Rising to the challenge of data ethics
We talk to Frankl advisor Dr Sue Fletcher-Watson about ethics and open science in autism research
A few weeks back I had the chance to catch up via Skype with Frankl advisor Dr Sue Fletcher-Watson.
Sue is a Chancellor’s Fellow at the Centre for Clinical Brain Sciences and the Patrick Wild Centre at The University of Edinburgh. Her research focuses on child development, how kids learn, and how various interventions can help babies born prematurely and kids diagnosed with autism.
Sue is also an advocate for including autistic people in the research process as consultants or mentors to non-autistic researchers. As she noted in a recent interview for Spectrum magazine:
All of the measures and the tools we build are based on a certain set of norms and expectations. I think it’s just good science to identify those assumptions and test them, and that’s something that you can do with input from the [autism] community that you can’t really do on your own.
I spoke to Sue about the needs of the autistic community when it comes to research, and what scientists such as herself are looking for from a platform like Frankl. Here’s what she thought.
#1 iPads can improve cognitive assessment but it’s important to keep the assessments engaging
Sue is particularly interested in how iPad apps can be used to help autistic kids. She’s reviewed dozens of apps designed for autistic people and their families, and she’s published a number of academic papers on the topic.
So I asked her what she thought about the use of iPads for cognitive assessment. There are, says Sue, “a lot of benefits” to automating data collection in this way. However, we shouldn’t think that kids can complete the tests on their own. At least for typically developing kids, there’s an audience effect, where they try harder because someone is watching. Sue believes that the trick with iPads will be to “gamify” the test so that kids find it intrinsically interesting and rewarding.
#2 We need to make it easier to share data
One of the main barriers to data sharing is the time it takes to prepare the data for sharing:
“The thing that stops me share my data is making it tidy. I don’t know if the labels are completely transparent to someone else. Are they going to understand what a particular variable is?”
Data collected via Frankl apps will have these kinds of annotations built in. And because the data will have a consistent organisation it’s easy to integrate it with other data. As Sue put it, “It’s not just sitting in the back of someone’s filing cabinet in someone’s office or on someone’s desk computer.” That will be a major step forward!
#3 We need to make it easier for researchers to share methods
As well as data sharing, Frankl also provides a platform for researchers to share their experimental protocols.
Sue notes that this happens informally already but on an ad hoc basis.
“I can contact an individual researcher and say, for example, can I have your stimuli because I want to replicate your experiment? And I would offer them something in return like putting their name on a paper. But I’d have to set up those arrangements individually with different people.
The Frankl platform simplifies this exchange by providing tokens as a payment or reward for re-using someone’s material.
“It allows you to do that in a much more flexible way across a community of users.”
#4 It’s important to consult the community on data sharing
We also talked about the ethics of data sharing. Sue is part of a neurodevelopmental disorders research network in the UK which is attempting to align research protocols to facilitate data sharing and pooling across research groups. In setting up this network, Sue asked participants and their families about their views on data sharing and was heartened by their positive response.
“First of all, people in the community are very positive about data sharing. Having spent an hour and a half in the lab doing assessments they want that information to be used more widely.”
But the people Sue spoke to were also concerned about data being repurposed for things other than what the consent has been given for. For example, lots of autistic people are against research focused on preventing or curing autism.
“There needs to be ethical policy for the data that extends beyond anonymisation and good science,” says Sue. “You need to have a community advisor on your team. You need to think, ‘How do we communicate to people? How do we explain it accurately but also understandably?’”
That’s the challenge for Frankl as we integrate consent features for data sharing into Frankl apps. We need ethical policies that are well thought-through, inclusive and backed by community support. And we need to make sure all app participants have a clear understanding of the potential research applications of any data sharing they consent to. Challenge accepted Sue!
If you’d like to know more about Frankl, you can read our whitepaper, check out our website, and follow us on Facebook and Twitter. You can also chat directly with us via our Telegram channel.
