JP2002022748A - Clinical inspection system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize a clinical inspection system which can judge whether the reinspection is necessary automatically and with high fidelity for an inspection result. SOLUTION: Criterion logic of requirement for reinspection is optimized by using stored information of the reinspection practice and the inspection results, which comprises the reinspection result for the specimen judged as in need of reinspection and the reinspection result for a part of the specimen judged as in no need of reinspection and the like.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a clinical test system and, more particularly, to a method for determining whether or not a retest is necessary based on test results.

[0002]

2. Description of the Related Art In a laboratory or an inspection center such as an inspection center, a test is performed on various items of a subject, and when a test result is obtained, the validity of the value is verified. I have. For example, if an abnormal value is measured as a result of the test, it is determined whether the test value of the subject is really abnormal or an abnormal value due to a measurement defect (e.g., an abnormality in the apparatus or a human error), and measures. If the possibility of the above failure is low, the test result is reported to the clinic as being correct, and if the possibility of the measurement failure is high, a retest is performed. The necessity of retesting was determined by a laboratory technician based on past experience with reference to a combination of test values and the previous test result for the patient (examinee).

In recent years, clinical tests have been automated, and clinical test systems that accumulate and manage test results as electronic data have become widespread. With the spread of clinical testing systems,
Logic (retest necessity determination means) for determining whether retest is necessary has been incorporated into clinical test systems. As such re-examination necessity determination means, a single item check, a previous value check, and an inter-item check are known as typical methods.

As a single item check, see, for example,
A technique disclosed in JP-A-151282 is known. This technique sets a reference range for each test item according to gender, age, and pregnant woman, and reports an abnormal result if the test value exceeds this reference range. Also,
Japanese Patent Application Laid-Open No. Hei 7-271873 describes a method of providing two types of reference ranges for each inspection item: an abnormal value range and a panic value range.

As the previous value check, see, for example,
Delta check described in US Pat.
A cumulative delta check described in JP-A-11-296605 is known. In the delta check, a reference value is set for the difference between the previous value and the current value for each inspection item, and an error is reported if the difference between the previous value and the current value exceeds the reference value. In the cumulative delta check, the standard deviation within an individual (the standard deviation of the test value for each test item for the same patient and patient) is determined for each test item, and the difference between the previous value and the current value is calculated. Calculate the ratio to the deviation. An abnormality is detected by providing a reference value to a value obtained by adding the above ratio obtained for each inspection item.

As a check between items, for example, Japanese Unexamined Patent Application Publication
As described in JP-A-296605, a reference value is provided for the ratio between items, and a method for judging abnormality or a logical operation expression using a plurality of inspection items is defined, and based on the truth of the logical operation expression. A method for determining an abnormality is known.

[0007]

SUMMARY OF THE INVENTION By applying the above-described technique, labor saving and efficiency in re-examination necessity determination are being realized. However, the test result that the technician determines that the retest is necessary does not completely match the result of the automatic retest necessity determination, and further improvement in accuracy is desired. Since it is difficult for the tester to make an error in the test result, in the automatic retest necessity judgment, generally, the reference value and the judgment logic are set so that the test result to be retested is not overlooked (missing detection). Test results that do not actually need to be retested are often used as criteria for determining the need for retesting. Therefore, it becomes necessary for the laboratory technician to confirm again the test result determined to be necessary for the retest by the automatic retest necessity determination, and to extract the test result that really needs the retest. In order to save labor of a laboratory technician, a more accurate retest necessity determination method is required.

[0008] Further, when applying the automatic retest necessity judgment,
Not only is it necessary to set a large number of reference values or logical expressions in advance, but also the composition of patients (diagnosis, age, etc.) of patients who visit each facility (hospital) differs, and a technician conducts a reexamination. It is difficult to set optimal reference values and logical expressions because the reference to be performed is different. Therefore, the reference value,
The logical formula cannot be easily optimized, and the automatic retest necessity determination cannot be effectively applied.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and realizes a more accurate automatic retest necessity determination and can easily optimize a reference value and a logic used in the retest necessity determination. It aims to provide a system.

[0010]

In a clinical test system having a retest necessity judging means for judging the necessity of a retest on a test result, a subject judged to need retest is already examined by a subject. Perform all retests using samples obtained from. The result of the retest is compared with the result of the previous test, and the retest necessity determination logic is corrected using the result.

[0011]

Embodiments of the present invention will be described below in detail with reference to the drawings.

