JPH04169849A - Threshold setting method for discriminating data - Google Patents

Abstract

PURPOSE:To make it possible to set a threshold value readily by providing a function for setting the threshold value by utilizing the set of the specified measured data, and providing a corrective function for the threshold value. CONSTITUTION:At first, logic for judgment is formed. Namely, a discriminating expression is entered into the condition part of a rule, and the result of the discrimination is entered into an executing part on a screen 21. In the threshold- value input part of the discriminating expression, the part of a threshold value 22 and a part of a threshold value setting 23 are selected with a mouse 24. The selected specified data are inputted from an analyzing device or other devices. Then, the threshold value is set and the threshold value is corrected. Then, measurement is performed, and the discrimination processing with the threshold value is performed. The threshold value is corrected in real time. The above described working is repeated, and the discriminating working is performed. Thus, the threshold value can be readily set.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to an automated system for checking a large amount of test data generated from an automatic biochemical analyzer, and in particular, to a system for automating the checking of test data generated in large quantities from an automatic biochemical analyzer, and in particular, for setting a discrimination threshold in the checking logic. Regarding the method of

[Conventional technology]

Conventionally, regarding the discrimination threshold, analysis errors obtained from daily quality control have been used as described in Nishihata et al. etc., based on
We have determined the base value, and as you can see in Fig. 2, we have adopted a fixed value. Also, what is the threshold value you have set?

It can be fixed until all data is checked.

[Problem to be solved by the invention]

The conventional technique described above has the problem that it is necessary to perform trial and error when setting the threshold value, which is time-consuming. In addition, the threshold value is always fixed and cannot follow fluctuations in measurement data, resulting in problems such as a drop in 1 discrimination accuracy.

SUMMARY OF THE INVENTION An object of the present invention is to provide a means for easily setting a threshold value in order to solve the above-mentioned problems. Additionally, in line with fluctuations in measurement data.

The purpose of this paper is to provide a method for correcting a threshold value.

[Means to solve the problem]

In order to solve the above problems, we provided a function to set a threshold value and a function to correct the threshold value using a specific set of measurement data. Furthermore, a function is provided to correct the threshold value in real time according to fluctuations in the sequentially output measurement data.

[Effect]

FIG. 1 shows a flowchart of threshold setting work and discrimination work according to the present invention. In step 11, discrimination logic is created. In step 12, the type of measurement data to be used for threshold setting is selected. In step 13, selected specific measurement data is input from one analytical device or another processing device. In step 14, threshold setting and threshold correction are performed. Step 15
In step 16, a measurement is performed, and in step 16, a determination process using a threshold value is performed. In step 17, threshold correction processing is performed in real time. The determination operation is performed by repeating steps 15 to 17.

below, (1) Creation of discriminant logic (2) Threshold setting (3) Seed correction The effect of this will be explained.

1. Creation of discriminant logic First, we will explain the operation of creating discriminant logic using rules on the screen.

As shown in FIG. 2, on the screen 21, the discriminant is written in the condition part of the rule, and the result of the judgment is written in the execution part. In the discriminant threshold input section 22 in the figure, it is also possible to set a threshold value using conventional numerical values. When performing threshold setting work according to the present invention, move the mouse 22 and 23 in the figure to
Select with 4.

The above is the operation of creating discriminant logic.

2. Threshold Setting Once the threshold setting process is selected by the above action, the type of measurement data to be used for threshold setting is then selected. These measurement data are data accumulated through previous measurements.

The types of measurement data used for threshold setting are: (1) Past measurement data (including normal and abnormal objects) (2) Past measurement normal object data (3) Past measurement abnormal object data (4) ) Targets standard measurement object data, etc. These are displayed on the screen as a menu, and the user selects the target measurement data.

Once the measurement data to be used for threshold setting is selected, data from the database of the analyzer or other processing device is selected.
Data is input and displayed in a graph format (32, (2) in FIG. 3) on the screen (31, (1) in FIG. 3). Here, as an example, a distribution map of past accession data is shown.

Furthermore, as threshold setting methods, (1) the minimum value or maximum value of the measured data, (2) the value that satisfies the required detection rate α% from the measured data, and (3) the +α or − of the measured value with the highest frequency value. α (4) Indicate the position of the threshold value on the screen, not the numerical value.

etc.

