A method and system for managing prenatal screening reports

Abstract

The present application belongs to the technical field of medical data processing, and particularly relates to a kind of antenatal screening report management method and system, its method includes: obtaining the physiological characteristic data and detection index data of the pregnant woman to be screened;According to the physiological characteristic data, match dynamic reference subgroup in historical health sample library, calculate the physiological characteristic dynamic relative fluctuation rate of detection index data;Based on physiological characteristic data and detection index data, construct feature vector, search nearest neighbor sample in historical health sample library, calculate comprehensive risk management score;According to the comprehensive risk management score, the sample is divided into different risk levels, and the corresponding report management strategy of intelligent release, gray area review or high-risk interception is executed, so as to realize the accurate grading management of antenatal screening report.The present application can complete the dynamic correction of risk and the fine management of gray area to improve the screening accuracy.

Description

Technical Field

[0001] This invention relates to the field of medical data processing technology. More specifically, this invention relates to a method and system for managing prenatal screening reports. Background Technology

[0002] Prenatal screening is an important means of preventing birth defects in infants. The accuracy of prenatal screening is directly related to the diagnostic decisions of pregnant women and the rational allocation of medical resources. At present, the core functions of prenatal screening report management systems are often limited to the input, transfer, printing and storage of sample information. In the core medical risk assessment process, they generally rely on testing instruments for judgment. When existing testing instruments make judgments, they generally use a fixed median multiple combined with a static risk cutoff value for binary classification. For example, a risk value greater than a certain fixed value is marked as high risk.

[0003] Existing testing instruments tend to overlook the physiological differences of individual pregnant women when using static thresholds for judgment. That is, they use a uniform correction formula for all pregnant women. However, the weight difference of pregnant women will affect blood dilution and the slight deviation of gestational age will also affect the hormone secretion baseline. Both blood dilution and hormone secretion baseline will have a significant impact on the test results.

[0004] Furthermore, existing technologies lack refined management of gray zone samples. Current logic treats risk values ​​as isolated numbers, and a large number of gray zone samples at the critical value are often simply classified as low risk or pushed to retesting without intelligent auxiliary judgment based on historical big data. This extensive management approach can easily lead to the missed detection of abnormal samples in the gray zone or cause a large number of actually healthy pregnant women to undergo unnecessary retesting, resulting in a waste of medical resources. Summary of the Invention

[0005] To address the problem that existing technologies neglect physiological differences and have poor gray area management, resulting in low screening accuracy, this invention provides a prenatal screening report management method and system that can dynamically correct risks and refine gray area management to improve screening accuracy.

[0006] In a first aspect, the present invention provides a method for managing prenatal screening reports, comprising: acquiring physiological characteristic data and test indicator data of pregnant women to be screened from a laboratory information management system; performing standardized preprocessing on the test indicator data and preparing data based on a historical health sample database; matching a dynamic reference subgroup in the historical health sample database based on the physiological characteristic data; calculating the dynamic relative volatility of the physiological characteristics of the test indicator data based on the statistical data of the dynamic reference subgroup and the gestational age information in the physiological characteristic data; constructing a feature vector based on the physiological characteristic data and the test indicator data; searching for nearest neighbor samples in the historical health sample database; calculating a comprehensive risk management score based on the dynamic relative volatility of the physiological characteristics of the test indicator data and the distance information of the nearest neighbor samples; classifying samples into different risk levels based on the comprehensive risk management score, and implementing corresponding intelligent release, gray zone review, or high-risk interception report management strategies, thereby achieving precise hierarchical management of prenatal screening reports.

[0007] By employing the aforementioned technical solution, physiological characteristic data and test indicator data of pregnant women to be examined are obtained. A dynamic reference subgroup is matched based on a historical health sample database. The statistical data of this subgroup is used to calculate the dynamic relative volatility of the physiological characteristics of the test indicator data. This is then used to construct a feature vector and search for nearest neighbor samples, ultimately calculating a comprehensive risk management score to implement a tiered management strategy. This method abandons the traditional single static cutoff value, effectively solving the problem of misjudgment caused by ignoring individual physiological differences. Simultaneously, by performing dynamic evaluation in a multi-dimensional feature space, refined management of gray-zone samples is achieved, thereby significantly improving the accuracy of prenatal screening and the overall reliability of the system.

