Clinical Decision Support System and Method in Personalized Treatment

Abstract

The invention relates to a clinical decision support system (100) and method that provides physicians with personalized treatment recommendations. The system (100) includes numerous modules (110, 120, 130, 140, 150) that analyze the patient's current laboratory data from national systems such as E-Nabız, pharmacogenetic profile, drug interactions, Türkiye-specific geriatric compliance criteria based on local population data (e.g., dietary habits), and national pharmacovigilance data from TÜFAM. The key novelty of the invention is that it includes a Recommendation and Alert Engine (200) that takes the raw analysis outputs from all these modules and checks them primarily against absolute "hard rules" that bypass scoring and can block prescriptions, then prioritizes, combines, and simplifies the remaining alerts according to their clinical risks.This engine solves the problem of "warning fatigue" caused by systems in the known state of the art by providing the physician with a single, understandable and actionable warning list (240), and improves patient safety.

Description

TARIFF Clinical Decision Support System and Method in Personalized Treatment Definition and Relevant Technical Field The invention provides patient-specific data, especially in patients with polypharmacy (the use of multiple drugs). (laboratory results, pharmacogenetic profile, demographic information) and current By integrating pharmacovigilance data, it provides physicians with real-time, personalized, and Computer-aided clinical decision-making that provides prioritized treatment recommendations It is related to support systems and methods. State of the Art To improve patient safety and optimize treatment effectiveness in medical applications, This is one of the fundamental goals of healthcare services. One of the biggest challenges in achieving these goals. someone, especially someone who uses multiple medications (polypharmacy) and has more than one chronic disease. In patients, the complexity of the prescribing process is a concern. Physicians consider the patient's needs in the decision-making process. current medications, laboratory results, genetic makeup, age-related physiological changes and evaluating numerous factors simultaneously, such as the potential side effects of medications. It is necessary. In the current state of the art, various Clinical Decision Support (CDS) tools are used to support this process. Electronic Health Records (EHRs) systems have been developed. These systems generally refer to Electronic Health Records. It works in integration with (ESK) and offers various warnings to physicians. For example; drug-drug databases that control their interactions (e.g., Medscape, Micromedex) and these Systems that generate alerts using databases are common. With advancements in personalized medicine, pharmacogenetic (PGx) data Systems integrating with KKDS (Customs Decision Making Systems) have also emerged. Patent number US20160239636A1 The application is a database that establishes relationships between genetic markers and drug responses. While explaining its creation, patent number US8392220B2 also includes the patient's genomic information. A cloud-based drug management system that predicts adverse outcomes using clinical data. - 1 - It describes the system. These systems use genetic data in the prescription process. This highlights the importance of its use. Similarly, it was developed to prevent inappropriate drug use in elderly patients. There are internationally recognized sets of criteria such as STOPP / START or Beers. These By converting the criteria into a software algorithm, automatic alerts are provided to physicians within the KKDS (Physician Health Information System). Systems that offer these have also been developed. However, these solutions in the current state of the art have significant limitations and cannot solve the problems they address. There are technical problems: • Lack of Data Integration: Existing systems often focus on a single problem. focuses on (e.g., only drug interaction or only geriatric suitability). The patient Real-time laboratory data from national health record systems such as e-Nabız. data from national adverse event centers such as the Turkish Pharmacovigilance Center (TÜFAM) data, based on the patient's pharmacogenetic profile and local population characteristics An integrated system that analyzes adapted criteria simultaneously and synergistically. It is not available. The data remains in different silos, which prevents a holistic understanding. It prevents the evaluation. • Alert Fatigue: Existing systems often struggle with each other. It generates numerous unrelated and unprioritized alerts. This situation causes problems for physicians. a condition that, over time, leads to ignoring warnings and is known as "warning fatigue". This is causing a serious technical problem. The physician, based on all the patient's data... A combined analysis showing which situation is the most critical and requires urgent intervention. And there is no prioritized warning mechanism. • Lack of Localization: International geriatric criteria, genetic differences, or It may not be optimal for every population due to local health system dynamics. The genetic makeup, dietary habits, or healthcare system of the Turkish population Special rules developed taking into account the infrastructure (such as E-Nabız, TÜFAM) There is no existing KKDS that