Ebook: Detection and Prevention of Adverse Drug Events
When someone enters hospital for examination or treatment, is given a doctor’s prescription, receives medication from a pharmacist or has his drugs administered by a nurse, he assumes that his health and welfare will benefit from this medical intervention. But the effectiveness of available therapies should also be viewed from the perspective of potentially negative consequences: adverse drug events (ADE). ADEs endanger patients’ safety and increase hospital costs, making reduction of preventable ADEs a crucial and challenging Public Health issue. This book presents papers from the workshop “Patient Safety through Intelligent Procedures in medication” focusing on the following topics: the identification of ADE and medication errors in hospital settings; the role of human and organisational factors on ADE and medication errors; Information and Communication Technologies to prevent or correct ADEs and medication errors. The papers in this book are the work of active scientific experts in the field and confront ideas and experiences arising worldwide, and in particular from several EU projects directed at resolving ADE problems. Here is an opportunity to find common ways for solving shared difficulties and make improvements happen.
Along with the continuously increasing sophistication and refinement of diagnostic procedures and therapeutic processes, the risk of Adverse Events occurring during a patient's hospitalization is also steadily rising up.
Most of the modern medications have a powerful therapeutic impact balanced by equally threatening poisonous potential side effects. Thus the risk of Adverse Drug Events (ADE) is rocketing, partly due to medication errors.
That is why Quality of Care and Patient Safety have become a general concern shared by a large community of healthcare professionals and patients.
Different strategies have been tried for tracking ADEs: reporting systems have proven to be useful but they are not exhaustive; records and chart reviews are effective in the detection of ADEs and for an evaluation of their prevalence, but these methods are time-consuming and their reproducibility is questionable. Automatic detection of ADE is still in the research domain.
Nevertheless, at the bottom line, all the teams in the world are facing identical problems: how to reliably detect ADEs, how to efficiently prevent them. War stories about ADEs are all alike, whatever the country and whatever the setting: hemorrhage following an inattentive prescription of NSAID on a bleeding ulcer, renal insufficiency after heparin treatment, reactions to antibiotics and antiviral drugs, etc.
Properly managing the modern therapeutic drugs requires mastering a large quantity of information and a high level of complex knowledge, both probably exceeding human capacities. Therefore Information and Communication Technologies (ICT) are called in for additional help to access medical records, manage the data, enhance the knowledge, perform statistical analyses, and provide Decision Support.
But despite this powerful technology support many problems remain.
The characteristics of the Electronic Health Records (EHR) vary according to the suppliers, the hospitals, the countries. Coding systems can be different. Norms and standards vary from place to place. The extraction and exploitation of medical data from different sources then appear as a real challenge and cannot be solved without the definition of common data models, standard references, classifications and taxonomies.
The classification of drugs is in itself a tower of Babel: brand names, commercial names, ATC codes, various dosages, regimens, routes of administration, pharmacological effects, potential ADE or allergies, etc. The knowledge is mastered by a limited community of experts while the complexity of the interactions and the frequency of the side effects would require that this knowledge be shared by a large number of health-care professionals and by the patients themselves.
Moreover, hospitals are much more than a mere technological setting: they are organized by administrative managers, they are ruled by physicians and nurses who are not robots, they treat citizens unwillingly playing the patient role. That is why organizational and human factors must be carefully studied as they can be at the origin or contribute to the occurrence of an ADE.
The new possibilities offered by statistical methods, data mining applications able to exploit large amount of records open a new era in the identification of abnormal hospitalization stays and detection of potential ADEs. Exploitation and screening of free-text documents such as letters and reports are now possible by means of semantic mining applications, providing new sources of patient information.
When looking at the papers gathered in this book, it appears that the approaches used to identify and prevent ADE are diverse but confronted to similar problems of codification, data exploitation, statistical analysis, etc. The ICT bring in powerful resources but they have to be driven by clear scientific objectives.
