Personal tools

You are here: Home / Research / Biomedical Signal Processing & Instrumentation / Prof. L Tarassenko

Biomedical Signal Processing & m-Health

There are two main emphases in the group at present: real-time monitoring of patients in hospital in order to provide early warning of deterioration and the use of mobile-phone based telehealth (m-health) to improve the management long-term conditions such as diabetes, asthma, COPD or hypertension. The group has a substantial amount of funding from EPSRC, MRC, the Department of Health, the Wellcome Trust and the Oxford Biomedical Research Centre to support its research, most of which is carried out in partnership with clinical Departments in Oxford.

In-hospital monitoring for early warning of physiological deterioration

Within the hospital, acutely ill patients in critical areas (the Intensive Care Unit, the Emergency Department or the Coronary Care Unit) have their vital signs (heart rate, breathing rate, blood pressure, oxygen saturation and temperature) continuously recorded by multi-parameter patient monitors. We have developed multivariate pattern recognition techniques to learn a description of normality in multi-parameter space and abnormalities are subsequently identified by testing for novelty against this description. This then triggers an escalation of care when the patient begins to deteriorate. The vital sign data fusion technology developed in partnership with spin-out company OBS Medical has delivered improvements in patient outcomes, validated during clinical trials in both the US and the UK.

In less critical areas (for example, the general ward), patients are not continuously monitored. Instead, nursing staff record their vital signs on paper at regular intervals (typically every four hours) from which they calculate a simple, additive score known as an ‘Early Warning Score’ (EWS). The periodic observation of the vital signs (‘tracking’) is linked to pre-determined calling or response criteria (‘trigger’) which are applied when the score exceeds a heuristically-chosen threshold. In collaboration with clinical colleagues, we have design an evidence-based early warning score, using a large dataset of 64,000 hours of vital-sign data acquired in previous clinical trials in Oxford and in the US. Statistical models of normality were constructed to convert the vital sign values into an integer “abnormality score”, which is then summed over all vital signs. The new track-and-trigger charts were successfully deployed in the John Radcliffe Hospital in early 2011. Today all adult ward patients in the Oxford University Hospitals (OUH) Trust are monitored using our evidence-based track-and-trigger charts. In the first year following the introduction of these new charts, the number of cardiac arrests in the Trust decreased by 10%.

m-health for chronic disease management

The real-time analysis of patient data is enabling personalised healthcare for people with long-term conditions such as diabetes. Mobile technology and new methods of communicating information are playing an important role in the management of these conditions. Our vision for mobile health (m-health) is to deliver personalised interventions at scale, to enable self-management. The implementation of that vision is based on a twin-track strategy. Firstly, we continue to build the evidence, through clinical trials, that the use of m-health for focused interventions, lasting between three and six months, can deliver improved patient outcomes with respect to usual care. Having demonstrated this for blood pressure control following stroke and for the titration of medication in diabetes, we have now turned our attention to gestational diabetes.

Secondly, a greater challenge is to extend the lessons learnt from the development of m-health for short-term interventions to the long-term management of chronic illnesses such as Chronic Obstructive Pulmonary Disease and Heart Failure. Successful innovation in m-health for these elderly patients requires an understanding not only of the patients' physiology but also of their psychology. The need for patients to self-monitor is a daily reminder that they have a chronic illness for the rest of their lives. Providing patients with state-of-the art, multi-purpose tablets removes some of the stigma sometimes associated with dedicated telehealth equipment. Similarly, the sensor technology needs to be selected so as to provide the maximum amount of information on the patient's physiological status for the least possible cost to the patient (in terms of disturbance to the activities of daily life). Working in partnership with colleagues from Primary Care, we are developing sustainable m-health solutions which can be integrated within existing clinical pathways for the management of patients with Heart Failure or with COPD (the latter project being funded by the Wellcome Trust and the Department of Health).