Temperature fluctuation compensation algorithm for colorimetric blood culture system

Abstract

The invention discloses a colorimetric method blood culture system temperature fluctuation compensation algorithm, which specifically comprises the following steps: S1, data acquisition and pretreatment: arranging a high-precision temperature sensor in a blood culture instrument for acquiring temperature data in the instrument in real time at the acquisition frequency of 1 second to 10 minutes once, and meanwhile, acquiring the temperature data in the blood culture instrument; a detection device of the blood culture instrument collects a detection value of blood culture in real time, the detection value and temperature data are recorded synchronously, the collected temperature data and detection value are preprocessed, and data points which are obviously abnormal are removed. By establishing the corresponding relation between the internal temperature of the instrument and the detection value of the blood culture instrument, when the internal temperature of the instrument is influenced by severe fluctuation of the room temperature, the detection value is compensated and simulated through an algorithm to replace an actual detection value, so that the detection accuracy is improved, and the real-time compensation based on a temperature detection value quantitative model is realized. Detection value deviation caused by temperature fluctuation is effectively corrected, and the false positive rate is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of blood culture instrument detection, in particular to a colorimetric blood culture system temperature fluctuation compensation algorithm. BACKGROUND

[0002] As a high incidence and high mortality disease in hospital, sepsis is mainly caused by bloodstream infection (BSI), and blood culture is the core means for diagnosing BSI. The automatic blood culture system provides rapid diagnostic basis for life-threatening diseases such as sepsis and bacteremia by real-time monitoring of the growth and metabolism changes of microorganisms in the culture bottle, and is a key technology for clinical bacterial examination.

[0003] The detection accuracy of the existing blood culture instrument is easily affected by the environmental temperature: when the room temperature fluctuates sharply in a short time, the temperature stability of the culture environment inside the instrument is destroyed, resulting in abnormal detection values, and then false positive results are caused. To solve this problem, the existing technology usually adopts the method of rejecting the detection data in the temperature fluctuation period, however, this method will inevitably cause data loss. If in the scene where the room temperature fluctuates sharply and frequently, rejecting the detection data in the temperature fluctuation period will cause more data loss, and even miss most of the data, so that the subsequent data analysis is incomplete, finally leading to abnormal detection results, which cannot provide reliable basis for clinical diagnosis, especially in the scene where the room temperature fluctuates frequently, missing the data of the key microbial growth period may cause false negative, which seriously affects the reliability of diagnosis.

[0004] Therefore, a method is needed which can effectively correct the detection value when the room temperature fluctuates sharply, avoid data loss, and improve the detection accuracy. SUMMARY

[0005] The present application aims to provide a colorimetric blood culture system temperature fluctuation compensation algorithm to solve the problems raised in the background.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical solution: a colorimetric blood culture system temperature fluctuation compensation algorithm, specifically comprising the following steps:

[0007] S1, data acquisition and pretreatment: a high-precision temperature sensor is arranged inside the blood culture instrument to acquire the temperature data inside the instrument in real time, the acquisition frequency is 1 second-10 minutes once, at the same time, the detection device of the blood culture instrument acquires the detection value of the blood culture in real time, and the temperature data is recorded synchronously, and the acquired temperature data and detection value are pretreated to remove the obviously abnormal data points;

[0008] S2, establishing a corresponding relationship model: using the pre-processed data, adopting the method of multiple regression analysis, establishing the corresponding relationship model between the instrument internal temperature and the detection value, taking the instrument internal temperature T as the independent variable and the detection value D as the dependent variable, constructing the regression equation D=aT 2 +bT+c, wherein a, b, and c are regression coefficients, and the regression coefficients are solved by the least square method, so that the mean square error between the model prediction value and the actual detection value is minimized;

[0009] S3, real-time monitoring and judgment: real-time monitoring of the change of room temperature, when detecting that the fluctuation amplitude of room temperature within 10 minutes exceeds ±3℃, and the fluctuation range exceeds the instrument applicable temperature range, it is determined that the room temperature fluctuates violently, at this time the instrument internal temperature will be affected by the room temperature fluctuation and change;

[0010] S4, compensation simulation calculation: when it is determined that the room temperature fluctuates violently and the instrument internal temperature changes abnormally, according to the established corresponding relationship model, the corresponding compensation detection value is calculated by using the current real-time collected instrument internal temperature data to replace the actual detection value.

