Medical device and computer-implemented method for predicting the risk, occurrence or development of an adverse health condition in a test subject in a sub-population arbitrarily selected from a total population

Abstract

A system for generating a general model for adaptively predicting the occurrence or development of a first adverse health condition in a first sub-population arbitrarily selected from a total population, comprising: extracting information on the characterizing features of a plurality of second sub-populations, on the occurrence and / or severity of a first adverse health condition found therein, and / or on the corresponding prognostic outcomes, from a plurality of publications. One or more of the characterizing features are associated with a corresponding first factor, the first factor indicating a relationship with the adverse or beneficial contribution of the characterizing feature to the occurrence or development of the first adverse health condition, and one or more of the characterizing features are associated with a corresponding second factor, the second factor indicating the relative occurrence frequency in the corresponding second sub-population considered in the respective publication. The characterizing features, along with their first and second factors, are combined into a general model for the total population, including calculating the baseline risk of a patient. The general model is stored in a computer-accessible and readable medium in a retrievable manner.

Description

Technical Field

[0001] The present method and apparatus relate to predicting the occurrence or development of an adverse health condition in a test subject and are used to support the treatment control and regulation of such an adverse health condition. Background Art

[0002] Medical-related information can be derived from a plurality of different sources, such as clinical data or non-clinical data. Medical-related information can be used by healthcare professionals for prescription and test analysis and / or for the diagnosis and management of diseases or medical events or more generally adverse health conditions. Medical-related information can also be used to assess the risk of contracting a disease, exacerbation of an existing disease, or suffering an adverse medical event. Health risk prediction is a process aimed at analyzing the probability of occurrence of a particular type of medical risk based on a particular type of medical-related information. For example, health risk prediction can be used to analyze the likelihood of contracting a lung disease based on whether a person is a smoker.

[0003] Numerous studies and clinical trials have been published in which the development of adverse health conditions and their end results, as well as various medical and other parameters characterizing the subjects under study, have been documented. Some studies have attempted to find numerical algorithms for predicting the occurrence and development of adverse health conditions and their end results by matching the medical and other parameters of the test subjects with parameters previously recorded.

[0004] Some algorithms are tested based on data obtained from peer groups that are not representative of the entire population. Therefore, their predictive potential for test subjects whose medical and other parameters do not correspond to those of the peer group is limited. Additionally, certain current algorithms cannot be directly updated quickly or easily using emerging scientific evidence related to specific medical-related information. Summary of the Invention

[0005] Accordingly, it is desirable to provide a computer-implemented method that generates a general model for adaptively predicting the occurrence or development of a first adverse health condition in a first sub-population arbitrarily selected from a total population. Additionally, it is desirable to provide a computer-implemented method that adaptively predicts the occurrence or development of a first adverse health condition in a first sub-population arbitrarily selected from a total population. Further, it is desirable to provide a medical device implementing the computer-implemented method for generating a general model, and a medical device implementing the computer-implemented method for adaptively predicting the occurrence or development of a first adverse health condition in a first sub-population arbitrarily selected from a total population. Further, it is desirable to provide a treatment control support system that supports decisions regarding treatment measures in response to the expected occurrence or development of a first adverse health condition, and that supports the patient-specific adaptation or personalization of such treatment measures.

[0006] The expression "first adverse health condition" is used herein to describe one or more from a non-exhaustive list, including: cardiovascular events (CVE), chronic kidney disease (CKD), all hospitalizations, fractures, injuries, falls, infections, treatments requiring renal replacement (e.g., dialysis or transplantation), end-stage renal disease, acute kidney injury, graft failure, vascular access complications, other disease outcomes, etc.

[0007] In a portion of the description below, reference is exemplarily made to the prediction of the occurrence of chronic kidney disease (CKD) in a subpopulation and / or the prediction of the ultimate development of CKD across various stages of severity to end-stage renal disease (ESRD) and renal failure. However, the same principles, methods, and systems discussed below herein can be used for other adverse health conditions exemplarily listed above. For example, the same method can be applied to the prediction of the occurrence of cardiovascular events / hospitalizations in patients with chronic kidney disease.

[0008] As used herein, the expression "first subpopulation" refers to an individual having a characteristic, property, or other distinguishing feature that allows the individual to be distinguished from other members of the total population. In addition to referring to an individual, it can also refer to a group of individuals sharing the same characteristic, property, or other distinguishing feature, thereby distinguishing the group from other members of the total population. The total population includes a plurality of second subpopulations, such as subpopulations considered in corresponding scientific studies, literature, etc. The total population can include all second subpopulations for which data can be used for analysis. Different second subpopulations may or may not share one or more characteristics, properties, or other distinguishing features with the first subpopulation or other second subpopulations. However, each subpopulation will differ from any other subpopulation in at least one characteristic. Second subpopulations analyzed separately for the same adverse health condition in an individual study or literature segment may partially overlap or may even be identical.

[0009] As described above, the expression "characterizing feature" as used herein refers to a feature, attribute that permits an individual or group of individuals to be distinguished from other individuals or groups of individuals. Characterizing features can include general information characterizing an individual or group, genetic information (e.g., obtained from a genetic marker analysis system), medical events and conditions, treatments, diagnostic and prognostic characterizations, etc. Characterizing features can thus include demographic data such as age, race, gender, work location, environmental factors, residential location and environment, lifestyle, etc. Similarly, for example, self-reported data from surveys intermittently obtained from an individual or group members (the survey may relate to perceived quality of health status, prescription drug information (e.g., type and / or amount of prescription drugs ingested by the individual or group), data obtained from diagnostic records (e.g., previous hospitalizations, clinical tests and results), and treatment data (e.g., disease, type, time and location of treatment, hospital and / or doctor, etc.)) can be used as features for distinguishing an individual or group of objects from other individuals or groups of objects. Additionally, medical data can be used as characterizing features, such as glomerular filtration rate, proteinuria, blood pressure, comorbidities (e.g., diabetes, hypertension or congestive heart failure), and etiology of a disease (e.g., glomerulonephritis). Other medical data that can be used as characterizing features include changes in glomerular filtration rate over time, phosphate level, bicarbonate, albumin, cholesterol, C-reactive protein, serum creatinine or calcium level. Such medical data can be obtained from any kind of conventional or advanced diagnostic tests, including but not limited to any imaging technique (e.g., magnetic resonance imaging, ultrasound, X-ray, CT scan, scintigraphy, etc.), electrophysiological tests, physical examination results, immunoassay and radioimmunoassay systems, biochemistry, polymerase (PCR) chain reaction analysis systems, chromatographic analysis systems, and / or receptor analysis systems, etc. It can also include data from other analysis systems, such as tissue analysis systems, cytology and histotyping systems, and immunocytochemistry and histopathology analysis systems. Characterizing features can be time-invariant or related to the current and / or past moments. Characterizing features can also relate to a rank or level of severity of an adverse health condition. Data can also be provided as a time series or as a derivative thereof, indicating changes over time. Generally, any kind of information suitable for classifying patients can be used as a characterizing feature.

