A method for accurately evaluating drug causal utility in clinical trials

Abstract

The present application relates to a kind of clinical trial drug causal effect accurate evaluation method, belong to clinical trial drug evaluation technical field, solve the problem of lack of unified accurate evaluation method for all real world drug clinical test scene in prior art.The method includes the following steps: according to the clinical test problem of interest, determine the concomitant event type, establish concomitant event dataset;Based on the clinical test problem of interest, determine the treatment intervention measure attribute, and establish treatment intervention measure attribute dataset;Based on different treatment intervention measures in treatment intervention measure dataset, establish population attribute dataset;Based on treatment intervention measure dataset and population attribute dataset, establish potential outcome variable attribute dataset;Based on potential outcome variable attribute dataset, establish outcome variable effectiveness statistical attribute dataset in target population.It is realized that for all real world drug clinical scene, a kind of unified accurate evaluation method is provided.

Description

Technical Field

[0001] This invention relates to the field of clinical trial drug evaluation technology, and in particular to a precise method for evaluating the causal efficacy of clinical trial drugs. Background Technology

[0002] The efficacy evaluation method for clinical drug trials should be established before data collection and analysis. This method should not be changed during data analysis. Clinical trial estimates are easily influenced by the estimation method used during data analysis, deviating from the original estimated target and leading to misinterpretations of drug efficacy. A major reason for this deviation is that the handling of comorbidities during clinical trials was not considered when establishing the efficacy evaluation method during the clinical trial design phase.

[0003] The supplementary document to the ICH E9(R1) guidelines for international clinical trials proposes five strategies for addressing comorbidities. Different types of comorbidities require different strategies when establishing methods for evaluating drug efficacy.

[0004] In response, existing studies have listed some efficacy evaluation targets for clinical trials based on ICH E9(R1) and targeting an unlimited sample population. However, these studies directly present the final efficacy evaluation expression function without providing guidance on how to progressively construct efficacy evaluation methods under various comorbidity management strategies. Furthermore, existing studies only provide efficacy evaluation expression functions for specific scenarios, not for all real-world drug clinical trial scenarios, thus lacking universality. Summary of the Invention

[0005] Based on the above analysis, the embodiments of the present invention aim to provide a precise evaluation method for the causal efficacy of drugs in clinical trials, in order to solve the problem of the lack of a unified and precise evaluation method for all real-world drug clinical trial scenarios.

[0006] On one hand, embodiments of the present invention provide a method for precise evaluation of the causal efficacy of drugs in clinical trials, comprising the following steps:

[0007] Step S0: Based on the clinical trial issues of interest, determine the types of comorbid events and establish a comorbid event dataset;

[0008] Step S1: Based on the clinical trial questions of interest, determine the attributes of treatment interventions and establish a dataset of treatment intervention attributes;

[0009] Step S2: Based on the different treatment interventions in the treatment intervention dataset, establish a population attribute dataset;

[0010] Step S3: Based on the treatment intervention dataset and the population attribute dataset, establish a potential outcome variable attribute dataset;

[0011] Step S4: Based on the potential outcome variable attribute dataset, establish a statistical attribute dataset of the effectiveness of the outcome variable in the target population.

[0012] Furthermore, the intervention attribute dataset in step S1 includes placebo intervention and clinical drug intervention under comorbid events.

[0013] Furthermore, the population attribute dataset in step S2 includes a target population, which is a sub-population determined based on the clinical drug treatment effects of interest.

[0014] Furthermore, step S3 includes the following steps:

[0015] S301: Based on the comorbidity event dataset and the treatment intervention attribute dataset, establish a dataset of potential outcome variables under a composite strategy for real-world scenarios where there are factors that affect the interpretation of outcome variables;

[0016] S302: Based on the potential outcome variable dataset established in step S301, for real-world scenarios where there is no influence on the interpretation of outcome variables, establish a potential outcome variable dataset that incorporates the treatment strategy.

[0017] S303: Based on the potential outcome variable dataset established in step S302, for hypothetical scenarios, establish a potential outcome variable dataset that combines hypothetical strategies. The outcome variable dataset is a dataset that contains the final potential outcome variables.

[0018] Furthermore, for the evaluation of drug causal utility with individual-level causal explanation, step S4 includes the following steps:

[0019] S401: Based on the potential outcome attribute dataset obtained in step S3, establish an individual-level drug causal efficacy evaluation dataset;

[0020] S402: Based on the individual-level drug causal efficacy evaluation dataset established in step S401, establish a drug efficacy dataset for the target population.

[0021] The drug efficacy dataset established in step S402 includes target data for estimating drug causal utility.

[0022] Furthermore, for the evaluation of drug causal utility where there is no individual-level causal explanation, step S4 includes establishing a drug efficacy dataset for the target population based on the potential outcome attribute dataset described in step S3.

[0023] Furthermore, the latent outcome variable data in the latent outcome variable dataset established in step S4 satisfy the following:

[0024] Y i =W i Y i S302 (1)+(1-W i )·Y i S302 (0).

