KR20210155883A - Analysis System and Method for Time of Change Point of Spontaneous Vaccine Adverse Event Reports - Google Patents

Abstract

The present invention relates to a temporal variation point analysis system and an analysis method thereof in spontaneous vaccine adverse event reports. According to the present invention, disclosed in the present invention is a technique capable of statistically searching or analyzing a time point showing a difference from the overall aspect of adverse event reports that change over time by comprising: a database unit storing spontaneous adverse event report data; a data preprocessing unit pre-processing target vaccine adverse event report data from the spontaneous adverse event report data; and a variation point analysis unit testing for the significance of temporal variations in the distribution of the cumulative sum width of index deviations of multiple samples.

Description

Analysis System and Method for Time of Change Point of Spontaneous Vaccine Adverse Event Reports

The present invention relates to a signal information retrieval method that takes into account temporal variability using spontaneous adverse event report data reported after vaccination, and supplements the limitations of the signal information retrieval method based on imbalance analysis used to discover drug safety problems. It relates to a temporal variation point analysis system and analysis method of spontaneous vaccine adverse event reporting that can be done.

As interest in drug safety has recently been increasing and its importance is being emphasized, the Food and Drug Administration, medical institutions where drugs are prescribed and manufactured, and pharmaceutical companies that produce drugs, etc. this is being done

To this end, drug side effects data are collected at various institutions and the data are being made into a database. Recently, the Food and Drug Administration has established a database of about 80,000 drug side effects by reorganizing drug side effects reported to the Food and Drug Administration from 1989 to early 2010. As of 2010, general or specialized hospitals, including 15 regional drug monitoring centers, are accumulating drug side effects data at each hospital. In addition, pharmaceutical companies are also building a database while reporting side effects that occurred in patients administered with their developed drugs to government agencies through clinical trials and post-marketing investigations. Therefore, using this database, information on clues about adverse reactions of drugs is being searched.

Drugs, especially vaccines, are generally administered to prevent disease in healthy people, and administration is recommended for most citizens in the case of national mandatory vaccination. Therefore, in the case of vaccines from a public health point of view, if the number of adverse events is rapidly increasing compared to the number of cases reported normally, it is necessary to quickly identify this and take measures such as an epidemiological investigation or recall of a vaccine product.

However, the conventional imbalance-based clue search is an approach that analyzes the property of whether a specific adverse event is reported particularly among all reported cases, and there is a limitation in that it cannot be determined whether the number of reports is rapidly increasing at a specific point in time. .

Republic of Korea Patent No. 10-1611687

An object of the present invention is to solve the problems of the prior art as described above, and to enable the search or analysis of clue information from the voluntary adverse event report data in consideration of temporal variability, a temporal change point analysis system for voluntary vaccine adverse event reports and To provide an analysis method.

In order to achieve the above object, the temporal variation timing analysis system of spontaneous vaccine adverse event report according to an embodiment of the present invention for achieving the above object is one or more vaccines belonging to a predetermined search target period and one or more adverse event information reported about the vaccine A database unit for storing voluntary adverse event report data including; a data preprocessing unit for extracting analysis target vaccine-adverse event report data from the spontaneous adverse event report data stored in the database unit, and constructing the analysis target vaccine-adverse event report data divided by a predetermined period; And based on the analysis target vaccine-adverse event report data constructed by dividing the data pre-processing unit by period, a plurality of samples are derived for the analysis target vaccine-adverse event report data, and the timing of change according to the significance of temporal variation a change point analysis unit that defines ; It may be characterized as one of including.

Here, the data pre-processing unit removes the redundancy due to the tracking report of individual adverse event reports from the spontaneous adverse event report data including the one or more vaccines and one or more adverse event information, and then prepares the analysis target vaccine-adverse event report data. Extraction may be another feature.

Here, the change point analysis unit calculates an indicator for each period unit using the analysis target vaccine-adverse event report data preprocessed by the data preprocessing unit, and accumulates the indicator deviation for each period unit using the indicator for each period unit By calculating the total width, a plurality of samples are derived for the analysis target vaccine-adverse event report data, and the significance of temporal variation in the distribution of the cumulative sum width of the index deviation of the plurality of samples is tested to define the point of change It can also be characterized as another feature.

In order to achieve the above object, the method for analyzing the temporal variation of the spontaneous vaccine adverse event report according to an embodiment of the present invention is a spontaneous adverse event report including one or more vaccines and one or more adverse event information reported with respect to the vaccine A pre-processing step of extracting analysis target vaccine-adverse event report data reported within a predetermined search target period from the data, and constructing the analysis target vaccine-adverse event report data divided by a predetermined period unit; And, based on the analysis target vaccine-adverse event report data constructed by dividing the period in the pre-processing step, a plurality of samples are derived for the analysis target vaccine-adverse event report data, and the time of change is determined according to the significance of temporal variation. It may be characterized as including; a change point analysis step to define.

Here, the change point analysis step comprises: an index calculation step of calculating an index for each period unit using the analysis target vaccine-adverse event report data constructed in the period unit; a variation width calculation step of calculating the width of the index fluctuation for each period unit or the cumulative sum width of the index deviation for each period unit among the indicators for each period unit calculated in the index calculation step; The analysis target vaccine, including the index for each period unit calculated in the index calculation step, the width of the index change for each period unit calculated in the fluctuation range calculation step, or the cumulative sum width of the index deviation for each period unit - report an adverse event a sampling step of deriving a plurality of samples for the data and deriving the cumulative sum width of the index deviation for each sample; a significance test step of testing the significance of temporal variation in the distribution of the cumulative sum width of the index deviations of the plurality of samples derived in the sampling step; And when the significance probability is less than or equal to the reference value by testing the significance of temporal fluctuations in the significance test step, the time point following the point at which the absolute value of the cumulative sum of the index deviations in each of the plurality of samples is the maximum is defined as the time point of variation It may be characterized as another feature to include;

Here, a period resetting step of dividing the search target period into a period T1 that is a time period before the change time and a period T2 that is a time period after the change point on the basis of the change time defined in the change time point definition step; It may be characterized as another feature to further include.

