A method and system for dynamic risk assessment of drug use in elderly patients

Abstract

The present invention discloses a method and system for dynamic risk assessment of medication for elderly patients, relating to the technical field of medication risk assessment. The method comprises collecting static factor data and dynamic factor data of elderly patients, pre-processing the collected static factor data and dynamic factor data, and generating a feature vector for risk assessment; calculating a comprehensive risk score based on the feature vector using preset assessment rules; comparing the comprehensive risk score with a preset risk grading threshold to determine the medication compliance risk level of the elderly patients. The system comprises data collection, pre-processing, risk assessment, and grading modules. The present application realizes multi-dimensional risk assessment by integrating static and dynamic data, thereby improving the comprehensiveness and timeliness of medication compliance risk assessment for elderly patients and providing technical support for medication intervention.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of drug risk assessment, and particularly relates to a method and system for dynamically assessing the drug risk of an elderly patient. BACKGROUND

[0002] With the aggravation of population aging, the drug safety problem of elderly patients is increasingly prominent. Clinical practice shows that the drug compliance (i.e. the degree to which patients take medicine according to the doctor's order) of elderly patients is significantly lower than that of other groups due to factors such as cognitive decline and multiple drug use. Non-compliance with medication may lead to poor treatment effect, increased adverse reactions and rising medical costs.

[0003] Traditional evaluation methods are mostly based on the age and cognitive score of elderly patients to make risk judgments, and the collection and analysis of dynamic changes in the drug-taking behavior of elderly patients and abnormal physiological indicators are not comprehensive enough, which leads to difficulty in reflecting the current drug-taking compliance risk state of elderly patients in a timely manner. SUMMARY

[0004] To solve the technical problem that the dynamic data of the drug-taking of elderly patients is not fully utilized in the evaluation of the drug-taking compliance of elderly patients, and the evaluation result is difficult to accurately reflect the current risk state in a timely manner, the present application provides a method and system for dynamically assessing the drug risk of an elderly patient.

[0005] The technical solution adopted by the present application is as follows:

[0006] The first aspect of the present application provides a method for dynamically assessing the drug risk of an elderly patient, comprising the following steps:

[0007] Collecting static factor data and dynamic factor data of the elderly patient, wherein the static factor data includes age, cognitive ability score and history of missed medication, and the dynamic factor data includes medication record data and physiological indicator data.

[0008] Preprocessing the collected static factor data and dynamic factor data to generate a feature vector for risk assessment.

[0009] Calculating a comprehensive risk score based on the feature vector through a preset evaluation rule.

[0010] Comparing the comprehensive risk score with a preset risk classification threshold to determine the drug-taking compliance risk level of the elderly patient.

[0011] Preferably, preprocessing the collected static factor data and dynamic factor data to generate a feature vector for risk assessment comprises the following contents:

[0012] segmenting and quantifying the age in the static factor data to obtain an age segment quantitative value; standardizing the cognitive ability score to obtain a cognitive ability standardized score; binary identifying the previous missed medication history to obtain a previous missed medication history binary identification value.

[0013] extracting and quantifying the medication time interval and the medication dose deviation in the dynamic factor data to obtain a medication time interval quantitative value and a medication dose deviation quantitative value; detecting and normalizing the abnormal values in the physiological index data to obtain a physiological index normalized value.

[0014] combining the age segment quantitative value, the cognitive ability standardized score, the previous missed medication history binary identification value, the medication time interval quantitative value, the medication dose deviation quantitative value, and the physiological index normalized value in a preset order to form a feature vector.

[0015] Preferably, the calculating a comprehensive risk score based on the feature vector through a preset evaluation rule includes the following contents:

[0016] multiplying the age segment quantitative value, the cognitive ability standardized score, the previous missed medication history binary identification value, the medication time interval quantitative value, the medication dose deviation quantitative value, and the physiological index normalized value in the feature vector by corresponding weight coefficients, accumulating each multiplication result to obtain a comprehensive risk score, and the comprehensive risk score is used to represent a quantitative value of the medication compliance risk of the elderly patient.