FIGS. 1 to 3 are views showing an example of the configuration of a clinical test system to which the present invention can be applied. In FIG. 1, an analysis device server 20, an inspection server 30, and an inspection client 40 are connected to a LAN 50 laid in a hospital or an inspection room. Each server 2
For example, a personal computer or the like can be used as the inspection client 40. LAN
The first inspection device 11, the second inspection device 12, and the third inspection device 13 are further connected to 50. The analysis device server 20 controls each inspection device, performs necessary processing such as analysis on the inspection result output from each inspection device, and sends information to the inspection server 30 via the LAN 50. The test server 30 accumulates information such as the obtained test results and determines whether the test results need to be retested. The inspection client 40 can refer to information such as the inspection results stored in the inspection server 30 and the necessity of the re-inspection determined by the inspection server 30. In addition, an inspection result for an inspection that is not automated can be input using the inspection client 40. In this case, the input data such as the test result is transmitted from the test client 40 to the test server 30 and stored in the test server 30.

FIG. 2 shows an example of a system configuration in which the hospital 500 and the decision logic learning facility 550 in which the system described in FIG. 1 is constructed can transmit and receive data via the communication line 540. The learning server 60 is installed in the decision logic learning facility 550, and the learning server 60
Is connected to the communication line 540 via the first communication means 610. The inspection server 30 is connected to the communication line 540 via the second communication unit 620. The inspection server 30 and the learning server 60 can transmit and receive data via the communication line 540. Communication line 540
For example, various lines capable of data communication can be used, and for example, a telephone line can be used. Also, the first communication means 610 and the second communication means 620
Uses an appropriate communication means according to the type of line. For example, when a telephone line is used as the communication line 540, a modem can be used as the first communication unit 610 and the second communication unit 620. According to this example, in the example of FIG. 1, it is possible to entrust the learning server 60 with optimization of the criterion for determining whether or not the test result performed by the test server 30 needs to be retested. Therefore, the learning server 60 can perform the optimization using the data of other hospitals connected to the communication line 540, so that a more complete optimization can be expected than the system of FIG.

FIG. 3 shows an example in which each of the hospitals 520 and 530 connected via a communication line 540 uses an inspection center 510 in which the system described in FIG.
In each hospital, a test result reference terminal 70 is installed so that the test result of the requested test target can be known. According to this example, each hospital can perform a test including a higher optimized retest while reducing capital investment.

The present invention can be applied to any of the embodiments.
Can be further improved.

FIG. 4 shows a basic block diagram of the configuration of the inspection server 30 according to the present invention. 100 is a system bus. 101 is a CPU. Reference numeral 102 denotes a database, and an initial inspection result 111 and a re-examination necessity determination result 114
Alternatively, data of a result such as the retest result 115, various information such as the retest execution information 112, the diagnosis information 121, the contract information 122, the request information 123 or the retest necessity judgment logic 116, the retest necessity judgment reference value 118, Are re-examination necessity determination processing 113, re-examination necessity determination logic optimization processing 11
7, an execution program for the disease candidate estimation processing 124 and the like are stored. Reference numeral 103 denotes a RAM, which is a memory area used as a work area for executing various processes. Reference numeral 104 denotes an input / output interface for inputting various information such as test result input, diagnostic information input, retest execution information input, contract information input, and outputting various information such as test result output and retest results. I do.

This embodiment enumerates data, execution processing programs, and the like as functions that the inspection server 30 should have, and does not cover all of them. Further, in the system configuration as shown in FIG. 2, the reexamination necessity determination logic optimization processing and the like are executed in the determination logic learning facility 550, and the inspection server 30 advances the processing in cooperation therewith. .

A specific process for optimizing the reexamination necessity determination logic according to the present invention in a system having such a configuration will be described.

Here, optimization of a check between two items by a linear discriminant will be described as an example. In particular,
In the reexamination necessity determination process 113, the value of t is calculated according to (Equation 1) using the two inspection values x and y. If t ≧ 0, it is determined that reexamination is necessary, and if t <0, it is determined that reexamination is unnecessary.

T = ax + by + c (Equation 1) Here, a, b, and c are constants and are recorded together with (Equation 1) in the reexamination necessity determination logic 116 of the database of FIG. The initial value of this value uses data collected in advance,
Set an optimized value or set a value based on the engineer's past experience. In this case, since the optimal value differs depending on the characteristics of the patient at each hospital (how many patients have a disease) and the conditions of the testing method, testing device, reagents, etc., it is not possible to optimize at each hospital. Of course.

The principle of optimization of the retest necessity determination logic at this time will be described with reference to FIG.