Here, the pre-processing step includes: a duplicate case removal step of removing duplicates due to tracking reports of individual adverse event reports from voluntary adverse event report data including one or more vaccines and one or more adverse event information within a predetermined search target period; The analysis target vaccine-adverse event, which is the report data corresponding to the adverse event of the target vaccine and the target vaccine, in the voluntary adverse event report data from which the duplication due to the follow-up report of the individual adverse event report was removed in the duplicate case removal step an extraction step of extracting the report data; and a period unit construction step of dividing and constructing the analysis target vaccine-adverse event report data extracted in the extraction step into a predetermined period unit.

Here, the index calculation step, the fluctuation range calculation step by using each of the T1 th period and the T2 th period divided in the period resetting step as the search target period; the sampling step; the significance test step; and the step of defining the change time point; may be characterized in that it is sequentially re-performed.

Here, the indicator for each period unit calculated in the indicator calculation step includes the number of adverse event reports per period unit, an imbalance-based proportional reporting ratio (PRR), the lower limit of the reporting ratio ratio, and an imbalance-based report Reporting odds ratio (ROR), the lower limit of the reporting odds ratio, an information component (IC) of an imbalance-based Bayesian trust transmission neural network, the lower limit of the information component, and anomaly for each period unit within the search target period Another feature may be that at least one of the average of the number of case reports, the deviation of the number of adverse event reports per period unit, and the cumulative sum of the deviation of the number of adverse event reports per period unit is included.

Here, the derivation of the plurality of samples in the sample derivation step may be further characterized by deriving the plurality of samples using a bootstrap method of restoring and extracting the samples.

Here, as another feature, the reference value of the significance probability has a value less than 0.05.

Here, in the significance test step, when the significance of the temporal variation is tested for and the significance probability exceeds the reference value, it may be another feature of terminating the analysis at the point of time of variation.

The temporal variation time point analysis system and analysis method of spontaneous vaccine adverse event report according to an embodiment of the present invention can statistically search or analyze the time point showing the difference with the overall aspect of adverse event reports that change over time, so through epidemiological investigation It has the effect of helping to evaluate a causal relationship or to take measures such as suspension of vaccination or recall of vaccine products more quickly.

1 is a block diagram schematically showing a temporal variation time point analysis system of spontaneous vaccine adverse event report according to an embodiment of the present invention.
2 is a graph schematically showing the number of adverse events reported by period unit in the temporal variation time point analysis system of spontaneous vaccine adverse event reports according to an embodiment of the present invention.
3 is a graph schematically showing the cumulative total of deviations in the number of adverse events reported by period unit in the temporal variation timing analysis system of spontaneous vaccine adverse event reports according to an embodiment of the present invention.
4 is a diagram schematically illustrating sampling restoration from spontaneous adverse event report data in the temporal variation time point analysis system of spontaneous vaccine adverse event report according to an embodiment of the present invention.
5 is a diagram schematically showing the cumulative sum width of deviations in each sample as a probability density function in the temporal variation time point analysis system of spontaneous vaccine adverse event reports according to an embodiment of the present invention.
6 is a graph schematically showing the division of periods based on the time of change in the temporal change point analysis system of the spontaneous vaccine adverse event report according to an embodiment of the present invention.
7 is a graph schematically showing the analysis of the change point by repeatedly applying the change point analysis execution.
8 is a flowchart schematically illustrating a temporal variation time point analysis method of a spontaneous vaccine adverse event report according to an embodiment of the present invention.
9 is a graph schematically showing the number of adverse events reported by period unit in the temporal variation time point analysis method of spontaneous vaccine adverse event reports according to an embodiment of the present invention.
10 is a graph schematically showing the cumulative sum of deviations in the number of adverse event reports by period unit in the temporal variation time point analysis method of spontaneous vaccine adverse event reports according to an embodiment of the present invention.
11 is a diagram schematically showing the cumulative sum width of deviations in each sample as a probability density function in the temporal variation time point analysis method of spontaneous vaccine adverse event report according to an embodiment of the present invention.
12 is a graph schematically showing defining a change point according to the cumulative sum of deviations in the temporal change point analysis method of spontaneous vaccine adverse event report according to an embodiment of the present invention.
13 is a graph schematically showing the division into periods before and after the time of change in the temporal change point analysis system of the spontaneous vaccine adverse event report according to an embodiment of the present invention.
14 is a graph schematically showing the analysis of the change time point by repeatedly applying the change time point analysis implementation.

Hereinafter, a preferred embodiment will be described with reference to the accompanying drawings so that the present invention may be understood in more detail.

Since the present invention can have various changes and can have various embodiments, specific embodiments are illustrated in the drawings and described in detail. However, this is not intended to limit the present invention to specific embodiments, and it should be understood to include all modifications, equivalents and substitutes included in the spirit and scope of the present invention.

Terms such as first, second, etc. may be used to describe various elements, but the elements should not be limited by the terms. The above terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, a first component may be referred to as a second component, and similarly, a second component may also be referred to as a first component. and/or includes a combination of a plurality of related listed items or any of a plurality of related listed items.

When an element is referred to as being “connected” or “connected” to another element, it is understood that it may be directly connected or connected to the other element, but other elements may exist in between. it should be On the other hand, when it is said that a certain element is "directly connected" or "directly connected" to another element, it should be understood that the other element does not exist in the middle.

The terms used in the present application are only used to describe specific embodiments, and are not intended to limit the present invention. The singular expression includes the plural expression unless the context clearly dictates otherwise. In the present application, terms such as “comprise” or “have” are intended to designate that a feature, number, step, operation, component, part, or combination thereof described in the specification exists, but one or more other features It should be understood that this does not preclude the existence or addition of numbers, steps, operations, components, parts, or combinations thereof.

Unless defined otherwise, all terms used herein, including technical and scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Terms such as those defined in a commonly used dictionary should be interpreted as having a meaning consistent with the meaning in the context of the related art, and should not be interpreted in an ideal or excessively formal meaning unless explicitly defined in the present application. does not

Hereinafter, preferred embodiments of the present invention will be described in more detail with reference to the accompanying drawings. In describing the present invention, in order to facilitate the overall understanding, the same reference numerals are used for the same components in the drawings, and duplicate descriptions of the same components are omitted.

1 is a block diagram schematically showing a temporal variation time point analysis system of spontaneous vaccine adverse event report according to an embodiment of the present invention.