This workshop gathers knowledgeable and active persons in the domain, who are currently working at solving the problems. This is the opportunity to confront the ideas and the experiences stemming from various continents, and particularly from several EU projects financed to contribute to the resolution of the ADE problem. It is the occasion to find common ways for solving some of the difficulties everyone is sharing and make progress happen. Many thanks to Prof. David Bates, Prof. Johanna Westbrook for their participation and their keynotes; to Prof. Peter Elkin and Prof. André Kushniruk for their support and their clarifying interventions; to all the participants to the workshop; and to the European Commission which, by funding European Projects on this topics, allowed the organization of this workshop and the edition of this book.
Régis Beuscart, Werner Hackl and Christian Nøhr (Editors), September 2009
The European project Patient Safety through Intelligent Procedures in medication (PSIP) aims at preventing medical errors. The objective are: (1) to facilitate the systematic production of epidemiological knowledge on Adverse Drug Events (ADE) and (2) to improve the entire medication cycle in a hospital environment. The first sub-objective is to produce knowledge on ADE: to know, as exactly as possible, per hospital, per medical department, their number, type, consequences and causes, including human factors. Data Mining of structured hospital data bases, and semantic mining of free-texts will provide a list of observed ADE, with frequencies and probabilities, thus giving a better understanding of potential risks. The second sub-objective is to develop innovative knowledge based on the mining results and to deliver professionals and patients contextualized alerts and recommendations fitting the local risk parameters. This knowledge will be implemented in a PSIP-Platform independent of existing ICT applications.
This paper addresses the question of the integration of Human Factors (HF) methods and models within projects aiming at (semi-) automatically identifying and preventing Adverse Drug Events (ADE). While more traditional methods such as voluntary reporting systems of medication errors tend to focus on HF causes of preventable ADEs, computer-based screening and mining methods tend to rely on a medical model of ADEs. As a consequence, HF methods and concepts are rarely considered in those projects. The paper describes the way HF methods have been incorporated in the PSIP (Patient Safety through Intelligent Procedures in medication) project lifecycle. It provides some examples of the results obtained and demonstrates their relevance to improve the entire detection and prevention process.
The number of medication errors reported to The Danish National Board of Health in Denmark exceeds 5000 per year. It is well known that computerized physician order entry (CPOE) with addition of decision support tools may reduce the frequency of medication errors. The primary scope of the work in Denmark has been to help health care professionals avoiding harmful errors. Using data primarily from The Danish National Board of Health, based on the reports of errors from Danish hospitals, and with our previous foundation in the international literature, we analyzed the errors which led to harmful conditions or death. In the process we developed a methodical consensus for identifying which medicines should have a warning attached, and we systematized the different kind of warnings. The following validation of the data resulted in a final list of 14 classes of drugs or drug substances, which all have been involved in serious medication errors. At present time there is a total of 136 different medicines with warnings found in the drug database for health professionals from Infomatum A/S (www.medicin.dk). In a parallel matter other decision support tools from Infomatum A/S
Infomatum A/S is a company jointly owned by The Danish Drug Information and The Danish Medical Association. Infomatum A/S publishes drug information for health professionals and the public.
Infomatum A/S is a company jointly owned by The Danish Drug Information and The Danish Medical Association. Infomatum A/S publishes drug information for health professionals and the public.
The ReMINE project aims at building a high performance prediction, detection and monitoring platform for managing Risks against Patient Safety (RAPS). The project will contribute to the optimization of RAPS management process in a healthcare system through the development of a platform allowing the (semantically based) fast and secure extraction of RAPS-related data and their correlation across several domains. In this respect the REMINE platform will promote early RAPS detection and mitigation by supporting the process of RAPS management both when a RAPS is foreseen, and the objective is the determination of the best set of preventive actions; and when a RAPS is detected, and the objective is the determination of the best possible reaction, the reliable distribution of the related action list to all involved parties, and the monitoring of the reaction effectiveness. These capabilities will be achieved by means of the establishment of an associated methodology and a framework/platform for integrated RAPS prediction/detection, analysis and mitigation. The overall platform structure assumes the presence of an “info-broker patient safety framework” connected with the Hospital Information System, which will support the process of collecting, aggregating, mining and assessing related data, distributing alerts, and suggesting actions to mitigate (or avoid) RAPS effects or occurrence. The underlying ontological system will support the semantic correlation of data with the hospital processes.