[0010] According to one aspect, a computer-implemented method for generating a general model for adaptively predicting the occurrence or development of a first adverse health condition randomly selected from a total population as described herein includes executing computer program instructions in a computer, the computer including one or more microprocessors, volatile and / or non-volatile memory, and one or more data and / or user interfaces for extracting from a plurality of publications and / or primary clinical data recorded in an electronic database and prepared according to a probability statistical model information about characterizing features of a plurality of second subpopulations, about the occurrence and / or severity of the first adverse health condition found therein, and / or about corresponding prognostic outcomes. One or more features (e.g., those known to be useful in the respective context) may be provided as additional inputs for initializing the extraction, but this is not a strict requirement since the extraction itself can identify other or additional characterizing features that are equally or equivalently relevant, or even more relevant. The computer program instructions may configure the computer executing those instructions to provide an extraction module. The extraction module may control or cooperate with various interfaces to access publications and / or clinical data records, etc. The interfaces include one or more of a data communication interface, a camera, a scanner, etc.

[0011] As used herein, the term "publication" is used to describe scientific studies and papers, general literature, or literature focused on health issues, datasets relating health issues to patient attributes that can be clearly linked to one or more patients: such attributes include, but are not limited to, age, sex, height, weight, BMI, substance or alcohol use, smoking, history of hypertension or hypotension, diabetes, COPD, lung cancer, CKD stage, cerebrovascular history, coronary artery disease, peripheral artery disease, chronic heart failure, chronic obstructive pulmonary disease, autoimmune disease, anxiety / depression, cancer, liver disease, BMI, albumin, glucose, HDL, LDL, triglycerides, CRP, IL-6, serum uric acid, HsTNT, phosphate, iPTH, proteinuria and albuminuria, other known chronic diseases, past cured diseases, behavior or lifestyle, psychological profile, morphological characteristics evaluated via any imaging technique, or functional characteristics evaluated via any appropriate diagnostic test. The list provided above is not exhaustive and may represent a subset of the characterizing features further described above.

[0012] Information about characterizing features extracted from a plurality of publications and / or primary clinical data recorded in an electronic database and prepared according to a probability statistical model may include those patient attributes that show a positive or negative relationship with respect to the first adverse health condition.

[0013] The method further includes executing computer program instructions in a computer for associating one or more of the identified characterization features with a corresponding first factor based on data from each of a plurality of publications and / or primary clinical data recorded in an electronic database and generated according to a probabilistic statistical model, the first factor indicating a relationship of the adverse or beneficial contribution of the characterization feature to the occurrence or development of a first adverse health condition. Thus, the first factor may be represented, for example, by an effect size measure such as odds ratio, hazard ratio, relative risk.

[0014] The relationship indicated by the first factor may express an increase or decrease in the risk of occurrence or development of the first adverse health condition in a patient, for example, expressed relative to a reference population of healthy or appropriately selected peers, expressed as a value generated by comparison with a baseline model in which a specific characterization feature is absent or at a "normal" level, or expressed as an absolute risk.

[0015] The method further includes: associating one or more of the identified characterization features with a corresponding second factor from each of a plurality of publications and / or primary clinical data sources recorded in an electronic database and generated according to a probabilistic statistical model, the second factor indicating the relative occurrence frequency or prevalence in a corresponding second subpopulation considered in the corresponding publication. The data on the characterization features in combination with the associated first and second factors on the plurality of publications and / or primary clinical data sources recorded in an electronic database and generated according to a probabilistic statistical model may provide an indication of the likelihood of occurrence or development in the total population. The computer program instructions may configure the computer executing those instructions to provide an associator module adapted to perform the association step. The associator module may implement various probabilistic statistical models selectable according to the first adverse health condition, the type of data, and / or the characterization features, etc.

[0016] The method further includes executing computer program instructions in a computer for combining the characterization feature and its first and second factors into a general model for the total population, wherein the combination includes calculating a baseline risk for a virtual "general" member of the total population and estimating the conditional probability of the first adverse health condition for all possible configurations of known health conditions contributing to the risk of the first adverse health event. The baseline risk may be regarded as the general risk for a virtual "general" member of the total population that does not have all the identified and known risk factors or has the risk factors at their lowest possible values. The computer program instructions may configure the computer executing those instructions to provide a combiner module adapted to perform the combination step. The combiner module may implement various probabilistic statistical models selectable according to the first adverse health condition, the type of data, and / or the characterization features, etc.

[0017] The characterization features and the combinations of their first and second factors can follow the steps described in the following examples.

[0018] In the case where the effect size is expressed as an odds ratio, the inputs to the process are:

[0019] The incidence rate of adverse consequences I

[0020] The prevalence rates of risk factors P1, P2,..., P n

[0021] Effect size measurements: OR1, OR2,..., OR n

[0022] The output of the process is the conditional probability of the risk factor:

[0023] P RF_1 , P RF_2 ,..., P RF_n

[0024] For a given outcome

[0025] The process can be briefly summarized as:

[0026] 1. Obtain I

[0027] 2. Obtain P1, P2,..., P n

[0028] 3. Obtain OR1, OR2,..., OR n

[0029] 4. Calculate the contribution to the outcome risk for each risk factor as:

[0030] W_RF i = ln(OR i ) × P i

[0031] 5. Calculate the baseline risk as λ = -ln((1 - I) / I) - Σ i W_RF i

[0032] 6. Given each risk factor, calculate the probability of the outcome as:

[0033] P_O RF_i = exp(Σ j≠i W_RF j + ln(OR i ) + λ) / (1 + exp(Σ j≠i W_RF j + ln(OR i ) + λ))

[0034] 7. The combined probability of the calculation result and each risk factor is:

[0035] P_O_RF i = P_O RF_i ×P i (Bayes' theorem)

[0036] 8. Given the result, the probability of each risk factor is calculated as:

[0037] P RF_i = P_O_RF i / I

[0038] Note that the same method can be extended to various statistical models derived from the exponential family of distributions. Similarly, depending on the set of characterization features considered, other statistical calculations can be used to combine various characterization features and their respective first and / or second factors into a general model or for calculating baseline risk.

[0039] The method also includes storing the general model in a retrievable manner on a computer-accessible and readable medium. To this end, the method can include executing computer program instructions in a computer to provide a communication module. The communication module can control or cooperate with various interfaces to access a data storage medium accordingly.

[0040] In other words, aspects of the method proposed above: selecting the most relevant information, such as clinical and patient-reported data, from a large number of available publications and / or primary clinical data recorded in electronic databases and made according to probability statistical models; evaluating the contribution of patient-specific health status attributes to future risks; creating a general model including a baseline reference; and using the attributes of a subpopulation as input values to store the model for future applications. The general models for different adverse health conditions may differ in the characterization features represented therein, mainly depending on their respective impacts on the likelihood of occurrence or development of a specific adverse health condition. The general model can include a baseline risk and data indicating the contribution of each or a selected number of characterization features to an increase or decrease in risk.