FIG. 5A shows a data group 9 in which the inspection value x and the inspection value y are set on the horizontal axis and the vertical axis, respectively, and the distribution of the inspection values is indicated by x.
20, data groups 940 indicated by ○. A straight line 950 indicates a straight line ax + by + c = 0, and data on the upper side of this straight line is determined to be required for retesting, and data on the lower side is determined to be unnecessary for retesting. Therefore, a data group 920 indicated by ×
Is recorded in the re-examination necessity determination result 114 as “re-examination required”, and for the data group 940 indicated by ○, “re-examination required” is recorded in the re-examination necessity determination result 114.

In the present invention, for the data group 920 indicated by X, each sample in the data group 920 is re-examined. In general, a specimen becomes inappropriate as a specimen after a long period of time. Therefore, it is assumed that the retest is performed as soon as the first test result 111 is obtained. Of course, the technician may review whether or not re-examination is actually necessary and omit the one that is determined to be unnecessary.

FIG. 5B shows the result of the retest. FIG.
The data group indicated by x in the bold line in (b) indicates a data group in which the result of the re-examination did not significantly differ from the initial inspection result. In other words, even if the retest is performed, the same result is obtained, that is, the retest is useless. These data are recorded again in the retest necessity determination result 114 as not requiring retest. On the other hand, the samples in the data group indicated by x surrounded by a dashed line with the original thickness indicate that the test value has changed to the position △ indicating the correspondence by the broken line with an arrow as a result of the retest. In other words, the initial test of these specimens is unreliable due to some abnormality, and the data in the range of the test value which is likely to be obtained by the retest is evaluated. Therefore, it can be evaluated that the judgment that the retest is necessary is appropriate.

Therefore, in the re-test necessity determination logic optimizing process 117, of the data of the re-test necessity judgment result 114, the re-test necessity is determined from the distribution of the test result in which the re-test necessity is inappropriate and the distribution of the appropriate test result. Can be changed to a straight line 960 represented by (Equation 2). The contents of (Equation 2) are adopted to modify the judgment logic recorded in the retest necessity judgment logic 116.

A'x + b'y + c '= 0 (Equation 2) The reexamination according to the present invention and the evaluation of the results are exactly the same as the results obtained by the technician reviewing whether or not the original examination results actually require reexamination. However, in the present invention, it is possible to reduce the burden on the subject and the burden on the technician, and obtain a more reasonable reexamination necessity determination logic.

A major cause of the change in the evaluation of the test result of the specimen is mostly caused by the measuring instrument or the reagent. Of course, facilities that perform this type of testing routinely perform certain preliminary checks on measuring instruments or reagents, but in practice this type of useless retesting is being evaluated. Is a fact. Therefore, in FIGS. 5A and 5B, an example is described in which a more reasonable optimization of the reexamination necessity determination logic is searched for by reexamination of those determined to be reexamination.
Conversely, there is a possibility that the data group 940 indicated by た that is not required to be retested should be determined to require retesting.

Referring to FIGS. 6A and 6B, a description will be given of an example in which the determination that retest is necessary is changed from a data group for which retest is unnecessary. FIG. 6A is data showing the same initial inspection result as FIG. 5A. In this example, a straight line a in the data group 940 once determined to be unnecessary
For data existing near x + by + c = 0,
Retest. First, all specimens in the data group indicated by x are re-examined because it is determined that re-examination is required. In this case, too, the sample becomes inappropriate as a sample after a long period of time, so this retest is performed immediately after the first result is obtained. In the comparison between the result of the re-examination and the result of the first initial examination, most of the crosses indicate that the inspection value has changed and the
Suppose it has changed to below 50. That is, it is evaluated that the determination that reexamination is necessary was correct. Then, from this distribution, it can be said that it is suspicious whether the test value is also correct for the data once determined to be unnecessary for the re-test indicated by ○ in the previous judgment. Therefore, in the present invention, in the data group indicated by a thick circle in FIG.
For the samples in the data group close to the threshold indicated by ○ surrounded by a broken line, the determination that the retest is unnecessary is corrected and the retest is determined. That is, the necessity of reinspection is recorded in the reinspection necessity determination result 114 again.

More specifically, in the re-test necessity determination logic optimizing process 117, of the data of the re-test necessity judgment result 114, re-test is performed based on a change in a value obtained by comparing the re-test result with the first result. Was evaluated as appropriate, and the straight line 950 that was determined to be necessary for retesting from the distribution of the original test values evaluated as appropriate for retesting was corrected to a straight line 970 represented by (Equation 3). The necessity of re-examination is determined again by a straight line corrected to a predetermined range. The content of this (Equation 3) is adopted as modifying the judgment necessity logic recorded in the retest necessity judgment logic 116.