[0008] Preferably, the dynamic relative volatility of the physiological characteristics satisfies the following relationship: ; In the formula, Indicates the first The dynamic relative fluctuation rate of the physiological characteristics of each detection indicator. Indicates the current pregnant woman's number The measured values ​​of each detection indicator Represents the 1st statistical result obtained from the dynamic reference subgroup. The mean of each detection indicator, Represents the 1st statistical result obtained from the dynamic reference subgroup. The standard deviation of each detection indicator This represents a preset positive integer. Indicates the sensitivity coefficient based on gestational age. This indicates the current gestational age of the pregnant woman. Indicates the first The optimal gestational week for each testing indicator.

[0009] By adopting the above technical solution, a normal number and a gestational age sensitivity coefficient are introduced into the calculation formula for the dynamic relative volatility of physiological characteristics. The use of the normal number avoids the problem of the denominator becoming invalid when the sample standard deviation approaches zero, ensuring the robustness of the calculation. Simultaneously, by including a logarithmic term encompassing gestational age deviation, a more stringent evaluation is applied to samples from non-optimal sampling periods, effectively preventing missed detections due to poor sampling timing. This allows for accurate identification and characterization of the degree of abnormality of indicators based on statistical benchmarks of populations with similar physical characteristics.

[0010] Preferably, the comprehensive risk management score satisfies the following relationship: ; In the formula, This indicates the overall risk management score. This indicates the total number of testing indicators involved in the evaluation. Indicates the first The weight of each detection indicator, Indicates the first The dynamic relative fluctuation rate of the physiological characteristics of each detection indicator. This represents the health similarity compensation coefficient. This indicates the number of nearest neighbor samples. The feature vector of the current pregnant woman is related to the first... Euclidean distance between historical health sample vectors.

[0011] By adopting the above technical solution, a health similarity compensation mechanism is introduced into the calculation formula of the comprehensive risk management score. This mechanism calculates the Euclidean distance between the current pregnant woman's feature vector and the historical healthy sample vectors, forming a non-linear compensation term. This mechanism can effectively identify samples with high individual indicator values ​​but overall multi-dimensional feature patterns highly similar to those of healthy individuals, thereby significantly reducing the false positive rate, avoiding unnecessary re-examination of actually healthy pregnant women, and improving the accuracy and efficiency of screening.

[0012] Preferably, the physiological characteristic data includes the pregnant woman's age, actual gestational age, and weight; the detection index data includes alpha-fetoprotein concentration and free human chorionic gonadotropin concentration.

[0013] By adopting the above technical solution, the pregnant woman's age, actual gestational age, and weight are clearly defined as physiological characteristic data, while alpha-fetoprotein concentration and free human chorionic gonadotropin concentration are used as detection indicators. This design fully considers the impact of weight differences on blood dilution and the impact of gestational age deviations on hormone secretion baselines, thereby ensuring the medical rationality and physiological interpretability of the dynamic reference subgroup construction and providing reliable and clinically significant data support for subsequent risk assessment.

[0014] Preferably, the standardization preprocessing of the detection index data includes: performing maximum and minimum value normalization processing on the continuous variables in the detection index data.

[0015] By adopting the above technical solution, the continuous variables in the detection index data are normalized to their maximum and minimum values, eliminating the interference caused by differences in the units and numerical ranges of different indicators on the algorithm weights. This provides a unified and standardized numerical basis for the construction of feature vectors and the calculation of Euclidean distance, thereby ensuring the accuracy and comparability of the comprehensive risk management score calculation.

[0016] Preferably, the matching dynamic reference subgroup includes: setting a dynamic sliding window for gestational week and weight, and filtering out all historical samples in the historical health sample database that fall within the dynamic sliding window to form the dynamic reference subgroup.

[0017] Preferably, the search for the nearest neighbor sample includes: using the K-nearest neighbor algorithm to find the sample in the historical health sample database that has the closest Euclidean distance to the feature vector. One sample.