[0041] In one aspect of the method, weights are derived and associated with the first and / or second factors. The weights can be derived from: data from a single publication or regarding multiple publications, or a mixture of multiple publications and primary clinical data recorded in multiple electronic databases and made according to probability statistical models.

[0042] In one aspect of the method, the extraction includes automated extraction using one or more of electronic text processing, optical character recognition, natural language processing, and manual data entry. In the case of automated extraction of characterizing features, rule-based extraction or artificial intelligence-assisted extraction and other methods may be used. Computer program instructions may configure an extraction module to provide one or more of the above functions or provide access thereto.

[0043] In one aspect of the method, prior to generating a general model, characterizing features and associated factors are adjusted, ranked, and / or selected for a plurality of publications. The quality of the publications and the primary dataset sources may be evaluated by considering the number of participants analyzed, the risk of bias through normalization procedures, the absence of conflicts of interest among the authors, etc.

[0044] In one aspect of the method, characterizing features and associated factors are adjusted, ranked, and / or selected for the conditional probability of an outcome associated with a single characterizing feature across multiple publications. In other words, a summary view of the importance of the same characterizing feature is provided for a mixture of probabilistic analyses across multiple publications and / or publications and primary clinical data.

[0045] In one aspect of the method, characterizing features and associated factors are adjusted, ranked, and / or selected differently for different prognostic time periods, thereby accommodating the finding that individual characterizing features may have different values or reliabilities for different prognostic time periods. This may reduce the number of characterizing features or parameters that need to be monitored and fed into the prediction and thus reduce the computational effort. When no prognostic time period or prediction time period is provided, one or more preset prediction or prognostic time periods may be used and corresponding models may be generated.

[0046] For example, for a 2-year prediction time period for the development of chronic kidney disease into end-stage kidney disease, the list of the most influential characterizing features or parameters includes the glomerular filtration rate gradient, glomerular filtration rate (GFR), proteinuria, body weight, hemoglobin, Charlson index, and serum phosphate, with the first three being very influential for the prediction. For a 5-year prediction time period for the development of chronic kidney disease into end-stage kidney disease, the list of the most influential characterizing features or parameters includes systolic blood pressure, diastolic blood pressure, serum phosphate, GFR, serum calcium, proteinuria, PTH, serum albumin, heart rate, with the first four being very influential for the prediction. It is not difficult to see that different prediction time periods require different sets of characterizing features or parameters, or at least different weights are given to the characterizing features or parameters. When using a general model to predict the occurrence or development of an adverse health condition, this information itself can be used to support a physician in determining or deciding on the characterizing features to be more closely monitored or for which additional tests are to be applied. By calculating the change in the predicted probability associated with each characterizing feature for each individual patient, the method also allows for the creation of a patient-specific ranking of the prognostic values of each characterizing feature contributing to the outcome prediction for each individual patient. In this way, it supports the physician in planning a personalized diagnostic assessment strategy for each given patient, thereby maximizing the prognostic accuracy while reducing the number of diagnostic tests required to describe the patient's health condition.

[0047] Similarly, in one aspect of the method, characterizing features and associated factors are differently adjusted, ranked, and / or selected for a specific first adverse health condition and / or the severity or current stage of the first adverse health condition. For example, when the first adverse health condition can take on different levels of severity or different stages, the characterizing features can have different values or reliabilities for predicting the risk of the adverse health condition progressing to the next level of severity or stage.

[0048] Thus, the method can include executing computer program instructions in a computer to provide a filling module, the filling module being adapted to receive an input corresponding to the selection of a prediction time period and / or the current level of severity or stage of an adverse health condition. The input from the filling module can be used to select one of a plurality of statistical models to generate a general model based on the filled input. For example, when an adverse health condition is typically divided into five different stages of severity, the filled input can be used to generate a general model for the progression of the adverse health condition from stage 3 to stage 4. This input may thus result in the selection of a second subpopulation of the data input, disregarding those subpopulations that are not appropriate, for example, because the adverse health condition does not progress by jumps.

[0049] Multiple publications subject to the above method may be limited to those that have mentioned or considered adverse health conditions, regardless of the positive, negative, or no correlation found between the first adverse health condition and the patient characteristics analyzed in the publication. This can be used as a preselection to enhance the quality of the prediction or to accelerate the execution of the method. Thus, in one aspect of the method, as an additional step, the results of the first round of the method, e.g., the characterization features related to the adverse health condition, including those found across multiple publications and not individually visible from a single publication, can be used as input for re-running the method, including those publications that were not previously considered, e.g., to find unlikely and previously unknown associations between the adverse health condition and additional characterization features.

[0050] Similarly, in one aspect of the method described above, periodic or event-triggered checks are performed on new publications, e.g., by scanning the publications to find the occurrence of terms or expressions or synonyms of one or more characterization features of an adverse health condition or previously identified characterization features related to an adverse health condition. Whenever such a new publication is found, data is extracted in the manner described above and the data is input into the method for re-running due to the steps of generating the general model.

[0051] By adding the new summary data to the initial evidence base extracted from clinical studies, real-world evidence, or literature fragments, the method allows for easy iterative adjustment and improvement. Such continuous fine-tuning of the existing model is feasible because the model generation process can uniquely be based on population summary measurements. A significant advantage of this literature-based analysis method compared to traditional data-driven methods is that the method is optimized to rely entirely on published, publicly available summary data for model derivation or update without the need to enumerate and track new cohorts for model derivation or update, which is an expensive, time-consuming, and often infeasible task. Even if not strictly necessary, the analysis of primary clinical data can optionally be used to supplement the model obtained through the literature review. By simply adding the new summary data measurements extracted from any eligible new study to the complete evidence base previously used to derive the model, the model parameters can be easily recalculated using the same method as described in the above sections.

[0052] Several aspects of the method described above can be combined with each other and / or can be integrated into the underlying method. For example, the two-round aspect can be combined with adjusting, ranking, and / or selecting characterization features.

[0053] Once a general model has been generated, it can be used in a computer-implemented method that adaptively predicts the occurrence or development of a first adverse health condition for any selectable first sub-population within a total population. In one aspect, the method includes executing computer program instructions in a computer to receive, from a general model generated and stored in a computer-accessible and readable medium by a method for generating a general model for adaptively predicting the occurrence or development of a first adverse health condition arbitrarily selected from a total population, one or more characterizing features. The computer executing the computer program instructions can represent a treatment control support system. The characterizing features can be those features identified as risk factors that have at least some relevance to the adverse health condition currently under consideration, either beneficial or exacerbating. The selection of the characterizing features to be received can reflect the current level or degree of severity present in the first sub-population, e.g., for predicting the development of the adverse health condition to the next stage, or can reflect the prediction time period. The selection of the characterizing features for the development of the adverse health condition to the next stage or for the prediction time period can effectively be a selection of a limited set of characterizing features or even a single characterizing feature that has been found to be a useful predictor. The selected characterizing features can be different for predicting the development of the adverse health condition to different next stages or for different prediction time periods. The computer program instructions can configure the computer executing those instructions to provide and / or control a communication module. The communication module can control or cooperate with various hardware and / or software interfaces to receive one or more characterizing features and / or the general model from a database.

