Atrial fibrillation (AF) is the most common tachyarrhythmia that requires treatment and represents constant clinical problem for general practitioners and cardiologists. Several bleeding risk scores have been developed for estimating bleeding risk in patients with AF. These include: HAS-BLED (hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile international normalized ratio, age >65 years, drugs/alcohol concomitantly), ORBIT (older age, reduced hemoglobin/hematocrit/anemia, bleeding history, insufficient kidney function, treatment with anti-platelets), ABC (age, biomarkers, clinical history), ATRIA (anemia, severe renal disease, age ≥75 years, previous hemorrhage, and diagnosed hypertension), and HEMORR(2)HAGES (Hepatic or Renal Disease, Ethanol Abuse, Malignancy, Older Age, Reduced Platelet Count or Function, Re-Bleeding, Hypertension, Anemia, Genetic Factors, Excessive Fall Risk, and Stroke). The use of oral anticoagulants is still the standard in stroke prevention in AF but should be balanced against the associated bleeding risk. The aim of this article was to describe the development of a clinical decision support system (CDSS) that will enable clinicians to perform a quick assessment of bleeding risk in patients with AF in order to optimize anticoagulation therapy in patients with AF. The software was developed in the form of a web application. The responsive design of the interface was key to optimal user interaction, providing seamless control of every step of the process regardless of the type of device used, whether a laptop or a smartphone. The backend of the application was developed in Python. More specifically, a web framework named Flask was utilized. It is considered to be a good choice for rapid prototyping and development and deployment of small- to medium-sized applications. The application separates the decision process into three steps. Displaying the first step prompts the user to select the type of score they want calculated. The following step includes entering anamnestic data, laboratory findings, symptoms, and comorbidities. The final screen displays the calculated score, which assists the user in determining the course of the treatment. This software represents a CDSS that enables faster and easier assessment of bleeding risk in patients with AF in order to achieve a better therapeutic modality. The responsive design and the web application format makes the software easily accessible on a wide range of devices.
Fibrilacija atrija (FA) najčešća je tahiaritmija koja zahtijeva liječenje te je stalni klinički problem za liječnike obiteljske medicine i kardiologe. Razvijeno je nekoliko algoritama za procjenu rizika krvarenja u bolesnika s FA-om. Među njima su HAS-BLED (arterijska hipertenzija, abnormalna funkcija jetre i bubrega, moždani udar, anamnestički podatci ili predispozicija za krvarenje, labilna vrijednost INR-a, dob >65 godina, istodobno konzumiranje droge i alkohola), ORBIT (starija životna dob, snižena vrijednost hemoglobina/hematokrita/anemija, anamnestički podatci o krvarenju, snižena bubrežna funkcija, liječenje antitromboticima), ABC (životna dob, biomarkeri, anamnestički podatci), ATRIA (anemija, teško smanjenje bubrežne funkcije, dob >75 godina, prethodno krvarenje i dijagnosticirana arterijska hipertenzija) i HEMORR(2)HAGES (bolest jetre ili bubrega, alkoholizam, zloćudna bolest, starija životna dob, smanjen broj ili funkcija trombocita, ponovno krvarenje, arterijska hipertenzija, anemija, genski čimbenici, znatan rizik od pada i moždani udar). Primjena oralnih antikoagulanasa još je uvijek standard u prevenciji moždanog udara u FA, ali je treba uravnotežiti s rizikom od krvarenja koji je s njom povezan. Svrha je ovoga članka opisati razvoj sustava za podršku pri donošenju kliničkih odluka (CDSS; eng. clinical decision support system) koje bi liječnicima omogućile brzu procjenu rizika od krvarenja u bolesnika s FA-om kako bi optimizirali liječenje antikoagulansima. Spomenuti je računalni program razvijen u obliku mrežne aplikacije. Responzivni ustroj korisničkog sučelja bio je ključan u postizanju optimalne interakcije korisnika s programom te korisniku omogućuje potpunu kontrolu pri svakom koraku postupka neovisno o vrsti uređaja koja se primjenjuje, bilo to prijenosno računalo bilo pametni telefon. Pozadinski sustav aplikacije razvijen je u programskom jeziku Python. Preciznije rečeno, rabi se mrežni kostur zvan Flask. On se smatra dobrim izborom za brzo prototipiziranje, razvoj i uvođenje malih do srednjih aplikacija. Aplikacija razdvaja postupak odlučivanja u trima koracima. Prikaz prvog koraka traži od korisnika da izabere vrstu zbroja rizika koji želi izračunati. Sljedeći korak uključuje unošenje podataka o povijesti bolesti, laboratorijskim nalazima, simptomima i komorbiditetu. Posljednji ekran prikazuje izračunani zbroj rizika, koji pomaže korisniku u odabiru tijeka liječenja. Ovakav program nudi CDSS koji omogućuje bržu i lakšu procjenu rizika krvarenja u bolesnika s FA-om kako bi se postigao bolji terapijski modalitet. Responzivni ustroj i sučelje u obliku mrežne aplikacije osiguravaju lako pristupanje programu s pomoću širokog raspona uređaja.