[0011] Preferably, in S1, the data points whose temperature values exceed the reasonable range caused by temperature sensor failure, below 0℃ or above 60℃, and the data points whose detection values appear mutation and do not conform to the growth rule of blood culture are rejected.

[0012] Preferably, in S2, in order to improve the accuracy of the model, the time factor can also be introduced as an auxiliary variable to construct a multiple regression model containing time t D=aT 2 +bT+ct+d, so as to improve the compensation accuracy of different culture stages.

[0013] Preferably, in S3, the temperature data inside the instrument is collected in real time to judge whether the change of the instrument internal temperature is abnormal change caused by the violent fluctuation of room temperature.

[0014] Preferably, in S4, if the current instrument internal temperature is T0, according to the regression equation D0=aT0 2 +bT0+c, the compensation detection value D0 is calculated, and D0 is used to replace the actual detection value for subsequent data analysis and judgment.

[0015] Preferably, the preprocessing of S1 also includes normalizing the detection value, specifically taking the detection value at the normal setting temperature of the instrument, preferably 36℃, as the reference, calculating the normalized value of the detection value at different temperatures, the formula is: normalized value=(detection value at certain temperature-reference temperature detection value) / reference temperature detection value, so as to eliminate the basic influence of temperature on the detection value and improve the stability of model training.

[0016] Preferably, the fitting method of the multiple regression analysis in S2 includes linear fitting, binomial fitting, trinomial fitting, tetranomial fitting or pentanomial fitting, and the optimal model is selected by comparing the determination coefficients of different fitting methods, wherein the R 2 ≥ 0.95.

[0017] Preferably, the basis for judging whether the change of the internal temperature of the instrument is an abnormal change caused by the severe fluctuation of the room temperature in S3 is the synchronism of the change trend of the internal temperature of the instrument and the fluctuation trend of the room temperature, the synchronous increase of the internal temperature of the instrument when the room temperature rises, and the positive correlation between the temperature change amplitude and the room temperature fluctuation amplitude, and the correlation coefficient is greater than or equal to 0.8.

[0018] Preferably, the compensated detection value is used to draw a microorganism growth curve in S4, and when the compensated detection value presents an increasing trend in line with the microorganism growth rule for 3 or more consecutive collection periods, it is determined that the blood culture is positive, so as to ensure the clinical application effectiveness of the compensated data.

[0019] Compared with the prior art, the present application has the following beneficial effects:

[0020] 1. The present application establishes the corresponding relationship between the internal temperature of the instrument and the detection value of the blood culture instrument, and when the severe fluctuation of the room temperature affects the internal temperature of the instrument, the detection value is compensated and simulated by an algorithm to replace the actual detection value, so as to improve the accuracy of detection, the real-time compensation based on the temperature detection value quantization model effectively corrects the detection value deviation caused by the temperature fluctuation, and reduces the false positive rate.

[0021] 2. The present application establishes the corresponding relationship model between the internal temperature of the instrument and the detection value, and when the severe fluctuation of the room temperature affects the internal temperature of the instrument, the detection value can be compensated and simulated by an algorithm in time to replace the actual detection value, so as to avoid the data missing problem caused by eliminating the detection value in the temperature fluctuation period, ensure the integrity of the data, and improve the accuracy and reliability of the detection result, and the multiple regression model with the time variable can adapt to the microorganism growth characteristics in different culture stages, and improve the compensation accuracy in complex scenes. BRIEF DESCRIPTION OF DRAWINGS

[0022] Fig. 1 It is a blood volume and detection value relationship curve of the present application;

[0023] Fig. 2 It is an actual reading curve of the present application;

[0024] Fig. 3 It is a compensated reading curve of the present application. DETAILED DESCRIPTION

[0025] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of the present application.