includes this. Consequently, in the current state of the technique, heterogeneous data from different sources integrating data in real time, analyzing this data according to local and personal factors And most importantly, the findings provide the physician with actionable, unique, and - 2 - A technique that solves the problem of alert fatigue by converting it into a prioritized alert. No solution has been found. The Purpose of the Invention and the Technical Problem It Solves The main purpose of this invention is to address the aforementioned shortcomings in the current state of the art. to eliminate. One aim of the invention is to provide a non-resettable thermal safety system with a temperature sensing element. The goal is to combine its components within a single compact and robust housing. Another aim of the invention is to combine both the temperature sensing element and the thermal fuse in a single unit. By mounting them onto printed circuit boards (PCBs), structural integrity is increased, and production and The goal is to reduce assembly costs. Another purpose of the invention is to eliminate the assembly requirements of separate components and additional cabling. A single-piece solution that eliminates complexity and the possibility of errors, and is easy to install. to present. Another purpose of the invention is to create a disposable device that permanently interrupts the circuit in the event of a malfunction. The goal is to provide fail-safe safety by using a thermal fuse. Brief Description of the Figures Clinical decision support in personalized therapy developed to achieve the invention's purpose. The system and method are explained in detail with the following figures: Figure 1: General architecture of the system subject to the invention and data flow between modules. It is a block diagram illustrating this. Figure 2: Basic steps of the method implemented by the computer that is the subject of the invention. It is a flowchart that illustrates this. Figure 3: The Suggestion and Alert Engine processing data from different modules. This is a flowchart illustrating the logic behind creating the final warning list. - 3 - Detailed Description of the Invention The invention is designed to provide clinical decision support to physicians in personalized treatment. a designed, computer-implemented system (100) and a system that runs This is the method. As shown in Figure 1, the system (100) has a modular architecture and a one or more equipped with a central processing unit (CPU), memory and a database It runs on the server. The system (100) receives input from the physician via a user interface. and communicates with various external and internal data sources. The basic components of the system (100) are described below, with reference to Figure 1: Data Integration Module (110): This module enables the system to communicate with external data sources. It is responsible for establishing, in particular, the E-National Health Registry System, the national health record system in Türkiye. Secure API (Application) for Pulse and hospitals' Laboratory Information Systems (LIS). By making calls to the Programming Interface, the patient's most up-to-date laboratory results (e.g. It automatically extracts creatinine, ALT, and potassium values. It converts the incoming data (e.g., JSON or...) It standardizes the system by converting it into an internal data model (in HL7 format). Standardized data, other modules especially for drugs requiring dose adjustment. It is made ready for use by the system. For example, if a patient's creatinine clearance is 45... When measured in mL / min, this information is useful in the analysis of a drug that requires dose adjustment. It is passed to the relevant module for use. Drug Interaction Analysis Module (120): This module is used by the physician via the user interface. new prescription data entered via this method, along with the patient's existing list of medications registered in the system. It compares them. This comparison involves pharmacokinetic and pharmacodynamic interactions, an invention based on an internationally recognized database (e.g., Micromedex) using an internal database containing a unique set of rules developed within the scope Module (120) identifies potential drug-drug and drug-food interactions and this It classifies interactions according to their clinical significance (e.g., Major, Moderate, Minor). The classified interaction list is sent as an output to the Suggestion and Alert Engine (200). Geriatric Adaptation Module (130): This module is especially for geriatric patients aged 65 and over. It is designed to detect inappropriate medication use and treatment neglect. Module - 4 - (130) takes the patient's age, diagnostic codes and prescribed medications as input. These inputs, international criteria such as BEERS or START / STOPP in the known state of the art In addition to these sets, one of the fundamental innovations of this invention is the "Criteria Specific to Türkiye". It analyzes according to a rule set that includes these specific criteria, which are more common in the Turkish population. observed genetic polymorphisms lead to differences in drug metabolism or high levels Local practices and local nutrition regarding the use of drugs with an anticholinergic load. It relies on factors such as habits. For example, this module (called the TR-G3 rule) (as a rule), those using RAAS blockers (ACEi / ARB / spironolactone) and also Individuals aged 75 years and older, or with an eGFR