[0054] A method for adaptively predicting the occurrence or development of a first adverse health condition in any selectable first sub - population of a total population further includes executing computer program instructions in a computer to receive data characterizing the first sub - population. The data characterizing the first sub - population can include all or a subset of the characterizing features required, provided, used, analyzed, or considered in generating a general model. Thus, the data characterizing the first sub - population can include data representing physiological characteristics, demographics, comorbidities, complications, and medications, such as: age, gender, alcohol consumption, smoking, BMI, history of hypertension, diabetes, CKD stage, history of cerebrovascular disease, coronary artery disease, peripheral artery disease, chronic heart failure, chronic obstructive pulmonary disease, autoimmune disease, anxiety / depression, cancer, liver disease, albumin level, glucose, HDL, LDL, triglycerides, CRP, IL - 6, S - uric acid, HsTNT, phosphate, iPTH, proteinuria, and albuminuria. It should be noted that the first sub - population for this method does not have to be the same as the first sub - population referred to in the method of generating the general model, and not all of the characterizing data need to be used or even available. The data characterizing the first sub - population can be received via a communication module, and the computer program instructions can control the communication module accordingly.

[0055] The method further includes providing one or more received characterizing features from the general model and data characterizing the sub - population to a computing module implementing a probability model, or more generally, to a predictor module. The computing or predictor module can be implemented as a computer that executes corresponding software exclusively or in a dedicated thread running in parallel with other tasks on the computer. The probability model can be one of a plurality of known models, such as, for example, a Bayesian network. The probability model produces or calculates a summary score indicating the risk or probability of the occurrence or development of an adverse health condition for the first sub - population as an output signal. For example, in a population without all risk factors, the risk or probability can be provided as an absolute or relative value. Depending on the baseline risk for those characterizing features, characterizing features that exist in the general model but are not available for the considered first sub - population can be ignored or accounted for.

[0056] The output signal is provided to a user or a computer. Providing can include displaying the result on a display screen, printing the result, providing the result acoustically (e.g., via a speaker after text - to - speech conversion), and so on. Providing can also include transmitting the result to the user's computer via a digital communication channel. Providing can include controlling the communication module accordingly.

[0057] In one aspect, a treatment control support system implementing the method can include: providing a data signal representing the positive or negative impact of one or more characterizing features on a profile score or the probability of occurrence or development of an adverse health condition. The data signals can be provided in a sorted order according to the importance of their contribution to the profile score.

[0058] In one aspect of a treatment control support system implementing the method, if multiple characterizing features and their impact on the profile score or probability are provided, then these features and impacts are provided in a sorted order according to the importance of their respective contribution to the profile score for the first sub-population being evaluated (i.e., for an individual patient or for a group of patients). The sorting can be an absolute contribution or a relative contribution. The contribution of a characterizing factor can be beneficial, i.e., reducing the risk, or adverse, i.e., increasing the risk, and the sorting by importance can be by absolute importance, regardless of whether it is favorable or adverse, or grouped into beneficial characterizing factors and adverse characterizing factors.

[0059] In one aspect of a treatment control support system implementing the method, the predictor module is further configured to adjust, weight, sort, and / or select characterizing features and associated factors based on the current stage or severity of an adverse health condition in the first sub-population received as an additional input. This can improve the accuracy of predicting the development of an adverse health condition to the next more severe stage.

[0060] In one aspect of a treatment control support system implementing the method, characterizing features that have a positive or negative impact on the risk determined for an adverse health condition are highlighted, and the characterizing features can be changed by one or more corresponding actions from a non-exhaustive list including treatment, lifestyle change, dietary change, and medical intervention. Highlighting can include printing in bold, italic, different font, different font or background color. For example, selective highlighting can be applied according to the ranking of the ease of changing the diet or lifestyle for the sub-population or the likelihood of sustained compliance with the changed diet or lifestyle. Other rankings for highlighting can include the estimated cost of the change, e.g., in combination with the selection of a possible treatment or medical intervention.

[0061] In one aspect, a treatment control support system implementing the method further includes: providing information about the extent to which the risk or probability of occurrence or development of an adverse health condition can be changed by a corresponding action (e.g., by indicating that stopping smoking or losing weight will reduce the risk to a certain extent). Additionally, different degrees of risk reduction corresponding to the action can be indicated, e.g., compared to the best possible outcome, a recommended corresponding action of reducing sugar intake by only 50% will result in a 60% risk reduction. The information can be provided by the communication module in response to data provided by the predictor module.

[0062] In one aspect, the treatment control support system implementing the method is also adapted to provide a selection of treatment recommendations based on the type of adverse health condition and / or a risk score. The selection of treatment recommendations can include, for example, more closely monitoring the patient or subgroup, e.g., via telemedicine, a portable device for monitoring body functions, visiting a physician or clinic more frequently, closer supervision by a case manager, referral to a specialist, more aggressive drug treatment, etc. The selection and / or ranking of treatment recommendations can also be based on the characterizing features present in the first subgroup and their respective contributions to the overall risk of the occurrence or development of an adverse health condition. The selection of treatment recommendations can be provided in response to a previous request to a database or artificial intelligence system connected thereto. The database stores at least one treatment recommendation for each of a plurality of adverse health conditions.

[0063] In one aspect, the treatment control support system implementing the method is also adapted to recommend a selection of additional diagnostic tests and / or treatments that can yield additional characterizing features, and results or data related to the additional characterizing features can be provided as input to the method, i.e., a software module for prediction, to improve the accuracy of the profile score. The selection of additional tests can be provided in a sorted order, e.g., based on their invasiveness and level / cost to the patient, their availability within a region or location, the time until test results can be expected, etc. Recommending a selection of additional diagnostic tests and / or treatments can include controlling the predictor module and / or the communication module accordingly.

[0064] In one aspect, the method further includes receiving a prediction time period. The prediction time period can be received at an early stage of performing the method and can result in selecting a different set of characterizing features and / or associated factors or weights as input for the prediction, and / or a different general model determined for a specific prediction time period.

[0065] Another aspect of the present disclosure includes a medical device, for example implemented by a computer system that executes computer program instructions that implement one or more aspects of the methods described above. The medical device may represent a model generator of a treatment control support system and a treatment control support system, which together act as a system. Each of the medical devices in the medical device may include a display and a user interface, an interface for receiving digital data, and one or more microprocessors and associated volatile and / or non-volatile memories. According to the IEEE 802.11 standard family (also known by the trademark name "WiFi"), according to the IEEE 802.15 standard family (also known by the registered trademark name "Bluetooth"), the interface for receiving digital data may be of a conventional type and may provide a connection to a wired or wireless network such as Ethernet (LAN), but also has a portable data storage device through a serial or parallel connection such as a Universal Serial Bus (USB) or a connection according to IEEE 1394 (FireWire). Each of these interfaces includes a physical transmitter and receiver section and a logical transmitter and receiver section, some of which may be similar in structure and operation on various standards. The medical device may also include or be configured to access a database that contains data records associated with a plurality of medical risks and a plurality of health parameters and / or characterization features, as well as other data that describes a total population and / or a plurality of subpopulations. The database may also provide access to a plurality of publications and other medical literature.