[0017] Preferably, the comparing the comprehensive risk score with a preset risk classification threshold to determine the medication compliance risk level of the elderly patient includes the following contents:

[0018] pre-setting a plurality of risk classification thresholds to form a risk level interval set, comparing the calculated comprehensive risk score with each risk classification threshold to determine the risk level interval thereof, and determining the medication compliance risk level of the elderly patient according to a preset corresponding relationship between the risk level interval and the medication compliance risk level.

[0019] Preferably, the calculating a comprehensive risk score based on the feature vector through a preset evaluation rule includes the following contents:

[0020] multiplying the age segment quantitative value, the cognitive ability standardized score, the previous missed medication history binary identification value, the medication time interval quantitative value, the medication dose deviation quantitative value, and the physiological index normalized value in the feature vector by corresponding basic weight coefficients to obtain a basic weighted intermediate value set.

[0021] extract an age section quantitative value corresponding to an age section interval from the feature vector, and form an age metabolism pathway cross feature pair with a drug metabolism pathway identifier in the medication record data item;

[0022] When the cross feature pair matches the attenuation coefficient, the medication dose deviation quantitative value weighting item and the physiological index normalized value weighting item in the basic weighted intermediate value set are corrected according to the attenuation coefficient, to generate a corrected weighted intermediate value.

[0023] The medication dose deviation quantitative value weighting item and the physiological index normalized value weighting item in the basic weighted intermediate value set are removed, and the remaining basic weighted intermediate value is accumulated to obtain a first part risk score; the corrected weighted intermediate value is accumulated to obtain a second part risk score; and the first part risk score and the second part risk score are accumulated to obtain a comprehensive risk score.

[0024] Preferably, the age data and the previous missed dose history data of the elderly patient are obtained through an electronic medical record system.

[0025] The cognitive ability score data of the elderly patient is obtained by using the MMSE scale; the medication record data of the elderly patient is obtained from an electronic medication record system; and the physiological index data of the elderly patient is obtained by using a physiological index monitoring device.

[0026] Preferably, the method further comprises the following content: for different risk levels, an individualized medication guidance scheme library containing medication time adjustment strategies, dose adjustment suggestions and missed dose remedial measures is pre-configured; when the medication compliance risk level of the elderly patient is determined, the medication guidance scheme of the corresponding level is called from the individualized medication guidance scheme library.

[0027] The second aspect of the application provides an elderly patient medication dynamic risk assessment system, which applies the above-mentioned elderly patient medication dynamic risk assessment method, comprising:

[0028] A data acquisition module is configured to acquire static factor data and dynamic factor data of an elderly patient, wherein the static factor data includes age, cognitive ability score and previous missed dose history, and the dynamic factor data includes medication record data and physiological index data.

[0029] A data preprocessing module is configured to preprocess the acquired static factor data and dynamic factor data to generate a feature vector for risk assessment.

[0030] A risk assessment module is configured to calculate a comprehensive risk score based on the feature vector and a pre-set assessment rule.

[0031] a risk grading module for comparing the comprehensive risk score with a preset risk grading threshold to determine the medication compliance risk level of the elderly patient.

[0032] The beneficial effects of the present application are at least one of the following:

[0033] By simultaneously collecting static factor data such as age, cognitive ability score, and dynamic factor data such as medication record data and physiological index data, the evaluation mode of the traditional method relying only on static factors is changed, and multi-dimensional coverage of the influencing factors of the medication compliance of the elderly patient is realized, providing more comprehensive data support for risk assessment.

[0034] Through the collection and analysis of medication record data and physiological index data, dynamic change information such as fluctuations in the medication behavior of the elderly patient and abnormal physiological states can be captured, compared with the traditional static evaluation method, the problem of lagging behind the actual risk state of the patient in the evaluation result is effectively improved, and the risk assessment is more in line with the actual situation of the current medication compliance of the patient.

[0035] By preprocessing the collected data to generate a feature vector and calculating a comprehensive risk score based on a preset evaluation rule, systematic integration analysis of static factors and dynamic factors is realized. This method overcomes the defect of insufficient utilization of dynamic data in the traditional method, so that the risk grading result can more accurately reflect the medication compliance risk level of the elderly patient, and provide a basis for subsequent intervention. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 The method flowchart of the first embodiment of the present application is shown in the figure.

