Information processing device, information processing method, and program

Abstract

To provide technology that allows for the prediction of how effective a promotional campaign will be for different target groups. [Solution] The policy result acquisition unit 51 performs a process to acquire information obtained as a result of the implementation of a promotional policy targeting at least one of doctors or patients, as policy result information for each of the one or more policies. The policy effect prediction unit 52 performs a process to predict the policy effect based on the policy result information for each of the one or more policies acquired by the policy result acquisition unit 51. For a predetermined policy among the one or more policies, the policy result acquisition unit 51 takes the time when the policy is implemented after improvement considerations have been made based on the prediction of the previous policy effect as the current time, and performs a process to acquire the information obtained as a result of the current implementation as the policy result information for the current time. The policy effect prediction unit 52 then performs a process to predict the policy effect of the predetermined policy for the current time based on the policy result information for the current time.

Description

Technical Field

[0001] The present invention relates to an information processing apparatus, an information processing method, and a program.

Background Art

[0002] Conventionally, as a system for supporting the business activities of MR (Medical Representatives) personnel (medical information personnel) who belong to, for example, a pharmaceutical company and conduct sales and information transmission of the company's own pharmaceutical products to medical personnel such as doctors and pharmacists, the technique disclosed in the following Patent Document 1 is known.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In pharmaceutical marketing, promotion to doctors is an important activity, and by promoting to doctors, it is expected to have an effect of changing the behavior of doctors to prescribe the drug. Examples of promotion include face-to-face explanations by MRs and browsing WEB media. In addition, by promoting to patients, it is expected to increase the frequency of medical visits and continue treatment. However, until now, it has remained unknown which promotions are effective and which specific measures are effective for which promotion targets. Therefore, there has been a situation where the needs to predict in advance how effective promotion measures are for which targets have not been met.

[0005] This invention was made in view of these circumstances, and aims to provide a technology that makes it possible to predict in advance which target audience a promotional measure will be effective for and to what extent. [Means for solving the problem]

[0006] To achieve the above objective, an information processing apparatus according to one aspect of the present invention is: A means for acquiring policy results, which acquires information obtained as a result of implementing promotional measures targeting at least one of doctors or patients, as policy result information for each of the one or more measures, For each of the one or more measures described above, a measure effect prediction means is provided to predict the effect of the measure based on the measure result information, It is equipped with.

[0007] An information processing method and program corresponding to the above-mentioned information processing apparatus according to one aspect of the present invention are also provided as an information processing method and program according to one aspect of the present invention. [Effects of the Invention]

[0008] According to the present invention, it is possible to predict in advance which target groups will benefit from a promotional measure and to what extent. [Brief explanation of the drawing]

[0009] [Figure 1] This figure shows an overview of the service that can be realized by an information processing system to which a server according to one embodiment of the information processing device of the present invention is applied. [Figure 2] This is a block diagram showing an example of the hardware configuration of a server according to one embodiment of the information processing apparatus of the present invention. [Figure 3] Figure 2 is a functional block diagram showing an example of the server's functional configuration. [Figure 4] This figure shows an example of how to predict the effectiveness of a promotion (estimating the expected impact based on past performance). [Figure 5]This figure shows an example of predicting the effectiveness of a promotion (estimating the expected impact using machine learning). [Figure 6] This figure shows an example of the implementation of promotional measures and the continuous updating of machine learning models. [Figure 7] This figure shows an example of how to estimate the monetary effect of implementing promotional measures. [Figure 8] This figure shows an example of considering policy improvements based on the results of factor analysis. [Modes for carrying out the invention]

[0010] Embodiments of the present invention will be described below with reference to the drawings.

[0011] Figure 1 shows an overview of the service that can be realized by an information processing system to which a server according to one embodiment of the information processing device of the present invention is applied.

[0012] This service relates to predicting the effectiveness of promotional strategies targeting physicians and patients in pharmaceutical marketing. This service allows you to predict in advance how effective a promotional campaign will be for which target audience. In other words, it can meet the need to predict in advance how effective a promotional campaign will be and to what extent, targeting specific audiences.