[0018] Preferably, the report management strategy for implementing the corresponding intelligent release, gray zone review, or high-risk interception includes: when the comprehensive risk management score is less than the low-risk threshold, it is determined to be a low-risk confidence sample, a report is automatically generated and pushed to the user terminal; when the comprehensive risk management score is greater than or equal to the low-risk threshold and less than the high-risk threshold, it is determined to be a gray zone sample, the report is suspended and the abnormal indicators are highlighted, and it is pushed to the expert workstation for manual review; when the comprehensive risk management score is greater than or equal to the high-risk threshold, it is determined to be a high-risk interception sample, the report is locked and an early warning notification is triggered.

[0019] By adopting the above technical solution, setting low-risk and high-risk thresholds, samples are classified into different risk levels and processed with differentiated strategies. Low-risk samples are intelligently released, improving report processing efficiency; gray-zone samples are suspended and pushed to an expert workstation for manual review to avoid missed detections in borderline cases; and high-risk samples have their reports locked and trigger alerts to ensure medical safety. This tiered mechanism achieves a balance between automated processing efficiency and the rigor of medical judgment, optimizing resource allocation and management processes.

[0020] Preferably, the method of highlighting abnormal indicators includes: obtaining the dynamic relative volatility of physiological characteristics of each detection indicator in the feature vector, marking the detection indicators whose values ​​exceed a preset volatility threshold as key risk factors, and visually highlighting the key risk factors in the expert workstation.

[0021] Secondly, the present invention provides a prenatal screening report management system, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned prenatal screening report management method is implemented.

[0022] By adopting the above technical solution, a computer program for the prenatal screening report management method is generated and stored in a memory so that it can be loaded and executed by a processor. This allows for the creation of a terminal device based on the memory and processor, making it convenient to use.

[0023] The beneficial effects of this invention are as follows: This invention utilizes historical negative data accumulated in the laboratory to construct a health manifold space that dynamically changes with gestational age and physical condition. By calculating the dispersion of the current sample relative to this dynamic space, health similarity is introduced as a risk compensation, enabling precise hierarchical management of screening reports. This method abandons the single static cutoff value, effectively solving the misjudgment problem caused by static thresholds, and improving the current situation of extensive management of gray area samples.

[0024] Furthermore, this invention eliminates the interference of individual differences by establishing dynamic benchmarks and reduces the false positive rate by using similarity compensation. This hierarchical management strategy based on dynamic assessment of multidimensional feature space significantly improves the intelligence level of the prenatal screening report management system, thereby ensuring the accuracy and reliability of medical risk assessment. Attached Figure Description

[0025] Figure 1 This is a flowchart of a prenatal screening report management method according to the present invention; Figure 2 This is a schematic diagram of the dynamic relative volatility benchmark analysis of physiological characteristics in this invention; Figure 3 This is a distribution map of the risk confidence deviation index based on health similarity compensation in this invention; Figure 4 This is a comparative analysis chart of the screening effectiveness of the present invention and existing technologies. Detailed Implementation

[0026] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0027] This invention discloses a method for managing prenatal screening reports, referring to... Figure 1 This includes steps S1-S4: S1. Obtain physiological characteristic data and test indicator data of pregnant women to be screened from the laboratory information management system, perform standardized preprocessing on the test indicator data, and prepare data based on the historical health sample database.

[0028] In an optional embodiment, the physiological characteristic data and test indicator data of the pregnant women to be screened need to be obtained in real time from the hospital's laboratory information management system. The physiological characteristic data of the pregnant women includes their age. Actual gestational age and weight Data such as alpha-fetoprotein (AFP) concentration and free human chorionic gonadotropin (hCG) concentration are used for pregnant women. The data for pregnant women's testing indicators refer to key biochemical indicators in serological screening. Specifically, key biochemical indicators in serological screening include AFP concentration and free hCG concentration, and are recorded as follows: The measured values ​​of each indicator are .

[0029] While acquiring physiological characteristic data and test indicator data of pregnant women to be screened in real time, in order to carry out subsequent statistical modeling, the system backend needs to pre-clean and store a historical sample database of those diagnosed as negative in the past three years. At the same time, Min-Max normalization is performed on all continuous variables to eliminate the influence of different units.