Referring to FIG. 1 , the temporal variation point analysis system 100 of the spontaneous vaccine adverse event report according to an embodiment of the present invention includes a database unit 110 , a data pre-processing unit 130 , and a variation time point analysis unit 130 . and may further include a query unit 140 .

The temporal variation timing analysis system of the spontaneous vaccine adverse event report shown in FIG. 1 is according to an embodiment of the present invention, and some blocks may be added, changed, or deleted in another embodiment.

The database unit 110 stores voluntary adverse event report data including one or more drugs and one or more adverse event information belonging to a predetermined search target period.

The database unit 110 may use a stored database input by a user, or may register and use a database accumulated in a local drug monitoring center, hospital, pharmaceutical company, and the like. At this time, the database unit 110 may retrieve and store the one or more drugs and one or more adverse event information from a computer-readable file format. Here, the computer-readable file format may be in the form of a spread sheet, and may also be in the form of an Excel file or an Access file of Microsoft.

The inquiry unit 140 makes it possible to inquire the voluntary adverse event report data including one or more vaccines and one or more adverse event information stored in the database unit 110 within a predetermined search target period.

And, according to the needs of the user, the inquiry unit 140 may inquire the analysis target vaccine-adverse event report data pre-processed by the data pre-processing unit 120 . In addition, the variation time point analysis unit 130 tests for the significance of temporal variation by examining the distribution of the cumulative total width of the deviation in the number of adverse event reports of an index or multiple samples calculated using the target vaccine-adverse event report data. It is also possible to inquire the result with a defined time point.

In the inquiry unit 140 , when a user wants to check various data or data stored in the database unit 140 , the inquiry unit 140 may inquire it.

The data pre-processing unit 120 extracts the analysis target vaccine-adverse event report data, which is the report data corresponding to the analysis target vaccine and the abnormal case of the analysis target vaccine, from the voluntary adverse event report data stored in the database unit 110, and the analysis target Vaccine-adverse event report data is divided into predetermined period units and the number of adverse event reports per period unit is established to preprocess the vaccine-adverse event report data to be analyzed.

The data preprocessing unit 120 removes duplicates due to tracking reports of individual adverse event reports from the voluntary adverse event report data including one or more vaccines belonging to a predetermined search target period and one or more adverse event information reported about the vaccine. Analysis target vaccine-adverse event report data can be extracted.

Since the voluntary adverse event report has a follow-up report function that records logical errors or additional information of individual adverse event reports, the same vaccine and adverse event may be reported twice. Therefore, the final report is selected, excluding overlapping cases.

For example, after an AAA patient is vaccinated, an adverse event of skin rash headache occurs and is registered as individual adverse event report 1, and if an adverse event of vomiting occurs a few days later, report 2 with vomiting added The report is additionally registered. Therefore, the duplication of Report No. 1 and Report No. 2 occurs, and the analysis is based on the final report No. 2 in order to remove this duplication.

In addition, logical errors such as the date of occurrence of adverse events in the past prior to the date of vaccination are excluded.

And, the data preprocessing unit 120 extracts the analysis target vaccine-adverse event report data, which is the report data corresponding to the analysis target vaccine and the abnormal case of the analysis target vaccine.

That is, only the reports corresponding to 'A' in Table 1 below are extracted.

Specific Adverse Events (AEs) All other adverse events (AE ^* ) Target drug A B All other drugs (drug ^* ) C D AE (adverse event): A, B, C, D: the number of reports of adverse events in each column

And, the analysis target vaccine-adverse event report data is divided into predetermined period units and constructed. In other words, it is reconstructed as a temporal range to search or analyze for temporal fluctuations in the analysis target vaccine-adverse event report data. Here, examples of the temporal range may include quarterly, monthly, weekly, or daily, and may be represented by month as shown in Tables 2 and 3 below.

January February In March ... i month Number of AE reports, or average number of cases

...

January February In March ... i month Number of AE reports, or average number of cases 15 cases 20 cases 60 cases ... 18 cases

The change point analysis unit 130 calculates an index for each period unit using the analysis target vaccine-adverse event report data preprocessed by the data preprocessing unit 120 . Then, by using the index for each period unit, the cumulative total width of deviation of the adverse event index by period unit is calculated, and n samples, which are multiple for the vaccine-adverse event report data to be analyzed, are derived, and the deviation of the adverse event index of each sample is calculated. The time of change can be defined by testing the significance of temporal fluctuations in the distribution of the cumulative sum width of . Here, n is a natural number and represents the number of samples.

In addition, the indicators for each period unit include the number of adverse event reports per period unit, an imbalance-based proportional reporting ratio (PRR), the lower limit of the reporting fraction ratio, an imbalance-based reporting odds ratio (ROR), The lower limit of the reporting odds ratio, the information component (IC) of the imbalance-based Bayesian trust transmission neural network, the lower limit of the information component, the average of the number of adverse event reports per period unit within the search target period, the number of adverse event reports per period unit It is desirable that at least any one or more of the cumulative total of deviations and deviations in the number of adverse event reports per period unit is included. In this description, the number of adverse event reports per period unit is the example of the indicator for each period unit.

)를 토대로 하여 각 기간단위 별로, 전체 건수의 평균(

)과 이상사례보고건수 간 편차(

-

)를 구한다. 그리고, 건수편차를 각 기간단위 경과에 따라 누적하여, 각 기간단위 별로 건수편차의 누적합계(각 구간별 누적합계

=

+

-

, CUSUM)를 산출한다.The index calculation for each period unit for analysis of the point of change is, for example, the number of adverse event reports per period unit (monthly or quarterly, as shown in Table 3 below)

) based on the average (

) and the deviation between the number of adverse event reports (

-

) to find And, by accumulating the deviation in the number of cases as each period unit elapses, the cumulative sum of the deviation in the number of cases for each period unit (the cumulative total for each period)

=

+

-

, CUSUM) is calculated.

January February In March ... i month by period unit
Number of AE reports

...

by period unit
deviation in the number of cases

-

(

)

-

-

...

-

Cumulative sum of deviations
(CUSUM)

=

+

-

(only,

)

=

+

-

=

+

-

...

=

+

-

In addition, the index calculated by the change point analysis unit 130 may be visualized by the inquiry unit 140 and displayed so that the user can inquire it.