[0013] One approach to this service is to, for example, "predict the effectiveness of promotions" and then "execute promotional measures and continuously update the machine learning model." Regarding the former, "predicting the promotional effect," two methods can be used to predict (estimate) the expected impact, for example:

[0014] The first method is to predict (estimate) the expected impact based on past promotional examples (based on past performance). As a second method, there is a method of predicting (estimating) the expected impact through machine learning using sales data, promotion log data, and physician attribute data. Details of these two methods will be described later.

[0015] In the latter "Execution of promotion measures and continuous update of machine learning models", by continuously performing the cycle of measure execution, data acquisition, modeling, and consideration of improvement measures, it is possible to improve the effectiveness of promotion measures and the prediction accuracy of the machine learning model KM described later. Details of the latter "Execution of promotion measures and continuous update of machine learning models" will be described later.

[0016] As shown in the output information OJ of FIG. 1, in this service, based on the prediction effectiveness rates for various measures for physicians and patients, it is possible to output the predicted effect amount, investment amount, and ROI (Return on Investment). That is, in this service, as shown in the output information OJ, first, when the measure category is for physicians, for a measure such as "physician targeting based on the number of existing patients", for example, its prediction effectiveness rate is "+X%", the predicted effect amount is "X million yen", the investment amount is "X million yen", and the ROI is "XX", which is output by the server 1 described later. Also, in the case of measures for physicians, for measures such as "differentiated appeals based on prescription priorities", "measures to improve the corporate brand", "differentiated appeals based on the possibility of prescription change", and "resource allocation between real and digital", for example, their prediction effectiveness rates, predicted effect amounts, investment amounts, and ROIs are output as described above. On the other hand, next, when the measure category is for patients, for a measure such as "measures to promote medical treatment behavior", for example, its prediction effectiveness rate is "+X%", the predicted effect amount is "X million yen", the investment amount is "X million yen", and the ROI is "XX", which is output in this way. Also, in the case of measures for patients, for a measure such as "measures to improve adherence", for example, its prediction effectiveness rate, predicted effect amount, investment amount, and ROI are output as described above.

[0017] As can be seen from the output information OJ shown in Figure 1, this service makes it possible to predict in advance which target audiences a promotional campaign will be effective for and to what extent. In other words, it can meet the need to predict in advance how effective a promotional campaign will be and to what extent, targeting specific audiences.

[0018] Referring to Figure 2, the hardware configuration of server 1 according to one embodiment of the information processing device of the present invention will be described. Figure 2 is a block diagram showing an example of the hardware configuration of a server according to one embodiment of the information processing apparatus of the present invention.

[0019] Server 1 comprises a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, a bus 14, an input / output interface 15, an input unit 16, an output unit 17, a storage unit 18, a communication unit 19, and a drive 20.

[0020] The CPU 11 executes various processes according to the program recorded in the ROM 12 or the program loaded into the RAM 13 from the storage unit 18. RAM13 also stores data and other information necessary for the CPU11 to perform various processes.

[0021] The CPU 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output interface 15 is also connected to this bus 14. An input / output interface 15 is connected to an input unit 16, an output unit 17, a storage unit 18, a communication unit 19, and a drive 20.

[0022] The input unit 16 consists of various hardware buttons and the like, and inputs various information according to the operator's instructions and operations. The output unit 17 is composed of a display such as an LCD and displays various images.

[0023] The memory unit 18 is composed of DRAM (Dynamic Random Access Memory) and stores various types of data. The communications unit 19 controls communication with other devices (e.g., physician information systems, pharmaceutical company systems, MR personnel terminals, etc.) via a network including the Internet.

[0024] A drive 20 is provided as needed. A removable media 30, consisting of a magnetic disk, optical disk, magneto-optical disk, or semiconductor memory, is appropriately mounted on the drive 20. Programs read from the removable media 30 by the drive 20 are installed in the storage unit 18 as needed. The removable media 30 can also store various data stored in the storage unit 18, just like the storage unit 18.

[0025] Although not shown in the diagram, the aforementioned physician information system, pharmaceutical company system, MR (Medical Representative) terminals, marketing staff terminals, etc., can also have a configuration that is basically the same as the hardware configuration in Figure 2. Therefore, a description of the configuration of the physician information system, pharmaceutical company system, MR terminals, marketing staff terminals, etc. will be omitted.