[0025] Where i represents the individual subject; Y i Outcome variables observed in real-world clinical trials; W i The actual treatment intervention received by subject individual i; Y i S302 (1) Y is a potential outcome variable for receiving clinical medication in step S302; i S302 (0) is a potential outcome variable for receiving placebo in step S302.

[0026] Furthermore, the outcome variables in the outcome variable dataset established in step S303 satisfy the following:

[0027]

[0028]

[0029] Where N is the total number of subjects and i is the number of individual subjects; τ i The target for estimating the causal utility of drugs at the individual level; Y i S3 (1) is a potential outcome variable under clinical drug treatment; Y i S3 (0) represents the potential outcome variable under placebo treatment; l represents the individual subject, l≠i; Y l S302 (1) Y represents the potential outcome variable for receiving clinical medication, as established in step S302. l S302 (0) represents the potential outcome variable for placebo administration established in step S302; S l (1) = 0 indicates that the concomitant event of the hypothetical strategy applicable to step S303 did not occur when receiving clinical medication; S l (0) = 0 indicates that the co-occurrence event of the hypothetical strategy applied in step S303 did not occur when the placebo was received; f(Y l S302 (1)|S l(1) = 0, l ≠ il = 1, ..., N) represents the potential outcome variable in the scenario where the concomitant events of the hypothetical strategy applicable to step S303 do not occur when receiving clinical drugs; f(Y) l S302 (0)|S l (0) = 0, l ≠ i, l1, ..., N) are potential outcome variables in the scenario where the concomitant events of the hypothetical strategy applicable to step S303 do not occur when the placebo is received.

[0030] Furthermore, the drug causal efficacy evaluation data in the individual-level drug causal efficacy evaluation dataset established in step S401 satisfies:

[0031] τ i =Y i S3 (1)-Y i S3 (0)

[0032] in,

[0033]

[0034]

[0035] Where N is the total number of subjects and i is the number of individual subjects; τ i The target for estimating the causal utility of drugs at the individual level; Y i S3 (1) is a potential outcome variable under clinical drug treatment; Y i S3 (0) represents the potential outcome variable under placebo treatment; l represents the individual subject, l≠i; Y l S302 (1) Y represents the potential outcome variable for receiving clinical medication, as established in step S302. l S302 (0) represents the potential outcome variable for placebo administration established in step S302; S l (1) = 0 indicates that the concomitant event of the hypothetical strategy applicable to step S303 did not occur when receiving clinical medication; S l (0) = 0 indicates that the concomitant event of the hypothetical strategy applied in step S303 did not occur when the placebo was received; f(Y l S302 (1)|S l (1) = 0, l ≠ i, l = 1, ..., N) represents the potential outcome variable in the scenario where the concomitant event of the hypothetical strategy applicable to step S303 does not occur when receiving clinical drugs; f(Y l S302 (0)|S l(0) = 0, l ≠ i, l = 1, ..., N) are potential outcome variables in the scenario where the concomitant events of the hypothetical strategy applicable to step S303 do not occur when the placebo is received.

[0036] Furthermore, the drug causal efficacy evaluation data in the drug efficacy dataset of the target population established in step S402 satisfies:

[0037]

[0038] in,

[0039]

[0040]

[0041] Where N is the total number of subjects and i is the number of individual subjects; N S2 The target population size is defined in step S2; τ is the target for estimating the causal utility of the drug in the target population; Y i S3 (1) is a potential outcome variable under clinical drug treatment; Y i S3 (0) represents the potential outcome variable under placebo treatment; l represents the individual subject, l≠i; Y l S302 (1) Y represents the potential outcome variable for receiving clinical medication, as established in step S302. l S302 (0) represents the potential outcome variable for placebo administration established in step S302; S l (1) = 0 indicates that the concomitant event of the hypothetical strategy applicable to step S303 did not occur when receiving clinical medication; S l (0) = 0 indicates that the co-occurrence of the hypothetical strategy did not occur when the placebo was received; f(Y) l S302 (1)|S l (1) = 0, l ≠ i, l = 1, ..., N) represents the potential outcome variable in the scenario where the concomitant event of the hypothetical strategy applicable to step S303 does not occur when receiving clinical drugs; f(Y l S302 (0)|S l (0) = 0, l ≠ i, l = 1, ..., N) are potential outcome variables in the scenario where the concomitant events of the hypothetical strategy applicable to step S303 do not occur when the placebo is received.

