This page is under development
The exploratory phase of our Grand Challenge proposal started from the premise that IT support for clinical decision making, care planning and active workflow management can make major contributions to improving patient care. At the beginning of the project we knew that all of the technology that would be needed to achieve our vision did not yet exist, but some of it was within reach. We knew much about how to deliver decision-support in clinical settings - how to formalise clinical decisions and processes, and how to embed them as service modules in care pathways for example. The partners had significant expertise in representing medical knowledge, presenting decision options to clinicians in the form of the arguments and supporting evidence that are relevant to the circumstances of a particular patient, and how to use this information in order to make and explain decisions and plans to clinicians or their patients. We had demonstrated and validated these capabilities in a range of operational clinical applications.
The starting point therefore is that (1) we have proven techniques for using IT to support quality and safety of care, and (2) as a result of work on the exploratory phase we now know that we can deliver decision support, workflow and knowledge sharing services to clinicians, patients and researchers using centralised orchestration of all services or open choreography techniques, or a mixture of both. The hypothesis of this project is that such services can be delivered into real-world clinical settings in a robust, flexible, scalable and safe way.
Although our prototypes have been demonstrated using specific applications the methods and platforms can be applied in a wide range of specialist and multidisciplinary clinical settings. The software can significantly enhance conventional (human readable) clinical guidelines and care pathways such as DOAS and Map of Medicine pathways by providing patient data capture (e.g. patient histories, case report forms in research trials), active decision support (e.g. alerts and reminders, advice in diagnosis or treatment decisions) and coordinate users and service providers who are distributed in space and time. Furthermore, we have shown how many other services can be integrated into the pathways, to assist with tasks like scheduling tests and investigations, optimising referrals, planning care, and actively supporting clinical staff in following care plans.
Primary technical objectives relate to three primary themes: service integration, service delivery and safety.
Service integration.
The clinical scenario for our demonstrators was based on standard processes for the initial assessment of women with possible breast cancer. This offered specialist decision support and workflow management services for members of a multidisciplinary team. This scenario will be the starting point for the proposed clinical trials during year 1, but we plan to use the extended scenario to show how services can be also provided to patients and researchers as well as clinicians, and how service provision can be opened up to 3rd party service and content providers.
Services for clinicians will include management of patients according to established care pathways agreed with clinicians at the UCL Royal Free and Imperial Healthcare/St. Mary’s NHS Trusts, covering clinical data recording, workflow management, alerts and reminders, clinical order entry and so forth. Decision-support services will be incorporated into the care pathways based on national clinical guidelines, drawing on publications by the NICE Cancer Guidelines development programme, with whom we are already collaborating. Non-cancer applications will also be developed to demonstrate the generality of the approach and to demonstrate decision support and other services in managing co-morbidities and hazard management.
Services for patients will be driven by the need to personalise care plans and empower patients by facilitating their access to personally focused knowledge and other services which will inform their decision-making. An active electronic diary will be a key device for supporting communication and coordination between patients and clinicians, and facilitating patient monitoring during their treatment at home and long-term follow-up.
Services for researchers will address the need to overcome the current disconnect between routine practice and research. Discussions with the Director and Chair of Connecting for Health’s Research Capability Programme indicate that many services which we can already implement would be extremely valuable if properly integrated e.g. services to capture patient data before, during and after treatment, to support recruitment of patients into research trials, adverse event monitoring and pharmaco-vigilance, and services which provide clinician and patient feedback on effectiveness of therapy and patient experience and well being.
Service delivery
The second major technical theme of the project is concerned with how to distribute and control the delivery of clinical, patient and research services, and in particular whether and when to use a conventional centralised model and when to use a more open framework. Given that both of these service delivery models is very general we may ask why the project should include two separate approaches rather than settle on one? This is because the two service models will have different strengths and weaknesses, and a key aim of this project is to establish what these are. To illustrate this we will consider some of the possible strengths and weaknesses of the models from the point of view of the three anticipated groups of user: clinicians, patients and clinical researchers.
Clinician’s view. From the clinical user’s perspective orchestration is like any client-server system which provides centralised management of service delivery and content – client-server architectures are well understood so applications are likely to be simpler, cheaper, more secure and safer to deploy than on open service architectures. However centralised organisation has the likely consequence that services which have not been explicitly integrated into the care package are “locked out”. Choreography on the other hand allows third party suppliers to offer services (under appropriate quality and safety protocols) relatively independently, permitting clinicians to use their judgement flexibly and adopt non-standard treatment options if the standard care plan is inappropriate.