[0030] In this way, by acquiring and standardizing the data, the differences in dimensions between different indicators can be eliminated, providing a unified data foundation for subsequent accurate calculations and ensuring the accuracy of the assessment.

[0031] S2. Match a dynamic reference subgroup in the historical health sample database based on the physiological characteristic data. Calculate the dynamic relative fluctuation rate of the physiological characteristics of the test indicator data based on the statistical data of the dynamic reference subgroup and the gestational age information in the physiological characteristic data.

[0032] In an optional embodiment, the system retrieves samples from a historical health database based on the current gestational age of the pregnant woman being tested. and weight A dynamic sliding window is set up; in one optional embodiment, the dynamic sliding window can be the current gestational age of the pregnant woman being tested. Week, weight kg, then all historical samples falling within the dynamic sliding window are selected to form a dynamic reference subgroup. Finally, the k-th ... arithmetic mean of the indicators and standard deviation . and This represents the expected distribution of indicators for a normal person at that gestational age and weight, and finally, the dynamic relative volatility of physiological characteristics is calculated. : ; In the formula, Indicates the first The dynamic relative volatility of each detection indicator; the larger the value, the higher the degree of abnormality of the indicator. This indicates the measured concentration of this indicator in the current pregnant woman; This represents the mean of the indicator obtained from the dynamic reference subgroup. This represents the standard deviation of the indicator obtained from the statistics of the dynamic reference subgroup; It is a very small positive constant used to prevent the denominator from becoming invalid when the sample standard deviation approaches 0, ensuring the robustness of the formula. Its dimensions should be the same as... Maintain consistency; The gestational age sensitivity coefficient is a preset constant, which is set to 0.5 in this embodiment of the invention. This indicates the current gestational age of the pregnant woman; This indicates the optimal gestational week for sampling this indicator.

[0033] To more clearly illustrate the role of dynamic relative volatility of physiological characteristics and the calculation process, the following example will demonstrate this: First, let's assume the actual gestational age of pregnant woman A. The blood glucose level was 17.5, and the patient's weight was 70 kg. The measured value was alpha-fetoprotein (AFP) concentration. The optimal gestational week for alpha-fetoprotein (AFP) concentration is 45. The value is 16. The system matches the mean alpha-fetoprotein (AFP) concentration in a dynamic reference subgroup based on pregnant woman A's gestational age and weight. The standard deviation is 40. Set to 4. It is 0.0001. It is 0.5.

[0034] First, calculate the basic volatility component. ; The final dynamic relative volatility is then calculated. ; The example above shows that because the gestational age of 17.5 weeks deviates slightly from the optimal gestational age of 16 weeks, the volatility is slightly amplified. If the deviation from the optimal gestational age is even greater, the final dynamic relative volatility will be amplified by a larger factor.

[0035] Thus, by introducing dynamic group statistics and gestational age deviation penalties, abnormal indicators can be assessed based on a benchmark of people with similar physical conditions, avoiding misjudgments caused by a one-size-fits-all approach. At the same time, samples from non-optimal sampling periods are subject to more stringent scrutiny to prevent missed detections due to poor sampling timing.

[0036] S3. Construct feature vectors based on physiological characteristic data and detection index data, search for nearest neighbor samples in the historical health sample database, and calculate the comprehensive risk management score based on the dynamic relative volatility of physiological characteristics of detection index data and the distance information of nearest neighbor samples.

[0037] In an optional embodiment, the system combines the pregnant woman's age, weight, and various biochemical indicators into a feature vector. Subsequently, the K-Nearest Neighbors algorithm was used to find [the relevant data] in the historical health sample database. Euclidean distance nearest In this embodiment of the invention, one sample is provided. The score is 20. The overall risk management score is then calculated. Comprehensive risk management score The calculation method is as follows: ; In the formula, This represents the final risk score; the higher the score, the greater the risk. This indicates the total number of indicators involved in the evaluation; Indicates the first Clinical weight of each indicator; For single-term volatility; The health similarity compensation coefficient; Represents the current sample vector With the Euclidean distance between historical health sample vectors.

