Ramsey Systems Ltd

Agent Based Modelling in Healthcare

Thu, 30 Aug 2018 04:48:04 -0700

As ever when I attempt to write a blog here I have stumbled across several topics I want to delve into further and I will to my best not to ramble on at tangents. What I plan to talk about is why, in our current work, we are proposing the use of agent based models (ABM) along side other modelling techniques (tangent number one: inconsistent use of terminology).

We are sure that ABM can have substantial value healthcare but we need to be able to explain why and how without intimidating people with the complexity and fine intricacies. The level of implementation of modelling methods is disappointingly low (tangent two), so we want to make our proposals as easy to use and see the benefits of as possible.

An analysis of the academic literature on simulation and modelling in health care.

SC Brailsford, PR Harper, B Pate and M Pitt

So, what are the benefits of using agent based modelling in general?

As with all modelling, ABM provides a quick, cheap and risk-free way to explore both reality and possibilities, compared to exploring and experimenting with real assets.

ABMs have a very flexible level of detail, as they are built from the bottom up. The details of decisions made are available as well as the statistics of all agents.

This bottom up approach also means that we can discover emergent phenomena, resulting from the behaviour and interactions between agents, some of which can be surprising.

The kinds of visualisations that can be created from ABMs make it a lot easier to comprehend than other modelling methods. The kinds of 2D and 3D graphics produced can be familiar and easily understood by the typical user.

By Harley Johnson

DataThon reflections

Mon, 30 Apr 2018 04:40:09 -0700

Last week we flew to Gothenburg, Sweden for Medical Informatics Europe’s 2018 conference, hosted by the Swedish Society for Medical Informatics (SFMI) in collaboration with EFMI (European Federation for Medical Informatics). This year the conference was co-located with Vitalis and attracted more than 3500 attendees, making it the largest eHealth event in Scandinavia. It was a great event, with seminars, demonstrations, knowledge building, networking and important discussions on the future of healthcare.

The DataThon kicked off with a tutorial focused on FHIR, conducted by Mr Ewout Kramer from Firely, to introduce and train attendees in using HL7 FHIR standards. This meant the next day, after sorting into groups, we could dive right in to playing around with the data and developing our digital health solutions. The teams all engaged in fruitful discussions and by the end of the day had developed some promising prototypes. The final session of the DataThon provided participants with the opportunity to showcase what they had learnt and created to a panel of experts, receive feedback and engage in a discussion about the use of FHIR, data and digital solutions in health. A vote was held using sli.do and first, second and third place prizes awarded as follows:

ADR Sniffer - A web-based pharmacovigilance application for use in checking a patient’s chances of having an adverse drug reaction to their asthma medication, using the VigiAccess database
A public health dashboard for asthma, showing the population distributions and patient statistics (e.g. medication and hospital admissions) for the available dataset, and...Read More

MIE2018 HL7 DataThon

Tue, 10 Apr 2018 03:36:42 -0700

We are preparing for the HL7 DataThon at the Medical Informatics Europe conference in Gothenburg, Sweden from the 23rd-26th April. At the DataThon, attendees will be trained in the use of the HL7 FHIR standard and the International Patient Summary profile, with the aim of developing unique user experiences for innovative FHIR asthma apps. Ramsey Systems are supporting the DataThon by preparing a dataset, to include the FHIR resources and health records, for 100,000 synthetic Nordic patients, relating to asthma.

We have developed an asthma pathway and generated a synthetic population based on openly available statistics and Census data for the Nordic countries. This means that the digital outputs generated when the synthetic population are run through the pathway are free from privacy restrictions and hence safe and effective for use by developers, business analysts and start-ups building innovative healthcare solutions at such events. The pathway has been developed based on the GINA and UpToDate asthma guidelines, along with clinical input from a Medical Advisory Board.

It should be clear that this is not meant to provide a decision support tool for individual patient care. We are using pathways to identify the relevant information items and model the digital outputs to be used by the Digital Health companies to learn how to read and write FHIR data items that are relevant to the care of asthma sufferers.