[0026] Through the collaboration of the various hardware and software components of Server 1 shown in Figure 2, Server 1 becomes capable of executing the various processes described later. As a result, the service provider of this service can provide the various services.

[0027] Next, we will explain the functional configuration of Server 1 in Figure 2, referring to Figure 3. Figure 3 is a functional block diagram showing an example of the functional configuration of the server in Figure 2.

[0028] As shown in Figure 3, the CPU 11 of server 1 functions as follows: a policy result acquisition unit 51, a policy effect prediction unit 52, and a learning unit 53. Furthermore, one area of ​​the storage unit 18 of server 1 is provided with the policy results DB 71, the policy effectiveness DB 72, and the machine learning model KM.

[0029] The policy result acquisition unit 51 in the CPU 11 executes a process to acquire information obtained as a result of the implementation of a promotion policy targeting at least one of doctors or patients, as policy result information for each policy. The policy result information obtained for each of the one or more policies is stored in the policy result DB 71 by the policy result acquisition unit 51. Furthermore, the policy results DB71 will also store promotional case studies, which will be described later, as policy results information.

[0030] The policy effect prediction unit 52 in the CPU 11 executes a process to predict the policy effect for each of the one or more policies acquired by the policy result acquisition unit 51, based on the policy result information described above. The predicted effects of the measures are stored in the measure effect DB 72 by the measure effect prediction unit 52.

[0031] As described above, the functioning of the policy result acquisition unit 51 and the policy effect prediction unit 52 makes it possible to predict in advance which target groups will be affected by a promotional measure and to what extent. In other words, it can meet the need to predict in advance how effective a promotional campaign will be and to what extent, targeting specific audiences.

[0032] The policy result acquisition unit 51 and the policy effect prediction unit 52 also perform the following processing. In other words, the policy result acquisition unit 51 selects one or more policies, and for each policy, it takes the instance in which the policy was implemented after improvement studies were conducted based on predictions of the previous policy's effectiveness, and acquires the information obtained as a result of this instance of implementation as the policy result information for that instance. Furthermore, the policy effect prediction unit 52 performs a process to predict the effect of a predetermined policy based on the policy result information. The results of this initiative will be stored in the Initiative Results DB71. Additionally, the effectiveness of this initiative will be stored in the Initiative Effectiveness DB72.

[0033] The effectiveness of promotional measures can be enhanced by further functioning of the policy result acquisition unit 51 and the policy effect prediction unit 52. Furthermore, it is possible to improve the prediction accuracy of the machine learning model KM, which will be discussed later.

[0034] The policy effect prediction unit 52 performs a process for each policy, predicting the policy effect according to a predetermined algorithm based on past cases (based on past performance) (see Figures 4 and 7 below). Furthermore, the policy effect prediction unit 52 performs a process to predict the policy effect of a given policy based on the results of a predetermined machine learning operation performed using training data that includes policy result information for the given policy and past policy result information (as described later with reference to Figures 5 and 7). Furthermore, the policy effect prediction unit 52 performs a process to generate improvement measures as shown in the policy improvement measures SK described later, referring to Figure 8. The generated improvement measures are stored, for example, in the Policy Effectiveness DB72 (or, for example, in the Improvement Measures DB which is not shown).

[0035] The learning unit 53 in the CPU 11 performs predetermined machine learning using training data that includes information on past policy results. In this embodiment, past sales data, promotion log data, and physician attribute data for multiple physicians are also used as training data. The learning unit 53 performs a process to update the machine learning model KM as needed.

[0036] Next, we will explain how to predict the promotional effect, referring to Figures 4 and 5. Figure 4 shows an example of how to predict the effectiveness of a promotion (estimating the expected impact based on past performance). Figure 4 explains the first method of predicting (estimating) the expected impact as described above as an approach for this service (a method of predicting (estimating) the expected impact based on past promotion examples (based on past performance)).