[0066] According to one aspect of the present disclosure, when the processor executes computer program instructions, a medical device implementing a model generator of a treatment control support system may be configured to generate a general model for adaptively predicting the occurrence or development of a first adverse health condition of a first subpopulation arbitrarily selected from a total population. The general model may represent the interrelationship between multiple medical risks and multiple health parameters. According to this aspect, the medical device is configured to extract information about representative characteristics of multiple second subpopulations, about the occurrence and / or severity of the first adverse health condition, and / or about corresponding prognostic outcomes found in multiple publications. The publications may be obtained, for example, by accessing one or more databases. The medical device may also be configured to associate, in each of the multiple publications, one or more of the representative characteristics identified therein with a corresponding first factor, the first factor indicating a relationship with an adverse or beneficial contribution of the representative characteristic to the occurrence or development of the first adverse health condition, and also to associate one or more of the representative characteristics with a corresponding second factor, the second factor indicating the relative frequency of occurrence in the corresponding second subpopulation considered in the respective publication. The medical device may also be configured to combine the representative characteristics and their first and second factors into a general model for the total population, wherein the combination includes calculating a baseline risk of a virtual "general" member of the total population, and storing the general model and / or the baseline risk in a computer-accessible and readable medium in a retrievable manner. For an object that does not exhibit or has not been exposed to any representative characteristics that have been identified as having a negative impact on the occurrence or development of an adverse health condition, the baseline risk may represent the risk of the occurrence or development of the adverse health condition, or exhibits or is exposed to such representative characteristics to the lowest possible extent.

[0067] In one aspect of the present disclosure, the medical device may be configured to perform automatic extraction of information from publications using one or more of electronic text processing, optical character recognition, and natural language processing.

[0068] In one aspect of the present disclosure, the medical device may be configured to: prior to generating the general model, adjust, rank, and / or select representative characteristics and associated factors for the quality of the publications. The adjustment, ranking, and / or selection may be performed according to the quality of the publications, or according to the conditional probability of the results associated with a single representative characteristic across multiple publications.

[0069] In one aspect of the present disclosure, the medical device may be configured to: in a first round of the method, limit the multiple publications to those considering the first adverse health condition; and in a second round of the method, use the results of the first round as input to the method and use one or more publications that do not consider the first adverse health condition.

[0070] In one aspect of the present disclosure, when the processor executes computer program instructions, a medical device implementing the treatment control support system may be configured to adaptively predict the occurrence or development of a first adverse health condition for any selectable first sub-population in a total population. According to this aspect, the medical device is configured to receive one or more characterizing features from a general model, such as generated and stored in a computer-accessible and readable medium according to the method described above. The medical device is further configured to receive data characterizing the first sub-population, and provide the one or more received characterizing features from the general model and the data characterizing the sub-population to a software module implementing a probability model. The medical device is further configured to provide a summary score output from the software module, indicating the risk or probability of the occurrence or development of an adverse health condition for the first sub-population, and provide to the user one or more features and their positive or negative impact on the risk or probability of the occurrence or development of the adverse health condition.

[0071] In one aspect of the present disclosure, the medical device may be configured to provide, in a sorted order, one or more characterizing features and their positive or negative impact on the risk or probability of the occurrence or development of an adverse health condition, based on the importance of the contribution of the one or more characterizing features to the summary score.

[0072] In one aspect of the present disclosure, the medical device may be configured to highlight those characterizing features that have a positive or negative impact on the risk or probability of the occurrence or development of an adverse health condition, which can be changed or affected by one or more corresponding actions from a non-exhaustive list including treatment, lifestyle changes, dietary changes, and medical interventions.

[0073] In one aspect of the present disclosure, the medical device may be configured to provide information on the extent to which the risk or probability of the occurrence or development of an adverse health condition can be changed by a corresponding action.

[0074] In one aspect of the present disclosure, the medical device may be configured to provide a selection of treatment recommendations based on the type of adverse health condition and / or the risk score.

[0075] In one aspect of the present disclosure, the medical device may be configured to: provide options for additional diagnostic tests and / or treatments that can generate data related to or describing additional characterizing features, and the results or data related to the additional characterizing features can be provided to a software module to improve the accuracy of the profile score. The options for additional diagnostic tests and / or treatments may be provided in a sorted order, for example, according to their cost, availability within a region or location, the time until test results can be expected to be available or visible effects, etc.

[0076] Another aspect of the present disclosure includes a computer-readable medium for use on a computer system configured to implement a general model for adaptively predicting the occurrence or development of a first adverse health condition in a first sub-population arbitrarily selected from a total population. The computer-readable medium according to this aspect has computer-executable instructions for performing a method including: extracting information about characterizing features of a plurality of second sub-populations, the occurrence and / or severity of the first adverse health condition found therein, and / or about corresponding prognostic outcomes from a plurality of publications; associating one or more of the characterizing features identified in each of the plurality of publications with a corresponding first factor, the corresponding first factor indicating a relationship with the adverse or beneficial contribution of the characterizing feature to the occurrence or development of the first adverse health condition, and also associating one or more of the characterizing features with a corresponding second factor, the second factor indicating the relative occurrence frequency in the respective second sub-population considered in the respective publication; combining the characterizing features with their first and second factors into a general model for the total population, wherein the combination includes calculating the baseline risk of a patient; and storing the general model in a computer-accessible and readable medium in a retrievable manner.

[0077] Yet another aspect of the present disclosure includes a computer-readable medium for use on a computer system configured to adaptively predict the occurrence or development of a first adverse health condition in an arbitrarily selectable first sub-population of a total population. The computer-readable medium according to this aspect has computer-executable instructions for performing a method that includes: receiving one or more characterizing features from a general model generated and stored in a computer-accessible and readable memory, receiving data characterizing the first sub-population, providing the one or more received characterizing features from the general model and the data characterizing the sub-population to a software module implementing a probability model, providing a profile score from the software module indicating the risk or probability of the occurrence or development of the adverse health condition for the first sub-population, and providing to a user one or more characterizing features and their positive or negative impact on the risk or probability of the occurrence or development of the adverse health condition.

[0078] Storage, reception, and / or provision, as used throughout the specification, may include establishing a physical and / or logical digital communication channel between a processor and a memory, between a first computer and a second computer, via a digital communication link or network or a combination thereof.

[0079] Current medical devices, systems, and methods can provide an effective and accurate prediction of adverse health conditions for any arbitrarily selected subpopulation or individual, based on health information obtained from publications, literature, etc. that focuses on a limited or selected subpopulation. Such techniques can be used to predict and manage individual health risks as well as analyze and manage the health risks of groups or populations.