First, if you select the minimum value or maximum value in (1), the measurement data used for threshold setting will be automatically set.
The minimum and maximum values are calculated and set as thresholds. An example of operation is shown in FIG. 3 (1). In the case of threshold setting based on the minimum value, a dotted line 34 (a) representing the threshold is displayed, indicating the position of the threshold on the data distribution. In the case of threshold setting based on the maximum value, dotted line 34 (
b) is displayed.

When (2) is selected, when the user inputs the request detection rate, a threshold that satisfies the request detection rate is calculated from the measurement data used for threshold setting, and is set as the threshold. An example of operation is shown in FIG. 3 (2). The shaded area 35 represents the detection rate classified based on the threshold value from the entire measurement data. A dotted line 34 is displayed at the position of the automatically calculated threshold value.

If (3) is selected, α is set by the user.

An example of operation is shown in FIG. 3 (3). When -α'' is specified for α, a dotted line 34 (a) is displayed, and when +α″ is specified, a dotted line 34 (b) is displayed.

If (4) is selected, the threshold value is designated using the cursor 33 as shown in FIG. 3 (4).

The threshold value to be set is indicated by a dotted line 34.

The above is the operation of the threshold value determination process.

3. Threshold Correction Function Further, the function of threshold correction processing will be explained below.

When a certain amount of quality control data and measurement data is accumulated during online measurement, the data is corrected using the threshold setting method described above.

Execution or non-execution of this correction work or the correction cycle can be set by the user.

〔Example〕

Hereinafter, the present invention will be explained in detail using FIGS. 4 to 12.

In this example, the measured values output from the biochemical analyzer are
A threshold value is used to determine normality or abnormality.

FIG. 4 is a block diagram of an apparatus to which the present invention is applied. This system includes a biochemical analyzer 41 having an automatic reexamination mechanism, a display device 43, a keyboard 44.
A threshold discrimination processing device 42 having a measurement data/quality control data file 45 is connected thereto.

FIG. 5 shows a processing flow for threshold value determination in this embodiment. First, in step 501, measured values are input from the vitalization testing device to the threshold value discrimination processing device. step 50
In steps 2 to 504, a check is performed with a threshold value set by the threshold setting operation according to the present invention.

In step 502, a reaction process check is performed using time series data. In step 503, an alarm check is performed using a warning issued by the analyzer. In step 504,
Cross-checks are performed between items using data ratios and correlations between multiple items, and judgment results are output for each item. Furthermore, in step 505, the priority of each check result is determined.

In step 506, the final result is output.

The above discrimination processing results are sequentially displayed on the result display screen. An example of a result sequential display screen is shown in FIG.

This screen displays an asterisk in the checked area that is determined to be abnormal for each sample.

Next, a specific example of the inter-item cross-check threshold setting process will be shown using FIGS. 7 to 11.

FIG. 7 shows an example of cross-checking between items. here,
If the item names GOT, GPT, LDH (7) are within the normal range, the correlation between the measured values of each item is further checked.

FIG. 8 first shows an example of a judgment check rule input screen.

As for the items GOT and GPT, threshold values are set using five numerical values as in the past. As an example, an example of setting a threshold value for determining whether the item LDH is normal or abnormal is shown.

First, point to the portions 81 and 82 in FIG. 8 with the mouse 83.

Next, select the measurement data to be used for threshold setting.

An example of the selection screen is shown in FIG. A menu 92 of measurement data used for threshold setting is displayed on the screen 91.

For example, when selecting measurement data, the mouse 93 is used to point to that part.

Next, a threshold setting work screen is displayed.

FIG. 10 is an example of a threshold setting work screen.

First, if you enter the item name of the measurement data in 1'01 on the screen, the distribution map of the measurement data used for threshold setting will be displayed.
A graph is displayed in the 102nd section of the figure.

Next, a threshold setting method is selected from the menu 103 in the figure. In the figure, 104 displays a 1-judgment formula, and thresholds set by various operations are displayed. Further, 105 and 106 in the figure are sequentially displayed by the threshold setting function selected from the menu 103.

For example, if you select "Maximum value" or "Minimum value" from the menu 103 in FIG. 10, the maximum value or minimum value of the measurement data will be displayed on the screen 104 and set as the threshold value. Figure (1) shows a specific example of threshold setting when 'maximum value' is selected. Here, the maximum data value "535" is set and displayed at 112 in the figure, and at the same time, a dotted line 111 is displayed on the graph.