[0042] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects:

[0043] 1. This invention integrates five attributes for drug efficacy evaluation. The output of step 4 is a precise drug causal utility estimation target. At the beginning of a clinical drug trial, clinical trial personnel, drug sponsors, statistical analysts, and drug regulatory agencies can use this method to form a precise drug causal utility estimation target. After the trial data collection is completed, the collected data and the precise causal utility estimation target are used to evaluate drug efficacy. At this point, statistical analysts and drug regulatory agencies conduct data analysis based on the precise causal utility estimation target. Clinical trial personnel, drug sponsors, and drug regulatory agencies interpret and determine the same drug efficacy evaluation target. Especially when there are multiple co-occurring events requiring multiple handling strategies, clinical trial drug efficacy evaluation is quite complex. This invention can greatly avoid misinterpretations of drug efficacy. Thus, using the above-mentioned four-step evaluation method for precise evaluation of clinical trial drug causal utility, the application order of multiple co-occurring event handling strategies will be very clear and explicit when constructing drug causal utility evaluation targets. It provides a unified and precise evaluation method for all real-world clinical drug scenarios.

[0044] 2. The precise evaluation method for the causal efficacy of drugs in clinical trials of the present invention can help clinical trial designers clearly compare the characteristics of various co-occurrence management strategies in ICH E9(R1), which helps to promote the selection of more appropriate co-occurrence management strategies and to accurately evaluate the causal efficacy of drugs in clinical trials.

[0045] 3. Based on the characteristics of different co-occurring event strategies, this invention establishes the principles and basic order of using different strategies, which is conducive to promoting the consistency of drug causal utility evaluation objectives and clinical interpretation, and helps to strengthen communication between sponsors and regulatory agencies regarding the therapeutic effect in clinical trials. It avoids the possibility of misunderstanding of the "therapeutic effect" in clinical trial reports, thereby avoiding incalculable losses to patients and society. At the same time, it can effectively avoid arbitrary interpretation and mining of data and standardize clinical drug causal utility evaluation.

[0046] 4. Based on the basic sequence established by the four-step evaluation method, a precise mathematical expression is constructed, which integrates multiple processing strategies for different co-occurring events. In this way, during the process of gradually constructing the mathematical expression of the evaluation method, the mathematical expression of the evaluation method is not reconstructed due to changes in the co-occurring event processing strategies, thus obtaining a unified and precise evaluation method.

[0047] In this invention, the above-described technical solutions can be combined with each other to achieve more preferred combinations. Other features and advantages of this invention will be set forth in the following description, and some advantages may become apparent from the description or be learned by practicing the invention. The objects and other advantages of this invention can be realized and obtained from what is particularly pointed out in the description and drawings. Attached Figure Description

[0048] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts.

[0049] Figure 1 This is an overall flowchart of the clinical trial drug causal efficacy precise evaluation method of the present invention;

[0050] Figure 2 This is a detailed flowchart of the clinical trial drug causal efficacy precise evaluation method of the present invention. Detailed Implementation

[0051] Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which form part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not intended to limit the scope of the present invention.

[0052] Concomitant events refer to a series of events that occur after the start of treatment and affect the characteristics of the outcome variable or the explanatory power of the research question. In real-world clinical trials, concomitant events are common, and there are often multiple types. The ICH E9(R1) guidelines supplementary document proposes five strategies for managing concomitant events. However, in reality, multiple concomitant events often exist, and different concomitant events require different management strategies. The combination of different strategies makes drug efficacy evaluation particularly complex. When precise and specific methods for evaluating drug efficacy are lacking, inconsistencies in understanding efficacy evaluation can easily arise between statisticians evaluating drug efficacy, clinical trial personnel interpreting the results, drug sponsors, and regulatory agencies. This phenomenon occurs frequently in real-world clinical trials and can lead to misinterpretations of clinical trial results.

[0053] To address the above problems, this invention provides a method for precise evaluation of the causal efficacy of drugs in clinical trials, comprising the following steps:

[0054] Step S0: Based on the clinical trial issues of interest, determine the types of comorbid events and establish a comorbid event dataset;

[0055] Step S1: Based on the clinical trial questions of interest, determine the attributes of treatment interventions and establish a dataset of treatment intervention attributes;

[0056] Step S2: Based on the different treatment interventions in the treatment intervention dataset, establish a population attribute dataset;

[0057] Step S3: Based on the treatment intervention dataset and the population attribute dataset, establish a potential outcome variable attribute dataset;

[0058] Step S4: Based on the potential outcome variable attribute dataset, establish a statistical attribute dataset of the effectiveness of the outcome variable in the target population.

[0059] The result obtained in step S4 is the precise drug causal utility estimation target. Therefore, after the trial data collection is completed, the collected data and the formed precise causal utility estimation target are used to evaluate the drug efficacy, thereby promoting communication among statistical analysts, clinical trial personnel, drug sponsors and drug regulatory agencies regarding the treatment effect in clinical trials, and strengthening the consistency between clinical drug trial drug efficacy evaluation methods and drug efficacy estimation results.

