When I first started as a dementia researcher (long ago it feels like …) , I wasn’t much aware of existing datasets one could access. Being trained up in neuropsychology, when I thought of data, I thought of administering memory, problem solving, or attention tests, such as the MoCA, or the Doors and People.
But what you quickly realise, especially in participants who live with dementia (because everyone needs breaks in between), is that data collection can take a very long time. So how do you get up that power and amp up your sample size?
Fast forward a few years, and I am the biggest fan of utilising existing routine and cohort data. Obviously, you shouldn’t just rely on these. You should also collect primary data, be that quantitative or qualitative, to really make sense of the topic you are investigating.
The first datasets I ever accessed was the National Alzheimer’s Coordinating Centre (or NACC in short) database [1]. This collects data from 34 sites across the US, and collects data over several waves (it’s still ongoing). For example, I looked at the cognitive underpinnings of finance management in people with Alzheimer’s disease, behavioural variant fronto-temporal dementia, and Lewy Body dementia [2]. Also, recently, I’ve looked at longitudinal variations in everyday functioning between Alzheimer’s disease dementia and behavioural variant fronto-temporal dementia [3]. Another analysis has looked at medication usage, so watch this space!
One thing to bear in mind when accessing cohort data sets is the time it takes to get access (if you get access!), and then the slight confusion at first when you get face to face with the datasets. They can be pretty confusing, so stick to the data manual that accompanies every cohort database. You will most likely have to record certain variables as well, and never underestimate the level of missing data for some variables.
Another database I am currently using is the SAIL databank [4] from Wales. The SAIL is a database of various datasets, and you can apply to get access to specific ones. We did have to pay a fee to access this, and it took a little while to get access. Once you get access, you log yourself in to the remote server in Wales – at first I thought I had a virus because my desktop screen was gone!
Together with collaborators from Edinburgh, Swansea, and Liverpool, we are looking at socio-economic variations in access to care homes in dementia, with a few analyses going on at the moment. For those of you in Liverpool, I’ll be sharing the first findings in a seminar on the 28th of April, so let me know if you are around and fancy a data talk! What’s different about the SAIL database compared to the NACC is that in SAIL, you yourself have to then merge various datasets together by an individual ID code (or ALF).
I’m currently also in the process of applying for Australian routine data on home care services in older people, but luckily I’ve got the help of an Australian collaborator to access the data! Stay tuned!
Ed:- You will find a list of datasets we know to be available here in our essential tools [5], if you know of others, please drop us a line and we will update our listings.
Author
Dr Clarissa Giebel [6] is a Postdoctoral Research Associate at the University of Liverpool and NIHR CLAHRC North West Coast. She has been working in dementia care research for over 7 years focusing her research on on helping people with dementia live at home independently for longer.
You can follow Clarissa on Twitter Follow @ClarissaGiebel [7]