Atrial fibrillation (AF) is the most common tachyarrhythmia that requires treatment and presents an almost ubiquitous clinical problem for general practitioners and cardiologists. (
Incidence of AF is 0.4% in adults up to 60 years of age, in 2.0-5.0% in people who are older than 60 years, and 12.0% in patients over 75 years. It is more likely to occur in men. (
Lack of atrial contraction and an altered and dilated left ventricle result in thrombus formation. Treatment of AF is primarily based on elimination of symptoms or achievement of sinus rhythm in persistent AF or regulation of the ventricular frequency in permanent AF, as well as the prevention of thromboembolic diseases. (
The most commonly used oral anticoagulants are coumarin derivatives, of which warfarin is the most commonly used. (
Liver disease can influence the synthesis of coagulation factors. Furthermore, there are conditions where elevated levels of metabolic processes (and therefore degradation of coagulation factors), such as high temperature or thyrotoxicosis, can also increase the effect of anticoagulant drugs. Many drugs can potentially intensify the effects of warfarin. (
All patients with AF and a CHA2DS2VASc score (used to assess the risk of thromboembolic incident) of 1 or over must receive oral anticoagulant therapy if there are no contraindications. (
There is no need for therapeutic monitoring and dose adjustments, as in the case of warfarin derivatives. Clinical trials on the effects of NOAC proved them to be as effective as warfarin in the prevention of stroke or systemic embolism. (
Although the use of NOAC is recommended, the use of vitamin K antagonists or warfarin derivatives is not disallowed. The therapeutic modality of patients with AF requires an assessment of the risk of bleeding. Bleeding is potentially the most dangerous side-effect of anticoagulant drugs (especially in the intestine or brain). (
Several bleeding risk scores have been developed for estimating bleeding risk in patients with AF. These include: HAS-BLED (hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile INR, elderly (>65 years), drugs/alcohol concomitantly), ORBIT (older age, reduced hemoglobin/ hematocrit/anemia, bleeding history, insufficient kidney function, treatment with anti-platelets), ABC (age, biomarkers, clinical history), ATRIA (anemia, severe renal disease, age ≥75 years, previous hemorrhage, and diagnosed hypertension), and HEMORR(2)HAGES (Hepatic or Renal Disease, Ethanol Abuse, Malignancy, Older Age, Reduced Platelet Count or Function, Re-Bleeding, Hypertension, Anemia, Genetic Factors, Excessive Fall Risk, and Stroke). (
Information technologies can have a big impact on medical practice in terms of developing decision support systems, which include all types of systems whose primary function is to take the information that can be used in the decision-making process. These are interactive, computer-based systems based on data and models that help solve problems and make decisions. Decision-making support systems in medicine are tools that contain established clinical knowledge and patient-oriented information to improve patient care. Their purpose is to assist in interaction between the patient and the physician from the moment of initial consultation, throughout the diagnostic process, and in the follow-up period. Time constraints caused by constant evolution of the standard of care contribute to the number of medical mistakes and delays in clinical decisions. (
The use of decision support systems is widespread in all aspects of professional activity, and such systems that support decision-making for medical and healthcare purpose are called clinical decision support systems (CDSSs). (
CDSSs vary by type and complexity. The systems can be passive (the user explicitly sends the support request), semi-active (observation systems that are automatically executed but present information only at user request), and active (automatically activated, present information without waiting on request, and sometimes make a decision without interaction with a healthcare professional). (
