Machine learning promises to revolutionise clinical decision-making and diagnosis. The potential applications of machine learning are substantial, however identifying the best way to use these data analytics can prove challenging.

We can use machine learning to ensure you have the most rigorous and robust analysis, risk stratification models and forecasts so that you can prioritise and target healthcare at the individuals most at risk of developing ill health.

We can:

• Create risk scores that help you decide which patients can benefit from enhanced services
• Develop risk algorithms to ensure you can identify patients who are at highest risk of poor health outcomes
• Predict patient resource utilisation including:
  • Appointment DNAs
  • Re-admission rates (30 days)
• Use machine learning software to improve the data analytics process

Causal models can help healthcare providers and services to understand the relationship between one event (a cause) and another (its effect).

To find a solution to a healthcare challenge, we first need to identify the cause of the problem, not just what is associated or correlated with it. Causal analysis can find the root cause of the problem you want to address, help identify solutions and determine the steps you can take to benefit patient health.

Approaches that can be used include:

• Creating a causal model to establish cause and effect relationships between variables
• Using targeted maximum likelihood estimation (TMLE) and other estimators. These can be implemented using machine learning algorithms, which are advantageous in complex observational data
• Building a causal Bayesian network to compare the performance of different interventions

Once we have identified the main area to focus on, we can help to create an intervention that will help you address that particular issue.

We use visual tools including data dashboards and interactive visualisations to help you present and monitor your healthcare data in a dynamic and interactive way. Dashboards allow healthcare professionals to quickly understand the narrative and enhance their understanding and buy-in to a selected approach.

Our impactful dashboards are easy to use and accessible, and can be tailored for different audiences including your managers, external partners and clinicians. 

They offer a clear and cohesive presentation of your data, identifying the important insights and enhancing understanding.

We can use dashboards to facilitate ongoing monitoring of a service or health outcome in a coherent, useful way.

We use a range of hosted technical data visualisation tools including Tableau, PowerBI and LucidChart.

We identify, select, and critically appraise relevant research through literature reviews that help inform the input parameters for our economic models, and provide information on care pathways and research methods. By quickly and fully understanding the evidence base, we can support you in making the right decisions.

We develop research protocols for each project, perform the reviews and produce timely, high impact reports that offer you clear, evidenced direction.

We are supported by a medical writing team who generate our reports and other communications including helping publish our work in high impact journals. 

This is crucial in our aim of sharing the learning from our work; being able to report research finding and data to key audiences in clear and effective language is vitally important. 

We also develop academic posters and papers.

We offer teaching on evaluation methods, evidence generation for innovation, R courses, and quasi-experimental design. We also cover a range of health economic topics from population health management to health economics for non-economists. 

Our aim is to share our learning and further develop the skills of health analysts across the NHS and industry partners, improving capability and capacity in this important area.

The importance and impact of real-world data assets have never been greater. We work with our clients to identify and provide expert advice on the most appropriate data assets and analytical methods to achieve their project requirements.

In addition to using publicly available datasets from NHS Digital, Public Health England, and the Office for National Statistics, through our hosts the NHS Midlands and Lancashire Commissioning Support Unit (MLCSU), we have access to an unrivalled range of health data including integrated datasets that span primary, secondary, community and social care settings. Data linkage is often a key requirement to extract the maximum insights from disparate datasets, including those generated locally, and for this we work with the Data Services for Commissioners Regional Office (DSCRO), which uses data from GP practices and NHS hospital trusts in regional processing centres.

Throughout all our work we adhere rigorously to information governance standards and best practice. All staff follow strict rules on accessing, analysing, and processing data and our work is risk assured by the MLCSU.