[0037] Figure 2 The system block diagram of the second embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0038] The embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0039] Embodiment one provides a dynamic risk assessment method for elderly patients, as shown in the figure, including the following steps: Figure 1

[0040] Step 1, collect the static factor data and dynamic factor data of the elderly patient, the static factor data includes age, cognitive ability score and history of missed medication, and the dynamic factor data includes medication record data and physiological index data.

[0041] ​Considering that the traditional evaluation method only relies on static data such as age and cognitive score, it cannot capture dynamic risk signals such as fluctuations in medication behavior (such as missed doses, dose deviations) and physiological index abnormalities (such as sudden blood pressure rise), resulting in a disconnection between the evaluation results and the actual risk state of the patient. This step builds a "static + dynamic" data system covering multiple risk factors such as age, cognition, medication behavior, and physiological indicators.

[0042] In one possible implementation, the collection of static factor data and dynamic factor data of the elderly patient includes the following: obtaining age data and past missed dose history data of the elderly patient through an electronic medical record system; obtaining cognitive ability score data of the elderly patient using an MMSE scale; obtaining medication record data of the elderly patient from an electronic medication record system; and obtaining physiological indicator data of the elderly patient through a physiological indicator monitoring device.

[0043] For example, in the implementation process, the static factor data can be automatically read from the birth date field entered when the patient is registered through the patient basic information module of the hospital HIS electronic medical record system. The system background calculates the time difference between the current date and the birth date to generate an age value (accurate to the nearest year). For example, if the electronic medical record shows that the patient was born on March 12, 1945, and the current system time is June 12, 2025, the age is calculated to be 80 years old.

[0044] MMSE (Mini-Mental State Examination) is used for evaluation. The scale includes 7 dimensions such as orientation, memory, and attention, with a total of 30 questions. The patient gets 1 point for each correct answer, and the total score ranges from 0 to 30. After the evaluation, the system completes the standardization conversion (such as mapping the original score to the standardized interval of 0-1).

[0045] Through the medication history module of the electronic medical record system, the comparison results of the medication records and the actual medication records in the past 12 months are retrieved. If there is a system preset missed dose marker (such as not taking medication according to the doctor's order for 2 consecutive times), a binary identification is automatically generated (1 represents a history of missed doses, and 0 represents no history of missed doses).

[0046] For example, the medication record data includes the patient's medication information from the hospital PACS electronic medication record system, including drug name, specification, medication frequency (such as 2 times a day), and standard medication time points (such as 8:00 and 20:00).

[0047] For example, physiological indicator data collection, such as blood pressure, heart rate, and other indicators can be collected non-contactly through a wearable smart bracelet. The device automatically measures and stores data at a preset frequency (such as every 30 minutes); blood glucose indicators can be collected through a blood glucose meter;

[0048] Step 2, preprocessing the collected static factor data and dynamic factor data to generate a feature vector for risk assessment.

[0049] In one possible implementation, the preprocessing of the collected static factor data and dynamic factor data to generate a feature vector for risk assessment includes the following:

[0050] Segmented quantization is performed on the age in the static factor data to obtain an age segmented quantization value; the cognitive ability score is standardized to obtain a cognitive ability standardized score; the history of missed medication is binary identified to obtain a history of missed medication binary identification value;

[0051] The medication record data in the dynamic factor data is extracted and quantized in terms of medication time interval and medication dose deviation to obtain a medication time interval quantization value and a medication dose deviation quantization value; the physiological index data is subjected to abnormal value detection and normalization processing to obtain a physiological index normalized value;

[0052] The age segmented quantization value, the cognitive ability standardized score, the history of missed medication binary identification value, the medication time interval quantization value, the medication dose deviation quantization value, and the physiological index normalized value are combined in a preset order to form a feature vector.

