JP2007026172A - Subject selecting device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a subject selecting device capable of efficiently performing a clinical trial. <P>SOLUTION: A risk function deriving part 47 analyzes survival time on the basis of the epidemiologic data of a patient and derives a risk function for obtaining a risk value which is the probability that the subject of the clinical trial drops out within a clinical trial execution period due to early withdrawal or stop. A risk value calculation part 48 applies the risk function to clinical trial eligible person information in a clinical trial eligible person information data base 22 so as to obtain a risk value for each clinical trial eligible person. A subject selecting part 49 selects the clinical trial eligible persons for the scheduled number of cases as the subjects in the ascending order of the risk values. Thus, the increase of a clinical trial cost and the extension of the clinical trial execution period due to the drop-out of the subject are prevented and the clinical trial is efficiently performed. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a clinical trial support system, and more particularly to a subject selection device for selecting subjects in a clinical trial.

  When a new drug is developed, it is necessary to obtain a national approval after reviewing the effectiveness and safety of the drug based on the Pharmaceutical Affairs Law. Therefore, in the final stage of drug development, “clinical trials”, which are clinical trials aimed at obtaining drug approval, are performed.

  As an implementation standard for this clinical trial, “Ministerial Ordinance on Standards for Implementation of Pharmaceutical Clinical Trials” (GCP (Good Clinical Practice)) is established based on the Pharmaceutical Affairs Law. Under this GCP, the requester of a clinical trial (pharmaceutical company, etc.), when the investigational drug is decided, is a protocol (plan) that describes the purpose, design, method, statistical consideration, organization / responsibility system, etc. of the clinical trial. Create a clinical trial plan. The clinical trial is conducted according to this protocol (hereinafter referred to as clinical trial protocol).

In a clinical trial, it is necessary to select a subject who is a person who will receive a clinical trial (study subject), and the selection of the subject is performed according to the subject selection criteria described in the clinical trial protocol (for example, Patent Document 1). reference).
In addition to technologies related to clinical trial support such as clinical trials, in the medical field, there is a technology for analyzing the survival time of patients (see, for example, Patent Document 2).
JP-A-11-154185 (page 3, FIG. 7) JP 2000-268116 A (page 3)

As described above, the selection of subjects is performed in accordance with a clinical trial protocol that has been prepared in advance for conducting the clinical trial. However, when the clinical trial is actually started, the clinical trial is caused by worsening clinical symptoms or occurrence of complications in the subject. There are many early withdrawals and cancellations that make it impossible to continue.
When the subject is withdrawn or canceled early, a trial termination report is created, and the entry of a new subject who becomes a new subject is awaited.

Therefore, even if a subject is selected from candidates according to the subject selection criteria of the clinical trial protocol, if there is an early withdrawal or cancellation, all the costs and time that have been invested in the subject until then are wasted. Furthermore, since it becomes necessary to conduct a trial for a new subject, the trial period has been extended and the cost has been increased.
As described above, there is a disadvantage that the efficiency of the clinical trial is poor only by selecting the subject according to the subject selection criteria of the clinical trial protocol.
In addition, the survival analysis technique has no direct relationship with the clinical trial support technology and has not been used to support the clinical trial.

  This invention is made | formed in view of the said situation, and it aims at providing the test subject selection apparatus which can perform a clinical test efficiently.

In order to achieve the above object, a subject selection device according to the first aspect of the present invention includes:
A test subject selection device for selecting test subjects from candidates,
A risk function storage means for storing a risk function for obtaining a risk value, which is a probability that the clinical trial is suspended due to a change in the candidate's symptom during the clinical trial, for each candidate;
For each candidate, candidate information storage means for storing candidate information including at least the identification information of the candidate and an observed value of a predetermined factor corresponding to the covariate of the risk function;
An implementation period information acquisition means for acquiring the implementation period information of the clinical trial;
A risk function is read from the risk function storage means, candidate information is read for each candidate from the candidate information storage means, and an independent variable of the risk function includes a predetermined factor corresponding to the covariate of the candidate information. A risk value calculating means for substituting the observation value and the execution period information acquired by the execution period information acquisition means, and obtaining the risk value that is a dependent variable of the risk function for each candidate;
Planned case number storage means for storing the planned number N of clinical trials;
The planned number of cases N is read from the planned number of cases storage means, and N candidates are selected from the candidates in order of increasing risk values obtained by the risk value calculation means, and the selected candidates are selected. Identification information output means for outputting identification information;
It is characterized by providing.

Epidemiological data acquisition means for acquiring epidemiological data collecting patient information in past affected cases;
A risk function deriving unit for performing statistical analysis based on the epidemiological data acquired by the epidemiological data acquiring unit and deriving the risk function;
You may make it provide.

The risk function deriving means includes:
Means for performing survival time analysis based on the epidemiological data acquired by the epidemiological data acquiring means, and deriving a hazard function h (z _i , t) that is a function of the covariate vector z _i of the patient _i and the time t;
means for deriving a function p (i, t) as the risk function by calculating p (i, t) = 1−h (z _i , t);
You may make it have.

Further, the subject selection device according to the second aspect of the present invention is:
A test subject selection device for selecting test subjects from candidates,
A risk function storage means for storing a risk function for obtaining a risk value, which is a probability that the clinical trial is suspended due to a change in the candidate's symptom during the clinical trial, for each candidate;
For each candidate, candidate information storage means for storing candidate information including at least the identification information of the candidate and an observed value of a predetermined factor corresponding to the covariate of the risk function;
An implementation period information acquisition means for acquiring the implementation period information of the clinical trial;
A risk function is read from the risk function storage means, candidate information is read for each candidate from the candidate information storage means, and an independent variable of the risk function includes a predetermined factor corresponding to the covariate of the candidate information. A risk value calculating means for substituting the observation value and the execution period information acquired by the execution period information acquisition means, and obtaining the risk value that is a dependent variable of the risk function for each candidate;
Risk value storage means for storing the risk value for each candidate obtained by the risk value calculation means in association with identification information of the candidate;
Planned case number storage means for storing the planned number N of clinical trials;
The planned case number N is read from the planned case number storage means, all combinations for extracting N persons from the candidates are obtained, and the identification information of the extracted candidate who is the extracted candidate is obtained for each combination. Candidate extraction means to obtain;
The risk value stored in association with the identification information of the extraction candidate acquired by the candidate extraction unit is read from the risk value storage unit, and the clinical test is performed for each combination based on the read risk value. Expense expectation value calculation means for obtaining an expense expectation value that is an expectation value of the cost of implementation of
Identification information output means for outputting identification information of extraction candidates in the smallest combination of expected expenses calculated by the expected expenses calculation means;
It is characterized by providing.