value less than 45 mL / min, who reported consuming potassium-fortified salt. It can generate a specific warning for a patient with a particular condition. This rule applies to common potassium levels in Türkiye. the habit of consuming added salt, general drug-based warnings of international criteria Furthermore, it is based on the technical fact that it significantly increases the risk of hyperkalemia, and this Mandatory laboratory tests (K+ and creatinine) specific to the situation and alternatives It triggers concrete action steps such as antihypertensive recommendations. The output of the module (130) is potential These are warnings about inappropriate medication or neglected treatment. Pharmacogenetics Module (140): This module contains the patient's previously recorded data and data entered into the system. loaded pharmacogenetic test results (e.g., CYP450 enzyme family, HLA types) It analyzes drug-gene interactions using the metabolism of a prescribed drug. or if its effectiveness is known to be affected by genetic variants the patient has, Module (140) is activated. The module is the Clinical Pharmacogenetics Application Consortium. It runs a rule engine based on international guidelines such as (CPIC). For example, a patient's If the CYP2D6 genotype is "poor metabolizer," then this enzyme metabolizes... For a given medication, the system automatically reduces the standard dose by 50% or recommends an alternative dose. It recommends switching to a medication. If the patient's genetic data is not available in the system, the system The physician may recommend that the relevant test be performed. The output of the module (140) is for specific dose adjustment. These are recommendations or clear warnings such as "avoid use". Pharmacovigilance Module (150): This module is the Turkish Pharmacovigilance Centre's By connecting to the (TÜFAM) database via an API, national data regarding prescribed drugs can be accessed. It analyzes adverse event reports at the level in real time. Module (150) analyzes a specific drug There is a statistically significant increase in the number of adverse event reports made recently (e.g., in the last 3 months) for this purpose. It applies a time series analysis to determine whether there has been a significant increase. - 5 - If a significant increase is detected, this could raise potential concerns about the safety of the drug. This is interpreted as a signal and a warning is generated. This goes beyond individual patient data. It is a unique feature that provides security control at the population level. Suggestion and Warning Engine (200): This engine is the most critical and innovative component of the invention and It functions as the central brain of the system (100). All the analysis described above Raw data outputs (potential interactions, from modules (110, 120, 130, 140, 150)) It takes into account input such as dosage adjustment recommendations, warnings about inappropriate medication, adverse event trends, etc. This The primary function of the engine is to process this large amount of diverse information so that the physician can easily access it. The goal is to transform it into a single prioritized list that they can understand and act upon. This solves the "warning fatigue" technical problem. As shown in Figure 3, the prioritization logic of the motor (200) is a two-stage technique. It is based on process: i. Strict Rule Control (220): The engine first scores all incoming raw warnings. a system that completely bypasses the previous one and has absolute priority It monitors the patient according to a defined set of “hard rules”. These rules apply to the patient which could directly threaten his / her life or definitely prevent treatment It identifies situations. For example; • Pharmacogenetic (PGx) Contraindications: Patient's genetic profile (e.g. The use of the prescribed medication (e.g., abacavir) must absolutely comply with HLA-B57:01). contraindications. • Pregnancy Contraindications: The patient must be in the first trimester of pregnancy and The prescribed medication has a proven teratogenic effect in humans. • Breastfeeding Contraindication: The rate at which the prescribed drug passes into breast milk (RID) Such as if it is higher than 10% and the baby is premature or less than 1 month old. The presence of a combination of sensitive situations. When any of these strict rules are triggered, the engine (200) will not respond to all other warnings. ignores a command that technically blocks the approval of the relevant prescription. It produces and is the single most important indicator that explains the situation to the doctor and suggests alternative treatments. It presents the warning. ii. Risk Score Based Ranking (210): If there are no strict rules in the first stage - 6 - If not triggered, the motor (200) proceeds to the second stage. In this stage, each raw stimulus, A clinical risk score (RS) according to a predefined set of rules (e.g., 0-100) (between 0 and 10) are assigned. These scores are given on a scale of 0-10 for better understanding in the user interface. can be normalized to a scale. Motor (200) assesses all actions according to these risk scores. It categorizes the possible warnings (e.g., Critical, Moderate, Low) and informs the physician of the best course of action. It presents them in a list, ranked starting from the highest risk. For example; 80 and An RS score above 50 is classified as "Critical," while a score between 50-79 is classified as "Moderate." They can be