Healthcare and the innovation cycle — possibly the biggest challenge?

by Ben Bridgewater, M.D.

Over 25-plus years spent in the United Kingdom’s National Health Service (NHS) surgery, academic medicine fields and more recently at DXC Technology, the world-leading independent, cross-industry, end-to-end IT services company, I am becoming increasingly interested in differing approaches to research and innovation. As I look into this, I am starting to believe that this is probably the most important, and possibly one of the harder nuts to crack, to drive necessary healthcare transformation at scale.

In my 20-year experience of cardiac surgery, my colleagues and I got a bit better at doing everything to provide benefits for patients, but there were no changes that really “shifted the dial”. For example, when I started, mitral valve replacement was the norm for leaking mitral valves, but for many years there has been an evidence base for repair (keeping the patient’s own tissues) rather than inserting a new valve made from foreign material. Over time, there was on-going momentum in published evidence, and the gradual adoption of repair techniques. But uptake of these beneficial approaches remains patchy, even now.

Why is the pace of change so slow? Well, it starts with the quality of evidence — to date there has been no large-scale randomized trial (the best form of evidence) showing the benefit of repair over replacement. The combined evidence is overwhelming in my view, but the profession has taken the view that clinical freedom should be permitted — “a good replacement is a better than a bad repair” — because the volume of high-quality studies to close the argument down once and for all does not exist. The preference for repair procedures is now included in international guidelines, but these are not enforced at scale by payers, providers or regulators, leaving scope for the slow pace of change.

What about the university setting? My academic experiences were primarily about producing an evidence base to support the use of cardiac surgery outcome measures for quality improvement, professional regulation and external publication. This was not a popular agenda in the surgical community. Key to the argument against transparency is that it encourages surgeons to turn down the highest risk patients, and it is exactly those patients who stand to benefit most from successful surgery.

There was a requirement to develop methods to adjust for operative risk and apply those methods in a way that decreased the potential unwanted consequences of surgical benchmarking (the potential to turn down the highest risk cases for example) while maximizing the benefits of structured audit and transparency (patient choice, lower mortality and fewer complications). To ensure clinical buy-in and to mitigate legal risk required a robust, transparent and peer-reviewed evidence base for the methodology we used. Getting this done required the classic academic approach of raising funds; appointing people to deliver the outputs; publishing the research and applying the methods to the data. The problem, however, is that the time from the start of research to large-scale deployment can be prolonged, by which time things have moved on. The quality of care improved over time and, as a consequence, the mortality observed in the real world was much lower than that predicted by the model. A time-expired model which over-predicts mortality gives false reassurance to patients, surgeons and regulators. The slow innovation cycle has real consequences.

At DXC, I have been exposed to commercial approaches to product life-cycle management (PLM) and the software development life cycle (SDLC), which are key to an organization’s value delivery. DXC is a large organization but still runs agile processes for its software development. I have also been exposed to outputs from a number of start-up software businesses over the last 18 months, and their innovation life cycle can be even quicker.

The venture capital (VC) community is configured to spot value in novel approaches and to invest to support change and make money. For example, one of the big areas of VC investment in health at the moment is “digital therapeutics”, whereby apps and other approaches are used to directly modify individual behaviours for the benefit of citizens and the health economy. Evidence that these approaches deliver benefits is just starting to emerge, but these approaches are far from being deployed at scale — healthcare organizations are waiting to see the benefit on clinical effectiveness and return on investment before progressing.

It is inevitable that evidence of efficacy will run behind innovation, but the challenge here is to deliver “faster time to value”. Delays in implementing innovation perpetuate global healthcare challenges, which are currently proving very difficult to address, are affecting healthcare delivery for citizens and having a wider economic impact. It is of course right and proper that patient safety should come first, and some degree of inertia in the medical profession is important — I have seen a number of innovations that have promised much and delivered nothing, including an approach that drilled holes in the heart to increase the blood supply for patients with coronary artery disease, and novel approaches for wrapping the heart with skeletal muscle to augment its pumping ability.

Adopting modern PLM processes to accelerate innovation flow is key. This requires a structured approach through ideation, delivering proofs of concept, minimal viable products, and implementation with constant evaluation and refinement. Most of the things that need to be transformed in healthcare are already recorded (if not measured) through routine data flows, and these should be leveraged to accelerate initiatives and minimize cost. (Imagine, for example, how easy it would be to apply machine-learning approaches on routine data to the problem of mortality prediction described above). The other key area of patient-reported outcomes and experiences is now easy to collect from a technology perspective.

As always, culture is key and essential for embracing change. Be humble and accept that even the best ideas do not always bring the expected benefits. You will need to be able to “fail fast”, and avoid the “hippo” (highest paid person’s opinion) trap. A digital healthcare system should be able to run multiple trials on new approaches at any one time using routine data and advanced analytics (for both identifying cohorts and extracting insights), but the methodology will often be more akin to start-ups than to classical academia (even though the latter will always have its place).

But, of course, the final word must go to the patients — or to citizens for trials on maintenance of wellness. They badly need faster innovation, and we must assess the efficacy of new developments from their perspective on their data. Everyone involved needs to recognize this and back-up innovation with the trustworthiness acquired through appropriate consent, privacy and security approaches.

Ben Bridgewater, M.D., was the director of global advisory for DXC’s Healthcare and Life Sciences Build organisation. He is an expert on health informatics, national clinical audit, clinical governance, healthcare transparency, patient-experience measurement and digital transformation in healthcare. He is now chief executive of Health Innovation Manchester.






  1. […] healthcare organisations also require “adaptive execution”.  We previously referred to the slow innovation cycle in medicine, with an average time from ideation to change in practice at a scale of 17 years, and […]

Speak Your Mind


This site uses Akismet to reduce spam. Learn how your comment data is processed.