[0080] In subjects with CKD, predicting the progression of the disease as accurately as possible may help determine when to prepare the subject for renal supplementation or replacement therapy (dialysis), as placing a catheter in the abdomen or forming a fistula, graft, or other access point to the blood circulation typically requires surgery and time for healing and / or maturation.

[0081] Similarly, in patients with CKD or ESRD, predicting the risk of hospitalization can allow caregivers or physicians to initiate measures to mitigate or even eliminate the risk of hospitalization, particularly hospitalization for cardiovascular disease.

[0082] By the impact of risk factors and beneficial factors on absolute or relative risk, patient- or subpopulation-specific ranking of risk factors and beneficial factors allows physicians or other medical caregivers to provide more targeted and effective treatment or preventive measures.

[0083] Individual users can use the disclosed medical devices and methods to predict the occurrence or development of a potentially adverse health condition based on their own health data or characteristic features. Individual users can also obtain information to reduce the risk or likelihood of the occurrence or development of an adverse health condition corresponding to their contribution to the risk or likelihood by changing relevant behaviors (e.g., lifestyle).

[0084] Group or institutional users can use the disclosed medical devices and methods to calculate health risks between groups, such as specific distributions between groups. Institutional users can also optimize the distribution to reduce the health risks of the group and promote a healthy lifestyle. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] In the following sections, the method will be described with reference to the drawings. In the drawings:

[0086] Figure 1 A flowchart of an exemplary method 100 for generating a general model for adaptively predicting the occurrence or development of a first adverse health condition in a first arbitrarily selected subpopulation from a total population is illustrated;

[0087] Figure 2 describes Figure 1 a more detailed exemplary embodiment of the selection step shown in

[0088] Figure 3 illustrates Figure 1 a more detailed exemplary embodiment of the combining and calculating steps shown in

[0089] Figure 4 a data and result flow diagram depicting an exemplary method for predicting the occurrence or development of a first adverse health condition in any selectable first sub - population of a total population;

[0090] Figure 5 shows a flowchart of an exemplary aspect of a method for generating a general model for adaptively predicting the occurrence or development of a first adverse health condition in a first sub - population arbitrarily selected from a total population, wherein the general model is updated by analyzing new publications; and

[0091] Figure 6 shows an exemplary computer system adapted to perform one or more aspects of the present method. DETAILED DESCRIPTION

[0092] Reference will now be made in detail to the exemplary embodiments illustrated in the accompanying drawings. Whenever possible, the same reference numbers will be used throughout the drawings to refer to the same or like parts.

[0093] Figure 1 A flowchart of an exemplary method for generating a general model for adaptively predicting the occurrence or development of a first adverse health condition in a first sub - population arbitrarily selected from a total population is illustrated. In step 102, a systematic review of literature or other publications including medical records from hospitals or other health institutions is performed; and in step 104, the most reliable articles or meta - analyses are selected from the literature or publications. In step 106, the risk - factor prevalence and risk - factor odds ratios are extracted from these selected documents characterizing features (e.g., adverse - outcome incidence), and are combined in step 108. Then, in step 110, the combined information is used to calculate the model parameters for the general model, which are stored in step 112. Any suitable type of probability model (e.g., exponential - family function) can be used to calculate the model parameters.

[0094] Figure 2 describes in Figure 1A more detailed exemplary embodiment of steps 102-106 as shown. In steps 202 and 204, the endpoints of the target population and the prognostic inference tool are defined and forwarded as input to step 206, where inclusion and exclusion criteria for the publications are defined. Steps 208 to 216 require defining the sources of evidence and searching for terms that define the target population, endpoints of interest, and characterizing features. At the end of each cycle, the literature is qualitatively evaluated to identify potential new exposures / risk factors.

[0095] Once the search strategy and all required definitions have been made, the publications are screened and graded in steps 218 and 220 respectively, and a final selection of publications is made in step 222 for analysis and extraction. Then, the extraction of characterizing features etc. is performed in step 224.

[0096] Figure 3 Illustrated in Figure 1 A more detailed exemplary embodiment of the combining step 108 and the calculating step 110 as shown respectively. In step 302, an assessment is performed to generate a theory representing the causal process linking the exposure and the outcome of interest. In steps 304 and 306, the measures of association extracted from the literature are combined into a common effect size measure. This can include scaling for different sample sizes, different discretizations, etc. At this point, different meta-analysis techniques can be performed for each independent risk factor or marker in step 308.

[0097] Figure 4 A data and result flow diagram depicting an exemplary method for predicting the occurrence or development of a first adverse health condition in any selectable first sub-population of a total population. In the example, the sub-population is a single patient. Data 402 from the patient record is provided to the general probability model 404. Running an analysis using the general probability model 404 provides one or more of the following: a recommendation 406 for additional diagnostic tests, an indication 408 representing the patient's risk of the adverse health condition, and an indication 410 of how much each characterizing feature or risk factor affects the patient's risk of developing or having the adverse health condition.

[0098] Figure 5 A flowchart showing an exemplary aspect of a method for generating a general model for adaptively predicting the occurrence or development of a first adverse health condition in any arbitrarily selected first sub-population from a total population, where the general model is updated by analyzing new publications. In step 502, a periodic or event-triggered check of new publications is performed. The review can be performed in the same manner as Figure 1 step 102 of Figure 2Determine the appropriate keywords in the same manner as explained. As long as a publication that promises to provide new evidence regarding adverse health conditions or generally new data, such as a new subpopulation, new characterization features, etc., is found in step 504, the method follows the "yes" path to step 508, where, for example, data extraction is performed in the manner discussed with reference to Figure 1 The following steps 510 and 512 can similarly correspond to one or more of the steps 106 to 110 discussed with reference to Figure 1 Updating the general model using the new publication allows the model to become more accurate or effective in prediction without having to export the entire model generation process that was initially required again, and thus saves time and effort.

[0099] Figure 6 An exemplary computer system 600 suitable for performing one or more aspects of the present method is shown. The computer system 600 may include a processor 602, a random access memory (RAM) 604, a read only memory (ROM) 606, a console 608, an input device 610, a network interface 612, a database 614, and a storage device 616. It should be understood that the types and quantities of the listed devices are merely exemplary and are not intended to be limiting. The number of the listed devices may vary, and other devices may be added.

[0100] The processor 602 may include any suitable type of general purpose microprocessor, digital signal processor, or microcontroller. The processor 602 may execute a sequence of computer program instructions to perform the various processes or method steps described above. The computer program instructions may be loaded from the read only memory (ROM) or from the storage device 616 into the RAM 604 for execution by the processor 602. The storage device 616 may include any suitable type of mass storage device provided to store any type of information required for the processes executed by the processor 602. For example, the storage device 616 may include one or more hard disk devices, optical disc devices, or other storage devices to provide storage space.