[0038] To more clearly illustrate the role and calculation process of the comprehensive risk management score, the following example will demonstrate this: First, let's assume there are two indicators in total, namely... For 2, indicator 1 Its weight It is 0.4; Indicator 2 Its weight is 0.8. It is 0.6; Without a compensation mechanism, the calculated comprehensive risk management score would be: ; If a compensation mechanism is added, when K is 20, the average similarity of the 20 nearest neighbor healthy samples is calculated. Its numerical representation is highly similar to that of healthy individuals; a compensation coefficient is set. If the value is 0.5, then the final comprehensive risk management score is calculated as follows: .

[0039] The calculations above show that without a compensation mechanism, the overall risk management score is 1.007, which may fall into the gray zone; with a compensation mechanism, the overall risk management score is 0.607, which may fall into the low-risk zone.

[0040] Thus, by introducing a health similarity compensation mechanism, abnormal samples that, although their individual values ​​are high, have an overall multidimensional feature pattern that is highly similar to historical healthy samples can be identified, thereby reducing false positives.

[0041] S4. Based on the comprehensive risk management score, the samples are divided into different risk levels, and corresponding intelligent release, gray zone review or high-risk interception report management strategies are implemented to achieve accurate hierarchical management of prenatal screening reports.

[0042] In an optional embodiment, the present invention sets a low-risk threshold. and high-risk threshold This allows the system to score based on comprehensive risk management. The calculation results will be used to automatically implement a hierarchical management strategy: when When the system determines that the sample is a low-risk confidence sample, it will perform intelligent release operation. At the same time, the system will automatically generate an electronic report, affix an electronic seal, and push it directly to the user terminal without manual review. when When the system identifies the sample as a gray area, it performs a gray area verification operation, suspends the report, and highlights the cause. For specific indicators that are too high, the data is pushed to the expert workstation, requiring mandatory manual review. Specifically, the system obtains the dynamic relative volatility of the physiological characteristics of each detection indicator in the feature vector, marks the detection indicators whose values ​​exceed the preset volatility threshold as key risk factors, and visually highlights the key risk factors in the expert workstation to help experts quickly locate the cause of the anomaly.

[0043] when When a sample is identified as high-risk, the system will perform a high-risk interception operation, automatically lock the report, trigger an alert SMS notification to the doctor, and recommend that the patient be transferred to the prenatal diagnosis process.

[0044] In this way, by implementing hierarchical management based on comprehensive risk scores, differentiated processing of samples with different risk levels is achieved, which significantly improves the automation level and management efficiency of report processing while ensuring the safety of screening.

[0045] To illustrate the technical effects of this solution more intuitively, the following explains the charts generated by this method: Reference Figure 2The distribution of discrete points representing historical healthy samples shows a slow upward trend with increasing gestational age. The static threshold of existing technologies remains constant with increasing gestational age, while the dynamic adaptive baseline calculated by this invention closely follows the changes in sample distribution. Points representing samples with special physical conditions are located above the static risk cutoff value in existing technologies, meaning they would be misjudged as high-risk under existing techniques. However, they are located below the dynamic adaptive baseline of this invention, indicating that this solution can identify samples with higher values ​​but within the normal fluctuation range relative to their physical condition.

[0046] Reference Figure 3 The figure shows a high-volatility, high-similarity sample. Its volatility is very high, but its health similarity is also high. In the prior art, this point would fall into the high-risk zone. However, in this invention, due to the compensation of the similarity of historical health sample features, this point is pulled back to the manual review zone or the low-risk pass zone, thus explaining the working principle of health similarity compensation.

[0047] Reference Figure 4 Compared to the bar chart representing the existing management system, the bar chart representing the system of the present invention shows that the false alarm rate of the system of the present invention is significantly reduced, while the first-pass rate is greatly improved, which intuitively demonstrates the beneficial effects of the present invention in improving management efficiency and accuracy.

[0048] This invention also discloses a prenatal screening report management system, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement a prenatal screening report management method according to the present invention.

[0049] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0050] In the description of this specification, "multiple" or "several" means at least two, such as two, three or more, unless otherwise expressly and specifically defined.