の演算を行うことにより求める手段と、を有するようにしてもよい。 The candidate information stored in the candidate information storage means includes, for each candidate, facility identification information for identifying a facility where a clinical trial of the candidate is performed,
The expected cost calculation means is:
Damage amount storage means for storing the amount of damage A that occurs when the clinical trial of one subject is stopped;
Fixed cost storage means for storing a fixed cost B, which is a cost that is uniformly generated when requesting a clinical trial from the facility, for each facility specified by the facility identification information;
Unit-contracted cost storage means for storing unit-consigned cost C, which is a cost incurred when requesting one facility to perform a clinical trial for each subject, for each facility specified by the facility identification information;
The facility identification information stored in association with the identification information of the extraction candidate i acquired by the candidate extraction means is read from the candidate information storage means, and the extraction candidate i is read based on the read facility identification information. A facility identification means for identifying a facility j used when a clinical trial is performed in combination of
Wherein each facility identification means has identified facility j, and accession number of cases counting means for counting the accession number of cases M _j is performed the number of facilities j when clinical trials carried out in a combination of the extraction candidate i ,
Read the risk value p (i) stored in association with the identification information of the extraction candidate i from the risk value storage means, read the damage amount A from the damage amount storage means, from the fixed cost storage means, The fixed value B (j) of the facility j specified by the facility specifying means is read, the unit trust cost C (j) of the facility j specified by the facility specifying means is read from the unit trust cost storage means, and the trust case The number M _j of entrusted cases of the facility j specified by the facility specifying means is acquired from the number counting means, and the expected cost value S is obtained for each combination.

And a means for obtaining by performing the above calculation.

  According to the present invention, clinical trials can be performed efficiently.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
First, a first embodiment of the present invention will be described.
FIG. 1 is a block diagram showing a configuration example of a clinical trial support system according to the first embodiment of the present invention.
As shown in the figure, the clinical trial support system according to the present embodiment includes a hospital information processing apparatus 1 provided in each of a plurality of hospitals that are clinical trial facilities, and a subject selection provided in a center that selects subjects for clinical trial implementation. The apparatus 2 and an epidemiological data storage device 3 for storing epidemiological data which is patient information in past disease cases collected in epidemiological studies, and a plurality of hospital information processing devices 1 and subject selection devices 2 are provided. A subject selection device 2 and an epidemiological data storage device 3 are connected via a communication network 4 via a communication network 4.

The hospital information processing apparatus 1 provided in each hospital stores a patient information database that stores patient information of patients who are in the hospital. Then, the hospital information processing apparatus 1 anonymizes the patient information stored in the patient information database and transmits it to the center via the communication network 4.
Anonymization of patient information is performed, for example, by deleting the patient's name, address, date of birth, etc., from the personal identification information in the patient information, and patient identification by hospital that is patient identification information given to each hospital Is done by leaving As a result, the center that receives the patient information cannot specify the patient's name or the like from the patient information.

  The subject selection device 2 provided in the center stores a plurality of anonymized patient information received from each hospital in the anonymized patient information database 21. Then, the subject selection device 2 extracts patients that match the subject selection criteria described in the trial protocol from the patients whose information is stored in the anonymized patient information database 21. Furthermore, the subject selection apparatus 2 sets the extracted patient as a trial eligible person who is a candidate for the subject, and records the patient information in the trial eligible person information database 22 as trial eligible person information.

In addition, the subject selection device 2 acquires epidemiological data from the epidemiological database provided in the epidemiological data storage device 3 via the communication network 5, performs survival time analysis that is a statistical analysis method based on the acquired epidemiological data, and performs a risk function Ask for.
Here, the risk function is a function for obtaining the probability that the subject will drop out within the trial implementation period due to early withdrawal or cancellation, and details thereof will be described later.

  The subject selection device 2 applies the risk function obtained by the survival time analysis to the clinical trial person information recorded in the clinical trial person information database 22 and obtains a dropout rate for each trial person. Then, in the descending order of the dropout rate, the same number of eligible patients as the number of planned cases described in the trial protocol is selected as subjects, and subject information that is specific information of the selected subjects is output.

Next, the subject selection apparatus 2 provided in the clinical trial support system according to the present embodiment will be specifically described.
FIG. 2 is a block diagram illustrating a configuration example of the subject selection device 2.
The subject selection device 2 includes a communication unit 31, an input unit 32, an output unit 33, a storage unit 34, and a control unit 35, and each unit is connected to each other via a bus 36 as illustrated. Yes.

  The communication unit 31 communicates with the hospital information processing apparatus 1 and the epidemiological data storage apparatus 3 via the communication networks 4 and 5 and includes a communication interface and the like. For example, the communication unit 31 receives the anonymized patient information from the hospital information processing apparatus 1 and the epidemiological data from the epidemiological data storage device 3.

The input unit 32 is used by a person who operates the subject selection device 2 to input various information, and includes input devices such as a keyboard and a mouse. For example, information or the like described in the clinical trial protocol is input to the input unit 22.
The output unit 33 outputs various information and includes a display device such as a display. For example, the output unit 33 displays and outputs the subject information of the selected subject.

  The storage unit 34 stores various information, programs, and the like, and includes an auxiliary storage device such as a hard disk. Further, the storage unit 34 includes an anonymized patient information database 21 and a clinical trial person information database 22. Furthermore, the memory | storage part 34 memorize | stores the description information etc. of a trial protocol.

The control unit 35 performs calculation processing of data and controls the communication unit 31, the input unit 32, the output unit 33, and the storage unit 34 via the bus 36, and includes a CPU (Central Processing Unit) 351, a ROM ( A read only memory (352), a random access memory (RAM) 353, and the like. Specifically, the CPU 351 executes the control program stored in the ROM 352 while temporarily storing various data using the RAM 353 as a work area. Done.
For example, the subject selection process or the like is performed in the subject selection device 2 by the control unit 35 controlling the above-described units according to the control program stored in the ROM 352 or the storage unit 34.

Next, a subject selection operation in the subject selection device 2 will be described.
FIG. 3 is a block diagram illustrating a functional configuration example of the subject selection device 2.
In addition, as for each component of the test subject selection apparatus 2 shown in FIG. 3, the control part 35 shown in FIG. 2 controls the communication part 31, the input part 32, the output part 33, or the memory | storage part 34 similarly shown in FIG. Can be realized.
In addition, portions having the same functions as those already described are denoted by the same reference numerals and description thereof is omitted.

The anonymized patient information receiving unit 41 receives anonymized patient information from each hospital and accumulates the received plurality of patient information in the anonymized patient information database 21.
FIG. 4 is a diagram illustrating a data configuration example of patient information stored in the anonymized patient information database 21.
As shown in the figure, the patient information includes information such as individual specifying information for each patient, life history, past history, test values, and the like. Each information has further subdivided information. The patient information stored in the anonymized patient information database 21 is anonymized patient information stored in the patient information database of the hospital information processing apparatus 1. Information such as date of birth, occupation, place of birth, etc. has been deleted.
The anonymized patient information receiving unit 41 stores the patient information in the anonymized patient information database 21. The center ID identifies the patient ID and the hospital ID for identifying the transmission source hospital. Are added to the personal identification information.