Furthermore, when "detection rate" is selected from the menu 103 in FIG. 10, a frame 105 in FIG. 10 is displayed. The user
When an abnormality or a proportion (detection rate) of data to be detected using a discriminant is input, a threshold value corresponding to this detection rate is calculated and displayed at 104 in the figure. A specific example is shown in FIG. 11 (2).

For example, if you want to determine 95% from the entire measurement data, enter the detection rate of 95% in 113 in the figure, and the shaded area shown in 114 in the figure will automatically be used as the threshold value indicating 95% of the distribution map. It is calculated as 112 parts in the figure, "487
", and a dotted line 111 is also displayed. Specifically, in the logic shown in FIG. 8, regarding item LDH, the proportion of measured data detected by the judgment formula is 95%.
It shows the meaning.

Further, when "α value" is selected from the menu 103 in FIG. 10, a frame 106 in the figure is displayed. The user inputs the movement amount α value from the measurement value having the highest frequency distribution value. A specific example is shown in FIG. 11 (3). In this case, if "+80" is set at 115 in the figure as the α value, the value "400"', which is the highest frequency distribution value (in this case "320"') plus 80, is displayed at 112 in the figure. A dotted line 111 is displayed.

Furthermore, when the user selects "Specify Manual" from the menu 103 in FIG. 10, an arrow-shaped cursor 107 is displayed on the screen, and when the user selects an appropriate position on the screen, the threshold value is set at 104 in the figure. Figure 11 (
A specific example is shown in 4).

A threshold value setting operation is performed using the cursor 116. The dotted line 111 also moves in conjunction with the movement of the cursor 116. Then, the threshold value indicated by the dotted line 111 is displayed at 112 in the figure.

In addition, a portion 109 on the screen in FIG. 10 allows selection of execution or non-execution of real-time correction, and the threshold value is corrected by the specified threshold setting method during real-time measurement.

FIG. 12 shows an example of the rules created by the above setting work. This is the threshold value of 35 for item LDH.
5'' is an example in which the maximum value of normal data is set as the threshold, and instructions are given to correct the threshold in real time.

According to this example, when GOT=30 GPT=46 LDH=367 are given as input data, the determination result based on this determination rule is output as "abnormal" when only the influence of LDH is considered.

Further, when a certain amount of data determined to be normal is accumulated during online measurement, the maximum value of the accumulated data is compared with an already set threshold value. in this case,
If the threshold values are different, the maximum value is adopted as the threshold value and corrected.

The above is an example of an automatic biochemical analyzer to which the present invention is applied.

This is an example of a data discrimination method.

[Effects of the Invention] According to the present invention, threshold setting work based on a specific set of measurement data is facilitated, and the reliability of the threshold value is improved. At the same time, since the threshold value is corrected based on the fluctuation trend of measurement data that is sequentially given, flexible discrimination is realized and the accuracy of discrimination processing is improved.

[Brief explanation of drawings]

FIG. 1 is a diagram showing the threshold setting processing flow of the present invention;
Figure 2 is a diagram showing a hexagonal example of discriminant logic in rule format;
The figure shows how to set a threshold value from the system management data displayed on the screen, Figure 4 is a block diagram to which the present invention is applied, and Figure 5 is a diagram to which the present invention is applied. A diagram showing the biochemical test data discrimination processing flow, Fig. 6 is a diagram showing an example of a screen displaying the discrimination processing results, Fig. 7 is a diagram showing an example of discrimination processing, and Fig. 8 shows discrimination logic in rule format. Figure 9 is a diagram showing the screen for selecting the type of system management data, Figure 10 is a diagram showing the screen for threshold setting work, and Figure 11 is a diagram showing the screen for selecting the type of system management data. A diagram showing a specific example of setting a threshold value from the displayed quality control data,
Figure 12 was created. 2 Figure

Claims (1)

[Claims] 1. A data discrimination method that uses a specific set of measurement data and has a threshold setting function in a device that has a function of discriminating measurement data output from an analysis device using a threshold value. . 2. A data discrimination method that uses a specific set of measurement data and has a threshold correction function in a device that has a function of discriminating measurement data output from an analyzer based on a threshold value. 3. Data that has the function of varying the threshold value in real time using the sequentially outputted measurement data in a device that has the function of determining the measurement data sequentially output from the analyzer using a threshold value in real time. Discrimination method. 4. The measurement data in Items 1, 2, and 3 above is a data discrimination method that has a function of importing data from a processing device other than the analyzer.