Technical Description

Human-centred design

The HeartMan project follows a human-centred and iterative approach. This means that patients and caregivers (end-users) are actively involved in identifying and addressing issues that could otherwise compromise a positive user experience, and therefore the value of the system.

In particular, patients and caregivers are involved in identifying user requirements, designing and evaluating prototypes, and participating in field trials. For this, the consortium partners apply ethnographic, co-design and user-experience methods, such as diary studies, interviews, ideation, paper prototyping, cognitive walkthroughs, and usability tests.

This will ensure that the HeartMan system will be truly meaningful for each individual patient and caregiver, capable of adapting to, and solving real life challenges associated with Congestive Heart Failure (CHF).

Sensing in HeartMan

The HeartMan project has developed sensor technology and intelligent interpretation methods to monitor patients in the least obstructive possible way and to maximise the value of the data thus collected.

The HeartMan device, a tailor-made wristband, includes sensors for PPG - Photoplethysmogram, a way of optically obtaining a volumetric measurement of an organ-, GSR - Galvanic Skin Response, a method to understand various types of activity in certain parts of the body - and acceleration, and it can also calculate heart rate, HRV - heart rate variability - and respiratory rate from the sensor data.

Processing the sensor data with models built with classical machine learning and deep learning enables the following advanced features:

  • The PPG measurements can be used to estimate the blood pressure without the need for a traditional measurement (with an inflatable cuff).
  • The accelerometer measurements are used to recognize the patients’ physical activities and their intensity, and trigger actions in specific contexts as needed.
  • Physiological signals, like heart rate variability, respiratory rate as well as the user’s voice - recorded during weekly conversations with an informal caregiver -, assess the patients’ psychological state.

As the wristband can be connected to mobile phones, via the purposely built HeartMan application, patients can also transfer data to their caregivers to ensure continued monitoring without requiring a face-to-face consultation.

The HeartMan wristband is worn in a similar way to (other) normal well-being products but its design and functionality follow medical device requirements and regulations.

Decision support & predictive models

The HeartMan decision support system (DSS) is designed to support patients in managing their disease by providing both sound medical advice and psychological support.

Some of the medical interventions are recommended by a model that predicts the patients’ self-perceived state of health from objective parameters (e.g., heart rate, or surrounding environment temperature). In case of a poor predicted state of health, the system may recommend an action to improve it. Most of the interventions are recommended by models based on expert knowledge, derived from medical guidelines on CHF management and other literature, as well as knowledge generated in the project.

The DSS guides patients through a comprehensive and personalised exercise programme; additional interventions relate to nutrition, medication, physiological measurements and management of the surrounding environment. As far as psychological interventions are concerned, these are:

  • Cognitive behavioural therapy, which acts on cognitive processes with a view to improve adherence to the prescribed medication scheme. Moreover, other strategies, such as prompting and reinforcement, will also be adopted to ensure adherence to the prescribed physical exercise and dietary plans;
  • Exercises of relaxation and mindfulness, which will aim to make patients more focused on the present moment (thus softening feelings of fear and anxiety) as well as to help them better cope with their disease. The exercises will be based on biofeedback games (i.e., games that will respond to a patient’s feelings of, for example, anxiety or fear as mentioned), mindfulness text messages and soothing sounds.

HeartMan applications

Besides the mobile app for patients, the HeartMan project also provides a flexible and user-friendly web user interface. Although intended primarily for medical professionals, this interface could also be relevant to other stakeholders.

This web interface distils the results of the heterogeneous data from the Interoperability Layer, hospital visits and sensors, to outline the patient evolution and present an insight into his/her overall health status. Medical professionals can further use the interface to review the suggestions made by the DSS.

Backend and integration

A testbed cloud infrastructure - a platform to conduct test on scientific theories, computational tools and new technologies - is installed in, and deployed for, the backend platform. Health data integration is done mainly via a module named Interoperability Layer, which acts as a key component for sharing standardised medical data. The Interoperability Layer is built with a HAPI-FHIR server, an implementation of the HL7 FHIR specification. This component operates as a collector of information from patients, while the database of the IoTool framework brings data to the DSS in both environments, mobile and cloud, to run the proper algorithms. The health concepts and parameters involved in HeartMan are modelled as FHIR resources and shareable through an Interface layer.

Clinical trial

The HeartMan trial will assess, in a relatively short-term (six months), whether the system is actually capable of improving patients’ quality of life in comparison with the current standard of care. The assessment will also include sexual activity and the patients’ self-perceived state of health, which as important but rarely studied issues. The trial will consider the collective impact of the three actions prompted by HeartMan (i.e., improved diet, psychological support and enhanced physical condition) rather than their individual interventions. The trial will also be used to evaluate the user experience of the system.

A six-month randomised control trial with a repeated measures experimental design will be conducted to evaluate the effect of the HeartMan system on disease management and health-related quality of life (HRQoL). Other endpoints are mortality and hospitalisation, the effect on behavioural outcomes, self-perceived state of health and clinical outcomes which may ultimately have an impact on disease management and HRQoL. A secondary objective is related to the effects of cognitive behavioural therapy and mindfulness exercises embedded in the HeartMan application in order to reduce psychological co-morbidities such as anxiety and depression as well as to improve behavioural outcomes as described above.

A total of 120 patients (60 patients in Belgium and 60 in Italy) will be enrolled through convenience sampling. Patients meeting the inclusion criteria are patients living with ischaemic stable CHF, with at least one hospitalisation due to their illness (as long as this hospitalisation has not been in the month preceding enrolment), with no planned surgical interventions, New York Heart Association (NYHA) functional class II-III, reduced left ventricular function (LVEF ≤ 40%), good cognitive functioning and sufficient knowledge of the national language.

All patients will undergo a baseline assessment by completing a questionnaire on a variety of topics (e.g. self-care behaviour, therapy adherence and quality of life) and an exercise capacity test. Following the baseline assessment, 40 patients (randomised) in the control group will continue to receive the standard care, while those in the intervention group (80 patients) will be provided with the necessary instructions and equipment. At the end of the intervention, the patients will undergo a final clinical examination similar to the baseline assessment.

The HeartMan trial will be implemented from April 2018 onwards, and patients will use the system, application and wristband for a minimum of five and a maximum of seven months.

If the trial is successful, hospitalisations and other healthcare-related costs should be reduced. These will also be considered proxies for improved outcomes.