A''x + b''y + c '' = 0 (Equation 3) In this case as well, the reexamination according to the present invention and the evaluation of the results are based on the fact that the technician reconfirms whether or not the original examination results actually require reexamination. Although the result may not exactly match the result obtained by the above, in the present invention, the burden on the subject and the burden on the technician can be reduced, and a more reasonable retest necessity determination logic can be obtained. Although this example is not advantageous in terms of reducing the cost of reexamination, it can be said that reexamination evaluation without oversight can be realized.

When the majority of the evaluation results indicate that the retest was appropriate, instead of correcting the retest necessity determination logic and re-determining the past data as described above, the "re-test" is simply performed. A similar effect can be obtained by issuing a warning that the result of the re-examination is unnecessary.

FIG. 7 is a flowchart showing the optimization processing of FIG. 5 among the above-described processing. First, a test for all samples is performed for a predetermined item (step 201). This inspection result is input via the input / output interface 104 and stored in the initial inspection result 111 of the database 102 (step 202). Of course, some of these data are automatically input from the analyzer server 20 and some are input manually by a laboratory technician. Next, a retest necessity judgment reference value 118 for retest necessity judgment and a retest necessity judgment logic 116 are read into the RAM 103 (steps 203 and 204).
In accordance with the retest necessity judgment logic, judgment of necessity of retest is performed for each sample (step 205). Here, the retest necessity determination result 114
"Re-examination required" is recorded (step 206). On the other hand, for the sample determined to be unnecessary for retesting, “No retesting” is recorded in the retesting execution information 112 (step 20).
7). With respect to the sample determined to require retesting in step 205, a retest is performed (step 208).
“Retest execution” is recorded in the retest information 112 (step 209). Further, a retest result is input (step 2).
10). Based on the magnitude of the difference between the input retest result and the initial test result, it is determined whether or not the sample really needs retest (step 211). Here, for the sample determined to be unnecessary for retest, “retest unnecessary” is recorded in the retest necessity determination result 114 (step 213). Then
The re-examination necessity determination logic optimization processing 117 is executed with reference to the re-examination necessity determination result 114 and the record of the re-examination execution information 112 (step 214). The reexamination necessity logic is corrected according to the result (step 215).

FIG. 8 is a flowchart showing the optimization processing of FIG. 6 among the above-described processings. In this case as well, the processing up to step 205 for executing the determination of the necessity of retest for each sample according to the retest necessity determination logic is the same. Here, for a sample determined to require retesting,
“Retest required” is recorded in the retest necessity determination result 114 (step 206), and a retest is performed (step 208). On the other hand, for samples determined not to require retesting here, evaluate how far the data resulting in the determination that retesting is not required are separated from the identification boundary that separates samples requiring retesting from unnecessary samples. (Step 220). For example (Equation 1)
Is used to evaluate whether or not t ≧ −0.1,
If t ≧ −0.1, it is handled in the same manner as the determination of “retest required” in the process of step 205 for executing the determination of the need for retest described with reference to FIG. Sufficient from the discrimination boundary, for example t is-
If the distance is less than 0.1, the processing when it is determined that "retest is unnecessary" in the processing of step 205 described in FIG. 7 is performed.

For the sample determined as t ≧ −0.1 in step 220, “retest required” is recorded in the retest necessity determination result 114 (step 206). On the other hand, here, t <
For the sample determined to be -0.1, retest execution information 1
12 is recorded as "no retest" (step 207).
With respect to the sample determined as t ≧ −0.1 in step 220, a retest is performed (step 208), and
“Retest execution” is recorded in the retest information 112 (step 209). Further, a retest result is input (step 2).
10). Based on the magnitude of the difference between the input retest result and the initial test result, it is determined whether retest is really necessary (step 211). Here, for the sample determined to be unnecessary for retest, “retest unnecessary” is recorded in the retest necessity determination result 114 (step 213). Next, the re-examination necessity determination logic optimization processing 117 is executed with reference to the re-examination necessity determination result 114 and the record of the re-examination execution information 112 (step 214). The reexamination necessity determination logic is modified according to the result. (Step 215).

In the embodiment described with reference to FIGS. 6 and 8, the number of retesting targets is increased as compared with the embodiment described with reference to FIGS. Although this is disadvantageous, it is advantageous from the viewpoint of preventing the retest target from being overlooked by mistake.

Although an example using a linear discriminant has been described here, other retest determination methods are also applicable.
For example, in retesting using a neural network,
A similar optimization can be performed.

Using a network having a plurality of input elements for inputting test results, diagnostic information and the like and one output element for judging the necessity of a retest, the test results requiring a retest,
The learning is performed in advance so that the output s becomes 1 when the diagnostic information is input, and becomes 0 when the inspection result of unnecessary data is input. Then, when unknown data is input, it is determined that reexamination is necessary when output s ≧ v, and that reexamination is unnecessary when s <v. v is a dark value of the judgment, and 0.5 is set as an initial value.

