medication and generates an order.

[0311] In some embodiments, see Appendix Figure 2 The method further includes: retrieving drug procurement data from the database in response to a query command sent by the front-end webpage.

[0312] The drug procurement data is sent to the front-end page, where the following parameters of the drugs are displayed in a list format:

[0313] The information includes the drug name, manufacturer, statistical period, drug code, specifications, negotiation data volume, negotiation duration, monthly task status, yearly task status, and overall task progress.

[0314] When making a search, users enter search keywords on the front-end page. Keywords can include, but are not limited to, the following: drug name, manufacturer, code, and statistical time. These can be combined in a search.

[0315] In some embodiments, the method further includes: receiving newly added drug data input from the front-end page in response to a new command sent by the front-end webpage.

[0316] In this embodiment, the newly added drug data includes: drug name, manufacturer, drug code, and specifications.

[0317] In some embodiments, see Appendix Figure 3 The method further includes: in response to a save command sent by the front-end timeliness setting webpage, receiving the real-time time entered in the front-end timeliness setting webpage; the front-end timeliness setting webpage also includes selectors and radio buttons.

[0318] Among these, "timeliness" refers to the time standards used in the procurement process, aiming to ensure that drug supply is timely, efficient, and meets medical needs. It mainly involves the following aspects:

[0319] Setting an order response timeline specifies the latest deadline for delivery companies to confirm orders. For example, Shandong Province stipulates that delivery companies must confirm orders within two working days after a medical institution places an order, while Sichuan Province requires delivery companies to provide a response of delivery or refusal within 48 hours after a medical institution initiates a drug order.

[0320] Regulations specify delivery time limits, clearly defining the time requirements from order confirmation to delivery to the designated location. For example, the "Hunan Province Centralized Procurement of Medicines and Medical Consumables Distribution Service Management Measures" stipulates that emergency and rescue medicines must be delivered within 12 hours of order confirmation, and general medicines should be delivered within 48 hours.

[0321] The system monitors and evaluates on-time performance. It monitors the actual execution status according to the set on-time time and generates data such as order response time rate and delivery time rate, which can be used as a basis for evaluating delivery companies and optimizing procurement processes.

[0322] The system triggers an early warning mechanism. If the corresponding steps are not completed within the set time frame, the system will automatically trigger an early warning and notify relevant personnel to handle the situation in order to avoid delays and ensure timely supply of medicines.

[0323] In this embodiment, selectors and radio buttons are used to allow users to flexibly configure rules or filtering conditions for calculating timeliness.

[0324] Selectors are used to select one or more options from multiple options. They are adapted to clearly categorized configuration items. Common selectors include dropdown selectors, date selectors, and tree selectors. In timeliness settings, their uses may include:

[0325] Select the time range: such as "monthly on-time rate", "quarterly on-time rate", "custom time", use the drop-down selector; if you want to be precise to a specific date, use the date selector (for example, select "2024-08-01 to 2024-08-31" to calculate the on-time rate for the current month).

[0326] Select the statistical objects: for example, select from dimensions such as "department", "doctor" and "equipment" (for example, select "internal medicine" and calculate the task timeliness rate of internal medicine separately).

[0327] Select data source: such as "system automatically recorded data" or "manually entered data", and specify which type of data to use to calculate the timeliness rate through the selector.

[0328] Radio buttons: Used for "choose one of two or more (and only one can be selected)", adapting to mutually exclusive rule items.

[0329] Radio buttons are typically represented by a circle with the text "option". Selecting one prevents you from selecting another. Common uses include:

[0330] Select calculation logic: For example, when calculating on-time rate, the two rules "unfinished tasks are not included in the statistics" and "unfinished tasks are counted as 'not on time'" are mutually exclusive, so use a radio button to let the user select one.

[0331] Choose the display format: for example, whether to display the results as "only percentage (e.g., 98%)" or "display details + percentage".

[0332] Select whether to include special cases: for example, "whether to include holiday tasks", select "yes" or "no" to determine whether holiday tasks are included in the on-time rate calculation.

[0333] In some embodiments, the method further includes generating a two-dimensional graph of the epidemic and drug procurement; in the two-dimensional graph, the horizontal axis represents time, and the vertical axis represents the epidemic and drugs.

[0334] In some embodiments, see Appendix Figure 4 The method may further include the following steps: displaying data related to smart procurement applications, a two-dimensional chart of drug procurement application trends, and two-dimensional charts of epidemic trends and drug procurement trends on a large screen.

[0335] The data related to smart procurement applications include: total number of applications, month-on-month growth and year-on-year growth of the total number of applications, number of suppliers, month-on-month growth and year-on-year growth of the number of suppliers, number of types of drugs procured, month-on-month data on the number of types of drugs procured, year-on-year data on the number of types of drugs procured, application completion rate, month-on-month and year-on-year data on the application completion rate, on-time delivery rate, month-on-month and year-on-year data on the on-time delivery rate.

[0336] The various parameters mentioned above, including month-on-month and year-on-year data, are represented by different colors. For example, parameters are represented in yellow, month-on-month data in green, and year-on-year data in red.

[0337] The two-dimensional graph of drug procurement application trends includes curves representing the quantity of drugs, inventory, consumption, and estimated demand, each with different colors.

[0338] The technical solution of this application achieves the following technical effects:

[0339] Improving procurement efficiency can achieve full-process automation, reduce manual operations, increase drug procurement efficiency, reduce procurement costs, and enable the system to automatically grab scarce drugs.

[0340] Business data analysis can monitor procurement data in real time, analyze drug usage and inventory status, provide decision-making support for managers, and improve management efficiency.

[0341] Reduce operating costs by improving procurement accuracy through smart procurement, reduce inventory buildup and capital tied up, avoid tying up hospital working capital, and reduce labor costs.