Claims

1. A method for managing prenatal screening reports, characterized in that, include: Physiological characteristic data and test index data of pregnant women to be screened are obtained from the laboratory information management system. The test index data is standardized and preprocessed, and data preparation is carried out based on the historical health sample database. Based on the physiological characteristic data, a dynamic reference subgroup is matched in the historical health sample database. Based on the statistical data of the dynamic reference subgroup and the gestational age information in the physiological characteristic data, the dynamic relative volatility of the physiological characteristics of the detection index data is calculated. Based on the physiological characteristic data and the detection index data, a feature vector is constructed, and the nearest neighbor sample is searched in the historical health sample database. Based on the dynamic relative volatility of the physiological characteristics of the detection index data and the distance information of the nearest neighbor sample, a comprehensive risk management score is calculated. Based on the comprehensive risk management score, the samples are divided into different risk levels, and corresponding intelligent release, gray zone review or high-risk interception report management strategies are implemented to achieve accurate hierarchical management of prenatal screening reports.

2. The method for managing prenatal screening reports according to claim 1, characterized in that, The dynamic relative volatility of the physiological characteristics satisfies the following relationship: ； In the formula, Indicates the first The dynamic relative fluctuation rate of the physiological characteristics of each detection indicator. Indicates the current pregnant woman's number The measured values ​​of each detection indicator Represents the 1st statistical result obtained from the dynamic reference subgroup. The mean of each detection indicator, Represents the 1st statistical result obtained from the dynamic reference subgroup. The standard deviation of each detection indicator This represents a preset positive integer. Indicates the sensitivity coefficient based on gestational age. This indicates the current gestational age of the pregnant woman. Indicates the first The optimal gestational week for each testing indicator.

3. The method for managing prenatal screening reports according to claim 1, characterized in that, The comprehensive risk management score satisfies the following relationship: ； In the formula, This indicates the overall risk management score. This indicates the total number of testing indicators involved in the evaluation. Indicates the first The weight of each detection indicator, Indicates the first The dynamic relative fluctuation rate of the physiological characteristics of each detection indicator. This represents the health similarity compensation coefficient. This indicates the number of nearest neighbor samples. The feature vector of the current pregnant woman is related to the first... Euclidean distance between historical health sample vectors.

4. The method for managing prenatal screening reports according to claim 1, characterized in that, The physiological characteristics data include the pregnant woman's age, actual gestational age, and weight; the detection index data include alpha-fetoprotein concentration and free human chorionic gonadotropin concentration.

5. The method for managing prenatal screening reports according to claim 1, characterized in that, The standardization preprocessing of the detection index data includes: performing maximum and minimum value normalization processing on the continuous variables in the detection index data.

6. The method for managing prenatal screening reports according to claim 2, characterized in that, The matching dynamic reference subgroup includes: setting a dynamic sliding window for gestational week and weight, and filtering out all historical samples from the historical health sample database that fall within the dynamic sliding window to form the dynamic reference subgroup.

7. The method for managing prenatal screening reports according to claim 3, characterized in that, The search for the nearest neighbor sample includes: using the K-nearest neighbor algorithm to find the sample in the historical health sample database that has the closest Euclidean distance to the feature vector. One sample.

8. The method for managing prenatal screening reports according to claim 1, characterized in that, The report management strategies for implementing corresponding intelligent release, gray zone review, or high-risk interception include: When the comprehensive risk management score is less than the low risk threshold, it is determined to be a low risk confidence sample, and a report is automatically generated and pushed to the user. When the comprehensive risk management score is greater than or equal to the low risk threshold and less than the high risk threshold, it is determined to be a gray zone sample, the report is suspended and the abnormal indicators are highlighted, and it is pushed to the expert workstation for manual review. When the comprehensive risk management score is greater than or equal to the high-risk threshold, it is identified as a high-risk intercept sample, the report is locked, and an early warning notification is triggered.

9. A method for managing prenatal screening reports according to claim 8, characterized in that, The highlighted display anomaly indicators include: The dynamic relative volatility of physiological characteristics of each detection indicator in the feature vector is obtained. Detection indicators whose values ​​exceed a preset volatility threshold are marked as key risk factors, and the key risk factors are visually highlighted in the expert workstation.

10. A prenatal screening report management system, characterized in that, include: A processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement a prenatal screening report management method according to any one of claims 1-9.

Images