  Returning to FIG. 3, the clinical trial protocol acquisition unit 42 acquires and stores the clinical trial protocol created by the clinical trial client such as a pharmaceutical company. Specifically, for example, the clinical trial protocol acquisition unit 42 acquires information described in the clinical trial protocol from the input unit 32 of FIG. 2, or data on the clinical trial protocol created by another information terminal. The clinical trial protocol is acquired and stored through the communication unit 31.

  The clinical trial qualified person extraction unit 43 searches the anonymized patient information database 21 according to the subject selection criteria described in the clinical trial protocol acquired by the clinical trial protocol acquisition unit 42, and patients who match the clinical trial protocol are candidates for the clinical trial. Extract as Specifically, for example, the trial-qualified person extraction unit 43 reads the subject selection criteria described in the trial protocol from the trial protocol acquisition unit 42, and the target disease name, age, sex, and the like indicated in the subject selection criteria are read. The selection conditions are compared with the patient information stored in the anonymized patient information database 21, and patients that match the subject selection criteria are extracted from the anonymized patient information database 21. Then, the extracted patient information of the patient is recorded in the clinical trial person information database 22 as clinical trial person information.

As described above, the trial-qualified person extraction unit 43 extracts patients who match the clinical trial protocol, so that the center that selects the subjects can grasp the presence of patients who can be the subject of the clinical trial.
Note that only the patient ID of the patient extracted by the clinical trial qualified person extraction unit 43 is recorded in the clinical trial qualified person information database 22, and other patient information of the extracted patient is anonymized patient information as necessary. You may make it acquire by referring the database 21. FIG.

The planned case number discriminating unit 44 discriminates whether or not the number of clinical trial qualified persons extracted by the clinical trial qualified person extracting unit 43 exceeds the planned number of cases described in the clinical trial protocol.
Specifically, for example, the planned case number determination unit 44 first reads the planned case number N described in the clinical trial protocol from the clinical trial protocol acquisition unit 42. Here, the planned number of cases is the number of subjects to be selected in the clinical trial.
Next, the planned case number discriminating unit 44 refers to the clinical trial qualified person information extracted by the clinical trial qualified person extracting unit 43 and stored in the clinical trial qualified person information database 22 and counts the number of clinical trial qualified persons. Then, the counted number of eligible clinical trials is compared with the planned number of cases N to determine whether the number of eligible clinical trials is greater than the planned number of cases N.

When the number of eligible clinical trials is less than or equal to the planned number of cases N, the planned number of cases discriminating unit 44 sets the clinical trial eligible person information database so that all clinically qualified persons extracted by the clinically qualified person extracting unit 43 are actual subjects. The patient IDs of all clinical trial qualified persons are read out from 22 and output to the selected subject display unit 45.
The selected subject display unit 45 displays the patient ID output from the planned case number discrimination unit 44 as information on the selected subject.

  On the other hand, when the planned case number discriminating unit 44 determines that the number of eligible clinical trials is greater than the planned number of cases N, the subject selection device 2 subsequently selects N as the number of planned cases from the eligible clinical trials. The operation of selecting a human subject is performed. Therefore, the planned case number determination unit 44 outputs a selection request signal to the risk value calculation unit 48 described later.

The subject selection device 2 according to the present embodiment selects subjects in consideration of the possibility of early withdrawal or cancellation after the start of the trial in order to efficiently conduct the trial. Therefore, the clinical symptoms worsening or the occurrence of complications, which are changes in symptoms subject to early withdrawal or discontinuation, are predicted before the start of the trial, and this prediction result is used as an evaluation value to determine the priority (priority) ).
Patients who are eligible for the trial are at risk of complications and severe transformation. Therefore, the subject selection apparatus 2 specifically simulates the probability of clinical symptoms worsening and complications within the trial period from the personal information of clinically qualified persons such as clinical data, and from the low value In order, it is determined that it is preferable as a trial subject, and priorities are given to those eligible for the trial. Then, subjects are selected based on this priority.

  Epidemiological data collected in epidemiological studies in the past are used to calculate the probability of clinical symptoms worsening and complications. The epidemiological data is stored in the epidemiological database of the epidemiological data storage device 3 of FIG. 1, and the personal data such as clinical information of patients is stored in the epidemiological data.

FIG. 5 is a diagram showing a data configuration example of epidemiological data stored in the epidemiological database.
Epidemiological data is created for each disease (affected disease name). In the epidemiological data, as variables shown in the figure, for example, as shown in the figure, for example, patient ID (sample ID), age, sex, affected disease name and onset date, diagnosis date and time (severity), Test items and results (test values), presence of complications, follow-up period, etc.

Hereinafter, the operation | movement which selects a test subject in the test subject selection apparatus 2 using epidemiological data is demonstrated.
Returning to FIG. 3, the epidemiological data receiving unit 46 receives epidemiological data from the epidemiological data storage device 3. Specifically, when the epidemiological data storage device 3 has, for example, an FTP (File Transfer Protocol) server function and FTP software is installed in the subject selection device 2, the epidemiological data receiving unit 46 is configured as shown in FIG. Epidemiological data can be downloaded as appropriate according to the operation to the input unit 32.

Based on the epidemiological data received by the epidemiological data receiving unit 46, the risk function deriving unit 47, as an evaluation function for selecting a subject, causes severe changes or complications in the trial period t in the patient i who can be the subject of the trial. The probability density function p (i, t) to be obtained is obtained.
In other words, the probability density function p (i, t) is a function for calculating the probability that the patient i will drop out within the trial period t, and is hereinafter referred to as a risk function. Then, by applying this risk function p (i, t) to patient information, a risk value p that is a probability (probability that the trial will be stopped) for each patient is dropped during the trial. Can be requested.
In addition, since real time processing is not required for the derivation calculation of the risk function p (i, t), a particularly fast solution method is not necessary, and a well-known survival time analysis method such as the Cox proportional hazard method can be used.

Here, a method for deriving the risk function p (i, t) by survival time analysis will be described.
The risk function p (i, t) can be calculated by using a statistical model (proportional hazard model) that is the basis of Cox regression, based on data obtained in a local epidemiological study (cohort study). it can.
In this regard, it is conventionally known to perform survival time analysis using the PHREG procedure of the SAS system (statistical analysis package).

When the covariate value of a solid i is represented by a vector z _i = (z _i1 , z _i2 ,..., Z _ip ), the probability function that the solid i will survive until a certain time t is Is represented by a hazard function h (z _i , t) such as
h (z _i , t)
= H ₀ (t) · exp (β ^T z _i )
= H ₀ (t) · exp (β ₁ z _i1 + β ₂ z _i2 +... + Β _p z _ip ) (1)

Here, the covariates are, for example, when performing survival time analysis using the patient epidemiological data shown in FIG. 5, age, sex, disease name and onset date, medical date and severity, examination items and examination Value etc.
Further, h ₀ (t) is a survival probability (baseline hazard) of the reference person at time t where the influence of all covariates is zero.