classified. This means the system will not only assign the highest-scoring warning, but also clinically. It allows the system to present all relevant warnings in order of importance. iii. Warning Merging and Simplification (230): The final step of the algorithm is to merge and simplify with each other. It is about combining warnings that are related or based on the same root cause. For example, a drug both because it was unsuitable according to geriatric criteria and because of renal dysfunction. If dose adjustment is required, the system displays these two warnings: “Elderly patient, kidney Due to its function (Creatinine Clearance: X mL / min), the dose of drug Y is changed to Z. a single, comprehensive statement such as "it should be reduced and its use should be re-evaluated" It can be turned into a warning. At the end of this process, the Suggestion and Warning Engine (200) is in the physician's user interface. It produces the final, ordered and simplified warning list (240) to be displayed. The application of the invention as a method is shown in the flowchart in Figure 2. The method begins with a physician entering a new prescription into the system (301). The system then enters the patient's It retrieves the existing data (medicines, diagnoses, etc.) from the database (302). Then, Data Current laboratory data from external systems such as E-Nabız via the Integration Module (110). It automatically retrieves the data (303). Simultaneously, the system retrieves Drug Interaction Analysis data. (304), Geriatric Compliance Analysis (305), Pharmacogenetic Analysis (306) and Pharmacovigilance It runs the steps of the analysis (307) in parallel or sequentially. All of this analysis The outputs obtained from the steps are sent to the Suggestion and Warning Engine (200) (308). Engine (200), primarily strict rule control and then risk score based as described above. It generates the final warning list by running the sorting and merging algorithm (309) and this The list is displayed on the physician's screen (310). The table below summarizes the system's modular structure, data flow, and technical logic: - 7 - Table 1: System Modules, Data Flow, and Technical Logic Element No. Element Name Function Data Inputs Processing Logic / Technique Features Data Outputs (Suggestion (To its engine) 100 Clinical Decision Support The system All modules containing and main operator system. User Interface (physician input: (Prescription), External Data Resources (APIs) Data flow between modules manages, on the processor It works, it communicates with the database. establishes. 110 Data Integrasy on Module E-Nabız and lab from their systems It extracts data and It standardizes. E-Nabız API call, LIS (Laboratory Information) (System) connection Incoming data (e.g. JSON, HL7) system internal data It converts it to a model. Specific parameters (Creatinine, ALT, K+) It separates. Standardization broken patient lab data. 120 Medicine Interaction Analysis Module Medicine, medicine, and medicine... food their interactions It detects. Prescription data, patient's current List of medications external / internal Interaction database Advanced rule set or external database (e.g. Micromedex API) using pharmacokinetics / pharmacodynamics micro interactions control It does. Clinical interactions. according to importance level (e.g. (Major, Moderate, Minor) It classifies. Classified was potential interaction list and importance degree. 130 Geriatric Rapport Module Geriatric in patients unsuitable prescriptions determines. Prescription data, patient age, diagnostic codes "Criteria Specific to Türkiye" (e.g., in the Turkish population) potassium salt usage + RAAS blocker local like combination dietary habits (rules based on) It applies the rule set. International Beers, START / STOP deviating from the criteria It includes the points. Potential unsuitable medicine usage or treatment neglect warnings. 140 Pharmacologist genetics Module The patient genetic according to your profile drug-dose optimization does. Prescription data, patient pharmacogenetic test result (e.g. CYP2D6, HLA- B*1502) Based on CPIC guidelines A rule starts the engine. Genotype-phenotype According to the match, "Poor Metabolizer", "Ultra-rapid Metabolizer" etc. situations dose adjustment or It generates alternative medicine suggestions. Specific drug dose for adjustment suggestion or "in use" avoid" warning. 150 Pharmacovi Jilans TÜFAM analyze data The prescribed medication, From the TÜFAM API National for a specific drug adverse event reporting For this medicine in the last 3 months - 8 - Module increasing It identifies the risks. incoming immediate adverse Event reporting data statistically frequent a significant increase time series that it is not It determines this through analysis. adverse event notifications at X% increase" trend warnings. 200 Suggestion and Warning engine All from modules incoming raw The warnings work, prioritizes and to the doctor presents. Modules (110- 150) all incoming data outputs Two-stage prioritization It applies the algorithm: 1) Bypassing the scoring and a "hard" blocker that blocks the prescription rules" (e.g., PGx, pregnancy) (contraindications) control (220). 2) Strict rule Otherwise, despite all the warnings, the clinic... It assigns a risk score (RS) and in order of importance (Critical > Medium > Low) ranks (210). 3) The same action combines the necessary warnings (230). The doctor user in the interface will be shown final, sorted And unified warning list (240). - 9 -