[0053] For example, in the specific implementation process, the collected age value is mapped to the preset 5 age interval: 60-69 years old corresponds to the quantization value 0.2; 70-79 years old corresponds to the quantization value 0.4; 80-89 years old corresponds to the quantization value 0.6; 90-99 years old corresponds to the quantization value 0.8; ≥100 years old corresponds to the quantization value 1.0.

[0054] The MMSE original score (0-30 points) is converted to the [0, 1] interval by linear mapping: the standardization formula is S=1-(30-R) / 30, where R is the original score and S is the standardized score; for example: if the MMSE score is 24 points, the standardized score is 1-(30-24) / 30=0.8.

[0055] The history of missed medication is directly binary identified using the binary identification generated in step 1: a history of missed medication (labeled as "1" in step 1) is converted to 1.0; no history of missed medication (labeled as "0" in step 1) is converted to 0.0.

[0056] Example, medication record data quantification includes medication time interval extraction: calculate the deviation of actual medication time and order time (minutes), and quantify according to the following rules: ≤15 minutes corresponds to a quantification value of 1.0 (complete compliance) 16-30 minutes corresponds to a quantification value of 0.8 (basic compliance); 31-60 minutes corresponds to a quantification value of 0.6 (mild delay); 61-120 minutes corresponds to a quantification value of 0.4 (moderate delay); 120 minutes corresponds to a quantification value of 0.2 (severe delay); For example: the order time is 8:00, and the actual medication time is 8:25, the deviation is 25 minutes, and the quantification value is 0.8.

[0057] Example, medication dose deviation extraction includes calculating the percentage deviation of actual medication dose and order dose: deviation rate formula: D = | (actual dose - order dose) / order dose |; Map the deviation rate to the quantification value: ≤5% corresponds to a quantification value of 1.0 (complete compliance); 6-10% corresponds to a quantification value of 0.8 (basic compliance); 11-20% corresponds to a quantification value of 0.6 (mild deviation); 21-50% corresponds to a quantification value of 0.4 (moderate deviation); 50% corresponds to a quantification value of 0.2 (severe deviation); For example: the order dose is 100mg, and the actual dose is 92mg, the deviation rate is 8%, and the quantification value is 0.8.

[0058] Example, physiological indicator data normalization includes outlier detection: identify outliers using the 3σ criterion: calculate the mean μ and standard deviation σ of the last 7 days of data If the current value is outside the range [μ-3σ, μ+3σ], it is marked as an outlier For example: the mean systolic blood pressure of a patient for the last 7 days is 130mmHg, and the standard deviation is 5mmHg, then the outlier range is [115, 145]mmHg. Normalization processing includes mapping the physiological indicator to the [0, 1] interval:

[0059] For positive indicators such as blood pressure and blood sugar, the formula is: N = (V - V min ) / (V max -V min ); Where N represents the normalized value; V represents the current measurement value; V min表示 The clinical minimum value of the indicator (such as systolic blood pressure 70mmHg); V max表示 The clinical maximum value of the indicator (such as systolic blood pressure 220mmHg).

[0060] For indicators such as heart rate that have an ideal interval: if G is within the ideal interval, then N = 1.0;

[0061] If G is outside the ideal interval, then N = 1 - |G - G i | / (G max -G i ). Where N represents the normalized value; G represents the current measurement value; Gi represents the ideal value of the index (e.g. heart rate 75 beats / min), G max represents the clinically safe upper limit of the index (e.g. heart rate 150 beats / min); for example: patient heart rate 85 beats / min, G i = 75, G max = 150, the normalized value is 1- |85-75| / (150-75) = 0.87.

[0062] Example feature vector: assuming that the data of an elderly patient after pre-processing results in: age 80 years corresponding to the segmented quantification value 0.6; MMSE score 24 points corresponding to the standardized value 0.8; history of missed medication corresponding to the binary value 1.0; medication time deviation 25 minutes corresponding to the quantification value 0.8; medication dose deviation 8% corresponding to the quantification value 0.8; blood pressure 150 / 95 mmHg corresponding to the normalized value 0.53; heart rate 85 beats / min corresponding to the normalized value 0.87; blood glucose 6.8 mmol / L corresponding to the normalized value 0.7.