2 is a diagram schematically showing an example of visualizing the number of adverse event reports by period unit, and FIG. 3 is a diagram schematically showing an example of visualizing the cumulative sum of deviations in the number of adverse event reports by period unit. As shown in FIGS. 2 and 3 , the index calculated by the change point analysis unit 130 may be visualized by the inquiry unit 140 to be inquired by the user.

)의 최댓값(

)과 최솟값(

) 의 차이인 '편차의 누적합계 폭'(

)을 도출한다.The change point analysis unit 130 calculates the cumulative sum of deviations for each period unit (CUSUM,

) of the maximum (

) and the minimum (

), which is the difference between the 'cumulative sum of deviations' (

) is derived.

here,

i: number of ranges in the data (eg quarterly, monthly, weekly, daily)

n: number of samples

: 이상사례 보고자료 건수,

: 이상사례보고건수 평균

: The number of reports of adverse events,

: Average number of adverse event reports

: 편차의 누적합계

: Cumulative sum of deviations

: 편차의 누적합계 최댓값,

: 편차의 누적합계 최솟값

: Maximum cumulative sum of deviations,

: Minimum cumulative sum of deviations

: 편차의 누적합계 폭 이다.

: It is the width of the cumulative sum of deviations.

)를 추출한다. And the change point analysis unit 130 restores and extracts some data from the entire data, and through a bootstrap method of resampling several times, a plurality of samples n (

) is extracted.

For example, if balls 1 to 3 are drawn 3 times and the restoration is performed, the first drawn ball is drawn and put again, the balls 1 to 3 are drawn a second time, and the balls 1 to 3 are drawn a third time. That is, the probability that the balls 1, 2, and 3 will be drawn in all three rounds is equal to 1/3, respectively, and the number of cases is 3x3x3, which is 27 kinds. In this way, a large number of samples are extracted through the sampling restoration method.

4 is a diagram schematically illustrating sampling restoration from spontaneous adverse event report data in the temporal variation time point analysis system of spontaneous vaccine adverse event report according to an embodiment of the present invention.

)을 도출한다. 4, the 'cumulative sum width of deviations' (

) is derived.

: n번째 표본의 편차의 누적합계 폭 이다.)(here,

: It is the width of the cumulative sum of the deviations of the nth sample.)

)의 분포에서, 전체 데이터에서 도출된 편차의 누적합계의 폭(

)을 비교하여 유의확률을 계산한다.And the change point analysis unit 130 is the cumulative sum width of n samples (

), the width of the cumulative sum of deviations derived from all data (

) to calculate the significance probability.

For example, if the significance probability is less than 0.05, which is set as the reference value, it can be said that the data has time-dependent fluctuations.

5 is a diagram schematically showing the cumulative sum width of deviations in each sample as a probability density function in the temporal variation time point analysis system of spontaneous vaccine adverse event reports according to an embodiment of the present invention.

5 , statistical significance is tested under a significance level of 5%.

)인 시점(

)의 다음 시점(

)을 변동시점으로 정의한다.And, when the change point analysis unit 130 is significant in the temporal change in the data, the absolute value of the cumulative sum of the deviations in the entire data is the maximum (

) when (

) at the next point in time (

) is defined as the point of change.

6 is a graph schematically showing the division of the period based on the time of change in the temporal change point analysis system of the spontaneous vaccine adverse event report according to an embodiment of the present invention.

)의 이전 시점(

)까지의 구간1과 변경시점부터 이후까지를 구간2로 정의한다.Temporal section 1 and section 2, which are divided as shown in FIG. 6, are the time of change (

) at the previous point in time (

) is defined as section 1, and from the time of change to after is defined as section 2.

As shown in FIG. 6 , the change point analysis unit 130 repeatedly applies the change time point analysis execution to the data of each section by dividing it into segments before and after the change time point.

As shown in FIG. 6, for each of the temporal section 1 and section 2, the change point analysis is performed as follows.

Assuming that the index for analysis of the point of change for section 1 is calculated as shown in Table 4 below,

January February In March ...

월

month by period unit
Number of AE reports

...

number of deviations
Cumulative total
(CUSUM)

=

+

-

(only,

,

)

=

+

-

=

+

-

...

=

+

-

In temporal section 1, the range of variation by time point is calculated, and n samples are extracted by applying the bootstrap method to the report data of section 1. And, in the distribution of the width of the cumulative sum of n samples, the statistical significance of the width of the cumulative sum of data in section 1 is tested.

)인 시점(

)의 다음 시점(

)을 변동시점으로 정의한다. 그리고, 변동시점을 기준으로 전, 후 구간(segment)을 나누어 각 구간의 데이터에 대해서 상기 변동시점 분석 실시를 반복적으로 적용한다.If it is significant in the significance test, the absolute value of the cumulative sum of deviations in the data of section 1 is the maximum (

) when (

) at the next point in time (

) is defined as the point of change. Then, by dividing the before and after segments (segments) based on the change time point, the change time point analysis execution is repeatedly applied to the data of each section.

Assuming that the index for analysis of the change point for temporal section 2 is calculated as shown in Table 5 below,

+1월

+ January

+February

+March ... i month by period unit
Number of AE reports

...

number of deviations
Cumulative total
(CUSUM)

=

+

-

(only,

,

)

=

+

-

=

+

-

...

=

+

-

In section 2, the range of fluctuation range for each time point is calculated. Then, n samples are extracted by applying the bootstrap method to the report data of section 2. And, in the distribution of the width of the cumulative sum of n samples, the statistical significance of the width of the cumulative sum of the data of section 2 is tested.

)인 시점(

)의 다음 시점(

)을 변동시점으로 정의한다. 그리고, 변동시점을 기준으로 전, 후 구간(segment)을 나누어 각 구간의 데이터에 대해서 변동시점 분석 실시를 반복적으로 실행한다.If it is significant in the significance test, the absolute value of the cumulative sum of deviations in the data of section 2 is the maximum (

) when (

) at the next point in time (

) is defined as the point of change. Then, by dividing the before and after segments (segments) based on the time of change, analysis of the time of change is repeatedly performed on the data of each section.

In this way, the change time analysis is performed in the T1 period (Segment1), which is a time period before the change time, and the change time analysis is also performed on the T2 period (Segment2), which is a time period after the change time.

7 is a graph schematically showing the analysis of the change point by repeatedly applying the change point analysis execution.

