Health Informatics 3.0
Total Page:16
File Type:pdf, Size:1020Kb
8 © 2011 IMIA and Schattauer GmbH Health informatics 3.0 Dipak Kalra University College London It’s time for health informatics to evolve it and in the formats (modality, device, Summary from enabling the information society level of detail) that are best suited to Web 3.0 promises us smart computer services that will interact with to creating the knowledge society: that individual at that time. This is in- each other and leverage knowledge about us and our immediate health informatics 3.0. deed a grand vision! context to deliver prioritised and relevant information to support de- The vision of the semantic web (a Healthcare now needs us to provide such cisions and actions. Healthcare must take advantage of such new critical element of Web 3.0) is of in- knowledge-integrating and collaborat- knowledge-integrating services, in particular to support better co- formation objects that are richly in- ing services: not as prototypes and pi- operation between professionals of different disciplines working in dexed by knowledge models and bound lots, but on a wide scale. different locations, and to enable well-informed co-operation be- to collaborating components within a Increasing sophistication of care, tween clinicians and patients. To grasp the potential of Web 3.0 we distributed computing ecosystem [1], with an ageing population [2] and more will need well-harmonised semantic resources that can richly con- bundled as smart services that interact and more illnesses becoming long-term nect virtual teams and link their strategies to real-time and tailored with each other to deliver personalised conditions (such as HIV [3]), requires evidence. Facts, decision logic, care pathway steps, alerts, education (tailored, relevant and prioritised) in- greater co-operation between profes- need to be embedded within components that can interact with mul- formation that supports the decisions sionals of different disciplines work- tiple EHR systems and services consistently. Using Health and actions of individuals and commu- ing in different locations, knowing Informatics 3.0 a patient’s current situation could be compared with nities. Services will collaborate to de- each others care goals, progress made the outcomes of very similar patients (from across millions) to de- liver an integrated and holistic solution and any difficulties encountered. This liver personalised care recommendations. The integration of EHRs to its users. In other words, individuals needs more than sharing the record of with biomedical sciences (‘omics) research results and predictive will no longer utilise computers to what has been done (EHRs), but shar- models such as the Virtual Physiological Human could help speed search for and locate specific items of ing each other’s clinical care strategy, up the translation of new knowledge into clinical practice. The mis- information to digest and weight up, options and logistic constraints, to de- sion, and challenge, for Health Informatics 3.0 is to enable healthy but will pose problems and receive an- termine optimal ways of aligning ef- citizens, patients and professionals to collaborate within a knowl- swers to questions like: forts. We accomplish this best today edge-empowered social network in which patient specific informa- • What are my best care management by discussing patients in team meet- tion and personalised real-time evidence are seamlessly interwoven. options for this particular patient? ings. Physical (or online) meetings • What will happen to renal function could only cover a minority of patients, Keywords if I discontinue this drug? and team collaboration needs a way of Semantic web, semantic interoperability, patient engagement, per- • What is the optimal dose in this pa- scaling up virtually, through knowl- sonalised medicine, trusted services tient to maximise effectiveness and edge services. minimise adverse effects? The rapid pace of bio-science dis- Yearb Med Inform 2011:8-14 • How much should we budget for covery, the advent of personalised children’s cancer services next year? therapies, and the overwhelming vol- ume of new clinical evidence, com- Web 3.0 computers will work out what bined with demands for safe, high qual- information they need, where to get it ity and equitable standards of care, has from and how to access it. They will already encouraged professionals to filter and prioritise and personalise replace historic sources of knowledge what they find, and present the user (textbook plus apprenticeship plus an- with a polished set of reasoned options ecdote) with evidence based care. The from which to make a final choice. access to evidence has itself evolved Smart services will present individu- from individuals searching the litera- als with information when they need ture (with greater or lesser skills at IMIA Yearbook of Medical Informatics 2011 9 Health informatics 3.0 critical appraisal) to peer-developed Much of health informatics, and guidelines that have filtered and com- also the needs of health care, have Harmonised Semantic bined published evidence into a speci- focussed up to now on managing col- Resources fication of best practice (50+ pages lections of data: for example improv- each!), and now progressively being ing interfaces and devices for data cap- The cornerstone of health informatics mapped into clinical workflows (as care ture, information models and databases 3.0, of collaborating components and pathways) to make them more usable. to store and process the data, inter- smart services, is knowledge integra- But we need to advance from these operability standards to communicate tion. Yet, today, our knowledge repre- forms of evidence (which are ulti- data and allow us to grow longitudinal sentations across healthcare, clinical mately informed by large expensive health records, single condition guide- research and bio-medicine are limited, studies that take years to deliver re- lines and simple decision support, and poorly interoperable. sults) to empirical data integrated analysis of reporting data sets and data Investments in health informatics re- across large populations, collected sys- warehouses, and a little bit of near pa- search, standards, product develop- tematically and therefore comparable: tient tele-care. The interpretation of com- ment and deployment are gradually real time outcomes. plex images and signals has perhaps had enabling the integration of data within Individualised evidence would take the greatest real-world impact so far. Bio- and between organisations. This is this a step further, in which a patient’s informatics is bringing us fresh ways to known colloquially as syntactic inter- current situation could be compared consider diseases and treatments, but we operability: the ability to exchange with the patterns and outcomes of very need to improve its integration with rou- data, but only being able to use it similar patients (from across millions) tine health care (beyond inherited disor- meaningfully if parties have agreed in in order to deliver point of care recom- ders and chemotherapy). advance on how they will represent mendations that are highly relevant. Such Health IT has so far helped us to meaning within the data. In the hier- case based systems already exist, but are cope with the increasing data volumes, archy of semantic interoperability lev- usually focused on a single condition, but as we finally succeed in joining els defined in the SemanticHEALTH intensively curated and only within one up disjointed parts of our health sys- Report [5], Web 3.0 services will re- (innovative) hospital. Linking case based tem how will we avoid information quire Level 3 interoperability: full se- analyses to emergent knowledge could overload? Will our clinicians and our mantic inter-operability with sharable radically shorten the “bench to bedside” managers simply get more data, or will context and seamless co-operability translation from over a decade to months. they have useful information presented (i.e. received data can be combined Such capability should be available to in ways that can be applied efficiently seamlessly with local data and proc- inform the routine care of all patients, and effectively? Will they spend more essed homogeneously). whenever needed. time sifting through health informa- We need guideline and decision sup- Most of our current guidelines are tion and published knowledge, or less? port systems, notification and alerting for single diseases, but patients rarely Will service management decisions components, and analytic tools to be have only one. The most realistic way (e.g. commissioning) be more wisely able to consistently process integrated of building up an understanding of informed or remain simplistic but health data drawn from multiple EHR how diseases and treatments interact, backed by larger volumes of data, as systems. New generation personalised and the optimal ways to manage mul- poor in quality as today? medicine needs the integration of EHRs tiple conditions, is to study (data mine) Next generation health care poses with biomedical sciences (‘omics) re- large-scale EHR repositories. Euro- some quite powerful challenges that re- search results and predictive models pean level disease specific repositor- quire us to harness and exploit knowl- such as the Virtual Physiological Hu- ies are already proving useful for edge at a much greater scale than we man. Clinical and public health re- knowledge discovery (e.g. [4]), but presently can. To grasp the potential search, and clinical trials, need to