[0063] The generated feature vector is: [0.6, 0.8, 1.0, 0.8, 0.8, 0.53, 0.87, 0.7]. The feature vector converts heterogeneous data sources into standardized inputs that can be directly used for risk calculation through unified quantification standards, solving the problem of inconsistent data dimensions and ineffective fusion in traditional evaluation.

[0064] Step 3, calculating the comprehensive risk score based on the feature vector through a pre-set evaluation rule.

[0065] In the first possible implementation, the calculation of the comprehensive risk score based on the feature vector through a pre-set evaluation rule includes the following contents:

[0066] The age segmented quantification value, cognitive ability standardized score, history of missed medication binary identification value, medication time interval quantification value, medication dose deviation quantification value, and physiological index normalized value in the feature vector are multiplied by the corresponding weight coefficients respectively, and the multiplication results are accumulated to obtain the comprehensive risk score, which is used to represent the quantification value of the medication adherence risk of the elderly patient.

[0067] wherein, wherein the weight coefficients are pre-set according to the influence degree of each data item on the medication adherence risk.

[0068] For example, according to the clinical expert consensus, the basic weight coefficients of each data item are set as follows: age segmented quantification value: 0.15; cognitive ability standardized score: 0.20; history of missed medication binary value: 0.15; medication time interval quantification value: 0.15; medication dose deviation quantification value: 0.15; blood pressure normalized value: 0.05; heart rate normalized value: 0.05; blood glucose normalized value: 0.05.

[0069] The weighted calculation and accumulation, the feature vector is: [0.6, 0.8, 1.0, 0.8, 0.8, 0.53, 0.87, 0.7]; Each data item weighted calculation process: 0.6x0.15=0.09; 0.8x0.20=0.16; 1.0x0.15=0.15; 0.8x0.15=0.12; 0.8x0.15=0.12; 0.53x0.05=0.0265; 0.87x0.05=0.0435; 0.7x0.05=0.035. Accumulation to get the comprehensive risk score: 0.09+0.16+0.15+0.12+0.12+0.0265+0.0435+0.035=0.745. The comprehensive risk score of the patient is 0.745.

[0070] Considering that in the first embodiment, the basic weighting does not consider the cross effects of the metabolic capacity of the elderly patients declining with age and the age of the drug metabolism pathway (such as the risk is higher when the elderly patients take kidney metabolism drugs), which may lead to the evaluation result not matching the clinical practice, in the second possible implementation, the calculation of the comprehensive risk score based on the feature vector by the pre-set evaluation rule includes the following contents:

[0071] The age segmentation quantitative value, the cognitive ability standardized score, the binary identification value of the history of missed taking, the drug taking time interval quantitative value, the drug dose deviation quantitative value, and the physiological index normalized value in the feature vector are multiplied by the corresponding basic weight coefficients respectively to obtain a basic weighting intermediate value set. The basic weight coefficients are pre-set according to the influence degree of each data item on the drug adherence risk.

[0072] The age segment interval corresponding to the age segmentation quantitative value and the drug metabolism pathway identifier in the drug record data item are extracted from the feature vector to form an age metabolism pathway cross feature pair; the metabolic capacity attenuation coefficient set corresponding to the cross feature pair is retrieved through the elderly drug risk factor association database.

[0073] When the cross feature pair matches the attenuation coefficient, the drug dose deviation quantitative value weighting item and the physiological index normalized value weighting item in the basic weighting intermediate value set are modified according to the attenuation coefficient to generate a modified weighting intermediate value.

[0074] The drug dose deviation quantitative value weighting item and the physiological index normalized value weighting item in the basic weighting intermediate value set are removed, and the remaining basic weighting intermediate values are accumulated to obtain a first part risk score; the modified weighting intermediate value is accumulated to obtain a second part risk score; the first part risk score and the second part risk score are accumulated to obtain a comprehensive risk score.

[0075] For example, the basic weighted intermediate value set is calculated using the same basic weight coefficients as in the first embodiment. Details are not repeated here.

