|
PSIP: Patient Safety through Intelligent Procedures in Medication
The prevention of Adverse Drug Events that may harm patients and generate unproductive extra healthcare costs is a major public health issue. The PSIP project aims at developing innovative computer-based applications able to automatically detect situations at risk and to deliver healthcare professionals and patients relevant “ad hoc” information helping them to prevent medication errors.
1.1 PROJECT CONTEXT AND OBJECTIVES
There are few statistics on medication errors and adverse drug events. In Europe, it is estimated that more citizens die from adverse drug events than by car accidents, the risk being more important for ageing patients suffering from multiple pathologies.
The first goal of PSIP project is to automatically generate knowledge on adverse drug events. Most of the patient’s medical information is now available in electronic format and stored in large medical databases.
The second goal of the PSIP project is to deliver to the healthcare professional and to the patients the contextual knowledge helping them characterizing the problem and adapting the treatment to avoid the upcoming adverse drug events.
The overall PSIP project general objectives are to develop services (procedures, decision systems, prototypes) for Adverse Drug Events (ADE) detection, analysis and prevention:
- For retrospective ADE detection and analysis:
- To automatically detect past ADE cases, without the need for reporting systems
- To identify, by innovative Data and Semantic techniques, Healthcare situations where patient safety is at risk
- For prospective prevention of ADEs:
- To improve the decision support tool related to medication cycle
- To deliver to healthcare professionals and patients usable, efficient and contextualized alerts and just-in-time and point of care relevant information
- To demonstrate a significant reduction of patient risk in a subset of diseases and practices in the hospital setting
- To implement standardized knowledge based tools.
More in details, scientific objectives are:
- To get a better knowledge of the prevalence of Adverse Drug Events and of their characteristics, per hospital, per region and per country
- To develop concepts and methods to achieve the contextualization of Clinical Decision Support System functions
- To model the architecture enabling both the independency and the interrelation between the knowledge (CDSS modules) and the connecting applications (Hospital Information System, Computerized Physician Order Entry system)
Technical objectives are:
- To run semi-automatically Data and Semantic mining techniques on existing healthcare data repositories
- To develop a platform incorporating the CDSS modules and easy to connect to healthcare IT applications, primarily CPOE
- To design and develop a prototype of a support system for healthcare professionals integrated in the medication workflow
- To design and develop a prototype of a support system to help patients monitor their medication process
In the PSIP project, a human factors approach was adopted not only for the design of the Human Computer Interfaces, but also for supporting and orienting the knowledge elicitation task and a usercentered design of the prototypes. The Human Factor analysis of the sociotechnical system provides the consortium with the best possible analysis of the workflow of the medication process, and highlights the differences observable between departments, hospitals or regions. Usability methods have been intensively used for the evaluation of the quality of development of the prototypes, and their potential usage.
In 2010, the PSIP+ extension has been granted by the experts of the European Commission. The major task of PSIP+ is to validate the PSIP approach in another clinical setting (the University Specialized Hospital for Active Treatment of Endocrinology, USHATE), in the context of different hospital practices, with a different architecture of the Hospital Information System, and somewhat different (local) drugs. A semantic mining solution for extracting drug names and clinical data from hospital patient records had to be developed to be used to enrich automatically the computerized data. New rules had to be extracted from Data Mining to fit the endocrinology diseases treated in the USHATE hospital. Specific problems linked with the Cyrillic and Latin alphabets, both used by clinicians had to be solved in a generic way.
Finally, the implementation of the ADE Scorecards, and the integration of the Cx-CDSS with the Hospital Information System of the Bulgarian hospital had to be realized, and evaluation performed.
1.2 PROJECT PHASES
According to the Document of Work, the project is divided into three phases:
- Phase I (Year 1): Improve the Knowledge on Adverse drug Events
- Phase 2 (Year 2): Build Decision Support Services
- Phase 3 (Year 3- month 43) : Integration in existing solutions and usages.
This Periodic report concerns the third phase of the project during which many goals had to be reached and several applications provided and evaluated.
1.3 PROJECT RESULTS
The figure below provides an overview of the scientific and technical work carried out in the PSIP project and identifies the main corresponding scientific and technical results.
