When I first started my job (a long time ago it seems!) and then my PhD (ditto…), I collected new data – so, primary data. I interviewed people living with dementia, carers, and health and social care professionals, and ran neuropsychological test batteries with people living with dementia. That ranged from asking them to remember different types of doors and people, to remember different lengths of digits.
Only later on in my jobs at UEA and now particularly at the University of Liverpool, I have become very much used to use what already exists. Genius, isn’t it, saves a lot of data collection time 😊 Obviously, we can’t just rely on existing data bases and routinely collected data – we equally need to research issues newly. But for many larger scale understandings, such as how often people go to hospital and why, or how many people get a diagnosis of dementia and enter a care home (to be very simplistic), we can utilise routine or cohort data.
Cohorts like the English Longitudinal Study of Ageing (or ELSA) (link: https://www.elsa-project.ac.uk/ [1]) have been specifically designed to collect specific types of information on a regular basis from the same participants. This allows to recognise trends and changes over time. The only downside can be the costs and staffing involved in ensuring that data can be collected on a regular basis, which is why it’s not so easy to set up effective cohorts. Plus, many participants might not wish to continue at some point.
Routine data, such as the Hospital Episode Statistics (or HES), collect data automatically, anonymously, and provide a great insight into, in this case, hospital usage.
Photo by Alexander Sinn on Unsplash
James Watson [2], fellow Dementia Researcher blogger, actually looked into how much routine and cohort datasets are utilised in dementia research, specifically when looking at inequalities in care. In his PhD systematic review (link: https://onlinelibrary.wiley.com/doi/10.1002/gps.5419 [3]), he found numerous different databases which have been used across the globe to explore how people with dementia access anti-dementia medication for example, or enter a care home.
Now imagine you can merge different routine datasets together, to create some sort of mega dataset. Sounds great, doesn’t it? And totally possible. One way of doing that is via the SAIL databank – Secure Anonymised Data Linkage (link: https://saildatabank.com/ [4]). The SAIL databank offers the opportunity to link together routinely collected data in Wales, such as how often people see their GPs to where they live and whether they have a dementia diagnosis.
So, I worked together with colleagues at Swansea and Edinburgh, and using a specific care home data base (when people have entered a care home) and a database created by Edinburgh on dementia diagnosis. We explored whether where people live (rural/urban) and whether their socio-economic background (living in more or less disadvantaged neighbourhoods) is associated with the time to care home entry in dementia, across Wales (link: https://onlinelibrary.wiley.com/doi/10.1002/gps.5446 [5]).
Considering you have data from across Wales at your laptop keyboard fingertips, there are a lot of data you have to manage and sift through and ensure you have the correct data selection rules. For example, by linking all these different databases together, we found that apparently some people received their dementia diagnosis within their first year of life. Clearly this was a glitch in the system in terms of data entry, but that meant we had to check through the data and ensure these cases for example were excluded. Not such a big deal though, as we still included nearly 35,000 people living with dementia who had entered a care home in Wales between 2000 and 2018. That’s another benefit of using (linked) routine data, the power you get from those large numbers.
After a lot of managing the data, running the analysis seems more or less straightforward. In particular, we showed that living in more disadvantaged neighbourhoods and in rural regions is associated with a slower rate of care home entry – so people took longer to enter a care home. People with dementia who lived alone, were older, and more frail, entered a care home faster.
As I said, routine and cohort data can’t answer all of our questions, but it’s definitely a great way to complement your research.
Author
Dr Clarissa Giebel
Dr Clarissa Giebel [6] is a Research Fellow at the University of Liverpool and NIHR ARC 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.