Here we talk with Dr Rose Penfold, shortly after completing her PhD research in May 2026. Her thesis can be found in the Edinburgh Research Archive: Rose’s thesis
1. What was your PhD research about?
I was looking at whether delirium during an emergency hospital admission is associated with a subsequent dementia diagnosis, and importantly, whether this relationship differs based on the people's existing long-term conditions.
Delirium is a sudden change in brain function – usually triggered by an acute illness – that results in confusion and difficulty with attention. It affects around one-in-four older people admitted to hospital in an emergency. However, we still do not fully understand whether delirium is primarily a complication of acute illness or to what extent it is an indicator of future dementia and brain health.
2. Why were electronic health records necessary for your research?
Firstly, I needed to study a very large population of older adults. Delirium is common, but only a small proportion of people will go on to receive a dementia diagnosis, so a large dataset was essential.
Secondly, these records provided detailed information about people’s health before the hospital admission – especially through GP records – and I was also able to follow people for several years after they left hospital.
Thirdly, some groups of people can be excluded from traditional research – such as those who are very frail, who are unable to consent, or they lack capacity during an episode of sudden illness. Using electronic health records allowed their experience to be represented in my research.
3. Why did you choose to work with DataLoch?
DataLoch brings together multiple health data sources and I wanted to use real data for my research and represent people's genuine experiences of health care and the health care system.
As an NHS Lothian geriatrician, I know how delirium assessment is undertaken locally and – after speaking with the DataLoch team – I knew that key information would be available in DataLoch data.
Importantly for my study this included GP data linked to hospital records so I could understand people’s baseline health, as well as what happened years after hospital admission.
4. How was your experience of working with DataLoch?
There are two aspects to this: working with routine health records and working with DataLoch.
Although working with health data can be challenging, my experience was generally very positive. My key variables were made available within the Trusted Research Environment workspace, including long-term follow-up data, which cannot be done in many traditional research studies. As a practising doctor, I have an advantage in being familiar with how data are captured in hospitals, so I have insights into potential limitations, such as that dementia (in this case) may not be consistently recorded. For any research team seeking access to health data, having clinical experience within the team is vital to better understand the data.
My experience of working with DataLoch was also generally positive. The application process is detailed, and the analysis platform is very secure, which is reassuring for patients and the public. However, this meant it took a while for data to become available, and some clarification steps were needed too: for example, around the format that date of birth would be provided in to preserve confidentiality.
There’s also a robust process for checking any results before they can be removed from the Trusted Research Environment: researchers should be aware that this results-release process can take a little time (so try not to do last-minute requests before a conference presentation)!
5. If you had to start over, would you do anything differently?
I would possibly spend a bit more time understanding the structure and characteristics of the whole dataset before analysing. At first, it's very tempting just to start answering your research question, particularly when you are excited about the project! Later, I came across some potentially avoidable issues such as understanding data gaps (missingness) and how some diagnoses are recorded.
Speaking with NHS Lothian clinical coders was particularly valuable and gave me a much better understanding of how clinical diagnoses are recorded through the health system. I would involve people with this type of expertise earlier in future projects.
6. After your PhD – what's next?
I want to develop a broader research programme around acute illness, delirium, longer-term brain health and dementia.
Linked health records will remain central to this work, potentially alongside other types of data, to help us understand the mechanisms connecting delirium with subsequent dementia. For example, are particular types of dementia more commonly diagnosed after delirium, and what might this tell us about the underlying processes?
Ultimately, I hope this research can improve how we identify, support and follow up people who experience delirium during a hospital admission.
7. Is there anything else you would like to say?
I’m very happy to be contacted by those interested in a similar research area, as well as those considering using DataLoch who would like a researcher perspective.
My final thought is for clinical colleagues. Research using health records depends on how we record things in hospital and other healthcare settings. My research experience has made me think about high-quality data recording and capture. The information recorded today may help answer important research questions in the future, so high-quality data capture is an important part of both good clinical care and good research.
For more about Rose’s research interests and publications so far, visit:
Rose Penfold research profile
Case Studies