Claims

REQUESTS 1. A clinical device containing a processor and memory, intended for use in personalized treatment. decision support system (100) and its feature is; i. a patient's national health record system (E-Nabız) ​​and / or a laboratory Automatically retrieve up-to-date laboratory data from the information system (LIS). at least one Data Integration Module (110) structured to ii. Potential drug interactions between the patient's current medications and the newly prescribed medications. to identify drug and drug-food interactions and to determine their clinical significance At least one Drug Interaction structured to classify according to its degree. Analysis Module (120), iii. mentioned Data Integration Module (110) and Drug Interaction Analysis Module (120) including analysis outputs from multiple analysis modules to take as input and the aforementioned analysis outputs, primarily, scoring bypassing the system and blocking prescription approval, previously controlling according to a defined set of hard rules (220) and a hard rule If not triggered, each analysis output will have a predefined value. by operating according to a scoring algorithm (210) that assigns a clinical risk score (RS), a single combined and prioritized alert to be presented to a physician A Suggestion and Alert Engine (200) configured to generate a list (240) includes It is characterized by...

2. According to claim 1, the system is (100) and its characteristic is; inappropriate medication in geriatric patients. the use or neglect of treatment, pharmacotherapy specific to the Turkish population at least structured to detect according to a rule set that includes its parameters. It includes a Geriatric Adaptation Module (130) and the aforementioned Suggestion and Warning Engine (200) is that it also processes the outputs from this module.

3. According to claim 2, the system is (100) and its feature is; the aforementioned Geriatric Adaptation Module. (130) the set of rules included in the patient’s RAAS blocker use, potassium salt based on consumption and age or renal function (eGFR) value - 10 - an assessment of the risk of hyperkalemia and taking local dietary habits into account It contains rule (TR-G3).

4. System (100) according to Claim 1 or Claim 2, and its characteristic is; the patient's pharmacogenetic test to generate recommendations for drug selection and / or dose optimization based on the results It should include at least one structured Pharmacogenetic Module (140) and the aforementioned Recommendation and the Warning Engine (200) also processes the outputs from this module.

5. According to claim 4, the system is (100) and its characteristic is; the aforementioned Pharmacogenetic Module (140), Based on Clinical Pharmacogenetics Application Consortium (CPIC) guidelines It includes a rule engine. According to claim 1, the system is (100) and its characteristic is; Turkish Pharmacovigilance Centre. By retrieving real-time adverse event reporting data from the (TÜFAM) database, prescribed a statistically significant increase in the national adverse event reporting frequency for a drug at least one Pharmacovigilance Module configured to determine whether or not it exists (150) includes and the aforementioned Suggestion and Warning Engine (200) comes from this module It also processes the outputs.

7. According to Claim 1, the system is (100) and its feature is the aforementioned Suggestion and Warning Engine. (200) includes a strict set of rules (220), pharmacogenetic contraindications, pregnancy rules regarding contraindications and / or breastfeeding contraindications It includes.

8. According to Claim 1, the system is (100) and its feature is the aforementioned Suggestion and Warning Engine. (200), different analyses that are based on the same root cause or are related to each other. Multiple alerts from the modules are compiled into a single, comprehensive report for the physician. It includes an alert merging logic (230) that converts it into an alert text.

9. A device implemented by a processor, for use in personalized treatment. It is a clinical decision support method, and its characteristic feature is; - 11 - i. a patient’s national health record via a Data Integration Module (110). updated from the system (E-Nabız) ​​and / or a laboratory information system (LİS) Automatic acquisition of laboratory data (303), ii. through a Drug Interaction Analysis Module (120) with the patient’s current medications Potential drug-drug and drug-nutrient interactions among newly prescribed medications Identifying their interactions and classifying them according to their clinical significance classification (304), iii. and by a Suggestion and Alert Engine (200), the aforementioned data retrieval (303) and multiple analyses including interaction detection (304) steps Obtaining the analysis outputs from step (308) and the engine mentioned above The analysis results, primarily, bypass the scoring system and the prescription. to a predefined set of strict rules (220) that blocks its approval According to the analysis, if a strict rule is not triggered, each analysis a scoring system that assigns a predefined clinical risk score (RS) to its output by operating according to its algorithm (210), a single one to be presented to a physician. (309) producing a combined and prioritized alert list (240) It includes the steps.

10. The method according to Claim 9 is characterized by being via a Geriatric Adaptation Module (130), Inappropriate drug use or neglect of treatment in geriatric patients, Turkish determined according to a rule set that includes population-specific pharmacotherapy parameters (305) includes the step of doing.

11. Method according to Claim 9 or Claim 10, and its characteristic is a Pharmacogenetic Module (140) through this method, drug selection and / or based on the patient's pharmacogenetic test results. This includes the step of generating dose optimization recommendations (306).

12. The method according to Claim 9 is characterized by being a Pharmacovigilance Module (150), Real-time adverse event data from the Turkish Pharmacovigilance Center (TÜFAM) database. By extracting notification data, national adverse event reporting is performed for a prescribed drug. to determine whether there is a statistically significant increase in frequency (307) It includes the step. - 12 -

Images