Then, using the epidemiological data stored in the epidemiological data storage device 3, performs proportional hazard analysis, beta ₁ in the above formula _{(1), β 2, ···} , by estimating the beta _p, hazard function Deriving h (z _i , t).
Next, regarding the relationship between the hazard function and the risk function, “survival” in the case of performing survival time analysis using the epidemiological data shown in FIG. 5 means that no complications occur at the same time. The risk function p (i, t), which is a probability function of clinical trial discontinuation due to clinical symptom deterioration or the like, is obtained as follows.
p (i, t) = 1−h (z _i , t) (2)

As described above, the risk function deriving unit 47 determines the covariate values associated with the patient i (age, sex, diseased disease name and date of onset, treatment date and severity, test items and test values, etc.) ) And time (trial period) t are independent variables, and a risk function p (i, t) having a risk value as a dependent variable is derived from epidemiological data.
Note that the risk function p (i, t) derivation in the risk function deriving unit 47 does not require real-time processing as described above, and may be performed and stored in advance at an appropriate time.

  The risk value calculation unit 48 operates the risk function p (i, t) for each clinical trial eligible person i when the planned clinical trial number judgment part 44 determines that the number of clinical trial qualified persons is larger than the planned case number N. The risk value p, which is the dropout rate during the trial period t, is calculated. By using this risk value p as an evaluation value, it is possible to determine the priority order of patients entering the actual clinical trial.

Specifically, when the risk value calculation unit 48 receives the selection request signal from the planned case number determination unit 44, the risk value calculation unit 48 acquires the risk function p (i, t) from the risk function derivation unit 47. Then, the trial period described in the trial protocol is read from the trial protocol acquisition unit 42 as a value to be substituted for the time t of the risk function.
Furthermore, the risk value calculation unit 48 reads the observed value of the factor corresponding to the covariate of the risk function p (i, t) from the clinical trial eligible person information database 22 for each clinical trial eligible person i. Specifically, for example, when the risk function is derived based on the epidemiological data illustrated in FIG. 5, the risk value calculation unit 48 determines the clinical trial eligible person from the patient information stored in the clinical trial eligible person information database 22. The observed values of factors such as age, sex, disease name and date of onset, date and time of medical treatment and severity, test items and test values are read out.
Then, the risk value calculation unit 48 substitutes the read clinical study period and the observed value of each factor into the risk function p (i, t), calculates the risk value p for each clinical trial qualified person, and The ID and the risk value p are associated and output to the subject selection unit 49.

  The subject selection unit 49 receives the risk values p calculated for each clinical trial eligible person from the risk value calculation unit 48, and sorts all clinical trial eligible persons in the order of ascending (smallest) risk value p (ascending order). Furthermore, the subject selection unit 49 reads the planned number of cases N from the trial protocol acquisition unit 42, selects the top N sort results, and outputs the selected patient IDs of eligible clinical trials to the selected subject display unit 45.

FIG. 6 is a diagram schematically illustrating a subject selection operation by the subject selection unit 49.
As shown in the figure, in the subject selection section 49, patient IDs of eligible clinical trials are sorted in ascending order of risk value p (ascending order). For example, when the planned number of cases N = 4, the top four people with the lowest risk values are selected as subjects.

The selected subject display unit 45 displays the patient ID output from the subject selection unit 49 as information on the selected subject.
As a result, the center can preferentially select subjects who have a low probability of early withdrawal or cancellation, and the person who selects the subjects looks at the display of the selected subject display unit 45 for the subjects selected as such. You can know.

  The selected subject display unit 45 may display the hospital ID and patient-specific patient ID of the patient selected as the subject. For this purpose, the selected subject display unit 45 that has received the selected patient ID may read and display the hospital ID and the patient-specific patient ID stored in association with the patient ID from the clinical trial qualified person information database 22. . In this case, the person who selects the subject at the center can know the hospital where the patient selected as the subject is located and the identification information of the patient at the hospital.

Next, an example of a subject selection process in the subject selection device 2 will be described using the flowchart shown in FIG.
In the subject selection process described here, the control unit 35 in FIG. 2 controls the communication unit 31, the input unit 32, the output unit 33, or the storage unit 34 according to the control program stored in the ROM 352 or the storage unit 34. It is executed by controlling. In the following description of the flowcharts, the same processing as that performed by a normal computer, such as communication control by the control unit 35 and memory control, is avoided in order to facilitate understanding.

  In this subject selection process, patient information transmitted from the hospital information processing apparatus 1 of each hospital is accumulated in the anonymized patient information database 21, and the storage unit 34 is based on the epidemiological database provided in the epidemiological data storage apparatus 3. In the state where the risk function and the trial protocol derived in the above are stored, for example, when the operator of the subject selection device 2 inputs information to start the subject selection to the input unit 32, the process is started. .

  First, the control unit 35 extracts patients that match the trial protocol from the anonymized patient information database (DB) 21 (step S11). Specifically, for example, the control unit 35 reads the subject selection criteria described in the trial protocol from the storage unit 34, extracts patient information that matches the subject selection criteria from the anonymized patient information database 21, and is eligible for the trial. Store in the information database 22.

  Next, the control unit 35 determines whether or not the number of extracted patients is larger than the planned number of cases described in the clinical trial protocol (step S12). Specifically, for example, the control unit 35 refers to the clinical trial qualified person information database 22 and counts the number of extracted patients, and reads and reads the number of planned cases described in the clinical trial protocol from the storage unit 34. Compare the number of planned cases with the number of extracted samples.

  When determining that the number of extractions is equal to or less than the number of planned cases (step S12; No), the control unit 35 outputs all the extracted patients (step S13) and ends the subject selection process. Specifically, for example, the control unit 35 displays the patient IDs of all the patients recorded in the clinical trial qualified person information database 22 by extraction on the display device of the output unit 33.

  On the other hand, if it is determined that the extracted number is larger than the planned number of cases (step S12; Yes), the control unit 35 causes the risk function to act on the clinical trial person information database 22 and sorts the risk values in ascending order. Extract (step S14). Specifically, for example, the control unit 35 reads the risk function and the clinical trial implementation period described in the clinical trial protocol from the storage unit 34, and observes the factor corresponding to the covariate of the risk function from the clinical trial eligible person information database 22. Read the value. Then, by substituting the read clinical trial period and the observed value of each factor into the risk function, the risk value is obtained for each clinical trial qualified person, and the patient IDs are sorted in ascending order of the risk value. Further, the planned number of cases N described in the trial protocol is read from the storage unit 34, and the patient IDs of the top N patients with the lowest risk values are extracted.

  Then, the control part 35 outputs the patient of the number of the planned cases extracted (step S15), and complete | finishes a subject selection process. Specifically, for example, the control unit 35 displays the extracted patient IDs for N planned cases on the display device of the output unit 33.

  As described above, according to the subject selection system according to the first embodiment of the present invention, the survival time analysis is performed based on the epidemiological data of the patient, and the subject is selected based on the analysis result. Cases (test candidates) that drop out during the trial period can be rejected with high accuracy before the trial. As a result, it is possible to prevent the cost for the trial from being increased due to the dropout, further to prevent the extension of the trial implementation period, and to conduct the trial efficiently.

