The term evidence-based medicine was coined more than 30 years ago. The central idea was that clinical practice guidelines should be based on available evidence, rather than expert opinion or reigning standard-of-care. Over the past three decades, statistics, engineering and computer science have advanced tremendously. The result is that both decisions for individual patients and public health policies can now be informed by insights obtained from ever-more nuanced and complicated algorithms and models.

While increased analytical power may have given us the ability to get more accurate answers to harder questions, it hasn’t necessarily made it easier for medical professionals at all levels to convince their audiences of patients, funders and politicians of the best way forward. In fact, the use of black box artificial intelligence may leave both the decision-maker and people affected by their decisions rather disempowered and confused.

The rapidly growing movement of explainable AI will certainly help in restoring confidence in the output of advanced health analytics. But merely understanding the algorithm isn’t enough. For data-driven decisions to carry broad and deep support from all relevant stakeholders, these stakeholders need to buy in to the decision. This is where data storytelling can make the difference.

Storytelling is at least 50,000 years old. Stories are an essential part of what it means to be human. They help us make sense of the world by transforming the intricate details of new information into a recognisable, relatable pattern: the story arc. Invariably, after setting the scene and introducing the main characters, every story revolves around a point of conflict, a sense of discomfort, uncertainty, or fear that gets resolved. Not every story ends well, but most stories end with decisions being made, actions taken and the initial unrest making way for clarity and an emotional release (catharsis).

A data story is no different. Initially, there is confusion or a dilemma about what medical decision is in the best interest of a patient population. Through creativity, ingenuity, (computing) power and perseverance, the conflict or fear is resolved and the way forward is clear. 

Of course, data plays a central role in the story. It is the path of breadcrumbs leading us out of the dark forest. For the audience to listen with empathy, it is essential that the data gets the attention it deserves. From the outset it should be made clear how, when, where and why the data came about. Next, graphs and tables can communicate what the data looks like. Most importantly, after the data has been wrangled and analysed, it’s the storyteller’s job to explain clearly and succinctly how this new information aligns with or contradicts what we believed or wondered about before we had the data. In this way, what we should decide or do, in light of the data, becomes logical, sensible, obvious.

A data story that is told well, engages the brain of the listener in multiple areas simultaneously for an immersive experience with maximal impact: information isn’t just transferred as a matter of fact, but it is packaged in emotion, language, logic and mathematics, and visual cues. Which medium, tone of voice, and format to choose for “storing” and sharing the story depends primarily on the storyteller’s motivation, the audience and the complexity and status of the story – completed or still unfolding. Data stories can be told in a static report with an executive summary, a short article like the one you are reading now, a peer reviewed paper, an interactive dashboard, or an explainer video. Often it’s not one but many or most of these, because the same story needs to be told to many different audiences, for different reasons.

Data storytelling is both an evolving branch of modern data science and the rediscovery of an ancient form of art. When done well, it is cathartic for the audience and a touch of magic that transforms medical professionals into empowered, effective decision makers. At Wimmy, we have developed an analytics portal (screenshot above) to help our clients share the data stories that we co-create in a format best suited to their audience: a short summary aimed at executives, the full, detailed report, an interactive dashboard for further exploration beyond the curated story, and additional, downloadable resources such as slide decks and pdf files. We can’t wait to see the impact this new platform will have on the health and wellbeing of populations around the world through more effective medical decision-making with data stories.