The EU-ADR project aims to exploit different European electronic healthcare records (EHR) databases for drug safety signal detection. In this paper we describe the project framework and the preliminary results. Methods: As first step we created a ranked list of the events that are deemed to be important in pharmacovigilance as mining on all possible events was considered to unduly increase the number of spurious signals. All the drugs that are potentially associated to these events will be detected via data mining techniques. Data sources are eight 8 databases in four countries (Denmark, Italy, the Netherlands, and the United Kingdom) that are virtually linked through harmonisation of input data followed by local elaboration of input data through custom-built software (Jerboa©). All the identified drug-event associations (signals) will be thereafter biologically substantiated and epidemiologically validated. To date, only Upper gastrointestinal bleeding (UGIB) event has been used to test the ability of the system in signal detection. Results: An initial ranked list comprising 23 adverse events was identified. The top-ranking events were: cutaneous bullous eruptions, acute renal failure, acute myocardial infarction, anaphylactic shock, and rhabdomyolysis. Regarding the UGIB test, a total of 48,016 first-ever episodes were identified. The age-standardized incidence rates of UGIB varied between 40-100/100,000 person-years depending on country and type of healthcare database. A statistically significant association between use of NSAIDs and UGIB was detected in all of the databases. Conclusion: a dynamic ranked list of 23 adverse drug events judged as important in pharmacovigilance was created to permit focused data mining. Preliminary results on the UGIB event detection demonstrate the feasibility of harmonizing various health care databases in different European countries through a distributed network approach.
The expansion of clinical information systems and the reduction in computing costs have led to an explosion of patient data available for reuse. However, this data is rarely combined and analyzed in an integrated manner. The DebugIT project is a large-scale integrating project funded within the 7th EU Framework Programme (FP7). The main objectives of the project are to build IT tools that should have significant impacts for the monitoring and the control of infectious diseases and antimicrobial resistances in Europe; this will be realized by building a technical and semantic infrastructure able to a) share heterogeneous clinical data sets from different hospitals in different countries, with different languages and legislations; b) analyze large amounts of this clinical data with advanced multi-modal data mining; c) apply the obtained knowledge for clinical decisions and outcome monitoring. The concepts and architecture underlying this project are discussed.
Our main objective is to detect adverse drug events (ADEs) in former hospital stays. As ADEs are rare, that supposes to screen thousands of electronic health records (EHRs). For that purpose, we need to define a data model that has two main objectives: (1) being able to describe hospital stays from various hospitals (2) being tuned so as to prepare the data mining process: as ADEs are not flagged in the datasets, the data model must be optimized for ADE detection. The article presents the phases of the design and the data model that results from this work. It is compatible with many hospitals. It deals with diagnoses, drug prescriptions, lab results and administrative information. It allows for data mining and ADE detection in EHRs.
Adverse drug events (ADEs) are a public health issue. The objective of this work is to data-mine electronic health records in order to automatically identify ADEs and generate alert rules to prevent those ADEs. The first step of data-mining is to transform native and complex data into a set of binary variables that can be used as causes and effects. The second step is to identify cause-toeffect relationships using statistical methods. After mining 10,500 hospitalizations from Denmark and France, we automatically obtain 250 rules, 75 have been validated till now. The article details the data aggregation and an example of decision tree that allows finding several rules in the field of vitamin K antagonists.