The simplest example of the CDSS checks the input data entered by the healthcare professional and checks for values within the reference range. The end result is a certain form of notice or reminder. CDSSs of medium complexity include prognostic calculators and automated guidelines in clinical practice. Prognostic calculators are used for automatic forecasting, usually based on clinical scoring systems. Complex CDSSs use artificial intelligence, data mining, or statistical methods for classifying or predicting a disease or condition of a patient. (
These methods automatically identify the key features that are important for clinical classification or anticipation of problems and use mathematical tools to determine the way these characteristics should be combined with the goal of creating output that represents classification or prediction.
CDSS are classified into six categories: medication dosing support, order facilitators, point-of-care alerts/reminders, relevant information display, expert systems, and workflow support systems. (
The aim of article was to describe the development of software that will enable clinicians to perform a quick assessment of bleeding risk in patients with AF in order to optimize anticoagulation therapy in patients with AF (especially in patients who use vitamin K antagonists in therapy).
The software was developed in the form of a web application. Responsive design of the interface was key to optimal user interaction, providing seamless control of every step of the process regardless of the type of device used, whether a laptop or a smartphone. The backend of the application was developed in Python. More specifically, a web framework named Flask was utilized. It is considered to be a good choice for rapid prototyping and development and deployment of small- to medium-sized applications. To deliver optimal user experience across different devices, the front-end was built using the Bootstrap open source library, currently in version 4. The application was deployed on the Heroku platform and can be accessed via the following link:
The process is separated into three steps. The first step prompts the user to select the type of score they want calculated (
First step – the user has to choose which score to use.
The following step includes entering anamnestic data, laboratory findings, symptoms, and comorbidities (
Input of data for score calculation.
The final screen displays the calculated score, which assists to user in determining the course of the treatment (
Calculated score.
CDSSs are common in medical practice, and in essence represent a computer-aided diagnosis systems, which assist physicians in establishing the diagnosis for a patient or directly participate in the process of treatment or in the follow-up period. (
The use of software that is easily accessible and often made by non-medical professionals requires clear information regarding references and data on the basis of which the software was made (this is clearly noted in this software). The clinical validation of such software must be something that is in the interest of the software engineer, which makes their product important, but also in the interest of the physician, who will consequently be confident in using this kind of software. The disadvantage of the software, besides the generational resistance towards its use, is that the choice of using a particular score depends on clinical experience of the physicians. It must also be considered that the choice of the score itself is not a solution and only facilitates the obtainment of a solution, in addition to of course being, in a subjective sense, a loss of authority through the use of information technology in clinical work.
Although the benefits and potential of CDSSs are quite large, their use is still seen as dubious, especially among older generations. The fact is that CDSS must be based on correct, clinically proven information. CDSSs can only be as effective as the strength of the underlying evidence base, meaning that up-to-date information from the best reference databases must be used. (
This software represents a CDSS, which enables faster and easier assessment of bleeding risk in patients with AF, leading to a better therapeutic modality. Easy access to this type of CDSS is of great help to physicians in everyday practice and allows them a higher quality decision-making regarding the therapeutic modality of the patient. The responsive design and delivery in the form of a web application makes the software easily accessible on a wide range of devices.