[0060] The aforementioned datasets are essentially various mathematical expressions. Once the clinical trial questions of interest are identified and the types of comorbidities are clarified, the evaluation function for drug efficacy can be constructed step-by-step according to the following steps. See below for details:

[0061] Step 1: Based on the clinical trial questions of interest, establish mathematical expressions for the attributes of treatment interventions;

[0062] Step 2: Based on the attributes of different treatment interventions, establish mathematical expressions for population attributes;

[0063] Step 3: Based on the attributes of the treatment intervention and the population, establish mathematical expressions for the attributes of potential outcome variables;

[0064] Step 4: Based on the attributes of the potential outcome variables, establish a mathematical expression for the statistical properties of the outcome variables' effectiveness in the target population.

[0065] In step 1, the treatment intervention attribute refers to the treatment intervention to be compared, including placebo intervention and clinical drug intervention under comorbid events.

[0066] In step 2, the population attributes include the target population attribute. The target population is a sub-population determined based on the clinical drug treatment effect of interest, and the mathematical symbol of the target population may depend on the treatment intervention defined in step 1.

[0067] In step 3, the potential outcome variables are the mathematical expressions for each intervention in step 1 and for each individual in the target population in step 2. The specific steps include the following steps.

[0068] S301: Based on the type of comorbid event and the attributes of treatment interventions, establish mathematical expressions for potential outcome variables in real-world scenarios where there are factors that affect the interpretation of outcome variables;

[0069] S302: Based on the mathematical expressions of the potential outcome variables in step S301, establish mathematical expressions of the potential outcome variables for real-world scenarios where there are no factors affecting the interpretation of the outcome variables;

[0070] S303: Based on the mathematical expressions of the potential outcome variables in step S302, establish mathematical expressions of the potential outcome variables for the hypothetical scenario.

[0071] The mathematical expression for the statistical properties of the outcome variable in the target population, established in step 4, yields the final output, which is the target for precise estimation of drug causal utility. Specifically, for the evaluation of drug causal utility with individual-level causal explanation, step 4 includes the following steps:

[0072] S401: Based on the potential outcome variables obtained in S303, establish a mathematical expression for evaluating the causal efficacy of drugs at the individual level;

[0073] S402: Based on the individual-level mathematical expression for drug causal efficacy evaluation established in S401, establish a mathematical expression for drug efficacy in the target population.

[0074] Specifically, for the evaluation of drug causal efficacy where there is no individual-level causal explanation, S401 includes directly establishing a mathematical expression for drug efficacy in the target population based on the potential outcome variables obtained in S303.

[0075] Because the target population and outcome variables may change with the treatment intervention, it is necessary to determine the treatment intervention and define its mathematical expression. Furthermore, since potential outcome variables may only exist within the target population, their mathematical expression must be established after the target population is identified. Therefore, the following four-step evaluation method for precise causal efficacy assessment of drugs in clinical trials is employed: first, mathematical symbols representing the attributes of the treatment intervention are established; then, mathematical symbols representing the target population are established; next, mathematical symbols representing the potential outcome variables are established; and finally, a mathematical expression representing the distribution characteristics of the outcome within the target population is established.

[0076] Compared to existing technologies, the four-step evaluation method of this invention integrates five attributes of drug efficacy evaluation. The output of step S401 is a precise drug causal utility estimation target. At the beginning of clinical drug trials, clinical trial personnel, drug sponsors, statistical analysts, and drug regulatory agencies can use this method to form a precise drug causal utility estimation target. After the trial data collection is completed, the collected data and the precise causal utility estimation target are used to evaluate drug efficacy. At this time, statistical analysts and drug regulatory agencies conduct data analysis based on the precise causal utility estimation target. Clinical trial personnel, drug sponsors, and drug regulatory agencies interpret and judge the same drug efficacy evaluation target. Especially when there are multiple co-occurring events requiring multiple handling strategies, clinical trial drug efficacy evaluation is quite complex. This invention can greatly avoid misinterpretation of drug efficacy. Thus, using the above-mentioned four-step evaluation method for precise evaluation of clinical trial drug causal utility, the application order of multiple co-occurring event handling strategies will be very clear and explicit when constructing drug causal utility evaluation targets. It provides a unified and precise evaluation method for all real-world clinical drug scenarios.

[0077] In the four-step evaluation method, the first step is to clearly define the treatment intervention and determine its mathematical symbol. This is because the target population and outcome variables may change with the treatment intervention. For example, the target population under the principal layer analysis strategy is the subgroup that will not experience mortality under any treatment. If the treatment intervention is not clearly defined, then the target population under the principal layer analysis strategy is also unclear. Therefore, it is necessary to first clearly define the treatment intervention and determine its mathematical symbol.

[0078] In the four-step evaluation method, the potential outcome variables and their mathematical expressions need to be determined only after the population attributes are identified. This is because potential outcome variables may only exist within the target population. For example, the HbA1c level of people who will die under treatment intervention is not clearly defined. If the target population is uncertain, the potential outcome variables cannot be accurately determined. Therefore, the mathematical expressions of potential outcome variables need to be established only after the population attributes are determined.

[0079] The effectiveness of drugs in clinical trials involves five attributes, as follows:

[0080] (A) The treatment interventions used for comparison.