[0076] For example, the cross-feature pair extraction and attenuation coefficient retrieval assume that the patient: the age segmentation quantitative value 0.6 corresponds to the age interval 80-89 years old; the medication record shows that the patient is taking a kidney metabolism drug (such as metformin). Through the old drug risk factor association database retrieval, the metabolic capacity attenuation coefficient of the 80-89 year old population taking the kidney metabolism drug is obtained: 1.35 (i.e. the risk increases by 35%).

[0077] The weight of the two affected data items is corrected: the original weighting value of the medication dose deviation quantitative value is 0.12; the corrected weighting value is 0.12x1.35=0.162; the original weighting value of the blood pressure normalization value (the kidney metabolism drug has a strong correlation with blood pressure) is 0.0265; the corrected weighting value is 0.0265x1.35=0.035775.

[0078] For example, the partial accumulation calculation: the first part of the risk score (unmodified data items) includes: age: 0.09; cognitive ability: 0.16; missed dose history: 0.15; medication time interval: 0.12; heart rate: 0.0435; blood glucose: 0.035; total: 0.09+0.16+0.15+0.12+0.0435+0.035=0.6085.

[0079] The second part of the risk score (corrected data items): medication dose deviation: 0.162; blood pressure: 0.035775 total: 0.162+0.035775=0.197775.

[0080] The patient's comprehensive risk score is 0.6085+0.197775=0.806275.

[0081] The score improvement of the second embodiment reflects the potential risk increase of the elderly patient taking kidney metabolism drugs; the dynamic weight correction mechanism makes the evaluation result closer to the clinical practice, and provides a quantitative basis for personalized medication guidance.

[0082] Step 4, compare the comprehensive risk score with the preset risk classification threshold to determine the medication adherence risk level of the elderly patient.

[0083] In one possible implementation, comparing the comprehensive risk score with the preset risk classification threshold to determine the medication adherence risk level of the elderly patient includes the following:

[0084] A plurality of risk classification thresholds are preset to form a risk level interval set; the calculated comprehensive risk score is compared with each risk classification threshold to determine the risk level interval in which it is located; and the medication adherence risk level of the elderly patient is determined according to a preset correspondence between the risk level interval and the medication adherence risk level.

[0085] For example, four risk levels and corresponding intervals are preset:

[0086] Risk level Risk score interval Clinical significance Low risk [0,0.4) Good medication adherence Medium risk [0.4,0.6) Mild adherence problems exist High risk [0.6,0.8) Significant adherence risk requiring intervention Very high risk [0.8,1.0] Serious medication safety concerns exist

[0087] Taking the comprehensive risk score 0.745 calculated in the first implementation in step 3 as an example; threshold comparison: 0.745≥0.6 and 0.745<0.8; corresponding risk level interval [0.6, 0.8), determined as high risk.

[0088] Taking the comprehensive risk score 0.806275 calculated in the second implementation in step 3 as an example: 0.806275≥0.8 and 0.806275≤1.0; corresponding risk level interval [0.8, 1.0], determined as extremely high risk.

[0089] Further, in a possible implementation, the following content is further included: for different risk levels, an individualized medication guidance scheme library containing medication time adjustment strategies, dose adjustment suggestions, and missed dose remediation measures is preconfigured; when the medication adherence risk level of the elderly patient is determined, the medication guidance scheme of the corresponding level is retrieved from the individualized medication guidance scheme library.

[0090] For example, the medication time adjustment strategy corresponding to the low risk level (risk score 0-0.4) is to maintain the current medication time, and a fixed alarm reminder is suggested. Example: “Continue to take antihypertensive drugs at 8:00 and 20:00 every day, and suggest using the phone alarm function”.

[0091] The dose adjustment suggestion is to maintain the current dose, and to review liver and kidney function every 3 months. Example: “Continue to take aspirin 100mg / day, and suggest checking blood coagulation function every quarter”.

[0092] The missed dose remediation measure is to take the missed dose immediately if the missed time is less than 12 hours, and to skip the missed dose and take it normally next time if it is more than 12 hours. Example: “If you forget to take antihypertensive drugs, take them immediately if the next medication time is more than 12 hours, otherwise skip them”.