Next, a second embodiment of the present invention will be described.
In the first embodiment, the risk value p for the individual i is obtained by performing survival time analysis, and the subjects are selected using the risk value p.
However, for example, in the following case, the risk value p cannot be simply obtained by the method described in the first embodiment.
・ If there is a missing value in the variable observed from the patient ・ There is some interoperability in the data column of the regional cohort analyzed by epidemiology and the Eligible Person Information Database 22 (compatibility even if similar items exist) If there is no)

Therefore, in the clinical trial support system according to the second embodiment of the present invention, the subject is selected by comprehensive judgment including parameters other than the risk value in consideration of information other than the patient's physical information.
There are parameters other than risk values that can be considered in the selection of subjects. For example, in the hospital (clinical trial organization) where patients who are the trial site are attending, the consent acquisition rate, the unit cost (cost required for the trial), There are parameters such as the subject entry rate (whether it progresses according to the schedule), and these parameters are closely related to the total trial cost (trial cost).
In the second embodiment, these parameters are combined with risk values, and a target evaluation function is derived by using a method such as a mathematical calculation method, and subjects are selected so as to minimize the total cost of the clinical trial. To do.

First, preconditions a, b, and c when selecting subjects using the clinical trial support system according to the second embodiment are shown.
a. Each patient is always in one of the hospitals that transmitted the patient information to the anonymized patient information database 21.
b. Let A be the average amount of damage (cost spent on the patient) caused by one subject dropping out (discontinuation of the trial).
c. The cost required when a clinical trial requester such as a pharmaceutical company requests a clinical trial from hospital j (the number of cases scheduled to be commissioned M _j ) is divided into the following two.
・ Fixed cost B (j) incurred uniformly: Labor cost for coordinating clinical trials ・ Consignment cost proportional to the number of cases {C (j) × M _j }: where C (j) is a unit for one subject Commission costs And, there is a difference between hospitals j in fixed costs B (j) and unit commission costs C (j).

  In the subject selection device 2 provided in the clinical trial support system according to the second embodiment, the evaluation calculation is performed by combining such a hospital-side parameter and the risk value p already described in the first embodiment. I do. Then, in order to incorporate the hospital parameters into the evaluation function, the evaluation value is not the risk magnitude but the cost magnitude.

FIG. 8 is a block diagram illustrating a functional configuration example of the subject selection device 2 included in the clinical trial support system according to the second embodiment of the present invention.
The configurations of the clinical trial support system and subject selection apparatus 2 according to the first embodiment shown in FIGS. 1 and 2 are the same as those of the clinical trial support system and subject selection apparatus 2 according to the second embodiment.
Further, in FIG. 8, anonymized patient information receiving unit 41, clinical trial protocol acquiring unit 42, clinical trial qualified person extracting unit 43, planned case number determining unit 44, selected subject display unit 45, epidemiological data receiving unit 46, risk function deriving unit. 47, since the risk value calculation part 48 has the same function as the test subject selection apparatus 2 which concerns on 1st Embodiment shown in FIG. 3, the same code | symbol is attached | subjected and the description is abbreviate | omitted.

The risk value storage unit 51 stores the risk value p calculated by the risk value calculation unit 48 for each clinical trial eligible person in association with the patient ID of the clinical trial eligible person.
The clinical trial qualified person combination unit 52 reads the planned number N of cases described in the clinical trial protocol from the clinical trial protocol acquisition unit 42. Then, the trial eligible persons registered in the clinical trial eligible person information database 22 are used as candidate test subjects, and all combinations for selecting the planned number N of cases are obtained. Furthermore, the clinical trial qualified person combination unit 52 reads out and stores the patient ID and hospital ID of the clinical trial qualified person (extraction candidate) constituting the combination from the clinical trial qualified person information database 22 for each obtained combination.

The premise cost storage unit 53 calculates the trial cost using the damage average amount A described above as a precondition for subject selection in the present embodiment, the fixed cost B (j) and the unit consignment cost C (j) for each hospital j. Remembering as a premise cost for.
FIG. 9 is a diagram illustrating an example of a data table that stores clinical trial costs for hospitals.
FIG. 9A is a data table for storing a fixed cost B (j) for each hospital j out of clinical trial costs for the hospital. The fixed cost of the hospital is stored in association with the hospital ID that identifies the hospital. Yes. FIG. 9 (b) is a data table for storing case proportional remuneration, which is the unit commission cost C (j) for each hospital j, out of the clinical trial costs for the hospital, and the case proportional remuneration of the hospital is related to the hospital ID. It is remembered.
The data tables shown in FIGS. 9A and 9B are stored in the storage unit 34 of FIG. 2 constituting the premise cost storage unit 53, and the clinical trial cost calculation unit 54 described below calculates the clinical trial cost. At this time, it is read as necessary.

The clinical trial cost calculation unit 54 calculates the clinical trial cost, which is the cost for conducting the clinical trial, for each combination obtained by the clinical trial qualified person combination unit 52. Specifically, the trial cost calculation unit 54 reads the premise cost from the premise cost storage unit 53, reads the patient ID and hospital ID of the clinical trial qualified person who constitutes the combination for each combination from the trial qualified person combination unit 52, The risk value p corresponding to the patient ID is read from the risk value storage unit 51. Then, based on each read data, the trial cost is calculated for each combination obtained by the trial qualified person combination unit 52. Further, the clinical trial cost calculation unit 54 associates the patient IDs of the clinical trial qualified persons constituting the combination with the clinical trial cost calculated by the combination and outputs them to the combination selection unit 55.
The details of the operation for calculating the clinical trial cost in the clinical trial cost calculation unit 54 will be described later.

The combination selection unit 55 selects the combination that minimizes the clinical trial cost calculated by the clinical trial cost calculation unit 54 from the combinations obtained by the clinical trial qualified person combination unit 52. Specifically, for example, when receiving the clinical trial cost calculated for each combination from the clinical trial cost calculation unit 54, the combination selection unit 55 stores the minimum value of the clinical trial cost and the combination constituent. Then, after receiving the clinical trial costs for all combinations from the clinical trial cost calculation unit 54, the patient IDs of the clinical trial qualified persons constituting the combination stored as the clinical trial cost being the minimum value are output to the selected subject display unit 45. .
The selected subject display unit 45 displays the patient ID output from the combination selection unit 55 as information on the selected subject.

Next, the operation for calculating the clinical trial cost in the clinical trial cost calculating unit 54 will be described.
In the present embodiment, since the magnitude of the cost is used as the evaluation value, the clinical trial cost calculation unit 54 performs an operation for converting the risk value p of the patient into the cost.
As already explained, the risk value of the individual i is p (i). Under the precondition b described above, the cost (expected value of damage) generated in the clinical trial for the individual i is p (i) × A.
Therefore, when considering the expected value of damage for each combination obtained by the trial-qualified person combination unit 52, the expected value of damage in one combination is calculated as p (i) × A calculated for each clinically qualified person i constituting the combination. The sum of

Here, for a patient whose risk value p (i) cannot be calculated, 1-h ₀ (t) is substituted for the risk value p (i) of the patient using the baseline hazard h ₀ (t). Use.
In addition, as an alternative to the risk value of the patient that cannot be calculated, an average value of the patient risk value that can be calculated may be used.