In the case corresponding to FIG. 5 and FIG. 7, optimization is performed by learning the neural network using the data of the need for reexamination and the necessity of reexamination recorded in the reexamination necessity determination result 114 in step 214 as teacher data. be able to. In the case corresponding to FIGS. 6 and 8, in step 220, instead of the condition of t ≧ −0.1, for example, s ≧ 0.4. As described above, by detecting the inspection value when the neural network outputs a value slightly larger than the initial value v of 0.5 (here, 0.4), the inspection near the discrimination boundary is performed. The value can be determined to be a retest.

5 and 6 when the linear discriminant is used.
As shown in (1), the identification boundary of the necessity of reexamination on a plane with x and y as axes is a straight line. In the neural network, since the discrimination boundary is a curve, a more accurate retest determination can be performed.

At present, computerization of information on diagnosis handled in hospitals is progressing, and such computerized information is being accumulated in various hospital information systems. Although not described, various information can be referred to for optimizing the retest necessity determination logic. For example, the diagnosis information 121 records a diagnosis result by a doctor, information on medication, and the like, and it is possible to determine whether a retest is necessary by referring to these. For example, even if a certain patient's blood glucose level is higher than normal, if the patient has already been diagnosed as having diabetes,
It can be determined that the possibility of a measurement error is low.
In addition, when a diagnosis of hyperlipidemia is made, the tendency of a specific test value according to the type of the disease is shown, such as a high level of cholesterol and neutral fat. Therefore, by using the information regarding the diagnosis, it is possible to appropriately determine the possibility of the measurement error. Furthermore, although the disease name is described as an example of the information regarding the diagnosis, information other than the disease name may be used. For example, when a certain test value is known to show a high value or a low value when taking a specific medicine, by using the medication information as information regarding diagnosis, it is possible to determine whether or not a retest is necessary. Accuracy can be increased. Alternatively, the information may be information on the stage of medical treatment, such as before the disease name is determined, whether the disease name has been determined and the patient is undergoing treatment, and how long has passed since the operation. Laboratory values may fluctuate significantly before the disease name is determined or immediately after surgery, but after the disease name is determined and treatment is started, or after a long time has passed since surgery, Is stable, the fluctuation of the inspection value becomes small.
By using such information, it is possible to increase the accuracy of the retest necessity determination using the change from the previous value.

In FIG. 5 and FIG. 6, for the sake of simplicity, the necessity of re-examination is described as an example of linear evaluation between two items. Various techniques can be extended and used. Furthermore, reference values or logical expressions for judgment were set for each gender and age, but when using disease names as information related to diagnosis as in the above example, additional Further, a reference value and a logical expression may be set for each disease name. In this case, the age, sex, and disease name of the examinee are read from the diagnosis information 121 as information related to the diagnosis, and the corresponding reference value and logical formula are re-examined.
And from the retest necessity determination reference value 118, it is possible to determine the necessity of the retest. In this case, the re-examination necessity determination logic 116 or the re-examination necessity determination reference value 118 is, for example, in a format as shown in FIG.
Record reference values and logical expressions for each disease name. As described above, in the reexamination necessity determination logic 116 and the reexamination necessity determination reference value 118, parameters such as a reference value used for reexamination necessity determination and a logical expression describing the determination logic are recorded for each disease. Can improve the accuracy of the necessity of reexamination.

If the patient's disease name is not available at the initial stage of medical treatment or the like in which the doctor's diagnosis has not been determined, as shown in FIG. The retest logic is described, and this retest reference value,
The retesting necessity determination process may be performed using the retesting logic. In addition, in addition to the above methods, retest reference values for all diseases,
The determination using the retest logic may be performed, and the retest reference value for at least one kind of disease, and when it is determined that the retest is unnecessary by the determination using the retest logic, it may be determined that the retest is unnecessary.
In this case, the patient is more likely to be a disease corresponding to the retest reference value when the retest is unnecessary, the retest logic, and as a result, the retest reference value for this disease, the retest logic is determined to be unnecessary. You can think.

When the disease name is not available, FIG.
After estimating the patient's disease name from the input test results and the like using the disease candidate estimation processing 124 shown in (1), re-test necessity determination is performed using a re-test reference value corresponding to the estimated disease name and re-test determination logic. Processing may be performed. In this case, if no disease candidate is output, the input test value is considered to be incorrect, and it can be determined that the sample requires retesting.