[0093] For example, the medication time adjustment strategy corresponding to the medium risk level (risk score 0.41-0.6) is to adjust to once a day long-acting preparation to reduce the risk of missing a dose. Example: “Suggest changing nifedipine sustained-release tablets to nifedipine controlled-release tablets, once a day, and take them after a fixed breakfast”.

[0094] Dose adjustment suggestion: maintain current dose, but increase monitoring frequency. Example: "Continue taking metformin 500 mg bid, recommend monitoring fasting blood glucose 3 times per week."

[0095] Missed dose remedy: if the missed dose is < 6 hours, take it immediately; if > 6 hours, skip the missed dose and take the next one normally. Example: "If you forget to take your diabetes medication, take it immediately if it is < 6 hours until the next dose; otherwise, skip it."

[0096] High-risk level (risk score 0.61-0.8) corresponding to medication time adjustment strategy: use intelligent medicine box to manage divided doses, set multiple reminders (such as mobile phone + voice alarm). Example: "Suggest using an intelligent medicine box (such as LifePod) and setting up three voice reminders for taking medicine in the morning, afternoon, and evening."

[0097] Dose adjustment suggestion: reduce the basal dose by 20% and increase the frequency of divided doses. Example: "Adjust warfarin from 3 mg / day to 2.5 mg / day, take it twice a day in the morning and evening."

[0098] Missed dose remedy: if missed, immediately contact the doctor or pharmacist for individualized guidance.

[0099] Extremely high-risk level (risk score > 0.8) corresponding to medication time adjustment strategy: inpatient observation adjustment, or supervised medication by caregivers. Example: "Suggest adjusting the medication regimen in the hospital or having family members supervise medication daily." Dose adjustment suggestion: suspend high-risk drugs and switch to alternative treatment options. Example: "Suspend the use of non-steroidal anti-inflammatory drugs and switch to acetaminophen for pain relief." Missed dose remedy: immediately seek emergency treatment. Example: "If you miss your blood pressure medication, go to the nearest hospital emergency department immediately."

[0100] Embodiment two provides a dynamic risk assessment system for medication of elderly patients, which applies the above-mentioned dynamic risk assessment method for medication of elderly patients, as shown in Figure 2 , comprising:

[0101] A data acquisition module is configured to acquire static factor data and dynamic factor data of the elderly patient, wherein the static factor data includes age, cognitive ability score, and previous missed dose history, and the dynamic factor data includes medication record data and physiological index data.

[0102] A data preprocessing module is configured to preprocess the acquired static factor data and dynamic factor data to generate a feature vector for risk assessment.

[0103] A risk assessment module is configured to calculate a comprehensive risk score based on the feature vector through a preset assessment rule.

[0104] a risk stratification module for comparing the composite risk score to a preset risk stratification threshold to determine a medication adherence risk level for the geriatric patient.

[0105] The above embodiments only express the specific implementation of the present application, which is described in more detail and in more detail, but it cannot be understood as a limitation on the scope of the patent of the present application. It should be noted that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of protection of the present application.

Claims

1. A method for dynamic risk assessment of medication use in elderly patients, characterized in that: The following steps are involved: Collecting static factor data and dynamic factor data of elderly patients, wherein the static factor data includes age, cognitive ability score and previous missed dose history, and the dynamic factor data includes medication record data and physiological index data; Preprocess the collected static factor data and dynamic factor data to generate feature vectors for risk assessment; Calculating a comprehensive risk score based on the feature vector using preset evaluation rules; comparing the comprehensive risk score with a preset risk grading threshold to determine the medication adherence risk level of the elderly patient; The collected static factor data and dynamic factor data are pre-processed to generate a feature vector for risk assessment, including the following: age in the static factor data is segmented and quantified to obtain age segmentation quantification values; cognitive ability scores are standardized and converted to obtain cognitive ability standardized scores; previous missed medication history is binary-identified to obtain a previous missed medication history binary identification value; Extract and quantify the medication time interval and medication dosage deviation of the medication record data in the dynamic factor data to obtain the quantified value of the medication time interval and the quantified value of the medication dosage deviation; perform abnormal value detection and normalization processing on the physiological index data to obtain the normalized value of the physiological index; The age segmentation quantification value, cognitive ability standardized score, previous missed medication history binary identification value, medication time interval quantification value, medication dosage deviation quantification value, and physiological index normalization value are combined in a preset order to form a feature vector; The calculation of the comprehensive risk score based on the feature vector through preset evaluation rules includes the following: multiplying the age segmentation quantified value, cognitive ability standardized score, previous missed medication history binary identification value, medication time interval quantified value, medication dosage deviation quantified value, and physiological index normalized value in the feature vector with the corresponding weight coefficient respectively, and accumulating the multiplication results to obtain a comprehensive risk score. The comprehensive risk score is used to characterize the quantitative value of the medication compliance risk of elderly patients.