Further, as shown in the precondition c, when conducting a clinical trial, a clinical trial requesting fee for requesting the clinical trial to a hospital (clinical trial implementation facility) where the patient is located is generated.
When calculating the trial request cost for one combination, the trial cost calculation unit 54 first obtains the hospital ID of each clinical trial eligible person constituting the combination, and calculates the number of clinical trial eligible persons with the same hospital ID in the combination. Count. Thereby, the clinical trial cost calculation unit 54 acquires the hospital j when the clinical trial is performed with the combination, and the scheduled commission number M _j indicating the number of subjects for which the clinical trial is performed at the hospital j.
In this case, the clinical trial request cost to the hospital j is the sum of the fixed cost B (j) and the commissioned cost {C (j) × M _j }.
For this reason, the clinical trial request cost in one combination is the sum of B (j) + C (j) × M _j calculated for each hospital j in which the clinical trial is performed in that combination.

  Considering the above expected value of damage and clinical trial request cost, the evaluation function for obtaining the total cost (expected cost value) S for the clinical trial in one combination obtained by the trial qualified person combination unit 52 is as follows: become.

  The clinical trial cost calculation unit 54 reads out the average loss A from the premise cost storage unit 53 and stores it in the premise cost storage unit 53 based on the hospital ID of the clinical trial person who constitutes the combination acquired from the trial person combination unit 52. The fixed cost B (j) and the commissioned cost C (j) of the hospital j are read out from the data table shown in FIG. Furthermore, the clinical trial cost calculation unit 54 reads the risk value p of the patient from the risk value storage unit 51 based on the patient ID of the clinical trial qualified person who constitutes the combination acquired from the clinical trial qualified person combination unit 52. Then, the clinical trial cost calculation unit 54 uses the evaluation function of the above formula (3) based on the read values, and costs for the clinical trial, which is the clinical trial cost, for each combination obtained by the clinical trial qualified person combination unit 52. Is calculated as S.

  The combination selection unit 55 selects a combination of eligible clinical trials that minimizes the S calculated by the clinical trial cost calculation unit 54, so that the clinical trial requester such as a pharmaceutical company can reduce the cost of entrusting the clinical trial. it can.

Next, the subject selection process in the subject selection apparatus 2 of the clinical trial support system according to the second embodiment of the present invention will be described with reference to the flowcharts shown in FIGS.
FIG. 10 is a flowchart showing an example of a subject selection process according to the second embodiment of the present invention.
Note that the subject selection process shown in FIG. 10 corresponds to the subject selection process according to the first embodiment shown in FIG. 7 with steps 24 and 25 added.

When the operator of the subject selection apparatus 2 inputs information indicating that the subject selection is started, for example, to the input unit 32, the subject selection process is started, and the control unit 35 selects a patient that matches the trial protocol. Extract from the anonymized patient information database 21 (step S21).
Next, the control unit 35 determines whether or not the number of extracted patients is larger than the planned number of cases described in the clinical trial protocol (step S22).
When determining that the number of extractions is equal to or less than the number of planned cases (step S22; No), the control unit 35 outputs all the extracted patients (step S23) and ends the subject selection process.

  On the other hand, if it is determined that the extracted number is larger than the planned number of cases (step S22; Yes), the control unit 35 determines whether or not the risk function can be applied to all the patients extracted in step S21 (step S24). Specifically, for example, the control unit 35 refers to the patient information of the clinical trial qualified person stored in the clinical trial qualified person information database 22 and has a factor item that can be interoperated with a factor corresponding to the covariate of the risk function. If there is no patient information, or if there is a missing value even if there is a factor item that can be interoperated, it is determined that the risk function cannot be applied to all the extracted patients.

  When the control unit 35 determines that the risk function cannot be applied to all the extracted patients (step S24; No), the control unit 35 selects a subject by performing a comprehensive determination process including parameters other than the risk value ( Step S25), the subject selection process is terminated. The details of the “total judgment process including parameters other than the risk value” will be described later.

  If the control unit 35 determines that the risk function can be applied to all extracted patients (step S24; Yes), the control unit 35 causes the risk function to act on the clinical trial person information database 22 and sorts the risk values in ascending order. The top N people are extracted (step S26). Then, the control part 35 outputs the patient of the number of the planned cases extracted (step S27), and complete | finishes a subject selection process.

Next, “overall judgment processing including parameters other than risk values” in step S25 will be described with reference to FIG.
In this “comprehensive judgment process including parameters other than risk values”, the control unit 35 first, as a loop (repetition) 1, the patient extracted from the anonymized patient information database 21 in step S21 of FIG. ), The processes of steps S32 to S34 are performed for all combinations from which N people are extracted (step S31). Specifically, for example, the control unit 35 reads the planned number N of cases described in the clinical trial protocol from the storage unit 34, and from the clinical trial qualified persons registered in the clinical trial qualified person information database 22, the number of planned cases N Find all combinations that extract.

  In the loop 1, the control unit 35 performs a total cost calculation process for the clinical trial for one combination (step S32). With this process, the control unit 35 calculates the total cost S of clinical trials for one combination. The details of this “total cost calculation processing for clinical trials” will be described later.

Next, the control unit 35 determines whether or not the total cost S of the clinical trial calculated in step S32 is the minimum value in the loop (step S33). Specifically, for example, the control unit 35 calculates S that has been calculated by the loop 1 that has already been executed and is minimized among other combinations, and S that is calculated in step S32 in the same loop 1; Are compared to determine whether the value is the minimum value.
When the control unit 35 determines that S calculated in step S32 is the minimum value (step S33; Yes), the control unit 35 saves a list of N persons constituting the combination of S and the S in the storage unit 34 (step S34). ). Then, the termination condition of loop 1 shown in step S31 is determined. On the other hand, if it is determined that the value is not the minimum value (step S33; No), step S34 is skipped, and the end condition of the loop 1 is determined.

If the control part 35 performs the process of these steps S32-S34 with respect to all the combinations which extract N persons from the patient (clinical trial qualified person) extracted from the anonymized patient information database 21, it will satisfy | fill an end condition. , Exits loop 1 (step S35).
Subsequently, the control unit 35 outputs the list of N persons finally stored in the storage unit 34 to the output unit 33 (step S36), and ends the “overall determination process including parameters other than the risk value”. Then, the process returns to the subject selection process of FIG.

If it is determined in step S24 in FIG. 10 that the risk function cannot be applied to all extracted patients by this “total judgment process including parameters other than risk values” (step S24; No), the clinical trial is performed. N trial eligible persons who constitute a combination that minimizes the sum S of costs are selected as subjects.
Therefore, it is possible to select subjects who minimize the clinical trial cost from the standpoint of the clinical trial client such as a pharmaceutical company.
Further, in this “comprehensive judgment processing including parameters other than risk values”, S is calculated for all combinations selected from N eligible clinical trial subjects (subject candidates) registered in the clinical trial eligible person information database 22, By storing the minimum value, a solution that minimizes S can be easily derived.