[0037] In the effect prediction KY shown in Figure 4, past promotion examples (considered as just one example) are as follows: Specifically, there was an example of a policy called "Targeting physicians based on the number of patients they treat" (an example obtained by the policy results acquisition unit 51), and in the case of drug A, the effectiveness rate was "+5%". Furthermore, the efficacy rate for drug B was "+4%", while the efficacy rate for drug C was "+3%". Given this example, if we apply the rules of this embodiment (the rule of basing on past performance), for instance, since the effectiveness rate of similar drug A is +5%, the effectiveness rate for the measure of "targeting doctors based on the number of patients" is predicted to be +5% in this case. The predicted effectiveness rate of +5% will be stored in the policy effectiveness DB 72 by the policy effectiveness prediction unit 52. An example of predicting (estimating) the expected impact, such as the amount of effect, based on the effectiveness rate will be discussed later with reference to Figure 7.

[0038] Figure 5 shows an example of how to predict the effectiveness of a promotion (estimating the expected impact using machine learning). Figure 5 explains the second method for predicting (estimating) the expected impact as described above as an approach for this service (a method that uses machine learning to predict (estimate) the expected impact using sales data, promotion log data, and physician attribute data).

[0039] In Figure 5, first, in the model building MK, which shows the model building process, the machine learning model KM is built by the learning unit 53 using sales data by physician, promotion log data, and physician attribute data. Next, in the model application MT, which indicates the time when the model is applied, the number of promotions and physician attribute data are applied to the machine learning model KM, and the policy effect prediction unit 52 of server 1 functions and outputs a sales forecast value. In other words, it is predicted how much impact each promotion will have if it is implemented for each physician.

[0040] In this embodiment, a standard machine learning modeling algorithm is used (a representative method such as a decision tree among regression and classification algorithms is preferred (because it has the advantages of enabling highly accurate modeling for tabular data, fast processing, distributed processing, and the ability to calculate importance in predictions, resulting in high explainability and interpretability)). Furthermore, as a precision verification method, this embodiment employs a standard method similar to that used for the modeling algorithm (a typical method is preferred, which involves splitting the dataset into training and evaluation sets and swapping them (for example, creating multiple patterns for splitting the entire dataset into training and evaluation sets. Specifically, patterns with precision levels from 1 to 5 are created, and their average value is used)).

[0041] Next, we will explain the implementation of promotional measures and the continuous updating of the machine learning model KM, referring to Figure 6. Figure 6 shows an example of the implementation of promotional measures and the continuous updating of the machine learning model KM.

[0042] Figure 6, the continuous cycle KS, illustrates the execution of promotional measures and the continuous updating of the machine learning model KM. For example, in the policy implementation step S1, the policy is implemented for doctors and patients, and the data obtained from the implementation of this policy is acquired in the data acquisition step S2. The data acquired in data acquisition step S2 will include not only survey data and sales data, but also prescription data and medical data from real-world data (RWD) obtained from medical treatment, for example. The data acquired in data acquisition step S2 is used in modeling step S3, and sales data and factor analysis results are output, such as the effect amount related to the expected impact mentioned above. Based on these output results, for example, in improvement measure consideration step S4, improvement measures are considered and reflected in the implementation of the next measures. As shown in Figure 6, the continuous cycle KS aims to improve the effectiveness of promotional measures and the predictive accuracy of the machine learning model KM by continuously performing the cycle of implementing measures, acquiring data, modeling, and considering improvement measures, which is essentially a PDCA cycle.

[0043] Next, we will explain how to estimate the monetary effect of implementing promotional measures, referring to Figure 7. Figure 7 shows an example of how to estimate the monetary effect of implementing promotional measures.

[0044] Regarding the estimation of the effect amount from the implementation of promotional measures, as shown in the effect amount KG in Figure 7, in the case-based JB, the effect amount is estimated based on the effectiveness rate of the measures and the number of target physicians. For example, if the effectiveness rate of "Measure A" is "+4%" and the number of target doctors is "30,000", the estimated effect is "1 million yen". And if the investment amount is "20 million yen", the ROI will be "5%". Here, as an example, let me explain in more detail the estimated effect of "100 million yen" mentioned above. The effect amount is estimated based on the sales volume of the drug, the number of target physicians, and the effectiveness rate. For example, if a drug has sales of 25,000 million yen when prescribed by 300,000 physicians, and the number of target physicians is set at, for example, 30,000 (10% of 300,000 is 30,000 (10% of sales of 25,000 million yen is 2,500 million yen)), then the effect amount would be sales * effectiveness rate = 2,500 * 0.04 = 100 million yen. On the other hand, as shown in the effect amount KG in Figure 7, in the model-based MB, the machine learning model KM is used to estimate the effect amount. For example, in the case of "Measure A," similar to the example-based JB mentioned above, the estimated effect is "1 million yen." If the investment is "20 million yen," the ROI will be "5%." As can be seen from the estimated effect amount in KG in Figure 7, Server 1 can predict in advance which target groups will benefit from a promotional measure and to what extent.