DataThon agenda

The DataThon will start with a tutorial focused on FHIR, conducted by Read More

Pathway editing with Node-RED

Fri, 06 Apr 2018 06:59:21 -0700

Trying to find an appropriate tool for editing pathways has not been easy. Even the search terms have been hard to work out. You’ll be presented will specialised biology pathway editors when you search for ‘pathway’ editor and you can get some very pretty and complex graphs searching for ‘network’ editors, these do not let you add data to the nodes however.

I have found a couple of promising programs that we will have to look into more in the future (Dia and DotEditor). For now, however, we have settled on using Node-RED.

Node-RED was certainly not built for this use but we have been able to make a pathway editor out of it anyway. After ignoring that you can incorporate JavaScript functions and dumping all of the default nodes it feels like we’re maybe selling Node-RED a bit short. Nevertheless, the GUI was what we really needed.

And so, we set about making some new custom nodes for our purposes. We needed some nodes into which we could put the data that describes what is happening at each step of the pathway. The guides provided by Node-RED made it a simple task and I got on with designing the data forms. The further you delve into customising the nodes the trickier it gets to find help and I did hit my limit on trying to code multiple tables and nest them. I was probably being a bit ambitious there. In the end, it has been more efficient for us to extract a lot of this data out into spreadsheets anyway so I have been able to achieve everything we needed.

I have been impressed by the amount of customisation that can be done in the creation of new nodes; everything can be changed dynamically if you want it to be. There are a couple of things we could have done with. The ability to add more than one line to the title displayed on the nodes and the ability to move where the inputs and outputs are on the...Read More

Synthetic versus anonymised health data

Fri, 01 Dec 2017 07:28:44 -0800

Anonymisation techniques – either encrypting or removing personally identifiable information - enable the wider use of personal data. However, using anonymised data comes with the risks of re-identifying patients, and thus involves a strict ethical approval process.

Using synthetic health data helps to overcome these privacy and confidentiality issues.

Synthetic health data is generated to represent real patient data, using publicly available open data sources, like NHS England and Public Health England statistics. Using synthetic health data also means:

You are able to practise writing queries and defining policies prior to accessing anonymised data, which has a strict ethical approval process
It holds no personal information and cannot be traced back to any individual; therefore, the use of synthetic data reduces confidentiality and privacy issues
You are given another perspective, helping to define a project because the synthetic data acts as a testbed
Tracking through synthetic data and the assumptions used to generate the synthetic data may assist with literature discovery – identifying the cohort and impact of research. For instance, researchers doing clinical trials may generate synthetic data to aid in creating a baseline for future studies and testing.

Synthetic data can also be used in testing and creating different scenarios within a system, for example, illustrating the impact of a policy change, both at present-time and in the future.

Synthetic electronic health care records

A group of researchers in Massachusetts have developed Synthea: an approach, method, and software mechanism for generating synthetic patients and the synthetic electronic health care record. Based on publicly available information, the records can be simulated with pathways of disease progression, care plans, personal workflows (of the citizen and the healthcare professional), and the lifecycle of...Read More

Health 2.0 Manchester

Fri, 24 Nov 2017 07:37:44 -0800

Cross-sectoral action for health is crucial. The transport networks of our cities, the amount of green space in our neighbourhoods and the quality of our housing all influence our health outcomes.

This week we attended a Health 2.0 meetup in Manchester. Health 2.0 is a global initiative that has grown into a movement of over 100,000 entrepreneurs, developers and healthcare stakeholders. Groups meet to share, discuss, showcase and develop ideas to improve patient care.

Image of Bright Building meeting space (https://mspl.co.uk/meetings-events/)

Health 2.0 Manchester is a meetup group focused on innovations in the delivery of healthcare and health related technology. At this event we heard an interesting range of speakers, from nurses to app developers, talking about the work they are doing to improve the delivery of healthcare in the North West region. Presenters included; Dan Gippell, the founder of Betterpoints app, a health behaviour change app developed to incentivise citizens to become more active, Rebecca Morris-Chippendale, Business Development Manager from NorthWest EHealth Ltd., who offer an alternative approach to clinical trial design and management around the world, and many more. It was particularly refreshing to hear Rebecca emphasising the need for collaborations between public and private organisations in healthcare.

The event enabled us to find out more about new and existing healthcare solutions in the North West region. We were able to see how agent-based simulations fit into and could contribute to existing projects.