Next, the “total cost calculation processing for clinical trial” in step S32 in FIG. 11 will be described with reference to FIG.
In the “total cost calculation process for clinical trials”, the control unit 35 first performs a damage calculation process due to dropout of the subject (step S41). By this processing, the control unit 35 calculates the damage p1 that occurs when the subject drops out for one combination. The details of the “damage calculation process due to dropout of subject” will be described later.

  Next, the control unit 35 obtains the sum ΣB (j) of all fixed hospitals required for entrusting to the hospital j and assigns the value to p2 (step S42). In other words, the control part 35 calculates | requires the sum total of the fixed cost over all the hospitals which entrust clinical trial with respect to one combination, when clinical trial is implemented by the combination.

Next, the control part 35 _calculates | requires the sum total (SIGMA) {C (j) * _Mj } over all the hospitals of the commission expense {C (j) * _Mj } required when entrusting to the hospital j, and the value Is substituted for p3 (step S43). In other words, the control part 35 calculates | requires the sum total of the consignment expense over all the hospitals which entrust clinical trial, when clinical trial is implemented with the combination with respect to one combination.
Subsequently, the control unit 35 substitutes P1 + P2 + P3 into S as the sum of costs (step S44), ends the “damage calculation processing due to dropout of the subject”, and completes “comprehensive determination including parameters other than the risk value” in FIG. Return to Processing.

Next, the “damage calculation process due to dropout of subject” in step S41 in FIG. 12 will be described with reference to FIG.
In the “damage calculation process due to dropout of subject”, the control unit 35 first assigns 0 to P1 (step S51).
Next, in the loop 2, the control unit 35 sets the initial value of the variable i to 0, increases the variable i by 1, and executes the processes of steps S53 to S55 while the variable i is smaller than N (step S52). . Here, the variable i represents N trial eligible persons (i = 0 to N−1) constituting one combination.

In the loop 2, the control unit 35 determines whether or not the risk value p (i) corresponding to the variable i can be calculated (step S53).
When determining that the risk value p (i) can be calculated (step S53; Yes), the control unit 35 calculates P1 + p (i) × A, substitutes the calculated value for P1, and sets the value of P1. Update (step S54). Then, the end condition of the loop 2 shown in step S52 is determined. In addition, A is the amount of loss (average) generated by dropout per subject as described above.

On the other hand, if it is determined that the risk value p (i) cannot be calculated (step S53; No), the control unit 35 uses 1−h ₀ (t) as an alternative to the risk value p (i), and P1 + {1 -H ₀ (t)} × A is calculated, the calculated value is substituted into P1, and the value of P1 is updated (step S55). Then, the end condition of the loop 2 is determined. . Note that h ₀ (t) is a baseline hazard as described above.

The control unit 35 repeats the loop 2 and when the variable i reaches N (when the termination condition is satisfied), the control unit 35 exits the loop 2 (step S56), ends the “damage calculation processing due to dropout of the subject”, and finishes in FIG. Return to "Cost total calculation processing for clinical trials".
With this “damage calculation process due to dropout of subject”, p1 becomes the amount of damage due to dropout of subject in one combination.

  As described above, according to the subject selection system according to the second embodiment of the present invention, the survival time analysis is performed based on the patient epidemiological data, and the clinical trial results and clinical trial costs at each hospital are reduced. Consider subjects to be selected. As a result, cases that drop out during the clinical trial can be repelled in advance, the cost of the clinical trial can be reduced, and the clinical trial can be performed efficiently.

Although the embodiments of the present invention have been described above, the present invention can be modified and applied in various forms and is not limited to the above embodiments.
For example, in the second embodiment described above, the patient information has a missing value and the risk value cannot be simply calculated. However, there is no missing value in the patient information, and all patients have Even when the risk value can be obtained, the subject may be selected by calculating the clinical trial cost described in the second embodiment. In that case, the cost that becomes the evaluation value can be calculated more accurately, and the subject can be selected for conducting the clinical trial more efficiently.

Further, parameters to be considered in selecting a subject are not limited to the parameters shown in the above embodiment.
For example, the presence / absence of a patient's own car, the distance from the patient's home to the trial hospital, age, workplace, etc. may be considered as parameters for subject selection. When these parameters are taken into consideration, subjects who are likely to drop out due to a car accident during the trial period can be rejected in advance, and the cost of the trial due to the dropout of the subject and the extension of the trial period can be reduced. Can be prevented.

In the first embodiment, the evaluation function mainly focuses on minimizing the probability that the subject will drop out during the trial. In the second embodiment, the total cost of the trial is minimized. The subjects were selected based on the evaluation function whose main viewpoint is to make the evaluation function, but there are various other ways of taking the evaluation function.
For example, for the purpose of setting up a clinical trial plan that takes into account the requests of clinical trial contractors such as pharmaceutical companies, evaluation functions such as the viewpoint of minimizing the number of contracted institutions and the viewpoint of minimizing the trial period are set. In addition, subjects may be selected based on these evaluation functions.

Moreover, in the said embodiment, the clinical trial support system, subject selection apparatus, etc. for performing a clinical trial efficiently were demonstrated. However, the present invention is not limited to clinical trials, and can also be applied to other clinical trials that do not aim to obtain approval for pharmaceuticals such as post-marketing clinical trials.
In the above embodiment, the risk function is derived using the proportional hazard model. However, the present invention is not limited to this, and the risk function may be derived using, for example, another survival time analysis method using a logistic model.

  In the above embodiment, the trial protocol is stored in advance in the storage unit, and information such as the subject selection criteria, the number of planned cases, the trial period, etc. described in the trial protocol is read and acquired as necessary. . However, you may acquire such information by receiving from an input part as needed.

  In the above embodiment, the communication network that connects the center and the hospital and the communication network that connects the center and the epidemiological data recording device are separated, but the connection is made using a common communication network such as the Internet. May be. The center may be provided with an epidemiological data recording device.

The subject selection device of the present invention is not limited to dedicated hardware, and can also be realized by a normal computer system.
For example, in the above embodiment, the program of the subject selection device has been described as being stored in advance in a memory or the like. However, a program for executing the above-described processing operation is recorded on a computer-readable recording medium such as a flexible disk, a CD-ROM (Compact Disk Read-Only Memory), a DVD (Digital Versatile Disk), or an MO (Magneto-Optical disk). A subject selection apparatus that performs the above-described processing operation may be configured by storing and distributing the program in a medium and installing the program in a computer.

Further, the program may be stored in a disk device or the like included in a predetermined server device on a communication network such as the Internet, and may be downloaded onto a computer by being superimposed on a carrier wave, for example. Furthermore, the above-described processing can also be achieved by starting and executing a program while transferring it via a communication network.
In addition, when the above functions are realized by sharing an OS (Operating System), or when the functions are realized by cooperation between the OS and an application, only the part other than the OS may be stored in a medium and distributed. Alternatively, it may be downloaded to a computer.