[0101] The console 608 may provide a graphical user interface (GUI) to display information to the user of the computer system 600. The console 608 may include any suitable type of computer display device or computer monitor. An input device 610 may be provided for the user to input information into the computer system 600. The input device 610 may include a keyboard, a mouse, or other optical or wireless computer input devices, etc. In addition, the network interface 612 may provide a communication connection such that the computer system 600 can be remotely accessed via a computer network through various communication protocols (such as Transmission Control Protocol / Internet Protocol (TCP / IP), Hypertext Transfer Protocol (HTTP), etc.).

[0102] The database 614 can include model data and / or any information related to the data records being analyzed, such as model parameters and test data. The database 614 can include any type of commercial or custom database. The database 614 can also include analysis tools for analyzing the information in the database. The processor 602 can also use the database 614 to determine and store performance characteristics of the generalized model.

[0103] Other embodiments, features, aspects, and principles of the disclosed exemplary medical devices and methods will be apparent to those skilled in the art and can be implemented in various environments and systems.

Claims

1. A computer-implemented method for treatment control support that generates a general model for adaptively predicting the occurrence or development of a specific first adverse health condition in a first subpopulation arbitrarily selected from a total population, the method comprising: - Controlling an extraction module to identify and extract information about characterizing features of a plurality of second subpopulations, about the occurrence and / or severity of the first adverse health condition found therein, and / or about corresponding prognostic outcomes from a plurality of publications and / or primary clinical data sources, wherein the extraction module is configured to identify additional characterizing features among the characterizing features that are individually unknown in any of the plurality of publications and / or primary clinical data sources; - Controlling an associator module to: from each of the plurality of publications and / or primary clinical data sources, associate one or more of the characterizing features identified therein with a corresponding first factor that indicates a relationship of an adverse or beneficial contribution of the characterizing feature to the occurrence or development of the first adverse health condition; and also associate one or more of the characterizing features with a corresponding second factor that indicates a relative occurrence frequency in the respective second subpopulation considered in the respective publication; - Controlling a combiner module to combine the characterizing features and the first and second factors of the characterizing features into a general model for the total population, wherein the combination includes calculating a patient's baseline risk; and - Controlling a communication module to store the general model in a computer-accessible and readable medium in a retrievable manner.

2. The method according to claim 1 further comprises: Controlling the extraction module to implement one or more of electronic text processing, optical character recognition, and natural language processing.

3. The method according to claim 1 or 2 further comprises: Before generating the general model, controlling the extraction module to adjust, rank, and / or select the characterizing features and associated factors according to a value representing the quality of the publications and / or primary clinical data sources.

4. The method according to any one of claims 1 to 3, further comprising: Adjusting, ranking, and / or selecting the characterizing features and associated factors according to a value of a conditional probability representing a result associated with an individual characterizing feature across the plurality of publications and / or primary clinical data sources.

5. The method according to any one of claims 1 to 4, further comprising: Adjusting, ranking, and / or selecting the characterizing features and associated factors according to a received prediction time target or for at least one preset prediction time target.

6. The method according to any one of claims 1 to 5, further comprising: Adjusting, ranking, and / or selecting the characterizing features and associated factors according to a received current stage or severity of the first adverse health condition in the first subpopulation or for at least one preset stage or severity.

7. The method according to any one of claims 1 to 6, further comprising: In a first round of the method, limiting the plurality of publications and / or primary clinical data sources to those that consider the first adverse health condition; And in a second round of the method, using the results of the first round as input to the method and using one or more publications and / or primary clinical data sources that do not consider the first adverse health condition.

8. A computer-implemented method for adaptively predicting the occurrence or development of a first adverse health condition in any selectable first sub-population of a total population, comprising: - Controlling a communication module to receive one or more characterizing features from a general model generated by the method according to any one of claims 1 to 7 and stored in a computer-accessible and readable memory; - Controlling the communication module to receive data characterizing the first sub-population; - Controlling the communication module to provide the received one or more characterizing features from the general model and the data characterizing the sub-population to a predictor module implementing a probability model; - Controlling the predictor module to provide a summary score from a software module indicating the risk or probability of the occurrence or development of the adverse health condition for the first sub-population; And - Controlling the communication module to provide to a user one or more characterizing features and the positive or negative impact of the one or more characterizing features on the risk or probability of the occurrence or development of the adverse health condition.

9. The method according to claim 8, wherein, The predictor module is further configured to: adjust, rank, and / or select the characterizing features and associated factors according to the current stage or severity of the first adverse health condition in the first sub-population received as an additional input.

10. The method according to claim 8 or 9 further comprises: Provide the one or more characterizing features and the positive or negative impact of the one or more characterizing features on the risk or probability of the occurrence or development of the adverse health condition in a sorted order according to the importance of the contribution of the one or more characterizing features to the summary score.

11. The method according to claim 10, further comprising: Highlight those characterizing features that have a positive or negative impact on the risk or probability of the occurrence or development of the adverse health condition, and the characterizing features can be changed or affected by one or more corresponding actions from a non-exhaustive list including treatment, lifestyle changes, dietary changes, and medical interventions.

12. The method according to claim 11 further comprises: Provide information on the extent to which the risk or probability of the occurrence or development of the adverse health condition can be changed by the corresponding actions.

13. The method according to any one of claims 8 to 12, further comprising: Provide a selection of treatment recommendations based on the type of adverse health condition and / or the risk score.

14. The method according to any one of claims 8 to 13, further comprising: Provide a selection of additional diagnostic tests and / or treatments that can generate data involving or describing additional characterizing features, and the data can be provided to the predictor module to improve the accuracy of the summary score.

15. The method according to claim 14 further comprises: Provide the selection of additional diagnostic tests and / or treatments in a sorted order.

16. A model generator for a treatment control support system, configured to generate a general model for adaptively predicting the occurrence or development of a specific first adverse health condition in any arbitrarily selected first sub-population from a total population, the model generator and / or the component parts of the model generator include a microprocessor, volatile and / or non-volatile memory, one or more data and / or user interfaces or cooperate with them, and the model generator further includes: - An extraction module adapted to identify and extract information about the characterizing features of a plurality of second sub - populations, about the occurrence and / or severity of the first adverse health condition found therein, and / or about the corresponding prognostic outcomes from a plurality of publications and / or primary clinical data sources, wherein the extraction module is configured to identify additional characterizing features among the characterizing features that are individually unknown in any of the plurality of publications and / or primary clinical data sources; - An associator module adapted to: from each of the plurality of publications and / or primary clinical data sources, associate one or more of the characterizing features identified therein with a corresponding first factor, the first factor indicating a relationship of an adverse or beneficial contribution of the characterizing feature to the emergence or development of the first adverse health condition; and also associate one or more of the characterizing features with a corresponding second factor, the second factor indicating the relative occurrence frequency in the respective second sub - population considered in the respective publication; - A combiner module adapted to combine the characterizing features and the first and second factors of the characterizing features into a general model for the total population, wherein the combination includes calculating the baseline risk of a patient; and - A communication module adapted to store the general model in a computer - accessible and readable medium in a retrievable manner.