A primary scientific task was to clarify the concept of ADE and the scope of ADEs targeted by the project.
This resulted in the PSIP Taxonomy that provides a structured description of ADE for ICT based identification and prevention, along with a Data Model supporting the technical task of export of medical data from hospital databases into exploitable repositories.
The second major scientific task is knowledge discovery, carried out by applying data and semantic mining techniques on these exploitable data repositories. This task issued two major scientific results. The first one is the set of PSIP rules that are used (1) to retrospectively screen the medical records to find suspected ADE cases and (2) to design and feed the CDSS aiming at prospectively prevent ADEs. The second scientific result of this task is the innovative context model used to contextualize the CDSS.
The subsequent scientific tasks are (i) the systematic screening of databases to provide statistics and epidemiology information on ADEs and (ii) the design of the PSIP knowledge based system, that benefit from the results of the screening task to contextualize the CDSS modules. The corresponding scientific results are the statistic tables on ADEs, reporting their prevalence and characteristics, and the Cx-CDSS modules that will help to prevent them in the future. Another technical result at this stage is the Connectivity Platform that ensures proper interoperability with existing ICT systems.
All along these phases the Human Factors Engineering research process has been going on, analyzing the existing work systems to support the design of the ICT products aiming at improving their safety in terms of ADE prevention. The scientific result of this task is a set of recommendations and design principles for ICT based interventions, services and applications.
The next technical tasks consist in designing and developing all the PSIP prototypes to be implemented in the clinical or patient environment. The corresponding technical results or products are the ADE scorecards application and the CDSS prototypes: Web prototype, IBM-DK prototype, Medasys prototype and patient modules.
All along the project, scientific evaluation activities have been planned and carried out.
1.4 PROJECT ACHIEVEMENTS AND RESULTS DURING THE PERIOD
The two last years of the project (2010-11) have been devoted mainly to the execution of phase 3: integration in existing IT solutions and usages and to the evaluation of the potential impact of PSIP.
Therefore a number of tasks carried out during phase 1 and phase 2 went to iterations and generalization:
- Generalization of data extraction from various databases as in phase 1 & 2 (French and Danish Hospitals), and application to Bulgarian databases (USHATE Hospitals) through the PSIP+ extension have been performed.
- New iterations of data and semantic mining, new knowledge data discovery (KDD), refinement of the Knowledge Base, developments of the version 2 of the Knowledge Base have been performed. The Multi-Terminology Server (MTS) has been maintained and improved.
In the last two years, the following work has been achieved:
- Supports for Healthcare Professionals showing different ways to implement Contextualized Clinical Decision Support have been designed and developed (WP9): the PSIP-DK prototype, the Medasys prototype, the Web prototype for the hospital conditions and for a anytime-anywhere conditions (French, English and Bulgarian versions).
- Patient Support Service prototypes providing PSIP patient information have been designed and developed (WP10): the Patient component for the hospitalization context (French, English and Danish versions), the Web Patient component for the home context (French and English versions), the Mobile Patient component (English and Danish versions)
- Humans Factors and Usability tasks have been performed to ensure the user-centered design, development and evaluation of the PSIP applications (WP11): usability evaluations have been conducted on the PSIP applications as the Web prototype, the PSIP-DK prototype, the Medasys Prototype, the Patient component for hospitalization context, the Mobile Patient component, and ADE Scorecards.
- Prospective Impact Assessment assessing the potential impact (usefulness) of the PSIP solutions in terms of patient safety has been performed (WP12). All in all, nine different evaluation studies have been carried out including user surveys, Delphi study, simulation studies, and prospective case studies. The studies indicate that PSIP applications, especially the ADE Scorecards and the concepts implemented in the prototypes, are appreciated by the majority of future users and are promising with regard to medication safety. According to international experts’ opinions, all developed PSIP applications have the potential to prevent up to 20% of the ADEs. In addition, PSIP tools appear to be easily transferable to other settings.
- Dissemination and Exploitation tasks assuring the dissemination (workshops, conferences, papers, articles, websites, and flyers) and the exploitation (exploitation plan, medico-economical assessment, business plan) have been achieved (WP13). The dissemination work was increased during these last years, with the participation to many conferences and workshops, and the successful organization of the Second PSIP international Workshop. The minutes, papers and articles of this workshop have been published in the book: “Patient Safety Informatics”, subtitled “Adverse Drug Events, Human Factors and IT Tools for Patient Medication Safety”.