Health 2.0 is a great initiative, demonstrating the need for collaboration and innovation in the delivery of healthcare. It was very encouraging to see first-hand how Manchester are promoting the sharing of ideas and collaboration, offering many exciting opportunities for SME’s and public health partners...Read More

Health in Ecology Policy

Fri, 13 Oct 2017 07:39:48 -0700

A healthy environment contributes to a healthy population. For example, when planning housing or other developments, considering how the environment can be best shaped to encourage active play and physical activity, reduce social isolation and encourage community resilience is important to improve health and wellbeing.

However, health data is currently siloed, inconsistent and inaccessible, making it difficult for other sectors, like ecology, to demonstrate and evaluate the impacts their initiatives have on health.

For ecology, the levels of access an individual has to green space in their environment has a significant impact on their health. The World Health Organisation’s review of the evidence linking urban green spaces to better health outcomes includes improved mental health, reduced cardiovascular morbidity and mortality, obesity and risk of type 2 diabetes, and improved pregnancy outcomes.

Mechanisms leading to these health benefits include psychological relaxation and stress alleviation, increased physical activity, reduced exposure to air pollutants, noise and excess heat.

With such strong evidence linking levels of green space to improved health outcomes, it seems nonsensical that there is not a standardised system in place which allows ecology departments to safely and effectively use health data to inform and evaluate their green space initiatives.

Such a system would require an information architecture, that defines how digital health information can be shared between people, organisations and their computers systems. This is something we are interested in developing, as well as using agent-based models to simulate the digital health information flows. Together these would allow policy-makers from all sectors to understand and illustrate digital health information flows,...Read More

Health in Transport Policy

Wed, 20 Sep 2017 04:01:40 -0700

Last week we attended the Birmingham in Real Time (BiRT) Network Resilience Challenges Workshop at the iCentrum on Innovation Birmingham Campus. The workshop brought together a range of stakeholders in a series of sessions, to brainstorm, highlight, detail and formalise a number of specific challenges to work on to start delivering the West Midlands Network Resilience agenda.

The workshop was led by Birmingham in Real Time project manager Alan Dolhasz, with support from Andy Radford (Birmingham City Council) and Teresa Jolley (SME Engagement and Innovation Facilitator for the Transport Systems Catapult/Transport for West Midlands Intelligent Mobility Incubator at iCentrum), building on the work already done with Birmingham’s UTMC traffic data with Birmingham City University, and the priorities required to deliver the Mayor’s performance dashboard for transport.

There was also a demo from CEO Robin North of Immense Simulations, on the tool that is being used to demo a Strategic Decision Support Tool for TfWM through the Keeping the West Midlands Moving project (by Transport Systems Catapult).

It was really great to see everybody, from policy-makers to developers, data analysts, researchers, and SME’s, so interested in working together to figure out how to collaboratively solve some of the challenges being faced, with a specific focus on the need for data to support the delivery and evaluation of projects, which would require open access to not only transport, but also health data…

The crossovers...Read More

Health in All Policies

Sun, 20 Aug 2017 03:53:18 -0700

Health inequalities are largely driven by socio-economic factors, and thus determined outside of healthcare.

As illustrated by Dahlgreen and Whitehead's social determinants of health rainbow:

It is therefore imperative for all sectors to work together in order to improve the health outcomes of a population.

Health in All Policies (HiAP) is the strategy, promoted by WHO, to include health considerations in policy making across different sectors that influence health. These include transportation, agriculture, land use, housing, public safety and education.

As a concept, HiAP reflects the principles of: legitimacy, accountability, transparency and access to information, participation, sustainability, and collaboration across sectors and levels of government.

WHO recognise that data sharing needs to be at the core, by developing cross-cutting information and evaluation systems.

Agent-based models to support data sharing

Agent-based models are a useful tool for simulating the digital health information flows within and between sectors that influence health. If standards-based, simulations could be developed as common monitoring and evaluation tools, providing policy-makers, across all sectors, with:

The ability to request health data in a clear, simple and unambiguous way
A consistent use of health data across all projects
A standardised, scalable and reusable way of monitoring and evaluating the health impacts of each initiative

WHO provide a Health in All Policies: Framework for Country Action on their website

By Allie Short