[0045] Next, referring to Figure 8, we will explain how to consider whether to continue, the amount of investment for the next period, and improvement measures based on the estimated effect amounts for each measure. Figure 8 shows an example of considering policy improvements based on the results of factor analysis.

[0046] Figure 8, which shows the proposed policy improvements SK, displays the improvement measures for each policy, generated by the policy effect prediction unit 52 of server 1. In this embodiment, the improvement measures are generated by the policy effectiveness prediction unit 52 based on the results of a survey (such as the survey data to be acquired as described above), and outputted. For example, in the example in Figure 8, the estimated effect amount (predicted effect amount) of measure A for doctors is calculated to be "100 million yen," so this measure is judged to be "continued," and the next investment amount is "5 million yen," and the result of the review of improvement measures is judged to "continue the current measure." On the other hand, in the case of patient-oriented measure E, the estimated effect amount (predicted effect amount) is calculated to be "100 million yen," so it is judged that this measure should be "abolished," and the next investment amount is "0 million yen," and the result of the review of improvement measures is that it should be "abolished."

[0047] Although embodiments of the present invention have been described above, the present invention is not limited to the embodiments described above. Furthermore, the effects described in these embodiments are merely a list of the most preferred effects arising from the present invention, and the effects of the present invention are not limited to those described in these embodiments.

[0048] For example, the above series of processes can be executed by hardware or by software. In other words, the functional configuration shown in Figure 3 is merely an example and is not particularly limited to the functional configuration shown in Figure 3. In other words, it is sufficient to have a function that can execute the above series of processes as a whole, and the type of functional block used to realize this function is not particularly limited to the examples in Figures 1 to 8. Furthermore, the location of the functional blocks is not particularly limited to Figures 1 to 8 and can be any location. For example, the functional blocks and database of Server 1 may be transferred to a physician information system, a pharmaceutical company system, a medical representative terminal, a marketing staff terminal, etc. Conversely, the functional blocks and database of a physician information system, a pharmaceutical company system, a medical representative terminal, a marketing staff terminal, etc. may be transferred to a server, etc. Furthermore, a single functional block and database may consist of hardware alone, software alone, or a combination of both.

[0049] When a series of processes are executed by software, the programs that make up that software are installed on a computer or other device from a network or storage medium. The computer may be a computer that is built into dedicated hardware. Furthermore, a computer can be any computer capable of performing various functions by installing various programs, such as a server, a general-purpose smartphone, or a personal computer.

[0050] Such recording media containing programs may consist not only of removable media (not shown) distributed separately from the main unit to provide programs to users, but also of recording media provided to users, etc., that are pre-installed in the main unit.

[0051] In this specification, the step of describing a program to be recorded on a recording medium includes not only processes that are performed chronologically in that order, but also processes that are not necessarily performed chronologically, but are executed in parallel or individually. Furthermore, in this specification, the term "system" refers to an overall system composed of multiple devices, means, etc.

[0052] In other words, the information processing device to which the present invention applies only needs to have the following configuration, and can take various forms. In other words, the information processing device to which the present invention is applied (for example, Server 1 in Figure 3) is: A means for acquiring policy results (e.g., the policy result acquisition unit 51 in Figure 3) acquires information obtained as a result of implementing promotional measures targeting at least one of the following: doctors (e.g., the group of doctors in Figure 6) and patients (e.g., the group of patients in Figure 6), for each of one or more measures (e.g., multiple measures shown in the policy name column in Figure 1), as policy result information, For each of the one or more measures mentioned above, there is a measure effect prediction means (for example, the measure effect prediction unit 52 in Figure 3) that predicts the measure effect (for example, the numerical values ​​shown in the columns for predicted effect amount, investment amount, ROI in Figure 1) based on the measure result information, It is equipped with.

