DSxHE received one of the first opportunities to participate in a conference from the Royal Statistical Society (RSS). We held a highly interactive, hybrid workshop on 14 September, with one of the DSxHE theme leads dialling in from South Africa, and the rest of us heading to Aberdeen, Scotland for the conference.
Despite stiff competition from concurrent sessions and an early morning start, the workshop room was jam-packed. To warm up, each of us in the room introduced ourselves and identified the inequality that matters to us most.
The workshop set out to discuss why should data science be for health equity (in short, DS x HE), rather than just for its own sake, and each of the 4 theme leads took turns to discuss:
An introduction to their theme
What are the current benefits of data science?
What are the potential risks of “ignoring” health equity in data science?
Reasons to be optimistic about improving health equity
In addition to these discussions, we received great advice from fellow attendees that led to further topics being discussed, triggered by two of the polls we ran:
Poll 1: What questions surface when you think of 'Data Science for Health Equity'?
Poll 1 uncovered a huge breadth of topics, across data, modelling, policy, health system, societal and research considerations. Some of the key discussions following from Poll 1 include:
Quantitative bias analysis as a new approach for mitigating inequity
The point on ‘design optimisation’ relates whose and what data should be collected in order to reduce disparities
Whether we need ‘more’ data or better integrated data, as a lot of data is already out there, but it’s difficult to draw any conclusions without data linkage
The point on ‘removing blame’ point was around not putting the onus on individuals to improve their health prospects (‘why don’t you just eat more healthily?’), which links to the problem of free will.
Poll 2: To what extent do you consider equity in your research?
It was reassuring to note that many had considered these factors, even if we didn’t necessarily account for equity in these factors. Some of the key discussions following from Poll 2 centred around:
Drivers of whether a factor is accounted for or not, e.g. regulation on these factors and the relative impact of these factors on the outcome of interest
How to account for equity in practice - this can be done, for example, via adjusting for these factors in statistical models.
With all these engaging discussions, the 1 1/2 hours came to an end very quickly. we closed the workshop with a call to action, and were glad to see that several workshop attendees have since the DSxHE community.
Since then, the 4 themes have been joined by a 5th theme, Standards for Health Data Equity, and all 5 themes officially launched on 28 September. We look forward to furthering the discussions that started at this workshop, either at theme events or wider DSxHE activities.
RSS DSxHE workshop slide deck: RSS-DSxHE - Google Slides