[0081] (B) The target population for the clinical problem.

[0082] (C) Outcome variables measured to address clinical problems.

[0083] (D) Strategies for handling accompanying events.

[0084] (E) The statistics of the outcome variable of interest in the target population.

[0085] Attribute (D) relates to the choice of co-occurrence handling methods, and its impact on the evaluation of drug causal efficacy is reflected in its impact on the other four attributes.

[0086] Specifically, the four-step evaluation method establishes the sequential construction process of mathematical expressions for different attributes, as well as the order of consideration for different scenarios for a unified attribute, namely the outcome attribute.

[0087] Specifically, among the five strategies, the therapy strategy targets the treatment intervention attribute (A), while the primary strategy targets the target population attribute (B). Therefore, according to the four-step evaluation method, the therapy strategy should be used before the primary strategy when constructing the objective for evaluating the causal efficacy of a drug. Furthermore, these two strategies are used before the other three strategies.

[0088] Among them, the three strategies—combined strategy, in-treatment strategy, and hypothetical strategy—are all aimed at the outcome variable attribute (C). The order in which they are constructed follows S301, specifically: the combined strategy takes precedence over the in-treatment strategy, and the hypothetical strategy is considered last.

[0089] Furthermore, the sequential construction process of the different strategies in step 3 is as follows:

[0090] First, hypothetical scenarios should be considered last. Specifically, the outcome variables in hypothetical scenarios are unobservable; they can only be extrapolated from real-world outcome variables. Therefore, only by first establishing the evaluation form of the real-world outcome variables can the expression of the outcome variables in hypothetical scenarios be determined based on this.

[0091] Secondly, among strategies targeting real-world outcome attributes, strategies that do not affect the clinical interpretive meaning of outcome variables should be considered later, while strategies that do affect the clinical interpretive meaning should be considered first. Specifically, strategies that do not affect the clinical interpretive meaning of outcome variables rely on the given outcome variables; the evaluation form of the clinically significant outcome variable needs to be established first before its corresponding evaluation form can be determined.

[0092] For example, if a treatment strategy is concerned with a predetermined outcome variable over a certain period of time (i.e., during the period when the subject receives the intervention assigned in the trial), it is first necessary to clarify the predetermined outcome variable it targets. When using both treatment strategies and composite strategies, the predetermined outcome variable here should be a composite outcome variable that affects the clinical interpretability.

[0093] Compared with existing technologies, this invention proposes a precise evaluation method for the causal utility of drugs in clinical trials. This method can help clinical trial designers clearly compare the characteristics of various co-occurrence management strategies in ICH E9(R1), which helps to promote the selection of more appropriate co-occurrence management strategies and to accurately evaluate the causal utility of drugs in clinical trials.

[0094] This invention establishes the principles and basic order of using different co-occurring event strategies based on their characteristics. This is conducive to promoting the consistency between the goals of drug causal utility evaluation and clinical interpretation, and helps to strengthen communication between sponsors and regulatory agencies regarding the therapeutic effect in clinical trials. It avoids the possibility that the "therapeutic effect" in clinical trial reports may be misunderstood, causing incalculable losses to patients and society. At the same time, it can effectively avoid arbitrary interpretation and mining of data and standardize the clinical drug causal utility evaluation.

[0095] Based on the established four-step evaluation method, a precise mathematical expression is constructed, which integrates multiple processing strategies for different co-occurring events. In this way, during the process of gradually constructing the mathematical expression of the evaluation method, the mathematical expression of the evaluation method is not reconstructed due to changes in the processing strategies of co-occurring events, thus obtaining a unified and precise evaluation method.

[0096] The four-step evaluation method of this invention is applicable to the quantitative needs of drug efficacy evaluation in both finite and infinite sample sizes in real-world clinical trials with comorbidities, thus making up for the lack of drug efficacy evaluation methods under finite sample conditions.

[0097] Example 1

[0098] Regarding the efficacy of novel coronavirus vaccines:

[0099] The clinical trial issue of concern is: to study the infection protection efficacy of the vaccine in real-world clinical trials with multiple comorbidities.

[0100] Concomitant events include: death, necessary vaccine interruption, and participant voluntary discontinuation of the trial.

[0101] The study focuses on the infection protection efficacy of "vaccine plus necessary vaccination interruption" versus "placebo plus necessary placebo interruption" in individuals who do not die from either vaccination or placebo, and assumes that participants are not permitted to discontinue the trial at their own discretion.

[0102] Assume that the drug sponsor and the drug regulatory agency deem the trial interruption, as determined by the physician, necessary.

[0103] Specifically, we first define the basic mathematical symbols.