, i-2).As shown in FIG. 7 , the variation time analysis unit 130 repeats until the temporal variation is not statistically significant in all sections. That is, the analysis ends when it cannot be said that there is a change in each interval. Therefore, as a result of the analysis of the time of change, it can be derived as k change points and k+1 sections (however, k = 0,

, i-2).

Using the temporal variation timing analysis system for spontaneous vaccine adverse event reporting as above, the temporal variation timing analysis method for spontaneous vaccine adverse event reporting can be implemented as follows.

Next, with reference to FIG. 8, an analysis method using the temporal variation time point analysis system of the spontaneous vaccine adverse event report according to an embodiment of the present invention will be described.

8 is a flowchart schematically illustrating an analysis method using a temporal variation time point analysis system of spontaneous vaccine adverse event report according to an embodiment of the present invention.

Referring to FIG. 8 , the temporal variation timing analysis method of the spontaneous vaccine adverse event report according to an embodiment of the present invention includes a pre-processing step (S10) and a variation timing analysis step (S20).

In the pre-processing step (S10), the analysis target vaccine-adverse event report data reported within a predetermined search target period is extracted from the voluntary adverse event report data including one or more vaccines and one or more adverse event information reported about the vaccine, and It is the step of constructing the analysis target vaccine-adverse event report data divided by the period unit of This pre-processing step includes a duplicate case removal step (S110), an extraction step (S120), and a period unit construction step (S130).

In the variation point analysis step (S20), a plurality of samples are derived for the analysis target vaccine-adverse event report data based on the analysis target vaccine-adverse event report data, which is divided by period in the pre-processing step (S10), and time fluctuations This is the stage to define the point of change according to the significance of

This variation point analysis step (S20) includes an index calculation step (S140), a variation width calculation step (S150), a sampling step (S160), a significance test step (S170), and a variation point definition step (S180).

And, as a step performed after the step of defining the change point (S180), an additional period resetting step (S190) may be further included.

Hereinafter, the pre-processing step (S10) will be described sequentially.

As mentioned above, in the pre-processing step (S10), in the voluntary adverse event report data including one or more vaccines and one or more adverse event information reported with respect to the vaccine within a predetermined search target period, the vaccine to be analyzed and the vaccine to be analyzed Extracting the analysis target vaccine-adverse event report data, which is the report data corresponding to the adverse event of It includes an extraction step (S120) and a period unit construction step (S130).

<<S110>>

Duplicate case removal step (S110) is a step of removing duplicates due to tracking reports of individual adverse event reports from voluntary adverse event report data including one or more vaccines and one or more adverse event information within a predetermined search target period.

The voluntary adverse event report has a follow-up reporting function that records logical errors and additional information of each individual adverse event report, so that the same vaccine and adverse event may be reported twice. Therefore, the final report is selected, excluding duplicate cases.

<<S120>>

Next, in the extraction step (S120), the report data corresponding to the adverse event of the target vaccine and the analysis target vaccine from the voluntary adverse event report data in which the duplication due to the tracking report of the individual adverse event report is removed in the duplicate case removal step (S110) This is the stage of extracting the vaccine-adverse event report data for the analysis of phosphorus.

Extract the vaccine to be analyzed and the relevant adverse events. That is, only the reports corresponding to 'A' in Table 6 below are extracted.

Specific Adverse Events (AEs) All other adverse events (AE ^* ) Target drug A B All other drugs (drug ^* ) C D AE (adverse event): A, B, C, D: the number of reports of adverse events in each column

<<S130>>

Next, the period unit construction step (S130) is a step of dividing the analysis target vaccine-adverse event report data extracted in the extraction step (S120) into a predetermined period unit and constructing it.

For example, it may be constructed as shown in Table 7 below.

2007
Q4 2008
First quarter 2009
Q2 2009
Q3 2009
Q4 ... 2017
Q4 Number of AE reports 1 case 1 case 6 cases 2 cases 0 ... 8 cases

Next, in the variable timing analysis step (S20), a plurality of samples are derived for the analysis target vaccine-adverse event report data based on the analysis target vaccine-adverse event report data constructed by dividing by period in the pre-processing step (S10), and time This is the stage to define the timing of change according to the significance of the change. As mentioned above, this variation point analysis step (S20) is an index calculation step (S140), a variation width calculation step (S150), a sampling step (S160), a significance test step (S170), and a variation point definition step (S180). includes

<<S140>>

The index calculation step (S140) is a step of calculating an index for each period unit by using the analysis target vaccine-adverse event report data constructed by period unit.

The indicators for each period unit calculated in the indicator calculation stage include the number of adverse event reports per period unit, the proportion-based reporting ratio (PRR), the lower limit of the reporting ratio, and the reporting odds ratio based on the imbalance. , ROR), the lower limit of the reporting odds ratio, the information component (IC) of the imbalance-based Bayesian trust transmission neural network, the lower limit of the information component, the average of the number of adverse events reported by period unit within the search target period, anomaly by period unit It is desirable to include at least one of the deviation in the number of case reports and the cumulative total of the deviation in the number of adverse event reports per period unit. In this description, the number of adverse event reports per period unit is the example of the indicator for each period unit.

). 예를 들어, 월별 이상사례보고건수를 산출한다는 것이다.Calculate the number of reports of adverse events by period unit (

). For example, it is to calculate the number of adverse event reports per month.

)과 이상사례보고건수(

) 간 편차(

-

)를 구한다. 그리고, 건수편차를 각 기간단위 경과에 따라 누적하여, 각 기간단위 별로 건수편차의 누적합계(각 구간별 누적합계

=

+

-

, CUSUM)를 산출한다.For each period unit, the average of the total number of cases (

) and the number of adverse event reports (

) between the deviations (

-

) to find And, by accumulating the deviation in the number of cases as each period unit elapses, the cumulative sum of the deviation in the number of cases for each period unit (the cumulative total for each period)

=

+

-

, CUSUM) is calculated.

January February In March ... i month by period unit
Number of AE reports, or average number of cases

...

by period unit
deviation in the number of cases

-

(

)

-

-

...

-

number of deviations
Cumulative total
(CUSUM)

=

+

-

(only,

)

=

+

-

=

+

-

...