2. A method for dynamic risk assessment of medication use in elderly patients according to claim 1, characterized in that: Comparing the comprehensive risk score with the preset risk grading threshold to determine the elderly patient's medication adherence risk level includes the following: Pre-set several risk classification thresholds to form a set of risk level intervals; Compare the calculated comprehensive risk score with each risk classification threshold to determine the risk level range in which it falls; The medication compliance risk level of elderly patients is determined based on the preset correspondence between the risk level range and the medication compliance risk level.

3. A method for dynamic risk assessment of medication use in elderly patients according to claim 1, characterized in that: The calculation of the comprehensive risk score based on the feature vector using a preset evaluation rule includes the following: Multiplying the age segmentation quantization value, cognitive ability standardized score, previous missed medication history binary identification value, medication time interval quantization value, medication dosage deviation quantization value, and physiological index normalization value in the feature vector by the corresponding basic weight coefficient to obtain a basic weighted intermediate value set; Extracting the age range corresponding to the age segmentation quantization value and the drug metabolic pathway identifier in the medication record data item from the feature vector to form an age-metabolism pathway cross-feature pair; retrieving a metabolic capacity attenuation coefficient set corresponding to the cross-feature pair through a geriatric medication risk factor association database; When the cross-feature pair matches the attenuation coefficient, the weighted item of the quantized value of the medication dosage deviation and the weighted item of the normalized value of the physiological index in the basic weighted intermediate value set are corrected according to the attenuation coefficient to generate a corrected weighted intermediate value; The weighted item of the quantified value of the medication dosage deviation and the weighted item of the normalized value of the physiological index in the basic weighted intermediate value set are removed, and the remaining basic weighted intermediate values ​​are accumulated to obtain the first part of the risk score; The second part of the risk score is obtained by summing up the modified weighted median values; Add the risk score of the first part and the risk score of the second part to get the comprehensive risk score.

4. A method for dynamic risk assessment of medication use in elderly patients according to claim 1, characterized in that: The age data and previous missed dose history data of elderly patients were obtained through the electronic medical record system; the cognitive ability score data of elderly patients were obtained using the MMSE scale; the medication record data of elderly patients were obtained from the electronic medication record system; and the physiological indicator data of elderly patients were obtained through physiological indicator monitoring equipment.

5. A method for dynamic risk assessment of medication use in elderly patients according to claim 1, characterized in that: Also included: For different risk levels, a personalized medication guidance program library is pre-configured, which includes medication time adjustment strategies, dosage adjustment recommendations, and missed dose remedial measures; after the elderly patient's medication compliance risk level is determined, the corresponding level of medication guidance program is retrieved from the personalized medication guidance program library.

6. A dynamic risk assessment system for medication use in elderly patients, characterized by: A method for dynamic risk assessment of medication use in elderly patients according to any one of claims 1 to 5 is applied, comprising: A data acquisition module, which is used to collect static factor data and dynamic factor data of elderly patients, wherein the static factor data includes age, cognitive ability score, and previous missed medication history, and the dynamic factor data includes medication record data and physiological index data; A data preprocessing module, which is used to preprocess the collected static factor data and dynamic factor data to generate a feature vector for risk assessment; a risk assessment module, configured to calculate a comprehensive risk score based on the feature vector using preset assessment rules; The risk grading module is used to compare the comprehensive risk score with a preset risk grading threshold to determine the medication compliance risk level of the elderly patient.

Images