[0104] Here, the total number of subjects is defined as N, and the number of individual subjects is defined as i;

[0105] Define the placebo intervention as t=0 and the vaccine intervention as t=1;

[0106] The treatment intervention regimen actually received by subject individual i is defined as W. i ;

[0107] Define the original potential outcome variables Y under placebo and vaccination conditions as follows: i (0), Y i (1);

[0108] Define the necessary trial interruption events as determined by the physician under placebo and vaccination conditions as follows: Potential comorbidities of death under placebo and vaccination were respectively The self-determined interruption events for participants receiving a placebo and those receiving a vaccine were respectively The value is 1 when the aforementioned accompanying event occurs, and 0 when it does not occur;

[0109] Define the potential outcome variables as follows: receiving a placebo and considering necessary trial interruption, and receiving a vaccine and considering necessary trial interruption.

[0110] in, and

[0111] Furthermore, a precise evaluation method for the causal utility of clinical trial drugs is used to establish the estimation target for the causal utility of vaccines. This includes the following steps.

[0112] Step 1: Establish the mathematical expression for the intervention attribute;

[0113] Specifically, the set of treatment interventions is established as follows:

[0114] Where t=0 and t=1 represent placebo intervention and vaccination intervention, respectively;

[0115] This indicates that the intervention was assigned as "vaccine plus necessary vaccination interruptions";

[0116] This means "a placebo plus a necessary placebo interruption".

[0117] i is used to refer to the individual subject i, i = 1, ..., N.

[0118] In step 1, the treatment strategy is handled and the physician determines the concomitant events that would lead to the interruption of the trial.

[0119] This allows for the precise expression of intervention measures in clinical trial studies.

[0120] Step 2: Establish mathematical expressions for population attributes;

[0121] Specifically, the target audience is defined as a sub-group.

[0122] in, This refers to a subgroup of people who did not die under either placebo or vaccination interventions.

[0123] Step 2 uses a primary strategy to handle the accompanying event of death. The total number of people...

[0124] This allows for a more precise targeting of the desired audience.

[0125] Step 3: Establish mathematical expressions for potential outcome attributes;

[0126] In this embodiment, there are no concomitant events associated with the application of the combined strategy and the treatment strategy. No processing is performed in steps S301 and S302, and the process proceeds directly to step S303.

[0127] Specifically, the outcome variable is defined as This refers to the infection state under the hypothetical scenario where subjects are not allowed to interrupt the experiment on their own.

[0128] in,

[0129]

[0130]

[0131] In step 3, the outcome observed in the real-world clinical trial is denoted by Yi, where,

[0132] in, This indicates that when receiving a placebo, subject i did not voluntarily discontinue the potential infection state in the trial scenario;

[0133] in, This indicates that when the vaccine is administered, subject i does not voluntarily discontinue the potential infection state in the trial scenario.

[0134] In this way, potential outcome variables can be precisely expressed for multiple comorbidities, and the five strategies proposed in the new clinical trial guidelines can be met.

[0135] Step 4: Establish a mathematical expression for the statistical properties of the outcome variable in the target population.

[0136] Specifically, the target for estimating the causal utility of drugs at the individual level is first defined as τ. i ,

[0137] in,

[0138]

[0139] Secondly, based on the individual-level drug causal utility estimation target, the drug causal utility estimation target for the target population is defined as τ;

[0140] in,

[0141]

[0142] The result obtained from the mathematical expression of the drug causal utility estimation objective based on the target population is the precise expression of the drug causal utility estimation objective. In this way, after the completion of the trial data collection, the collected data and the formed precise causal utility estimation objective can be used to evaluate the drug efficacy. This can promote communication among statistical analysts, clinical trial personnel, drug sponsors and drug regulatory agencies regarding the treatment effect in clinical trials, and strengthen the consistency between the drug efficacy evaluation methods and drug efficacy estimation results in clinical drug trials.

[0143] Example 2

[0144] The PIONEER 1 phase 3a clinical trial investigating the efficacy of the oral medication semaglutide in the treatment of type 2 diabetes:

[0145] The clinical trial question of concern is: in a hypothetical scenario where other hypoglycemic drugs are unavailable and adverse drug reactions are considered ineffective treatment, what is the therapeutic effect of oral semaglutide relative to placebo during the course of treatment?

[0146] Treatment efficacy is assessed by monitoring whether HbA1c levels are below 48 mmol / mol (6.5%).

[0147] Concomitant events include emergency administration of other hypoglycemic drugs, interruption of treatment intervention, and adverse drug reaction events.

[0148] Specifically, we first define the basic mathematical symbols.

[0149] Here, the total number of subjects is defined as N, and the number of individual subjects is defined as i;

[0150] Define the placebo intervention as t=0 and the semaglutide intervention as t=1;

[0151] The treatment intervention regimen actually received by subject individual i is defined as W. i ;

[0152] Define the original potential outcome variables Y under oral placebo and semaglutide as follows: i (0), Y i (1);

[0153] Adverse drug events were defined as follows for oral placebo and semaglutide: and Whether other hypoglycemic drugs were used when administering placebo and semaglutide were respectively and The value is 1 when the aforementioned accompanying event occurs, and 0 when it does not occur;

[0154] The time to treatment interruption was defined as T for oral placebo and semaglutide. i dcon (0) and T i dcon (1).