An important part of adverse drug events (ADEs) detection is the validation of the clinical cases and the assessment of the decision rules to detect ADEs. For that purpose, a software called “Expert Explorer” has been designed by Ideea Advertising. Anonymized datasets have been extracted from hospitals into a common repository. The tool has 3 main features. (1) It can display hospital stays in a visual and comprehensive way (diagnoses, drugs, lab results, etc.) using tables and pretty charts. (2) It allows designing and executing dashboards in order to generate knowledge about ADEs. (3) It finally allows uploading decision rules obtained from data mining. Experts can then review the rules, the hospital stays that match the rules, and finally give their advice thanks to specialized forms. Then the rules can be validated, invalidated, or improved (knowledge elicitation phase).
The objective of this research is to assess the suitability of the Apriori association analysis algorithm for the detection of adverse drug reactions (ADR) in health care data. The Apriori algorithm is used to perform association analysis on the characteristics of patients, the drugs they are taking, their primary diagnosis, co-morbid conditions, and the ADRs or adverse events (AE) they experience. This analysis produces association rules that indicate what combinations of medications and patient characteristics lead to ADRs. A simple data set is used to demonstrate the feasibility and effectiveness of the algorithm.
Adverse drug events are a public health issue (98,000 deaths in the USA every year). Some computerized physician order entry (CPOEs) coupled with clinical decision support systems (CDSS) allow to prevent ADEs thanks to decision rules. Those rules can come from many sources: academic knowledge, record reviews, and data mining. Whatever their origin, the rules may induce too numerous alerts of poor accuracy when identically applied in different places. In this work we formalized rules from various sources in XML and enforced their execution on several medical departments to evaluate their local confidence. The article details the process and shows examples of evaluated rules from various sources. Several needs are enlightened to improve confidences: segmentation, contextualization, and evaluation of the rules over time.
Objective: The objective of this work is to create a bilingual (French/English) Drug Information Portal (DIP), in a multi-terminological context and to emphasize its exploitation by an ATC automatic indexing allowing having more pertinent information about substances, organs or systems on which drugs act and their therapeutic and chemical characteristics. Methods: The development of the DIP was based on the CISMeF portal, which catalogues and indexes the most important and quality-controlled sources of institutional health information in French. DIP has created specific functionalities and uses specific drugs terminologies such as the ATC classification which used to automatic index the DIP resources. Results: DIP is the result of collaboration between the CISMeF team and the VIDAL Company, specialized in drug information. DIP is conceived to facilitate the user information retrieval. The ATC automatic indexing provided relevant results in 76% of cases. Conclusion: Using multi-terminological context and in the framework of the drug field, indexing drugs with the appropriate codes or/and terms revealed to be very important to have the appropriate information storage and retrieval. The main challenge in the coming year is to increase the accuracy of the approach.
A concept-based terminology that covers all features of healthcare is essential for the development of an Electronic Health Record (EHR). Since a significant percentage of the EHR can be drug related information, we decided to implement the controlled drug terminology provided by SNOMED CT to achieve the potential benefit to promote Patient Safety that a fully functional pharmacy system can offer. One of the expected advantages of our Project is to establish a bridge between reference terminology and the drug knowledge databases. There is also an economic advantage of implementing a “clinical drug product”, the one defined by the drug name, its strength and dose form, instead of the manufactured drug product. The Pharmacy economic management of stocks and response to the offers from the pharmaceutical companies is another expected asset of the Project. This Project is intended as well to give support to a more widespread objective of interoperability with the Primary Care systems.
Adverse Drug Events (ADEs) are currently considered as a major public health issue, endangering patients' safety and causing significant healthcare costs. Several research efforts are currently concentrating on the reduction of preventable ADEs by employing Information Technology (IT) solutions, which aim to provide healthcare professionals and patients with relevant knowledge and decision support tools. In this context, we present a knowledge engineering approach towards the construction of a Knowledge-based System (KBS) regarded as the core part of a CDSS (Clinical Decision Support System) for ADE prevention, all developed in the context of the EU-funded research project PSIP (Patient Safety through Intelligent Procedures in Medication). In the current paper, we present the knowledge sources considered in PSIP and the implications they pose to knowledge engineering, the methodological approach followed, as well as the components defining the knowledge engineering framework based on relevant state-of-the-art technologies and representation formalisms.