Patient’s view. Most patients in the UK place a high trust in the NHS and the services it provides, and many mistrust services which are provided by the independent or alternative medicine sectors. For patients who are increasingly active in their own care decisions and management (diabetes is an established example; cancer is becoming one) NHS web sites offer a “one stop shop” which simplifies self-education and decision-making. For the many NHS patients who do wish to use independent or alternative services, however, choreography will offer greater choice by allowing 3rd party services to be visible to the patient from within the NHS core. A patient who generally prefers to use core cancer services for example, may prefer to select diet-management or psychological services from a local charity or accept an invitation to participate in a research trial from Cancer Research UK.
Researcher’s view. Connecting for Health’s Research Capability Programme seeks to realise the long held ambition for the NHS to be a national laboratory for clinical research. An advantage of the orchestration approach is that the NHS could straightforwardly maintain a central database of open clinical trials, assess patients for eligibility in those trials, and capture data for data mining and research and impose data and service standards on all NHS accredited trials. At any one time however there are many trials in progress which are not being carried out within the NHS (indeed patients may wish to enrol in a non-UK trial). A further attraction of an open choreographic approach is that such trials can be visible to and, under appropriate terminology and ontology mapping protocols, patients may enrol in them while their care records, decisions and health outcomes remain accessible to NHS clinicians and researchers.
Safety management
Patient safety has a number of dimensions including quality assurance, at both software engineering and knowledge content levels, strategies for promoting and supporting a safety culture around the design, implementation and use of clinical services. From a software engineering perspective our claim that the services we are proposing can be “sound” is based on the fact that PROforma and LCC are well-defined formal languages, whose semantics are well understood; this opens up the use of formal specification and verification techniques, model checking etc where applications have safety-critical components. We have experience of applying these techniques at both the software and content levels. The claim that applications can be “safe” is based partly on the same soundness argument but we also propose to apply standard safeware methods such as HAZOP together with use of active monitoring for adverse events or potentially hazardous clinical actions, and triggering alarms or mitigations at run time In previous work we have extensively investigated and reported on safety management generally and specifically in clinical decision support systems and will apply the general protocol we proposed in this project.
Supporting R&D
In order to carry out trials we need a robust delivery platform for the orchestration and choreography applications. The team is uniquely positioned to provide this. OpenClinical.net is a software platform and infrastructure for deploying decision support, workflow and knowledge sharing services for clinical research and routine patient care jointly developed by Oxford, UCL and Edinburgh. The parent project OpenClinical (www.openclinical.org) was established in 2002 as an open information service and portal for clinicians, technologists, healthcare providers, commercial IT suppliers and other users; the user base is conservatively estimated at 20,000 individuals worldwide. The OpenClinical.net web site builds on this success by providing a suite of tools and services for deploying PROforma decision support and knowledge management applications via the web (www.openclinical.net). This will allow us to significantly increase the impact of our research by supporting an application hosting service for application testing and clinical evaluation by third parties.
We are currently able to automatically deploy CDS, workflow and other services on OpenClinical.net so long as service management is entirely centralised. A key enhancement will therefore be to support an open service model based on LCC choreographies. Among the enhancements will be a library of standard choreographies which is accessible by 3rd party providers who wish to offer their services to users of other OpenClinical applications.
Technology assessment It is generally accepted that new technologies cannot be judged purely on technical criteria (performance, scalability etc. ) or even clinical criteria (e.g. patient outcomes, adverse events, timeliness of treatment and quality of life) but must also reflect personal, social and organizational requirements and constraints. The services that we envisage will be used at the point of care and will therefore impact many participants in the process and will directly influence what, when and how each person’s care is delivered. Consequently there is an important need for sophisticated technology assessment, and use of appropriate tools to identify and assess different kinds of impact, from changes in working patterns and safety management, to ethical and legal liability issues. Partners at Imperial College have reviewed the requirements for technology assessment in decision support and other knowledge based services and finalising the development of a technology assessment procedure based on this. It will be used (and validated and refined) during the project to guide the design of clinical services and to assess their strengths and weaknesses in use.