[0155] Define the time point at which the test specifies the measurement of the outcome HbA1c level as: Whether HbA1c levels are below 48 mmol / mol (6.5%) is expressed as I(Y) i (0)≤6.5), where I is an indicator function.

[0156] Furthermore, the estimation target for the causal utility of the oral drug semaglutide was established based on the precise evaluation method for drug causal utility in clinical trials. This specifically includes the following steps.

[0157] Step 1: Establish the mathematical expression for the intervention attribute;

[0158] Specifically, the set of treatment interventions is established as {t|t=0,1}.

[0159] Here, t=0 and t=1 represent oral placebo and semaglutide, respectively.

[0160] This allows for the precise expression of intervention measures in clinical trial studies.

[0161] Step 2: Establish mathematical expressions for population attributes;

[0162] Specifically, the target audience is defined as a sub-group.

[0163] in, This indicates the target group of type 2 diabetes patients who meet the inclusion criteria for the trial.

[0164] This allows for a more precise targeting of the desired audience.

[0165] Step 3: Establish mathematical expressions for potential outcome attributes;

[0166] Specifically, the potential outcome variable is defined as Y. i cwh (0) and Y i cwh (1) represents the final potential outcome variable for the three co-occurring events.

[0167] Based on the precise evaluation method for drug causal efficacy, this variable is expressed precisely in three steps. The details are as follows.

[0168] S311: The potential outcome variables under a combined strategy for adverse drug events are: and

[0169] in, This indicates that HBA1C is not higher than 48 mmol / mol (6.5%) under placebo or no adverse events have occurred;

[0170] in, This indicates that HBA1C was not higher than 48 mmol / mol (6.5%) under oral semaglutide administration or no adverse events occurred.

[0171] S312: Based on the outcome variable determined in S311, apply in-treatment strategies targeting events associated with treatment intervention interruption. Combining these two approaches, the potential outcome variable is Y. i cw (0) and Y i cw (1), specifically expressed as:

[0172]

[0173]

[0174] in, This indicates that there has been no interruption in treatment;

[0175] This indicates that treatment has been interrupted.

[0176] S313: Based on the outcome variables determined in step S312, consider hypothetical strategies for which other hypoglycemic drugs are unavailable. The final potential outcome variables are:

[0177]

[0178]

[0179] Among them, the outcomes observed in real-world clinical trials are represented by Y. icw It means that Y i cw =W i Y i cw (1)+(1-W i )·Y i cw (0).

[0180] in, This represents the outcome variable when no other blood glucose-lowering medication is taken, while the patient is given a placebo orally.

[0181] in, This represents the outcome variable when oral semaglutide is taken without other hypoglycemic drugs.

[0182] In this way, potential outcome variables can be precisely expressed for multiple comorbidities, and the five strategies proposed in the new clinical trial guidelines can be met.

[0183] Step 4: Establish a mathematical expression for the statistical properties of the outcome variable in the target population.

[0184] Specifically, we first define the causal efficacy of the oral drug semaglutide at the individual level as τ. i ,in,

[0185] τ i =Y i cwh (1)-Y i cwh (0);

[0186] Based on the individual-level causal efficacy of oral semaglutide, the individual-level causal efficacy of oral semaglutide in the target population is defined as τ.

[0187] in,

[0188]

[0189] The result obtained from the mathematical expression of the drug causal utility estimation objective based on the target population is the precise expression of the drug causal utility estimation objective. In this way, after the completion of the trial data collection, the collected data and the formed precise causal utility estimation objective can be used to evaluate the drug efficacy. This can promote communication among statistical analysts, clinical trial personnel, drug sponsors and drug regulatory agencies regarding the treatment effect in clinical trials, and strengthen the consistency between the drug efficacy evaluation methods and drug efficacy estimation results in clinical drug trials.

[0190] It should be noted that although this invention only lists five types of co-occurrence management strategies mentioned in ICH E9(R1), it is not limited to these five strategies; other co-occurrence management strategies are also applicable. Furthermore, this invention is not only applicable to clinical trials for drug development, but also to causal utility evaluations in all clinical trials.