[0026] Please refer to Figs. 1-3 A colorimetric blood culture system temperature fluctuation compensation algorithm, specifically comprising the following steps: S1, data acquisition and preprocessing: a high-precision temperature sensor is arranged inside the blood culture instrument to collect temperature data inside the instrument in real time, the collection frequency is 1 second-10 minutes once, at the same time, the detection device of the blood culture instrument collects the detection value of the blood culture in real time, and the temperature data is recorded synchronously, the collected temperature data and detection value are preprocessed to remove obviously abnormal data points, eliminate temperature value exceeding the reasonable range caused by temperature sensor failure, data points below 0℃ or above 60℃, and data points with sudden change in detection value and not conforming to the growth rule of blood culture, the preprocessing also includes normalizing the detection value, specifically taking the detection value of the normal setting temperature of the instrument, preferably 36℃, as the reference to calculate the normalized value of the detection value at different temperatures, the formula is: normalized value=(detection value at certain temperature-reference temperature detection value) / reference temperature detection value, to eliminate the basic influence of temperature on the detection value and improve the stability of model training, the measurement accuracy of the high-precision temperature sensor is ±0.1℃, to ensure the accuracy of temperature data collection.

[0027] S2, establish a corresponding relationship model: using the preprocessed data, a corresponding relationship model between the temperature inside the instrument and the detection value is established by using the method of multivariate regression analysis, taking the temperature T inside the instrument as the independent variable and the detection value D as the dependent variable, and constructing the regression equation D=aT 2 +bT+c, wherein a, b and c are regression coefficients, the regression coefficients are solved by the least square method, so that the mean square error between the model prediction value and the actual detection value is minimized, in order to improve the accuracy of the model, the time factor can also be introduced as an auxiliary variable to construct a multivariate regression model D=aT 2 +bT+ct+d, to improve the compensation accuracy of different culture stages, the fitting method of multivariate regression analysis includes linear fitting, binomial fitting, trinomial fitting, tetranomial fitting or pentanomial fitting, the optimal model is selected by comparing the determination coefficients of different fitting methods, wherein the R 2 of the optimal model is greater than or equal to 0.95.

[0028] S3, real-time monitoring and judgment: real-time monitoring of the change of room temperature, when detecting that the fluctuation amplitude of room temperature within 10 minutes exceeds ±3℃ and the fluctuation range exceeds the applicable temperature range of the instrument, it is determined that the room temperature fluctuates violently, at this time the internal temperature of the instrument will be affected by the room temperature fluctuation and change, real-time collection of the internal temperature data of the instrument, judgment of whether the change of the internal temperature of the instrument is abnormal change caused by the violent fluctuation of the room temperature, the basis for judging whether the change of the internal temperature of the instrument is abnormal change caused by the violent fluctuation of the room temperature is the synchronism of the internal temperature change trend of the instrument and the room temperature fluctuation trend, when the room temperature rises, the internal temperature of the instrument rises synchronously, and the temperature change amplitude is positively correlated with the room temperature fluctuation amplitude, the correlation coefficient is greater than or equal to 0.8, the applicable temperature range of the instrument is 15-30℃, when the room temperature exceeds the range and the fluctuation amplitude within 10 minutes exceeds ±3℃, the violent fluctuation determination is triggered, the multiple regression model updates the coefficients based on the newly collected pretreatment data every 24 hours to adapt to the performance drift of the instrument in long-term use.

[0029] S4, compensation simulation calculation: when it is determined that the room temperature fluctuates violently and the internal temperature of the instrument changes abnormally, according to the established corresponding relationship model, the corresponding compensation detection value is calculated by using the real-time collected internal temperature data of the instrument, to replace the actual detection value, if the current internal temperature of the instrument is T0, according to the regression equation D0=aT0+bT0+c, the compensation detection value D0 is calculated, D0 is used to replace the actual detection value for subsequent data analysis and judgment, the compensation detection value is used to draw the microbial growth curve, and when the compensation detection value presents an increasing trend in line with the growth rule of microorganisms for 3 or more consecutive collection periods, it is determined that the blood culture is positive, so as to ensure the effectiveness of the clinical application of the compensated data. 2 +bT0+c, the compensation detection value D0 is calculated, and D0 is used to replace the actual detection value for subsequent data analysis and judgment, the compensation detection value is used to draw the microbial growth curve, and when the compensation detection value presents an increasing trend in line with the growth rule of microorganisms for 3 or more consecutive collection periods, it is determined that the blood culture is positive, so as to ensure the effectiveness of the clinical application of the compensated data.

