It's that time of year when online commentators offer their insights on what we can expect to be big in healthcare during the coming twelve months. Even a casual glance at the latest entries informs us that Big Data is something that health and medical device and technology companies should ignore at their peril.
Amongst other uses, we are told that Big Data will enable innovators to “uncover unmet needs”.
To assess the validity of this assertion, let’s consider three types of healthcare unmet need:
- Therapeutic unmet need that could be addressed by pharmaceuticals or therapeutic devices, e.g., improve insulin sensitivity in a diabetic patient
- Clinical unmet need of healthcare practitioners and providers, e.g., reduce the number of hypoglycemic episodes experienced by a diabetic patient
- User unmet need or need in use of practitioners and patients, e.g., measure personal blood glucose levels in a low light environment
The target set of therapeutic needs for a medical condition is typically relatively well understood. We know what therapies should be provided for patients; Big Data has the potential to tell us, at scale, where they are failing, or what the unmet therapeutic needs are.
It is the case of unmet clinical and user needs that raises greater questions of the potential of Big Data. It is also the case that highlights a misuse or, at the very least, a misinterpretation of the language of unmet needs, and even of innovation itself.
Big Data insight is not unmet need
Big Data advocates refer to the reams of insight it can reveal. What we need to understand, however, is that whilst unmet need represents insight, insight is not necessarily unmet need.
Certainly, Big Data can provide us with plenty of insight in respect of the users and usage of data-enabling medical devices, technologies, apps or software, such as:
- which practitioners or patients use or are monitored by the technology
- their past behaviour, actions, responses and consequences of their use
- their likely behaviour and potential actions
- their physical and emotional status
- their spatial or temporal context of use
- the individual and aggregate performance or outcomes arising from use
- user feedback
- the costs and resource consumption related to current and potential behaviour, practice and use cases, and
- even (heaven forbid) failures in use
All of this is useful information, however, it does not tell us about unmet need, or insight that can reveal hidden opportunities for new product innovation altogether.
This is because Big Data is based on current and historical device and technology selection decisions and behaviors that, necessarily, are undertaken within the confines of the devices, software platforms and technologies that are available or being used at the time. Insight on what somebody has done or is doing now is not, however, necessarily insight on what they are striving to do; that is, what choices they would make without limitation set by existing devices and technology platforms.
By way of illustration, a diabetic will select a blood glucose meter based on how well they believe it will satisfy their own set of needs when trying to control their blood glucose levels. An orthopaedic surgeon will select or influence the procurement of a total hip replacement system based on how well it enables them to achieve a defined clinical goal and on preferences when interacting with or using the system. Both select and use the best options available to them at the time. Whilst they may only select the best available, the best isn’t necessarily perfect – it could even be “the best of a bad bunch”. Knowing what someone has chosen and how they use it doesn’t inform us about what they would choose or how they would use it if they were unlimited in choice.
In other words, Big Data insight does not elicit gaps in understanding in what patients and practitioners can achieve, with the devices and technologies currently available, and what they want or would like to achieve, i.e., their unmet needs.
A more direct approach to discovering unmet need is through mining of first-hand user feedback from social media, online forums and internal customer service sources. At present, the cited limitations of this approach are of a technical nature: “it’s extremely difficult (or not practically efficient) for computers to understand and process natural language, automate sentiment analysis, or determine ambiguous context”. It could be argued, however, that human behaviour will ultimately limit its potential. Users typically articulate what they want from a product in the form of solutions; their expectations are influenced by their experiences of what has existed in the past. The result is incremental innovation at best and, at worst, products that simply fail in the market place precisely because they delivered what users said they wanted.
Discovering unmet needs
To deliver breakthrough innovation and develop new products that address unmet clinical and user needs, companies must address the following questions:
- What clinical, therapeutic or personal / professional goals are practitioners, health care providers and patients actually trying to achieve?
- What are the actual activities they wish to do or are trying to do to achieve these goals?
- How are they making decisions to achieve these goals?
- What capabilities and resources do practitioners and patients have - or do not have - to perform these activities, whether actual or perceived?
- And, importantly, what outcomes or needs on current activities can they not yet achieve or realise, and why?
Only by asking these questions explicitly will such information be garnered and novel insights and opportunities revealed to inform concept generation, selection and the development of innovative health and medical devices and technologies. For that, we need to rely on human approaches to define scope, observe, probe, gather and interpret insight on patient, practitioner and health provider practices, goals and behaviour. Whether, in time, such insight could be truly assimilated into an actionable form without the empathy and reasoning afforded by human input is another consideration for another day.
Big Data has the potential to help uncover unmet therapeutic needs. It has power to reveal new insight into the behaviour of practitioners and patients when using existing medical devices, software and health technologies; insight that can also be used to inform the design and delivery of better user experiences from analysis of the behavioural patterns of all users (such as patient and practitioner alerts or event notifications). However, Big Data is a tool that must be utilized correctly. Until we ask the right questions of it, can we really expect it to provide us with the actionable market insight, in the form of unmet clinical and user needs, that medical device and technology companies require to drive growth through product innovation?
Finally, while we have described the limitations of Big Data in the context of health markets and users, we suggest that the same limitations apply in other markets too.