17. The model generator according to claim 16 further includes a filling module adapted to receive an input corresponding to the selection of a specific first adverse health condition and / or the severity or current stage of the first adverse health condition, wherein, The specific first adverse health condition and / or the severity or current stage of the first adverse health condition is used to select one statistical model from various statistical models to generate the general model.

18. The model generator according to claim 16, wherein, The extraction module is configured to implement one or more of electronic text processing, optical character recognition, and natural language processing.

19. The model generator according to any one of claims 16 to 18, further configured to: before generating the general model, adjust, rank, and / or select the characterizing features and associated factors according to a value representing the quality of the publication and / or primary clinical data source.

20. The model generator according to any one of claims 16 to 19, further configured to: adjust, rank, and / or select the characterizing features and associated factors according to a value representing the conditional probability of the result associated with a single characterizing feature across the plurality of publications and / or primary clinical data sources.

21. The model generator according to any one of claims 16 to 20, further configured to: adjust, rank, and / or select the characterizing features and associated factors according to a received prediction time target or for at least one preset prediction time target.

22. The model generator module according to any one of claims 16 to 21, further configured to: adjust, rank, and / or select the characterizing features and associated factors according to the received current stage or severity of the first adverse health condition in the first sub - population or for at least one preset stage or severity.

23. The model generator module according to any one of claims 16 to 22 is further configured to: in a first round, control the extraction module to limit the plurality of publications and / or primary clinical data sources to those publications and / or primary clinical data sources that actually consider the first adverse health condition; and in a second round, use the result of the first round as an input and use one or more publications and / or primary clinical data sources that do not consider the first adverse health condition.

24. A treatment control support system configured to adaptively predict the occurrence or development of a first adverse health condition in any selectable first sub-population in a total population, the system and / or components of the system including a microprocessor, volatile and / or non-volatile memory, one or more data and / or user interfaces or cooperating therewith, and the system further comprising: - A communication module configured to receive an input identifying or selecting the first adverse health condition, a general model suitable for the first adverse health condition generated by the model generator of the treatment control support system according to any one of claims 16 to 23 and stored in a computer-accessible and readable memory, and one or more characterizing features used in the general model; the communication module is further configured to receive corresponding data characterizing the first sub-population; - A predictor module configured to process the one or more characterizing features received according to the general model and the data characterizing the sub-population, wherein the predictor module includes computer program instructions implementing a probability model executable by a computer, wherein the probability model and / or parameters of the probability model are selected or adapted according to the first adverse health condition, and wherein the predictor module is configured to apply the probability model according to the received general model, the characterizing features, and the data characterizing the first sub-population for outputting a summary score indicating the risk or probability of the occurrence or development of the adverse health condition in the first sub-population; and - wherein the communication module is further configured to provide the summary score and / or one or more characterizing features and the positive or negative impact of the one or more characterizing features on the risk or probability of the occurrence or development of the adverse health condition to a user or another computer system.

25. The treatment control support system according to claim 24, wherein, The predictor module is further configured to: adjust, weight, rank, and / or select the characterizing features and associated factors according to the current stage or severity of the first adverse health condition in the first sub-population received as an additional input.

26. The treatment control support system according to claim 24 or 25, wherein, The predictor module or the communication module is further configured to: provide the one or more characterizing features and the positive or negative impact of the one or more characterizing features on the risk or probability of the occurrence or development of the adverse health condition in a sorted order according to the importance of the contribution of the one or more characterizing features to the summary score.

27. The treatment control support system according to any one of claims 24 to 26, wherein, The communication module is further configured to: highlight those characteristic features that have a positive or negative impact on the risk or probability of occurrence or development of the adverse health condition, and the characteristic features can be changed or affected by one or more corresponding actions from a non-exhaustive list including treatment, lifestyle changes, dietary changes, and medical interventions.

28. The treatment control support system according to any one of claims 24 to 27, wherein, The predictor module or the communication module is further configured to: provide information on the extent to which the risk or probability of occurrence or development of the adverse health condition can be changed by the corresponding actions.

29. The treatment control support system according to any one of claims 24 to 28, wherein, The communication module is further configured to: in response to a request issued to the database, or an artificial intelligence system connected to the communication module, provide a selection of treatment recommendations based on the type of adverse health condition and / or the risk score, and the database stores at least one treatment recommendation for each of the multiple adverse health conditions.

30. The treatment control support system according to any one of claims 24 to 29, wherein The predictor module is further configured to: provide, through the communication module, a selection of additional diagnostic tests and / or treatments that can generate data involving or describing additional characteristic features, and the data can be provided to the predictor module to improve the accuracy of the profile score.

31. A system for generating a general model for adaptively predicting the occurrence or development of a first adverse health condition in a first sub-population randomly selected from a total population, the system comprising: - an extraction module configured to: receive a first data signal representing a plurality of publications and / or primary clinical data sources; identify and extract, from the data signal derived from the plurality of publications and / or primary clinical data sources, information on characteristic features of a plurality of second sub-populations, on the occurrence and / or severity of the first adverse health condition found therein, and / or on corresponding prognostic outcomes; and generate a corresponding second data signal, wherein the extraction module is configured to identify other characteristic features among the characteristic features that are individually unknown in any of the plurality of publications and / or primary clinical data sources; - a correlation module configured to: receive the second data signal; correlate one or more of the characteristic features identified in each of the plurality of publications and / or primary clinical data sources and represented by the second data signal with a corresponding first factor, the first factor indicating a relationship with the adverse or beneficial contribution of the characteristic feature to the occurrence or development of the first adverse health condition; generate a corresponding third data signal; further correlate one or more of the characteristic features with a corresponding second factor, the second factor indicating the relative occurrence frequency in the corresponding second sub-population considered in the respective publication and / or primary clinical data source; and generate a corresponding fourth data signal; - A combination module, configured to: receive the second data signal, the third data signal, and the fourth data signal; combine the characterization features represented by the second data signal, the third data signal, and the fourth data signal, and the first and second factors of the characterization features into a general model for the total population, wherein the combination includes calculating the baseline risk of a patient; and provide a fifth data signal representing the general model; and - A data access module, configured to: access a computer-accessible and readable medium; and store the fifth data signal representing the general model in the computer-accessible and readable medium in a retrievable manner.

32. A system for adaptively predicting the occurrence or development of a first adverse health condition in any selectable first sub-population of a total population, comprising: - A first data receiver module, configured to receive a data signal representing one or more characterization features from a general model generated by the method according to any one of claims 1 to 7 and stored in a computer-accessible and readable memory; - A second data receiver module, configured to receive a data signal representing the characterization of the first sub-population; - A software module, configured to receive and process the data signal representing the received one or more characterization features and the data signal representing the characterization of the sub-population from the general model, the software module including an implementation of a probability model, and configured to provide a summary score from the software module indicating the risk or probability of the occurrence or development of the adverse health condition for the first sub-population; and - A providing module, configured to provide to a user or another medical device one or more characterization features and the positive or negative impact of the one or more characterization features on the risk or probability of the occurrence or development of the adverse health condition.

Images