[0342] Secondly, this application proposes a smart procurement system for pharmaceuticals and consumables, see appendix. Figure 5 It includes a processor 51 and a memory 52 for storing processor-executable instructions.

[0343] The processor 51 is configured to execute the intelligent procurement method for pharmaceuticals and consumables described in any of the above-mentioned methods.

[0344] The processor 51 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0345] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of this disclosure. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the following claims.

[0346] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A smart procurement method for pharmaceuticals and consumables, characterized in that, include: Obtain historical drug procurement data, patient visit data, and epidemiological data from hospitals; Using a drug demand forecasting model, the first drug procurement data is predicted and determined based on the hospital's historical drug procurement data, patient visit data, and epidemiological data. Determine the adaptability coefficient for drug procurement adjustments, specifically including: The adaptability coefficient for drug procurement adjustments is determined using the following methods: Determine the doctor coefficient, patient coefficient, and seasonal coefficient; The doctor coefficient is used to quantify the impact of doctor factors on drug procurement volume. The patient coefficient is used to quantify the degree of influence of patient factors on drug procurement volume. The seasonal coefficient is used to reflect the impact of seasonal factors on drug procurement volume; the doctor coefficient, the patient coefficient, and the seasonal coefficient are nonlinearly refined and adjusted according to external characteristics to obtain the drug procurement adjustment adaptability coefficient; The drug procurement adjustment adaptability coefficient is obtained by jointly calculating the doctor coefficient, the patient coefficient, and the seasonal coefficient; or, The drug procurement adjustment adaptability coefficient is determined based on the deviation between the predicted and actual values ​​of the hospital's historical drug procurement data. Based on the drug procurement adjustment adaptability coefficient, the first drug procurement data is adjusted to obtain the second drug procurement data; The joint calculation of the doctor coefficient, the patient coefficient, and the seasonal coefficient includes: The interaction coefficient weights are determined based on any combination of the doctor coefficient, the patient coefficient, and the seasonal coefficient. The interaction coefficient is determined based on the interaction coefficient weights and any combination of the doctor coefficient, the patient coefficient, and the seasonal coefficient; A joint coefficient is determined based on any combination of the doctor coefficient, patient coefficient, and seasonal coefficient, and the interaction coefficient, and the joint coefficient serves as the adaptability coefficient for adjusting drug procurement. The determination of the drug procurement adjustment adaptability coefficient based on the deviation between the predicted and actual values ​​of the hospital's historical drug procurement data specifically includes: The hospital's historical drug procurement data is grouped according to preset dimensions to obtain multiple groups of hospital historical drug procurement data. Determine the predicted and actual values ​​of historical drug procurement data for each group of hospitals; The correction coefficient for each group is determined based on the predicted and actual values ​​of the historical drug procurement data of each hospital. The drug procurement adjustment adaptability coefficient is determined based on each set of correction coefficients; The refinement and adjustment of the doctor coefficient based on external characteristics includes: The external feature is that it is segmented into weekdays and holidays; On weekdays, the doctor's coefficient is the base value; The baseline value is a benchmark value for the doctor's coefficient set in a weekday scenario, providing a reference scale for the calculation of the doctor's coefficient; During holidays, the doctor's coefficient is the product of the fluctuation coefficient and the baseline value; or... External characteristics include: drug classification and physician scheduling; The doctor coefficients mentioned above have been refined and adjusted based on the drug classification and doctor scheduling.

2. The intelligent procurement method for pharmaceuticals and consumables according to claim 1, characterized in that, The nonlinear refinement adjustment of the seasonal coefficient based on external characteristics includes: Based on the aforementioned seasonal coefficient, the adjusted seasonal coefficient is determined by adjusting the weights according to time and seasonal coefficient; The seasonal coefficient adjustment weights are determined based on holidays or peak flu seasons.

3. The intelligent procurement method for pharmaceuticals and consumables according to claim 1, characterized in that, The refinement and adjustment of the patient coefficients based on external characteristics includes: The external characteristic is the degree of urgency; When the urgency level is normal, the patient coefficient is the baseline value of the patient coefficient; The patient coefficient baseline value is a preset value corresponding to the normal level of urgency. When the urgency level is expedited, the patient coefficient is determined based on the patient coefficient baseline value and the first patient coefficient weight value; In cases where the urgency level is sudden, the patient coefficient is determined based on the patient coefficient baseline value and the second patient coefficient weight value.

4. The intelligent procurement method for pharmaceuticals and consumables according to claim 1, characterized in that, The method further includes: In response to a query command sent from the front-end webpage, retrieve drug procurement data from the database; The drug procurement data is sent to the front-end page, where the following parameters of the drugs are displayed in a list format: The information includes the drug name, manufacturer, statistical period, drug code, specifications, negotiation data volume, negotiation duration, monthly task status, yearly task status, and overall task progress.

5. The intelligent procurement method for pharmaceuticals and consumables according to claim 1, characterized in that, The method further includes: In response to the add command sent by the front-end webpage, it receives the newly added drug data entered on the front-end page.

6. The intelligent procurement method for pharmaceuticals and consumables according to claim 1, characterized in that, The method further includes: In response to the save command sent by the front-end time rate setting webpage, receive the time input in the front-end time rate setting webpage; The aforementioned "timeliness" is used to define the time standard in the procurement process; The front-end timeliness settings webpage also includes selectors and radio buttons; The selector and the radio button are used to flexibly configure the rules or filtering conditions for calculating the timeliness.

7. A smart procurement system for pharmaceuticals and consumables, characterized in that, processor; Memory used to store processor-executable instructions; The processor is configured to execute the intelligent procurement method for pharmaceutical consumables as described in any one of claims 1 to 6.

Images