[0053] According to the present invention, it is possible to predict in advance which target groups will benefit from a promotional measure and to what extent.

[0054] In the above-mentioned information processing device (for example, Server 1 in Figure 3), The means for acquiring the results of the measures described above acquires information obtained as the results of the measures for the current time for a predetermined measure among the one or more measures described above, after an improvement study has been conducted based on the prediction of the effect of the previous measure, and this is taken as the current time (for example, acquisition during a PDCA cycle as shown in the continuous cycle KS in Figure 6). The measure effect prediction means predicts the effect of the predetermined measure based on the information regarding the results of the measure in the current instance (for example, predicting the effect of the measure while running a PDCA cycle, as shown in the continuous cycle KS in Figure 6). It is possible.

[0055] In the above-mentioned information processing device (for example, Server 1 in Figure 3), The aforementioned policy effect prediction means is For each of the one or more measures mentioned above, the effect of the measure is predicted according to a predetermined algorithm based on past performance (see, for example, the rules shown in Figure 4, Effect Prediction KY) (see, for example, Model-Based MB in Figure 7). It is possible.

[0056] In the above-mentioned information processing device (for example, Server 1 in Figure 3), The aforementioned policy effect prediction means is Based on the results of a predetermined machine learning operation performed using training data that includes the results information of the predetermined measures and past results information of the predetermined measures, the effectiveness of the predetermined measures is predicted (see, for example, the model building MK and model application MT in Figure 5). It is possible.

[0057] The above-mentioned information processing device (for example, Server 1 in Figure 3) A learning means (for example, the learning unit 53 in Figure 3) that performs the predetermined machine learning (for example, see the model building MK in Figure 5) using the learning data which includes the past results of the aforementioned measures, It can provide even more. [Explanation of symbols]

[0058] 1...Server, 11...CPU, 18...Storage unit, 51...Measure result acquisition unit, 52...Measure effect prediction unit, 53...Learning unit, 71...Measure result DB, 72...Measure effect, OJ...Output information, KY...Effect prediction, MK...Model construction, MT...Model application, KM...Machine learning model, KS...Continuous cycle, JB...Case-based, MB...Model-based, KG...Effect amount, SK...Measure improvement measures

Claims

1. A means for acquiring policy results, which acquires information obtained as a result of implementing promotional measures targeting at least one of doctors or patients, as policy result information for each of the one or more measures, For each of the one or more measures described above, there is a measure effect prediction means that predicts the effect of the measure based on the measure result information, An information processing device equipped with the following features.

2. The means for acquiring the results of the measures described above acquires information obtained as the results of the measures described above for the current implementation of a predetermined measure among the one or more measures described above, after an improvement study has been conducted based on the prediction of the effect of the previous measure, and the implementation of that measure is considered the current implementation. The measure effect prediction means predicts the effect of the predetermined measure based on the information regarding the results of the measure in the current instance. The information processing apparatus according to claim 1.

3. The aforementioned policy effect prediction means is For each of the one or more measures mentioned above, the effect of the measure is predicted according to a predetermined algorithm based on past performance. The information processing apparatus according to claim 1.

4. The aforementioned policy effect prediction means is Based on the results of a predetermined machine learning operation performed using training data that includes the results of the predetermined measures and past results of the predetermined measures, the effect of the predetermined measures is predicted. The information processing apparatus according to claim 1.

5. A learning means that performs the predetermined machine learning using the learning data which includes past policy result information, The information processing apparatus according to claim 4, further comprising:

6. In an information processing method executed by an information processing device, A step to acquire policy results, in which information obtained as a result of implementing promotional measures targeting at least one of doctors or patients is acquired as policy result information for each of the one or more measures, For each of the one or more measures described above, a measure effect prediction step is performed to predict the effect of the measure based on the measure result information, Information processing methods including

7. On the computer, A step to acquire policy results, in which information obtained as a result of implementing promotional measures targeting at least one of doctors or patients is acquired as policy result information for each of the one or more measures, For each of the one or more measures described above, a measure effect prediction step is performed to predict the effect of the measure based on the measure result information, A program that executes control processes, including those mentioned above.

Images