It is a block diagram which shows the structural example of the clinical trial support system which concerns on the 1st Embodiment of this invention. It is a block diagram which shows the structural example of a test subject selection apparatus. It is a block diagram which shows the function structural example of a test subject selection apparatus. It is a figure which shows the data structural example of the patient information memorize | stored in the anonymization patient information database. It is a figure which shows the data structural example of the epidemiological data memorize | stored in the epidemiological database. It is a figure which shows typically a subject's selection operation by a subject selection part. It is a flowchart which shows the example of the test subject selection process which concerns on the 1st Embodiment of this invention. It is a block diagram which shows the function structural example of the test subject selection apparatus which concerns on the 2nd Embodiment of this invention. It is a figure which shows the example of the data table which memorize | stores the clinical trial cost with respect to a hospital. It is a flowchart which shows the example of the test subject selection process which concerns on the 2nd Embodiment of this invention. It is a flowchart which shows the example of the comprehensive judgment process including parameters other than a risk value. It is a flowchart which shows the example of the sum total calculation process of the cost concerning a clinical trial. It is a flowchart which shows the example of the damage calculation process by a test subject drop | exit.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Hospital information processing apparatus 2 Test subject selection apparatus 3 Epidemiology data storage apparatus 4,5 Communication network 21 Anonymized patient information database 22 Clinical trial person information database 31 Communication part 32 Input part 33 Output part 34 Storage part 35 Control part 36 Bus 41 Anonymous Patient Information Receiving Unit 42 Clinical Trial Protocol Acquisition Unit 43 Clinical Trial Qualified Person Extraction Unit 44 Planned Case Number Discrimination Unit 45 Selected Subject Display Unit 46 Epidemiological Data Receiving Unit 47 Risk Function Deriving Unit 48 Risk Value Calculation Unit 49 Subject Selection Unit 51 Risk Value Storage Part 52: Trial eligible person combination part 53: Premise cost storage part 54: Clinical trial cost calculation part 55: Combination selection part 351 CPU
352 ROM
353 RAM

Claims (5)

A test subject selection device for selecting test subjects from candidates,
A risk function storage means for storing a risk function for obtaining a risk value, which is a probability that the clinical trial is suspended due to a change in the candidate's symptom during the clinical trial, for each candidate;
For each candidate, candidate information storage means for storing candidate information including at least the identification information of the candidate and an observed value of a predetermined factor corresponding to the covariate of the risk function;
An implementation period information acquisition means for acquiring the implementation period information of the clinical trial;
A risk function is read from the risk function storage means, candidate information is read for each candidate from the candidate information storage means, and an independent variable of the risk function includes a predetermined factor corresponding to the covariate of the candidate information. A risk value calculating means for substituting the observation value and the execution period information acquired by the execution period information acquisition means, and obtaining the risk value that is a dependent variable of the risk function for each candidate;
Planned case number storage means for storing the planned number N of clinical trials;
The planned number of cases N is read from the planned number of cases storage means, and N candidates are selected from the candidates in order of increasing risk values obtained by the risk value calculation means, and the selected candidates are selected. Identification information output means for outputting identification information;
A subject selection device comprising: Epidemiological data acquisition means for acquiring epidemiological data collecting patient information in past affected cases;
A risk function deriving unit for performing statistical analysis based on the epidemiological data acquired by the epidemiological data acquiring unit and deriving the risk function;
The subject selecting device according to claim 1, comprising: The risk function deriving means includes:
Means for performing survival time analysis based on the epidemiological data acquired by the epidemiological data acquiring means, and deriving a hazard function h (z _i , t) that is a function of the covariate vector z _i of the patient _i and the time t;
means for deriving a function p (i, t) as the risk function by calculating p (i, t) = 1−h (z _i , t);
The subject selecting device according to claim 2, wherein A test subject selection device for selecting test subjects from candidates,
A risk function storage means for storing a risk function for obtaining a risk value, which is a probability that the clinical trial is suspended due to a change in the candidate's symptom during the clinical trial, for each candidate;
For each candidate, candidate information storage means for storing candidate information including at least the identification information of the candidate and an observed value of a predetermined factor corresponding to the covariate of the risk function;
An implementation period information acquisition means for acquiring the implementation period information of the clinical trial;
A risk function is read from the risk function storage means, candidate information is read for each candidate from the candidate information storage means, and an independent variable of the risk function includes a predetermined factor corresponding to the covariate of the candidate information. A risk value calculating means for substituting the observation value and the execution period information acquired by the execution period information acquisition means, and obtaining the risk value that is a dependent variable of the risk function for each candidate;
Risk value storage means for storing the risk value for each candidate obtained by the risk value calculation means in association with identification information of the candidate;
Planned case number storage means for storing the planned number N of clinical trials;
The planned case number N is read from the planned case number storage means, all combinations for extracting N persons from the candidates are obtained, and the identification information of the extracted candidate who is the extracted candidate is obtained for each combination. Candidate extraction means to obtain;
The risk value stored in association with the identification information of the extraction candidate acquired by the candidate extraction unit is read from the risk value storage unit, and the clinical test is performed for each combination based on the read risk value. Expense expectation value calculation means for obtaining an expense expectation value that is an expectation value of the cost of implementation of
Identification information output means for outputting identification information of extraction candidates in the smallest combination of expected expenses calculated by the expected expenses calculation means;
A subject selection device comprising: The candidate information stored in the candidate information storage means includes, for each candidate, facility identification information for identifying a facility where a clinical trial of the candidate is performed,
The expected cost calculation means is:
Damage amount storage means for storing the amount of damage A that occurs when the clinical trial of one subject is stopped;
Fixed cost storage means for storing a fixed cost B, which is a cost that is uniformly generated when requesting a clinical trial from the facility, for each facility specified by the facility identification information;
Unit-contracted cost storage means for storing unit-consigned cost C, which is a cost incurred when requesting one facility to perform a clinical trial for each subject, for each facility specified by the facility identification information;
The facility identification information stored in association with the identification information of the extraction candidate i acquired by the candidate extraction means is read from the candidate information storage means, and the extraction candidate i is read based on the read facility identification information. A facility identification means for identifying a facility j used when a clinical trial is performed in combination of
Wherein each facility identification means has identified facility j, and accession number of cases counting means for counting the accession number of cases M _j is performed the number of facilities j when clinical trials carried out in a combination of the extraction candidate i ,
Read the risk value p (i) stored in association with the identification information of the extraction candidate i from the risk value storage means, read the damage amount A from the damage amount storage means, from the fixed cost storage means, The fixed value B (j) of the facility j specified by the facility specifying means is read, the unit trust cost C (j) of the facility j specified by the facility specifying means is read from the unit trust cost storage means, and the trust case The number M _j of entrusted cases of the facility j specified by the facility specifying means is acquired from the number counting means, and the expected cost value S is obtained for each combination.

Means to obtain by performing the operation of,
The subject selection apparatus according to claim 4.

Images