When using medication information, reference values and logical expressions are recorded for each type of drug, and different reference values and logical expressions are selected according to the type of drug administered to the patient. By using this, it is possible to improve the accuracy of the necessity of retesting. In addition, different reference values and logical expressions are recorded according to the number of days elapsed since the operation, and the necessity of reexamination is determined by selecting and using different reference values and logical expressions according to the number of days elapsed since the patient's operation. Can be improved in accuracy.

As the re-test necessity determination logic, a neural network may be used in which the present test value, past test values, diagnostic information, etc. are input and the necessity of re-test is output. In this case, the re-examination necessity determination logic 116 records parameters describing the network structure such as the number of layers of the neural network, the number of elements in each layer, and the connection weight between the elements, and the like. When making a negative determination, a network is configured based on the recorded information.

When fuzzy inference is used as the reexamination necessity determination logic, the reexamination necessity determination logic 116 records a membership function used for fuzzy inference.

Determination based on reexamination necessity determination logic (step 2
The result of 05) is recorded in the re-examination necessity determination result 114, and it is preferable that information serving as the basis of the re-examination necessity determination result is also recorded here. The basis of the retest necessity judgment result explains why it was judged necessary if retest was necessary, and why it was judged unnecessary if retest was unnecessary. This is the information that provides the basis for doing so. The reasons for determining that retesting is necessary include, for example, inspection items and their values that exceeded the reference value in a single item check, inspection items and their values that exceeded the reference value in the previous value check, and abnormalities in the inter-item check , A test value used for the logical operation, information on diagnosis such as a disease name, and the like. In addition, the grounds when it is determined that the retest is unnecessary are, for example, information that there is no abnormal value when there is no abnormal value (inspection value exceeding the reference value) in the single item check. In addition, if there is no abnormal value and the retest is unnecessary, the previous value of the test item indicating the abnormal value or information on diagnosis such as a disease name can be used as the basis for the determination.

FIGS. 10 and 11 show screen configuration examples for displaying the test result and the retest necessity judgment result on the display of the test client 40. In FIG. 10, on the screen 800 of the inspection client 40, the reception date and time 810, the inspection order number 815, the patient name 820, the patient's age 825, the patient's gender 830, the patient's consultation department 835, and the inspection status 84 are displayed.
An example in which 0, presence / absence of abnormal value 845, necessity of retest 850, and the like are displayed in a list format. In the test status 840, the progress of the test is shown for each type of test such as a biochemical immunoassay (described as "biochemical immunity" in the figure) and a general test (described as "general" in the figure). I have. In the figure, “complete” indicates that the inspection has been completed and the inspection result has been reported, and “not yet” indicates that the inspection has not been completed yet. "-" Indicates that no order has been placed for that type of inspection. The presence / absence 845 of an abnormal value indicates that there is no abnormal value when “absent”, and indicates that there is an abnormal value when “present”.
If no test result has been output yet, "-" is displayed. The retest necessity 850 includes the retest necessity determination processing 113.
The necessity of a retest determined by is displayed, and in the case of "No", it indicates that there is no need for a retest,
"Necessary" indicates that reexamination is necessary.
When no test result is output, “−” is displayed.

Here, the above information is listed in the order of the order number, but it is also possible to extract the information according to conditions. For example, by setting conditions for specific medical departments, patient names, presence / absence of abnormal values, necessity of reexamination, and extracting only rows satisfying the conditions, the target information can be efficiently referred to. .

In the example of the screen configuration shown in FIG. 10, when one of the lines is selected by operating a mouse, a keyboard or the like attached to the inspection client 40, more detailed information on the selected line is displayed. FIG. 11 shows a configuration example of a screen displayed at this time.

FIG. 11 shows an example of a screen configuration on the display 800 of the inspection client 40. The screen includes a patient basic information display area 860 and an examination result display area 86
5, a retest necessity determination result display area 870 and a retest execution information input area 875.

[0052] In the patient basic information display area 860, information about the patient, such as the patient ID, name, age, and gender, is displayed. In the inspection result display area 865, a current value, a previous value, and a previous value for each inspection item such as TP and Alb are displayed in a table format together with the inspection date. In the re-examination necessity determination result display area 870, a result of the necessity of re-examination by the re-examination necessity determination processing 113 and information serving as a basis for the determination are displayed. In the reinspection execution information input area 875, the operator of the present system uses the mouse, keyboard, or the like attached to the inspection client 40 to select whether or not to perform the reinspection (whether or not to perform the reinspection). Can be entered.