=

+

-

)은 5.78이며, 각 분기별 건수와 건수 편차의 누적합계는 다음 표 9와 같다.For example, the average number of reported cases from the fourth quarter of 2007 to the fourth quarter of 2017 (

) is 5.78, and the cumulative sum of the number of cases per quarter and the deviation in the number of cases is shown in Table 9 below.

2007
Q4 2008
First quarter 2009
Q2 ... In 2013
First quarter ... 2017
Q4 Number of AE reports 1 case 1 case 6 cases ... 3 cases ... 8 cases deviation in the number of cases -4.78 -4.78 0.22 ... -2.78 ... 2.22 number of deviations
Cumulative total
(CUSUM) -4.78 -9.56 -9.34 ... -58.17 ... 0

<<S150>>

The fluctuation range calculation step (S150) is a step of calculating the width of the variation in the number of adverse events reported by period unit or the cumulative total width of the deviation in the number of adverse events reported by period unit among the indicators for each period unit calculated in the index calculation step.

)의 최댓값(

)과 최솟값(

)의 차이인 '편차의 누적합계 폭'(

)을 도출한다.For all data, the cumulative sum of deviations calculated for each period unit (CUSUM,

) of the maximum (

) and the minimum (

), which is the 'cumulative sum width of deviations' (

) is derived.

)은 2017년 4분기 0이며, 최솟값(

)은 2013년 1분기 ??58.17이다. 따라서 편차의 누적합계 폭(

)은 58.17이다.Maximum cumulative sum of deviations (

) is 0 in the fourth quarter of 2017, and the minimum value (

) is ??58.17 in the first quarter of 2013. Therefore, the width of the cumulative sum of deviations (

) is 58.17.

<<S160>>

In the sampling step (S160), the width of the variation in the number of adverse event reports by period unit calculated in the index calculation step for each period unit and the variation width calculation step in the index calculation step (S150) or the cumulative total width of the deviation in the number of adverse event reports by period unit It is the step of deriving multiple samples for the vaccine-adverse event report data to be analyzed including

)을 도출한다.Here, for the derivation of a plurality of samples, a plurality of n samples can be derived by using the bootstrap method of restoring and extracting the samples. From all data, using the bootstrap method, extract 1,000 samples and calculate the width (

) is derived.

○ 1st sample

)은 5.02이며, 각 분기별 건수와 건수 편차의 누적합계는 다음과 같다. 편차의 누적합계 최댓값(

)은 2011년 1분기 14.66이며, 최솟값(

)은 2009년 4분기 -19.22이다. 따라서 편차의 누적합계 폭(

)은 33.88이다.The average number of reported cases from Q4 2007 to Q4 2017 in the first sample outlined in Table 10 (

) is 5.02, and the cumulative sum of the number of cases per quarter and the deviation in the number of cases is as follows. Maximum cumulative sum of deviations (

) is 14.66 in the first quarter of 2011, and the minimum value (

) is -19.22 in the fourth quarter of 2009. Therefore, the width of the cumulative sum of deviations (

) is 33.88.

2007
Q4 2008
First quarter 2009
Q2 ... In 2013
First quarter ... 2017
Q4 Number of AE reports 6 cases 1 case 0 ... 7 ... 18 cases deviation in the number of cases 0.98 -4.02 -5.02 ... 1.98 ... 12.98 number of deviations
Cumulative total
(CUSUM) 0.98 -3.05 -8.07 ... -1.54 ... 0

○ 2nd sample

)은 6.44이며, 각 분기별 건수와 건수 편차의 누적합계는 다음과 같다. 편차의 누적합계 최댓값(

)은 2012년 2분기 28.66이며, 최솟값(

)은 2008년 4분기 -12.20이다. 따라서 편차의 누적합계 폭(

)은 40.86이다.The average reported cases from Q4 2007 to Q4 2017 in the second sample outlined in Table 11 (

) is 6.44, and the cumulative sum of the number of cases and deviations in the number of cases for each quarter is as follows. Maximum cumulative sum of deviations (

) is 28.66 in the second quarter of 2012, and the minimum value (

) is -12.20 in the fourth quarter of 2008. Therefore, the width of the cumulative sum of deviations (

) is 40.86.

2007
Q4 2008
First quarter 2009
Q2 ... In 2013
First quarter ... 2017
Q4 Number of AE reports 5 cases 2 cases 10 cases ... 4 cases ... 4 cases deviation in the number of cases -1.44 -4.44 3.56 ... -2.44 ... -2.44 number of deviations
Cumulative total
(CUSUM) -1.44 -5.88 -2.32 ... 14.34 ... 0

In this way, the cumulative sum width of the deviations of the 3rd to 999th samples is calculated.

○ 1,000th sample

)은 4.88이며, 각 분기별 건수와 건수 편차의 누적합계는 다음과 같다. 편차의 누적합계 최댓값(

)은 2009년 4분기 17.1이며, 최솟값(

)은 2013년 2분기 -6.2이다. 따라서 편차의 누적합계 폭(

)은 23.3이다.The average number of reported cases from the fourth quarter of 2007 to the fourth quarter of 2017 in the 1,000th sample outlined in Table 12 (

) is 4.88, and the cumulative sum of the number of cases per quarter and the deviation in the number of cases is as follows. Maximum cumulative sum of deviations (

) is 17.1 in the fourth quarter of 2009, and the minimum value (

) is -6.2 in the second quarter of 2013. Therefore, the width of the cumulative sum of deviations (

) is 23.3.

2007
Q4 2008
First quarter 2009
Q2 ... In 2013
First quarter ... 2017
Q4 Number of AE reports 5 cases 2 cases 10 cases ... 1 case ... 1 case deviation in the number of cases 0.12 -2.88 5.12 ... -3.88 ... -3.88 number of deviations
Cumulative total
(CUSUM) 0.12 -2.76 2.37 ... -4.32 ... 0

<<S170>>

The significance test step (S170) is a step of testing the significance of temporal variations in the distribution of the cumulative sum width of the deviation in the number of adverse event reports of the n samples derived from the sampling step.

It is preferable that the reference value of the significance probability has a value less than 0.05.

)의 분포에서 본 데이터에서 도출된 편차의 누적합계의 폭(

) 58.17의 유의확률을 계산하여 유의수준 5%하에서 통계적 유의성을 검정한다.The width of the cumulative sum of 1,000 samples (

), the width of the cumulative sum of deviations derived from this data (

) Calculate the significance probability of 58.17 and test the statistical significance under the significance level of 5%.