[0030] The blood culture temperature fluctuation compensation algorithm establishes the corresponding relationship between the internal temperature of the instrument and the detection value of the blood culture instrument, when the room temperature fluctuates violently and affects the internal temperature of the instrument, the algorithm is used to compensate the detection value to replace the actual detection value, so as to improve the detection accuracy.

[0031] A temperature sensor and a detection device are arranged in the blood culture instrument, the internal temperature data and the detection value of the instrument are collected in real time, and the collected data is pretreated to remove obvious abnormal data points.

[0032] Using the pretreated data, a corresponding relationship model between the internal temperature of the instrument and the detection value is established by using multiple regression analysis method, the model includes a regression equation taking the internal temperature of the instrument as the independent variable and the detection value as the dependent variable, and a time factor can also be introduced as an auxiliary variable.

[0033] Real-time monitoring of room temperature changes, when the room temperature fluctuation amplitude exceeds the set threshold within the set time, it is determined that the room temperature fluctuates violently, and the instrument internal temperature data is collected in real time to determine whether the instrument internal temperature change is abnormal change caused by violent room temperature fluctuation.

[0034] When it is determined that the room temperature fluctuates violently and the instrument internal temperature changes abnormally, according to the established corresponding relationship model, the current real-time collected instrument internal temperature data can be used to calculate the corresponding compensation detection value, which can include time factor, to replace the actual detection value.

[0035] First, collect the data of the influence of different temperatures on the detection data changes. The following is the influence of different temperatures on the detection data of the same blood culture bottle collected by the indigo blood culture instrument:

[0036]

[0037]

[0038] Take 36℃, the normal setting temperature of the instrument, as the reference reading for normalization processing. The specific scheme is to subtract the 36℃ reading from the detection data at different temperatures, and then divide by the 36℃ reading. The normalized data is as follows:

[0039]

[0040]

[0041] Establish a calculation model of temperature and detection data. Assuming that the temperature is T℃ and the reading is y, the simulated normal temperature reading is Y(T) = y / (1-0.0264*T+0.9468).

[0042] The following data is the growth period monitoring data when the temperature fluctuates. The instrument detects that the data fluctuates when the temperature is 36.7℃, and the compensated temperature reading is as shown in the third column:

[0043]

[0044]

[0045]

[0046] The temperature suddenly rises to 36.7℃ at the 6th to 11th readings, causing the data to deviate significantly. After using the temperature compensation algorithm, the curve deviation is significantly reduced. The data shows that it can effectively reduce the detection data deviation caused by temperature fluctuation and improve the accuracy of the detection result.

[0047] While embodiments of the application have been shown and described, it is to be understood that the embodiments described are merely exemplary of the principles and application of the present application. Numerous modifications and adaptions can be effected without departing from the spirit and scope of the present application, which is not limited to the exact construction and arrangement described. It is intended, therefore, to cover all modifications and adaptions that fall within the scope of the claims and their equivalents.