[0191] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for precise evaluation of drug causal effect in clinical trials, characterized in that, Includes the following steps: S0: Based on the clinical trial issues of concern, determine the types of comorbid events and establish mathematical expressions for comorbid events; S1: Based on the clinical trial questions of concern, determine the attributes of the treatment intervention and establish mathematical expressions for the attributes of the treatment intervention; S2: Based on the different treatment interventions in the mathematical expression of treatment interventions, establish mathematical expressions for population attributes; S3: Based on the mathematical expressions of treatment intervention measures and population attributes, establish mathematical expressions for potential outcome variable attributes; S4: Based on the mathematical expressions of the attributes of potential outcome variables, establish mathematical expressions for the statistical attributes of the effectiveness of outcome variables in the target population; Step S3 includes the following steps: S301: Based on the mathematical expressions of comorbid events and the mathematical expressions of the attributes of treatment interventions, establish mathematical expressions of potential outcome variables under a composite strategy for real-world scenarios where there are factors that affect the interpretation of outcome variables; S302: Based on the mathematical expressions of potential outcome variables established in step S301, for real-world scenarios where there are no real-world scenarios that affect the interpretation of outcome variables, establish mathematical expressions of potential outcome variables that combine with treatment strategies; S303: Based on the mathematical expression of the potential outcome variable established in step S302, for the hypothetical scenario, establish a mathematical expression of the potential outcome variable combined with the hypothetical strategy. The mathematical expression of the potential outcome variable is a mathematical expression that includes the final potential outcome variable. Step S4 includes the following steps: S401: Based on the mathematical expressions of the potential outcome variable attributes obtained in step S3, establish a mathematical expression for evaluating the causal efficacy of drugs at the individual level; S402: Based on the individual-level mathematical expression for drug causal efficacy evaluation established in step S401, establish a mathematical expression for drug efficacy in the target population. The mathematical expression for drug efficacy established in step S402 includes target data for estimating drug causal efficacy.

2. The method of claim 1, wherein: The mathematical expression for the intervention attributes in step S1 includes placebo intervention and clinical drug intervention under comorbid events.

3. The method of claim 1, wherein: The mathematical expression for the population attributes in step S2 is for the target population, which is a sub-population determined based on the clinical drug treatment effects of interest.

4. The method of claim 1, wherein: For the evaluation of drug causal efficacy where there is no individual-level causal explanation, step S4 includes establishing a mathematical expression for drug efficacy in the target population based on the mathematical expression of the potential outcome variable attribute described in step S3.

5. The method of claim 1, wherein, The latent outcome variable data and observed outcome variables in the mathematical expression of the latent outcome variables established in step S3 satisfy the following: wherein, is the individual subject; is the observed outcome variable in a real-world clinical trial; is the individual subject is the actual received treatment intervention regimen; is the potential outcome variable under the clinical drug received in step S302; is the potential outcome variable under the placebo received in step S302.

6. The method of claim 1, wherein: The latent outcome variables in the mathematical expression of the latent outcome variables established in step S303 satisfy the following: in, The total number of subjects; For individual subjects; To accept potential outcome variables under clinical drug treatment; For potential outcome variables under placebo conditions; For individual subjects, ; For the potential outcome variables under clinical drug treatment established in step S302; For the potential outcome variables under placebo treatment established in step S302; When receiving clinical medication, the concomitant event of the hypothetical strategy applicable to step S303 did not occur; When the placebo was administered, the concomitant event of the hypothetical strategy applicable to step S303 did not occur; This is a potential outcome variable in a scenario where the concomitant events of the hypothetical strategy applicable to step S303 do not occur when receiving clinical drugs. This refers to the potential outcome variable in a scenario where the concomitant events of the hypothetical strategy applied in step S303 do not occur when a placebo is received.

7. The method of claim 1, wherein: The drug causal efficacy evaluation data in the individual-level mathematical expression for drug causal efficacy evaluation established in step S401 satisfy the following: in, in, The total number of subjects; For individual subjects; Target for estimating the causal utility of drugs at the individual level; To accept potential outcome variables under clinical drug treatment; For potential outcome variables under placebo conditions; For individual subjects, ; For the potential outcome variables under clinical drug treatment established in step S302; For the potential outcome variables under placebo treatment established in step S302; When receiving clinical medication, the concomitant event of the hypothetical strategy applicable to step S303 did not occur; When the placebo was administered, the concomitant event of the hypothetical strategy applicable to step S303 did not occur; This is a potential outcome variable in a scenario where the concomitant events of the hypothetical strategy applicable to step S303 do not occur when receiving clinical drugs. This refers to the potential outcome variable in a scenario where the concomitant events of the hypothetical strategy applied in step S303 do not occur when a placebo is received.

8. The method of claim 1, wherein: The drug causal efficacy evaluation data in the mathematical expression for drug efficacy in the target population established in step S402 satisfy the following: in, in, The total number of subjects; For individual subjects; The target population size established in step S2; To estimate the causal utility of the drug in the target population; To accept potential outcome variables under clinical drug treatment; For potential outcome variables under placebo conditions; For individual subjects, ; For the potential outcome variables under clinical drug treatment established in step S302; For the potential outcome variables under placebo treatment established in step S302; When receiving clinical medication, the concomitant event of the hypothetical strategy applicable to step S303 did not occur; When the placebo was administered, the concomitant event of the hypothetical strategy applicable to step S303 did not occur; This is a potential outcome variable in a scenario where the concomitant events of the hypothetical strategy applicable to step S303 do not occur when receiving clinical drugs. This refers to the potential outcome variable in a scenario where the concomitant events of the hypothetical strategy applied in step S303 do not occur when a placebo is received.

Images