Societal impact:
The ultimate goal of PSIP is to improve the quality of care and patient safety through the use of Information and Communication Technology. The impact of PSIP applications will be significant if their implementation is properly integrated in quality procedures. They can efficiently support and enhance these procedures by different ways: (1) by regularly providing information on the causes of these ADE in a medical unit (2) by providing information on the causes of these ADE (3) by supporting the discussion with the healthcare professionals.
ADEs endanger Patient Safety as they represent a risk for the patient. If ADEs can be prevented, then it is a significant progress. PSIP applications are expected to have such an impact that will be measurable through the statistical analyses and the ADE Scorecards application.
New policies are susceptible to be introduced to rationalize the functioning of hospitals or to reorganize clinical units. The PSIP methodology, by providing indicators on ADEs, can deliver a first set of indicators on the impact of these policies.
Patients are more and more involved in a cooperative management of their own disease. This dimension has been studied in the project through the development of a Patient Module. New recommendations for the patient information are issued from this study.
Economical impact:
A cost-benefit analysis has been performed to evaluate the financial benefits associated with the implementation of the PSIP Core Applications in the medical departments of hospitals. It has been estimated that for the 3 main niches identified in PSIP (Hyperkalemia, Renal Insufficiency, Coagulation Disorders), the appearance of an ADE during an hospitalization results in a prolongation of the duration of the hospitalization of 5.3 days, resulting in an increasing cost of 4,500 Euros per ADE.
If we estimate that a better quality management of the ADE would improve the number of ADEs by a ratio of 3%, the implementation of the PSIP application would be financially positive.
Following the recommendations of the industrial partners and in agreement with the results of the cost benefit analysis, we have studied the exploitation possibilities of the two main core applications:
- the PSIP ADE Scorecard,
- and the PSIP Contextualised CDSS.
A plan for the installation of the PSIP ADE Scorecards Applications has been based on a improvement in the prevention of the ADEs by 3% per year during 5 years. This plan shows that the benefits if the application is displayed in 500 hospitals in Europe would be at minimum up to 2,8 millions of Euros.
This evaluation demonstrates that adoption of these services at a large scale would be beneficial for the healthcare community.
A plan for the installation of the PSIP Cx-CDSS has been described. It is not related with large financial risks, as there is no need for large investments in fixed costs like buildings, machineries etc. The only large cost relates with salaries. In five years, based on the installation of the system in 30 different sites, the return on investment would be sufficient to cover the expenses justified by the maintenance and continuous development of Cx-CDSS.
1.5 PSIP PATIENT SAFETY THROUGH INTELLIGENT PROCEDURES IN MEDICATION
Project coordinator: University Hospital of Lille (France, Centre Hospitalier Régional Universitaire de Lille)
Contact person: Pr. Régis Beuscart
Tel: +33 3 20 44 60 38
Fax: +33 3 20 44 60 43
Email: regis.beuscart@univ-lille2.fr
Website: http://www.psip-project.eu
PSIP Partners:
- Centre Hospitalier Régional Universitaire de Lille (France),
- Centre Hospitalier Universitaire de Rouen (France)
- Centre Hospitalier de Denain (France)
- Capital Region of Denmark (Denmark)
- Oracle France SAS (France)
- IBM Danmark A/S (Denmark)
- Medasys SA (France)
- Vidal SA (France)
- Kite solutions, S.N.C. (Italy)
- Ideea Advertising SRL (Romania)
- Aristotle University of Thessaloniki (Greece)
- Aalborg University (Denmark)
- Private Universität für Gesunheitswissenschaften, Medizinische Informatik und Technik GmbH (Austria),
- University Specialized Hospital for Active Treatment of Endocrinology (Bulgaria)
- Institute of Information and Communication Technologies (Bulgaria)
Timetable: from 01/2008 – to 07/2011
Total cost: 10,235,008 €
EC funding: 7,496,328.00 €
Instrument: IP
Project Identifier: FP7-2007-ICT-1-5.2
|