In some cases, a laboratory technician refers to the test result of a sample that the system determines that a retest is necessary, and determines whether or not a retest is really necessary.
If the necessary reexamination is automatically performed according to the reexamination necessity determination result as in the present invention, the judgment by the laboratory technician can be omitted in many cases. Also in the case where the judgment by the laboratory technician is made, the re-examination necessity judgment result 114 for recording the basis of the re-examination necessity judgment is provided, the inspection result is displayed in the inspection result display area 865, and the re-examination necessity judgment result display area 8 is displayed.
By displaying the re-examination necessity determination result and the basis of the re-examination necessity determination in 70, when the technician makes a final decision, it is possible to indicate which test value to pay attention to as the basis of the re-examination necessity determination. This allows the technician to make quick decisions.

The reexamination information input in the reexamination information input area 875 is recorded in the reexamination information 112. The re-examination necessity determination logic optimization processing 117 uses the inspection result recorded in the initial inspection result 111, the re-examination execution information recorded in the re-examination execution information 112, the re-examination result 115, and the re-examination necessity determination result 114, and The information for optimizing the retest necessity determination logic 116 used in the rejection determination processing 113 is output.

For example, when optimizing a reference value used in a single item check, a previous value check, an item-to-item check, etc. according to the prior art, a histogram of the inspection value distribution is calculated using past inspection results recorded in the inspection result recording means. An optimization method is well known in which a reference value is created and a range including 95% of the entire value around the average value is used as a reference range. As another method, there is a method in which the average value μ and the standard deviation σ of the inspection values are obtained, and the range (μ−kσ, μ + kσ) is used as a reference range.
Here, k is an appropriate constant. However, such a method of setting the reference range does not use information as to what kind of test value the specimen that needs to be retested is. Therefore, when the retest is required using the reference range actually set by this method, it is necessary to retest the samples that have been retested in the past or to retest those that do not need to be retested. In many cases. Therefore, at the time of setting the reference range, in addition to the past test results, by using a retest necessity determination result indicating whether the test result is a result to be retested,
It is possible to set the maximum reference range in which all the test results that really need to be re-tested among the past test results are judged to be necessary to be re-tested. By setting the reference range in this manner, oversight of test results to be retested is reduced, and cases in which test results that do not actually need to be retested are judged to be retesting are reduced. Can be improved in accuracy.

In the re-examination necessity determination processing 113, when other methods are used, various optimization methods and learning methods corresponding to the respective methods may be used. For example, when a neural network is used as the retest necessity determination process 113, the past test value read from the test result 111 is used as learning data, and whether or not the sample to be retested read from the retest necessity determination result 114 is determined. Can be used as teacher data to learn a neural network. When, for example, a feed-forward type network is used as the neural network, a back propagation method can be used as a learning method.

The determination of the necessity of reinspection can be considered as a pattern recognition problem in which a pattern having an inspection value as a feature amount is classified into two types, a pattern that requires reinspection and a pattern that does not require reinspection. Therefore, as the re-examination necessity determination method of the re-examination necessity determination logic 116, various known pattern recognition methods can be used. As the re-examination necessity determination logic optimization processing 117, various known pattern recognition logics can be used. Construction methods can be used. Therefore, for example, even when fuzzy inference is used in the reexamination necessity determination logic 116, the inference method can be optimized by using the inspection result and the corresponding reexamination necessity determination result. In addition, by using a known pattern recognition logic construction method, it is possible to obtain an optimal logic expression by learning.

As described above, by providing the initial inspection result 111, the re-inspection result 115, and the re-inspection execution information 112, the above-described re-inspection necessity determination logic can be optimized, and the re-inspection necessity determination accuracy is improved. be able to.

Although the optimization of the retest logic has been described as being always executed in the flow of the retest in FIGS. 7 and 8, it is executed when the user of the present system inputs the retest necessity determination logic change. It is good to do.

Also, it goes without saying that the system operator can change the retest necessity logic regardless of the retest result. For example, the retest reference value or the retest logic of FIG. 9 is updated using the input device.

[0061]

According to the present invention, the re-test necessity determination logic is optimized by using the re-test result and the test result accumulated for the re-test result for the sample judged to be re-test required or the re-test result for a part of the sample judged to be unnecessary. By doing so, the accuracy of the retest can be increased.

[Brief description of the drawings]

FIG. 1 is a diagram showing one configuration example of a clinical test system to which the present invention can be applied.

FIG. 2 is a diagram showing another example of the configuration of the clinical test system to which the present invention can be applied.

FIG. 3 is a diagram showing still another example of the configuration of the clinical test system to which the present invention can be applied.

FIG. 4 is a diagram showing a basic block configuration of an inspection server according to the present invention.

FIG. 5 is a view for explaining one embodiment of the principle of optimization of the retest necessity determination logic according to the present invention.