Claims

1. A temperature fluctuation compensation algorithm for a colorimetric blood culture system, characterized in that, Specifically, the following steps are included: S1. Data Acquisition and Preprocessing: A high-precision temperature sensor is installed inside the blood culture instrument to collect the internal temperature data in real time. The acquisition frequency is once every 1 second to 10 minutes. At the same time, the detection device of the blood culture instrument collects the detection value of blood culture in real time and records it synchronously with the temperature data. The collected temperature data and detection values ​​are preprocessed to remove obviously abnormal data points. S2. Establish a Correspondence Model: Using the preprocessed data, a multiple regression analysis method is employed to establish a correspondence model between the instrument's internal temperature and the measured value. The instrument's internal temperature T is used as the independent variable, and the measured value D is used as the dependent variable. The regression equation D = aT is constructed. 2 +bT+c, where a, b, and c are regression coefficients. The regression coefficients are solved using the least squares method to minimize the mean square error between the model's predicted values ​​and the actual detected values. S3. Real-time monitoring and judgment: Real-time monitoring of room temperature changes. When the room temperature fluctuation exceeds ±3℃ within 10 minutes and the fluctuation range exceeds the instrument's applicable temperature range, it is judged as a drastic fluctuation in room temperature. At this time, the internal temperature of the instrument will change due to the fluctuation in room temperature. S4. Compensation Simulation Calculation: When it is determined that the room temperature fluctuates drastically and the internal temperature of the instrument changes abnormally, the corresponding compensation detection value is calculated based on the established correspondence model and the currently collected internal temperature data of the instrument in real time, so as to replace the actual detection value.

2. The temperature fluctuation compensation algorithm for the colorimetric blood culture system according to claim 1, characterized in that: S1 excludes data points where the temperature value exceeds the reasonable range due to temperature sensor malfunction, is below 0℃ or above 60℃, and data points where the detection value changes abruptly and does not conform to the growth pattern of blood culture.

3. The temperature fluctuation compensation algorithm for the colorimetric blood culture system according to claim 1, characterized in that: In S2, to improve the accuracy of the model, a time factor can be introduced as an auxiliary variable to construct a multiple regression model D = aT that includes time t. 2 +bT+ct+d is used to improve the compensation accuracy at different stages of cultivation.

4. The temperature fluctuation compensation algorithm for the colorimetric blood culture system according to claim 1, characterized in that: The S3 collects real-time temperature data inside the instrument to determine whether changes in the instrument's internal temperature are abnormal changes caused by drastic fluctuations in room temperature.

5. The temperature fluctuation compensation algorithm for the colorimetric blood culture system according to claim 1, characterized in that: In S4, if the current internal temperature of the instrument is T0, according to the regression equation D0 = aT0 2 +bT0+c, calculate the compensated detection value D0, and use D0 to replace the actual detection value for subsequent data analysis and judgment.

6. The temperature fluctuation compensation algorithm for the colorimetric blood culture system according to claim 1, characterized in that: S1's preprocessing also includes normalizing the detection values. Specifically, based on the instrument's normal set temperature, preferably 36℃, the normalized values ​​of the detection values ​​at different temperatures are calculated. The formula is: Normalized value = (Detection value at a certain temperature - Detection value at the reference temperature) / Detection value at the reference temperature. This eliminates the basic influence of temperature on the detection values ​​and improves the stability of model training.

7. The temperature fluctuation compensation algorithm for the colorimetric blood culture system according to claim 1, characterized in that: In S2, the fitting methods for multiple regression analysis include linear fitting, binomial fitting, trinomial fitting, tetranomial fitting, or quinomial fitting. The optimal model is selected by comparing the coefficients of determination of different fitting methods, where the R-squared value of the optimal model is... 2 ≥0.

95.

8. The temperature fluctuation compensation algorithm for the colorimetric blood culture system according to claim 1, characterized in that: The criteria for determining whether the internal temperature change of the instrument is an abnormal change caused by drastic fluctuations in room temperature in S3 are: the synchronicity between the trend of internal temperature change of the instrument and the trend of room temperature fluctuation; when the room temperature rises, the internal temperature of the instrument rises synchronously, and the temperature change amplitude is positively correlated with the room temperature fluctuation amplitude, with a correlation coefficient ≥ 0.

8.

9. The temperature fluctuation compensation algorithm for the colorimetric blood culture system according to claim 1, characterized in that: In S4, the compensated detection value is used to plot the microbial growth curve. When the compensated detection value shows an increasing trend consistent with the microbial growth pattern for three or more consecutive collection cycles, it is judged as a positive blood culture, so as to ensure the clinical application effectiveness of the compensated data.

Images