Introducing Tim Shepheard-Walwyn: Director at Kensa Health

Tell us about your background and how you ended up working in healthcare analytics.

With a background in economics and finance, I specialised in regulation and risk management. As financial products grew in complexity, I focused on developing and implementing the statistical models necessary to manage them. Upon leaving banking, I sought to apply this expertise to other sectors with complex processes, leading me to focus on healthcare.

What kinds of healthcare or operational challenges have you worked on most closely?

Initially, we supported UK ambulance services with emergency response design and planning, conducting strategic reviews and business plans, especially for out-of-hospital care.

I also worked with Acute trusts and the Emergency Care Support Team, addressing issues like excess mortality at Mid Staffordshire Trust.

I later worked in Australia and New Zealand, supporting the Canterbury District Health Board after the 2012 earthquake by helping develop strategic plans and integrated health systems. On returning to the UK, I assisted Welsh health boards with system implementation and led the Health Insights project in South East England.

What makes healthcare analytics different from standard reporting or dashboards?

Healthcare, like finance, is a complex adaptive system that requires clinician engagement for change. Standard reports use top-down metrics and rarely address root causes.

Effective analytics must answer the 'Why?' by examining individual patient pathways relevant to clinicians.

Can you talk about a piece of work or insight that genuinely changed decision-making or outcomes?

A key example was after the Canterbury earthquake, which reduced Christchurch Hospital's capacity by 60 beds just before winter's surge. I led a live session in which clinicians and managers reviewed admissions and bed-occupancy predictions for different patient groups.

We identified that COPD patients arriving by ambulance drove winter bed use. Recognising that many could be cared for at home, a new partnership enabled stable patients to remain there, reducing bed occupancy by over 40 beds.

What excites you most about where NHS analytics is heading over the next few years?

The NHS has long aimed to be data-driven, but decisions often rely on anecdote. AI now allows staff to interact with data in natural language, letting non-technical staff engage directly. This shift could transform how the NHS operates.

What do you think operational leaders most need from analytics today?

With an ageing population and tight budgets, the NHS can't keep doing the same things. Leaders need analytics to clarify challenges, uncover new opportunities, and communicate effectively to drive change.

What’s one misconception people have about NHS analytics?

Probably the biggest misconception about analytics in the NHS is that only the BI team can do analytics. This creates long queues for the BI team's queries, and frequently, the answers it provides just lead to another question. It also means that there is very little challenge to the BI team. There is also a continual tension between the BI team's reporting and analytical priorities.

What signals or patterns do you look for first when something starts going wrong operationally?

The key questions to ask when something starts to go wrong are: firstly, is this a demand issue – i.e., did we have more patients than we predicted? - or is it a process issue – i.e. has something happened which means that it is taking longer to manage patient flow through the system. Once we know if it’s a demand or a process issue (and it could be a bit of both), we then need to go to the next level of investigation - i.e. which cohort of patients and which part of the process lies at the root of the problem.

What makes system-wide analytics difficult in practice

David Meates, the Chief Executive of the Canterbury District Health Board, made an important observation that ‘Change moves at the pace of trust’.

This is particularly true in system-wide working, where every organisation has its own perspective on the issues and who is standing in the way of change.

The starting point for success is to develop trust in a ‘single version of the truth’  – a commonly shared system-wide data set and approach to analytics that means that debate and discussion can be about what the data is telling us and not about whether we believe the data.

What’s the most valuable thing operational teams can gain from real-time analytics?

The availability of real-time data in an operational environment has always been a goal for our operational teams. But data is only useful if decisions and actions can be taken in response to it as it changes, and frequently, operational staff have no ability to act in real time based on what the data is telling them.

In this context, operational teams should define clear expectations and conduct regular status checks to quickly identify and manage deviations from the plan.

In well-run organisations, these daily and weekly huddles, where staff can reflect on what the data is telling them and what they can do about it, are the basis for effective operational management .  

Can you describe a moment where the data told a very different story from the assumptions people were making?

In 2008, ambulance services in England were required to reach 75% of urgent calls within 8 minutes. This put pressure on services and fostered the assumption that those missing the target were inadequate. Our reviews revealed that the standard was set without considering feasibility or costs, particularly in rural areas, where meeting the time frame was unrealistic.

This meant that services were concentrating their activity in areas where they could achieve the new standard, and patients living outside these areas received a lower standard of service than before the change. Overtime, it became increasingly clear that the call connect standard was impractical, and new standards were introduced that distinguished between life-threatening incidents that required an immediate response and other types of calls, with standards set at both the mean and the 90th percentile to ensure that all patients would receive an appropriate response.      

To discuss how these insights apply to your system, connect with Tim via LinkedIn.

If you’re ready to manage demand, capacity, and variation with greater confidence, contact the Kensa Health Analytics team today.    

‍