FIG. 6 is a diagram illustrating another embodiment of the principle of optimization of the retest necessity determination logic according to the present invention.

FIG. 7 is a view showing a flowchart of a process of optimizing the retest necessity determination logic according to FIG. 5;

FIG. 8 is a view showing a flowchart of processing for optimizing the retest necessity determination logic according to FIG. 6;

FIG. 9 is a diagram illustrating an example of a parameter such as a reference value used for retest necessity determination in the retest necessity determination logic according to the present invention, and a logical expression describing the determination logic.

FIG. 10 is a view showing one example of a screen configuration for displaying an inspection result and a re-inspection necessity determination result on a display of the inspection client according to the present invention.

FIG. 11 is a view showing another example of the screen configuration for displaying the test result and the retest necessity determination result on the display of the test client according to the present invention.

[Explanation of symbols]

11: first inspection device, 12: second inspection device, 13:
Third inspection apparatus, 20: analysis apparatus server, 30: inspection server, 40: inspection client, 50: LAN, 60:
Learning server, 70: test result reference terminal, 100: system bus, 101: CPU, 102: database, 10
3: RAM, 104: input / output interface, 111:
Initial inspection result, 112: Retest execution information, 113: Retest necessity determination processing, 114: Retest necessity result data, 115: Retest result data, 116: Retest necessity determination logic, 117: Retest Necessity judgment logic optimization processing, 1
21: diagnosis information, 122: contract information, 123: request information, 124: disease candidate estimation processing, 500: hospital, 51
0: Inspection center, 520: First hospital, 530: Second
540: communication line, 550; judgment logic learning facility, 610: first communication means, 620: second communication means, 800: display screen of the examination client 40, 810: column displaying reception date and time, 815 : Column for displaying the order number, 820: column for displaying the patient's name, 82
5: column displaying age, 830: column displaying gender, 8
35: a column for displaying a medical department, 840: a column for displaying an examination status, 845: a column for displaying the presence or absence of an abnormal value, 850: a column for displaying the necessity of reexamination, 860: a patient basic information display area, 865: Inspection result display area, 870: Retest necessity judgment result display area, 875: Retest execution information input area,
920, 940: data group, 950: ax + by + c =
Straight line satisfying 0, 960: a'x + b'y + c '= 0
970: a straight line that satisfies a ″ x + b ″ y + c ″ = 0.

 ──────────────────────────────────────────────────の Continuing on the front page (72) Inventor Hitoshi Matsuo 1-280 Higashi Koigakubo, Kokubunji-shi, Tokyo Inside the Central Research Laboratory of Hitachi, Ltd. F-term in Hitachi Central Research Laboratory (reference) 2G058 AA08 GD05 GD07

Claims (6)

[Claims] 1. A clinical test system having a function of judging the necessity of a retest based on a retest necessity judgment logic having predetermined parameters for test results of a plurality of test subjects.
For the subject of the subject re-examined by the re-examination necessity determination logic, a re-examination is performed, and the re-examination execution information indicating whether the re-examination has been performed, the first inspection result and the re-examination result are referred to, A clinical test system, wherein a parameter of a retest necessity determination logic is corrected. 2. A clinical test system having a function of judging the necessity of a retest based on retest necessity judgment logic having predetermined parameters for test results of a plurality of test subjects.
Means for individually storing the test results, and a comparison between the results of the respective retests for the plurality of samples determined to be retested by the retest necessity determination logic and the corresponding stored first test results, respectively. According to the means for determining whether the re-examination is appropriate and the distribution of the test results determined to be appropriate for the re-examination, the re-examination necessity determination theory has determined that re-examination is unnecessary. A clinical test system having a need for retesting a sample and a means for instructing correction. 3. A means for instructing the correction by correcting a parameter of the retest necessity determination logic and re-determining the stored first inspection result by the corrected retest necessity determination logic. The clinical test system according to claim 2, wherein 4. A clinical test system having a function of determining the necessity of a retest based on predetermined retest necessity determination logic with respect to test results of a plurality of test samples, and A clinical test system using a different retest necessity determination logic. 5. A clinical test system having a function of determining the necessity of a retest based on predetermined retest necessity determination logic with respect to test results of a plurality of test samples of a subject. A clinical test system, wherein a different retest necessity logic is used depending on the type of the retest. 6. A clinical test system having a function of determining the necessity of a retest based on predetermined retest necessity determination logic for test results of a plurality of test subjects, A clinical test system comprising a disease candidate extracting means for outputting a suspected disease name, and using a different reexamination necessity determination logic according to a type of a disease outputted by the disease candidate extracting means.