Clinical decision support systems (CDSS) are the new generation clinical support tools that ‘make it easy to do it right’. Despite promising results, these systems are not common practice, although experts agree that the necessary revolution in health care will depend on its implementation. To accelerate adoption a strategy is handed for structured development and validation of CDSS' content (clinical rules). The first results show that the proposed strategy is easily applicable for creating specific and reliable rules, generating relevant recommendations.
In this article, we will try to address the most basic requirements for facilitating the knowledge management challenges through the elaboration of medical documentation/ record keeping with several implications on patient safety/medication safety and research quality aspects, the main purpose being the simplification of utilizing the usable outputs of ontology development efforts. This simplification is of vital importance from KM implementation in medical and healthcare domains. Because, as Ceusters et al  elaborate, reaching consensus on even the most basic concepts will become an intricate work in terms of the wide-scale implementation of ontology-based KM solutions in clinical practice and other healthcare related processes. EHR (Electronic Health Records) standards developed by various SDOs
SDOs: Standards Development Organizations
SDOs: Standards Development Organizations
The purpose of this study is to examine how everyday use of the Computerised Physician Order Entry (CPOE) system in the Capital Region of Denmark has led to medication errors. The study is based on clinicians' reporting of patient safety incidents. It was found that the immediate causes of the patient safety incidents primarily relates to a) a mismatch between clinical work routines and the structure of the CPOE system, b) the complexity of the user interface, and c) lack of barriers against commonly occurring, severe errors in some areas of the CPOE system. The following was concluded: A well designed CPOE system should be intuitive, provide barriers against serious mistakes, and make the correct choice an easy one. Furthermore it was concluded that it is important that the CPOE system closely supports accepted clinical work routines and that risk assessment is performed prior to implementing new design or functionality.
Ten years ago research of the impact of health information technology (HIT) on medical work practices started at Erasmus MC. The research is characterized by practice driven field research. From the beginning computerized physician order entry systems (CPOE) were a major topic. Research questions were how implementation of CPOE could be understood, how physicians were responding to reminders and alerts and how CPOE impacted professional workflow and collaboration. Studies of CPOE implementation aimed to understand why the adoption rate is so low and riddled with difficulties. Studies of reminders and alerts addressed the problem of alert fatigue. Finally, studies of workflow explored how CPOE influenced clinical workflow and how simplistic and linear models underlying CPOE may lead to poor designed systems and even compromise patient safety. Findings include the need for a shared understanding of medical challenges when implementing CPOE, conceptual models to understand alert fatigue and medical workflow and the impossibility of agreeing which alerts to suppress hospital-wide. The underlying research principle is the sociotechnical approach, which stipulates that technology, people and organizations should be studied from a single theoretical framework. This paper summarizes the results of ten years of research.
The European project PSIP (Patient Safety through Intelligent Procedures in Medication) aims at semi-automatically identifying and preventing ADE. Automatically detected Adverse Drug Events have to be reviewed and validated by human experts. Existing methods usually have the experts review the cases and document their rating in a structured form. One of the limitations of these methods is their poor ability to analyze and clear the disagreements between the experts and the system. This paper presents an innovative Human Factors based method to support the review by clinicians and pharmacologists of these automatically detected ADE. We use think aloud methods and portable labs to track and record the experts reasoning and their reviewing cognitive procedures. We present preliminary results obtained with this method, which allows identifying the key data and information used to characterize the ADE. This method provides useful feedbacks allowing a continuous refinement and improvement of the automated detection system.