...

The difference between the maximum and minimum values of the cumulative sum of deviations 33.88 40.86 26.71 35.93 68.37 ... 23.3

On the other hand, in the significance test step, the significance of temporal variation is tested, and when the significance probability exceeds the reference value, the analysis of the point of variation is terminated.

<<S180>>

In the step of defining the time of change (S180), the significance of temporal variation is tested in the significance testing step (S180), and if the probability of significance is less than or equal to the reference value, the absolute value of the cumulative sum of deviations in the number of adverse event reports in each of the multiple n samples is the maximum. This is the stage of defining the next point in time as the point of change.

11 is a diagram schematically showing the cumulative sum width of deviations in each sample as a probability density function in the temporal variation time point analysis method of spontaneous vaccine adverse event report according to an embodiment of the present invention.

11 , since the significance probability is 0.01, which is less than the significance level of 0.05, it is statistically significant. Therefore, it can be said that there is variation within the data.

12 is a graph schematically showing defining a change point according to the cumulative sum of deviations in the temporal change point analysis method of spontaneous vaccine adverse event report according to an embodiment of the present invention.

)의 다음 시점인 2013년 2분기(

)를 '변동시점1'로 정의한다.12, the time of change at which the change occurred is the first quarter of 2013 (

) in the second quarter of 2013 (

) is defined as 'change point 1'.

<<S190>>

The period resetting step (S190) is a step of dividing the search target period into a period T1, which is a time period before the change time, and a period T2, a time period after the change time, based on the change time defined in the change time point definition step. to be.

Based on 'Time of Change 1', the temporal segments before and after are divided and the analysis of the time of change is repeatedly performed on the data of each section.

13 is a temporal variation point analysis system of spontaneous vaccine adverse event report according to an embodiment of the present invention, based on the variation point, the interval before the change time is period 1 (segment 1), and the period from the time of change to the period thereafter It is a graph diagram schematically showing division into 2 (segment2).

13 , it is divided into a first section (segment1) from the fourth quarter of 2007 to the first quarter of 2013, and a second section (segment2) from the second quarter of 2013 to the fourth quarter of 2017. And, analysis of the point of change is performed for each section.

The index calculation step and the variation range calculation step as described above by using each of the T1 th period and the T2 th period divided in the period resetting step (S190) as a search target period; Re-perform the sampling step, significance test step, and change time point definition step in sequence. Here, in the significance test step, the significance of temporal variation is tested, and when the significance probability exceeds the reference value, the analysis of the time point of variation is terminated.

Assuming that the index for analysis of the change point in section 1 is calculated as shown in Table 14 below,

2007
Q4 2008
First quarter 2009
Q2 ... In 2013
First quarter Number of AE reports 1 case 1 case 6 cases ... 3 cases deviation in the number of cases -2.14 -2.14 2.86 ... -0.14 number of deviations
Cumulative total
(CUSUM) -2.14 -4.27 -1.41 ... 0

9.37 is obtained by calculating the range of variation by time in section 1, and 1,000 samples are extracted by applying the bootstrap method to the report data of section 1.

And, in the distribution of the width of the cumulative sum of 1,000 samples, the statistical significance of the width of the cumulative sum of the section 1 data is tested. If it is judged not to be significant in the significance test, it cannot be said that there is a change in the data from the 4th quarter of 2007 to the 1st quarter of 2013.

Assuming that the index for analysis of the point of change for section 2 is calculated as shown in Table 15 below,

In 2013
Q2 In 2013
Q3 In 2013
Q4 ... 2017
Q4 Number of AE reports 14 cases 18 cases 3 cases ... 8 cases deviation in the number of cases 5.16 9.16 -5.84 ... -0.84 number of deviations
Cumulative total
(CUSUM) 5.16 14.32 8.47 ... 0

37.9 can be obtained by calculating the range of variation by time in section 2, and 1,000 samples are extracted by applying the bootstrap method to the report data of section 2.

And, in the distribution of the width of the cumulative sum of 1,000 samples, the statistical significance of the width of the cumulative sum of the section 1 data is tested. If it is determined to be significant in the significance test. From the second quarter of 2013 to the fourth quarter of 2017, there was a change in the data. And the fourth quarter of 2014, which is the point after the third quarter of 2014, when the absolute value of the cumulative sum of deviations is the maximum, is defined as the change point.

And, based on the 4th quarter of 2014, it is divided into section 3 from the second quarter of 2013 to the third quarter of 2014 and section 4 from the fourth quarter of 2014 to the fourth quarter of 2017. Repeat until no more.

14 is a graph schematically showing the analysis of the change point by repeatedly applying the change point analysis execution.

As shown in FIG. 14 , as a result of repeatedly applying the change time analysis to the sections divided based on the change time point, statistically significant points were found in the second quarter of 2013, the fourth quarter of 2014, the third quarter of 2016, and 2017. As of the second quarter of the year, 4 change points and 5 sections are derived.

On the other hand, it goes without saying that the technical idea of the present invention can also be applied to a computer-readable recording medium containing a computer program for performing the analysis method according to the present embodiment. In addition, the technical ideas according to various embodiments of the present invention may be implemented in the form of computer-readable programming language codes recorded on a computer-readable recording medium. The computer-readable recording medium may be any data storage device readable by the computer and capable of storing data. For example, the computer-readable recording medium may be a ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical disk, hard disk drive, flash memory, solid state disk (SSD), or the like. In addition, computer-readable codes or programs stored in a computer-readable recording medium may be transmitted through a network connected between computers.

As described above, according to the temporal variation timing analysis system and analysis method of spontaneous vaccine adverse event reports according to an embodiment of the present invention, it is possible to statistically search or analyze a time point showing a difference from the overall aspect of adverse event reports that change over time Therefore, it has the advantage of helping to evaluate a causal relationship through an epidemiological investigation or to take measures such as suspension of vaccination and recall of vaccine products more quickly.

As described above, the detailed description of the present invention has been made by the embodiments with reference to the accompanying drawings, but since the above-described embodiments have only been described with reference to the preferred embodiments of the present invention, the present invention It should not be construed as being limited only to the embodiments, and the scope of the present invention should be understood as the following claims and their equivalents.

100: A system for time-variation analysis of spontaneous vaccine adverse event reporting.
110: database unit 120: data pre-processing unit
130: change point analysis unit 140: inquiry unit

Claims (13)

a database unit for storing voluntary adverse event report data including one or more vaccines and one or more adverse event information reported with respect to the vaccine;
a data preprocessing unit for extracting analysis target vaccine-adverse event report data from the spontaneous adverse event report data stored in the database unit, and constructing the analysis target vaccine-adverse event report data divided by a predetermined period unit; and
Based on the analysis target vaccine-adverse event report data divided by period in the data preprocessing unit, a plurality of samples are derived for the analysis target vaccine-adverse event report data, and the time of change is determined according to the significance of temporal variation. a defining change point analysis unit; characterized in that it comprises,
Temporal variation point analysis system for spontaneous vaccine adverse event reporting.
The method of claim 1,
The data pre-processing unit,
Characterized in extracting the analysis target vaccine-adverse event report data after removing duplication due to follow-up reports of individual adverse event reports from the spontaneous adverse event report data including the one or more vaccines and one or more adverse event information,
Temporal variation point analysis system for spontaneous vaccine adverse event reporting.
3. The method of claim 2,
The change point analysis unit,
By using the analysis target vaccine-adverse event report data pre-processed in the data preprocessing unit, an indicator for each period is calculated, and the cumulative sum width of the indicator deviation for each period unit is calculated using the indicator for each period unit, and the analysis target Vaccine - characterized by deriving a plurality of samples for adverse event report data, and defining the point of change by testing the significance of the temporal variation in the distribution of the cumulative sum width of the index deviation of the plurality of samples,
Temporal variation point analysis system for spontaneous vaccine adverse event reporting. Extracts the target vaccine-adverse event report data reported within a predetermined search target period from the voluntary adverse event report data including one or more vaccines and one or more adverse event information reported about the vaccine, and divides it into a predetermined period unit A pre-processing step of constructing the analysis target vaccine-adverse event report data; and
Based on the analysis target vaccine-adverse event report data constructed separately for each period in the pre-processing step, a plurality of samples are derived for the analysis target vaccine-adverse event report data, and the time of change is defined according to the significance of temporal variation A change point analysis step to
A method of analyzing the temporal variation of spontaneous vaccine adverse event reporting.
5. The method of claim 4,
The change point analysis step is,
an indicator calculation step of calculating an indicator for each period unit using the analysis target vaccine-adverse event report data constructed in the period unit;
a variation width calculation step of calculating the width of the index fluctuation for each period unit or the cumulative sum width of the index deviation for each period unit among the indicators for each period unit calculated in the index calculation step;
The analysis target vaccine including the index for each period unit calculated in the index calculation step and the width of the index change for each period unit calculated in the fluctuation range calculation step or the cumulative sum width of the index deviation for each period unit - Report an adverse event a sampling step of deriving a plurality of samples for the data and deriving the cumulative sum width of the index deviation for each sample;
a significance test step of testing the significance of temporal variation in the distribution of the cumulative sum width of the index deviations of the plurality of samples derived in the sampling step; and
In the significance test step, if the significance probability is less than or equal to the reference value by testing the significance of temporal fluctuations in the significance test step, the time point following the point at which the absolute value of the cumulative sum of the index deviations in each of the plurality of samples is the maximum is defined as the time point of variation as the point of change. characterized in that it contains;
A method of analyzing the temporal variation of spontaneous vaccine adverse event reporting.
6. The method of claim 5,
A period resetting step of dividing the search target period into a period T1 that is a time period before the change time and a period T2 that is a time period after the change time; on the basis of the change time defined in the change time definition step; characterized in that it comprises
A method of analyzing the temporal variation of spontaneous vaccine adverse event reporting.
7. The method of claim 6,
The pre-processing step is
Duplicate case removal step of removing duplicates due to tracking reports of individual adverse event reports from voluntary adverse event report data including one or more vaccines and one or more adverse event information within a predetermined search target period;
The analysis target vaccine-adverse event, which is the report data corresponding to the adverse event of the target vaccine and the target vaccine, in the voluntary adverse event report data from which the duplicate due to the follow-up report of the individual adverse event report was removed in the duplicate case removal step an extraction step of extracting the report data; and
A period unit construction step of dividing the analysis target vaccine-adverse event report data extracted in the extraction step into a predetermined period unit and constructing;
A method of analyzing the temporal variation of spontaneous vaccine adverse event reporting.
8. The method of claim 7,
using each of the T1 th period and the T2 th period divided in the period resetting step as the search target period, the index calculation step, the variation width calculation step; the sampling step; the significance test step; and the step of defining the change time point; characterized in that it is sequentially re-performed,
A method of analyzing the temporal variation of spontaneous vaccine adverse event reporting.
9. The method of claim 8,
In the indicator for each period unit calculated in the indicator calculation step, the number of adverse event reports per period unit, an imbalance-based reporting ratio (PRR), a lower limit of the reporting fraction ratio, and an imbalance-based reporting odds ratio (reporting odds ratio, ROR), the lower limit of the reporting odds ratio, the information component (IC) of the imbalance-based Bayesian trust transmission neural network, the lower limit of the information component, the report of anomalies by the period unit within the search target period characterized in that at least one or more of the average of the number of cases, the deviation of the number of adverse event reports per period unit, and the cumulative sum of the deviation of the number of adverse event reports per period unit is included,
A method of analyzing the temporal variation of spontaneous vaccine adverse event reporting.
10. The method of claim 9,
The derivation of the plurality of samples in the sampling step is,
Characterized in deriving the plurality of samples by using a bootstrap method for restoring and extracting samples,
A method of analyzing the temporal variation of spontaneous vaccine adverse event reporting.
11. The method of claim 10,
The reference value of the significance probability is characterized in that it has a value less than 0.05,
A method of analyzing the temporal variation of spontaneous vaccine adverse event reporting.
12. The method of claim 11,
In the significance test step, when the significance of the temporal variation is tested for and the significance probability exceeds the reference value, the analysis at the time of variation is terminated,
A method of analyzing the temporal variation of spontaneous vaccine adverse event reporting.
A computer-readable recording medium in which a program capable of executing the voluntary vaccine adverse event analysis method according to any one of claims 4 to 12 is recorded.

Images